article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
by a lvy - feller diffusion process we mean a markovian process governed by a stable probability density function ( @xmath2 ) evolving in time , @xmath3 whose spatial fourier transform ( the characteristic function ) reads @xmath4 where @xmath5 the two relevant parameters , @xmath6 called the _ index of stability _ , and @xmath7 ( related to the asymmetry ) , improperly referred to as the _ skewness _ , are real numbers subject to the conditions , see @xcite , @xmath8 by introducing the similarity variable @xmath9 we can write @xmath10 where @xmath11 is the stable @xmath2 at @xmath12 the specific form of the characteristic function ( 1.1 ) allows us to recognize @xmath13 as the green function ( fundamental solution ) of the cauchy problem @xmath14\ , , \q u(x,0 ) = \delta ( x)\ , , \q x\in { \rr}\,,\q t>0\ , , \eqno(1.3)\ ] ] where @xmath15 is the pseudo - differential operator with symbol @xmath16 let us recall that a generic pseudo - differential operator @xmath17 , acting with respect to the variable @xmath18 is defined through its fourier representation , namely @xmath19 \ , dx = \hat a(\kappa ) \
, \hat \phi ( \kappa ) \ , ,
$ ] where @xmath20 denotes a sufficiently well - behaved function in @xmath21 and @xmath22 is referred to as symbol of @xmath23 given as @xmath24 with the names of lvy and feller we have intended to honour both paul lvy , @xcite , who first introduced the class of stable distributions , see @xcite , @xcite , @xcite , and william feller @xcite , who first investigated the semigroups generated by a pseudo - differential equation of type ( 1.3 - 4 ) . for @xmath25 and @xmath26
( with @xmath27 ) we recover the standard gaussian and cauchy @xmath28 @xmath29 in physics , the first recognition that the ( symmetric ) lvy distribution could be characterized via a pseudo - differential operator of type ( 1.3 - 4 ) was made explicitly by west & seshadri @xcite .
recently , gorenflo & mainardi @xcite , @xcite , @xcite , have revised feller s original arguments by interpreting ( 1.3 ) as a space - fractional diffusion equation ( of order @xmath30 and `` skewness '' @xmath7 ) and have provided a variety of related random walk models , discrete in space and time , which by properly scaled transition to vanishing space and time steps converge to the corresponding continuous lvy - feller processes . in other words the discrete probability distributions generated by the random walk models have been proved to belong to the domain of attraction of the corresponding stable distribution . here , limiting ourselves to the symmetric case ( @xmath27 ) , we present the main features of the random walk models by gorenflo & mainardi , and we display preliminary results of a few numerical case studies , which can be of some interest in econophysics . the lvy statistics in modelling fluctuations of economical and financial variables , formerly used by mandelbrot in the early sixties ,
is still nowadays adopted with success , see @xcite , @xcite , @xcite .
let @xmath31 be an integer - valued random variable and let the random variables @xmath32 be i.i.d .
(= _ independent identically distributed _ ) , all having their probability distribution common with @xmath31 .
we define a spatial - temporal grid @xmath33 by @xmath34 @xmath35 where @xmath36 and @xmath37 then we consider the sequence of random variables @xmath38 with ( for convenience ) @xmath39 and interpret it as follows .
a particle , sitting in @xmath40 at time @xmath41 finds itself at a later instant @xmath42 in point @xmath43 which is an integer multiple of @xmath44 we recognize the @xmath45 ( for @xmath46 ) as `` transition probabilities '' : @xmath47 is the probability of a particle jumping from a point @xmath48 to a point @xmath49 as time proceeds from @xmath50 to @xmath51 all @xmath47 are non - negative , and their sum equals 1 .
the probability @xmath52 of sojourn of our particle in point @xmath53 at instant @xmath54 obeys the _ transition law _ @xmath55 which has the form of a discrete convolution .
hence , introducing _ generating functions _
@xmath56 we obtain @xmath57^n = [ \tilde p(z)]^n \ , , \ ; n\in \nn_0 \ , .
\eqno(2.4)\ ] ] the power series in ( 2.3 ) and ( 2.4 ) are absolutely and uniformly convergent on @xmath58 and assume the value 1 at @xmath59 putting @xmath60 and observing @xmath61 we recognize @xmath62 and @xmath63 as _ characteristic functions _ of the random variables @xmath64 and @xmath65 respectively .
our aim is to approximate the lvy - feller diffusion process , which is governed by the evolution equation ( 1.3 ) , arbitrarily well . to this purpose
we introduce a strictly monotonic _ scaling relation _ @xmath66 as @xmath67 we will fix @xmath68 and let @xmath69 ( and likewise @xmath70 ) go to zero over such values that always @xmath71 is a positive integer .
then we have the equivalences @xmath72 and @xmath69 depends on @xmath70 , finally on @xmath73 , so that @xmath74 replacing @xmath69 by @xmath75 in ( 2.1 ) we obtain a sequence of random variables @xmath76 with characteristic functions @xmath77^n\,$ ] ( note that now @xmath78 is fixed ) . invoking theorem 3.6.1 of lukacs @xcite , what remains to be shown is that @xmath79 as @xmath80 the characteristic function of the corresponding symmetric lvy - feller process . for this
it suffices that , for fixed @xmath81 @xmath82 \equiv { t\over \sigma ( h)}\ , \log \big [ \tilde{p}(\e^{i\kappa h})\big ] \to -t\ , |\kappa |^\alpha \ , , \q \hbox{\rm{as}}\q h\to 0\ , .
\eqno(2.5)\ ] ] our random walk can be interpreted as a `` difference scheme '' to approximate the evolution equation ( 1.3 ) , if we write ( 2.2 ) in the equivalent form , observing the scaling relation , @xmath83\ , , \eqno(2.6)\ ] ] in fact , whereas the l.h.s .
is the explicit discrete approximation to the first - order time derivative @xmath84 the r.h.s can be considered as a particular discrete approximation to the space pseudo - differential term @xmath85\ , , $ ] provided we mean @xmath86 with @xmath87
from the previous section we have learnt that , in order to construct discrete random walk models which are convergent ( in distribution ) to the ( symmetric ) stable @xmath88s , the clue points are : 1 ) to guess a suitable _ generating function _ @xmath89 whose coefficients of its power series expansion provide the transition probabilities , 2 ) to determine the corresponding _ scaling relation _ @xmath90 which ensures the required convergence . in the classical case of the gaussian distribution ( @xmath25 ) the matter is easily treated if we remember that the corresponding density is the fundamental solution of the standard diffusion equation , which is known to be well approximated via the finite - difference equation @xmath91 in this case , introducing the scaling parameter @xmath92 so @xmath93 the transition probabilities turn out to be @xmath94 subject to the condition @xmath95 thus the generating function is @xmath96 \ , .
\eqno(3.3)\ ] ] the proof of the convergence to the gaussian is simple since one easily finds @xmath97\ , \log [ \widetilde p(\e^{i\kappa h } ] \to -t\kappa^2 $ ] as @xmath98 the scheme ( 3.2 ) means that for approximation of the standard gaussian process the corresponding random walk model exhibits only jumps of one step to the right or one to the left or jumps of width zero . for the stable non gaussian processes we expect to find a non - polynomial generating function with infinitely many transition coefficients which imply the occurrence of arbitrarily large jumps .
it is common practice to refer to the corresponding random walks as to _
lvy flights_. in the following , limiting ourselves to the symmetric cases ( @xmath27 ) , we shall resume the main features of three different random walk models , referred to as ( rw1 ) , ( rw2 ) , ( rw3 ) , of which gorenflo and mainardi have proved the convergence to the corresponding continuous processes . for each model
we give the generating function @xmath99 with the transition probabilities @xmath47 and the scaling relation @xmath100 , referring to the original papers for details . from the analysis of the classical gaussian case we find it natural to first introduce the scaling parameter @xmath101 with @xmath1 but , as we shall show later
, this will not necessarily imply @xmath102 with @xmath103 constant for fixed @xmath104 the model ( rw1 ) has been introduced and discussed in @xcite , starting from the identification of the operator @xmath105 in the framework of fractional calculus , see @xcite , and then applying , in the authors original approach , the grnwald - letnikov discretized scheme . for this model
we need to keep distinct the two cases ( a ) @xmath106 and ( b ) @xmath107 being the case @xmath26 excluded in this treatment .
the generating function is @xmath108\ , , & $ 0<\alpha < 1\,,$ \cr 1-{\ds{\mu\over 2\cos(\alpha\pi/2)}}\ , [ z^{-1}(1-z)^\alpha+ z(1-z^{-1})^\alpha]\ , , & $ 1<\alpha
\le 2\,.$ \cr } \eqno(3.4)\ ] ] for both cases the scaling relation is confirmed to be @xmath109 but the parameter @xmath103 is subject to different restrictions to ensure that @xmath110 in the case ( a ) we have @xmath111 and @xmath112 in the case ( b ) we have @xmath113 and @xmath114\,,$ \cr p_{\pm k}= ( -1)^{k+1}\ , { \ds{\mu\over 2|\cos(\alpha\pi/2)|}}\ , { \ds{\alpha \choose k+1 } } \ , , & for $ \ ;
k=2,3,4,\dots$\cr } \eqno(3.5b)\ ] ] we note that , whereas the classical gaussian random walk ( 3.2 ) is promptly recovered from ( 3.5b ) for @xmath25 , the random walk for the cauchy process ( @xmath26 ) can not be obtained , neither directly nor by a passage to the limit @xmath115 indeed , in both the limits @xmath116 and @xmath117 the permissible range of the scaling factor @xmath103 is vanishing . in numerical practice the consequence will be that if @xmath30 is near 1 the convergence is slow : for good approximation we will need a very small step - time @xmath70 with respect to the step - length @xmath44 differently from ( rw1 ) we shall show how the two other random walk models , ( rw2 ) and ( rw3 ) , that we are going to briefly discuss , exhibit a smooth scaling law in the range @xmath118 , so the case @xmath26 is no longer singular . however , `` there is no free lunch '' .
we have to pay for the good behaviour at @xmath26 with bad behaviour at @xmath25 in a sense to be seen later .
the model ( rw2 ) is obtained by imposing for the transition probabilities an expression in terms of binomial coefficients , as suggested from ( 3.5b ) , which is required to be regular in the limit as @xmath119 namely @xmath120 where @xmath121 subject to the condition @xmath122 is to be determined .
the model has been extensively discussed in @xcite whereas the particular case @xmath26 ( related to the cauchy process ) has been formerly presented in @xcite . to guarantee the convergence for all @xmath30 ( @xmath1 ) to the corresponding continuous process
, the asymptotics as @xmath123 still requires a scaling relation of the kind @xmath124 ( with @xmath103 constant ) , namely @xmath125 we recognize that for @xmath126 the model ( rw2 ) exhibits a smooth scaling law but a discontinuity is present at @xmath25 as can be seen by taking the limit as @xmath127 this allows us to recover for @xmath25 the gaussian model ( 3.2)-(3.3 ) , but it shows that in numerical practice when @xmath0 is near 2 the convergence is expected to be slow .
for @xmath128 we obtain the generating functions and the transition probabilities as follow . putting @xmath129\,$ ] for @xmath130 and introducing the scaling parameter @xmath131 we have @xmath132 , & $ \alpha \ne 1\ , ,
$ \cr 1 - { \ds { \mu \over \pi}}\ , [ ( 1-z^{-1})\,\log ( 1-z)+(1-z)\,\log ( 1-z^{-1 } ) ] , & $ \alpha = 1\ , . $ \cr } \eqno(3.8)\ ] ] for @xmath133 we have @xmath134 and @xmath135 for @xmath136 we have @xmath137 and @xmath138 in both models ( rw1 ) , ( rw2 ) the generating function is expressed in terms of elementary functions and the transition coefficients , for @xmath139 exhibit an asymptotic behaviour as @xmath140 consistent with that of the power - law tails of the stable densities as @xmath141 see @xcite .
indeed we obtain @xmath142 @xmath143 the asymptotic behaviour of the stable densities is the starting point for the random walk model ( rw3 ) in that , following gillis & weiss @xcite , we require that all @xmath47 for @xmath144 are proportional to @xmath145 however , the simplicity of the starting point leads to difficulties for treating this model since the generating function is no longer elementary , a fact that makes the convergence proof a really hard affair .
we need to recall the following special functions @xmath146 respectively known as the riemann _ zeta _ function and the _ polylogarithmic _ function .
this model has been extensively discussed in @xcite ( see also @xcite ) , where the following generating function is derived @xmath147\,.\eqno(3.14)\ ] ] here @xmath148 so @xmath149 is by its power series also defined on the periphery @xmath150 of the unit circle .
then we get a pure power - law random walk with @xmath151 where @xmath121 subject to the condition @xmath152\ , , $ ] is to be determined . to ensure the convergence to the corresponding continuous process the following scaling relation must hold @xmath153 it is interesting to note that in the case @xmath25 the classical random walk model ( 3.2)-(3.3 ) is no longer recovered , since now arbitrarily large jumps occur ( with a probability decaying as @xmath154 ) as in the lvy flights .
nevertheless , through the scaling relation ( 3.16 ) , this random walk converges to the continuous gaussian process .
thus this model gives us the opportunity to verify that a random walk with infinite variance steps may ( slowly ) converge to the gaussian , in agreement with a general theorem on the domain of attraction of the normal law , see @xcite .
in general the random walk models are not only valuable from the conceptual point of view for visualizing what the diffusion means but also for numerical calculations , either as monte carlo simulation of particle paths in a diffusion process or as discrete imitation of the process in form of redistribution ( from one time level to the next ) of clumps of an extensive quantity ( across the spatial grid points ) .
our models can be used in at least three different ways : @xmath155 as _ finite difference schemes _ for approximate calculation of symmetric stable densities ; @xmath156 for producing _
sample paths _ of individual particles performing the random walk ; @xmath157 for producing _ histograms _ of the approximate realization of the densities @xmath158 by simulating many individual paths with the same number of time steps and making statistics of the final positions of the particles . for numerical simulations of stable random variables
different algorithms have been provided by a number of specialists , including chambers _
et al _ @xcite , bartels @xcite , mantegna @xcite , janicki & weron @xcite , samorodnitsky & taqqu @xcite .
our present approach for treating lvy statistics has been carried out independently from all the above references but uniquely based on the random walk models presented here , so , as far as we know , our results would be in the great part original .
having preliminarily checked a sufficient level of accuracy for our finite difference schemes with the existing tables of stable densities , here we present some results on the simulation of the sample paths and histograms corresponding to some typical values of the index of stability , namely @xmath159 in practice , in our numerical studies there is required truncation in two ways .
it is impossible to simulate all infinitely many discrete probabilities , so the size of possible jumps must be limited to a maximal possible jump length .
the other truncation is required if a priori one wants a definite region of space to be considered in which the walk takes place .
then , particles leaving this space have been ignored .
our simulations , based on one million of realizations , have been carried out in the interval @xmath160 all the histograms refer to stable densities at @xmath161 for @xmath162 the space interval being reduced to avoid the border effects .
the sample paths are plotted against the time steps , up to 1200 for @xmath26 and up to 2000 for @xmath163 so they refer to different final times , namely @xmath164 respectively .
the transition probabilities have been chosen from our random walk models as follows : @xmath26 from ( rw2 ) , see ( 3.10 ) , with scaling parameter @xmath165 ; @xmath166 from ( rw1 ) , see ( 3.5b ) , with @xmath167 for @xmath168 from the standard model , see ( 3.2 ) , with @xmath169 the cases @xmath26 ( cauchy process ) and @xmath25 ( normal process ) have been considered for a possible comparison with the standard and accurate algorithms existing in the literature , whereas @xmath170 has been chosen in view of possible applications in econophysics where usually the index of stability ranges from 1.4 to 1.7 , see @xcite , @xcite .
for the simulation of markovian processes characterized by symmetric lvy probability densities evolving in time we have presented three different random walk models , discrete in space and time , by giving their respective transition probabilities .
we have indicated how , by use of generating functions , convergence for properly scaled transition to vanishing steps of space and time can be analyzed , and we have hinted at peculiarities occurring at the particular stability indices @xmath26 ( the cauchy process ) and @xmath25 ( the gauss process ) .
we have displayed preliminary results of a few numerical case studies concerning sample paths and histograms to check the efficiency of our algorithms . from the sample paths one can recognize the `` wild '' character of the lvy flights with respect to the `` tame '' character of the brownian motion .
we expect that our arguments can be relevant in different fields of physics including the emerging one of econophysics , where stable distributions are becoming more common . in statistical physics
the stable distributions play a key role in the ( wonderful ) world of random walks constructed by the late montroll and continued through his school , see @xcite , @xcite , @xcite .
also tsallis and his associates have recognized the key role of stable distributions with respect to a generalized theory of thermostatics , see @xcite , and references therein
. here , we have ( only briefly ) pointed out the relation between lvy statistics and space - fractional diffusion equations .
however , stable distributions turn out to be related also with time - fractional diffusion equations , see @xcite . furthermore
, this topic is relevant for fractal phenomena , where differential equations of fractional order are usually adopted to describe their evolution , see @xcite , @xcite , @xcite , and reference herein .
w. feller , on a generalization of marcel riesz potentials and the semi - groups generated by them , _ meddelanden lunds universitets matematiska seminarium _ ( _ comm .
universit de lund _ ) tome suppl .
ddi a m. riesz , lund ( 1952 ) 73 - 81 .
r.n . mantegna and h.e .
stanley , physics investigation of financial markets , in : f. mallamace and h.e .
stanley , eds .
, _ proceedings international school of physics `` e. fermi '' , course cxxxiv _ ( io press , amsterdam , 1997 ) 473489 .
montroll and b.j .
west , on an enriched collection of stochastic processes , in : e.w .
montroll and j. leibowitz , eds . , _ fluctuation phenomena _ ( north - holland , amsterdam , 1979 ) , pp .
montroll and m.f .
shlesinger , on the wonderful world of random walks , in : j. leibowitz and e.w .
montroll , eds .
, _ nonequilibrium phenomena ii : from stochastics to hydrodynamics _ ( north - holland , amsterdam , 1984 ) , pp . 1 - 121 .
j. klafter , m. f. shlesinger and g. zumofen , beyond brownian motion , _ physics today _ * 49 * no 2 ( 1996 ) 3339 .
f. mainardi , p. paradisi and r. gorenflo , probability distributions generated by fractional diffusion equations , in : j. kertesz and i. kondor , eds . ,
_ econophysics : an emerging science _ ( kluwer , dordrecht , 1998 ) in press , pp .
a. carpinteri and f. mainardi , eds . ,
_ fractals and fractional calculus in continuum mechanics _ ( springer verlag , wien , 1997 ) .
zaslavsky , _ physics of chaos in hamiltonian systems _ ( imperial college press and world scientific , london and singapore , 1998 ) .
|
we propose a variety of models of random walk , discrete in space and time , suitable for simulating stable random variables of arbitrary index @xmath0 ( @xmath1 ) , in the symmetric case .
we show that by properly scaled transition to vanishing space and time steps our random walk models converge to the corresponding continuous markovian stochastic processes , that we refer to as lvy - feller diffusion processes . , and random walks , stable probability distributions , diffusion # 11
|
dwarf elliptical ( de ) galaxies ( @xmath9 ) , and the fainter dwarf spheroidal ( dsph ) galaxies ( @xmath10 ) , are the most numerous galaxy type in the universe .
as such , any galaxy evolution / formation scenario must clearly be able to predict and describe the properties of these galaxies . as well as being numerous , dwarfs have low masses ( @xmath11 m@xmath12 ) , and are therefore more easily influenced by their environment than more massive galaxies , with environment likely playing a vital role in their formation and evolution .
one well - known environmental dependence that dwarfs follow is the morphology - density relation defined by giant galaxies @xcite , such that the fraction of early type galaxies increases with increasing environmental density .
it was noted by @xcite that dwarf irregulars were not found in great numbers in the virgo cluster , despite such galaxies being numerous in the local group , providing the first hints that an environmental dependence of dwarf galaxy morphology exists .
the morphology - density relation was subsequently shown to extend to the dwarf galaxy regime in @xcite , with the number fraction of dwarf ellipticals increasing with local projected density .
this relationship hints that the formation of dwarfs may be linked to the environment in which they reside .
however , this environmental dependence of dwarf evolution is complicated by the fact that dwarfs in the cores of rich clusters are not a homogeneous population @xcite , showing variations in their star formation histories .
we previously examined the star formation history of dwarfs in the core of perseus in @xcite using deep keck / lris spectroscopy down to @xmath13 .
these dwarfs exhibit a range of ages and metallicities , with an old , metal poor population with ages @xmath148 gyr , and a young , metal - rich population with ages @xmath15 gyr .
this spread in ages and metallicities suggests multiple formation scenarios are required to explain their origin .
it is hypothesized that the spread in ages and metallicities can be explained if at least some dwarfs are the result of the morphological transformation of infalling galaxies .
several formation scenarios have been proposed to explain the infall origin of such cluster dwarf ellipticals , with a number of environmental processes thought to influence the formation and evolution of these galaxies .
dwarf galaxies are likely susceptible to high speed encounters with other galaxies ( harassment , @xcite ) , ram pressure stripping by the hot ism ( e.g. grebel et al .
2003 ) , or by interactions with the tidal potential of their host cluster , group , or galaxy .
these processes can cease star formation and remove mass from an infalling progenitor galaxy , transforming it from late to early type .
the efficiency of these processes depend on the density of the environment in which the dwarf resides .
however , the degree to which environment is important in shaping these low mass galaxies is largely unknown .
we began to investigate how environment shapes the evolution of dwarf galaxies in @xcite .
we investigated the dsph population in the core of the perseus cluster , using deep _
hubble space telescope ( hst ) _ advanced camera for surveys ( acs ) imaging in the f555w and f814w bands .
this imaging reveals that dwarfs in the core of the perseus cluster are remarkably smooth in appearance , lacking the internal features such as spiral arms and bars seen in observations of brighter dwarf ellipticals ( @xmath16 ) in the virgo and fornax clusters @xcite .
this result suggests these galaxies have not recently been transformed from late - type progenitors .
similarly , @xcite find that dwarfs on circular orbits in the virgo cluster are on average rounder than those on more elongated orbits , with no evidence of disks or residual star formation .
based on these results , they imply that these round des are part of the original cluster population , whereas the flatter des on elongated orbits originate from an infalling population . this study of how environment shapes low mass galaxies was continued in @xcite , where we compared the photometric scaling relations for early - type galaxies ( @xmath17 ) in different environments .
this paper is based on our @xmath1 acs results , as well as data gathered from the literature , comprising of data for a range of environments from the local group to dense clusters .
the results of this paper reveal that scaling relations for dsphs / des are almost independent of environment , however these scaling relations change at @xmath18 .
this suggests that below this luminosity , the low masses , and therefore shallow potential wells , of these galaxies mean different processes likely drive their evolution .
in this paper we continue our study of dwarf galaxies in the cluster environment , extending our study of the perseus cluster to its outer regions . using deep
_ hubble space telescope ( hst ) _ wide field planetary camera 2 ( wfpc2 ) observations in the f555w and f814w bands , we target seven fields in the outer regions of the perseus cluster down to @xmath19 = -12 .
this imaging allows us to compare dwarfs in the core and outskirts of a rich galaxy cluster , to examine the effect of environmental density on the cluster dwarf galaxy population . using this imaging
, we identify eleven new de / dsph systems in the outer regions of perseus , all located at cluster - centric radii @xmath20 kpc .
we also identify in this new imaging ten possible dirr candidates , which we also discuss in this paper .
we examine the structures of our des and compare them to those of des in the cluster core .
we find that the structures of des in perseus become progressively more disturbed with increasing cluster - centric radius , which we interpret as evidence that these dwarfs are undergoing a morphological transition from late to early type galaxies as they infall into the cluster .
however , despite being morphologically disturbed , these objects show no evidence for recent star formation , following the same colour magnitude relation ( cmr ) as those in the cluster core .
this result suggests that these dwarfs have recently transformed from late - to - early type galaxies , with their colours transforming prior to their structures .
we also compare number counts for both the core and outer regions of the cluster , and find that very faint dwarfs ( @xmath21 ) are preferentially found in the outer regions of perseus .
this paper is organized as follows . in section 2
we discuss our observations , including cluster membership in section 2.5 .
the analysis of our data is presented in section 3 , with the morphologies of the des are covered in section 3.1 , the colour - magnitude relation as a function of environment discussed in section 3.4 , along with _ galaxy evolution explorer ( galex ) _ uv observations in section 3.5 .
a comparison of the number counts in both the core and outskirts of the cluster are presented in section 4 , with a discussion of our results in section 5 .
our results are summarized in section 6 . throughout this paper
we refer to both cluster des and cluster dsphs as des .
we examine the dwarf galaxy population of the perseus cluster ( abell 426 ) , one of the nearest rich galaxy clusters , with a redshift @xmath22 kms@xmath23 @xcite , and a distance @xmath24 mpc . despite its proximity
, it has not been studied in as much detail as other clusters such as fornax , virgo or coma , most likely due to its low galactic latitude ( @xmath25 = -13@xmath26 ) . to examine how local environmental density shapes
the cluster dwarf galaxy population involves deep , high resolution imaging sampling the whole cluster , from its dense core to sparse outer regions . using @xmath1 acs and wfpc2 observations spanning 800 kpc of the perseus cluster , we examine the effect of local environment on cluster dwarf ellipticals .
our survey design is shown in fig .
1 , with the acs observations targeting the core of the cluster , and wfpc2 observations focusing on the cluster outskirts .
the observations are explained in more detail below . as part of the original study of dwarfs in the cluster core
, we have obtained high - resolution @xmath1 acs wide field camera imaging in the f555w and f814w bands for 5 fields in the core of the perseus cluster , resulting in a survey area of @xmath1457 arcmin@xmath27 ( program go-10201 ) .
these pointings are in the vicinity of the peculiar galaxy ngc 1275 ( taken to be central galaxy of perseus ) and the giant elliptical ngc 1272 , the two brightest members of the cluster .
exposure times are one orbit each in the f555w and f814w bands .
the reduction of these data are explained in detail in @xcite and @xcite .
the positions on the sky for these five pointings are shown in fig .
[ survey ] . due to the failure of acs in 2007 , we obtained @xmath1 wfpc2 imaging for 7 fields in the outer regions of the perseus cluster in the f555w and f814w bands ( program go-10789 ) .
the total wfpc2 survey area is @xmath1440 arcmin@xmath27 , and extends to a distance of 600 kpc from the centre of the perseus cluster , which is taken to be the galaxy ngc 1275 ( @xmath28 , @xmath29 ) .
exposure times are one orbit per field in the f555w and f814w bands , resulting in a total exposure time of 14 orbits .
these data were obtained in four exposures per field , to remove the effect of cosmic rays , and other defects , which could masquerade as features in the dwarfs .
the wfpc2 images were processed using standard iraf stsdas techniques for the reduction of space telescope data , with a drizzle.scale parameter of 0.5 .
a drizzle.pixfrac scale of 0.9 was selected .
this produced a final pixel scale of 0.05 , corresponding to 18 pc at the distance of the perseus cluster total magnitudes in the f555w for all dwarfs in our acs and wfpc2 imaging are measured out to their petrosian radii @xcite .
the colours of each dwarf are then measured within the central 1 aperture .
we correct these magnitudes for interstellar reddening adopting the colour excess @xmath30 of @xcite .
the magnitudes and colours for the dwarfs in our acs imaging are then converted to the @xmath31 system using the transformations of @xcite , allowing their ( @xmath2 colours to be determined .
the transformations to the @xmath32 and @xmath33 bands for the wfpc2 dwarfs are those of @xcite for the gain @xmath34 setting . to determine if any of the dwarfs in either the core or outer regions of perseus contain recent star - formation , we utilise deep _ galaxy evolution explorer ( galex ) _ data in the far - ultraviolet ( fuv ) and near - ultraviolet ( nuv ) bands down to @xmath35 @xmath1424.5 .
when combined with structural information , star formation can provide an indicator of whether or not a dwarf is in the process of being transformed from late to early - type .
if this is the case , it may retain a low level of star formation until the transformation is complete and all star - forming material is removed .
we detect 3 dwarfs ( cgw40 , cgw45 and cgw46 ; @xcite ) in the nuv , all of which are in the cluster core .
these dwarfs are all non - detections in the fuv .
the uv data are in ab magnitudes .
we correct these magnitudes for interstellar reddening using the colour excess @xmath30 of @xcite .
the @xmath35 magnitudes and @xmath36 colours of these dwarfs are presented in table 1 .
unfortunately the @xmath37 observations are not deep enough to detect the dwarfs in our wfpc2 observations .
however , had any of the more luminous dwarfs in this region supported intense star formation , they would have shown up in the @xmath37 imaging with anomalously blue colours .
the three dwarfs that we did detect in the perseus cluster core can be compared to those of dwarfs in other clusters , to investigate if the @xmath36 colour - magnitude relation is consistent between clusters ( @xmath38 3.5 ) , as is observed for optical colours ( e.g. @xcite ) .
these dwarfs have nuv colours of normal inactive des .
.dwarf galaxies in our acs imaging with @xmath37 observations . the @xmath35 magnitudes and @xmath36 colours
are extinction corrected .
[ cols="<,^,^,^,^ " , ] the most reliable method for confirming cluster memberships is by measuring radial velocities , though this is not easily done for low surface brightness objects such as cluster dwarf ellipticals , requiring long integration times .
the high resolution and depth of our hst imaging allows for the separation of cluster dwarfs and background galaxies based on sizes , structures and morphologies ( e.g. @xcite , @xcite ) .
we also have confirmed cluster membership for six of the dwarfs in our acs imaging based on spectroscopic observations @xcite .
these spectroscopically confirmed cluster members all have similar morphologies , with smooth , extended light envelopes , little or no internal structure and a low surface brightness .
all dwarfs in this study are well resolved , thus if a galaxy has the same morphology as a spectroscopically confirmed cluster member , we assume it is also a cluster member .
dwarf ellipticals are diffuse and symmetric in appearance , whereas background galaxies often contain spiral structures or tidal features .
dwarf ellipticals can also be separated from background galaxies as they have a low surface brightness , whereas background galaxies do not .
this method was utilized in @xcite , resulting in a sample of 29 dwarf ellipticals in the core .
we use the high resolution of our wfpc2 imaging to find des in a similar way to those in the cluster core @xcite , identifying galaxies with similar morphologies to those with confirmed cluster membership in our acs imaging .
this yields a sample of 11 nucleated dwarf ellipticals ( de , n ) in the cluster outskirts .
the positions of these dwarfs within the perseus cluster are shown in fig .
[ survey ] , with a selection of images of these dwarfs presented in fig .
[ montage ] . a detailed assessment of these dwarfs , along with those in the cluster core , will be presented in penny et al .
( in prep ) .
the presence of nuclei in all the galaxies in our cluster outskirts sample is consistent with our sample in the core , where 24 out of 29 ( @xmath39 ) dwarfs identified in the cluster core are nucleated down to @xmath40 .
previous surveys have suggested that below @xmath18 , nucleated dwarf ellipticals ( de , ns ) are rare , and more concentrated towards the cluster centre @xcite ; however more recent studies have disputed this ( e.g. @xcite ) .
the high fraction of faint nucleated dwarfs we have identified is comparable with recent space and ground based surveys of des ( @xcite ) .
there is no evidence for this clustering in our de sample , comparable to the findings of @xcite , who speculate that previous claims that de , ns are preferentially found in the cluster center may be due to selection effects , with only deep ground - based and space - based imaging able to resolve the nuclei of faint des @xcite .
, and the lower right - hand number is the @xmath41 colour . dwarfs 2 , 3 and 7 are very red , and we can not rule out that these are background objects.,width=309 ] we further constrain cluster membership by quantifying the morphologies of all the dwarfs we identify by eye .
cluster dwarfs can be separated from background ellipticals based on quantitative measures of their structures , such as though their concentration indices @xcite .
this index describes the ratio of the radii containing 80% and 20% of the galaxy s light , with the higher the value of c , the more centrally concentrated the light distribution is .
background ellipticals will have @xmath42 , whereas low surface brightness objects such as dwarfs will have c @xmath142.5 . all dwarfs we identify as des have @xmath43 , showing they are low concentration galaxies rather than background ellipticals .
however , one dwarf in the cluster core sample is found to have an unusually high concentration index ( c@xmath44 ) , due to the presence of an unusually large nucleus , leading to ambiguity in its classification . this galaxy could therefore be a compact dwarf or a background galaxy .
we therefore exclude this dwarf from our morphological analysis of the de population of the cluster core and outskirts , though its colour places it on the red sequence of the perseus cluster colour magnitude relation .
size was also used to place constraints on cluster membership .
dwarf ellipticals have typical sizes @xmath45 kpc @xmath46 kpc .
background spiral and elliptical galaxies at @xmath47 are similar in angular size to dwarfs in perseus ( 3.0 ) , though the high resolution of @xmath1 imaging allows us to easily reject background spirals as cluster members . likewise , background giant ellipticals will not have such a low surface brightness , will be more concentrated , and will be generally redder than cluster des . )
the candidate dirrs do not exhibit the symmetric structures seen for cluster des .
, width=317 ] we identify in our wfpc2 imaging a sample of dwarf sized galaxies ( 0.5 kpc @xmath48 2 kpc ) with irregular morphologies that we can not rule out as cluster members based on their colours , which are likely dirrs .
two examples are shown in fig .
[ dirrs ] .
however , their cluster membership will need to be confirmed via spectroscopy as these objects can in some cases appear similar to background spirals .
we find no dirrs in the acs imaging of the cluster core .
it is hypothesized that infalling lower mass late - type galaxies , such as dwarf irregulars ( dirrs ) , are the progenitors of des . through tidal interactions
as they infall into the cluster , these late - type galaxies lose their gas and cease star formation , and morphologically transform into early - type galaxies @xcite .
we identify dirrs in our wfpc2 imaging based on their morphologies and colours . while we identify the dirr candidates in fig .
[ survey ] , further results concerning the nature of these galaxies will be presented in penny et al .
( in preparation ) once we can better assess the probability of their cluster membership .
it is hypothesized that some dwarf ellipticals are formed via tidal processes such as ram pressure stripping and the harassment of galaxies infalling into the cluster @xcite .
these transformations are thought to take place in environments with low velocity dispersions , such as galaxy groups which are later accreted by the cluster .
the outskirts of a rich cluster represent a region where galaxies that have been recently accreted into the cluster are likely to reside .
if the transformation from a late type galaxy to an early type dwarf elliptical does indeed take place in groups that are later accreted by the cluster , we might expect to find in the cluster outskirts a population of dwarfs that retain evidence of this transformation , such as remnant bars or tidal features ( e.g. @xcite ) .
previous studies have found example objects , termed transition dwarfs ( e.g. @xcite ) , which are galaxies that show characteristics of both early - type dwarfs ( des & dsphs ) and late - types ( dirrs ) .
it is likely that dirrs are the precursors of des , with des being dirrs that have lost their gas or converted it into stars long ago @xcite .
transition objects have little or no star formation and low luminosities ( typical of des or dsphs ) , as well as containing hi gas as is typical for dirrs .
for example , @xcite find dwarfs in the centaurus a group with morphologies intermediate between early and late type that contain hi reservoirs .
similarly , @xcite find that dwarfs in the virgo cluster with hi gas content are more similar to local group transition dwarfs than regular cluster des .
transition dwarfs likely represent a class of dwarfs undergoing the evolutionary transformation between late and early type galaxies , and may provide important clues as to the origin of dwarf ellipticals / spheroidals .
these galaxies will have morphologies intermediate between dirrs / des @xcite , and we therefore examine the structures of the dwarfs identified in our wfpc2 imaging to identify such objects .
morphological evidence for an infall origin will be erased over time as any internal sub - structure is removed ( see @xmath38 5 ) . for infalling dwarfs to retain evidence of the transformation from late - type to early - type , they must have fallen into the cluster in recent times . by looking at the distances of our dwarfs from the cluster center
, we can get an idea of the minimum time a dwarf must have been in the cluster for .
a dwarf located at the edge of the cluster will either have just started to infall into the cluster , or be on an eccentric or highly radial orbit .
this is because galaxies on circular orbits are likely part of the original cluster population , whereas those on radial orbits likely originate from an infalling population @xcite .
combining this information with the morphologies of the dwarfs allows us to see if any of the dwarfs are being morphologically transformed as they infall into the cluster .
for example , if these dwarfs are infalling into the cluster for the first time , and have disturbed morphologies , this would suggest they are in the process of transforming from late - type to early - type galaxies , with the cluster environment driving this transformation .
the dwarfs in our wfpc2 imaging are all located at projected distances @xmath49 kpc from the cluster centred galaxy ngc 1275 , with the most distant dwarfs at distances @xmath14600 kpc from the cluster centre .
if a morphologically disturbed dwarf is at a large distance from the cluster centre , it is either infalling into the cluster core for the first time , or else it is on a highly radial orbit . the typical crossing time for a dwarf entering a cluster such as perseus will be @xmath141.5 gyr , therefore a dwarf at a distance of @xmath14600 kpc from the cluster centre is either entering the cluster for the first time , or has been in the cluster for @xmath43 gyr and is on a highly radial orbit , as it must have passed through its orbit pericenter already and returned to the cluster outskirts .
therefore the dwarfs in our wfpc2 imaging are either old objects or have been in the cluster for more than 3 gyr . by quantifying their morphologies , we can see if the structures of the dwarfs in the cluster outskirts sample are consistent with a late - infall scenario .
we quantify the morphologies of our newly identified dwarfs in the outer regions of perseus using the concentration , asymmetry and clumpiness ( cas , @xcite ) parameters .
these parameters are measured using the inverted form of the petrosian radius @xcite for each galaxy .
each dwarf is manually cleaned of background galaxies and foreground stars using the task imedit prior to the calculation of the cas parameters , as these could masquerade as features in the dwarfs , or contaminate the measurements .
we compute the concentration index ( c ) for each galaxy , which is defined as the ratio of the radii containing 80@xmath50 and 20@xmath50 of a galaxy s light .
the higher the value of c , the more concentrated the light of the galaxy is towards the centre , with dwarf ellipticals having an average value of 2.5 @xmath51 0.3 @xcite .
the asymmetry indices ( a ) for the dwarfs in our sample are also computed .
this index measures the deviation of the galaxies light from perfect 180@xmath26 symmetry .
the less symmetric the galaxy , the higher the asymmetry index .
asymmetric light distributions are produced by features such as star formation , galaxy interactions / mergers and dust lanes @xcite , which are not found in dwarf ellipticals in the core of perseus @xcite .
the third index we measure is the clumpiness index ( s ) .
this index describes how patchy the distribution of light within a galaxy is .
the clumpiness ( s ) index is the ratio of the amount of light contained in high frequency structures to the total amount of light in the galaxy . to measure s , the original effective resolution of the galaxy
is reduced by smoothing the galaxy by a filter of width @xmath52 to create a new image .
the original image is then subtracted by the new , lower resolution image to produce a residual map .
the flux of the residual light is then summed , and this sum is divided by the sum of the flux in the original galaxy image to give the galaxy s clumpiness ( s ) . for dwarf ellipticals in the core of perseus , this is found to be near zero . however , if dwarf ellipticals in the outskirts of perseus have only recently been transformed from late to early - type galaxies , then we might expect some internal substructure to remain .
this is because young , blue stars have short life - times ( @xmath53 myr , @xcite ) compared to the time - scale for star clusters and large scale star forming regions to dissolve ( @xmath54 gyr , e.g. @xcite ) .
we give a detailed analysis of this situation in 5 .
this index is very sensitive to signal - to - noise , and we found it necessary to compute this index at half the petrosian radius of each dwarf .
we compare the cas values for dwarfs in the core and outer regions in fig .
dwarf ellipticals in the outskirts of perseus have higher values of both a and s than those in the cluster centre , with dwarfs in the outer regions of the cluster having asymmetry and clumpiness values @xmath3a@xmath4 @xmath5 , and @xmath3s@xmath4 @xmath6 , compared to @xmath3a@xmath4 @xmath7 , @xmath3s@xmath4 @xmath8 for those in the cluster core .
these mean values were calculated using a bootstrap method .
this method selects dwarfs at random for both the inner and outer regions of the cluster with replacement , drawing samples with the same size as the original data sets .
this was repeated 10,000 times for both the inner and outer region data sets , giving bootstrap samples for the inner and outer regions of the cluster .
we then compute a mean for each of the bootstrap resampled data sets , and calculate the dispersions for these means .
the errors quoted are 1@xmath55 from the mean .
the dispersions for the means of these bootstrap samples are 0.024 and 0.010 for the outer and inner asymmetry means respectively , with dispersions for the outer and inner clumpiness bootstrap means of 0.024 and 0.013 respectively .
these results indicate that des in the outer parts of perseus are significantly more disturbed than those in the cluster core .
we also investigate the significance of this result using a monte carlo method .
the monte carlo simulation we preform changes the values of the asymmetry and clumpiness values for dwarfs in the outer regions of the cluster by randomly varying the asymmetry and clumpiness values within the error bars for each galaxy .
this is repeated 10,000 times , with the mean for each simulated outer data set compared to the mean for the inner dwarfs .
what we find is that in none of these 10,000 realisations does the mean asymmetry of the simulated values for the outer dwarf asymmetries become lower than that of the inner dwarf galaxies .
in fact , we have to increase the sizes of the error bars by a factor of four before we see any increase in the fraction of the simulated data sets having an average asymmetry or clumpiness lower than those found for the inner sample . for the clumpiness values ,
the error bars must be increased by a factor of ten .
this implies the increase we see in the asymmetry and clumpiness values is robust to at least 10 @xmath55 significance . while the concentration index c
does not reflect irregularities in a galaxy s light distribution , we compare the results for the two samples for completeness .
we find that c is consistent between the two environments ( c.f fig .
[ cas ] ) , with @xmath3c@xmath4 @xmath56 for dwarfs in the cluster outskirts , compared to @xmath3c@xmath4 @xmath57 for those in the cluster core .
these means were computed using the same method as for the asymmetry and clumpiness values .
therefore , within the error bars , the two populations have the same concentration values . local density decreases with increasing cluster - centric radius , such that the cores of clusters are more densely populated than the cluster outskirts .
galaxy morphology is dependent on local galaxy density , with spiral and irregular galaxies more commonly found in regions of low local density ( the morphology - density relation , @xcite ) .
our hst survey areas are not large enough to accurately determine local galaxy density , so we instead use projected cluster - centric distance as an indicator for local galaxy density . to determine how the morphologies of our dwarfs change with cluster - centric distance , we compute the projected distance of each dwarf from the cluster centre ( taken to be the cluster centred galaxy ngc 1275 ) , and compare these distances to their a and s values
. the values of c are not investigated as a function of environment , as this index provides a useful way of separating cluster des from background ellipticals , but does not give a measure of features due to recent star formation or interactions @xcite .
the results are shown in fig .
[ casr ] .
we fit the relationship between both a and s with r using a weighted least squares fit , using bootstrap samples .
we generate 10,000 bootstrap samples , each of the same size as the original data set .
a weighted least squares fit was then performed to each of the bootstrap samples .
this process is repeated 10,000 times , with a new least - squares fit found for each resampling .
we then find the mean and dispersions of these fits , with these best fit lines shown as red lines on fig .
the errors quoted below are the dispersions of the least squares fits to the bootstrap samples .
the following fits to the data are found : @xmath58 @xmath59 where @xmath60 is the projected distance from the cluster centre .
there is a correlation between both @xmath61 and @xmath62 with projected cluster centric distance @xmath60 , such that the more distant a dwarf is from the cluster centre , the more disturbed its morphology is .
we confirm this result with a two - sided kolmogorov - smirnov test comparing the morphologies of dwarfs located at @xmath63 kpc with those at @xmath64 kpc from the cluster center .
we compare the cumulative distributions of a and s for the core and outskirts environments .
the likelihood that the values of a for the two environments are drawn from the same population is @xmath65 , and @xmath66 for s. this low significance ( @xmath67 ) shows that dwarfs in the outer parts of the cluster are indeed more disturbed than those in cluster core to a @xmath68 confidence level .
however , not all dwarfs in the cluster outskirts have disturbed structures , with some des smooth like those in the core .
this is to be expected as groups accreted into clusters will contain some smooth dwarfs ( e.g. the local group dsphs ) .
the slopes of the two relationships are similar , and this result suggests that the asymmetries in the dwarfs are the result of star formation rather than dynamical processes .
the asymmetry and clumpiness indices are both good indicators of recent star formation , as star formation produces irregularities in the structure of the dwarf .
the fact that the morphologies of the dwarfs become more disturbed with increasing cluster - centric radius can be explained by environmental effects .
for example , ram pressure stripping varies as @xmath69 , where @xmath70 is the density of the intracluster medium and @xmath71 is the speed of the satellite galaxy relative to the icm .
therefore this process is more effective towards the core of the cluster as the density of the intracluster medium increases .
galaxies at smaller clustercentric distances will have their gas removed more quickly than those in the cluster outskirts , so they cease star formation earlier .
as star formation has ceased more recently for des in the cluster outskirts , their morphologies have not yet completely transformed from late to early type , unlike those at smaller clustercentric radii .
alternatively , harassment @xcite can morphologically transform a small disk galaxy into a dwarf elliptical as it falls into the cluster .
high - speed close encounters with massive cluster galaxies cause gravitational shocks that disrupt the disks of small sc - sd galaxies .
the rate of these encounters will increase as the dwarf progenitor galaxy infalls into the cluster , as the density of the cluster ( and therefore number of massive galaxies ) increases . as a result , if a galaxy is transformed from late to early type via harassment , a galaxy that has just entered the cluster for the first time will have a more disturbed morphology than one in the core , as it will have had fewer encounters to destroy its internal substructure .
therefore a harassment formation scenario can also explain the more disturbed morphologies found for dwarfs in the cluster outskirts compared to those in the cluster core .
a well - defined colour - magnitude relation ( cmr ) exists for elliptical galaxies , such that the brightest galaxies are the reddest in colour , as first noted by @xcite for ellipticals in the virgo and coma clusters .
this correlation is generally regarded as a relationship between a galaxy s mass ( as traced by luminosity ) , and metallicity ( traced by colour , @xcite ) .
dwarf galaxies follow this same colour magnitude relation ( e.g. @xcite ) . using the newly identified dwarfs in the outer regions of perseus
, we investigate whether dwarfs in the lower density regions of rich clusters lie on this same colour - magnitude relation .
given that the morphologies of the dwarfs in the outskirts of perseus suggest they have been recently transformed from late to early type , their colours may also reflect this , in that a recently transformed galaxy might be expected to be bluer in colour due to having a younger stellar population .
our colour - magnitude relation is shown in fig .
we compare the cmr for the core and outer region dwarf populations , to examine any environmental dependence on the cmr .
the cmrs for the two regions are fitted using a least squares method with the points weighted according to the error on their colour .
this provides the following best fits to the data : core : @xmath72 and for the outskirts : @xmath73 galaxies redder than @xmath74 are excluded from this fit , due to the ambiguity of their cluster membership ( circled points on the cmr , dwarfs 2,3 and 7 ) .
however , dwarfs this red have been found by @xcite for fainter objects in the fornax cluster , which are more distant from the colour - magnitude relation than our candidate dwarfs , so they are included in our colour magnitude diagram . within the error bars , the colour magnitude relations for the two environments are identical .
this suggests that given the disturbed morphologies of the dwarf galaxy population the outskirts of perseus , it appears that the colours of dwarfs transform prior to their morphologies .
if this were not the case , we would expect the dwarf population in the outer regions to be significantly bluer than those in the core , with their colours matching better their disturbed morphologies . since dwarf ellipticals in both the core and outer regions of the perseus cluster follow the same colour magnitude relation ( fig .
[ cmd ] ) , this implies similar star formation histories and stellar populations ( and therefore metallicities , @xcite ) of dwarfs in the two environments .
the dwarfs have colours consistent with passive stellar populations that stopped forming stars long enough ago for there to be no age information left @xcite
. recent work by @xcite also suggests that all early - type galaxies follow a single cmr , regardless of environment , and we investigate if this relation extends to the outer regions of rich clusters .
we perform a least squares fit to the colour magnitude relation for the cluster dwarf elliptical population as a whole ( again excluding des redder than @xmath75 ) , giving @xmath76 the relationship between colour and absolute magnitude found by @xcite for dwarfs in the fornax cluster core is @xmath77 , with a colour - magnitude relation of @xmath78 found for the hydra i cluster @xcite .
a similar relation of @xmath79 is found for the centaurus cluster . within the error bars ,
the three clusters have identical colour - magnitude relations as perseus for their dwarf galaxy populations .
this also matches the colour magnitude relation found for dwarfs in the local group @xcite .
therefore it appears that the universality of the cmr extends to des in the low - density , infall regions of clusters .
we also include for comparison our dirr candidates . like the cluster des , all these galaxies are red in colour , and we can reject any candidates with ( @xmath80 as these are probable background galaxies .
the red colours of these galaxies suggests they are either background galaxies or dirrs that have entered the cluster at recent times , and have been stripped of their gas and ceased star formation .
we extend our study of the cluster colour - magnitude relation to the uv , by examining the @xmath36 colours of our dwarfs using galex data for the perseus cluster .
if a dwarf has been recently transformed from late - to - early type , its stellar population will likely retain evidence of recent or ongoing star formation . during ram presssure stripping ,
the high pressure of the hot intracluster medium ( icm ) can trigger the collapse of molecular clouds in an infalling late - type galaxy , leading to a burst of star formation @xcite .
likewise , tidal perturbations in the dense cluster core can trigger star formation episodes . an indication of whether recent star formation has taken place in our early - type dwarfs can therefore be gained by examining their ( @xmath81 ) colours .
however , a flux in uv does not necessarily identify a young stellar population .
evolved stars , such as old , hot horizontal branch and post asymptotic giant branch stars with high metallicity @xcite can produce a uv excess in early - type galaxies , and it is important to determine if such a population is producing the uv flux in the dwarfs . if a dwarf has undergone a recent burst of star formation , its ( nuv@xmath82v ) colour will reflect this , being bluer than the colour magnitude relation for early type galaxies . )
colour magnitude relation for dwarfs in the perseus cluster ( triangles ) , overplotted on the results of @xcite for des in the virgo cluster .
the perseus dwarfs for which we have @xmath35 data are cgw 40 , cgw 45 and cgw 46 .
the line is the cmr for the virgo cluster des , which we extend to fainter magnitudes.,width=309 ] @xcite investigate the ( @xmath81 ) colour - magnitude relation for dwarf ellipticals in the virgo cluster . by comparing the colour - magnitude relation for the core of perseus and for virgo
, we can see if the ( @xmath81 ) cmr is consistent between clusters . to a limiting magnitude of m@xmath83 ,
the ( @xmath81 ) colour index for virgo dwarfs reddens monotonically with increasing luminosity .
we extrapolate this colour magnitude relation to @xmath84 l@xmath12 , to cover the faintest dwarf ( cgw46 ) for which we have nuv data , with the results presented in fig .
the three dwarfs in the perseus core lie on the same cmr as those in the virgo cluster , and are not excessively blue , suggesting they have not undergone a recent burst of star formation . given that these dwarfs have remarkably smooth morphologies @xcite , and that their stellar populations show no evidence for formation via a recent infall , it is unlikely that the uv detections of these objects are due to recent star formation . instead , this result suggests that the uv detections of these dwarfs is due to the presence of an old stellar population .
this fact can be used to place a limit on the maximum star formation rate these detected dwarfs can have .
the limiting magnitude of our @xmath37 data is nuv = 24.5 ( @xmath85 ) , and given that most of our dwarfs are non - detections in the uv , or they are detected due to the presence of an old , evolved stellar population , the uv flux due to star formation must therefore be fainter than this .
the nuv luminosity of an object can be used to calculate a star formation rate , using the method of @xcite .
this gives a maximum star formation rate of @xmath86 m@xmath12yr@xmath23 for the dwarfs i.e. they must be forming fewer stars than this per year to be non - detections .
due to their low masses , dwarf galaxies are thought to be particularly susceptible to tidal forces , ram pressure stripping and harassment . in the centre of the cluster ,
the cluster tidal potential becomes stronger , so it is expected that smaller spheroidal galaxies will be disrupted in this region . in the outer parts of the cluster , these tidal forces are weaker and dwarfs are therefore more able to survive the cluster potential , resulting in higher number counts and a steeper faint end to the luminosity function ( e.g. @xcite ) .
however , in @xcite we argue that most dwarfs currently in the cores of rich clusters must have a large dark matter component to prevent such disruption .
we compare the relative numbers of low mass galaxies in the core and outskirts of perseus using number counts .
all sources in our acs and wfpc2 imaging were identified using sextractor @xcite .
objects with more than 50 contiguous pixels 2@xmath55 above the sky rms were identified .
we exclude from our acs catalog all objects with central coordinates less than 50 pixels from the ccd edges and gaps , to prevent the inclusion of incomplete objects in the catalog .
edge objects were likewise excluded from the wfpc2 catalog . to provide accurate number counts , stars must be removed from our object catalogs . as the perseus cluster is at a low galactic latitude ( @xmath87 ) ,
the object counts will be contaminated by large numbers of foreground stars . at the distance of perseus
, a dwarf galaxy with m@xmath88 will have an apparent magnitude @xmath89 , so we therefore require a reliable method to separate stars and galaxies down to this faint magnitude .
one way to separate stars and galaxies is to use the sextractor stellar classification index class_star , where a value of 0 corresponds to an extended source , and a value of 1 to a point source .
however , for fainter sources the reliability of this classification breaks down , making the separation of stars and galaxies difficult for sources fainter than @xmath90 ( e.g. @xcite ) .
the separation of galaxies and stars is instead done using the relationship between the sextractor parameters mum_max and mag_auto , where mag_auto is the total magnitude , and mu_max is the peak surface brightness of an object against the background level .
this method was utilized by @xcite to separate stars and galaxies in the cosmos survey , and is based on the fact that the light distribution of point sources scales with magnitude .
point sources such as stars occupy a well defined region in the mum_max vs mag_auto plane as they follow the point spread function ( psf ) , as shown in fig .
[ sep ] , allowing us to separate stars and galaxies down to f814w @xmath91 .
this method does not distinguish between des and dirrs , so both are included in our number counts .
the galaxies are binned according to magnitude , with the results shown in fig .
[ ncounts ] for both the core and outer regions of the cluster , along with the stellar counts for both regions .
the higher stellar counts in the acs data at faint magnitudes due to the presence of large numbers of globular clusters around the giant ellipticals , with these globular clusters contributing to the point source counts . to measure an accurate faint - end slope to the galaxy luminosity function
, it is essential to ensure that number counts are not contaminated by background galaxies .
this can be done using spectroscopy to measure the redshift of the galaxies in question , or via the statistical subtraction of background galaxies . however , given that the same background counts should apply for both the core and outer regions , we can make a direct comparison for both the core and outer regions without making a background subtraction . to allow comparison between the two surveys.,width=309 ]
we compare the number counts for the core and outer regions in fig .
[ ncountcomp ] .
the error bars are the standard deviation for each bin .
number counts below f814w = 23 are excluded as these are fainter than the completeness limits for the wfpc2 imaging . within our error bars ,
the number counts for the two populations are identical for dwarfs brighter than @xmath92 . at f814w @xmath4 21.5 ,
the number counts in the outskirts of perseus exceed those in the cluster core , suggesting that for at least the very faintest galaxies ( @xmath93 , @xmath14@xmath94@xmath95 ) environment plays an important role in their evolution .
one such role is that some dwarf ellipticals are unable to survive in the cores of rich clusters unless they are sufficiently dark matter dominated to prevent their disruption @xcite .
dwarfs in the cluster outskirts are subject to smaller direct environmental influences , therefore low mass galaxies are more able to survive in this region .
this results in a steeper faint - end slope to the luminosity function in lower density environments e.g. the ` dwarf galaxy - density relation ' suggested by @xcite .
we compare the morphologies and structures of des in both the core and outskirts of perseus , and find those in the cluster outskirts to be on average more disturbed than those in the cluster core .
this can be interpreted in two ways : either the dwarfs are systems undergoing the transformation from late - to - early type galaxies as they infall into the cluster , or they are on extremely radial orbits with small pericentric distances , resulting in them being morphologically disturbed via high speed interactions as they pass through the cluster core
. such high speed interactions are unlikely to destroy these galaxies unless they pass very close to the cluster core @xcite , but are likely to significantly modify their structures . however
, if the structures of such galaxies were indeed modified as they pass through their orbit pericenter , we would expect to see morphologically disturbed dwarfs in the cluster core , whereas in @xcite we identify a remarkably smooth population of des in this region , with no evidence of such morphological disruption .
however , not all the dwarfs in the outskirts have disturbed morphologies- some des are smooth as are all of those in the core , with a @xmath96 and s @xmath96 .
it is thought that clusters grow via the accretion of galaxy groups ( e.g. @xcite ) , and groups such as the local group contain a fraction of des and dsphs , as well as dirrs .
therefore at least some of the dwarfs entering the cluster will have smooth morphologies , and only those transforming from dirrs or spirals will have disturbed morphologies .
the des we identify lie on the colour magnitude relation expected for des in all environments , suggesting that if they do originate from an infalling population of galaxies transforming from late - type to early - type , they have lost their gas and therefore ceased star formation .
we infer that these galaxies in the outer regions of perseus are therefore `` transition dwarfs '' , with the cluster environment likely playing an important role in modifying these galaxies ( e.g. @xcite ) .
other examples of such dwarfs have been found in the virgo and fornax clusters @xcite .
these dwarfs lie on the red sequence , but do not have the smooth structures found for the dwarf ellipticals in the core of perseus .
this can be explained due to the timescales required for the colour and morphological transformation of a late - type to early - type galaxy .
the process of removing gas from a star forming dwarf can take place over short timescales given the shallow potential wells of these galaxies .
for example , ram - pressure stripping can remove the gas reservoir , shutting off star formation in a dwarf galaxy entering a cluster over a timescale of @xmath14100 myr @xcite .
evidence for this gas removal has been seen , with the majority of des in the virgo cluster do not contain hi gas @xcite , suggesting that the process of gas removal is very efficient in low mass galaxies , causing star formation to halt rapidly .
the colours of star forming galaxies , and star forming regions within these galaxies , are dominated by young , blue o and b type stars . however , once their gas has been removed via ram pressure stripping or some other method , the reservoir of gas required to fuel this star formation no longer exists .
therefore the colour of the galaxy will rapidly redden due to the short main - sequence life - times of massive stars .
we can use the @xcite models to find over what timescale a galaxy will transform from blue to red once star formation ceases .
we plot ( @xmath97 ) and ( @xmath98 ) colours versus time in fig .
[ bruzualcharlot ] .
the ( @xmath97 ) colours are included for reference as this colour will evolve faster than the ( @xmath98 ) colour .
it can be seen that a blue stellar population evolves rapidly to a red one , with a galaxy transforming from blue to red over a timescale of @xmath140.5 gyr .
if ram pressure stripping is the dominant mechanism driving the transformation from blue to red galaxies , then the complete transformation from a blue , star forming dwarf to a red , passive dwarf will take 0.5 - 1 gyr . ) and ( @xmath97 ) colour indices with time .
we include models for [ fe / h]@xmath99 and @xmath100 .
the shaded region represents the range in colours for our dwarfs in both environments , excluding the very red cluster candidates in our wfpc2 imaging due to the uncertainty in their cluster membership.,width=309 ] however , to completely transform from a late to early type , internal sub - structure must be removed from the galaxy .
the dwarfs in the outer regions contain asymmetries and clumpiness in their light distribution , whereas those in the core do not , so if the two populations of dwarfs have the same origin , clearly this structure must somehow be removed .
we can estimate over what timescale star clusters within the dwarf will dissipate , removing substructure caused by star formation .
stars typically form in clusters of @xmath101 m@xmath12 , and these will disperse and dissolve over time .
such clusters in dense environments such as the center of the mw will dissolve rapidly , on a timescale of @xmath102 myr @xcite . however , dwarf galaxies typically have stellar densities much lower than in more massive galaxies , and therefore the timescale for star clusters to evaporate will be longer .
open clusters in the outer regions of the mw are found to have ages between 0.5 - 6 gyr ( e.g. @xcite ) .
also , substructure such as spiral arms and bars has been found for two des in the fornax cluster @xcite , with ages @xmath103 gyr determined for these dwarfs via spectroscopy @xcite .
it is likely therefore that structure will remain in a dwarf galaxy that has ceased star formation for at least a gyr , though detailed simulations will be required to obtain a more precise life - time .
therefore , the morphological transformation from late to early type galaxy will take longer than the colour transformation .
other processes will also play a role in the transformation from late - to - early type galaxy to a lesser extent .
for example galaxy harassment @xcite caused by encounters with larger galaxies and the cluster s gravitational potential , will transform a disky late - type galaxy into a de or dsph over a timescale of 3 - 4 gyr .
such an encounter will cause the internal kinetic energy of the infalling galaxy to increase , modifying its energy and mass profile . to counteract this and regain equilibrium , the size of the galaxy will increase and it loses its most energetic stars . over time , the galaxies density decreases , resulting in the morphological transformation of a spiral to a de over several orbits ( @xmath104-@xmath105 gyr ) .
the number counts of faint dwarfs ( m@xmath106 ) suggest that in the dense cluster core , environmental influences become very important , with only the most dark matter dominated dwarfs with sufficient mass to survive the cluster tidal forces able to survive in this region .
these forces are lower in the cluster outskirts , resulting in higher number counts compared to the core region .
the result that the lowest mass dwarfs are preferentially found in less dense regions has been observed for other clusters .
@xcite find that the luminosity function in abell 2218 changes from @xmath107 in the cluster core to @xmath108 outside the radius @xmath109 kpc . likewise @xcite find a similar result for the hercules cluster ( a2151 ) , with the radial lf showing a trend of increasing faint - end slope as cluster - centric radius increases .
this result is likely due to tidal processes , which are more effective at smaller cluster - centric radii .
we identify from @xmath1 acs imaging a total of 11 dwarf ellipticals in the outskirts of the perseus cluster , all of which are newly discovered .
we compare the colours , structures and morphologies of these newly discovered dwarfs in the cluster outskirts to those in the core region , for which we have @xmath1 acs imaging . to quantify their morphologies
, we compute their concentration , asymmetry and clumpiness ( cas ) parameters and compare these results to those obtained in @xcite for the dwarfs in the core .
the dwarfs in the cluster outskirts have , on average , more disturbed structures than those in the core , with @xmath3a@xmath4 @xmath5 and @xmath3s@xmath4 @xmath6 , compared to @xmath3a@xmath4 @xmath7 , @xmath3s@xmath4 @xmath8 for those in the cluster core .
the fact that the dwarfs in the cluster outskirts are morphologically disturbed whereas those in the core are not suggests that those in the outskirts are part of an infalling population of galaxies transitioning from dirr to de , and not part of an original population .
we show that these galaxies lie on the same red sequence as those in the cluster core , showing they have ceased star formation .
further evidence for this lack of star formation comes from our @xmath37 imaging , whereby none of the dwarfs in the outer regions are detected in our nuv imaging , indicating they have sfrs of less than 0.001 m@xmath12yr@xmath23 . from this , and the fact that these dwarfs have more disturbed morphologies than those in the cluster core , we infer and show that the colours of these galaxies transform over a quicker timescale than their morphologies . by comparing the colours of these dwarfs to @xcite models of colour versus the time since star formation ceased , we find that the colours of these galaxies can transform rapidly over a timescale of @xmath14500 myr , whereas substructure likely remains in the galaxies for several gyr after star formation ceases . finally , to examine
if local galaxy density has an effect on the faint end slope of the luminosity function , we investigate how the number counts in the dense core of perseus compare to those in the outskirts .
the number counts for faint galaxies ( @xmath21 ) are higher in the cluster outskirts than in the core , suggesting that the faintest galaxies are unable to survive in the very dense parts of the cluster unless they are sufficiently dark matter dominated to prevent their disruption by the cluster potential .
this provides further evidence for the existence of a dwarf galaxy density relation , whereby a steeper faint - end slope of the luminosity function is found in the lower density environments .
sjp acknowledges the support of a stfc studentship .
evh acknowledges support from an asi - cofis grant .
rwo acknowledges support from the galex guest investigator program administered by nasa
goddard space flight center .
support for the us participants in programs go-10201 and go-10789 was provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
we would like to thank ruth grtzbauch and asa bluck for useful discussions .
99 aguerri j.a.l .
, gonzlez - garca , a.c . , 2009 , a&a , 494 , 891 beasley m.a . ,
cenarro a.j . , strader j. , brodie j.p . , 2009 , aj , 137 , 5146 bekki k. , couch w.j . , 2003 , apj , 596 , l13 bertin e. , arnouts s. , 1996 , a&as , 117 , 393 binggeli b. , tammann g.a .
, sandage a. , 1987 , aj , 94 , 251 boselli a. et al . , 2005 ,
apj , 629 , l29 bouchard a. , jerjen h. , da costa g.s .
, ott j. , 2007 , aj , 133 , 261 bressan a. , chiosi c. , fagotto g. , 1994 , apjs , 94 , 63 bruzual g. , charlot s. , 2003 , mnras , 344 , 1000 conselice c.j . ,
bershady m.a .
, jangren a. , 2000 , apj , 529 , 886 conselice c.j . , 2003 ,
apjs , 147 , 1 conselice c.j .
, gallagher j.s .
iii , wyse r.f.g .
, 2001 , apj , 559 , 791 conselice c.j .
, gallagher j.s .
iii , wyse r.f.g .
, 2002 , aj , 123 , 2246 conselice c.j . , gallagher j.s .
iii , wyse r.f.g .
, 2003 , aj , 125 , 66 conselice c.j . , gallagher j.s .
iii , wyse r.f.g .
, 2003 , apj , 591 , 167 cortese l. , gavazzi g. , boselli a. , franzetti p. , kennicutt r.c . , oneil k. , sakai , s. , 2006 , a&a , 453 , 847 ct p. , et al . , 2006 , apjs , 165 , 57 de rijcke s. , dejonghe h. , 2002 , mnras , 329 , 829 de rijcke s. , dejonghe h. , zeilinger w.w .
, hau g.k.t . , 2003 , a&a , 400 , 119 de rijcke , s. , penny s.j . , conselice c.j .
, valcke s. , held e.v .
, 2009 , mnras , 393 , 798 dressler a. , 1980 , apj , 236 , 351 ferguson h.c .
, sandage a. , 1989 , apj , 346 , l53 gallagher j.s .
iii , hunter d.a .
, 1989 , aj , 98 , 806 grant n.i . , kuipers j.a .
, phillipps s. , 2005 , mnras , 363 , 1019 grebel e.k . , gallagher j.s .
iii , harbeck d. , 2003 , aj , 125 , 1926 hasegawa t. , sakamoto t. , malasan h.l . , 2008 , pasj , 60 , 1267 holtzman a. , burrows c.j .
, casertano s. , hester , j.j . , trauger , j.t . ,
watson , a.m. , worthey g. , 1995 , pasp , 107 , 1065 jerjen h. , kalnajs a. , binggeli b. , 2000 , a&a , 358 , 2000 kennicutt r.c .
, 1998 , araa , 36 , 189 kim s. s. , figer d. f. , lee h. m. , morris m. , 2000 , apj , 545 , 301 kormendy j. , 1985 , apj , 295 , 73 kormendy j. , fisher d. b. , cornell m. e. , bender r. , 2009 , apjs , 182 , 216 kron r.g . , 1995 , the deep universe : saas - fee advanced course 23 .
springer , new york leauthaud a. , et al .
, 2007 , apjs , 172 , 219 lisker t. , grebel e.k .
, binggeli b. , 2006 , aj , 132 , 497 lisker t. , fuchs b. , 2009 , a&a , 501 , 429 lisker t. , et al .
, 2009 , apj , 706 , l124 lotz j.m . ,
miller b.w .
, ferguson h.c . , 2004 ,
apj , 613 , 262 michielsen d. , de rijcke s. , dejonghe h. , 2004 , astron .
, 325 , 122 michielsen d. , koleva m. , prugniel p. , zeilinger w.w . , de rijcke s. , dejonghe h. , pasquali a. , ferreras i. , debattista v.p . , 2007 , apj , 670 , 10 michielsen d. , et al .
, 2008 , mnras , 385 , 1374 mieske s. , hilker m. , infante l. , mendes de oliveira c. , 2007 , a&a , 463 , 503 misgeld i. , mieske s. , hilker m. , 2008 , a&a , 486 , 697 moore b. , katz n. , lake g. , dressler a. , oemler a. , nature , 1996 , 379 , 613 moore b. , lake g. , katz n. , 1998 , apj , 495 , 139 mori m. , burkert a. , 2000 , apj , 538 , 559 penny s.j .
, conselice c.j . , 2008 , mnras , 383 , 247 penny s.j . , conselice c.j .
, de rijcke s. , held e.v .
, 2009 , mnras , 393 , 1054 petrosian v. , 1976 , apj , 209 , l1 phillipps s. , parker q.a . , schwartzenberg j.m . , jones j.b
, 1998 , apj , 493 , l59 phillips s. , driver s.p . , couch w.j . ,
smith r.m . , 1998 ,
apj , 498 , l119 poggianti b.m .
et al , 2001 , apj , 562 , 689 pracy m. b. , de propris r. , driver s. p. , couch w. j. , nulsen p. e. j. , 2004 , mnras , 352 , 1135 reaves g. , 1962 , pasp , 74 , 392 snchez - janssen r. , iglesias - pramo j. , muoz - tun c. , aguerri j. a. l. , vlchez j. m. , 2005 , a&a , 434 , 521 sandage a. , 1972 , apj , 176 , 21 sandage a. , hoffman g.l .
, 1991 , apj , 379 , 45 schlegel d.j .
, finkbeiner d.p . , davis m. , 1998 , apj , 500 , 525 sirianni m. , et al . , 2005 ,
pasp , 117 , 1049 smith castelli a.v .
, bassino l.p .
, richtler t. , cellone s.a . , aruta c. , infante l. , 2008 , mnras , 386 , 2311 struble m.f . ,
rood h.j . , 1999 ,
apjs , 125 , 36 thompson l.a .
, gregory s.a . , 1993 ,
aj , 106 , 2197 trentham t. , tully r.b . , 2009 , mnras , 398 , 722 van den bergh s. , 1994 , aj , 107 , 1328
|
we present the results of a _ hubble space telescope ( hst ) _ study of dwarf galaxies in the outer regions of the nearby rich perseus cluster , down to @xmath0 , and compare these with the dwarf population in the cluster core from our previous @xmath1 imaging . in this paper
we examine how properties such as the colour magnitude relation , structure and morphology are affected by environment for the lowest mass galaxies .
dwarf galaxies are excellent tracers of the effects of environment due to their low masses , allowing us to derive their environmentally based evolution , which is more subtle in more massive galaxies .
we identify 11 dwarf elliptical ( de ) and dwarf spheroidal ( dsph ) galaxies in the outer regions of perseus , all of which are previously unstudied .
we measure the ( @xmath2 colours of our newly discovered des , and find that these dwarfs lie on the same red sequence as those in the cluster core .
the morphologies of these dwarfs are examined by quantifying their light distributions using cas parameters , and we find that des in the cluster outskirts are on average more disturbed than those in the core , with @xmath3a@xmath4 @xmath5 and @xmath3s@xmath4 @xmath6 , compared to @xmath3a@xmath4 @xmath7 , @xmath3s@xmath4 @xmath8 for those in the core . based on these results , we infer that these objects are `` transition dwarfs '' , likely in the process of transforming from late - type to early type galaxies as they infall into the cluster , with their colours transforming before their structures .
when we compare the number counts for both the core and outer regions of the cluster , we find that below @xmath0 , the counts in the outer regions of the cluster exceed those in the core .
this is evidence that in the very dense region of the cluster , dwarfs are unable to survive unless they are sufficiently massive to prevent their disruption by the cluster potential and interactions with other galaxies .
galaxies : dwarf galaxies : clusters : general galaxies : clusters : individual : perseus cluster
|
experiments with tunnel junctions using ferromagnetic metals@xcite have been an interesting topic since a long time .
this subject has grown again@xcite because of the new field of spintronics where spin dependent currents are an important requisite of many possible devices.prinz , boeck this implies that control and measurements of spin polarized currents are needed .
spin - polarized electron tunnelling@xcite is a key tool to measure the current polarization and to understand the physics involved in these effects .
most of the recent experimental works have focused on the suppression of the andreev reflection and have used the point contact geometry with a superconducting electrode@xcite . however
the local information extracted by point contact or scanning tunnelling microscope techniques seems to be less suitable for devices than planar tunneling junctions .
in addition , the intrinsic difficulty of fabricating a perfect uniform oxide layer can jeopardize the latter technique .
recently kim and moodera@xcite have reported a large spin polarization of 0.25 from polycrystalline and epitaxial ni ( 111 ) films using meservey and tedrows s technique@xcite and standard al electrode and oxide barriers , that allows for an almost ideal barrier behaviour . in this work
we show that the current polarization of ferromagnets can be extracted without applying a magnetic field to the junction and using an oxide barrier that is far from ideal .
the experimental data are obtained for nb / nb@xmath2o@xmath3/ni planar tunnel junctions , where the barrier is fabricated using the nb native oxides . in this case , different nb@xmath2o@xmath4 oxides are present in the barrier .
this fact usually prevents the analysis of the dv / di characteristics in terms of perfect tunnelling .
indeed , we show below that our junctions are neither in the tunnelling nor in the transparent regime , but rather in a regime intermediate between both .
we shall argue that the current polarization can be obtained even in this intermediate regime by use of a simple model .
nb(110 ) and ni(111 ) films , grown by dc magnetron sputtering , were used as electrodes . the structural characterization of these films
was done by x - ray diffraction ( xrd ) and atomic force microscopy ( afm ) , see for instance villegas _
et al_.@xcite briefly , the junction fabrication was as follows : first , a nb thin film of 100 nm thickness was evaporated on a si substrate at room temperature .
an ar pressure of 1 mtorr was kept during the deposition . under these conditions ,
the roughness of the nb film , extracted from xrd and afm , is less than 0.3 nm@xcite and superconducting critical temperatures of 8.6 k are obtained .
after this , the film was chemically etched to make a strip of 1 mm width .
a tunnel barrier was prepared by oxidizing this nb electrode in a saturated water vapor atmosphere at room temperature.@xcite the thickness of the oxide layer , extracted from the simulations performed with the suprex program,@xcite is 2.5 nm .
x - ray photoelectron spectroscopy analysis performed in these oxidized films reveals that dielectric nb@xmath5o@xmath6 is the main oxide formed .
there are also other oxides , such as metallic nbo , but in much less amount .
taken into account grundner and halbritter studies,@xcite nb@xmath5o@xmath6 is the outermost oxide layer on nb , whereas nbo is located closer to the nb film .
the characterization by afm reveals a rms roughness of around 0.7 nm
. a detailed account of the structural and compositional characterization of the barrier has been reported elsewhere.@xcite on top of this film ( nb with the oxide barrier ) , the second electrode of ni was deposited under the same conditions as nb ( up to 60 nm thickness ) using a mask to produce cross strips of 0.5 mm width , so that the overlap area @xmath7 of the two electrodes is 0.5 mm@xmath8 .
junctions fabricated using these materials and geometry , will not show a good tunneling behaviour , neither a point - contact behaviour . as we will see , the junctions will lie on the intermediate regime .
perpendicular transport in tunnelling configuration was investigated by means of characteristic dynamic resistance ( @xmath9 ) versus voltage ( @xmath10 ) using a conventional bridge with the four - probe method and lock - in techniques .
the measured lock - in output voltage was calibrated in terms of resistance by using a known standard resistor .
we have measured the conductance , defined as the inverse of the differential resistance @xmath9 , of three different tunnel junctions .
moreover , junction 1 ( j1 ) has been measured and analyzed at temperature @xmath11 k , junction 2 ( j2 ) , at temperatures @xmath12 and @xmath13 k , and junction 3 ( j3 ) , at temperatures @xmath14 , @xmath15 and @xmath16 k. these measurements have allowed us to access and assess the behavioiur and quality of our samples in the low ( j1 and j2 ) and intermediate ( j2 and j3 ) temperature range of this heterojunction .
each data set presented in this article has been normalized with respect to the background conductance @xmath17 of the corresponding junction .
the conductance @xmath18 across a normal / superconducting junction may be expressed in terms of the reflection probabilities @xmath19 of quasiparticles transversing the junction , and of andreev processes @xmath20 as @xmath21 where @xmath22 is the fermi function ; and @xmath10 is the applied voltage .
andreev reflection processes are proportional to the square of the conventional transmission coefficient of the barrier , @xmath23 , and therefore are strongly suppressed for highly resistive barriers .
junctions with transmission coefficients smaller than about 0.1 show small subgap conductances , and can be classified as belonging to the tunneling regime .
our experimental results , shown as black circles in [ fig1 ] and [ fig2 ] , exhibit a significant conductance below the superconducting gap even at the lowest temperatures . therefore we expect that our effective oxide barriers should be neither too high nor too thick , and we classify them as belonging to an intermediate transparency regime .
while there are fairly complete descriptions of the transmission across ferromagnet / superconducting junctions@xcite , we have decided to describe it by three models that share the virtue that are conceptual and algebraically simple .
these models are : i ) the generalization of the blonder - tinkham - klapwijk ( btk ) model @xcite to ferromagnetic electrodes proposed by strijkers and coworkers;@xcite ii ) a description of the effects of a finite current polarization in terms of spin dependent transmission coefficients , as discussed by prez - willard _
et al_;@xcite and iii ) a very simple generalization of the btk model to ferromagnetic electrodes with finite bulk magnetization .
we define the current polarization @xmath24 as the imbalance between the current intensity of majority and minority carriers@xcite in a metal / ferromagnet junction , measured when the voltage tends to zero , @xmath25 where both spin channels contribute equally to the total current intensity and conductance .
k. black circles indicate experimental data ; dashed - dotted green lines are fits to model i , dashed blue , fits to model ii , and solid red , fits to model iii.,width=302,height=264 ] [ fig1 ] strijkers model uses as adjustable parameters the current polarization @xmath24 , the height of barrier @xmath26,@xcite which is modeled by a delta function , and the size of the superconducting gap at the interface @xmath27 .
the process of electron transfer across the junction is splitted into a fully polarized channel , for which the andreev reflection coefficient @xmath28 is zero , and @xmath29 where @xmath30 and @xmath31 are the coherence coefficients of the superconducting wave function ; and a paramagnetic channel described by the coefficients @xmath32 and @xmath33 of the btk model in its conventional form.@xcite the total conductance is then written in terms of the conductance of the fully polarized channel ( @xmath34 ) and the conductance of the paramagnetic channel ( @xmath35 ) : @xmath36 this model interpolates between the paramagnetic case ( btk model ) , and the half metal , where it predicts correctly that the amplitude for andreev reflection vanishes .
k ( a ) and 3.945 k ( b ) , together with a fit to models i ( dash - dotted green ) , ii ( dashed blue ) and iii ( solid red).,width=302,height=264 ] [ fig2 ] we now use a simple model based on quasiclassical theory,@xcite in which boundary conditions have been dumped into the spin - dependent transmission coefficients @xmath37@xcite .
the conductance for each spin channel in this model is determined by @xmath38 where the effective reflection coefficients depend now on spin .
these can be expressed in terms of the normal @xmath39 and anomalous @xmath40 components of the green s function , evaluated right at the superconducting side of the interface , as follows @xmath41 where the reflection amplitude of the barrier satisfies the sum rule @xmath42 .
the explicit functional form of @xmath43 and @xmath44 may be obtained from the equation of motion of the quasiclassical green function that , for a bulk superconductor , reduces to @xmath45 where disorder and pair breaking effects , parametrized by the dynes parameter @xmath46,@xcite are dealt within the t - matrix approximation .
( 5 ) is used to fit the conductance data , using the two transmissions @xmath37 , @xmath47 and @xmath48 as adjustable parameters .
4.53 k ( a ) , 5.0 k ( b ) and 5.39 k ( c ) , together with a fit to models ii ( dashed blue ) and iii ( solid red).,width=302,height=264 ] [ fig3 ] eqs .
( 7 ) can also be solved when the gap @xmath48 is zero .
this gives @xmath49 and @xmath50 , and corresponds the solution for the normal state .
then , the conductance per spin chanel is , simply , @xmath51 , and the current polarization is obtained from it as @xmath52 perez - willard et al.@xcite used this model to analyze al / co point contacts , where the proximity effect may be discarded since the size of the junction is negligible compared with the coherence length .
they indeed found close agreement with their experimental data for a wide range of temperatures .
.parameters used to fit junctions 1 , 2 , and 3 with model ii . here
@xmath53 [ cols="^,^,^,^,^,^,^,^",options="header " , ]
it is also apparent from the table that model ii provides values for the superconducting gap that are too small , and actually do not seem to follow the temperature dependence expected for a bcs superconductor .
for instance , the model predicts a zero - temperature gap for j3 , @xmath54 mev . using the measured critical temperature , we find that the ratio @xmath55 is equal to 3.25 , which is much smaller than that of bulk nb ( 3.8 ) . to understand better why the model fails to fit the zero voltage conductance of j2 and j3 , we plot in fig .
4 @xmath56 for several sets of the parameters
@xmath57 , @xmath24 and @xmath47 , that cover most of the parameter space .
we have chosen for @xmath57 the values 0.1 , 0.4 , 0.55 and 0.7 .
for each @xmath57 , we have taken two representative values of @xmath58 ( 0.15 and 0.35 ) and three different values of @xmath47 , ( 0.1 , 0.3 and 0.5 ) .
the figure shows that the 24 curves cluster in four different sets , according to the value of @xmath57 .
this implies that the value and temperature dependence of the zero - voltage conductance is determined in this model essentially by the average transmision .
the figure demonstrates in any case that the experimental values of @xmath59 for junctions j2 and j3 show a stronger dependence with temperature than the estimates provided by model 2 .
we therefore believe that the model , while very appealing due to its simplicity , is actually too simple to describe the physical behavior of these junctions that belong to the intermediate transparency regime and have been measured at low and intermediate temperatures . as a function of reduced temperature .
red circles and black squares correspond to the experimentals data of j2 and j3 , respectively . dashed red and solid black lines correspond to the values of @xmath59 obtained with model iii , using the parameters shown in table ii.,width=302,height=226 ] we turn the attention now to our generalized btk model ( model iii ) .
we note that the model is able to fit well the conductance data of junctions 1 and 2 .
in addition , it also provides a good fit to the data of j3 , in contrast to the quasiclassical model .
more importantly , the model provides a temperature - independent barrier height @xmath26 , that translates into temperature - independent transmission coefficients ( see table ii ) .
the values of z so obtained allow us to clasify the junctions in an intermediate regime between tunnelling and point contact .
junction 2 is actually the closest to the tunneling regime , while junction 1 is closest to the transparent regime .
junction 3 , that we failed to fit with model ii , lies well within this intermediate regime .
5 clearly demonstrates that our model provides a temperature dependence of @xmath59 that fits well the experimental data .
( a ) shows the temperature dependence of the gap @xmath48 obtained in the fits performed with model iii .
we find values for the gap slightly smaller than those of bulk nb , but consistent with the measured critical temperature and with the conventional temperature dependence of a bcs superconductor .
the extrapolated zero - temperature gap , @xmath60 mev , provides a superconducting ratio @xmath61 in close agreement with the ratio for nb , 3.8 .
( a ) also shows the temperature dependence of the disorder parameter .
we find that @xmath47 shows a smooth and linear dependence within the studied range of temperatures .
the values of @xmath47 are large , as should be expected , since our nb samples are highly disordered .
we plot in fig
. 6 ( b ) the current polarization @xmath24 obtained using model iii , as a function of the height of the barrier z. we find that @xmath24 increases with z from 0.13 to 0.19 , which shows that the current polarization increases with the tunelling quality of the junction .
we note that soulen et al .
@xcite have found polarizations of about 45% using ni / nb junctions in a point contact geometry .
more recently kim and moodera @xcite have performed experiments for ni / al plannar tunnel junctions .
they have found that @xmath24 grows from 11% to 33% as the tunneling quality of the junctions increases .
our theoretical calculations confirm that @xmath24 should increase with the strength or the thickness of the tunneling barrier .
the preceding analysis shows that the main advantage of model iii ( btk ) over model ii ( adjustable spin dependent transmission coefficients ) is the stronger dependence on temperature of the conductance in model iii .
the main difference between the two models lies in the description of the ferromagnetic electrode .
model iii uses two different fermi surfaces , one for each spin .
hence , the proximity effect is suppressed , reducing the andreev reflection at low temperatures . of the delta - function barrier entering our generalized btk model for the three junctions studied in this article .
lines have been added to aid the eye.,width=302,height=226 ]
we have fabricated and measured fairly large and far from ideal superconducting / ferromegnetic tunel junctions .
a simple theoretical model allows us to extract tunnelling related parameters for such junctions .
we have grown nb / nb@xmath0o@xmath1/ni planar tunnel junctions , and measured its conductance at different temperatures .
we have found that they belong to the regime of intermediate transparencies .
we have been able to fit the conductance curves with a simple generalization of the btk model,@xcite that provides a sensible set of temperature - independent transmission probabilities .
our calculations suggest that the results can depend significantly on the description of the bulk ferromagnetic electrode , as andreev reflection depends both on the barrier transmission coefficients and on the exchange field inside the ferromagnet .
we have been able to give reasonable values of the current polarization without the need to apply a magnetic field .
we have also studied the relationship between these two quantities , showing that the current polarization depends significantly on the height of the barrier .
this simple generalization of the btk model to a superconducting / ferromagnetic junction could be improved considering a 3d model .
however , we have shown that the simple one dimensional model can fit the experimental data and also provides good estimates of the current polarization .
we wish to thank very useful discussions with j. moodera , m. eschrig and m. fogelstrom .
this work was supported by mec ( mat2002 - 04543 , mat2002 - 12385e , mat2002 - 0495-c02 - 01 , bfm2003 - 03156 and ap2002 - 1383 ) , and cam ( gr / mat/0617/2004 ) .
additionally , e.m.g . acknowledges spanish ministerio de educacin y ciencia for a ramn y cajal contract , and r. escudero thanks universidad complutense and ministerio de educacin y ciencia for a sabbatical professorship .
f . thanks ministerio de educacin y ciencia for a fpu grant ( ap2002 - 1383 ) .
|
we have measured the tunnelling current in nb / nb@xmath0o@xmath1/ni planar tunnel junctions at different temperatures .
the junctions are in the intermediate transparency regime .
we have extracted the current polarization of the metal / ferromagnet junction without applying a magnetic field .
we have used a simple theoretical model , that provides consistent fitting parameters for the whole range of temperatures analyzed .
we have also been able to gain insight into the microscopic structure of the oxide barriers of our junctions .
|
the dark matter ( dm ) density in our galaxy is about ( see e.g. ref . ):
@xmath0 however , only upper limits on the level of @xmath1 ( see below ) are known for the density of dark matter particles ( dmp ) in the solar system ( ss ) .
meanwhile , information on their density is of great importance for the experiments aimed at the detection of dm .
the question of capture of weakly interacting massive particles ( wimp ) or dmp by the ss was pioneered in refs . .
very recently these studies are pushed further by extensive numerical simulations performed in refs . .
our interest to this problem was attracted by a recent ref . where an estimate is given for the dm density in ss , as resulting from the gravitational capture of galactic dmp s .
according to the conclusions of ref . , the density of the captured dm , for instance , at the earth orbit is about @xmath2 , only about an order of magnitude below the best upper limits on it . in the present note
we perform new analytical analysis of the gravitational capture of galactic dmp s by the ss . according to our results , the increase of the dm density in the ss due to
this capture is small , certainly well below @xmath2 .
the solar system is immersed in the halo of dark matter and moves together with it around the center of our galaxy . to simplify the estimates ,
we assume that the sun is at rest with respect to the halo . the dark matter particles in the halo
are assumed to have in the reference frame , co - moving with the halo , the maxwell distribution ( see ref . ):
@xmath3 with the local rms velocity @xmath4 km / s .
let us elucidate what looks to be the most efficient mechanism of the dmp capture .
it was pointed out and partly analyzed ( though for the capture of comets , but not of dmp s ) , by petrosky in ref . and chirikov and vecheslavov in ref . .
of course , a particle can not be captured by the sun alone .
the interaction with a planet is necessary for it , this is essentially a three - body problem of the sun , planet and dmp .
obviously , the capture is dominated by the particles with orbits close to parabolic ones with respect to the sun , and with the distances between their perihelia and the sun comparable with the radius of planet orbit @xmath5 .
the capture can be effectively described by the so - called restricted three - body problem ( see for instance refs . ) . in this approach the interaction between two heavy bodies ( the sun and a planet in our case )
is treated exactly . as exactly is treated the motion of the third , light body ( a dmp in our case ) in the gravitational field of the two heavy ones .
one neglects however the back reaction of a light particle upon the motion of the two heavy bodies .
obviously , this approximation is fully legitimate for our purpose .
still , the restricted three - body problem is rather complicated , and requires in the present case both subtle analytical treatment and serious numerical calculations ( see ref . ) . under certain conditions
the dynamics of light particle ( e.g. dpm ) becomes chaotic .
however , the amount of dm captured by the ss can be found by means of simple estimates .
the total mass captured by the sun ( its mass is @xmath6 ) together with a planet with mass @xmath7 , during the lifetime @xmath8 of the ss , can be written as follows : @xmath9 here @xmath10 is the capture cross - section .
the product @xmath11 is averaged over distribution ( [ ma ] ) ; with all typical velocities in the ss much smaller than @xmath12 , this distribution simplifies to @xmath13 to estimate the average value @xmath14 , we resort to dimensional arguments , supplemented by two rather obvious physical requirements : the masses @xmath7 and @xmath6 of the two heavy components of our restricted three - body problem should enter the result symmetrically , and the mass @xmath15 of the light component ( dmp ) should not enter the result at all .
thus , we arrive at @xmath16 here @xmath17 is the newton gravitation constant ; an extra power of @xmath18 , inserted into this expression , is perhaps inherent in @xmath10 .
the final estimate for the captured mass is @xmath19 since the capture would be impossible if the planet were not bound to the sun , it is only natural that the result is proportional to the corresponding effective `` coupling constant '' @xmath20 .
thus obtained values for the masses of dm captured due to the planets of the ss , are presented in table 1 .
we quote also therein the corresponding results of ref . for these masses .
the disagreement is huge for all planets , and especially for the light ones where it exceeds two orders of magnitude .
we can not spot exactly its origin since the calculations of ref . involve rather complex numerical simulations ( it is possible that their assumption of capture radius @xmath21 does not correspond to reality ) . on the other hand , however , we can not see any reasonable possibility for a serious increase of our results .
moreover , in a sense they can be considered as upper limits for the amount of the captured dm , at least because we have neglected here the inverse process , that of the ejection of a captured dmp due to the same three - body gravitational interaction .
the result ( [ de1 ] ) is given for the three body problem .
the dynamical mechanism of capture is described below in next section .
the contribution of the diffusive ( non - dark ) matter in the ss should be significantly smaller since a homogeneously distributed dust gives compensation of gravitational forces acting on dmp .
the total mass @xmath22 of the dm captured by the planets is strongly dominated by the heavy jovian planets , jupiter , saturn , uranus , neptune , and constitutes according to table 1 about @xmath23 g. this value is small as compared to the total mass @xmath24 g of the common matter in the ss . it is small even as compared to the total non - captured mass of the dm in the ss : this total mass , calculated with value ( [ rho ] ) for the dm density , constitutes @xmath25 g ( we assume here that the effective radius of the ss is about @xmath26 au ) .
however , it is an order of magnitude larger than the dm mass of density ( [ rho ] ) inside the radius of neptune orbit @xmath27 au . the contribution to the discussed effect of the diffuse ( non - dark ) matter in the ss should be significantly smaller since in a homogeneous dust the gravitational forces acting on dmp are compensated .
the dynamical mechanism of capture is described in the next section .
for the restricted three - body problem , in a close similarity to dynamics of comets ( see refs . ) , the dmp dynamics can be described by a symplectic area - preserving map @xmath28 here @xmath29 is the energy of dmp with mass @xmath15 rescaled by its gravitational energy on distance @xmath5 from the sun ; @xmath30 is the phase of the planet on its circular orbit at the moment when the dmp is at the perihelion and @xmath31 is a certain periodic function of @xmath30 .
bars denote the new values of variables after one rotation around the sun .
the physical meaning of this dynamical map is rather simple : the first equation gives the change of dmp energy after one passage near the sun , the second equation gives the change of the planetary phase between two passages of dmp and is essentially determined by the kepler law .
the first equation is valid also for scattering particles with positive energy ( @xmath32 ) .
thus dmp can be captured by the sun and the planet only if its rescaled energy @xmath33 .
after the capture , the dmp dynamics is described by map ( [ map ] ) until ejection . to compute the captured dm mass @xmath34 we assume that , after being once captured , the dmp remains captured for the whole life time @xmath35 of ss . in this way we obtain the maximum bound for @xmath34 .
the kick function @xmath31 was computed in ref . for the case
when a comet ( or dmp ) and a planet move in one plane and when the perihelion distance @xmath36 . in this case
@xmath37 and the function @xmath38
so that @xmath39 . effectively the function @xmath40 is determined by the frequency fourier component of the force between the planet and dmp , since the rotation of the planet is rapid compared to the rotation of dmp the amplitude of the component is exponentially small for @xmath41 when the dmp motion is smooth and analytical . in this case
@xmath42 is exponentially small and there is practically no trapping of dmp .
for @xmath43 the motion is not analytic due to close passage between the planet and dmp and @xmath42 is relatively large . in this case
a dmp with rescaled energies @xmath44 can be captured by the planet .
it is interesting to note that the map ( [ map ] ) with @xmath45 is known as the kepler map .
it describes the process of microwave ionization of rydberg atoms and chaotic autoionization of molecular rydberg states ( see ref . and refs .
therein ) .
we note that the energy change of dmp given by @xmath31 results from the integration over the whole orbit rotation of dmp around the sun which includes many orbital periods of the planet . thus this energy change appears from long - range interaction and has qualitatively different origin compared to local close collisions between dmp and planet which were assumed to give the main contribution for dmp energy change in refs . .
let us now estimate the capture cross - section @xmath10 assuming that for all dmp the dynamics is described by the kepler map with fixed @xmath46 , then only dmp with energies @xmath47 are captured under the condition that @xmath48 ( here @xmath49 is the velocity of the planet ) .
the value of @xmath50 can be expressed via the dmp parameters at infinity where its velocity is @xmath51 and its impact parameter is @xmath52 and hence @xmath53 ( see ref . ) . since @xmath43 we obtain the cross - section @xmath54 where the last relation is taken for those typical velocities @xmath55 at which the capture of dmp takes place ( for @xmath56 we have @xmath57 ) .
( [ de ] ) , ( [ sigma ] ) give the captured mass @xmath34 of ( [ de1 ] ) with an additional numerical factor @xmath46 . according the above estimates dpm captured by jupiter have typical velocities at infinity @xmath58 for typical @xmath59 and @xmath60 , @xmath61 .
this value of @xmath51 is in a good agreement with the numerical simulations of peter ref . which give typical captured dpm velocities for jupiter of @xmath62 .
another interesting feature of the analytical expression for the cross - section of captured particles @xmath10 ( [ sigma ] ) is that it is much larger than the area of the planet orbit .
in fact , @xmath10 diverges at small velocities as @xmath63 but this divergence is weaker compared to the divergence of the reserford cross - section . in our case of the restricted three - body problem
the divergence appears due to the property of the kepler motion where the dmp distance at perihelion is proportional to square of the orbital momentum which in its turn is proportional to the product of the velocity @xmath51 and impact parameter @xmath52 at infinity .
in addition it is important to use the value of typical dmp velocity captured by the planet for perihelion distance of the order of @xmath5 .
this leads to the analytical equation ( [ sigma ] ) for the capture cross - section in the restricted three - body problem .
while the total masses @xmath34 of the captured dm can be ( hopefully ) described by simple dimensional estimate ( [ de1 ] ) , the situation for the corresponding dm densities @xmath64 is more subtle .
the reason is as follows .
the captured dmp s had initial trajectories predominantly close to parabolas with respect to the sun , and the velocities of these dmp s change only slightly as a result of scattering .
therefore , it is quite natural that the final , elliptical trajectories of these dmp s have large semi - major axes .
indeed , dmp captured into an elliptic trajectory had initially a hyperbolic trajectory , focussed at the sun and close to a parabolic one . as a result of the capture , the eccentricity @xmath65 of the trajectory changes from @xmath66 to @xmath67 , with @xmath68 .
it is quite natural that the final , elliptical trajectories of the captured dmp s have large semi - major axes . to estimate their typical values , we recall ( see ref . ) that the radius - vector @xmath69 of a captured dmp ( counted off the sun ) is related to the azimuthal angle @xmath30 as follows : @xmath70 where @xmath71 is the so - called orbit parameter ( its value is irrelevant for our line of reasoning )
, the maximal @xmath72 and minimal distance @xmath73 from the sun correspond to @xmath74 , so that their ratio is @xmath75 in the numerator of this ratio , we can safely put with our accuracy @xmath76 . as to its denominator , we recall that the difference @xmath77 is related to the gravitational perturbation by planet , and therefore is proportional to @xmath7 . thus , by dimensional reasons , @xmath78 the minimal distance between dmp and the sun @xmath73 should be on the same order of magnitude as the radius @xmath5 of the planet orbit .
therefore , the semi - major axis @xmath79 of resulting ellipse is huge : @xmath80 in particular , in the case of jupiter our estimate gives @xmath81 .
a similar numerical factor appears in ref . .
according to the numerical calculations of ref .
, in the case of jupiter the values of the semi - major axes @xmath82 for the resulting trajectories belong to the interval 10@xmath83 10@xmath84 au for @xmath85 @xmath86 . the fact that it is comets that are considered in ref . , but
not dmp s , is obviously of no importance for this conclusion .
the minimum value of @xmath82 is defined by the maximum @xmath87 value which can be reach by an injected dmp during its chaotic motion .
in fact @xmath88 is the chaos border and according to the chirikov resonance - overlap criterion ( see ref . ) we have @xmath89 as it was shown in ref . . for jupiter @xmath60 and at @xmath90 corresponding to @xmath56 we have @xmath91 .
we note that this value of @xmath42 gives the maximum @xmath92 corresponding to the similar value found for the comet halley ( see ref . ) .
in fact , the data presented in ref . show that the comet halley have the chaos border around @xmath93 ( see fig.3 in ref . ) .
of course the values of @xmath82 linked to the chaos border in @xmath94 are the minimum ones since during its chaotic dynamics dmp have also @xmath95 with larger @xmath82 . however , we are interested in orbits captured for very large times @xmath35 ( [ t ] ) .
such times are by two orders of magnitude larger than a typical diffusive life time of comet halley found to be of the order of @xmath96 years ( see ref . ) .
it is known that chaotic trajectories may be sticking to boundaries of integrable islands for very long times ( see ref . and refs .
therein ) and hence we can expect that those orbits will be somewhere in vicinity of the chaos border around @xmath97 with @xmath98 .
in fact , for the case when the inclination angle between the planes of dmp and planet @xmath99 and when @xmath48 the function @xmath31 contains higher harmonics of @xmath30 ( see the case of the comet halley in ref . ) .
this leads to easier emergence of chaos so that even for light planets one may have the chaos border @xmath97 .
therefore we can make an assumption resulting in the most optimistic prediction for the `` partial '' dark matter densities @xmath100 .
we assume that each of the total masses @xmath34 of the captured dm occupies the volume @xmath101 where @xmath5 is the orbit radius of the corresponding planet .
we do not claim that this assumption is correct , but believe , however , that the comparison of its ( almost certainly , overoptimistic ) results with the observational limits will be instructive .
the corresponding values of the `` partial '' dark matter densities @xmath102 ( in g / cm@xmath83 ) are presented in table 2 .
we omit in it the densities due to uranus and neptune , tiny even at the discussed scale .
then , in accordance with the accepted model , the total dark matter density @xmath104 at a given radius does not coincide with the corresponding @xmath64 .
it includes , in line with it , the sum of the contributions to the density due to all the planets , outer with respect to the given one .
the estimates given above neglect ejection of dmp from the ss .
such an assumption is not justified if the dmp dynamics in ss becomes completely ergodic on a time scale @xmath105 .
then after the time @xmath106 the detailed balance principle becomes valid and the density of captured dmp becomes as its galactic density as it was argued in ref .
( see also discussion in refs . ) .
however , the estimate of @xmath106 have certain subtle points . in the frame of the map ( [ map ] ) it is given by the diffusion time from @xmath107 to @xmath108 . for the kepler map
the diffusion coefficient is @xmath109 and hence @xmath110 where @xmath111 is an average period of dmp . for the case of jupiter such an estimate gives a satisfactory value of @xmath112 years for the case of the comet halley as it is discussed in ref .
however , @xmath42 sharply decreases with the increase of perihelion distance @xmath113 , where @xmath114 is the orbital momentum of dmp . as a result
a growth of @xmath114 can give sharp increase of @xmath106 which can become comparable with @xmath35 . the effects linked to variations of @xmath114 were not considered in ref . .
they are properly treated in the numerical simulations of refs .
but there only the effect of jupiter is considered .
other planets and fluctuating galactic fields can give stronger growth of @xmath114 with a significant increase of @xmath106 .
indeed , from the studies of rydberg atoms in a microwave field it is known that time - oscillating space homogeneous fields can produce strong variation of eccentricity of orbits ( see ref . and fig . 11 there ) .
also it is known that noise generates penetration of chaotic trajectories inside integrable islands and very slow decay of poincar recurrences with diverging trapping time ( see ref . ) .
in addition to that recent large scale numerical simulations of ref . show significant changes of eccentricity of planets on a time scale of the order of @xmath35 .
therefore the question about @xmath106 value for dmp captured by the ss requires further studies .
it is not excluded that it is comparable or even larger than @xmath35 . in such a situation
the upper bound ( [ de1 ] ) will be close to the real value of total captured mass . in any case
it is clear that there is practically no ejections of captured particles on a time scale of dmp orbital period @xmath115 .
a typical captured dmp rescaled energy is @xmath116 corresponding to one iteration of the map ( [ map ] ) which gives a change of @xmath94 from negative to positive values .
the rotation period of such dmp is rather large compared to the period @xmath117 of the planet : @xmath118 . for the case of jupiter
@xmath119 years and at @xmath46 we have @xmath120 years .
this is by a factor @xmath121 shorter than the ss life time @xmath35 .
this gives the lower bound of the captured dm mass which is obtained by replacing @xmath35 by @xmath115 in the equation ( [ de1 ] ) .
at last , let us consider the observational data on the dm in our ss . the most reliable and accurate information on it follows from the studies of the perihelion precession of venus , earth , and mars . under the assumptions that the dm density @xmath104 is distributed spherically symmetric with respect to the sun and that the eccentricity of the planetary orbit is small , the corresponding relative shift of the perihelion per period is ( see ref . and refs . therein ) : @xmath122 where @xmath69 is the radius of the orbit
. this relation gets almost obvious ( up to an overall numerical factor ) if one recalls that , in virtue of the gauss theorem , for a spherically symmetric density @xmath123 the action of the dm inside the orbit reduces to that of a point - like mass , and therefore does not induce the perihelion precession . on the other hand , for such density @xmath123 , the dm outside the orbit does not influence at all the motion of a planet .
the recent , most precise observational data ( see ref . ) on the precession of perihelia are presented in table 3 ( therein the theoretical values @xmath124 of the perihelion rotation and the results of observations @xmath125 are given in angular seconds per century ) . with these data ,
one arrives at the upper limits on the dm density at the distances from the sun , corresponding to the orbit radii of venus , earth , and mars , on the level of @xmath126 this observational upper limit exceeds by about two orders of magnitude the results ( almost certainly overestimated ) presented in table 2 .
our results do not mean , however , that the searches for the dark matter in the ss are senseless .
of course , the capture of the galactic dm analyzed here is not the only conceivable source of the dark matter in the solar system .
it is quite possible in particular that the solar system itself has arisen due to a local high - density fluctuation of the dark matter . now on the related theoretical problems . to obtain more firm results for the captured dm mass and density one needs to take into account that the kick function @xmath31 in the map ( [ map ] )
depends on an inclination angle between planes of dmp and planet orbits and also on the dmp perihelion distance .
however , a typical case of the comet halley analyzed by chirikov and vecheslavov in ref . gives the map function of a form similar to one discussed here so that the estimates presented should be also applicable for such more general dmp orbits .
further analytical and numerical studies are required for a better understanding of dmp dynamics inside the solar system .
we thank j. edsj and a.h.g .
peter for useful remarks and for pointing to us relevant works on wimp and dmp capture by the ss . the work of i.b .
khriplovich was supported in part by the russian foundation for basic research through grant no .
08 - 02 - 00960-a .
|
we study the capture of galactic dark matter by the solar system .
the effect is due to the gravitational three - body interaction between the sun , one of the planets , and a dark matter particle .
the analytical estimate for the capture cross - section is derived and the upper and lower bounds for the total mass of the captured dark matter particles are found .
the estimates for their density are less reliable .
the most optimistic of them give an enhancement of dark matter density by about three orders of magnitudes compared to its value in our galaxy .
however , even this optimistic value remains below the best present observational upper limits by about two orders of magnitude .
|
at low temperature the energy dissipation in the vortex cores is the main source of resistivity in the mixed state of type - ii superconductors@xcite .
if a supercurrent flows through a superconductor , it exerts a force on the vortices .
unless pinned by impurities or inhomogeneities , the vortices are brought into motion , which in turn leads to dissipation .
depending on the level of disorder , the vortices may move at different angles with respect to the direction of the supercurrent . at weak disorder ,
the vortices move together with the supercurrent ( `` ballistic limit '' , with the hall angle close to @xmath0 ) . at sufficiently strong disorder ,
the vortex motion is directed perpendicular to the supercurrent ( `` dissipative limit '' , small hall angle ) . both limits
are well understood within the quasiclassical description @xcite . a simplified approach describing
the dissipative limit goes back to the theory of bardeen and stephen treating the vortex as a region of normal phase inside a superconductor @xcite . in spite of neglecting the structure of the quasiparticle excitations in the vortex core @xcite , the bardeen
stephen theory gives the same result ( up to a numerical factor ) as the accurate quasiclassical calculation @xcite .
it has been recently suggested that the microscopic structure of the core excitations may play a much more prominent role in that part of the dissipative regime where the excitation spectrum remains discrete ( sharp quasiparticle levels ) , specifically in layered superconductors@xcite . in the clean limit ( i.e. , for scattering rates @xmath1 much smaller than the superconducting gap @xmath2 )
, the motion of quasiparticles in the vortex core is ballistic : they cross the vortex core many times before scattering off impurities .
therefore in this limit the spectral properties are sensitive to the details of disorder realization . in the superclean regime (
@xmath3 , where @xmath4 is the fermi energy ) the levels inside a two - dimensional vortex split into two sets ( ` combs ' ) of equally spaced levels @xcite .
the transformation of this correlated spectrum into a featureless uncorrelated one with increasing disorder proceeds in distinct steps : within the moderately clean regime ( @xmath5 ) a new intermediate region ( @xmath6 ) has been found @xcite where the comb structure remains preserved but is randomly shifted in energy ( the number of impurities in the core has to be small enough to preserve the ` combs ' , while being large enough to randomize the overall shift ) . increasing the impurity concentration further ( already in the moderately clean limit for weak impurities @xcite , and ultimately in the dirty limit @xcite ) , the ` comb ' structure is destroyed with a crossover to the class @xmath7 random - matrix ensemble@xcite . keeping within the moderately clean regime with a discrete spectrum , in two - dimensional ( or layered ) superconductors there further
exist two limits of dissipation @xcite : the first one applies to the slowly moving vortex with a discrete quasiparticle spectrum ( `` discrete - spectrum regime '' ) where the dissipation is due to landau zener transitions between individual levels @xcite .
the second limit is with levels smeared into a continuum either by the vortex motion or by inelastic processes ( `` continuum - spectrum regime '' ) ; the dissipation then is given by the linear - response kubo formula @xcite . at low temperatures ,
the inelastic smearing can be neglected and the crossover between these two regimes is controlled by the vortex velocity with the characteristic velocity given by @xmath8 ( here , @xmath2 is the superconducting gap and @xmath9 denotes the elastic mean free path@xcite ) . in the framework of random - matrix models with time - dependent hamiltonians ,
the dissipation in the discrete - spectrum and continuum - spectrum regimes was considered by wilkinson @xcite .
he finds that , in the case of the unitary wigner
dyson ensemble , the linear dissipative response remains valid in the whole range of velocities , both in the continuum - spectrum ( high velocity ) and in the discrete - spectrum ( low velocity ) regimes .
based on this fact and on the similarity between the unitary and class @xmath7 ensembles , it was shown that with the class @xmath7 level statistics the dissipation rate nearly follows the bardeen - stephen prediction , even in the limit of small velocities @xmath10 where the quasiclassical description is no longer valid @xcite .
contrary , it was claimed in ref . that the additional correlation between levels ( the ` two - comb ' structure ) may lead to an anomalously high vortex viscosity in the moderately clean regime in the low - velocity limit . in our paper
, we reconsider the problem of the vortex viscosity in a two - dimensional @xmath11-wave superconductor , taking full account of the discreteness of the vortex spectrum and of the microscopic structure of the quasiparticle levels .
we assume the moderately clean limit with an appropriate number of impurities in the core , such that the spectrum possesses the randomly shifted two - comb structure ( the precise condition is given in section [ circular - ensemble ] ) .
we find that in spite of the level correlations derived in refs . , the vortex viscosity does not differ from the well - known quasiclassical result @xcite for three - dimensional superconductors . within the same microscopic model we consider separately the discrete - spectrum and continuum - spectrum regimes of dissipation :
in both limits we arrive at the bardeen
stephen result for the dissipation , _ xx = ~_n , [ sigma - xx ] where @xmath12 is the electron density , @xmath13 is the spacing between the levels in the core@xcite , and @xmath14 is the magnetic field .
the two - comb level structure found in refs .
may be classified as the circular unitary random - matrix ensemble of dimension two@xcite .
therefore our findings may be considered as a generalization of wilkinson s results @xcite to circular ensembles .
we also find the scattering time @xmath15 for quasiparticles in the core in the @xmath11-wave approximation : for a weak impurity strength , the effective @xmath16 is larger than the bulk scattering rate @xmath17 by the logarithm of the impurity strength .
a similar logarithmic correction was previously derived in refs . .
the paper is organized as follows . in section [ section - hamiltonian ]
we prepare for the calculation by deriving the microscopic hamiltonian projected onto the relevant subgap states in the vortex core . in section [ circular - ensemble ]
we review the results of refs . on the circular random - matrix ensemble appearing in a disordered vortex at the intermediate level disorder .
we then describe the two limiting regimes of dissipation ( sect .
[ two - regimes ] ) , the discrete - spectrum and the continuum - spectrum regimes , and set up the stage for the calculations . in section [ section - discrete ]
we treat the case of the discrete energy spectrum ( low velocities and no inelastic level broadening ) , while section [ section - continuum ] is devoted to the opposite continuum - spectrum limit where the levels are broader than the interlevel distance .
finally , we discuss our findings in section [ section - discussion ] .
the sensitivity of the energy levels to the vortex displacement is calculated in an appendix .
before discussing the moving vortex , let us review the excitation spectrum of a clean two - dimensional vortex at rest .
the corresponding wave functions are given by the solutions to the bogoliubov
degennes equations h_0 & ( ) + ^ * ( ) & -h_0 u + v = e u + v , [ h0 ] where @xmath18 .
we assume an axially symmetric vortex with the order parameter @xmath19 , where the modulus of the order parameter depends only on the radial component @xmath20 and the phase winds with the angular coordinate @xmath21 .
we neglect the magnetic field in the vortex core , assuming a large magnetic penetration depth @xmath22 , where @xmath23 is the superconducting coherence length ( here and below we choose units with @xmath24 ) . in the quasiclassical limit @xmath25 , the spectrum and the eigenfunctions may be easily found@xcite .
the eigenvalues form a spectrum of equidistant levels e_=_0 , where the angular momentum @xmath26 takes half - integer values , @xmath27 , and _ 0 = , k(r ) = _
0^r dr ; @xmath20 is the distance from the core center .
the basic electronic energy scale in the vortex core takes the value @xmath28 , up to a possible logarithmic prefactor @xmath29 due to the shrinkage of the vortex core at small temperatures @xmath30 ( kramer pesch effect@xcite ) .
the eigenstates @xmath31 take the form @xmath32 where @xmath33^{-1 } \sim k_f/\xi$ ] is the normalization factor , and @xmath34 are the bessel functions . in the following
, we will be interested in processes at energies far below the superconducting gap and thus project all the operators onto the subgap states ( [ subgap - states - mu ] ) .
it will be convenient to take the fourier transform of these eigenvectors in the variable @xmath26 and introduce a new angular variable @xmath35 labelling the direction of the quasiclassical motion of the quasiparticle , _
( ) _ _ ( ) e^i = a e^i/2 + e^-i/2 e^-k(r ) e^i _ , [ subgap - states - phi ] where we introduced the vector @xmath36 pointing perpendicular to the direction specified by the angle @xmath35 , with absolute value @xmath37 .
the plane - wave exponent in ( [ subgap - states - phi ] ) is then @xmath38 .
this basis of wave functions has a very simple structure : in addition to the phase winding @xmath39 ( providing the antiperiodic boundary conditions in @xmath35 ) , these wave functions are plane waves in the direction of the wave vector @xmath40 restricted to a region of size @xmath41 around the vortex center . note that @xmath42 is not an eigenfunction of the hamiltonian ( [ h0 ] ) .
prepared in such a state at @xmath43 , the wave function will rotate in the @xmath35-basis according to @xmath44 .
we now turn to the problem of the moving vortex with impurities .
it can be described by the time - dependent bogoliubov
i = h(t ) , where @xmath45 is the hamiltonian of the vortex at the position @xmath46 with @xmath47 the vortex velocity , h(t ) = h_0+u ( ) & ( -t ) + ^*(-t ) & -h_0-u ( ) , [ moving - vortex - hamiltonian ] and @xmath48 is the potential set up by the impurities . in the clean limit
( @xmath49 ) , the admixture of bulk states ( with energies greater than @xmath2 ) to the vortex states may be neglected .
we project the hamiltonian ( [ moving - vortex - hamiltonian ] ) onto the subgap states of a clean vortex ( [ subgap - states - phi ] ) by substituting ( , t ) = _ 0 ^ 2 a(,t ) _ ( -t ) . defined in this way , the amplitude @xmath50 has antiperiodic boundary conditions in @xmath35 , @xmath51 . the time evolution of the coefficients @xmath50 obeys the schrdinger - type equation i a = -i_0 a + t_ a( ) + l_ a( ) , [ equation - of - motion ] where the kernel @xmath52 is produced by the vortex motion , and the kernel l_ = _ ( -t ) | u ( ) _ z | _ (-t ) is due to the impurities .
the @xmath30-kernel may be easily computed from the explicit form ( [ subgap - states - phi ] ) of @xmath53 . in the limit
@xmath25 , the matrix element ( [ t - kernel ] ) takes the form t_ = - 2(- ) _ and we identify this term with the `` doppler shift '' .
the impurity potential @xmath48 is taken as a sum over point - like impurities , u()= _
i v_i ( -_i ) .
then the scattering kernel may be expressed as @xcite l_ = 2i a^2 _
i v_i e^-2 k(|_i - t| ) e^i(_-_)(_i - t ) .
[ scattering - l ] summarizing , we end up with the equation of motion ( [ equation - of - motion ] ) containing three terms : the first term describes the circular ( `` chiral '' ) motion of the quasiparticle in the vortex@xcite .
the second term is the doppler shift due to the vortex motion .
and the third term describes the scattering off impurities ( taking point - like impurities is equivalent to including only @xmath11-wave scattering ) .
we are interested in the energy pumping in this time - dependent model .
note that the physical energy is given by the first and third terms in the evolution operator ( [ equation - of - motion ] ) but does not include the second term ( @xmath30-term ) which arises from the time - derivative of the basis wave functions .
this discrepancy between the energy and the evolution operator may be resolved by an appropriate _ time - dependent gauge transformation _ ( ) = a ( ) e^-i_t .
[ tilded - basis ] this gauge transformation has a dual effect on the equation of motion ( [ equation - of - motion ] ) : firstly , it makes the energy operator coincide with the evolution operator .
the old @xmath30-kernel is now replaced by a time - dependent one , _
= 2(- ) _ 0
( _ z _ ) t , [ tilded - t ] where @xmath54 is the unit vector perpendicular to the plane ( this gauge transformation resembles the one in electrodynamics replacing a static electric field with a magnetic field linearly growing in time ) .
secondly , the @xmath55-kernel is transformed as well and the new kernel takes the form _
= 2i a^2 _
i v_i e^-2 k(|_i - t| ) e^i(_-_)_i , [ tilded - l ] which differs from eq .
( [ scattering - l ] ) by the cancellation of the velocity term @xmath56 in the last exponent .
the physical content of the two gauges may be understood in the following way : the basis wave functions ( [ subgap - states - phi ] ) are quasiclassical plane waves
( at the wave vector @xmath37 ) cut off by the long - wavelength envelope @xmath57 of the size of order @xmath41 . in the original gauge [ with variables @xmath50 ] , this basis was chosen by simply translating the basis ( [ subgap - states - phi ] ) together with the vortex . in the new gauge [ with variables @xmath59 , only the long - wavelength envelope is translated , without shifting the phases of the quasiclassical plane waves .
this difference in matching phases of plane waves at different vortex positions produces two different descriptions of the moving vortex . in the first gauge [ variables @xmath50 ] ,
the evolution equation ( [ equation - of - motion ] ) contains impurities moving with respect to the vortex ( fast oscillations in @xmath60 ) .
this description is similar to the approach taken in refs . , a static vortex subject to moving impurities .
however , in those references , the @xmath30-term was omitted .
we will show below that omitting this term does not change the result for the dissipation in the moderately clean limit , thus justifying the approach of refs . .
in the second gauge [ variables @xmath59 , there are no oscillating terms in @xmath60 ( except for the slowly varying envelope @xmath61 whose time derivative may be neglected in most cases ) .
all the oscillations in the @xmath55-term in the first gauge may be removed by a single gauge transformation , as all impurities move with respect to the vortex with the same velocity and in the same direction .
this fact has not been properly taken into account in ref . , which has lead to an unphysical result .
we shall comment in more detail on the derivation of ref . in sec .
[ section - discrete ] ; here we just remark that the parallel motion of the impurities with respect to the vortex requires special care in the calculations within the first gauge , but is automatically taken into account in the second gauge .
in this section , we review the derivation of the two - comb spectral statistics in the vortex from refs . and set up the notation for the calculations in the later sections . for the discussion of the spectral statistics in this section we take the vortex at rest ( @xmath62 ) .
then the hamiltonian in the equation of motion ( [ equation - of - motion ] ) contains only two terms , the kinetic term @xmath63 and the scattering term @xmath60 ( for a vortex at rest there is no difference between @xmath60 and @xmath64 ) . at not very high impurity concentration
[ see eq .
( [ upper - bound ] ) below ] , the scattering kernel @xmath60 may be approximated as a sum of two local terms [ this approximation is due to the rapidly oscillating exponent in ( [ scattering - l ] ) ] , l_ = 4i_0 _ i=1^n [ j_i _ , _ i _ ,_i+ - j_i^ * _ , _ i+ _ ,_i ] , [ imp - potential - kl ] where @xmath65 is the number of impurities in the core , the parameters @xmath66 specify the angular positions of the impurities , and @xmath67 are @xmath68-functions smeared over the width @xmath69 and antiperiodically continued in @xmath70 and @xmath71 ( @xmath72 ) .
note that the regularization ( [ imp - potential - kl ] ) as a product of two smeared @xmath68-functions is important for evaluating the scattering matrix ( [ m - i ] ) below @xcite .
the effective strength of the @xmath73-th impurity is j_i = ~i_i e^-2k(r_i)+2ik_fr_i , [ j - i ] where @xmath74 is the born parameter of the impurity ( an additional imaginary unit compared to the notation in ref .
is due to the antiperiodic boundary conditions employed ) .
= 0.6 the matrix elements ( [ imp - potential - kl ] ) couple only angles with difference close to @xmath75 . introducing the two - component vector ( ) = a ( ) + a(+ ) , , [ psi - doublet ] the scattering becomes local in @xmath35 .
the individual scattering events in ( [ equation - of - motion ] ) then may be integrated separately and formulated in terms of a boundary condition @xcite ( _ i+0)=m_i ( _ i-0 ) , [ bc - imp ] where m_i = 1-|j_i|^2 & 2j_i + -2j_i^ * & 1-|j_i|^2 .
[ m - i ] the boundary condition for @xmath76 going around the half circle is ( 0)=-i_y ( ) .
[ bc - pi ] thus the energy levels are determined by the eigenvalues of the matrix @xmath77 , where m = m_n m_2 m_1 ,
n > > _ 1 > 0 .
[ m - def ] the full `` scattering matrix '' @xmath77 is a unitary matrix with the eigenvalues @xmath78 , so that the energy levels @xmath79 are solutions to the equation ( e/_0 ) = -i2 _ y m. [ kl - spectrum ] the evolution of the wave function @xmath76 is schematically shown in fig .
[ fig ] . in the moderately clean limit with a sufficiently large ( but not too large ) number of impurities in the core ,
the matrix @xmath80 ( and thus also @xmath77 ) is random with a uniform distribution over the @xmath81 group .
such a random - matrix ensemble is classified as the `` circular unitary '' ensemble with dimension two @xcite .
the density of states and the level correlations may be easily computed from eq .
( [ kl - spectrum ] ) .
for example , the average density of states is given by@xcite ( e ) = 2_0 ^2 ( ) .
[ kl - dos ] the spectrum consists of two combs of equidistant levels : a level is characterized by the comb number @xmath82 and its position in the comb , @xmath83 , where @xmath84 is the eigenvalue of the lowest level with positive energy ( @xmath85 ) .
the eigenfunctions can be written explicitly as _
, k ( ) = e^ie_,k /_0 m_,0 _ ( = 0 ) , [ psi - f ] where @xmath86 are the two eigenvectors of @xmath87 , and the scattering matrices @xmath88 are defined as m_, = m_i_n m_i_1 , >
_ i_n > > _
i_1 > , [ m - two - point ] where the product is taken over all impurities between the angles @xmath89 and @xmath35 .
( [ m - two - point ] ) introduces @xmath88 for @xmath90 ; furthermore , it is convenient to define the scattering matrix for @xmath91 as @xmath92 .
the symmetry of the @xmath81 matrices m_,^ * = _ y m_, _ y [ m - symmetry ] allows us to translate easily between the eigenfunctions for the two series of levels via _ -,-k ( ) = _ y _ + , k^ * ( ) .
[ psi - symmetry ] in our calculations , we shall need not only the properties of the spectrum , but also some statistical properties of the wave functions in the regime of the circular unitary ensemble . at any given point @xmath35 and for any energy level
, the wave function @xmath76 is a spinor pointing in a random direction .
all directions are equally probable , allowing us to compute equal - point correlation functions such as [ ff ] @xmath93 the correlations between wave functions at different values of @xmath35 may be expressed in terms of the properties of the scattering matrices ( [ m - two - point ] ) .
the angular correlations between scattering matrices decay exponentially , m^_, _ z m_, = e^-|-|/ _ _ z , [ wf - decay ] where = 16n_imp _ 0^rdr , [ phitau ] and @xmath94 is the impurity concentration . in ( [ phitau ] ) we have assumed ( for simplicity ) that all impurities have equal born parameters , in which case the effective impurity strength @xmath95 as given by ( [ j - i ] ) becomes a function of @xmath96 and is denoted @xmath97 in ( [ phitau ] ) . in order to derive ( [ wf - decay ] ) and ( [ phitau ] )
one takes a small increment of @xmath98 in @xmath35 and averages over a single impurity scattering matrix , @xmath99 .\end{gathered}\ ] ] the last averaging is performed independently over @xmath88 and @xmath100 . in averaging ,
we first account for the rapidly oscillating off - diagonal elements in ( [ m - i ] ) and arrive at ( [ wf - decay ] ) , ( [ phitau ] ) . for strong impurities with the born parameter @xmath101 ( as above
, we assume that all impurities are of equal strength ) , the scattering time @xmath15 is of the order of the bulk normal - state elastic mean - free time @xmath102 . for weak impurities with @xmath103 , the integral in eq .
( [ phitau ] ) has a logarithmic dependence on the born parameter@xcite , 1~1_n 1 .
[ 1/tau ] a similar logarithmic behavior of the scattering rate was also derived in refs . .
the relaxation angle @xmath104 in ( [ phitau ] ) plays a central role in the context of vortex dissipation , as it enters the expression for the friction coefficient and hence also for the flux flow conductivity . in ref .
this quantity entered the _ numerator _ in the expression for the conductivity , while in our results it enters the _ denominator _ in the bardeen
stephen form ( [ sigma - xx ] ) .
we comment in detail on our disagreement with ref . in section [ section - discrete ] and in the appendix .
the physical meaning of @xmath104 is the correlation length of the scattering matrix @xmath88 and @xmath15 is the corresponding scattering time .
we shall see later that it is the correlation function ( [ wf - decay ] ) which determines the rate of interlevel transitions . in principle , one can derive a similar exponential decay `` in the @xmath105 and @xmath106 directions '' [ with @xmath107 replaced by @xmath108 or @xmath109 in ( [ wf - decay ] ) ] , in which case the correlation length @xmath110 is twice larger than @xmath104 `` in the @xmath111 direction '' .
this is due to the specific form ( [ m - i ] ) of the scattering matrix @xmath100 . in order to realize the regime of the circular unitary random - matrix ensemble it is necessary that @xmath112 , which defines the lower bound for the moderately clean regime , @xmath113
however , the region of the circular unitary ensemble does not extend over the whole moderately clean regime @xcite .
the additional restriction originates from the breaking of the instant - scattering approximation ( [ imp - potential - kl ] ) in the limit of too strong disorder .
indeed , the typical width of the smeared @xmath114-functions in this equation can be estimated as @xmath115 ( that corresponds to an impurity at the distance of the order of @xmath41 from the vortex center ) .
on the other hand , the number of impurities in the core is @xmath116 .
the solution for the spectrum discussed above is justified as long as the @xmath114-functions in eq .
( [ imp - potential - kl ] ) do not overlap , which is equivalent to the condition @xmath117 .
thus , the circular unitary ensemble is realized only in a relatively narrow range of disorder strengths 1_n ^2 [ upper - bound ] [ we assume that the born parameter @xmath103 and make use of eq .
( [ 1/tau ] ) ; for @xmath118 one should replace @xmath119 by 1 ] . for stronger disorder ( i.e. , larger scattering rate @xmath17 ) or smaller born parameter @xmath120 the instant - scattering approximation ( [ imp - potential - kl ] ) fails and the circular unitary ensemble crosses over to the class @xmath7 ensemble @xcite . a phenomenological approach to this crossover has been discussed in ref .
and the results of numerical simulations are available in ref . .
the time evolution of the core states is described by eq .
( [ equation - of - motion ] ) .
since the evolution operator is nonstationary ( i.e. , explicitly time - dependent ) , it causes transitions between levels with different energy . in a fermionic system ,
the rate of downward transitions is suppressed compared to the rate of upward transitions due to the pauli exclusion principle , leading to an increase of the average energy with time .
the energy pumped into the system will finally be transferred to the thermal bath via the interaction with phonons or other soft degrees of freedom , thus producing a finite dissipation .
there are two different mechanisms of dissipation depending on whether the individual energy levels can be resolved or not @xcite .
if the discrete spectrum is smeared into a continuous one , the energy pumping can be calculated with the help of the standard linear - response kubo formula @xcite .
this regime is naturally realized when the _ inelastic width _ @xmath121 of an energy level exceeds the mean level spacing @xmath122 . on the other hand
, the spectrum may turn out effectively continuous even at @xmath123 if the time dependence of the evolution operator on the right - hand side of eq .
( [ equation - of - motion ] ) is so fast that it destroys the instantaneous adiabatic spectrum @xcite . in this case , the frequency @xmath124 of perturbations due to the nonadiabaticity of the spectrum exceeds @xmath122 and plays the role of the effective level width . in the opposite case , when @xmath125 , the spectrum is essentially discrete and the dissipation is due to rare landau zener transitions @xcite taking place when two levels come very close to each other . in a normal metal at low temperatures ,
the inelastic widths due to the electron - electron and electron - phonon interactions are given by @xmath126 and @xmath127 , respectively@xcite ( @xmath128 is the debye temperature ) .
since @xmath129 is of the order of @xmath122 already at @xmath130 , it does not contribute to the level smearing at lower temperatures .
furthermore , for the states localized in the vortex core the interaction with phonons appears to be strongly suppressed compared to the normal - state rate @xcite ; one of the main reasons is that the quasiparticles with energy @xmath131 are composed of nearly equal mixtures of electron and hole components [ i.e. , @xmath132 , cf .
( [ subgap - states - mu ] ) ] with negligible net charge .
thus , at sufficiently small temperatures @xmath133 , the inelastic width of the core states @xmath134 and the regime of dissipation is determined solely by the vortex velocity . in ref .
the crossover velocity @xmath135 separating the regimes with discrete and continuous spectra was estimated as v_k ~ [ v_k ] and we will present a microscopic derivation of this result below . for model hamiltonians from the three wigner - dyson random - matrix ensembles @xcite ,
dissipation was calculated by wilkinson @xcite in the regimes of small ( @xmath136 ) and large ( @xmath137 ) velocities ( in the limit @xmath138 ) .
he finds that if the hamiltonian @xmath139 of the system depends on time through @xmath140 then the energy dissipation is determined by the variance of the ( off - diagonal , @xmath141 ) matrix elements @xmath142 normalized by the mean level spacing @xmath122 . in the continuous - spectrum regime specified by @xmath143 , the energy dissipation rate as given by the kubo formula , w_kubo = c(0)v^2
, [ w - kubo ] is the same for the orthogonal ( @xmath144 ) , unitary ( @xmath145 ) , and symplectic ( @xmath146 ) ensembles and describes viscous damping . in the discrete - spectrum regime , at @xmath147 ,
the result depends on the level - repulsion parameter @xmath148 , which determines the probability @xmath149 to find two levels at a distance @xmath150 .
in this situation the energy dissipation rate is given by the expression @xcite w_v^(+2)/2 , [ w - zener ] hence the dissipation is superohmic for the gaussian orthogonal ensemble , while for the gaussian unitary ensemble it remains ohmic with @xmath151 , exactly coinciding with @xmath152 despite a very different mechanism of dissipation . in the next sections we calculate the dissipation rate for the vortex motion in the regime of the koulakov
larkin ` two - comb ' spectrum @xcite , both in the limits of small and large velocities , thus extending wilkinson s considerations to the case of circular unitary ensembles .
in this section we calculate the dissipation for the two - comb level structure described in section [ circular - ensemble ] at small velocities .
we will work in the ` tilded ' basis @xmath153 introduced in eq .
( [ tilded - basis ] ) , where the energy coincides with the expectation value of the evolution operator ( [ equation - of - motion ] ) , with @xmath154 and @xmath60 substituted by their tilded counterparts ( [ tilded - t ] ) and ( [ tilded - l ] ) . in the low - velocity limit
, the energy is pumped into the system when two levels come very close to each other and a nonadiabatic landau
zener transition becomes possible .
due to the symmetry of the two - comb spectrum the rate @xmath155 of such transitions is the same for each neighboring pair of levels .
we will calculate @xmath155 by considering the lowest level with positive energy @xmath156 and its mirror image with energy @xmath157 . for simplicity
we assume that the vortex moves in the @xmath105 direction , @xmath158 . following the logic of ref .
, we diagonalize the hamiltonian at @xmath43 and restrict it to the @xmath159 matrix involving the pair of states considered , h(t ) = + h_11 t & h_12 t + h_12^ * t & -- h_11 t , where @xmath160 are the elements of the matrix @xmath161 ( neglecting the quadratic term @xmath162 is justified in the landau zener regime since the duration of a nonadiabatic transition is proportional to the gap of the avoided crossing and is small at @xmath147 ) .
the instantaneous adiabatic spectrum takes the form e(t ) = , [ ezener ] where @xmath163 equation ( [ ezener ] ) describes an avoided crossing with the minimal distance ( landau zener gap ) @xmath164 between the spectral branches realized at @xmath165 .
the probability of the landau
zener transition at such a crossing is @xmath166 .
the mean rate of transitions is given by r_lz = ( t_lz ) ( - ) , [ r - zener ] where the role of @xmath167 is to count each avoided crossing once .
the average in eq .
( [ r - zener ] ) is taken over the distribution of the parameters @xmath168 , @xmath169 , and @xmath170 describing the avoided crossing .
the energy @xmath168 is expressed through the transfer matrix @xmath80 according to eq .
( [ kl - spectrum ] ) .
the coefficients @xmath169 and @xmath170 are the matrix elements of the perturbation ( [ tilded - t ] ) over the exact wave functions @xmath171 [ which depend on the trajectory @xmath172 via eq .
( [ psi - f ] ) ] , @xmath173 we come to the crucial point : the quantities @xmath168 , @xmath169 , and @xmath170 have different dependencies on the transfer matrix @xmath172 . in the moderately clean limit ,
when the number of impurities in the core is sufficiently large and @xmath174 , the matrix @xmath172 performs many rotations over the @xmath81 group .
therefore , we conclude that ( i ) @xmath168 , @xmath169 , and @xmath170 are uncorrelated , and ( ii ) the distribution of @xmath169 and @xmath170 is gaussian . we further calculate the variances of @xmath169 and @xmath170 . for @xmath175
we obtain with the help of eq .
( [ wf - decay ] ) , @xmath176 integrating over @xmath177 and averaging over the spinor @xmath76 according to ( [ ff1 ] ) we find = ( _ + ^_z _ + ) ^2 = .
analogously , with the help of eq .
( [ ff2 ] ) we obtain = |_+^_x _ + ^*|^2 = . in the same way it can be shown that @xmath178 , thus proving the statistical independence of @xmath169 and @xmath170 .
we conclude that @xmath169 , @xmath179 , and @xmath180 are independently distributed with the same distribution ( h ) = 1_0k_fv ( - ) .
[ p(h ) ] the distribution function for @xmath168 is provided by the density of states ( [ kl - dos ] ) , in the limit @xmath181 , @xmath182 .
averaging eq .
( [ r - zener ] ) over @xmath168 , @xmath169 , and @xmath170 with the distributions @xmath183 and @xmath184 , we obtain for the mean rate of landau zener transitions r_lz = .
the energy dissipation rate is given by @xmath185 with the vortex viscosity = 12 k_f^2_= 12 k_f^2_0 .
[ eta - low - v ] as defined by eq .
( [ eta - low - v ] ) , @xmath186 is a two - dimensional viscosity . in a layered superconductor
, it determines the friction force @xmath187 exerted on the vortex from excitation of quasiparticles within one layer .
the statistics ( [ p(h ) ] ) of the matrix elements of @xmath188 determines the sensitivity of the spectrum to the vortex motion and allows one to find the critical velocity @xmath135 separating the regimes of discrete and continuum spectra .
the discrete spectrum can be resolved if the change of the hamiltonian during the time @xmath189 , @xmath190 , is smaller than @xmath122 .
taking for @xmath191 the dispersion of the distribution ( [ p(h ) ] ) , we find for the critical vortex velocity v_k ~ , [ v_k2 ] coinciding with the estimate ( [ v_k ] ) found in ref . . comparing with ref .
, we find that our result ( [ eta - low - v ] ) for the viscosity is smaller by a factor @xmath192 and the expression ( [ v_k2 ] ) for @xmath135 is larger by the factor @xmath193 . in ref .
, following the analysis in refs .
, it was assumed that the dissipation is produced by impurities moving with respect to the vortex , equivalent to neglecting the doppler - shift term @xmath154 in eq .
( [ equation - of - motion ] ) . in order to identify the origin of the discrepancy in the results and verify the validity of omitting the doppler shift , we reconsider below the treatment of ref .
we will see that the source of the disagreement is not in the neglect of the doppler - shift term but in the neglect of cross - correlations between the motion of different impurities . to demonstrate this
, we adopt the approach of ref .
( omitting the doppler - shift term ) and re - calculate the correlator of the spatial gradients of the matrix @xmath80 @xcite = 14 , [ eta - kl ] from which the viscosity follows via @xmath194 .
comparison between eqs .
( [ eta - kl ] ) and ( [ w - kubo ] ) shows that the quantity @xmath195 plays the role of the correlation function @xmath196 .
the matrix @xmath80 specified in eq .
( [ m - def ] ) is a product of transfer matrices of individual impurities .
therefore , there are two contributions to the correlator ( [ eta - kl ] ) , one originating from derivatives over the coordinates of the same impurity ( diagonal part ) and a second one from derivatives taken on different impurities ( off - diagonal part ) , _ diag = 14 _ i , + c_off - diag = 14 _ ij .
[ coffdiag - def ] the diagonal part is given by @xcite _
[ cdiag ] the cross term , expressing the correlated nature of the motion of the impurities with respect to the vortex , has been missed in ref . .
its calculation is presented in the appendix and the result takes the form _ off - diag = - .
[ coffdiag ] we thus see that the two contributions @xmath197 and @xmath198 nearly cancel each other and the net sensitivity of the spectrum to the vortex motion appears to be significantly lower than that found in ref . , = , where we employed the condition @xmath174 of the moderately clean limit . using eq .
( [ eta - kl ] ) one recovers the result ( [ eta - low - v ] ) for the vortex viscosity .
the above consideration not only corrects the result of ref . ; in addition , it serves as a microscopic justification of the model adopted in refs . , where the dissipation is due to impurities moving through the core and the doppler - shift term @xmath30 in eq .
( [ equation - of - motion ] ) is neglected .
if the vortex velocity @xmath199 exceeds @xmath135 the discrete spectrum is smeared and transitions between non - nearest levels become possible . in this limit , the energy dissipation can be calculated with the help of the kubo formula ; as in sec .
[ section - discrete ] we will use the ` tilded ' basis ( [ tilded - basis ] ) . neglecting the slow time dependence of @xmath200 , we write the evolution operator as @xmath201 , where @xmath202 is the hamiltonian of the vortex at rest and @xmath203 takes the form i_ = 2(- ) i ( ) , i ( ) = _ 0 k_f v , [ i ] where we again assumed that @xmath158 . the energy dissipation rate w = t | h(t ) | t is calculated as the linear response to the term @xmath204 with the help of the kubo formula w = i _
-^0 t dt 0 | [ i(t),i(0 ) ] | 0 , where ( t)=_0 ^ 2 i ( ) ^(,t ) ( , t ) and @xmath205 is a second - quantized operator in the interaction representation .
the latter can be rewritten in terms of the ( exact disorder dependent ) eigenfunctions @xmath206 and eigenvalues @xmath207 of the hamiltonian @xmath202 , ( , t ) = _ n c_n a_n ( ) e^-ie_nt , with @xmath208 the corresponding annihilation operators .
averaging over the initial state of the vortex at rest ( so that @xmath209 coincides with @xmath210 ) one arrives at the expression w = i
_ -^0 t dt i ( ) i( ) _ km ( n_k - n_m ) + a_k ( ) a_k^*( ) a_m( ) a_m^ * ( ) e^i(e_k - e_m)(t-t ) , where @xmath211 is the distribution function at the energy @xmath212 .
this expression can readily be represented ( cf .
, e.g. , ref . ) in terms of the green functions g^r , a_e(, ) = _ n , w = 12 i ( ) i( ) + [ g^r_e()-g^a_e( ) ] [ g^r_e()-g^a_e( ) ] .
[ w - gg ] the above formula for the dissipation rate is not restricted to the regime of the circular unitary ensemble but is valid in the whole moderately clean limit regime as long as @xmath137 .
in particular , one can easily recover the result of the standard quasiclassical analysis by assuming the @xmath15-approximation : within this approach the green function is diagonal in the momentum representation , g^r , a_e ( ) = , and evaluating the integral in ( [ w - gg ] ) one finds the energy dissipation rate @xmath213 with the vortex viscosity coefficient _ = 12 k_f^2 _ 0 . [ eta - tau ]
the @xmath15-approximation is the simplest approach to the problem where all spectral correlations are neglected .
below , we will calculate the vortex viscosity for the case of the koulakov - larkin two - comb statistics and check how the presence of the two - comb correlations in the energy levels modifies the result ( [ eta - tau ] ) . to this end
we rewrite eq .
( [ w - gg ] ) in terms of the wave functions @xmath214 introduced in ( [ psi - f ] ) , @xmath215 where @xmath216 .
we first re - express the wave functions @xmath214 via ( [ psi - f ] ) and perform the summation over @xmath217 and @xmath218 according to the summation formula _ k
e^i e_skt = e^i e_s,0 t _ m=-^(t - m /_0 )
( which produces a double infinite sum of @xmath68-functions of @xmath219 and @xmath220 ) .
we cut off the integrals over @xmath219 and @xmath220 at some time scale smaller than @xmath221 such a cut - off is equivalent to assuming a smearing of the energy levels with widths larger than @xmath122 . under this assumption , only one term of the double infinite sum survives with @xmath222 .
next , the integrations over @xmath219 , @xmath220 , and @xmath79 can be trivially performed .
finally , we sum over @xmath11 and @xmath223 and arrive at @xmath224 this average is calculated with the help of eq .
( [ wf - decay ] ) yielding @xmath225 the form of the above result coincides with ( [ eta - tau ] ) calculated within the @xmath15-approximation . the microscopic expression for the elastic relaxation time @xmath15
is given by eq .
( [ phitau ] ) . in case
the inelastic relaxation time is shorter than @xmath15 , it substitutes the latter in eq .
( [ eta - high - v ] ) . to conclude this section we mention that the same result ( [ eta - high - v ] ) can be obtained in the model where the doppler - shift term @xmath154 in eq .
( [ equation - of - motion ] ) is neglected and the dissipation is due to the motion of impurities with respect to the vortex . within that model
the energy dissipation rate is calculated as a linear response to a change in the impurity positions .
the derivation formally repeats the one presented above but with the operator @xmath226 . performing the same manipulations leading to eqs .
( [ w - kubo-1 ] ) and ( [ eta - high - v ] ) we obtain = , [ w - kubo-4 ] where @xmath80 is the product of scattering matrices defined in ( [ m - def ] ) .
the above correlation function for the random @xmath81 matrix @xmath80 can be related to the correlation function ( [ eta - kl ] ) and is equal to @xmath227 .
therefore , eq . ( [ w - kubo-4 ] ) exactly coincides with its low - velocity analogue ( [ eta - kl ] ) and reproduces the result ( [ eta - high - v ] ) .
the main conclusion of this paper is that the bardeen - stephen expression for the flux flow conductivity is extremely insensitive to the details of the level correlations in the vortex core .
we calculated the energy dissipation rate for the case of the koulakov larkin two - comb level structure when the spectral correlations are most pronounced .
such a statistics , classified as the circular unitary ensemble of dimension two , is obtained for layered superconductors within the region of the moderately clean limit specified by eq .
( [ upper - bound ] ) .
we found that the vortex viscosity is the same for the regimes of discrete ( @xmath10 ) and continuum ( @xmath228 ) spectra and coincides with the result ( [ eta - tau ] ) of the phenomenological @xmath15-approximation .
the viscosity @xmath186 determines the flux - flow dissipative conductivity via the standard relation @xmath229 . assuming a cylindrical fermi surface , we arrive at the result ( [ sigma - xx ] ) obtained previously via several quasiclassical approaches @xcite . despite we found the same result for @xmath230 in the limits of small and large vortex velocities
, it is worth emphasizing that the physics of energy dissipation is quite different in the two limits . in the kubo regime ( @xmath228 )
the energy is pumped continuously , whereas in the landau - zener regime ( @xmath10 ) it is absorbed in discrete portions of size @xmath122 .
the equivalence between the @xmath15-approximation and the result of the exact microscopic treatment is not very surprising in the kubo regime , as in this limit it is determined by the net sensitivity of the spectrum to the vortex displacement rather than by interlevel correlations . on the other hand , in the landau - zener regime , this equivalence is a matter of coincidence : it relies on the fact that the vortex hamiltonian belongs to the circular _ unitary _ universality class characterized by a level repulsion parameter @xmath145 .
the other outcome of the present work is that it provides a justification for neglecting the doppler - shift @xmath30-term , eq .
( [ t - kernel ] ) , that was implicitly assumed in earlier papers @xcite . without the @xmath30-term , the system is equivalent to a vortex at rest with impurities moving through its core .
the dissipation is then related to the change of the level positions due to the motion of the impurities . on the other hand
, the @xmath30-term describes the electric field in the core , which is the source of energy dissipation in the bardeen - stephen model omitting this term is a bit confusing .
nevertheless , our analysis indicates that the vortex viscosities calculated _ with _ and _ without _ this term coincide in the moderately clean limit .
one may expect , however , that a correct treatment of this term is crucial in the calculation of the hall conductivity and in the superclean limit . in this paper we considered a strictly two - dimensional superconductor ( or layered superconductor with negligible interlayer coupling ) . in the case of the koulakov
larkin circular unitary ensemble statistics , the effect of interlayer hopping is small provided that the effective quasiparticle temperature in the core , @xmath231 , is smaller than @xmath232 , where @xmath233 is the effective mass anisotropy .
this condition ensures that the interlayer hopping amplitude is always smaller than @xmath122 and can be neglected .
this behavior is to be contrasted with the one for the class @xmath7 random - matrix statistics realized in the moderately clean case for weak impurities ( @xmath234 ) and in the dirty ( @xmath235 ) limit . in that case
, the interlayer coupling leads to tunneling from the @xmath12-th level in one layer to the @xmath236-th level ( with @xmath237 ) in the adjacent layer , thereby opening an interlayer channel for energy dissipation .
the competition between the intralayer and interlayer channels may lead to a strong deviation from the bardeen
stephen formula ( [ sigma - xx ] ) and even to hysteretic behavior of the current - voltage curve@xcite . on the contrary ,
the two - comb spectrum is absolutely rigid : the @xmath12-th level in one layer can have an avoided crossing only with the @xmath12-th level in the adjacent layer and the tunneling between these two levels does not result in dissipation .
we thank m. v. feigelman for many useful discussions .
this research was supported by the scopes program of switzerland , the dutch organization for fundamental research ( nwo ) , the russian foundation for basic research under grant 01 - 02 - 17759 , the program `` quantum macrophysics '' of the russian academy of sciences , the russian ministry of science , the russian science support foundation ( m. a. s. ) , and the swiss national foundation .
m. a. s. thanks eth zrich for hospitality .
here , we calculate the off - diagonal correlation function @xmath238 defined in eq .
( [ coffdiag - def ] ) . to this end
we divide the angle interval @xmath239 $ ] into many small pieces @xmath240 $ ] of width @xmath241 so that each piece contains one impurity at maximum .
the transfer matrix of the @xmath217-th interval is hence either @xmath242 ( no impurities ) or @xmath243 ( if the angle of the @xmath73-th impurity @xmath244 $ ] ) .
then @xmath245 where @xmath246 and y(k ) = m^_^(k),^(k-1 ) , y(p ) = m^_^(p),^(p-1 ) .
the representation ( [ a1 ] ) is suitable for averaging over disorder , since one can independently average the matrices @xmath247 , @xmath248 , @xmath249 , and @xmath250 .
the statistical independence of @xmath251 , @xmath252 , and @xmath250 follows from the fact that the intervals @xmath240 $ ] , @xmath253 $ ] , and @xmath254 $ ] do not overlap .
as we will see below , the correlator ( [ a1 ] ) is essentially nonzero at @xmath255 .
the matrix @xmath247 , which couples @xmath256 and @xmath256 through the point @xmath257 , is the product of a large number of matrices @xmath100 and looses all correlations within the interval @xmath254 $ ] . in calculating @xmath258 and @xmath259 with the help of eqs .
( [ m - i ] ) and ( [ j - i ] ) only the fast phase of @xmath95 should be differentiated , y(k ) = -y(k ) = i _ z 2k_f ^(k ) _ 0 ^ 2nrdr = i _ z k_f
^(k ) , with the last relation following from the definition ( [ phitau ] ) of the angle coherence length @xmath260 . taking the continuum limit @xmath261 we obtain c_off - diag = _ 0^d_1 d_2 _ 1 _ 2 . averaging over @xmath262 according to eq .
( [ wf - decay ] ) and integrating over @xmath263 we find c_off - diag = _ 0^d e^-/ _ , where the upper limit is substituted by infinity due to the fast convergence of the integral .
finally , averaging over @xmath247 uniformly distributed over the @xmath81 group and evaluating the remaining integral , we arrive at eq .
( [ coffdiag ] ) .
|
we present a microscopic calculation of the energy dissipation in the core of a vortex moving in a two - dimensional or layered superconductor in the moderately clean regime . in this regime ,
the quasiclassical bardeen
stephen result remains valid in spite of the strong correlations between the energy levels .
we find that the quasiclassical expression applies both in the limit of fast vortex motion ( with transitions between smeared levels ) and in the limit of slow vortex motion ( with nearly adiabatic dynamics ) .
this finding can be related to the similar result known for the unitary random - matrix model .
|
the field of non - linear dynamics introduced the fascinating idea that an apparently random behavior of a time series might have been generated by a low dimensional deterministic system @xcite .
based on the notions of chaos theory , different algorithms have been invented to infer if an observed time series is a realization of a chaotic system , e.g. the estimation of the largest lyapunov - exponent @xcite , the correlation dimension @xcite and nonlinear prediction @xcite .
there is hope to gain deeper insights in complex systems like those from biology and physiology by applying these methods .
however , the application of these methods to a finite , often noisy set of measured data is not straightforward , see e.g. @xcite and references therein .
for example , in order to claim a finite , fractal correlation dimension , a scaling region of sufficient length has to be established .
determining this scaling region by eye or some algorithm may lead to an erroneous evidence of chaotic behavior . in order to evaluate the analysis ,
it has become popular to apply the method of surrogate data @xcite .
therefore , data are generated which have the same linear statistical properties as the original data but not the possible nonlinear ones .
for many realizations of these data , the same algorithm as to the original data is applied . a significant difference between the distribution of the nonlinear feature for the surrogate data and
the original data is taken as an indication that the process underlying the original data is deterministic @xcite , nonlinear @xcite or even chaotic @xcite .
the explicit null hypotheses of surrogate data testing for linearity is that the data were generated by a linear , stochastic , gaussian stationary process , including a possible invertible nonlinear observation function .
thus , a rejection of this hypothesis does not necessarily mean that the data come from a chaotic , i.e. some kind of stationary , nonlinear deterministic , process .
they might also originate from a nonlinear stochastic or even simply from a linear , stochastic , non - stationary process . in this paper , we investigate the power of surrogate data testing against non - stationarity . as nonlinear feature
we use the correlation dimension .
the behavior of correlation dimension estimates has been investigated for the @xmath0 , @xmath1 type of linear non - stationarity @xcite . for physiological data ,
such @xmath2 behavior has been observed in heart rate @xcite .
often , physiological data are characterized by some kind of oscillatory behavior like eeg , hormone secretion , breathing or tremor . for such data , types of non - stationarity introducing some time dependency of the oscillating dynamics ,
e.g a modulation of frequency or amplitude , seems to be a natural violation of the null hypothesis . if the process is linear and the time dependency of the parameters , and thus , the autocovariance function is periodically in time , these processes are called cyclostationary @xcite . many other types of non - stationarity in oscillatory processes are imaginable .
we choose cyclostationary processes because they allow in simple way for a parametric violation of the null hypothesis .
formally , these processes can be expressed as higher dimensional autonomous non - linear stochastic processes .
a special version of surrogate data testing acting on segments of the data has been suggested to analyze such data @xcite . in the next section
, we informally discuss the class of cyclostationary processes and introduce the two specific examples we use in section iii to investigate the power of surrogate data testing with respect to these types of non - stationarity .
the parameters @xmath3 and @xmath4 of a linear stochastic autoregressive ( ar ) process @xmath5 : [ ar_process ] x(t ) = _
i=1^p a_1 x(t - p ) + ( t ) , ( t ) ~(0,^2 ) determine the autocovariance function @xmath6 : r ( ) = < x(t ) x(t+ ) > .
the spectrum @xmath7 is given as fourier transform of the autocovariance function : s ( ) = e^-i r ( ) . a possible first step to non - stationarity is to define a time dependent spectrum @xmath8 and , correspondingly , a time dependent autocovariance function @xmath9 : r(t , ) = < x(t ) x(t+ ) > .
a cyclostationary process of periodicity @xmath10 is defined by : [ cyclo_def ] r(t , ) = r(t+l , ) . for the ar process of eq .
( [ ar_process ] ) this means that the parameters @xmath3 and @xmath4 may change periodically . as process satisfying the null hypothesis of surrogate data testing for linearity , we chose an autoregressive ( ar ) process of order two : x_t = a_1 x_t-1 + a_2 x_t-2 + _ t , _
t ~(0,^2 ) . in terms of physics ,
ar processes can be interpreted as a combination of linear relaxators and linear damped oscillators driven by noise . for an ar process of order two which describes a damped oscillator ,
the parameters are related to the relaxation time @xmath11 and period @xmath12 by : @xmath13 the variance of the process var(@xmath14 ) is given by : [ var_ar2 ] ( x_t)= .
we choose an ar2 process with @xmath15 , @xmath16 and @xmath17 as process @xmath18 that satisfies the null hypothesis .
[ ar2_dat]a displays a realization of this process .
= 8.4 cm = 8.4 cm = 8.4 cm the oscillatory behavior with a mean period of 10 time steps is clearly visible as well as the natural variability of period and amplitude .
[ ar2_spe ] ( solid line ) shows the estimated spectrum of the process .
the spectrum was estimated by averaging 100 periodograms , i.e. the squared absolute value of the fourier transform of the data .
= 8.4 cm a broad peak , typical for a stochastically driven linear damped oscillator can be seen . based on eqs .
( [ a1],[a2],[var_ar2 ] ) we now introduce two parameterized violations of this stationary , linear , stochastic process in order to investigate the power of surrogate data testing with respect to non - stationarity . for the first violation of stationarity in the frame of cyclostationary processes , we choose a simple amplitude modulation , corresponding by eq.([var_ar2 ] ) to a periodicity of the variance of the driving noise . based on the stationary ar2 process @xmath18 ,
the amplitude modulated process @xmath19 is given by : [ process1 ] x_amp(t ) = ( 1 + _ ( 2/t_mod t ) ) x_0(t ) .
@xmath20 , the modulation depth , parameterizes the violation of the null hypothesis .
@xmath21 determines the modulation period .
[ ar2_dat]b displays a realization of this process with @xmath22 and @xmath23 for three periods of the modulation .
compared to fig .
[ ar2_dat]a , the non - stationarity is hardly visible . due to the long modulation period compared to the period of the process ,
its spectrum is not distinguishable from that of the stationary process in fig .
[ ar2_spe ] . for the second violation of stationarity
, we chose a modulation of the period @xmath12 of the ar2 process with period @xmath21 and amplitude @xmath24 around the mean period @xmath25 .
this leads to a time dependency of the parameter @xmath26 of the ar2 process : @xmath27 @xmath28 parameterizes the violation of the null hypothesis .
according to eq .
( [ var_ar2 ] ) , the time dependency of @xmath29 causes a time dependency of the variance of the process .
the effect of a changing variance is already covered by the first process , eq .
( [ process1 ] ) . to investigate only the effect of a changing period of the process here , we use eq .
( [ var_ar2 ] ) to adjust the variance @xmath30 of the driving noise such that the variance of the process is constant : @xmath31 where @xmath32 and @xmath4 denote the parameters of the process @xmath33 satisfying the null hypothesis .
[ ar2_dat]c displays a realization of this process with @xmath34 and @xmath35 .
again , compared to fig .
[ ar2_dat]a , the non - stationarity is hardly visible .
[ ar2_spe ] ( dashed line ) shows the estimated spectrum of the process .
the spectrum shows two peaks at the corresponding frequencies due to the specific type of modulation chosen .
as nonlinear feature to investigate the power of surrogate data testing against the two violations of stationarity we use the correlation dimension .
the phase space is reconstructed by delay embedding .
the delay is chosen equal to the lag at which the autocorrelation function first crosses zero .
the correlation dimension @xmath36 is defined by : @xmath37 where c(r ) , the correlation integral , is given by : @xmath38 including the theiler correction @xmath39 @xcite which we chose equal to the mean period , i.e. 10 time steps .
the canonical procedure to establish a finite correlation dimension is to show the existence of a scaling region for small @xmath40 where eq .
( [ d2 ] ) holds and stays constant for a high enough embedding dimensions . for all processes investigated here , the true correlation dimension is infinity .
following the idea of surrogate data testing , we fix an algorithm to obtain a finite value from the correlation integral and look for differences to the original data .
therefore , we apply theiler and lookman s rule of five chord estimator @xcite and chose their @xmath41 equal to the standard deviation of the data . for such a large @xmath41
we do not examine the small scale behavior of eq .
( [ d2 ] ) anymore .
we are aware that we should not call this quantity correlation dimension anymore .
it has been termed dimensional complexity
@xcite .
the surrogate data are produced by the fft algorithm @xcite .
for each degree of violation of the null hypothesis 50 independent surrogate data sets of length 8192 were generated . denoting the `` correlation dimension '' of the original data by @xmath42 , the mean of the distribution of this feature for the surrogate data by @xmath43 and its the variance by @xmath44 , the result
is displayed as : [ zet ] z= .
it was confirmed that the distribution of the feature is sufficiently well described by a gaussian distribution .
thus , @xmath45 can be related to a confidence interval , since for 50 realizations the @xmath46-distribution of @xmath47 is well approximated by a gaussian distribution and @xmath48 corresponds to the 5 % level of significance . in general , in power of the test investigations a procedure different from that outlined above is chosen . for a certain significance level , e.g. 5% , and different degrees of violation of the null hypothesis ,
numerous realizations , e.g. 1000 , of the process are generated and the fraction of rejected null hypotheses is reported .
due to the high computational burden for calculating the correlation integral , this procedure is not feasible here .
the above procedure has the drawback that the results depend on the single realization that is used as basis for the surrogates .
we repeated the analysis reported below for independent realizations and found no qualitative differences for different realizations .
for the first violation of the null hypothesis , we increase @xmath49 in eq.([process1 ] ) from zero , i.e. no violation , to 0.5 in steps of 0.1 .
the distribution of these data are not gaussian for @xmath50 .
thus , the amplitude adjusted surrogate data algorithm @xcite was applied .
the deviation from gaussianity is weak for the range of violations chosen .
we also applied the algorithm without amplitude adjustment and did not found significant different results .
[ ar2amp_res ] displays the result of the simulation study . in dependence on the embedding dimension
, @xmath51 is displayed for different degrees of violation of the null hypothesis .
= 8.4 cm as expected , without any violation , the @xmath51 values stay within the 2@xmath52 region given by @xmath53 .
a modulation depth @xmath49 of @xmath54 and @xmath55 leads to results at the border of 5% significance .
starting from @xmath56 , see fig .
[ ar2_dat]b , the null hypothesis is clearly rejected at the 5% level of significance whenever the embedding dimension is large enough to reconstruct the second order process appropriately .
to investigates the effect of a variation in the period of the linear stochastic process , we increase @xmath28 in eq.([process2_1],[process2_2 ] ) from zero to three.the distribution of these data are gaussian independent from the value of @xmath28 .
thus , no amplitude adjustment was necessary .
again , the distribution of the feature is sufficiently well described by a gaussian distribution .
[ ar2fre_res ] displays the result of the simulation study .
= 8.4 cm for all degrees of violation , the violation is not detected when the embedding dimension is too small to unfold the dynamics in phase space .
otherwise , a modulation of the period of 15 % , see fig .
[ ar2_dat]c , leads to a clear rejection of the null hypothesis at the 5 % level of confidence .
the simulation studies reported in this paper indicate that surrogate data testing for linear , stochastic , gaussian stationary processes is powerful against a violation of the assumption of stationarity .
thus , a significant result of the test does not necessarily indicate a non - linear or even chaotic process underlying the data .
it might have simply be caused by a non - stationarity of the process .
|
surrogate data testing is a method frequently applied to evaluate the results of nonlinear time series analysis .
since the null hypothesis tested against is a linear , gaussian , stationary stochastic process a positive outcome may not only result from an underlying nonlinear or even chaotic system , but also from e.g. a non - stationary linear one .
we investigate the power of the test against non - stationarity .
|
there is currently a great deal of interest in the statistical properties of primordial perturbations from inflation , because measurements of any non - gaussianity will improve by about an order of magnitude over the next few years , for example with the atacama cosmology telescope @xcite and planck @xcite
. this will provide a key way to discriminate between the many models of inflation .
although single field models of slow - roll inflation typically generate a small level of non - gaussianity @xcite , there may be an observable level generated in multiple field inflation @xcite , for example the curvaton scenario @xcite , or some single field models such as dirac - born - infeld inflation ( dbi ) @xcite .
currently observations of the cmb have concentrated on constraining the 3-point function ( bispectrum ) @xcite , but the 4-point function ( trispectrum ) has also been considered @xcite , and in principle higher order @xmath0-point functions may also be observable if they are sufficiently large . in some models
the first signal of non - gaussianity may come through the trispectrum , for example some special cases of the curvaton scenario @xcite .
furthermore higher order statistics in principle carry more information and therefore could distinguish between different sources of non - gaussianity , for example in many cases the bispectrum can be paremetrised by a single non - linearity parameter , @xmath1 , while the trispectrum depends on two wavenumber independent parameters , @xmath2 and @xmath3 @xcite .
we will calculate the primordial curvature perturbation on uniform density hypersurfaces , @xmath4 , on large scales employing the @xmath5 formalism @xcite .
we use the separate universe approach @xcite .
this considers each super - hubble scale patch to be evolving like a separate friedmann - robertson - walker universe which are locally homogeneous . by patching these regions together
we can track the evolution of the perturbations on large scales just by using background quantities which greatly simplifies our calculations .
we also need to know the perturbations of the scalar fields at hubble - exit , but in many cases these are extremely close to gaussian in which case their statistical properties are given purely in terms of the power spectrum of the fields .
the number of @xmath6-foldings , @xmath7 , given by n=^t_fin_t_inih(t ) dt , is evaluated from an initial flat hypersurface to a final uniform - density hypersurface .
the perturbation in the number of @xmath6-foldings , @xmath5 , is the difference between the curvature perturbations on the initial and final hypersurfaces .
we wish to calculate primordial perturbations , hence we pick a final uniform density hypersurface to be at a fixed time during the standard radiation dominated era , for example during primordial nucleosynthesis .
the initial time is arbitrary provided it is after the hubble exit time of all relevant scales .
it is often convenient to pick this time to be shortly after hubble - exit .
we introduce the method for calculating the primordial curvature perturbation in terms of the @xmath5 formalism in more detail below . here
we present a feynman diagram type approach to calculating the primordial non - gaussianity , where every term in the @xmath0-point function of @xmath4 has a diagrammatic representation .
similar diagrams were introduced in the context of non - linear perturbation theory of large scale structure @xcite and further developed in @xcite ( see also @xcite ) .
it is possible to read off the corresponding mathematical term in the @xmath0-point function for every diagram and it is possible to draw all the diagrams in a systematic way so that all terms are included .
this approach extends to loop corrections of any required order .
although the loop corrections are expected to be small in general compared to the tree level terms there is no proof of this , except for a particular loop correction to the bispectrum in some specific models of inflation @xcite .
we will see that in general the loop corrections depend on a large scale cut off ; depending on this cut off the loop correction may contribute significantly to observable quantities , see for example @xcite .
in fact there is a specific model of inflation where the one loop term gives the dominant and potentially observable contribution to the bispectrum @xcite .
a diagrammatic method to calculate field fluctuations during inflation was presented in @xcite .
we first introduce the primordial curvature perturbation . in the next two sections of this paper
we assume the initial field perturbations are gaussian , which in many cases is a good approximation to make , for example slow - roll inflation . in sec .
[ sec : realspacediagrams ] the rules for drawing diagrams in real space are given , along with a discussion of why we want to calculate connected rather than disconnected diagrams . in sec .
[ sec : fourierspacediagrams ] the equivalent rules in fourier space are given .
an application of the fourier space rules for the power spectrum including second order loop corrections is given in sec .
[ sec : powerspectrum ] . in sec .
[ sec : renormalisation ] we discuss a way to renormalise the vertices of the diagrams such that all diagrams with dressed vertices can be absorbed , with a proof given in the appendix . in sec .
[ sec : ng ] the assumption of gaussian initial fields is dropped and the complete rules presented and applied to several examples . finally we conclude in sec .
[ sec : conclusion ] .
we can write the primordial curvature perturbation , @xmath4 , in terms of derivatives of @xmath7 with respect to the fields , @xmath8 and the field perturbation , @xmath9 , at the initial time @xmath10 . by neglecting any additional dependence on @xmath11 at hubble - exit
we are assuming the field perturbations on large scales are overdamped and thus we can neglect the decaying mode .
this is valid during slow roll , but also more generally holds for light fields during inflation .
the assumption that @xmath7 is a function of field values only is also valid in dbi inflation @xcite .
we use the notation that the fields are @xmath12 , where the perturbed field @xmath13 satisfies @xmath14 .
the curvature perturbation is given by @xmath15 the dummy index @xmath16 labels the light scalar fields relevant during inflation , and summation convention is used throughout this paper .
it is more convenient to work with the @xmath4 defined above since it satisfies @xmath17 , even though often @xmath18 is used in the literature .
this equation can hence be used to calculate the primordial @xmath0-point function of @xmath4 , although the result will depend on the @xmath0-point function of the fields at hubble exit .
this has so far been explicitly calculated for the 2- , 3- and 4-point functions , @xcite .
the connected 2- , 3- and 4-point functions of the fields are defined by @xmath19 note that only the 4-point function ( and higher ) depend on the direction of the @xmath20 vectors , the 2- and 3-point functions of the fields just depend on the magnitude of the vectors , @xmath21 .
homogeneity of the random fields implies that the sum of the @xmath20 vectors must be zero , while isotropy implies that the @xmath0-point functions are invariant under reorientations of this closed configuration . at lowest order in slow roll different fields
are uncorrelated at hubble exit and all light fields have the same power spectrum @xcite , so @xmath22 where @xmath23 is the kronecker delta - function , and the variance per logarithmic interval in @xmath24-space is given by @xmath25 where the hubble parameter @xmath26 is evaluated at hubble - exit , @xmath27 . at zeroth order in slow - roll parameters , @xmath28 is independent of wavenumber , i.e. we have a scale invariant spectrum for the field fluctuations .
the bispectrum satisfies @xmath29 @xcite , where @xmath30 is a slow - roll parameter , so it is zero at lowest order in slow roll .
however the trispectrum is not zero at lowest order in slow roll @xcite , @xmath31 , but is suppressed by @xmath32 compared to the power spectrum . for a discussion of the slow - roll order of higher order @xmath0-point functions of the fields
see @xcite .
initially we will assume that all of the fields have a purely gaussian distribution at an initial time , e.g. shortly after hubble exit .
then every odd @xmath0-point function of the fields is zero and every even @xmath0-point function of the fields can be reduced to a product of 2-point functions .
this assumption is not required and the extension to non - gaussian field perturbations will be presented in sec . [
sec : ng ] . however the approximation of initially gaussian field perturbations is frequently made in the literature and is often a good approximation , for example @xcite shows that the 3-point function of the fields adds an unobservably small contribution to the bispectrum around hubble exit assuming slow roll inflation with a standard kinetic term of the scalar fields , while @xcite show that the 4-point function of the fields adds an unobservably small contribution to the trispectrum around hubble exit under the same assumptions .
working to zeroth order in slow roll implies that ( [ cfreefields ] ) holds and that the 3-point function of the fields is zero .
so although working to lowest order in slow roll is not the same as working with gaussian field perturbations , in practice the conditions are related .
hence when assuming that the field perturbations have a gaussian distribution at hubble exit we will also work at zeroth order in slow roll .
it is likely that any observable amount of non - gaussianity can be calculated to sufficiently high accuracy when making these two approximations , assuming a standard kinetic term . assuming that the fields are uncorrelated gaussian variables , with identical distributions , as discussed above , the two point function of the fields is given by [ varphicorrelation ] _
x_i^a_x_j^b = \ { l ^a bg(|x_i - x_j| ) ij , +
^a b^2 i = j . .
all of the higher @xmath0-point functions of the fields can be written in terms of @xmath33 and the variance of the fields @xmath34 .
the two point function @xmath33 only depends on @xmath35 because we are assuming the background is homogenous and isotropic .
the diagrammatic rules for the connected @xmath0-point function , @xmath36 , are : 1 .
draw @xmath0 points representing the @xmath0 spatial points @xmath37 and connect them with @xmath38 propagators ( dashed lines ) which attach two of the positions @xmath39 ( it can be two distinct positions or it can have both ends attached to the same position ) . the diagram should be connected in order to calculate the connected @xmath0 point function , see sec .
[ sec : disconnecteddiagrams ] .
we require @xmath40 in order to draw a connected diagram and diagrams with @xmath41 are tree level , while those with @xmath42 include loop corrections .
2 . label each end of each propagator with the field indices @xmath43 .
3 . assign a factor @xmath44 to each spatial point , @xmath39 , where the number of derivatives of @xmath7 is the number of propagators attached to that @xmath39 .
4 . assign a factor of @xmath45 to each propagator , where @xmath46 are the appropriate field indices attached to the propagator and @xmath47 are the positions at either end of the propagator .
if both ends of the propagator are attached to the same @xmath39 instead assign a factor @xmath48 .
divide by the appropriate numerical factor . whenever @xmath49 propagators attach the same @xmath39 and @xmath50 at
both ends this gives a factor of @xmath51 . whenever @xmath49 propagators dress an @xmath39 this gives an additional factor of @xmath52 ( as well as the factor of @xmath51 due to the previous rule ) .
a propagator dresses an @xmath39 if both ends of the propagator are attached to the same @xmath39 .
add all permutations of the diagrams which is all of the distinct ways to relabel the spatial points .
the number of permutations depends on the symmetries of the diagram , a diagram with complete symmetry between all of the spatial points has only one term , while for a diagram with no symmetries between the spatial points there are @xmath53 permutations . in fig .
[ fig : realspace234pt ] the diagrams for the 2-point function at tree level and one loop level , plus the tree level terms for the 3- and 4-point functions are shown .
note that for the 4-point function there are two tree level terms .
the terms corresponding to the diagrams are given below , in the same order as the diagrams , [ 2ptfnreal ] _ x_1_x_2&= & ( n_a n^a + n_a n^ab_b)g(|x_1-x_2| ) + 12n_abn^abg(|x_1-x_2|)^2 , + _ x_1_x_2_x_3&= & n_a n_b n^ab ( g(|x_1-x_2| ) g(|x_1-x_3|)+2 ) , + [ 4ptfnreal ] _ x_1_x_2_x_3_x_4_c & = & n_a n_b n_c n^abc ( g(|x_1-x_2| ) g(|x_1-x_3| ) g(|x_1-x_4| ) + 3 ) + & & + n_a
n_b n^a_c n^bc ( g(|x_1-x_2|)g(|x_2-x_3|)g(|x_3-x_4|)+11 ) . although the second diagram for the two point function , fig .
[ fig : realspace234pt ] , has a dressed vertex with an associated numerical factor of @xmath54 , there is another permutation of this diagram with the @xmath55 dressed which gives an equal contribution .
the third diagram has two propagators attaching @xmath56 and @xmath55 and there are no other permutations of this diagram so there is a numerical factor of @xmath54 , as shown in the third term of ( [ 2ptfnreal ] ) .
the diagrams for the 3- and 4-point functions are all at tree level so they all have a numerical factor of 1 .
the first diagram for the 4-point function has 4 permutations since we have 4 choices of which of the @xmath57 has 3 propagators attached to it as given in the first term of ( [ 4ptfnreal ] ) , the second diagram of the 4-point function has 12 permutations because there are @xmath58 choices of which 2 of the @xmath57 should have 2 propagators attached to them in general any @xmath0-point function has a connected and a disconnected part . throughout this paper
we will only consider the connected @xmath0-point function because the disconnected contribution contains no new information over the lower @xmath0-point functions that it is split into .
there is no disconnected contribution to the 2- or 3-point function of the fields or zeta , because we assume @xmath59 .
the disconnected contribution is only non - zero if a subset of the @xmath20 vectors sums to zero .
for example the disconnected part of the primordial 4-point function , @xmath60 , is the product of two 2-point functions such as @xmath61 and in total there are three such terms .
this is non - zero even if @xmath4 is purely gaussian , but it contains no new information compared to the primordial 2-point function .
this term only contributes if @xmath62 for some @xmath63 , so provided the sum of any two @xmath20 vectors is not zero any contribution is from the connected part .
the connected part of the primordial 4-point function is only non - zero if @xmath4 is not purely gaussian and it contains information that is not observable from the primordial 2- or 3-point functions .
we present here the rules for drawing the connected diagrams ( see sec .
[ sec : disconnecteddiagrams ] ) of the @xmath0-point function of @xmath4 at @xmath38-th order ( i.e. @xmath64 ) , for @xmath40 .
the tree level terms correspond to @xmath41 . 1 .
draw all distinct connected diagrams with @xmath0-external ( solid ) lines and @xmath38 ( dashed ) propagators .
every vertex must consist of 1 external line and at least 1 propagator .
2 . label the external legs with incoming momenta @xmath65 for @xmath66 and label the propagators with internal momenta @xmath67 for @xmath68 .
label each end of each propagator with a field index @xmath69 .
3 . assign a factor @xmath70 to each vertex .
the number of derivatives of @xmath7 corresponds to the number of propagators attached to each vertex .
we use the convention that incoming momentum is positive .
the @xmath71 function ensures momentum is conserved at each vertex .
see fig .
[ fig : fourierspacerulesg ] .
4 . assign a factor @xmath72 to each propagator , where @xmath46 are the appropriate field indices that the propagator is labelled with at either vertex and @xmath73 is the momentum attached to the propagator .
see fig .
[ fig : fourierspacerulesg ] .
integrate over the propagator momenta , @xmath74 .
the first @xmath75 integrations are trivial because of the @xmath71 functions but any further integrations ( in the case of a diagram with loop corrections ) can not in general be evaluated analytically .
divide by the appropriate numerical factor . whenever @xmath49 propagators attach the same vertices at
both ends this gives a factor of @xmath51 .
whenever @xmath49 propagators dress a vertex this gives an additional factor of @xmath52 ( as well as the factor of @xmath51 due to the previous rule ) .
a propagator dresses a vertex if both ends of the propagator are attached to the same vertex .
add all permutations of the diagrams which is all of the distinct ways to relabel the @xmath65 attached to the external lines .
the number of permutations depends on the symmetries of the diagram , a diagram with complete symmetry between all of the external lines has only one term , while for a diagram with no symmetries between the external lines there are @xmath53 permutations .
going between the unlabelled diagrams in fourier or real space is straightforward , to go from fourier space to real space one simply removes the external lines and places an @xmath39 at each vertex . to go from real space to fourier space
one attaches an external line in place of every @xmath39 . in fourier space rule 1 for the diagram
is to draw every possible diagram at the appropriate order .
rule 2 gives the rules for labelling the diagrams .
rules 35 give the associated mathematical expression for each diagram , up to a numerical factor given in rule 6 .
a derivation of the numerical factors is given in sec .
[ sec : renormalisation ] and app .
[ sec : app : nptfn ] .
the numerical factor for all tree level diagrams is 1 .
finally rule 7 tells one to include every permutation of the labelling of the @xmath65 on the external legs that are distinct .
the primordial power spectrum in fourier space is defined by @xmath76 the tree level and one loop correction terms to the 2-point function is diagrammatically given by fig .
[ fig:2pt1loopg ] . after carrying out the trivial integration over the propagator momentum the tree level term is given by @xmath77 the one loop terms
are given by [ 2ptfn1loop ] p_^1(k)=d^3q ( 12n_abn^abp(q)p(|-|)+ n_an^ab_bp(k)p(q ) ) , where we have already carried out one integration .
both of the diagrams associated with these terms have a numerical factor of 2 , however there are also 2 permutations of the second term because it has a non symmetric diagram . because of momentum conservation , @xmath78 , it follows that @xmath79 . after enforcing @xmath80 the two permutations of the second diagram
give an equal contribution which cancels the numerical factor of 2 .
going to 2 loops , terms of order @xmath81 there are four terms .
diagrammatically the terms are given in fig .
[ fig:2pt2loopg ] which corresponds to [ 2ptfn2loop ] p_^2&=&d^3q_1d^3q_2 ( 14n_abc^abn^cp(q_1)p(q_2)p(k)+14n_ab^an^bc_cp(k)p(q_1)p(q_2 ) .
+ 12n_abn^abc_cp(q_1)p(q_2)p(|-| ) + 16n_abcn^abcp(q_1)p(|-|)p(|-| ) ) . only the fourth diagram of fig .
[ fig:2pt2loopg ] does not have a dressed vertex .
in fact at every level there is only one diagram without dressed vertices , as has already been shown explicitly at tree and one loop level in fig .
[ fig:2pt1loopg ] . in the case where a single field @xmath82 generates the primordial curvature perturbation , so @xmath83
, we have p_^tree(k)=(n)^2 p(k ) , where @xmath84 . in the case of a scale invariant spectrum
we can also evaluate the 1 loop integrals if we apply a large scale cut off @xmath85 . for
a discussion of how to evaluate the integral over the loop momenta and the dependence of this term on the cut off @xmath85 see for example @xcite .
we also apply a small scale cut off @xmath24 .
the result is p_^1loop(k)= p(k)((n)^2+nn ) ( kl ) , where the variance @xmath86 was defined by eq .
( [ p ] ) .
in analogy with eq .
( [ p ] ) we define @xmath87 which from observations of the cmb is of the order @xmath88 @xcite .
we hence have [ p1loop ] = _^tree ( + ) ( kl ) .
we can write the derivatives of @xmath7 in terms of the ( in principle ) observable non - linearity parameters @xmath1 and @xmath3 which in this case can be defined by @xcite [ fandgdefn ] = _ 1 + 35f_nl_1 ^ 2+g_nl_1 ^ 3 + , where @xmath89 is gaussian because it is directly proportional to the initial gaussian field perturbation , @xmath90 , and the dimensionless non - linearity parameters , @xmath1 and @xmath3 , are given by [ fnl1field ] f_nl&=&56 , + [ gnl1field ] g_nl&=&. for the extension of these two formula to the case of multiple field inflation see for example @xcite
. substituting ( [ fnl1field ] ) and ( [ gnl1field ] ) into ( [ p1loop ] ) we find [ p1loopfg ] = _^tree ( f_nl^2+g_nl ) ( kl ) .
if @xmath91 and we take the large scale cut off to be comparable to the present day hubble scale then we can take @xmath92 @xcite , and the observational bound on the bispectrum , @xmath93 @xcite bounds the first term of the ratio above to be less than @xmath94 .
there is not yet any equivalent observational bound on @xmath3 but it would have to be extremely large in order to give a significant contribution to eq .
( [ p1loopfg ] ) .
it therefore seems that the only way for the 1 loop contribution to be significant compared to the tree level term of the power spectrum is to take an exponentially large cut off @xmath85 . for a discussion of this possibility see for example @xcite
there is a way to reduce the number of diagrams significantly , by renormalising the vertices such that the diagrams with dressed vertices are automatically accounted for .
we do this by renormalising the factors @xmath44 , that are attached to each vertex .
the derivative of the number of @xmath6-foldings for the given background @xmath95 is @xmath96 , and we can relate this to the number of @xmath6-foldings at a general point @xmath97 by [ ntdefn ] n ( ( ) ) = n|__0+n_a^a+n_ab^a^b + n_abc^a^b^c+ .
the expectation of @xmath98 is given by [ nt ] _ abc = n_abc+ n_abcd^d^2 + 18 n_abcde^de^2 ^ 2 + , where we have used wick s theorem to decompose the expectation of powers of @xmath99 according to [ wick ] ^2n=(2n-1)!!^2^n , ^2n-1=0 .
the double factorial is defined by @xmath100 .
the coefficient of the general term in ( [ nt ] ) , @xmath101 is given by ( [ ntdefn ] ) and ( [ wick ] ) and is @xmath102 .
this explains the numerical factor associated with every dressed vertex , see rule 6 , sec .
[ sec : fourierspacediagrams ] . in fourier space
the ensemble average of @xmath98 is [ ntfourier ] _ abc = n_abc+ n_abcd^d d^3qp(q ) + 18 n_abcde^de d^3q_1 d^3q_2 p(q_1)p(|-| ) + , which in real space is given by eq . ( [ nt ] ) . in fig .
[ fig : ntprimed ] the graphical representation of @xmath103 is given up to 2 loops .
so far the derivatives of @xmath7 had been calculated at @xmath95 , however if we instead calculate the expectation of the derivative of @xmath7 at a general point , and attach this factor @xmath104 to each vertex ( which we call a renormalised vertex ) then all diagrams with dressed vertices are automatically included in the expansion ( [ nt ] ) and do nt need to be drawn .
we provide a proof of this in appendix [ sec : app : renn ] .
the new diagrammatic rules for the connected @xmath0-point function with renormalised vertices are : 1 .
draw @xmath0 points representing the @xmath0 spatial points @xmath37 and connect them with @xmath38 propagators ( dashed lines ) which attach two distinct positions @xmath39 and @xmath50 .
the vertices are drawn with a solid dot to show that they are renormalised , see fig .
[ fig : ren2ptfn2loop ] for examples of such diagrams .
the diagram should be connected in order to calculate the connected @xmath0 point function , see sec .
[ sec : disconnecteddiagrams ] .
we require @xmath40 in order to draw a connected diagram and diagrams with @xmath41 are tree level , while those with @xmath42 include loop corrections .
2 . label each end of each propagator with the field indices @xmath43 .
3 . assign a factor @xmath104 to each spatial point , @xmath39 , where the number of derivatives of @xmath105 is the number of propagators attached to that @xmath39 .
4 . assign a factor of @xmath45 to each propagator , where @xmath46 are the appropriate field indices attached to the propagator and @xmath47 are the positions at either end of the propagator .
divide by the appropriate numerical factor . whenever @xmath49 propagators attach the same @xmath39 and @xmath50 at
both ends this gives a factor of @xmath51 .
add all permutations of the diagrams which is all of the distinct ways to relabel the spatial points .
the number of permutations depends on the symmetries of the diagram , a diagram with complete symmetry between all of the spatial points has only one term , while for a diagram with no symmetries between the spatial points there are @xmath53 permutations . in fig .
[ fig : ren2ptfn2loop ] the diagrams for the 2-point function up to two loop level is presented .
the terms corresponding to the diagrams are given below , in the same order as the diagrams , _
x_1_x_2&= & _ a^ag(|x_1-x_2| ) + 12_ab^abg(|x_1-x_2|)^2 + & & + _
abc^abcg(|x_1-x_2|)^3 + .
note that with the renormalised diagrams there is only one diagram at every order in loops .
this result is also derived to all orders in app .
[ sec : app : renn ] .
rules for drawing the connected diagrams ( see sec .
[ sec : disconnecteddiagrams ] ) of the @xmath0-point function of @xmath4 at @xmath38-th order ( i.e. @xmath64 ) , for @xmath106 .
the tree level terms correspond to @xmath41 . 1 .
draw all distinct connected diagrams with @xmath0 external ( solid ) lines and @xmath38 ( dashed ) propagators .
every renormalised vertex must consist of 1 external line and at least 1 propagator , and drawn with a solid dot to show that the vertex is renormalised .
the propagators can not have both ends attached to the same vertex .
2 . label the external legs with incoming momenta @xmath65 for @xmath66 and label the propagators with internal momenta @xmath67 for @xmath68 .
label each end of each propagator with a field index @xmath69 .
3 . assign a factor @xmath107 to each vertex .
the number of derivatives of @xmath105 corresponds to the number of propagators attached to each vertex .
we use the convention that incoming momentum is positive .
the @xmath71 function ensures momentum is conserved at each vertex .
4 . assign a factor @xmath72 to each propagator , where @xmath46 are the appropriate field indices that the propagator is labelled with at either vertex and @xmath73 is the momentum attached to the propagator .
5 . integrate over the propagator momenta , @xmath74 .
the first @xmath75 integrations are trivial because of the @xmath71 functions but any further integrations ( in the case of a diagram with loop corrections ) can not in general be evaluated analytically .
divide by the appropriate numerical factor . whenever @xmath49 propagators attach the same vertices at
both ends this gives a factor of @xmath51 .
add all permutations of the diagrams which is all of the distinct ways to relabel the @xmath65 attached to the external lines .
the number of permutations depends on the symmetries of the diagram , a diagram with complete symmetry between all of the external lines has only one term , while for a diagram with no symmetries between the external lines there are @xmath53 permutations .
we again calculate the power spectrum up to 2 loop level , this time with renormalised vertices .
there is only one diagram at every loop level , we present the three lowest order diagrams in fig . [
fig : ren2ptfn ] , which correspond to p^_&= & _
a^ap(k ) + d^3q _
ab^abp(q)p(|-| ) + & & + d^3q_1 d^3q_2 _ abc^abcp(q_1)p(|-|)p(|-| ) .
we calculate the bispectrum including the leading order loop corrections .
the bispectrum is defined by [ bispectrum ] _
_ k_3b _ ( k_1,k_2,k_3 ) ^3 ( ) . at tree level
there is one term and at 1 loop level there are two terms , which are shown diagrammatically by fig .
[ fig : ren3ptfn1loop ] and these give the contribution @xmath108
if we do nt assume that the scalar fields are gaussian at the initial time ( e.g. hubble exit ) we have to take account of the higher @xmath0-point functions of the fields .
this is likely to be important in models with a non - standard kinetic term where the non - gaussianity of the fields may be large at hubble exit @xcite , for example dbi inflation inspired from brane world models , @xcite .
the connected 2- , 3- and 4-point function of the fields are defined in ( [ fieldcorrelators])([4ptfnfields ] ) .
our analysis relies on the @xmath5-formalism which assumes that @xmath7 is a function of the field values @xmath109 but not the field derivatives independently as well .
however we can still go beyond leading order in a slow - roll expansion and derive expressions which are valid to arbitrary order in slow roll , for example the different fields are in general correlated at hubble exit , at first order in slow roll , see @xcite .
the diagrams are more complicated since while the 2-point function is still drawn as a propagator the higher @xmath0-point functions are drawn by an @xmath0-point vertex of dashed lines .
the rules for the diagrams are now : 1 . draw all distinct connected diagrams with @xmath0-external ( solid ) lines and the appropriate number of internal correlators , which can be propagators ( 2-point functions ) or higher @xmath0-point correlators of the fields .
the order of the connected @xmath0-point function of the field perturbations is @xmath75 , i.e. @xmath110 .
there can not be more than one external line attached to each vertex .
every dashed line must be attached to an external line on at least one end , if it is attached to an external line on both ends it is a propagator ( a correlator of order 2 ) .
2 . label the external lines with incoming momenta @xmath65 for @xmath66 and label each dashed line with internal momenta @xmath67 for @xmath68 .
label each dashed line at a vertex containing an external line with a field index @xmath111 .
3 . assign a factor @xmath112 to each vertex which includes an external line , where the number of derivatives of @xmath7 corresponds to the number of dashed lines attached to that vertex .
the @xmath71 function ensures momentum is conserved at each vertex .
see fig .
[ fig : fourierspacerulesng ] .
4 . assign a factor @xmath113 to each propagator , where @xmath46 are the appropriate field indices that the propagator is labelled with at either vertex and @xmath73 is the momentum attached to the propagator .
see fig . [
fig : fourierspacerulesng ] .
5 . assign a factor @xmath114 to each vertex of three dashed lines and no external lines ( which corresponds to a 3-point function of the fields ) , where @xmath115 and @xmath116 are the field indices attached to the dashed lines where they meet a vertex with an external line .
see fig .
[ fig : fourierspacerulesng ] again .
similar rules hold for vertices of dashed lines at higher order .
integrate over all internal momenta , @xmath74 . for a tree diagram
all of the integrations are trivial , but in the case of a diagram with loop corrections there will be some integrals which in general can not be evaluated analytically .
divide by the appropriate numerical factor . whenever @xmath49 propagators or higher order correlators of the fields attach the same vertices ( and with the same number of lines to each vertex ) at
both ends this gives a factor of @xmath51 .
if @xmath117 legs from a higher order correlator are attached to the same vertex this gives an extra factor of @xmath118 .
it follows that if @xmath49 correlators , each of order @xmath119 , all dress the same vertex then this gives a numerical factor of @xmath120 .
see app .
[ sec : app : ng ] for a proof of these rules and see fig .
[ fig : ntng ] for two example of dressed vertices , the associated terms with the numerical factors are given in ( [ ntngeg1 ] ) , ( [ ntngeg2 ] ) .
add all permutations of the diagrams which is all of the distinct ways to relabel the @xmath65 attached to the external lines .
the number of permutations depends on the symmetries of the diagram , a diagram with complete symmetry between all of the external lines has only one term , while for a diagram with no symmetries between the external lines there are @xmath53 permutations .
we give again the power spectrum defined by ( [ powerspectrumdefn ] ) , but this time for non - gaussian initial fields and valid to all orders in slow roll .
there is one extra diagram at one loop level , and six extra diagrams at two loop level .
because of momentum conservation , @xmath78 , it follows that @xmath121
. therefore even the non - symmetric diagrams which would have two permutations only correspond to a single term after enforcing this symmetry .
the tree level term is given by @xmath122 notice that at leading order in slow roll using ( [ cfreefields ] ) this reduces to ( [ 2ptfntree ] ) .
the one loop terms are given by p_^1(k)&=&d^3q ( 12n_abn_cdc^ac(q)c^bd(|-|)+ n_an_bcdc^ab(k)c^cd(q ) + & & + n_an_bcb^abc(k , q,|-| ) ) , where the extra one loop term which does not appear for gaussian fields is shown in fig .
[ fig:2pt1loopng ] , and the first two terms were shown diagrammatically in fig .
[ fig:2pt2loopg ] .
going to 2 loops , terms of order @xmath81 there are ten terms , only the last four of which are non - zero for a gaussian field .
diagrammatically the terms involving a 3-point or 4-point function of the fields are given in fig .
[ fig:2pt2looptbng ] , which corresponds to the first six terms of [ 2ptfn2loopng ] p_^2&=&d^3q_1d^3q_2 ( 13n_an_bcdt^abcd(,-,- , ) .
+ & & + 14n_abn_cdt^abcd(,-,,- ) + & & + n_abn_cdeb^bde(|+|,|-|,|-|)c^ac(q_1 ) + 12n_abn_cdeb^abc(k , q_1,|+|)c^de(q_2 ) + & & + 13n_an_bcdeb^cde(q_1,|+|,q_2)c^ab(k ) + 12n_an_bcdeb^abc(q_1,|-|,k)c^de(q_2 ) + & & + 14n_an_bcdefc^ab(k)c^cd(q_1)c^ef(q_2 ) + 14n_abcn_defc^ad(k)c^bc(q_1)c^ef(q_2 ) + & & + 12n_abcdn_efc^ab(q_1)c^ce(|-|)c^df(q_2 ) + & & .
+ 16n_abcn_defc^ad(q_1)c^be(|-|)c^cf(|-| ) ) .
the final four terms were already shown diagrammatically in fig .
[ fig:2pt2loopg ] , they correspond to the last four terms of ( [ 2ptfn2loopng ] ) .
however the mathematical expressions for these four diagrams given in ( [ 2ptfn2loop ] ) are correct at leading order in slow roll , while the expression above is valid to all orders in slow roll .
note that 1 of the 1 loop diagrams and 4 of the 2 loop diagrams have dressed vertices , these can be removed by renormalising the vertices in a similar way to the case with gaussian fields , see app .
[ sec : app : ng ] .
the connected part of the primordial trispectrum is defined by @xmath123 diagrammatically all of the tree level terms for the 4-point function are given by fig .
[ 4ptfntreenew ] .
writing down these terms in order we get [ trispectrumfull ] t_^ ( , , , ) & = & n_an_bn_cn_d t^abcd ( , , , ) + & & + n_abn_c n_d
n_e ( c^ac(k_1)b^bde(|+|,k_3,k_4)+ ( 11 perms ) ) + & & + n_abn_cdn_e n_f ( c^ac(|+|)c^bd(k_3)c^df(k_4 ) + ( 11 perms ) ) + & & + n_abcn_d n_e n_f ( c^ad(k_2)c^be(k_3)c^cf(k_4)+(3perms ) ) . this result was first derived in @xcite without using diagrams ( see also @xcite ) .
note that all of the numerical coefficients are one , because all of the terms are tree level , and that only the final two terms are non - zero if the field fluctuations are gaussian .
the real space diagrams for these two terms were given in fig .
[ fig : realspace234pt ] .
we have presented a diagrammatic method for calculating any @xmath0-point function of the primordial curvature perturbation , @xmath4 , at tree - level or any required loop level .
rules are given for first drawing diagrams at the required order , and then for writing down the corresponding terms in the @xmath0-point function , in either real or fourier space . for example , we have drawn diagrams corresponding to all the 1- and 2-loop terms to the power spectrum and have given the corresponding corrections to the power spectrum .
our method is based on using the @xmath5-formalism which identifies the local curvature perturbation with the perturbation of the integrated expansion from an initial hypersurface .
the expansion on large scales ( typically larger than the hubble scale ) can be determined as a ( non - linear ) function of the initial local values of scalar fields during inflation using the homogeneous ( frw ) equations of motion in the separate universe approach . in practice
we are working with fields that are smoothed on scales @xmath124 .
in particular we neglect any ultraviolet divergence of the fields on small scales .
we show that it is possible to renormalise vertices by absorbing an infinite sum of terms , corresponding to loop corrections to vertices .
these are automatically accounted for if we work in terms of the average local expansion @xmath125 and its derivatives , defined in eq .
( [ nt ] ) for gaussian fields , instead of the background expansion , @xmath126 . this removes terms such as @xmath34 which is divergent for scale - invariant spectra , leaving all @xmath0-point functions finite in real space so long as @xmath125 and its derivatives are finite .
however divergent terms remain in fourier space due to integrals over loop momenta between vertices which may diverge in the infrared .
a discussion of the effect of the large scale cut off on observables is given in @xcite .
results for the primordial @xmath0-point function are given in terms of derivatives of @xmath7 with respect to the fields at hubble exit and the @xmath0-point functions of the fields evaluated at hubble exit .
in single - field models of inflation the curvature perturbation is constant on super hubble scales , so it is possible to calculate the derivatives of @xmath7 explicitly at hubble exit during inflation @xcite without needing to know the details of the end of inflation or reheating , etc . in this case
both the bispectrum and trispectrum have been shown to be small , regardless of whether the fields are assumed to be gaussian or not , but assuming the field has a standard kinetic term in the action . during inflation the derivatives of @xmath7
can be calculated explicitly even in multiple field models for a separable potential , this been done to second @xcite and third order @xcite .
several specific multiple field models have been studied in greater detail , e.g. @xmath127-flation @xcite and a model of multiple field inflation in which @xmath7 can be calculated exactly , @xcite .
for all examples considered so far in the slow roll regime the non - gaussianity has proved to be small ( see also @xcite ) , however there are many models which explicitly generate a large non - gaussianity after inflation .
examples include the curvaton scenario @xcite and modulated reheating @xcite , also preheating , for example with hybrid inflation @xcite .
in general a numerical calculation of the derivatives of @xmath7 is required , the advantage of the @xmath5 formalism is that we only need to solve the homogenous background equations .
if one assumes the fields are gaussian then only the 2-point function of the fields is required , but in general one requires the @xmath128-point function of the fields when working to @xmath38th order for any @xmath0-point function of @xmath4 .
so far the 3- and 4-point functions have been calculated , both at leading order in slow roll and ignoring any quantum mechanical loop corrections .
it is possible but not expected that these corrections may be important @xcite . in the most commonly considered case of slow - roll inflation , the field is extremely close to gaussian at hubble exit so it is a good approximation to assume that the field is gaussian and that any observable non - gaussianity is generated after hubble exit . methods to generate a large non - gaussianity during inflation include having a break in the potential , @xcite , or to have a non - standard kinetic term so that the non - gaussianity of the fields is large at hubble exit @xcite , for example dbi inflation inspired from brane world models , @xcite .
the authors are grateful to david lyth and robert crittenden for comments .
kk is supported by pparc ( stfc ) .
ms is supported in part by jsps grants - in - aid for scientific research ( a ) no .
18204024 and ( b ) no .
the authors acknowledge support by the jsps japan - uk collaboration program .
here we provide a proof that renormalising vertices removes the diagrams with dressed vertices . in real space
this is equivalent to removing terms with @xmath129 .
we first provide a proof for the power spectrum where the notation is simpler .
we then provide a proof for the general @xmath0-point function , initially assuming that the fields are gaussian at hubble exit , sec .
[ sec : app : nptfn ] , and finally extend this proof to allow for arbitrary field correlations at hubble exit , sec .
[ sec : app : ng ] .
the gauge invariant curvature perturbation is defined by ( [ deltan ] ) as = n - n= _
p=1^ n__p ( ^a^b^c- ^a^b^c ) , where the subscript @xmath119 means that there are @xmath119 field indices above the brace . from the definition of @xmath130 , the 2-point function
is given by [ app:2ptfn ] _
p , q=1^n__pn__q ( _ x^a_x^b_x^c-_x^a_x^b_x^c ) ( _ y^d_y^e_y^f- _ y^d_y^f ) .
because the fields are gaussian every @xmath0-point function of the fields can be split into products of 2-point functions .
we can choose the pairs to be connected pairs , i.e. each @xmath99 is evaluated at a different point or disconnected pairs which are @xmath129 .
we consider every possible way of splitting the term @xmath131 into pairs : 1 connected pair : there are @xmath132 different ways to select this term .
this gives & & _ p , q n__p
n__q ^ad g(|x - y| ) _
x^b_x^c _ y^e_y^f = _ a^ag(|x - y| ) , + & & _ a= _ p=1^n_a_p-1^b^c .
2 connected pairs : there are @xmath133 ways to choose these two connected pairs .
the factor of @xmath134 comes because the order in which we pick the pairs does not matter .
this gives & & _ p , q n_ab_p-2
n_de_q-2 ^ad^be g^2(|x - y| ) _
y^f + & = & _ ab^abg^2(|x - y| ) .
continuing this for every possible number of connected pairs , we find that ( [ app:2ptfn ] ) is equivalent to _ x_y= _ r=1^__r^abcg^r(|x - y| ) .
similar to the calculation of the 2-point function , we have to split the term @xmath135 into pairs in every possible way . in the case that some of the @xmath136 do not appear in a connected pair
there is no contribution to the @xmath0-point function , for example if @xmath137 terms do not appear in any connected terms then the last bracket of ( [ app : nptfn ] ) is @xmath138 .
we are left with [ app : reducednptfn ] _
x_y_z&= & _ _
n n__pn__q n__r + & & _ x^a_x^b_x^c _ z^g_z^h_z^i_cc , where the subscript @xmath139 to the angular brackets means that they must be decomposed into pairs such that every @xmath140 appears in at least one connected pair .
we introduce the notation that @xmath141 is the number of connected pairs of @xmath142 and @xmath143 and similarly for every other possible connected pair .
furthermore @xmath144 is the number of @xmath145 that form part of a connected pair , so @xmath146 .
note that we require @xmath147 .
we consider a term in ( [ app : reducednptfn ] ) from an arbitrary choice of @xmath148 .
the number of ways to choose the @xmath149 connected pairs of @xmath150 correlators is given by @xmath151 , then the number of ways to choose the @xmath152 connected correlators linking @xmath142 and @xmath153 is given by @xmath154 .
continuing this counting for all the pairs , we find the number of ways to split into the required number of connected pairs is .
this choice of @xmath155 therefore gives the following term to ( [ app : nptfn ] ) , & & _ p , q, ,r n_abcn_def n_ghi
^ad^be + & & g^l_xy(|x - y|)g^l_yz(|y - z|)_p - n_x_r - n_z + & = & _ _ n_x__n_z^ad^beg^l_xy(|x - y|)g^l_yz(|y - z| ) , where there are @xmath149 contraction of field indices of the first two factors of @xmath104 etc . repeating this for all allowed choices of @xmath156 we get _ x_y_z= _ l_xy,,l_yz _ _ n_x__n_z^ad^beg^l_xy(|x - y|)g^l_yz(|y - z| ) .
diagrammatically a term @xmath157 corresponds to @xmath149 propagators which all have both ends attached to the same spatial points @xmath142 and @xmath143 .
this explains the numerical factor of @xmath158 , as given by rule 5 in real space ( or equivalently rule 6 in fourier space ) .
here we extend the proof that renormalising vertices eliminates the diagrams with dressed vertices also holds for non - gaussian fields . by doing this
we also derive the numerical factors that are associated with every diagram .
the proof is more complex than before because we must consider every possible correlation of an arbitrary number of fields , rather than reducing every @xmath0-point correlation to products of 2-point functions .
we now have [ app : generalisednt ] _
abc&= & n_abc+ n_abcde^d^e+n_abcdef ^d^e^f + & & + n_abcdefg ^d^e^f^g+ .
the correlators here includes the connected and disconnected parts . again
the @xmath0-point function of @xmath130 is given by ( [ app : nptfn ] ) and we can reduce it to ( [ app : reducednptfn ] ) in a similar way to the case with gaussian fields . however we now have to decompose the term in angle brackets into connected correlators of arbitrary order in every possible way .
we extend the notation that @xmath149 is the number of terms like @xmath159 to more general correlators like @xmath160 is the number of terms like @xmath161 .
every @xmath49 must have at least 2 distinct @xmath39 in the subscript .
we define @xmath144 as the number of @xmath142 s contained in the correlators and again we require that @xmath162 and similarly for @xmath163 , etc .
we consider the term in ( [ app : reducednptfn ] ) coming from a particular choice of @xmath164 , which we denote by p^abcefgh & = & _ l_xy _ l_xxz + & = & _ p_w _ c^l_xxzz and the product on the last line is over all chosen correlators of the fields and we have also dropped the field indices on this line .
we can repeat the calculation for the number of ways to split the total correlator of ( [ app : reducednptfn ] ) into the appropriate connected pairs .
the only new point is that whenever an @xmath39 appears more then once in the same @xmath165 the order in which we pick those @xmath142 s does not matter , so we have to divide by the factorial of the number of @xmath142 s raised to the power @xmath165 . in total
we get , where the product in the denominator is over all @xmath49 s and the number of @xmath39 s repeated in each individual @xmath49 . therefore this choice of @xmath49 s gives the following contribution to ( [ app : nptfn ] ) , & & _ p , q,,r n_abcn_def n_ghi + & & p^abgh _ p - n_x_r - n_z + & = & _ _ n_x__n_zp^abgh .
repeating this for all allowed choices of the @xmath49 s this gives [ app : nptfnng ] _
x_y_z= _ l_xy,,l_xx z _ _ n_x__n_zp^abi .
( [ app : nptfnng ] ) explains the numerical factor of rule 7 in sec .
[ sec : ng ] , except for the case of dressed vertices .
the numerical factor for dressed vertices requires a calculation of the number of ways to reduce the @xmath0-point function in ( [ app : generalisednt ] ) into connected pairs of the required order .
if @xmath49 correlators of order @xmath0 each dress the same vertex then we need to know the number of ways to split the @xmath166-point function into @xmath49 lots of connected @xmath0-point functions .
the result is ^lnc_n^n(l-1)c_n^2nc_n = , where the denominator of @xmath51 arises because the order in which we pick the @xmath0-point functions does not matter .
the factor of @xmath167 is cancelled by the numerical factor in ( [ app : generalisednt ] ) .
the generalisation to more complicated dressed vertices is straightforward to calculate , for example the number of ways to split an @xmath168-point function into @xmath169 lots of connected @xmath170-point functions and @xmath171 lots of connected @xmath172-point functions is . in fig .
[ fig : ntng ] we give two examples of these rules , both of which are 4 loop terms of the dressed vertex of @xmath173 , the first diagram has @xmath174 and @xmath175 while the second diagram has @xmath176 and @xmath177 .
the associated mathematical expressions are & & [ ntngeg1 ] n_abcdefg^b^c^d^e^f^g , + & & n_abcdefgh^b^c^d^e^f^g^h .
[ ntngeg2 ] v. acquaviva , n. bartolo , s. matarrese and a. riotto , nucl .
b * 667 * , 119 ( 2003 ) [ arxiv : astro - ph/0209156 ] .
j. m. maldacena , jhep * 0305 * , 013 ( 2003 ) [ arxiv : astro - ph/0210603 ]
. b. a. bassett , s. tsujikawa and d. wands , rev .
phys . * 78 * , 537 ( 2006 ) [ arxiv : astro - ph/0507632 ] .
k. enqvist and m. s. sloth , nucl .
b * 626 * , 395 ( 2002 ) [ arxiv : hep - ph/0109214 ] ; d. h. lyth and d. wands , phys .
b * 524 * , 5 ( 2002 ) [ arxiv : hep - ph/0110002 ] ; t. moroi and t. takahashi , phys . lett .
b * 522 * , 215 ( 2001 ) [ erratum - ibid .
b * 539 * , 303 ( 2002 ) ] [ arxiv : hep - ph/0110096 ] ; k. enqvist and s. nurmi , jcap * 0510 * , 013 ( 2005 ) [ arxiv : astro - ph/0508573 ] ; k. a. malik and d. h. lyth , jcap * 0609 * , 008 ( 2006 ) [ arxiv : astro - ph/0604387 ] ; m. sasaki , j. valiviita and d. wands , phys .
d * 74 * , 103003 ( 2006 ) [ arxiv : astro - ph/0607627 ] .
e. silverstein and d. tong , phys . rev .
d * 70 * , 103505 ( 2004 ) [ arxiv : hep - th/0310221 ] ; m. alishahiha , e. silverstein and d. tong , phys .
d * 70 * , 123505 ( 2004 ) [ arxiv : hep - th/0404084 ] ; m. x. huang and g. shiu , phys .
d * 74 * , 121301 ( 2006 ) [ arxiv : hep - th/0610235 ] .
l. verde , l. m. wang , a. heavens and m. kamionkowski , mon . not .
soc . * 313 * , l141 ( 2000 ) [ arxiv : astro - ph/9906301 ] .
d. h. lyth , k. a. malik and m. sasaki , jcap * 0505 * , 004 ( 2005 ) [ arxiv : astro - ph/0411220 ] .
d. h. lyth and y. rodriguez , phys .
lett . * 95 * , 121302 ( 2005 ) [ arxiv : astro - ph/0504045 ] .
d. s. salopek and j. r. bond , phys .
d * 42 * , 3936 ( 1990 ) .
m. sasaki and t. tanaka , prog .
* 99 * , 763 ( 1998 ) [ arxiv : gr - qc/9801017 ] .
d. wands , k. a. malik , d. h. lyth and a. r. liddle , phys .
d * 62 * , 043527 ( 2000 ) [ arxiv : astro - ph/0003278 ] .
r. scoccimarro and j. frieman , astrophys .
j. suppl .
* 105 * , 37 ( 1996 ) [ arxiv : astro - ph/9509047 ] .
m. crocce and r. scoccimarro , phys .
d * 73 * , 063519 ( 2006 ) [ arxiv : astro - ph/0509418 ] . i.
zaballa , y. rodriguez and d. h. lyth , jcap * 0606 * , 013 ( 2006 ) [ arxiv : astro - ph/0603534 ] .
l. boubekeur and d. h. lyth , phys .
d * 73 * , 021301 ( 2006 ) [ arxiv : astro - ph/0504046 ] . m. musso , arxiv : hep - th/0611258 . m. x. huang , g. shiu and b. underwood , arxiv:0709.3299 [ hep - th ] .
d. seery and j. e. lidsey , jcap * 0509 * , 011 ( 2005 ) [ arxiv : astro - ph/0506056 ] .
d. seery , j. e. lidsey and m. s. sloth , jcap * 0701 * , 027 ( 2007 ) [ arxiv : astro - ph/0610210 ] . b. j. w. van tent , class . quant .
* 21 * , 349 ( 2004 ) [ arxiv : astro - ph/0307048 ] .
c. t. byrnes and d. wands , phys .
d * 74 * , 043529 ( 2006 ) [ arxiv : astro - ph/0605679 ] .
p. r. jarnhus and m. s. sloth , arxiv:0709.2708 [ hep - th ] .
d. h. lyth and i. zaballa , jcap * 0510 * , 005 ( 2005 ) [ arxiv : astro - ph/0507608 ] .
d. binosi and l. theussl , comput .
commun . * 161 * , 76 ( 2004 ) [ arxiv : hep - ph/0309015 ] .
d. h. lyth , arxiv:0707.0361 [ astro - ph ] .
d. n. spergel _ et al . _
[ wmap collaboration ] , astrophys .
j. suppl .
* 170 * , 377 ( 2007 ) [ arxiv : astro - ph/0603449 ] .
m. sasaki , j. valiviita and d. wands , phys .
d * 74 * , 103003 ( 2006 ) [ arxiv : astro - ph/0607627 ] .
d. seery , arxiv:0707.3377 [ astro - ph ] .
f. vernizzi and d. wands , jcap * 0605 * , 019 ( 2006 ) [ arxiv : astro - ph/0603799 ] .
t. battefeld and r. easther , jcap * 0703 * , 020 ( 2007 ) [ arxiv : astro - ph/0610296 ] .
k. y. choi , l. m. h. hall and c. van de bruck , jcap * 0702 * , 029 ( 2007 ) [ arxiv : astro - ph/0701247 ] .
s. a. kim and a. r. liddle , phys .
d * 74 * , 063522 ( 2006 ) [ arxiv : astro - ph/0608186 ] .
d. battefeld and t. battefeld , jcap * 0705 * , 012 ( 2007 ) [ arxiv : hep - th/0703012 ] . m. sasaki , class . quant .
* 24 * , 2433 ( 2007 ) [ arxiv : astro - ph/0702182 ] .
l. alabidi and d. h. lyth , jcap * 0605 * , 016 ( 2006 ) [ arxiv : astro - ph/0510441 ] .
g. i. rigopoulos , e. p. s. shellard and b. j. w. van tent , arxiv : astro - ph/0511041 .
s. yokoyama , t. suyama and t. tanaka , arxiv:0705.3178 [ astro - ph ] .
g. dvali , a. gruzinov and m. zaldarriaga , phys .
d * 69 * , 023505 ( 2004 ) [ arxiv : astro - ph/0303591 ] ; l. kofman , arxiv : astro - ph/0303614 .
n. barnaby and j. m. cline , phys .
d * 73 * , 106012 ( 2006 ) [ arxiv : astro - ph/0601481 ] ; n. barnaby and j. m. cline , phys .
d * 75 * , 086004 ( 2007 ) [ arxiv : astro - ph/0611750 ] .
s. weinberg , phys .
d * 74 * , 023508 ( 2006 ) [ arxiv : hep - th/0605244 ] .
x. chen , m. x. huang , s. kachru and g. shiu , jcap * 0701 * , 002 ( 2007 ) [ arxiv : hep - th/0605045 ] .
|
we present feynman type diagrams for calculating the @xmath0-point function of the primordial curvature perturbation in terms of scalar field perturbations during inflation .
the diagrams can be used to evaluate the corresponding terms in the @xmath0-point function at tree level or any required loop level .
rules are presented for drawing the diagrams and writing down the corresponding terms in real space and fourier space .
we show that vertices can be renormalised to automatically account for diagrams with dressed vertices .
we apply these rules to calculate the primordial power spectrum up to two loops , the bispectrum including loop corrections , and the trispectrum .
|
the propagation of vibrational excitations in disordered media is one of the important problems in condensed matter physics
. microscopic nature of these vibrational excitations is still poorly understood , despite that they are responsible for the specific heat , the thermal conductivity and the sound propagation in amorphous materials . in 1999
allen and feldman proposed a simple classification of vibrations in disordered media @xcite , which was based on the numerical simulations of amorphous silicon .
the low - frequency vibrations are longitudinal ( la ) or transverse ( ta ) plane waves ( acoustic phonons ) .
the mean free path of phonons decreases rapidly with frequency due to structural or dynamical disorder . at some frequency @xmath2
thz the mean free path of phonons becomes comparable with its wavelength .
it is a so - called ioffe - regel criterion .
however , the mobility edge @xmath3 for vibration localization is much higher then @xmath4 what is not the case for electron localization .
vibrations in amorphous materials in this relatively wide frequency range @xmath5 are delocalized but not propagated as phonons .
these vibrations were called _ diffusons _ because they spread by means of diffusion @xcite .
however , this picture can not clearly explain the prominent peak ( more than 3 times in the value ) in the diffusivity of vibrations in amorphous silicon between 7 and 13 thz @xcite .
allen and feldman noticed that the sharp rise of the diffusivity at 7 thz corresponds to the end of ta branch in crystalline si which is responsible for the local minimum in the vibrational density of states .
so the frequency 7 thz marks a point in the spectrum where vibrations change their character from somewhat ta - like , to somewhat la - like with much larger group velocity
. however the notions of transverse and longitudinal vibrations are ill defined for such big frequencies in amorphous silicon . in this paper
we will generalize the notions of transverse and longitudinal vibrations for disordered systems like amorphous silicon .
as an example we study numerically the model amorphous silicon ( a - si ) system consisting of @xmath6 atoms contained in a periodic cubic box of lengths @xmath7 of approximately 87 .
the technical details of the preparation of the a - si sample have already been presented in ref .
the si - si interaction in the amorphous silicon studied here is well described by the stillinger - weber potential @xcite .
low - frequency vibrations even in amorphous media are well - defined plane longitudinal and transverse waves . these vibrations can be described in continuous medium approximation .
the displacement fields @xmath8 for longitudinal ( l ) and transverse ( t ) waves have a form @xmath9 however , this definition contains the phonon wavevector @xmath0 , which is ill - defined for high - frequency vibrations in amorphous medium . to generalize this approach
let us note that @xmath10 the transverse displacement @xmath11 has zero divergence so it conserves local volumes .
the natural analog of the local volumes in amorphous media are voronoi cells around each atom .
the voronoi cell @xmath12 , associated with the atom @xmath13 is the set of all points in the space around this atom whose distance to the atom position @xmath14 is not greater than their distance to the other atoms @xmath15 @xcite .
this type of cells is also known as wigner - seitz cells in crystallography for regular crystals .
figure [ fig : sql ] shows a simple example of voronoi cells for longitudinal and transverse waves in a simple quadratic lattice .
a longitudinal wave evidently changes the volumes of voronoi cells while the transverse wave does not change the volumes despite the change in the form of the cells . a displacement of atoms in amorphous media may ( or may not ) change volumes of voronoi cells .
we will call the displacement of atoms @xmath16 to be transverse if it does not change the volumes of all voronoi cells . for
that let us introduce a matrix @xmath17 which is responsible for the relative change of the @xmath13th voronoi cell volume @xmath18 under @xmath19th atom displacement in the direction @xmath20
@xmath21 the explicit formula for the matrix @xmath17 will be derived in the next section . using this matrix the displacement of @xmath19th atom in the direction @xmath20
, @xmath22 produces the following relative change of the voronoi cell volumes @xmath23 in the matrix notation this equation reads @xmath24 where @xmath17 is a rectangular @xmath25 matrix ( with @xmath26 being the number of particles ) and @xmath27 is a displacement vector with @xmath28 elements .
the matrix @xmath17 is a discrete analog of the divergence operator . by definition
the transverse component @xmath29 of the displacement @xmath27 is such that @xmath30 , i.e. @xmath29 is the projection of the displacement @xmath27 to the _ null space _ of the matrix @xmath17 .
the longitudinal component @xmath31 is a remaining orthogonal component of the displacement field and it is the projection of @xmath27 to the _ row space _ of the matrix @xmath17 .
these projections have the following form @xcite @xmath32 indeed , @xmath29 does not change the voronoi cell volumes @xmath33 and @xmath31 is orthogonal to @xmath29 @xmath34
let us show how matrix @xmath17 can be derived from geometry only .
let we shift only the atom @xmath19 by the vector @xmath35 ( figure [ fig : vor ] ) .
the vector @xmath35 is small so the voronoi cells after shifting have almost the same structure , but their facets are slightly shifted and rotated in space .
the shifting @xmath35 of the atom @xmath19 can change the volume @xmath18 of a nearest neighbor cell @xmath13 . by definition the cells @xmath13 and @xmath19 are nearest neighbors if they have a common facet @xmath36 . by definition of the voronoi cell
the facet @xmath36 lies in the plane @xmath37 , which has the normal @xmath38 , where @xmath39 is the vector connecting atoms @xmath13 and @xmath19 and go through the point @xmath40 .
therefore the equation for this plane reads @xmath41 after shifting of the atom @xmath19 the new facet @xmath42 lies in the plane @xmath43 which satisfy the equation @xmath44 where @xmath45 , @xmath46 , and @xmath47 .
the signed distance from an arbitrary point @xmath48 to the plane @xmath43 is @xmath49 this distance has the sign `` @xmath50 '' if @xmath48 lies on the same side to the plane @xmath43 as the atom @xmath13 and the sign `` @xmath51 '' if @xmath48 lies on the opposite side to the plane .
if @xmath48 lies in the plane @xmath43 then the distance @xmath52 is equal to 0 and we get the eq .
( [ eq : pl ] ) . in the linear approximation on @xmath53 the change of the volume
@xmath18 is the integral of @xmath52 over the surface of the initial facet @xmath36 @xmath54 where @xmath55 and @xmath56 are correspondingly the area and the centroid of the facet @xmath36 . in the linear approximation on @xmath53 , the eq .
( [ eq : dvi1 ] ) reads @xmath57 where we have taken into account that @xmath58 because @xmath59 . and
@xmath60 denote the position of @xmath61 and @xmath56 respectively . in two dimensional case the point @xmath60
is the center of the segment @xmath36 .
b ) dashed lines shows the voronoi cells after shifting of the atom @xmath19 by the vector @xmath35 .
the point @xmath62 denotes the position of @xmath63 .
green and red areas show the increasing and decreasing of the @xmath18 respectively . ] if all neighboring atoms shift , the change of the volume @xmath18 can be written in the matrix form @xmath64 the nondiagonal elements of the matrix @xmath65 we have already found above @xmath66 the diagonal element @xmath67 means the change of the @xmath13th voronoi cell volume under shifting of the @xmath13th atom itself .
shifting of the all atoms by the same vector does not change the volumes of the voronoi cells .
therefore @xmath68 after dividing by the volume we finally get the matrix @xmath17 @xmath69 it is notable that finite - elements methods also often use voronoi cells and have a similar definition for the finite differences for the divergence operator @xcite .
( red line ) . the relative number of the transverse modes @xmath70 is shown by green hatching between red line and the value 1 . ]
the total vibrational density of states can be decomposed to longitudinal and transverse components @xmath71 where @xmath72 is an eigenfrequency , @xmath73 is the corresponding eigenmode , @xmath74 and @xmath75 are its transverse and longitudinal components defined by eqs .
( [ eq : prjns ] ) and ( [ eq : prjrs ] ) .
the total vibrational density of states and its longitudinal and transverse components was calculated for the numerical model of a - si consisting of @xmath6 atoms by the kernel polynomial method ( kpm ) @xcite .
the results are shown in the figure [ fig : doslt ] .
there are three clearly seen regions .
the vibrations are mostly transverse below 7 thz and above 15 thz . in the frequency interval in between the vibrations have the longitudinal nature .
the low number of longitudinal vibrations in the low - frequency region is explained by the debye law , which gives @xmath76 in our model the amorphous si has the longitudinal sound velocity @xmath77 and the transverse sound velocity @xmath78 which gives the ratio @xmath79 for @xmath80 .
this value coincides with the inset in the figure [ fig : doslt ] .
the domination of the longitudinal modes between 7 thz and 15 thz in amorphous silicon corresponds to the gap between the upper frequency of ta modes ( 7.5 thz ) and the lower frequency of to modes ( 13.9 thz ) in crystalline si @xcite .
this frequency region in crystalline si is totally occupied by la and lo modes without a gap . in the same frequency region
the vibrations of amorphous si have a small transverse component ( 1520% ) .
certainly , there are no optical phonons with well defined wavevector in amorphous silicon due to relatively strong disorder
. however , the short - range order of vibrational modes in amorphous phase can be similar to that in crystal phase . at the same time , our definition of the longitudinal and transverse vibrations is local because it is based on the voronoi cells , which depends on the neighbor atoms only .
the figure [ fig : doslt ] shows a sharp change of the nature of vibrations at 7 thz from almost transverse to almost longitudinal .
the longitudinal vibrations correspond to the stretching of the chemical bonds between atoms while the transverse vibrations correspond to the less rigid bond bending and rocking .
therefore longitudinal vibrations can transfer the vibrational energy much faster than transverse vibrations which leads to the sudden rise of the diffusivity at 7 thz observed in refs .
furthermore , our additional calculations show that the longitudinal modes still have a well - defined wavevector and a sound velocity up to @xmath81 thz unlike the transverse modes , which have the frequency of the ioffe - regel crossover @xmath82 thz .
we have shown that vibrations in amorphous solids can be decomposed to transverse and longitudinal components in a generalized sense . the transverse component does not change the volumes of voronoi cells around atoms while the longitudinal component is the remaining orthogonal component changing this volume .
we show that the rise of the diffusivity at 7 thz in amorphous silicon corresponds to the sharp change of the nature of vibrations from almost transverse to almost longitudinal ones having high sound velocity
. this decomposition can be fruitful for the investigation of the correlation function and the diffusivity of transverse and longitudinal components of vibrations in amorphous solids independently from each other .
the 3d - structure of several voronoi cells in amorphous silicon is available online in supplementary materials .
one of the authors ( ymb ) thanks the dynasty foundation for the financial support .
two authors ( dap and ymb ) thanks the university lyon 1 for hospitality .
99 allen p b , feldman j l , fabian j and wooten f 1999 _ phil . mag .
_ b * 79 * 1715 beltukov y m , kozub v i and parshin d a 2013 _ phys
134203 feldman j l , allen p b and bickham s r 1999 _ phys .
_ b * 59 * 3551 fusco c , albaret t and tanguy a 2010 _ phys . rev .
_ e * 82 * 066116 stillinger f h and weber t a 1985 _ phys .
_ b * 31 * 5262 aurenhammer f 1991 _ acm computing surveys _ * 23 * 345 meyer c d 2000 _ matrix analysis and applied linear algebra _
( siam ) mishev i d 1998 _ numer .
meth . part .
d. e. _ * 14 * 193 weie a et al .
2006 _ rev .
mod . phys_. * 78 * 275 tubino r , piseri l and zerbi g 1972 _ j. chem .
phys . _ * 56 * 1022
|
we show that harmonic vibrations in amorphous silicon can be decomposed to transverse and longitudinal components in all frequency range even in the absence of the well defined wave vector @xmath0 . for this purpose
we define the transverse component of the eigenvector with given @xmath1 as a component , which does not change the volumes of voronoi cells around atoms .
the longitudinal component is the remaining orthogonal component .
we have found the longitudinal and transverse components of the vibrational density of states for numerical model of amorphous silicon .
the vibrations are mostly transverse below 7 thz and above 15 thz . in the frequency interval in between the vibrations have a longitudinal nature .
just this sudden transformation of vibrations at 7 thz from almost transverse to almost longitudinal ones explains the prominent peak in the diffusivity of the amorphous silicon just above 7 thz .
|
self - consistent mean - field models with effective energy - density functionals ( skyrme - hartree - fock , gogny , relativistic ) are established as reliable tools for description of nuclear structure and dynamics , for a comprehensive review see @xcite . in particular , there is a rising implementation of these models to dynamical features of nuclei ( see , e.g. @xcite ) , which is caused , to a large extend , by exploiting the spectra and reaction rates of exotic nuclei in astrophysics @xcite .
however , applications of self - consistent models to nuclear dynamics are still limited even in the linear regime which is usually treated within the random - phase - approximation ( rpa ) .
the calculations are plagued by dealing with high - rank rpa matrices .
this is especially painful for deformed systems where lack of symmetry requires a huge one - particle - one - hole @xmath3 configuration space .
the rpa problem becomes much simpler if the residual two - body interaction is factorized ( i.e. reduced to a separable form ) : @xmath4 where @xmath5 is a hermitian @xmath6 operator and @xmath7 is a matrix of strength constants .
the factorization allows to reduce a high - rank rpa matrix matrix to much smaller one with a rank 4k ( where the coefficient 4 is caused by the isospin and time - parity factors , see discussion at the next section ) .
the separable expansion can be formulated in a fully self - consistent manner and provide a high accuracy with a small number ( k @xmath8 ) of the separable terms .
several self - consistent schemes for separable expansions have been proposed during the last decades @xcite . however , these schemes were not sufficiently general . some of them are limited to analytical or simple numerical estimates @xcite , the others are not fully self - consistent @xcite . between them
the approach @xcite for skyrme forces is quite promising .
however , it still deals with rpa matrices of rather high rank ( @xmath9 ) .
recently , we have developed a general self - consistent separable rpa ( srpa ) approach applicable to arbitrary density- and current - dependent functionals @xcite .
the method was implemented to the skyrme functional @xcite for both spherical @xcite and deformed @xcite nuclei . in srpa ,
the one - body operators @xmath10 and associated strengths @xmath11 are unambiguously derived from the given energy - density functional . there remain no further adjustable parameters .
the success of the expansion depends on an appropriate choice of the basic operators @xmath10 .
experience with spherical srpa gives guidelines for an efficient choice such that a few separable terms suffice to reproduce accurately the exact rpa spectra @xcite .
the success becomes possible due to the following factors : i ) an efficient self - consistent procedure @xcite based on sound physical arguments , ii ) proper inclusion of all parts of the residual interaction , time - even as well as time - odd couplings , iii ) incorporation of the symmetries ( translation , particle number , ... ) leading to a correct description of the related zero - energy modes , iv ) building the separable operators by such a way that they have maxima at different slices of the nucleus and thus cover both surface and interior dynamics .
furthermore , we note that srpa is very general and can be applied to a wide variety of finite fermion systems .
for example , srpa was derived for the kohn - sham functional @xcite and widely used for description of linear dynamics of valence electrons in spherical and deformed atomic clusters @xcite .
the enormous reduction of computational expense by srpa is particularly advantageous for deformed systems where 1ph configuration space grows huge .
srpa becomes here a promising tool for large scale studies .
it is the aim of this paper to present a first implementation of srpa for axially deformed nuclei with pairing . like the full rpa
, srpa follows two strategies to compute the dynamical response .
it can be calculated via determination of rpa eigenvalues and eigenstates or by a direct computation of the multipole strength functions related to experimental cross sections .
we will discuss both ways .
as a first test , we will apply the method to description of isovector e1 and isoscalar e2 giant resonance ( gr ) in the deformed nuclei @xmath0sm , @xmath1u , and @xmath2no .
these nuclei cover a broad size range from rare - earth to super - heavy elements .
four different skyrme forces ( skt6 @xcite , skm * @xcite , sly6 @xcite , and ski3 @xcite ) will be used to explore dependence of the results on the actual parametrization . for our aims
it is important that these forces represent different values of some relevant nuclear matter characteristics ( isoscalar and isovector effective masses and asymmetry energy ) .
we will discuss dependence of the collective strength on these characteristics , scrutinize the impact of time - odd coupling terms in the skyrme residual interaction and demonstrate important role of the landau fragmentation .
the paper is organized as follows .
the general srpa formalism is presented in sec .
[ sec : srpa ] and specified for skyrme forces in sec .
[ sec : skyrme ] . in sec .
[ sec : q_choice ] we discuss the choice of the input operators .
results of the calculations are analyzed in sec .
[ sec : results ] .
the summary is done in sec .
[ sec : summary ] .
some important details of the method are given in appendices a - c .
for more details , see documentation in @xcite .
in the most general form , the factorization of the residual interaction ( [ 1 ] ) reads @xmath12 where the indices @xmath13 and @xmath14 label neutrons and protons , @xmath15 are time - even hermitian one - body operators and @xmath16 are their time - odd counterparts , @xmath17 and @xmath18 are the strength matrices .
time reversal properties of the operators are @xmath19 t\hat{y}_{sk}t^{-1 } & = & \gamma_t^y \hat{y}_{sk } & , & \gamma_t^y=-1 & , \end{array}\ ] ] where @xmath20 is the operator of time reversal .
the expansion ( [ v_sep ] ) needs to take care of both classes of operators since the relevant skyrme functionals involve both time - even and time - odd couplings , see @xcite and appendix [ sec : dens_curr ] . though time - odd variables do not contribute to the
mean field hamiltonian of spin - saturated systems , they can play a role in time - dependent perturbations and nuclear dynamics .
altogether , the presence of time - even and time - odd couplings in the functional naturally leads to a formulation of the separable model in terms of hermitian operators with a given time - parity .
these operators have the useful property @xmath21>=0 } \langle [ \hat{a},\hat{b } ] \rangle \sim ( 1-\gamma_t^a\gamma_t^b)\ ] ] which means that the average value of the commutator at the ground state @xmath22 is not zero only for operators of the opposite t - parities @xmath23 .
this property will be widely used in the following .
rpa is the limit of small - amplitude harmonic vibrations around the ground state .
the dynamics is formulated in general by a time - dependent variation on the basis of a given energy functional @xmath24 @xmath25 we deal with a set of densities @xmath26 where @xmath27 denotes the type of density ( spatial density , kinetic density , current , spin density , spin - orbit density , ... ) and @xmath13 labels protons and neutrons . for reasons of compact notation , we combine the density type @xmath27 with the space point @xmath28 into one index @xmath29 such that @xmath30 the densities are related to the corresponding one - body operators @xmath31 ( see the list in appendix [ sec : dens_curr ] ) as @xmath32 further , the state @xmath33 is the underlying bardeen - cooper - schrieffer ( bcs ) mean - field state composed from the single - particle states @xmath34 and the corresponding pairing occupation amplitudes @xmath35 .
the time - evolution is determined by variation of the action @xmath36 .
till now we keep the occupation amplitudes @xmath35 fixed at their ground state values and consider the variation of the single particle states .
this yields the time - dependent mean - field equations as @xmath37 with the mean field hamiltonain @xmath38 being a functional of the local and instantaneous densities @xmath39 .
the freezing of the occupation amplitudes @xmath35 somewhat inhibits application of srpa for vibrational modes with a strong pairing impact ( e.g. for low - lying modes in neutron - rich light deformed nuclei ) @xcite . however , in the present study we are interested in giant resonances where pairing dynamics plays a minor role . in the linear regime of small amplitude oscillations , the time - dependent state consists of the static ground state @xmath40 and a small time - dependent perturbation @xmath41 where both @xmath42 and @xmath22 are bcs states .
hence , all dynamical quantities can be decomposed as a sum of the stationary ground state and small time - dependent parts : [ eq : dens_j(t ) ] @xmath43 inserting ( [ eq : dens_j(t ) ] ) into ( [ eq : e(j ) ] ) and keeping the terms up to the first order in @xmath44 , one obtains the single - particle hamiltonian @xmath45 with the static mean - field part @xmath46 and the time - dependent response @xmath47 related to the oscillations of the system .
note that the second functional derivative is to be taken at the ground state density as indicated by the index @xmath48 .
this holds for all second functional derivatives and so we will skip this explicit index in the following . for the brevity of notation , we will also skip the explicit dependencies on space coordinates and come back to these details in sec . [
sec : skyrme ] where the residual interaction for the skyrme functional is worked out .
the linearized equation of motion reads @xmath49 further steps require a more specific view of @xmath50 .
this will be done in the next subsections from different points of view , macroscopic and microscopic .
it is convenient to obtain the perturbed mean - field state @xmath51 by the scaling transformation @xcite @xmath52 \exp [ -ip_{sk}(t)\hat{q}_{s k } ] | \rangle_s\ ] ] where both @xmath53 and @xmath54 are the slater determinants and @xmath55 and @xmath56 are generalized coordinate ( time - even ) and momentum ( time - odd ) hermitian one - body operators .
these operators fulfill the properties [ eq : q_p ] @xmath57=\hat{p}_{sk}^{\dagger},\quad \gamma_t^p=-1\end{aligned}\ ] ] where @xmath58 stands for the full hamiltonian embracing both the one - body mean - field hamiltonian and the two - body residual interaction .
the commutator in ( [ eq : p ] ) is assumed to be mapped into the one - body domain ( see , e.g. , the mapping into @xmath59 in eq .
( [ eq : h_xy ] ) ) .
if the functional includes only time - even densities , then @xmath60 does not contribute to the commutator and so @xmath61 can be safely replaced by @xmath62 .
the operators ( [ eq : q_p ] ) generate time - even and time - odd real collective deformations @xmath63 and @xmath64 . using ( [ eq : scaling ] ) and assuming only small deformations , the transition densities ( [ eq : j(t)_psi ] ) read @xmath65 \rangle \\ & & \qquad\ ; + p_{s k}(t ) \langle [ \hat{q}_{s k},\hat j_{s}^{\alpha } ] \rangle\ } \nonumber\end{aligned}\ ] ] where , following ( [ eq:<[a , b]>=0 ] ) , time - even densities contribute only to responses @xmath66|\rangle$ ] while time - odd ones only to @xmath67|\rangle$ ] .
then the response hamiltonian ( [ eq : h_resp ] ) recasts as @xmath68 where the time dependence is concentrated in the amplitudes @xmath63 and @xmath64 while all time - independent terms are collected in the hermitian one - body operators [ eq : x_y ] @xmath69 \rangle { \hat j}_{s'}^{\alpha ' } , \ ] ] @xmath70 \rangle { \hat j}_{s'}^{\alpha'}\ ] ] with the properties [ xy - properties ] @xmath71 obviously , @xmath15 and @xmath16 are the reasonable candidates for the time - even and time - odd operators in the separable expansion ( [ v_sep ] ) .
the operator @xmath15 involves contributions only from the time - even densities while the operator @xmath16 only from the time - odd ones .
the upper index @xmath14 in the operators ( [ eq : x_y ] ) determines the isospin ( proton or neutron ) subspace where these operators act .
this is the domain of the involved density operators @xmath72 . to complete the construction of the separable expansion ( [ v_sep ] )
, we should yet determine the matrices of the strength constants @xmath17 and @xmath18 .
this can be done through variations of the basic operators [ eq : xyvar ] @xmath73 \rangle = - \sum_{s'k ' } q_{s'k'}(t ) \kappa_{s'k',sk}^{-1 } \ ; , \\ \label{eq : yvar } \delta { y}_{s k}(t ) & = & \langle \psi(t)|{\hat y}_{sk}|\psi(t ) \rangle - \langle { \hat y}_{sk } \rangle = i \sum_{s'k ' } p_{s'k'}(t ) \langle [ \hat{q}_{s'k'},\hat y_{sk}^{s ' } ] \rangle = - \sum_{s'k ' } p_{s'k'}(t ) \eta_{s'k',sk}^{-1}\end{aligned}\ ] ] where [ eq : kappa - eta ] @xmath74 \rangle = \sum_{\alpha \alpha ' } \frac{\delta^2 e}{\delta j_{s'}^{\alpha'}\delta j_{s}^{\alpha } } \langle [ \hat{p}_{s k},{\hat j}_{s}^{\alpha } ] \rangle \langle [ \hat{p}_{s ' k'},{\hat j}_{s'}^{\alpha ' } ] \rangle \ ; , \\ \label{eq : eta } \eta_{s'k',sk}^{-1 } = \eta_{sk , s'k'}^{-1 } & = & -i \langle [ \hat{q}_{s'k'},{\hat y}_{sk}^{s ' } ] \rangle = \sum_{\alpha \alpha ' } \frac{\delta^2 e } { \delta j_{s'}^{\alpha'}\delta j_{s}^{\alpha } } \langle [ \hat{q}_{s k},{\hat j}_{s}^{\alpha } ] \rangle \langle [ \hat{q}_{s ' k'},{\hat j}_{s'}^{\alpha ' } ] \rangle.\end{aligned}\ ] ] eqs .
( [ eq : kappa - eta ] ) represent the elements of the symmetric matrices which are inverse to the matrices of the strength constants in ( [ v_sep ] )
. indeed , eqs .
( [ eq : xyvar ] ) can be recast to @xmath75 from that we read off the response hamiltonian ( [ eq : h_xy ] ) as @xmath76 which leads within the collective space of the generators @xmath77 to the same eigenvalue problem as the separable hamiltonian @xmath78 with @xmath79 given in eq .
( [ v_sep ] ) ( see @xcite ) . in principle
, we already have in our disposal the macroscopic srpa formalism for linear regime of the collective motion in terms of real harmonic variables [ eq : q_p ] @xmath80 indeed , eqs .
( [ eq : x_y ] ) and ( [ eq : kappa - eta ] ) deliver the one - body operators and strength matrices for the separable expansion of the two - body interaction . by substituting the response hamiltonian ( [ eq : h_dxdy ] ) and the perturbed wave function ( [ eq : scaling ] ) into time - dependent hf equation ( [ eq : lintdhf ] ) one gets the eigenvalue problem .
the number @xmath81 of the collective variables ( and thus of the separable terms ) is dictated by the accuracy we need in the description of collective modes .
for @xmath82 , the method in fact is reduced to the sum rule approach with one collective mode @xcite . for @xmath83
, we have a system of k coupled oscillators and the method becomes similar to so - called local rpa @xcite suitable for description of branching and gross - structure of collective modes .
the macroscopic srpa as outlined in section [ sec : macro_srpa ] serves here as a convenient tool to derive the optimal separable expansion . but it can not describe the landau fragmentation of the collective strength . for this aim , we should build the microscopic part of the model . in what follows , we will consider the eigenvalue problem and the direct computation of the strength function .
the perturbation @xmath50 belongs to the tangential space of the variations of a mean field state . for the pure slater states , they are all conceivable one - particle - one - hole ( @xmath6 ) excitations @xmath84 in the bcs case ,
the elementary modes are reduced to the two - quasiparticle ( @xmath85 ) excitations @xmath86 where @xmath87 generate the bcs quasiparticle state @xmath88 and @xmath89 is its time reversal .
just this case is employed in our actual calculations .
both @xmath6 and @xmath85 excitations have much in common and result in the same microscopic srpa equations ( for exception of pairing peculiarities outlined in appendix [ sec : pairing ] ) .
so , in what follows , we will not distinguish these two cases .
in particular , we will use for both excitations one and the same notation @xmath90 . to formulate the eigenvalue problem
, we will exploit the standard rpa technique for the separable hamiltonian ( [ eq : h_sep ] ) where the separable operators and strength constants are delivered by the macroscopic srpa , see eqs .
( [ eq : x_y ] ) and ( [ eq : kappa - eta ] ) . following @xcite , the collective motion
is represented in terms of [ psi_thouless ] @xmath91 where @xmath92 creates the one - phonon eigenmode @xmath93 ; the operators @xmath94 and @xmath95 are defined in ( [ eq:1ph ] ) ; the perturbed @xmath96 and ground @xmath22 states have the form of slater determinants .
we employ here the thouless theorem @xcite which establishes connection between two arbitrary slater determinants .
the wave function ( [ psi_thouless ] ) is a microscopic counterpart of macroscopic scaling ansatz ( [ eq : scaling ] ) .
but here we aim a fully _ microscopic _ description of excitations covering all the @xmath6 space while in the previous sections we considered _ macroscopic _ flow as a benchmark for forming the separable interaction .
the time - dependent response hamiltonian @xmath97 for the mode @xmath98 and the separable interaction ( [ v_sep ] ) reads @xmath99 \rangle } _ { \mbox{$\bar{q}_{sk}^{\nu}$ } } \nonumber\\ & & + \sum_{sk}{\hat y}_{sk } \underbrace { \sum_{s'k'}\eta_{sk , s'k ' } \langle [ { \hat y}_{s'k'},\hat{c}^{\dagger}_{\nu } ] \rangle } _ { \mbox{-i$\bar{p}_{sk}^{\nu}$ } } \;,\end{aligned}\ ] ] where [ eq : barpqdef ] @xmath100 note that here the values @xmath101 and @xmath102 are new objects generated by the eigenmode @xmath92 .
they serve for notation of @xmath59 and have no relation to the collective generators @xmath103 , @xmath104 used in the section [ sec : macro_srpa ] . substituting the ansatz ( [ psi_thouless ] ) into the linearized time - dependent hf equation ( [ eq : lintdhf ] ) and employing the form ( [ eq : hres_sep ] ) for the response field yields the expression for @xmath6 expansion coefficients @xmath105 where @xmath106 is the unperturbed energy of @xmath6 pair . in the derivation above
we used the operator properties ( [ xy - properties ] ) .
then the matrix elements @xmath107 and @xmath108 are real and image , respectively , while all the unknowns @xmath109 , @xmath101 and @xmath102 are real . inserting the result ( [ eq : c_pm_qp ] ) into eqs .
( [ eq : barpqdef ] ) yields finally the set of srpa equations for the values @xmath101 and @xmath102 @xmath110 + \bar{p}_{\bar{s}\bar{k}}^{\nu } f_{s'k',\bar{s}\bar{k}}^{(xy ) } \}=0 \ ; , \nonumber\\ \sum_{\bar{s}\bar{k } } & \ { \bar{q}_{\bar{s}\bar{k}}^{\nu } f_{s'k',\bar{s}\bar{k}}^{(yx ) } + \bar{p}_{\bar{s}\bar{k}}^{\nu } [ f_{s'k',\bar{s}\bar{k}}^{(yy ) } - \eta_{\bar{s}\bar{k},s'k'}^{-1 } ] \}=0\end{aligned}\ ] ] where @xmath111 with @xmath112 and @xmath113 the system of linear homogeneous equations ( [ eq : x_c_qp4 ] ) has a non - trivial solution only if its determinant is zero .
this yields the dispersion equation to obtain the rpa eigenvalues @xmath114 . when exploring the response of the system to time - dependent external fields , we are often interested in the total strength function rather than in the particular rpa states .
for example , giant resonances in heavy nuclei are formed by thousands of rpa states whose contributions in any case can not be resolved experimentally . in this case , it is more efficient to consider a direct computation of the strength function which avoids the details and crucially simplifies the calculations . for an external electric field of multipolarity @xmath115 , we define the strength function as @xmath116 where @xmath117 is the lorentz weight with an averaging parameter @xmath118 and @xmath119 is the matrix element of @xmath115 transition from the ground state to the rpa state @xmath120 .
further , @xmath121 is an effective charge to be specified later .
the operator of the electric external field in the long - wave approximation reads @xmath122 the ( normalized ) amplitudes @xmath123 follow from ( [ eq : c_pm_qp ] ) . unlike the standard definition of the strength function with @xmath124 , we exploit here the lorentz weight .
it is convenient to simulate various smoothing effects .
the strength function ( [ eq : strength_function ] ) can be recast to the form which does not need information on the particular rpa states @xcite .
for this aim , we will use the cauchy residue theorem .
namely , the strength function is represented as a sum of the residues for the poles @xmath125 . since the sum of all the residues ( covering all the poles ) is zero , the residues with @xmath125
( whose calculation is time consuming ) can be replaced by the sum of residues with @xmath126 and @xmath127 whose calculation is much less expensive .
the explicit derivation is given in @xcite .
the final expression reads @xmath128_{z=\omega\!+\!i\delta /2}\ ] ] @xmath129 where @xmath130 means the image part of the value inside the brackets , @xmath131 is the determinant of the rpa matrix ( [ eq : x_c_qp4 ] ) with @xmath114 replaced by the complex argument @xmath132 , @xmath133 is the algebraic supplement of the determinant , and @xmath134 @xmath135 for the sake of brevity , we introduced in ( [ eq : sf_1 ] ) the new index @xmath136 where @xmath137=1 for time - even and 2 for time - odd quantities . for example , @xmath138 and @xmath139 . the first term in ( [ eq : sf_1 ] ) collects the contributions of the residual interaction .
it vanishes at @xmath140 .
the second term is the unperturbed ( purely two - quasiparticle ) strength function . before proceeding to further specification of the method
, it is worth to comment some its essential points .
+ @xmath141 to determine the unknowns @xmath142 of a full ( non - separable ) rpa , one requires diagonalization of the matrices of a high rank equal to the size of the @xmath6 basis .
the separable approximation allows to reformulate the rpa problem in terms of a few unknowns @xmath143 and @xmath144 and thus to reduce dramatically the computational effort . as is seen from ( [ eq : x_c_qp4 ] ) , the rank of the srpa matrix is equal to @xmath145 where @xmath81 is the number of the separable operators .
+ @xmath141 the number of the srpa states @xmath146 is equal to the number of the relevant @xmath6 configurations used in the calculations . in heavy nuclei
, this number can reach @xmath147-@xmath148 .
every state @xmath146 is characterized by the particular set of the values @xmath149 and @xmath150 .
+ @xmath141 eqs .
( [ eq : x_y ] ) and ( [ eq : kappa - eta ] ) relate the basic srpa values with the initial functional and input operators @xmath151 . after choosing @xmath151 ,
all other srpa equations are straightforwardly determined following the steps @xmath152\rangle \ ; \rightarrow \ ; \hat{y}_{sk } , \ ; \eta_{sk , s'k'}^{-1 } \ ; \rightarrow \ ; \hat{p}_{sk } \nonumber\\ & & \qquad \rightarrow \ ; \langle [ \hat{p}_{sk},{\hat j}^{\alpha}_s]\rangle \ ; \rightarrow \ ; \hat{x}_{sk } , \ ; \kappa_{sk , s'k'}^{-1 } .
\label{eq : q - initial}\end{aligned}\ ] ] as is discussed in sec .
[ sec : q_choice ] , the proper choice of @xmath151 is crucial to achieve a good convergence of the separable expansion ( [ v_sep ] ) with a minimal number of separable operators .
srpa itself does not provide a recipe to get @xmath151 but these operators can be introduced following intuitive physical arguments .
+ @xmath141 srpa restores the conservation laws ( e.g. translational invariance ) violated by the static mean field .
if a symmetry mode has a generator @xmath153 , then we keep the conservation law @xmath154=0 $ ] by including @xmath153 into the set of the input generators @xmath155 together with its complement @xmath156 $ ] .
+ @xmath141 the basic srpa operators can be expressed via the separable residual interaction ( [ v_sep ] ) as @xmath157_{ph } , \qquad \hat{y}_{sk } = -i [ \hat{v}^{\rm sep}_{\rm res } , \hat{q}_{sk}]_{ph}\ ] ] where the index @xmath90 means the @xmath6 part of the commutator .
it is seen that the time - odd operator @xmath155 retains the time - even part of @xmath158 to build @xmath159 .
vice versa , the commutator with the time - even operator @xmath151 keeps the time - odd part of @xmath158 to build @xmath160 .
we use the skyrme functional @xcite in the form @xcite @xmath161 where @xmath162^\frac{4}{3 } , \label{ecoul } \\ { \cal h}_{\rm pair}(\chi_s ) & = & \frac{1}{2}\sum_s v_{{\rm pair},s}\chi^*_s\chi^{\mbox{}}_s , \label{pair_functional}\end{aligned}\ ] ] @xmath163 are kinetic , coulomb , pairing and skyrme terms respectively .
the densities and currents used in this functional are defined in the appendices [ sec : dens_curr ] and [ sec : pairing ] .
densities without the index @xmath13 involve both neutrons and protons , e.g. @xmath164 .
parameters @xmath165 and @xmath29 are fitted to describe ground state properties of atomic nuclei , see e.g. @xcite . for the sake of brevity ,
we omit here derivation of the skyrme mean field which can be found elsewhere , e.g. in @xcite , but present only the main values entering the srpa equations .
the crucial ingredients of the skyrme srpa residual interaction are the second functional derivatives entering expressions for the basic operators ( [ eq : x_y ] ) and strength matrices ( [ eq : kappa - eta ] ) .
they read @xmath166 \nonumber \\ & & \quad - \delta_{s s_1 } \delta_{s p}\frac{1}{3}(\frac{2}{\pi})^{1/3}[\rho_p({\vec r})]^{-2/3 } \ } \ ; \delta ( { \vec r}-{\vec r}_1 ) \nonumber \\ & & \quad + \delta_{s s_1 } \delta_{s p } \;\frac{e^2}{|{\vec r}-{\vec r_1}| } \ ; , \\ \label{de_drho_dtau } & & \frac{\delta^2 e}{\delta \tau_{s_1 } ( { \vec r}_1)\delta \rho_s ( { \vec r } ) } = [ b_1 - b_1'\delta_{ss_1 } ] \delta ( { \vec r}-{\vec r}_1 ) \ ; , \\ \label{de_drho_dj } & & \frac{\delta^2 e}{\delta { \vec \im}_{s_1 } ( { \vec r}_1)\delta \rho_s ( { \vec r } ) } = [ b_4 + b_4'\delta_{ss_1 } ] { \vec \nabla}_{{\vec r}_1 } \delta ( { \vec r}-{\vec r}_1 ) \ ; , \\ \label{de_dj_drho } & & \frac{\delta^2 e}{\delta \rho_{s_1 } ( { \vec r}_1)\delta { \vec \im}_{s } ( { \vec r } ) } = - [ b_4 + b_4'\delta_{ss_1 } ] { \vec \nabla}_{{\vec r}_1 } \delta ( { \vec r}-{\vec r}_1)\end{aligned}\ ] ] for time - even densities and @xmath167 \delta_{mn } \delta ( { \vec r}_1-{\vec r } ) , \nonumber \\ \label{de_dj_dsigma } & & \frac{\delta^2 e}{\delta{\sigma}_{s_1,m}({\vec r}_1)\delta{j}_{s , n}({\vec r } ) } = \\ & & \qquad - [ b_4 + b_4 ' \delta_{s s_1}]\epsilon_{mnl } { \nabla}_{{\vec r}_1 , \ ; l}\delta ( { \vec r}_1-{\vec r } ) \nonumber\end{aligned}\ ] ] for time - odd densities .
the last two terms in ( [ de_drho_drho ] ) represent the exchange and direct coulomb contributions .
further , the indices @xmath168 in ( [ de_dj_dj ] ) and ( [ de_dj_dsigma ] ) run over the three basis spatial directions of the chosen representation ( in our case the cylindrical coordinate system ) and @xmath169 is the totally antisymmetric tensor associated with the vector product .
the responses in ( [ eq : x_y ] ) and ( [ eq : kappa - eta ] ) are expressed in terms of the averaged commutators @xmath21 > } \langle [ \hat{a},\hat{b}]\rangle \quad \mbox{with } \quad \gamma_t^a= - \gamma_t^b\ ] ] where @xmath40 is the quasiparticle vacuum .
calculation of these values can be considerably simplified if to express them through the matrix elements of the operators @xmath170 and @xmath171 .
to be specific , we associate the operator @xmath172 with a time - even operator @xmath173 or a time - odd operator @xmath174 , which have real or imaginary matrix elements , respectively .
then @xmath175 \rangle = 4i \sum_{ph \in s } \langle ph|\hat{q}_{sk } \rangle \im \langle ph|\hat{b}_s \rangle \ ; , \ ] ] @xmath176 \rangle = -4 \sum_{ph \in s } \langle ph|\hat{p}_{sk } \rangle \re \langle ph|\hat{b}_s \rangle \ ; , \ ] ] where both @xmath177 and @xmath178 run over all single - particle states and @xmath179 and @xmath180 mean the imaginary and real parts of the values to the right .
the pairing factors are included in the single - particle matrix elements . then elements of the inverse strength matrices are real and read @xmath181 \rangle = 4i \sum_{ph \in s ' } \langle ph|\hat{p}_{s'k ' } \rangle \ ; \re\langle ph|\hat{x}_{sk}^{s ' } \rangle \ ; , \ ] ] @xmath182 \rangle \ ; .\ ] ] the responses entering @xmath183 and @xmath184 in ( [ eq : x_y ] ) are also real and read @xmath185 \rangle \\
= -4i \sum_{ph \in s } \langle ph|\hat{p}_{sk } \rangle \ ; \re \langle ph|\hat{j}^{\alpha}_s \rangle \ ; , \ ] ] @xmath186 \rangle \\
= -4 \sum_{ph \in s } \langle ph|\hat{q}_{sk } \rangle \ ; \im \langle ph|\hat{j}^{\alpha}_s \rangle\ ] ] where @xmath187 are transition densities .
we actually deal with axially symmetric systems .
the modes then can be sorted into channels with a given angular - momentum projection @xmath188 to the symmetry axis @xmath132 . for even - even nuclei
@xmath188 takes integer values .
the explicit expressions for the responses in cylindrical coordinates ( see definition of these coordinates in the appendix [ sec : cylindr ] ) then read @xmath189\rangle \\ & = & ( \vec{e}_{\rho}j_{y , sk}^{\rho}(\rho , z ) + \vec{e}_{z}j_{y , sk}^{z}(\rho , z))\cos\mu\theta \nonumber\\ & + & \vec{e}_{\theta}j_{y , sk}^{\theta}(\rho , z)\sin\mu\theta \ ; , % \label{vec_j_y } \nonumber\\ \label{vec_s_y } \vec{s}_{y , sk}(\vec{r } ) & = & i \langle[\hat{q}_{sk},{\hat{\vec s}}_s]\rangle \\ & = & ( \vec{e}_{\rho}s_{y , sk}^{\rho}(\rho , z ) + \vec{e}_{z}s_{y , sk}^{z}(\rho , z))\sin\mu\theta \nonumber\\ & + & \vec{e}_{\theta}s_{y , sk}^{\theta}(\rho , z)\cos\mu\theta \ ; , % \label{vec_s_y } \nonumber\\ \label{rho_x } \rho_{x , sk}(\vec{r } ) & = & i \langle [ \hat{p}_{sk},{\hat \rho}_s]\rangle = \rho_{x , sk}(\rho , z)\cos\mu\theta \ ; , \\ \label{tau_x } \tau_{x , sk}(\vec{r } ) & = & i \langle [ \hat{p}_{sk},{\hat \tau}_s]\rangle = \tau_{x , sk}(\rho , z)\cos\mu\theta \ ; , \\ \label{vec_j_y } \vec{\im}_{x , sk}(\vec{r } ) & = & i \langle [ \hat{p}_{sk},{\hat { \vec\im}}_s]\rangle \\ & = & ( \vec{e}_{\rho}\im_{x , sk}^{\rho}(\rho , z ) + \vec{e}_{z}\im_{x , sk}^{z}(\rho , z))\cos\mu\theta \nonumber\\ & + & \vec{e}_{\theta}\im_{x , sk}^{\theta}(\rho , z)\sin\mu\theta \nonumber\end{aligned}\ ] ] where @xmath190-depending response components are real .
all the dependence on the spatial coordinates follows from the transition densities entering the responses .
the response components involving @xmath191 obviously vanish for modes with @xmath192 .
explicit expressions in cylindrical coordinates for strength matrices , responses , transition densities , matrix elements , coulomb contributions and other srpa values can be found in @xcite .
it is to be noted that the present calculations do not yet include the coulomb contribution to the residual interaction .
this introduces the uncertainty reaching up to @xmath1930.4 mev in an average peak position @xcite .
such effect is safely below the precision of the present investigation .
as was mentioned in sec . [
sec : general ] , srpa starts with the choice of appropriate generating operators @xmath151 , see the sequence of the model steps in ( [ eq : q - initial ] ) .
the srpa formalism itself does not provide these operators . at the same time , their proper choice is crucial to get good convergence of the separable expansion ( [ v_sep ] ) with a minimal number of separable terms .
the choice should be simple and universal in the sense that it can be applied equally well to different modes and excitation channels .
we propose a choice inspired by physical arguments .
the main idea is that the generating operators should explore different spatial regions of the nucleus , the surface as well as the interior .
the leading scaling generator should have the form of the applied external field in the long - wave approximation , which is most sensitive to the surface of the system .
since nuclear collective motion dominates in the surface region , already this generator should provide a good description .
next generators should be localized more in the interior to describe an interplay of surface and volume vibrations . for @xmath194 giant resonances in spherical nuclei
, we used a set of generators with the radial dependencies in the form of power and bessel functions .
@xcite . in the present study for deformed nuclei
, we will implement , for the sake of simplicity , only generators with the power radial dependence @xmath195 the separable operators @xmath196 and @xmath197 with @xmath198 are mainly localized at the nuclear surface while the operators with @xmath199 allow to touch , at least partly , the nuclear interior .
this simple set seems to be a good compromise for the first calculations of the giant resonances .
our analysis shows that already two first operators with @xmath200 suffice for a spectral resolution of 2 mev , as discussed below . in the next studies we plan to enlarge the list of the input generators so as to cover properly the nuclear interior and to take into account the coupling of the modes with different @xmath201 , induced by the deformation
the main aim of the present calculations is to demonstrate ability of srpa to describe multipole giant resonances ( gr ) in deformed nuclei and to explore dependence of the resonance strength distributions on the nuclear matter properties and some terms of the skyrme functional .
the isovector giant dipole resonance ( gdr ) and isoscalar giant quadrupole resonance ( gqr ) in the axially deformed nuclei @xmath0sm , @xmath1u , and @xmath2no will be considered .
we will employ a representative set of skyrme forces , skt6 @xcite , skm * @xcite , sly6 @xcite , and ski3 @xcite , with different features ( isoscalar and isovector effective masses , asymmetry energy , ... ) .
these forces are widely used for description of ground state properties and dynamics of atomic nuclei @xcite including deformed ones . as is seen from table 1 , the forces span a variety of nuclear matter properties while all correctly reproducing experimental values of the quadrupole moments in @xmath0sm and @xmath1u .
.[tab : skyrme ] nuclear matter and deformation properties for the skyrme forces under consideration .
the table represents the isoscalar effective mass @xmath202 , symmetry energy @xmath203 , sum rule enhancement factor @xmath204 , isovector effective mass @xmath205 , and quadrupole moments @xmath206 in @xmath0sm , @xmath1u , and @xmath2no .
experimental values of the quadrupole moment in @xmath0sm and @xmath1u are 6.6 b and 11.1 b , respectively @xcite . [ cols="^,^,^,^,^,^,^,^ " , ] the calculations use a cylindrical coordinate - space grid with the spacing of 0.7 fm .
pairing is treated at the bcs level and is frozen in the dynamical calculations . the collective response for the gdr ( @xmath207 ) and gqr ( @xmath208 ) is computed with two input operators ( [ eq : scale_op ] ) at k=1 and 2 .
both gr are calculated in terms of the energy - weighted ( l=1 ) strength function ( [ eq : sf_1 ] ) with the averaging parameter @xmath118= 2 mev ( as most suitable for the comparison with the experiment ) .
the factorization of the residual interaction and the strength function technique dramatically reduce the computational effort .
for example , by using a pc with cpu pentium 4 ( 3.0 ghz ) we need about 25 minutes for the complete calculations of the gdr in @xmath1u , including 1 minute for computation of the strength function itself .
the isovector gdr and isoscalar gqr are calculated with the effective charges @xmath209 , @xmath210 and @xmath211 , respectively . the isoscalar dipole spurious mode is located at 2 - 3 mev .
the deviation from the desirable zero energy is caused by several reasons .
first , we have neglected in the present study the contribution from the coulomb residual interaction .
the second ( more important ) reason is that the nucleus is treated in a finite coordinate box , which artificially binds the center of mass .
larger numerical boxes could help here but at quickly increasing computational cost . in any case
, it suffices for our present purposes that the center - of - mass mode lies at a low energy and thus is safely separated from the gdr .
we use a large configuration space including the single - particle spectrum from the bottom of the potential well up to @xmath212 mev .
this results in 7000 - 10000 dipole and 11000 - 17000 quadrupole two - quasiparticle configurations in the energy interval 0 - 100 mev .
the relevant energy - weighted sum rules are exhausted by @xmath213 .
such a basis is certainly enough for the present aims .
results of the calculations are presented in figs . 1 - 3 .
the first two figures compare the calculated gdr and gqr with the available photoabsorption @xcite and @xmath214 @xcite experimental data .
it is seen that all four skyrme forces provide in general an appropriate agreement with the experiment .
so , srpa indeed can give a robust treatment of the multipole gr and , what is important , does this with a minimal computational effort .
the isovector gdr in @xmath0sm , @xmath1u , and @xmath2no calculated with the skyrme forces skt6 , skm * , sly6 and ski3 .
the plots depict : the collective strength calculated with two ( solid curve ) and one ( dotted curve for @xmath2no ) input operators , the unperturbed quasiparticle strength ( dash curve ) and the photoabsorption experimental data @xcite ( triangles ) .
, width=302,height=396 ] the same as in fig . 1 but for isoscalar gqr and @xmath214 experimental data @xcite .
, width=302,height=396 ] to illustrate accuracy of the method , we give for @xmath2no the strength functions calculated with one ( k=1 ) and two ( k=1,2 ) input operators .
it is seen that both cases are about equal for the gqr with its simple one - bump structure . at the same time
, the second operator considerably changes gross structure of the more complicated gdr .
we have checked that inclusion of more operators ( k @xmath215 3 ) does not result in further significant modification of the gdr .
so , the approximation of two input operators seems to be reasonable , at least for the present study at a resolution of 2 mev .
the figures also show the unperturbed strengths , i.e. the mere two - quasiparticle ( 2qp ) distributions without the residual interaction .
comparison of these strengths with the fully coupled collective strengths ( solid curves ) displays the collective e1 and e2 shifts . as a trivial fact
, we observe the proper right shift for the isovector gdr and the left shift for the isoscalar gqr .
what is more interesting , the shifts for both resonances ( including isovector dgr ) depend on the _ isoscalar _ effective mass @xmath202 . to explain this , one should remind that the smaller @xmath202 , the more stretched the single - particle spectrum @xcite ( see also relevant examples in @xcite ) . and indeed the e1 and e2 unperturbed strengths in the figures exhibit a systematic shift to higher energy while moving from skt6 to ski3 ( for e1 the shift is essentially weaker than for e2 @xcite ) .
simultaneously , we have the corresponding evolution of the collective energy shifts .
namely , they decrease for gdr and increase for gqr ( for exception of gdr for skt6 ) .
all these trends lead to a remarkable net result : strong variations of the unperturbed strengths and collective shifts with @xmath202 considerably compensate each other so as the final gdr and gqr energies become much less sensitive to different skyrme forces and approach the experimental values .
the remaining energy differences of order @xmath216 mev show some other , sometimes known , trends for the gr .
first , we see a systematic ( for exception of skt6 ) downshift of the calculated gdr with increasing the symmetry energy @xmath203 .
this , at first glance , surprising result complies with experience of the systematic studies in spherical nuclei @xcite .
the case of skt6 looks as exception from this simple rule .
it has the same asymmetry energy as skm@xmath217 ( see table [ tab : skyrme ] ) but a lower gdr resonance peak .
this comes because the density dependence @xmath218 is abnormally low here ( for reasons whose discussion goes beyond the scope of the present paper ) . besides the trend with @xmath203 , fig .
[ fig : e1 ] also hints a connection of the gdr energy with the sum rule enhancement factor @xmath204 and the related value of the _ isovector _ effective mass @xmath219 .
namely , the smaller the effective mass ( larger the @xmath204 ) , the higher the gdr .
the trends and connections mentioned above should , however , be considered with a bit of care .
indeed , the variation of @xmath203 between the different skyrme forces in table [ tab : skyrme ] is rather small and probably not enough to demonstrate a strong and unambiguous trend . besides , a complicated dependence of gdr on _ different _ isovector factors can spoil and entangle the concrete trends . in any case , analysis of the dgr trends deserves more systematic study which is now in reach with the efficient srpa technique . instead , the evolution of isoscalar gqr is more clear and systematic .
we see a steady upshift of the gqr peak from skt6@xmath217 to ski3 , which complies with the known dependence on the isoscalar effective mass @xmath202 @xcite , namely , the lower the effective mass , the higher the gqr .
this trend has a simple explanation . as was discussed above , the low @xmath202 results in stretching the single - particle spectrum and thus in the upshift of the 2qp quadrupole strength . in the gqr case ,
the opposite dependence of the collective shift is not enough to compensate the strong 2qp upshift and hence we obtain the trend .
note that in our calculations skt6 with @xmath220 yields the best agreement for the gqr .
this confirms to some extend the findings @xcite for @xmath221pb that a good reproduction of the gqr requires a large effective mass .
shape and width of gr in deformed nuclei are mainly determined by the deformation splitting and the spectral fragmentation due to interference with energetically close @xmath6 states ( landau fragmentation ) .
following table [ tab : skyrme ] , the different skyrme forces provide quite similar quadrupole moments .
hence they result in close deformation splittings .
instead , the landau fragmentation depends sensitively on the spectral pattern of a model , determined in a large extent by isoscalar and isovector effective masses .
our samples of forces contain sufficient variation of both ones .
nonetheless , it turns out that the width and shape of the gqr are practically the same for all the four skyrme forces . at the same time
, the strongest variation is found for the gdr whose width and gross - structure significantly depend on the force .
it is seen that skm * gives an artificial third right peak ( especially in @xmath0sm ) and thus an overestimation of the gdr width .
this effect weakens for sly6 ( leading to the best description of gdr ) and vanishes at all for ski3 ( resulting in underestimation of the resonance width ) .
the appearance of the artificial right shoulder for skm * and sly6 was already noted for deformed rare - earth and actinide nuclei @xcite and @xmath221pb @xcite . for the particular skyrme forces
, this effect seems to be universal .
it takes place for gdr in heavy nuclei , independently on their shape .
as is shown in @xcite , the right shoulder is provoked by an excessive collective shift and further enforced by the presence in the region of the 2qp bunch composed from the particular high - moment configurations ( @xmath222 and @xmath223 for @xmath0sm and @xmath224 and @xmath225 for @xmath1u and @xmath2no ) @xcite .
these configurations represent the intruder @xmath226 states entering the valence shell due to the strong spin - orbital splitting .
they form a dense 2qp bunch which is easily excited and thus provides high sensitivity of the right gdr flank
. this feature can be useful for additional selection of the parameters of skyrme forces
. it would be very interesting to relate the above results for the gdr width and profile with the effective masses .
some possible correlations can immediately be noted .
for example , the evolution of the right shoulder from skm * to ski3 seems to correlate with a systematic decrease of @xmath202 . besides
, the largest right shoulder in skm * can be related with the smallest value of @xmath219 for this force .
this conclusion , however , requires still confirmation from more extended studies with an even broader basis of forces . at the same time , the results hint that the properties of the isovector gdr probably depend not only on the isovector @xmath219 but also on the isoscalar @xmath202 . for the first glance
, this statement looks surprising .
however , we should take into account that @xmath227 influences the mean level spacings @xmath228 in the mean field and hence can in principle affect the gdr .
the resonance should depend on the ground state properties which in turn are related to @xmath227 .
it worth also noting that both @xmath202 and @xmath219 are generated by one and the same term of the skyrme functional ( @xmath229 ) and so probably are not fully decoupled in the dynamics . some non - trivial relations between isoscalar and isovector parameters
are already discussed in literature , e.g. the relation @xmath230 and its connection with @xmath231 @xcite .
and we see here an interesting field for future investigations .
the isovector gdr and isoscalar gqr in @xmath1u calculated with the skyrme forces skt6 , skm * , sly6 and ski3 .
the strength is calculated with ( solid curve ) and without ( dash curve ) contribution of the time - odd current .
, width=264,height=396 ] as a next step , let s consider contributions of the time - odd densities to the gdr and gqr .
these results are illustrated in fig .
the calculations show that only the current - current contribution ( [ de_dj_dj ] ) is essential while the contribution ( [ de_dj_dsigma ] ) connected with the spin density is negligible .
so , in fig .
3 we display only effect of the current density .
first of all , it is worth noting that the most significant changes appear again at the right flank of the resonances . probably , the high - moment @xmath226-configurations mentioned above play important role not only for gdr but for gqr as well .
the impact of the time - odd current is different in gdr and gqr . in the quadrupole resonance , we see the systematic downshift and narrowing the strength . there is a clear correlation with the value of the isoscalar effective mass : the lower @xmath202 , the stronger the time - odd effect .
after inclusion of the time - odd coupling , the gqr for different @xmath202 become much closer .
so , the time - odd contribution weakens the influence of @xmath202 . in the dipole resonance ,
the time - odd shift is not so systematic .
we observe no shift for skt6 , the upshift for skm * and the downshifts for sly6 and ski3 .
again one may note some correlation with effective masses .
the specific skm * case can be connected with very small @xmath219 for this force . besides
, one may note ( for exception of skm * ) increasing the time - odd impact with lowering @xmath202 . in principle , the correlations between the influence of time - odd terms and effective masses in both gdr and gqr cases were more or less expected because the current density enters the term of the skyrme functional @xmath229 just responsible for generation of the effective masses .
this explains why the skt6 case with @xmath232 ( no effective mass effects ) does not demonstrate any time - odd impact .
it is worth noting that the dominant contributions to the collective response from the principle terms of the skyrme functional have different signs and thus , in a large extent , compensate each other ( this can be easily checked in the srpa by estimation of the different contributions to the inverse strength constants ( [ eq : kappa - eta ] ) . as a result , the smaller contributions ( time - odd , spin - orbital and coulomb ) become important .
the present study involves three nuclei from different mass regions and four skyrme forces with various bulk properties .
we have yet to disentangle more carefully the separate influences of them .
this requires systematic variations of forces in a large set of test cases .
such systematic studies can be now easily performed due to the computationally efficient srpa method .
a general procedure for the self - consistent factorization of the residual nuclear interaction is proposed for arbitrary density- and current - dependent functionals . following this procedure ,
the separable rpa ( srpa ) method is derived .
srpa dramatically simplifies the calculations while providing a reliable description of nuclear excitations .
the reduction of the computational effort is especially useful for deformed nuclei . in the present paper , srpa with skyrme forces
is specified for description of collective dynamics in axially - deformed nuclei .
for the first explorations , srpa is applied for description of isovector giant dipole resonances ( gdr ) and isoscalar giant quadrupole resonances ( gqr ) in deformed nuclei from rare - earth ( @xmath0sm ) , actinide ( @xmath1u ) , and superheavy ( @xmath2no ) regions .
four skyrme forces ( skt6 , skm * , sly6 , and ski3 ) with essentially different bulk properties are used .
the calculations show that srpa can successfully describe multipole giant resonances . a good agreement with available experimental data
is achieved , especially with sly6 for gdr .
we did not find any peculiarities of the gr in superheavy nuclei .
the behavior of the resonances in all three mass regions looks quite similar .
we analyzed dependence of gdr and gqr descriptions on various skyrme forces , in particular on the isoscalar and isovector effective masses and the symmetry energy .
the contribution of the time - odd couplings was also explored .
some known trends for gdr with @xmath231 and gqr with @xmath202 were reproduced . besides , the close relation between the time - odd contribution and @xmath202 was demonstrated for gqr .
the calculations also hint some interesting ( though not enough systematic ) trends for gdr .
altogether , the results point out correlations between isoscalar and isovector masses and time - odd contributions .
the correlation seem to be natural since all the items originate from one and the same term of the skyrme functional .
the time - odd and effective mass impacts manifest themselves mainly at the right flanks of the strength distributions .
the impacts are much stronger and diverse for gdr .
besides , this resonance exhibits the strong landau fragmentation .
the gdr gross structure considerably depends on the applied skyrme force and the related effective masses . as a result ,
the gdr structure can serve as an additional test for selection of the skyrme parameters and related nuclear matter values .
this work is a part of the research plan msm 0021620834 supported by the ministry of education of the czech republic .
it was partly funded by czech grant agency ( grant no .
202/06/0363 ) , grant agency of charles university in prague ( grant no .
222/2006/b - fyz / mff ) .
we thank also the support from dfg grant gz:436 rus 17/104/05 and heisenberg - landau ( germany - bltp jinr ) grants for 2005 and 2006 years .
are grateful for the bmbf support under contracts 06 er 808 and 06 dd 119 .
in our study the skyrme forces include time - even ( spatial , kinetic - energy , spin - orbit ) and time - odd ( current , spin ) densities associated with the hermitian operators @xmath233 where @xmath234 is the pauli matrix and @xmath235 is the number of protons or neutrons .
the densities read as @xmath236 where @xmath237 is the density operator , @xmath238 is the wave function of the single - particle state @xmath88 , and @xmath239 is the pairing occupation weight .
cylindrical coordinates @xmath240 are defined as @xmath241 then the single - particle particle wave function and its time reversal have the form of spinors @xmath242 @xmath243 where @xmath244 is the projection of the complete single - particle moment onto symmetry z - axis of the axial nucleus .
expressions for differential operators in cylindrical coordinates are elsewhere , see e.g. @xcite .
the pairing is treated with bardin - coopper - schriffer ( bcs ) method .
then the srpa values gain the pairing factors involving coefficients @xmath245 and @xmath246 of the bogoliubov transformation from particles to quasiparticles .
the densities and currents in the skyrme functional dominate in the particle - hole channel .
the pairing density falls into another channel provided by the two - particle excitations .
it involves a different pairing weight , namely @xmath247 rather then the @xmath239 for the standard densities ( [ eq : dens ] ) .
specifically the pairing density reads @xmath248 in the case of pairing , the hermitian one - body operators with a given time - parity ( @xmath249 ) obtain in the @xmath90-channel the form @xmath250 where @xmath251 are two - quasiparticle operators and @xmath252 are the pairing factors .
this is the case for time - even operators @xmath253 and @xmath254 and the time - odd operator @xmath255 . the time - odd operator @xmath256 = i \ { [ \hat h_{0 } , \hat q_{sk } ] + [ \hat v_{\rm res}^{\rm sep } , \hat q_{sk } ] \ }
\nonumber\\ & = & i [ \hat h_{0 } , \hat q_{sk } ] - \hat y_{sk}\end{aligned}\ ] ] have a more complicated form because of the additional term @xmath257 $ ] . namely it reads @xmath258 the srpa formalism in sections [ sec : srpa ] and [ sec : skyrme ] is presented in a general form equally suitable for the cases with and without pairing . in the case of pairing , the ground and perturbed @xmath6 many - body wave functions are replaced by their bcs counterparts , the summation indices @xmath177 and @xmath178 run the quasiparticle states , and the involved values ( densities , operators and matrix elements ) acquire the pairing factors given above .
the pairing is not important for giant resonances considered in the present study .
so , we freeze the pairing in the dynamics and do not present here the explicit form of the pairing contribution ( pairing vibrations ) to the residual interaction . some examples of this contribution can be found in @xcite .
99 m. bender , p .- h .
heenen , and p .-
reinhard , rev .
phys . * 75 * , 121 ( 2003 ) .
stoitsov , j. dobaczewski , w. nazarewicz , s. pittel , and d.j .
dean , phys .
c * 68 * , 054312 ( 2003 ) .
a. obertelli , s. peru , j.p .
delaroche , a. gillibert , m. girod , and h. goutte , phys .
c * 71 * , 024304 ( 2005 ) .
maruhn , p .-
reinhard , p.d .
stevenson , j. r. stone , and m.r .
strayer , phys .
c * 71 * , 064328 ( 2005 ) .
j. terasaki , j. engel , m. bender , j. dobaczewski , w. nazarewicz , and m. stoitsov , phys .
c * 71 * , 034310 ( 2005 ) .
k. langanke , nucl . phys . *
a687 * , 303 ( 2001 ) j. stone and p .- g .
reinhard , to appear in prog .
rowe , `` nuclear collective motion '' ( methuen and co. ltd .
, london , 1970 ) .
a. bohr and b.r .
mottelson , `` nuclear structure '' , vol.2 ( w.a .
benjamin , new york , 1974 ) .
e. lipparini and s. stringari , nucl . phys . *
a371 * , 430 ( 1981 ) . t.suzuki and h. sagava , prog .
phys . , * 65 * , 565 ( 1981 )
. t. kubo , h. sakamoto , t. kammuri and t. kishimoto , phys .
c * 54 * , 2331 ( 1996 ) .
nesterenko , w. kleinig , v.v .
gudkov , and j. kvasil , phys . rev .
c * 53 * , 1632 ( 1996 ) .
n. van giai , ch .
stoyanov and v.v .
voronov , phys .
c * 57 * , 1204 ( 1998 ) .
severyukhin , ch .
stoyanov , v.v .
voronov , and n. van giai , phys .
c * 66 * , 034304 ( 2002 ) .
nesterenko , w. kleinig , v.v .
gudkov , n. lo iudice , and j. kvasil , phys . rev . * a56 * , 607 ( 1997 ) .
w. kleinig , v.o .
nesterenko , and p .-
reinhard , ann .
( ny ) * 297 * , 1 ( 2002 ) .
nesterenko , j. kvasil , and p .-
reinhard , phys .
c * 66 * , 044307 ( 2002 ) .
j. kvasil , v.o .
nesterenko , and p .-
reinhard , in proceed . of 7th inter .
spring seminar on nuclear physics , miori , italy , 2001 , edited by a.covello , world scient .
publ . , p.437 , 2002 ; arxiv nucl - th/00109048 .
nesterenko , j. kvasil , p .-
reinhard , w. kleinig , and p. fleischer , proc .
of the 11-th international symposium on capture - gamma - ray spectroscopy and related topics , ed .
by j.kvasil , p.cejnar , m.krticka , prague 2 - 6 sept .
2002 , word scientific singapore , p. 143
nesterenko , j. kvasil and p .-
reinhard , prog .
, * 15 * , 127 ( 2006 ) .
nesterenko , j. kvasil , w. kleinig , p .-
reinhard , and d.s .
dolci , arxiv : nucl - th/0512045 .
skyrme , phil . mag . * 1 * , 1043 ( 1956 ) ; d. vauterin , d.m .
brink , phys .
c * 5 * , 626 ( 1972 ) .
engel , d.m .
brink , k. goeke , s.j .
krieger , and d. vauterin , nucl .
* a249 * , 215 ( 1975 ) . j. dobaczewski and j. dudek , phys .
c * 52 * , 1827 ( 1995 ) .
w. kohn and l. sham , phys .
rev . * 140 * a1133 ( 1965 ) .
o. gunnarson and b.i .
lundqvist , phys . rev . *
b13 * , 4274 ( 1976 ) .
w. kleinig , v.o .
nesterenko , p .-
reinhard , and ll .
serra , eur .
j. d * 4 * , 343 ( 1998 ) .
nesterenko , w. kleinig , f.f .
de souza cruz , and n. lo iudice , phys .
lett . * 83 * 57 ( 1999 ) .
nesterenko , w. kleinig , and p .-
reinhard , eur .
j. d*19 * , 57 ( 2002 ) .
v.o . nesterenko , p .-
reinhard , w. kleinig , and d.s .
dolci , phys .
a * 70 * , 023205 ( 2004 ) .
f. tondeur , m. brack , m. farine , and j.m .
pearson , nucl .
a420 * , 297 ( 1984 ) .
j. bartel , p. quentin , m. brack , c. guet , and h .- b .
hakansson , nucl . phys . *
a386 * , 79 ( 1982 ) .
e. chabanat , p. bonche , p. haensel , j. meyer , and r. schaeffer , nucl . phys . *
a643 * , 441(e ) ( 1998 ) .
reinhard and h. flocard , nucl .
a584 * , 467 ( 1995 ) .
we now work out the srpa extension to solve this problem . v.o .
nesterenko , w. kleinig , j. kvasil , p .-
reinhard , and p. vesely , under preparation for publication .
reinhard , ann .
( leipzig ) * 1 * , 632 ( 1992 ) .
reinhard , m. brack , o. genzken , phys .
a * 41 * , 5568 ( 1990 ) .
thouless , nucl . phys .
* 21 * , 225 ( 1960 ) .
malov , v.o . nesterenko , and v.g .
soloviev , teor .
fiz . * 32 * , 134 ( 1977 ) .
j. kvasil , n. lo iudice , v.o .
nesterenko , and m. kopal , phys .
c * 58 * , 209 ( 1998 ) . g. lauritsch and p .-
reinhard , int .
phys . * 5 * , 65 ( 1994 ) .
d. vauterin , phys .
c * 7 * , 296 ( 1973 ) .
stoitsov , j. dobaczewski , w. nazarewitcz and p.ring , comp .
, * 167 * , 4363 ( 2005 ) .
g.a . korn and t.m .
korn , `` mathematical handbook for scientists and engeneers '' , chapter 6.5 ( mcgraw - hill book company , new york , 1968 ) .
a.s goldhaber and g. scharff - goldhaber , phys .
c * 17 * , 1171 ( 1978 ) .
janis database : l0025.029 - 0 ( sac1971 ) .
dietrich and b.l .
bergman , at .
data nucl .
data tables , * 38 * , 199 ( 1998 ) ; iaea photonuclear data .
youngblood , y .- w .
lui , h.l .
clark , b. john , y. tokimoto , and x. chen , phys .
c * 69 * , 034315 ( 2004 ) .
nesterenko , v.p .
likhachev , p .-
reinhard , v.v .
pashkevich , w. kleinig , and j. mesa , phys .
c * 70 * , 057304 ( 2004 ) . the unperturbed e1 and e2 strengths are provided by transitions via one and two quantum shells , respectively .
hence the stretching effect of the effective mass @xmath202 is much stronger for e2 than for e1 .
t. sil , s. shlomo , b.k .
agrawal , and p .-
reinhard , phys .
c * 73 * , 034316 ( 2006 ) p .- g .
reinhard , nucl .
a649 * , 305c ( 1999 ) .
m. brack , c. guet , and h.b .
hkansson , phys . rep . * 123 * , 275 ( 1985 ) w. satula , r.a .
wyss , and m. rafalski , phys .
c * 74 * , 011301(r ) ( 2006 ) .
|
we formulate the self - consistent separable random - phase - approximation ( srpa ) method and specify it for skyrme forces with pairing for the case of axially symmetric deformed nuclei .
the factorization of the residual interaction allows to avoid diagonalization of high - rank rpa matrices , which dramatically reduces the computational expense .
this advantage is crucial for the systems with a huge configuration space , first of all for deformed nuclei .
srpa takes self - consistently into account the contributions of both time - even and time - odd skyrme terms as well as of the coulomb force and pairing .
the method is implemented to description of isovector e1 and isoscalar e2 giant resonances in a representative set of deformed nuclei : @xmath0sm , @xmath1u , and @xmath2no .
four different skyrme parameterizations ( skt6 , skm * , sly6 , and ski3 ) are employed to explore dependence of the strength distributions on some basic characteristics of the skyrme functional and nuclear matter . in particular , we discuss the role of isoscalar and isovector effective masses and their relation to time - odd contributions .
high sensitivity of the right flank of e1 resonance to different skyrme forces and the related artificial structure effects are analyzed .
|
machine - type communications ( mtc ) are typically characterized by a massive number of machine - type devices that connect to the network to transmit small data payloads . those features
present a significant challenge to cellular networks , whose radio access part is traditionally designed to deal with a rather low number of connections with high data requirements .
specifically , current cellular networks , such as lte - a , are connection - oriented @xcite , requiring a connection establishment between the device and the base station ( bs ) before the device can transmit its data packet . as an example
, the connection establishment in lte - a involves a high amount of signaling overhead , which is particularly emphasized when the data payload is small , e.g. , less than 1000 bytes @xcite .
therefore , in 3gpp it was proposed an approach to optimize the connection establishment by reducing the signaling overhead @xcite .
the resulting simplified connection establishment protocol starts with the contention - based access reservation protocol ( arp ) @xcite , depicted in the first four steps in fig .
[ fig : arpcomparison](a ) , followed by a fifth message where the signaling and a small data payload are concatenated .
the signaling exchanges related to the security mechanisms are omitted in the optimized version of the lte - a connection establishment , by reusing an a - priori established security context @xcite .
the throughput and blocking probability of the arp are rather sensitive to the number of contending devices .
specifically , the devices contend for access by sending their preambles in a designated and periodically occurring uplink sub - frame , here termed as random access opportunity ( rao ) .
when the number of contending devices is high @xcite , multiple devices activate the same preamble in a rao , which leads to collisions of their rrc connection requests , see fig .
[ fig : arpcomparison](a ) .
consequently , most devices are unable to establish a connection in the first attempt and perform subsequent attempts that , due to the high load , are also likely to result in collisions .
a solution put forward to cope with congestion , was the extended access class barring ( eab ) @xcite , where certain classes of devices are temporally blocked from participating in the arp , but at the cost of an increased access latency to those same devices .
another drawback of the arp is that the network learns the devices identities and connection establishment causes only after the rrc connection request is successfully received , as the contention is performed via randomly chosen preambles that do not carry information .
a solution that allows the network to learn the identities and connection establishment causes of the contending devices already at the beginning of the arp , could enable their differentiated treatment in later phases of the connection establishment and even skip some of the steps in the lte - a random access protocol , as indicated in fig .
[ fig : arpcomparison ] .
[ fig : arpcomparison ] in this paper we propose a new access method based on signatures and bloom filtering @xcite .
the method is demonstrated in the context of the lte - a arp , however , we note that it can be employed in the next generation arps @xcite following similar principles . in the proposed method , instead of contending with a single preamble in a rao , the devices contend by transmitting a predefined sequence of preambles in a frame composed of several raos , the transmitted sequence of preambles is denoted as the _ device signature_. the presented ideas are a conceptual extension of the work @xcite , where the devices contend for access by selecting a random signature , generated by combining random preambles over consecutive raos .
in contrast , in the method described here , each device contends with a unique signature generated using the international mobile subscriber identity ( imsi ) of the device and its connection establishment cause , in further text referred to as the device s identification .
specifically , we apply the bloom - filter @xcite principles for signature generation , where the device s identification is hashed over multiple independent hash functions and the resulting output used to select which preamble in which rao to activate . we introduce an analytical framework through which we tune the signature properties ,
i.e. , its length and the number of activated preambles , based on the number of expected arrivals and the target efficiency of the use of system resources , denoted as the goodput .
we also investigate the expected latency and signature detection probability of the proposed method .
finally , we show that , when the arrivals are synchronous , the proposed method outperforms the lte - a connection establishment procedure in terms of goodput , while achieving similar or lower average latency .
the rest of the paper is organized as follows .
section [ sec : lte_arp ] summarizes the standard arp in lte - a .
section [ sec : proposed_contention_modifications ] describes the proposed access method and section [ sub : analytical_performance_model ] presents the corresponding analysis .
section [ sec : system_performance_evaluation ] evaluates the performance of the proposed method , comparing it with the reference lte - a procedure for mtc traffic .
section [ sec : conclusions ] concludes the paper .
a successful lte - a access reservation entails the exchange of four messages , as depicted in fig . [
fig : arpcomparison](a ) .
initially , a device randomly chooses a preamble to be transmitted in a rao from a set of available preambles generated using zadoff - chu sequences @xcite .
the preambles are orthogonal and can be simultaneously detected by the bs .
we also note that the bs is able to detect a preamble even when it is transmitted by multiple devices @xcite , i.e. , a collision in the `` preamble space '' is still interpreted as an activated preamble .
this represents a logical or operation , since the preamble is detected as activated if there is _ at least _
one device that transmits the preamble .
this observation motivates the use of bloom filter , a data structure based on or operation for testing set membership .
the devices whose preambles are detected are notified via a random access response ( rar ) in the downlink and assigned a temporary network identifier .
the reception of the rar triggers the transmission of the rrc connection request in the allocated uplink sub - frame . at this point ,
the bs is able to detect the collision of the multiple connection requests , sent by the devices that originally sent the same preamble .
the successfully received connection requests are acknowledged , marking the start of the data transmission phase .
on the other hand , the devices whose connection requests collided , do not receive the feedback and either contend again by sending a new preamble or end up in outage when the number of connection attempts reaches the predefined limit . in the rrc connection request ,
the device informs the network of its temporary identifier , imsi , and the connection establishment cause . from these
, the network can confirm if the device is authorized for access , track the device s subscribed services and reestablish the preexisting security context @xcite .
as already mentioned , the channel over which the devices contend can be modeled as an or multiple access channel ( or - mac ) . by @xmath0 , denote the set of available preambles , where the absence of preamble activation is denoted by the idle preamble @xmath1 .
assume that there are @xmath2 devices in total .
we model the contention by assuming that the device @xmath3 , @xmath4 , transmits a binary word @xmath5,\end{aligned}\ ] ] where bit @xmath6 indicates if the device @xmath3 transmitted preamble @xmath7 .
note that only a single entry @xmath8 , @xmath9 , can be set to 1 since a device can only transmit a single preamble in a single rao .
the bs observes @xmath10 where @xmath11 denotes a bit - wise or operator and @xmath12 is the detected binary word of device @xmath3 .
in particular , the bs detects a transmitted preamble with probability @xmath13 and with probability @xmath14 falsely detects a non - transmitted preamble , which may cause that @xmath15 . in practice
, the preamble detection at the bs should ensure that @xmath16 and @xmath17 @xcite requirement in @xcite corresponds to the single activation of a preamble .
when a preamble is activated by multiple devices it is expected that the effective @xmath18 will be higher @xcite . ] .
finally , every non - zero entry in @xmath19 implies a detection of the corresponding preamble .
obviously , in the best - case scenario , the bs can detect up to @xmath20 different devices in a rao .
[ fig : lteormac ] the essence of the proposed method lies in the idea of devices contending with combinations of @xmath21 preambles transmitted over @xmath22 raos , denoted as signatures .
each preamble of a signature is sent in a separate rao , while @xmath22 raos define a signature frame , see fig .
[ fig : lteormac ] . extending the model introduced in section [ sec : lte_arp ] ,
the device @xmath3 contends by transmitting its signature @xmath23 , \end{aligned}\ ] ] where the binary words @xmath24 , @xmath25 , follow the structure introduced in .
obviously , the number of available signatures is @xmath26 , potentially allowing for the detection of exponentially more contenders compared to the case in which the preambles sent in each of the @xmath22 raos are treated independently and where the maximal number of detected contenders is @xmath27 .
similarly to , the bs observes @xmath28 where @xmath29 is the detected version of @xmath30 .
the bs decodes all signatures @xmath31 for which the following holds @xmath32 where @xmath33 is the bit - wise and . at this point
, we turn to a phenomenon intrinsically related to the proposed contention method @xcite .
namely , even in the case of perfect preamble detection ( @xmath34 ) and no false detections ( @xmath35 ) , the bs may also decode signatures that have _ not _ been transmitted but for which also holds . in other words , the bs may decode _
false positives_. an example of this is shown in fig .
[ fig : ormacsignaturetransmissiondetectionexample ] .
signatures when @xmath36 and @xmath37 and ( b ) erroneous decoding of a signature which was not present in the original transmission ( @xmath34 and @xmath35 ) . ]
[ fig : ormacsignaturetransmissiondetectionexample ] the performance of the random signature construction in terms of probability of decoding false positives was first analyzed in @xcite , where they are referred to as phantom sequences .
on the other hand , there is an extensive work on the construction of or - mac signatures @xcite based on the following criterion : if up to @xmath38-out - of-@xmath2 signatures are active , then there are no false positives .
however , these constructions are not directly applicable to the lte - a access , as they would ( 1 ) require that a device sends multiple preambles in the same rao , and ( 2 ) imply rather long signature lengths , i.e. , @xmath39 , which implies an increased access latency . inspired by bloom filters @xcite , we propose a novel signature construction that uses much lower signature lengths , at the expense of introducing false positives in a controlled manner . in the proposed method , the device signature is constructed in such a way that it provides a representation of the device s identification , which is assumed to be a - priori known to the network . to illustrate how a signature is constructed , we first consider the case where a single preamble is available at each of the @xmath22 raos dedicated to the signature transmission , i.e. , @xmath40 .
taking the view of the device @xmath3 , we start with the binary array @xmath30 of length @xmath22 , indexed from @xmath41 to @xmath22 , where all the bits are initially set to @xmath42 .
we then activate @xmath21 index positions in this array , i.e. , we set them to @xmath41 ; note that @xmath21 is a predefined constant valid for all devices .
this is done by using @xmath21 independent hash functions , @xmath43 , @xmath44 , whose output is an integer value between 1 and @xmath22 , corresponding to an index position in the array , and where @xmath45 is representation of the device identity .
the resulting binary array becomes the device signature .
this construction follows the same steps as the object insertion operation in a bloom filter @xcite .
when @xmath46 , the signature construction occurs in two stages .
the first stage corresponds to the selection of the @xmath21 active raos using hash functions @xmath43 , @xmath44 , as described previously . in the second stage , for each of the activated raos , a contending device selects and transmits randomly one of @xmath20 preambles .
this is performed by hashing the device identity using another set of independent hash functions @xmath47 , @xmath44 , i.e. , a separate hash function for each rao , whose output is an integer between @xmath41 and @xmath20 that corresponds to one of the available preambles .
the signature - based access reservation protocol is depicted in fig .
[ fig : arpcomparison](b ) , which starts by the devices transmitting their signatures . upon the successful decoding of a signature ,
the bs transmits the _ rrc connection setup _ message .
in contrast with the lte - a arp depicted in fig . [
fig : arpcomparison](a ) , the messages 2 and 3 are not required in the signature based access , since the bs is able to determine from the signature the imsi of the device and the connection establishment cause .
the protocol concludes with the transmission of the small data payload together with the completion of the rrc connection message .
the described signature generation raises two important issues : ( i ) out of @xmath21 hash functions @xmath43
, @xmath44 , there is a probability of @xmath48 that at least two of these functions generating the same output , leading to less than @xmath21 distinct raos active in a signature ; ( ii ) there is a non - zero probability that two or more devices share the same signature , given by @xmath49^{-1}\ ] ] and @xmath2 as the total number of devices .
the above probabilities can be minimized by increasing the signature length @xmath22 , which is the reason why these issues are commonly ignored within the bloom filter related literature , where @xmath22 is of the order of @xmath50 .
although we do not use such large ranges for @xmath22 , we note that for values of @xmath51 and @xmath52 that are used in the performance evaluation in section [ sec : system_performance_evaluation ] , the second probability can be neglected , as in this case @xmath53 .
[ alg : bloomfilterinsertion ] * input * : @xmath45 , @xmath22 , @xmath20 , @xmath21 ; + * initialize * : @xmath54 , @xmath55 , @xmath56 output @xmath30 ; + the first issue can be addressed by a signature construction that enforces @xmath21 distinct active raos per signature .
we provide in alg .
[ alg : bloomfilterinsertion ] a description of a practical signature construction that uses the modulus operation as basis for the hashing .
this construction ensures that @xmath21 distinct raos are active per signature , by removing the raos selected in previous iterations from the set of available raos .
further , the preambles activated in previously selected raos are removed from the set preambles available for the next iteration .
this operation limits the generation of signatures to @xmath57 active raos ; however , this is within the operating range of interest where @xmath58 and allows us to apply probabilistic tools , as presented in the analysis in section [ sub : analytical_performance_model ] , to design the signatures length @xmath22 and number of active raos @xmath21 . as it will be shown in section [ sec : system_performance_evaluation ] , the proposed signature generation algorithm matches well the derived analytical model .
finally , we note that an essential prerequisite for the proposed signature access scheme is that the signature generation algorithm and all the hash functions are known to all devices , including the bs .
this can be accomplished via the existing periodic broadcasts that include the network configuration ; an alternative would be to include this information already in the device s subscriber identity module .
we analyze a single instance of the contention process , assuming a synchronous batch arrival of @xmath59 devices .
we assume that the probability of an arrival of a device is @xmath60 /t$ ] , and denote the expected number of arrivals as @xmath61 $ ] .
the parameters of the proposed scheme are the signature frame size , denoted by @xmath22 , the number of active raos in the signature , denoted by @xmath21 , and the number of preambles per rao that are available for signature construction , denoted by @xmath20 .
the first two parameters are subject to design , and we analyze their dimensioning when on average @xmath38-out - of-@xmath2 signatures are active , such that the false positive rate is below a threshold .
in contrast , @xmath20 is assumed to be fixed , which corresponds to the typical scenario in lte - a systems .
we start by establishing the relationship between the correctly detected signatures and all detected signatures , which also includes the false positives , after all the contenders have completed @xmath62 step of the proposed method , see fig . [
fig : arpcomparison](b ) .
we denote this metric as the goodput @xmath63 .
in essence , the goodput reflects the efficiency of the subsequent small data transmission , as the bs will also attempt to serve the falsely detected signatures .
the expected goodput is @xmath64 = { \mathrm{e}}\left [ \frac{n_\text{a } } { n_\text{a } + p } \right ] \approx \frac{{\mathrm{e}}[n_\text{a } ] } { { \mathrm{e}}[n_\text{a } ] + { \mathrm{e}}[p ] } = \frac{n}{n + { \mathrm{e}}[p]}.\ ] ] where @xmath65 is the number of false positives . from
it follows @xmath66 \leq 1,\end{aligned}\ ] ] as there can be no more than @xmath2 detected signatures .
the mean number of false positives @xmath67 $ ] can be approximated as @xmath68 \approx p_\text{fa } ( t - n ) , \ ] ] where @xmath69 corresponds to the mean number of inactive signatures , while @xmath70 denotes the false positive probability , i.e. , the probability of an inactive signature being perceived as active .
. now becomes @xmath71 \approx \frac{n}{n + p_\text{fa } (
t - n ) } .\end{aligned}\ ] ] using , we proceed by setting the target goodput @xmath72 and establishing the relation between @xmath72 and the corresponding target @xmath73 @xmath74 to compute @xmath70 , we rely on approximations that hold when the number of simultaneously active signatures @xmath38 is high enough . specifically , @xmath70 is the probability that all @xmath21 preambles associated with an inactive signature , are detected as activated by the bs
. each of these @xmath21 preambles can be ( i ) actually activated by an active signature and detected as such by the bs , or ( ii ) not activated by any of the active signatures , but falsely detected as activated by the bs .
now , the probability that a particular preamble in a particular rao is not activated by any of the signatures , denoted by @xmath75 , is @xmath76 where @xmath27 is the total number of preambles in @xmath22 raos , @xmath21 is the number of preamble activations per user , @xmath38 is the number of active signatures , and it is assumed that the selection of any preamble in any rao is equally likely .
the detection of a preamble is non - ideal and therefore we have to distinguish between two events : ( i ) detection of a preamble transmitted by at least one device with probability @xmath18 ; ( ii ) false detection of a non - transmitted preamble with probability @xmath77 .
we approximate @xmath70 as @xmath78^k \\ \nonumber & = \left [ p_d + ( p_{f } - p_d ) \cdot p_\text{idle } \right]^k,\end{aligned}\ ] ] and where ( a ) becomes a lower bound when @xmath40 and @xmath34 and @xmath79 @xcite . from , the required signature frame size @xmath80 to meet the target
@xmath73 is @xmath81^{-1}\ ] ] [ alg : signaturedetection ] * input * : @xmath82 , @xmath19 , @xmath22 , @xmath20 , @xmath21 ; + * initialize * : @xmath83 , @xmath84 to compute the @xmath21 that minimizes @xmath80 in , we assume @xmath34 and @xmath79 .
then , for a given @xmath38 and @xmath22 , the value of @xmath21 that minimizes @xmath70 is given by @xcite @xmath85 we use to find the minimal required @xmath80 via .
furthermore , recall that each device can only activate up to a single preamble per rao , resulting in the constraint @xmath86 where we assume to work in the regime in which @xmath87 , i.e. , where @xmath88 .
now , the minimum @xmath80 can be obtained by solving iteratively the following fixed - point equation obtained from combining and @xmath89 { \frac{\ceil{k_{\min}}}{m } \left [ 1 - \left ( \frac{p_{fa}^{1 / \ceil{k_{\min } } } - p_d}{p_f - p_d } \right)^{1/n}\right]^{-1}},\ ] ] which converges for @xmath90 and @xmath91 , i.e. , the prescribed preamble detection performance @xcite .
a straightforward approach for signature decoding is to perform it after all raos of the signature frame have been received , i.e. , after the bs has observed the whole signature frame .
an alternative is to perform the decoding iteratively after every received signature rao , i.e. , the bs attempts to decode a signature while only having access to a partial observation of the signature frame .
the latter strategy is inspired with the fact that @xmath21 active raos constituting a signature are randomly spread over the signature frame and , in principle , the bs does not have to wait until the end of the frame to detect a signature .
the decoding performance is the same for both strategies when all @xmath22 raos in the signature frame have been received , but the average latency in the latter approach is lower .
we provide in alg .
[ alg : signaturedetection ] an algorithmic description of the iterative signature decoding , where the notation @xmath92 corresponds to the first @xmath93 entries of vector @xmath94 .
the key steps of the alg .
[ alg : signaturedetection ] are steps 5 and 7 . in particular , in step 5
the bs discards the signatures that could not have generated the partial observation @xmath95 from the set of potentially active signatures @xmath96 .
obviously , it is expected that @xmath96 will decrease with the additional received raos . in step 7
, the bs detects the signatures whose combinations of active raos and preambles are uniquely contributing to the partial observation @xmath95
. then the bs reports to the respective device that its signature has been decoded , which in the lte - a protocol realization would correspond to the rrc connection setup message , as shown in fig .
[ fig : arpcomparison](b ) . finally , in steps 1012 , when all raos have been received , the bs reports all the signatures within the set @xmath97 as decoded . , @xmath98 , @xmath99 , @xmath100 , @xmath101 , and @xmath102 from .
] in fig .
[ fig : advancedreceivertrace ] , we provide a simulation snapshot showing how many signatures are considered potentially active and how many have actually been decoded as the raos of the signature frame elapse .
obviously , the iterative signature decoding occurs in a spread manner , which leads to the spreading of the feedback messages acknowledging the decoding of each signature , i.e. , the rrc connection setup message in fig .
[ fig : arpcomparison](b ) . in this way
, the scenario in which a high number of devices attempt to complete the access reservation protocol simultaneously is avoided , i.e. , the occurrence of congestion at the later stages of the arp is reduced .
another important observation is that most of the signatures become decoded well before the end of the signature frame .
in order to evaluate the performance of the proposed signature based access and compare it with the proposed 3gpp lte - a solution for mtc traffic @xcite , we have implemented an event driven simulator where the main downlink and uplink lte channels are modeled . specifically , the simulator implements the both procedures depicted in fig .
[ fig : arpcomparison](a ) and fig .
[ fig : arpcomparison](b ) , while the downlink control and data channels ( pdcch and pdsch respectively ) and the uplink data and random access channels ( pusch and prach ) are modeled as in @xcite .
we consider a typical cell , configured with one rao every 1 ms , @xmath103 available preambles for contention @xcite .
we assume a total population of size @xmath104 , and a batch arrival of @xmath105 devices with a payload of @xmath106 bytes to transmit , the arrival probability of an individual device is given by @xmath107 , i.e. , @xmath105 is a binomially distributed random variable with mean @xmath108 = n$ ] .
the mean number of arrivals @xmath38 is assumed to be known , and the signature based scheme is dimensioned for it .
can be estimated , e.g. , using techniques that take advantage of the lte - a arp , such as the one proposed in @xcite . ]
the probability of preamble detection by the bs is set to @xmath100 and the probability of false detection of a preamble is set to @xmath101 @xcite . in the baseline ,
i.e. , 3gpp scheme , we assume the typical values for the backoff window of 20 ms and the maximum number of @xmath109 connection attempts @xcite . the devices upon becoming active contend for access by activating randomly one preamble in one of the available raos within the backoff interval , i.e. , the batch arrival is spread with the backoff interval . in case
that a device is the only one that selected a given preamble in a given rao and that this preamble has been detected , then the access procedure , as depicted in fig [ fig : arpcomparison](a ) , proceeds until completion . otherwise , the device will reattempt the access within the back - off window after the timer to receive the rar as elapsed .
when multiple devices select the same preamble within a rao , the resources assigned by the bs corresponding to the step 3 in the protocol are wasted due to the collided devices ; and the collided devices re - attempt access later by selecting a random rao within the backoff interval .
the devices re - attempt access until either successful or until exceeding the allowed number of retransmissions . in the proposed method ,
the devices contend by transmitting their signatures , where the signature frame length @xmath22 is obtained from .
for the sake of comparison , we also evaluate the performance of the random signature construction @xcite , where @xmath110 . each device upon its signature being decoded , even in the case of false positive , receives the feedback rrc connection setup message and is assigned uplink data resources for the transmission of the third and final message , see fig [ fig : arpcomparison](b ) .
the performance is evaluated in terms of : ( i ) the average goodput @xmath111 $ ] ; ( ii ) the average latency until the first step in both access schemes is successful , corresponding to a singleton preamble in the baseline and a successfully decoded signature in the proposed scheme ; ( iii ) the average latency until the small data transmission takes place , corresponding to step 5 in the baseline and to step 3 in the proposed scheme , see fig [ fig : arpcomparison ] ; and ( iv ) probability of device being successfully detected upon the completion of the access protocol .
the average goodput @xmath111 $ ] is evaluated as the ratio between the successfully used resources and the total resources spent in the third step of both access protocols .
it directly relates to the efficient use of resources , since the bs is only able to discern if there is a correctly detected device upon successful completion of the third step . in the baseline scheme ,
the system resources are wasted whenever two or more devices select the same preamble within a rao ; the goodput in this case is given as the ratio between the total number of messages that are exchanged successfully and the total number of exchanged messages at the third step , including the failed ones due to collisions .
in the case of the signature based access , the wasted resources in the third step occur whenever a false positive signature occurs , and the goodput is given by .
$ ] observed with increasing @xmath38 , for the 3gpp scheme , random signature construction @xcite and the proposed signature construction .
( @xmath112 ) ] [ fig : goodput ] the expected goodput is depicted in fig .
[ fig : goodput ] , where for the goodput target for the proposed method is set to @xmath113 .
we observe that the proposed method meets the actual goodput meets the design target at higher access loads . on the other hand , at lower @xmath38 , the performance deviates from the target value @xmath113 .
this is due to the assumption that the false positive signatures are independently and uniformly generated from the idle signatures , which is the basis of the approximation in .
we can also observe that the goodput performance of the proposed method is always superior to the 3gpp scheme . specifically , in the 3gpp scheme the devices re - attempt retransmission upon colliding and until they are either successful or the number of retransmissions is exceeded .
each subsequent failed retransmission results in additional wasted system resources , which results in the observed degradation of the baseline goodput with increasing number of active devices .
finally , the goodput achieved with the random signature construction @xcite is quite low , due to the high number of false positives . and minimum @xmath80 computed from , at different stages of the access procedures .
( @xmath112 ) ] [ fig : latency ] in fig .
[ fig : latency ] we depict the mean latency at step 1 in all schemes , as well as in steps 3 and 5 in the signature and 3gpp schemes , respectively .
an important observation is that the latency of the proposed method is always lower than the 3gpp scheme ; and the gap between these two schemes increases for higher @xmath38 .
this is a consequence of the more efficient detection of active users , as can be seen when comparing the latency of these two schemes at step 1 .
furthermore , the random signature construction has the worst performance , the reason being that a signature can not be decoded before all @xmath22 raos of the signature frame have been received @xcite . finally , in tab . [ table : probdetectiontable ]
we show the probability of a device being successfully detected at end of the access protocol . here
the proposed method has a slight performance degradation compared to the 3gpp scheme , but this degradation diminishes higher access loads .
the 3gpp scheme achieves higher detection performance due to only requiring one transmission out of all preamble retransmissions to be successful , making it more robust but at the cost of lower goodput and higher latency . on the other hand
, the random signature construction leads to a very low detection performance , as it requires the successful detection of all the active preambles @xcite .
following the insights provided by bloom filters , we have introduced the concept of signatures with probabilistic guarantees and applied it to a system model derived from the lte - a access reservation protocol .
the most important feature of the proposed method is in allowing the device to be identified already at the access stage .
moreover , the method is very efficient in terms of use of the system resources and has a favorable performance in terms of decoding latency . in the paper we assumed that the base station serves the successfully connected devices without preferences .
nevertheless , it is straightforward to modify the proposed solution to scenarios in which the bs serves devices based on the identifications inferred from the decoded signatures , i.e. , imsis and/or connection establishment causes . in such cases ,
the proposed access method enables differentiated treatment by the bs from the very beginning .
finally , we note that in the paper we assessed a simplified scenario of a synchronous bath arrival in order to present the key concepts and the related analysis . tuning the proposed scheme for the other typical models , like the beta arrival model for synchronous arrivals or the poisson arrival model for asynchronous arrivals
, is left for further work .
.probability of successfully detecting a device [ % ] .
( t = 1000 ) [ cols="^,^,^,^,^,^",options="header " , ] [ table : probdetectiontable ]
this work was performed partly in the framework of h2020 project fantastic-5 g ( ict-671660 ) , partly supported by the danish council for independent research grant no .
dff-4005 - 00281 `` evolving wireless cellular systems for smart grid communications '' and by the european research council ( erc consolidator grant nr .
648382 willow ) within the horizon 2020 program .
the authors acknowledge the contributions of the colleagues in fantastic-5 g .
d. t. wiriaatmadja and k. w. choi , `` hybrid random access and data transmission protocol for machine - to - machine communications in cellular networks , '' _ ieee trans . wireless commun .
_ , vol . 14 , no . 1 , p. 3346
, jan . 2015 .
g. c. madueno , j. j. nielsen , d. m. kim , n. k. pratas , c. stefanovic , and p. popovski , `` assessment of lte wireless access for monitoring of energy distribution in the smart grid , '' _ ieee journal on selected areas in communications _ , vol .
34 , no . 3 , pp .
675688 , march 2016 .
h. thomsen , n. k. pratas , c. stefanovic , and p. popovski , `` code - expanded radio access protocol for machine - to - machine communications , '' _ transactions on emerging telecommunications technologies _
24 , no . 4 , pp .
355365 , 2013 .
|
we present a random access method inspired on bloom filters that is suited for machine - type communications ( mtc ) .
each accessing device sends a _ signature _ during the contention process .
a signature is constructed using the bloom filtering method and contains information on the device identity and the connection establishment cause .
we instantiate the proposed method over the current lte - a access protocol .
however , the method is applicable to a more general class of random access protocols that use preambles or other reservation sequences , as expected to be the case in 5 g systems .
we show that our method utilizes the system resources more efficiently and achieves significantly lower connection establishment latency in case of synchronous arrivals , compared to the variant of the lte - a access protocol that is optimized for mtc traffic .
a dividend of the proposed method is that it allows the base station ( bs ) to acquire the device identity and the connection establishment cause already in the initial phase of the connection establishment , thereby enabling their differentiated treatment by the bs .
|
( hereafter ) is a 33ms radio pulsar discovered in the green bank telescope ( gbt ) 350 mhz drift - scan pulsar survey @xcite . with a dispersion measure of 3.27pc@xmath5
, it appeared to be one of the closest pulsars to the earth .
further observations showed was in a binary system with an orbital period of 2.45days and a minimum companion mass of about 1@xmath6 .
this sort of system straddles the line between potential companion types .
it could be a double - neutron star ( dns ) , of which there are only roughly 12 and whose study is crucial to understanding the formation of sources of khz gravitational waves ( e.g. , * ? ? ?
* ) and testing general relativity ( e.g. , * ? ? ?
, it could be a pulsar with a massive white dwarf companion a so - called `` intermediate - mass binary pulsar '' ( imbp)that descended from a binary with a more massive companion than in traditional systems with pulsars and low - mass white dwarfs @xcite .
imbp systems are rare , with fewer than 20 known , and massive white dwarfs are themselves rare , with fewer than 8% of the white dwarfs ( wds ) from optical surveys having masses above @xmath7 @xcite .
understanding the formation and evolution of imbp systems provides a crucial piece in our understanding of binary evolution and pulsar recycling , and helps delineate evolutionary paths between low - mass nss and high - mass white dwarfs @xcite .
@xcite used very long baseline interferometry astrometry to measure the parallax of with exquisite precision .
they find a distance of @xmath0pc ( it is the second closest binary pulsar system and one of the closest nss of any type ) .
the astrometric data also suggested an edge - on orbit , opening up the possibility of a measurement of the shapiro delay @xcite , which gives two post - keplerian @xcite parameters for the system and hence determines the component masses ( e.g. , * ? ? ? * ) .
here we present the detailed timing analysis of the system , including the measurement of the shapiro delay and the determination of the masses ( ) .
we then present deep optical and near - infrared searches for the companion to ( ) , which we use to constrain models of its formation and evolution ( ) .
we find that the system almost certainly must be an imbp system , but that we do not detect the companion , constraining it to be one of the coolest white dwarfs ever observed .
unlike some sources where temperature inferences are highly dependent on white dwarf model atmospheres ( e.g. , * ? ? ?
* ) , this measurement is robust , given the small uncertainties on the mass and ( especially ) distance .
we conclude in .
radio observations of to measure the shapiro delay occurred in the last week of 2011 may with the 100 m robert c. byrd gbt .
we had a 6hr observation taken around superior conjunction of the binary system augmented by five 2hr observations at each of the other five shapiro extrema , all using the green bank ultimate pulsar processing instrument ( guppi ; @xcite ) .
the 800mhz of bandwidth centered at 1500mhz in two orthogonal polarizations was separated into 512 nyquist - sampled frequency channels of width 1.5625mhz via a polyphase filter bank .
these channels , sampled at 8-bits , provided full polarization information and an effective time resolution of 0.64@xmath8s .
each channel was coherently dedispersed at the nominal dispersion measure ( dm ) of the pulsar ( 3.27761pc@xmath9 at the time , although we later refined this measurement ) .
each observing session was broken into 30-minute observations of separated by 60s calibration scans of the extragalactic radio source 3c 190 .
the calibration scans were taken in the same mode as the pulsar observations , but also included a 25hz noise diode inserted into the receiver . ) .
the position angle of the linear polarization is given in the upper panel . as is the case with most msps
, the polarization position angle variations do not permit a rotating vector model fit , so we can not constrain the emission geometry .
[ fig : profile ] , scaledwidth=40.0% ] data reduction was performed using the ` psrchive ` package @xcite .
flux calibration used the on- and off - source scans of 3c 190 .
this was followed by removal of radio frequency interference by the psrzap utility .
the calibrated pulse profile determined from the long observation covering conjunction is given in .
the data were aligned in time using the best ephemeris ( below ) , divided into 16 frequency channels , and re - fit for dispersion measure and rotation measure using a bootstrap error analysis .
we found that the period - averaged flux density varied by a factor of a few over the course of long observations due to scintillation , with an average of 12mjy at 1500mhz .
individual times - of - arrival ( toas ) were measured from the folded total - intensity profiles using the frequency domain algorithm in ` psrchive ` @xcite .
a template was created by fitting three gaussians to the summed pulse profile . from these gaussian components
, we created a noise - free template with the phase of the fundamental component in the frequency domain rotated to zero .
the observations were divided into 2minute segments , with one toa measured for each segment . note that since interstellar scintillation caused the flux to vary considerably , there was a proportional change in the toa precision that varied over the data set .
these data were combined with previous data taken for the discovery observations of @xcite to produce a timing model .
we used the `` dd '' model @xcite in ` tempo ` , which incorporates the shapiro delay .
the astrometric data for this model were taken from @xcite , and we used the de421 jpl ephemeris @xcite .
timing fits with no shapiro delay were statistically unacceptable , with an rms residual of @xmath10s ( @xmath11 for 931 degrees - of - freedom ) , and a clear shapiro delay signature was obvious in the residuals ( ) . with the shapiro delay included in the fit the rms residual was 4.2@xmath8s ( @xmath12 for 929 degrees - of - freedom ) , with no obvious remaining structure in the residuals ( varying the astrometric parameters within the uncertainties from @xcite changed the timing results by @xmath13 ) .
the shapiro delay determines the inclination of the orbit and the companion mass ; this is then combined with the binary mass function to determine the pulsar s mass .
due to the combination of several different and much less precise observing modes from earlier monitoring with the high - precision shapiro delay campaign , we estimated the timing parameters with a bootstrap error analysis .
we give the full timing results , with 1-@xmath14 error estimates from the bootstrap analysis , in .
timing residuals for , using the new data from this paper ( blue : mjd 55,60055,921 ) and older data ( gray ) , as a function of orbital phase ( true anomaly plus longitude of periastron ) .
top : residuals computed from the best - fit model without shapiro delay ( the rms residual is @xmath10s ) .
middle : residuals computed including shapiro delay .
the red curve is the best - fit shapiro delay profile .
bottom : residuals computed relative to the best - fit model including shapiro delay ( the rms residual is @xmath15s ) .
conjunction is at a phase of 0.25 . in all panels
the left axis shows the residuals in @xmath8s , while the right axis shows the residuals in milliperiods .
note the different @xmath16-axis scales.,scaledwidth=50.0% ] our data consist of high - quality coherently dedispersed data from an intensive 1 week campaign and a few other epochs .
the remainder of the data were both less precise and less uniform , with a wider range of observation frequency and instrumental setup .
this makes it difficult ( if not impossible ) to robustly constrain long - term secular changes like periastron precession ( @xmath17 ; @xcite ) . nonetheless , we tried a fit with @xmath17 fixed to the value predicted by general relativity ( @xmath18 ) .
the resulting fit was good , with the rms decreasing to 3.8@xmath8s .
the pulsar and companion masses each increased by about 1@xmath14 compared to the values in .
given the small eccentricity and inhomogeneous data set with large gaps we do not believe that fitting for @xmath17 is viable at this time , but encourage further long - term monitoring of this system to establish its secular behavior .
l c + spin period ( s ) & 0.032817859053065(3 ) + period derivative ( ss@xmath19 ) & @xmath20 + dispersion measure ( pc@xmath9 ) & 3.2842(6 ) + rotation measure ( radm@xmath21 ) & + 2.6(1 ) + reference epoch ( mjd ) & 55743 + right ascension ( j2000 ) & 22:22:05.969101(1 ) + declination ( j2000 ) & @xmath22:37:15.72441(4 ) + r.a . proper motion ( mas@xmath23 ) & 44.73(4 ) + dec proper motion ( mas@xmath23 ) & @xmath245.68(6 ) + parallax ( mas ) & @xmath25 + position epoch ( mjd ) & 55743 + span of timing data ( mjd ) & 5500555922 + number of toas & 943 + rms residual ( @xmath8s ) & 4.2 + + orbital period ( days ) & 2.4457599929(3 ) + projected semi - major axis ( lt - s ) & 10.8480276(12 ) + epoch of periastron ( mjd ) & 55742.13242(0 ) + orbital eccentricity & @xmath26 + longitude of periastron ( deg ) & 119.778(12 ) + mass function ( @xmath6 ) & 0.22907971(8 ) + @xmath27 & 0.9985(3 ) + companion mass ( @xmath6 ) & 1.05(6 ) + + distance ( pc ) & 267.3@xmath28 + transverse velocity ( km@xmath29 ) & 57.1@xmath30 + orbital inclination @xmath31 ( deg ) & 86.8(4 ) + shklovskii period derivative ( @xmath32 ) & @xmath33 + intrinsic period derivative ( @xmath32 ) & @xmath34 + surface magnetic field ( @xmath35 gauss ) & 0.719 + spin - down luminosity ( @xmath36 ) & 1.72 + characteristic age ( gyr ) & 33.8 + pulsar mass ( @xmath6 ) & 1.20(14 ) + flux density at 1500mhz ( mjy ) & 12 + we observed the position of at optical and near - infrared wavelengths , as listed in .
the deepest keck observations used the red side of the low - resolution imaging spectrometer ( lris ; @xcite ) on the 10 m keck i telescope .
the data were reduced using standard procedures in ` iraf ` , subtracting the bias , dividing by flatfields , and combining the individual exposures .
the seeing was about @xmath37 in the combined @xmath38 image , and @xmath39 in the combined @xmath40 image .
we computed an astrometric solution fitting for a shift and separate scales and rotations along each axis ( i.e. , a six - parameter fit ) using 100 non - saturated stars identified from the sloan digital sky survey ( sdss ) data release 10 ( dr10 ; @xcite ) , giving rms residuals of @xmath41 in each coordinate .
we did photometric calibration relative to sdss photometry , identifying 23 well - detected , well - separated , non - saturated stars , and transforming from the sdss filter set to johnson cousins using the appropriate transformation equations .
the zero - point uncertainty was @xmath42mag , although there are systematic uncertainties coming from our filter transformations .
we see no object at the position of the pulsar ( ) ; the closest object is about @xmath43 from the position of the pulsar ( about @xmath44 away ) and appears extended ( @xmath45 and statistical position uncertainties of @xmath46 in each coordinate ) .
we determined the 3@xmath14 upper limits using ` sextractor ` @xcite to determine the magnitude that gave a 0.3mag uncertainty ( verified with fake - star tests ) , which we give in .
we observed in @xmath47-band with the goodman spectrograph @xcite on the 4.1 m southern astrophysical research ( soar ) telescope over two nights in 2013 july .
all exposures were dithered and binned by a factor of two in both dimensions .
the frames were bias - subtracted and flattened with a dome flat .
we then used a median of the data ( having masked the scattered - light halos of three saturated stars ) from the second night constructed without registration to create a sky flat , which we smoothed with a @xmath48pixel boxcar filter .
this corrects for larger - scale brightness variations .
cosmic rays were interpolated on individual exposures using the ` lacosmic ` routine @xcite .
the seeing varied considerably over the course of the observations , going from @xmath49 to @xmath43 .
we then shifted each exposure by an integer number of pixels for registration and summed them .
the final summed image has an effective seeing of @xmath50 and a total exposure time of 2.6hr .
the photometric zero - point was again computed relative to the sdss dr10 data , using 31 stars .
the astrometric solution was done using six 30s exposures through http://astrometry.net @xcite . as with the keck data , we see no object at the position of the pulsar ( ) and
give a 3@xmath14 upper limit in .
l c c c c c soar / goodman & 2013 jul 2 & @xmath47 & @xmath51 & 26.4 & 19.2 + soar / goodman & 2013 jul 3 & @xmath47 & @xmath52 & & + keck
i / lris(red ) & 2013 aug 4 & @xmath38 & @xmath53 & 26.3 & 19.1 + keck
i / lris(red ) & 2013 aug 4 & @xmath40 & @xmath53 & 26.0 & 18.9 + keck ii / nirc2 & 2013 oct 12 & @xmath54 & @xmath55 & 21.0 & 13.9 + while they were taken through different filters and with very different instruments / resolutions , we tried combining the keck @xmath38-band and soar @xmath47-band images using ` swarp ` @xcite .
we still see no source at the position of the pulsar .
the data are sufficiently different that a limiting flux is difficult to compute , but it could be as much as 0.3mag fainter than the limits in .
the near - infrared observations come from the nirc2 camera square field of view . ] on the 10 m keck ii telescope , and used the laser guide star adaptive optics ( ao ) system @xcite .
the data were taken through thin clouds and the ao corrections were not optimal , resulting in a delivered image quality of @xmath41 fwhm .
the images were reduced using a custom pipeline implemented with ` python ` and ` pyraf ` using dark frames and dome - flats .
a sky fringe frame was created by combining dithered images of multiple targets with the bright stars masked .
we used ` sextractor ` @xcite for the preliminary detection and masking of stars .
the fringe frame was subtracted from the flat - fielded data after being scaled to the appropriate sky background level . before coadding the frames ,
each frame was corrected for optical distortion using a distortion solution measured for nirc2 .
a faint glare has been visible in the lower right ( south - west ) corner of the nirc2 wide camera images starting in 2009 august .
the shape and amplitude of the glare vary with telescope orientation , resisting correction through surface fitting or modeling .
instead we masked the glare using a triangular region .
there was no independent photometric calibration that night , and only a single star is visible on the co - added image .
to determine a photometric zero - point , we used photometry for that star from the sdss dr10 .
we then employed the empirical main - sequence color relations from @xcite , inferring the @xmath56 color from the observed @xmath57 color ( we ignore differences between @xmath58 and @xmath54 filters ) . for this star ( sdss [email protected] )
we infer a spectral type of k2.5 and predict @xmath59 . we expect zero - point uncertainties of @xmath60mag or so based on comparison of the other sdss colors to those predicted using @xcite .
again we see no object at the position of the pulsar , and give 3@xmath14 upper limits in .
since we do not detect the optical counterpart of the companion , the first inference is that the companion could be a low - mass ns .
it would be the lowest mass ns known @xcite , although it is only a roughly 23 @xmath14 excursion from the mean of the companions in dns systems @xcite : rare , given the @xmath61dns systems , but not impossible . in that case , its eccentricity of @xmath62 would be a factor of @xmath63 lower than any other dns system ( has the lowest eccentricity of @xmath64 , although this may be an ns wd system ; @xcite ; van leeuwen et al .
2014 , , submitted ) . in
we show the eccentricity versus component masses for all dns and ns wd systems with well - determined masses .
in fact there are three ns wd systems with higher eccentricities : , which was likely not recycled @xcite ; , which has had its eccentricity increased by dynamical interactions @xcite ; and ( likely an imbp , with the eccentricity the result of unstable mass transfer ; @xcite ) .
the normal formation scenario for a dns involves two core - collapse supernova explosions , with the eccentricity the result of the second explosion and its kick , and no final mass - transfer phase to circularize the orbit ( e.g. , * ? ? ?
in contrast , formation via an electron - capture supernova ( ecs ; @xcite ) could result in a significantly lower ns mass ( @xcite ; @xcite ) along with a lower supernova kick @xcite .
has a low transverse velocity ( @xmath65 ) , although higher than some systems thought to be the products of ecss ( given the age of the system , this velocity may be more related to motion in the galactic potential than birth conditions ) .
this may reflect the velocity dispersion of the progenitor systems .
however , the contrast between and other systems thought to be the results of ecss ( e.g. , psr j1906 + 0746 or ; @xcite ) is extreme , with the ratio of eccentricities above 200 as mentioned previously . in a scenario without a kick we can place an upper limit on the amount of material that could have been ejected by the explosion to @xmath66 ( with @xmath67 the pulsar mass and @xmath68 the current companion mass ; e.g. , * ? ? ?
this is a much tighter bound than in any of the other systems proposed for this mechanism , and difficult to reconcile with the change in binding energy needed to collapse to a ns @xmath69 ( with @xmath70 km for an ns ) , presumably released as neutrinos ( e.g. , * ? ? ?
* ) : this leads to the horizontal line in , above which all confirmed dns systems are found . in order to have a dns system with such a low eccentricity
, we need to invoke increasingly exotic ( and perhaps implausible ) evolutionary scenarios .
for instance , if the system began as a hierarchical triple @xcite , then the inner components could have formed a standard eccentric dns system early on
. later evolution of the outer member could have led to a circum - binary accretion disk that would have worked to circularize the inner system , after which the outer object would have exploded or otherwise been ejected from the system .
the other possible scenario is that the companion could be a massive wd , making the system an imbp .
its orbital eccentricity is somewhat high compared to most low - mass binary pulsars of similar periods ( based on @xcite ) , but not nearly as high as a dns , consistent with an imbp classification @xcite .
it falls in the locus of other co wds in the `` corbet '' ( binary period versus spin period ) diagram in @xcite . the pulsar mass is lower than most pulsar
wd binaries , but is consistent with the short orbital - period imbp discussed by @xcite which may indicate a similar formation mechanism involving a common envelope @xcite .
however , as a wd it would be extremely faint : far fainter than any of the optical companions to imbps currently known @xcite or indeed any wd companion to a millisecond pulsar ( msp ) with a similar mass @xcite ; it is perhaps the faintest wd ever observed . with the apparent magnitude limits from
, we can compute absolute magnitude limits in each band .
we use the distance @xmath71pc @xcite , and we estimate the extinction to be @xmath72mag from @xcite . in terms of bolometric luminosity the most constraining limit ends up coming from the @xmath38-band data , where we limit @xmath73 ( the @xmath47-band limit of @xmath74 is very similar , given slight differences in bolometric correction ) . for comparison , the companion to with a median companion mass of @xmath75 has @xmath76 @xcite . in
we plot the absolute magnitude against mass for pulsar+wd systems as well as select cool wds with parallax distances : even compared to the observed truncation of the cooling sequence in old halo globular clusters like ngc 6397 @xcite or m4 @xcite , the putative companion is far fainter : at the distance of ngc 6397 , our limit of @xmath73 translates to an apparent magnitude of @xmath77 , compared to @xmath78 , or @xmath79 for the coolest wds seen in ngc 6397 .
some of the difference comes from the change in radius : a @xmath80 wd has a radius about 65% of that of a typical @xmath81 wd , leading to a 1mag change in brightness at the same effective temperature .
but the difference in is more like 2.5mag , so the companion to must also be cooler than the known thick disk / halo wds . beyond the absolute magnitude , which is directly computable from observable quantities , we can limit the radius / temperature of a putative wd by using our @xmath38-band absolute magnitude limit to constrain the bolometric luminosity .
this is more complicated , as it involves atmosphere calculations in an uncertain and poorly tested regime , but it should be reasonably reliable .
we use the synthetic photometry and evolutionary models from @xcite and @xcite for h and he atmospheres , respectively . ] . for isolated wds pure he atmospheres
can be largely excluded because of bondi - hoyle accretion from the ism @xcite , and even small amounts of hydrogen mixed into the helium can cause near - infrared flux deficiencies like pure hydrogen ( see below ; @xcite ) . however , the binary orbit and msp wind in this system could have inhibited such accretion and therefore a he atmosphere is possible . in any case
a pure he atmosphere will serve as a limiting case compared to the h models .
these models are used to convert the absolute magnitude limits into temperature limits , so for simplicity we use the @xmath80 models ( differences in bolometric corrections as a function of mass are small , @xmath82mag ) .
the most constraining limit is again from the @xmath38-band data , where @xmath73 implies @xmath83k ( see ) for a h atmosphere .
the he - atmosphere models do not extend to sufficiently cool temperatures but stop at @xmath84k with @xmath85 . at lower temperatures the details of the atmospheric physics are rather uncertain , but a blackbody is likely an acceptable approximation ( p. bergeron , 2014 , private communication ) . with a he atmosphere
an effective temperature @xmath86k would be required ( ) . the h limits are more constraining since more of the flux appears in the optical regime rather than the near - infrared a consequence of collisionally induced absorption by molecular h@xmath87 @xcite .
these limits change slightly with mass given the small but finite mass uncertainties , since the radius would change with mass : going to the @xmath88 h model we can constrain @xmath89k ( at our nominal mass of @xmath90 the radius of a c / o wd is about @xmath91 , and it scales as @xmath92 ) . as inferred from
, the companion to would be far cooler than any known wd from other surveys ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , where the coolest objects tend to have @xmath93k .
however , we can not exclude such a very cool wd on age grounds .
wd cooling curves , which start out having more massive objects warmer at the same age , eventually cross to have more massive objects cooler at the same age (; this is also visible in ) .
this is because massive wds crystallize earlier , at a higher @xmath94 ( but at a similar internal temperature ) , at which point the faster debye cooling takes over @xcite .
cooling ages for these models may not be reliable , as the impacts of state changes , sedimentation , and chemical processes are not precisely known , and the atmospheres are not trivial to calculate @xcite .
but we believe conservatively that the cooling age is close to 10gyr , almost certainly @xmath95gyr . in we
show example cooling curves , computed for thin and thick da atmospheres and c / o wds ( likely irradiation is a negligible perturbation to the wds surface temperature , given the measured spin - down luminosity of the pulsar ) . for the model closest to the best - fit mass of we would infer that the true age is near 9gyr , with the possible range from 612gyr .
the upper limit provided by the pulsar s characteristic spin - down age ( 34gyr after correction for the shklovskii effect [ @xcite ] ) is not constraining ; the assumption that the pulsar s initial spin period is much shorter than the current spin period is clearly not valid .
instead , we take as our upper limit to the age that of the milky way s halo ( @xmath96gyr ; @xcite ) minus the @xmath97myr required for the main - sequence lifetime of a @xmath98 progenitor @xcite , although this does not really exclude any models .
such an age would , however , imply a lower limit to the ( re-)birth period of about 25ms , assuming spin - down with a braking index @xmath99 ( magnetic dipole radiation ) .
we note that the cooling models in may not be the only solution for this progenitor : changing the wd composition ( likely it is below the transition to o / ne / mg wds based on @xcite , although binary evolution could change that ; also see @xcite ) or atmosphere ( helium , carbon , etc ) could lead to different solutions , and to draw robust conclusions we need to explore a wider range of models with better observational constraints .
there are also considerable complications and uncertainties in models for these temperatures : for instance , the models of ( * ? ? ?
* the basti database ) give rather different ages as @xmath94 never drops below @xmath100k for @xmath80 models , even for ages of @xmath101gyr , while @xcite and @xcite do have @xmath102 models go below 4000k ( note that the models in @xcite are primarily o / ne rather than c / o ) .
however , we believe the @xmath94 upper limits to be more robust , as they do tend to agree between different calculations . while extreme , the companion to may not be especially unique .
similar ultra - cool wds are presumably present in globular clusters and in the field even if they are often too faint to identify on their own .
individual ultra - cool wds can be identified but only if very nearby , like the two objects in @xcite at @xmath103pc . if we correct roughly for the different progenitor masses between the @xcite systems and @xcite and use a @xcite initial mass function , we would estimate @xmath104 massive wds of a similar age within 300pc , which is of the same order as the luminosity function from ( * ? ? ?
* also see @xcite ) extrapolated and @xcite say that at most a few percent of wds are lost off the faint end of the luminosity function . ] to @xmath105 .
instead , binary systems are the best way to identify cold wds ( e.g. , * ? ? ?
* ) , which is effectively the technique used here .
but even in binary systems where we know that a source is present , the systems will often be too distant for good constraints ( i.e. , psr j1454@xmath245846 in ; @xcite ) .
we still require a fortuitously nearby system for useful observations .
the occurrence of a nearby massive wd like the companion to is reasonably consistent with expectations based on the observed binary population : there are five pulsar binaries from the atnf pulsar catalog @xcite within 300pc , and the other four have low - mass he wd companions .
this @xmath106 ratio is similar to that for co wd compared to he wd companions in the whole atnf catalog ( also see @xcite ) , and the pulsars spin - down ages appear to have similar distributions for both companion types .
finally , we can ask whether an ns is the most likely companion to an ultra - cool wd .
most binaries are assumed to have mass ratios near one ( @xcite , but see @xcite ) , but a binary composed of two ultra - cool wds would be just as hard to detect optically as a single object .
if the companion were a lower - mass wd or a main - sequence star the binary could be visible , although it would require spectroscopic follow - up to identify the companion and in the absence of _ gaia _ this has not been done for the majority of stars within a few hundred pc .
so the situation of , with an ns companion , is reasonably plausible as the initial mass ratio would have been close to one and the chances of companion follow - up and identification after discovery of the pulsar are high .
we have determined an accurate mass for the partially recycled pulsar and its companion ; the latter is value consistent with both an ns and a wd . despite not finding the companion in a deep optical / near - infrared search
, we reject a dns explanation as the binary system shows evidence of circularization requiring mass transfer after the last supernova .
instead the companion is likely a high - mass wd . using the extremely precise distance determination from @xcite
, we can set a robust limit of @xmath107 .
this implies an very old and cool wd : fainter than all other pulsar companions by a factor of about 100 , and fainter than the lower - mass `` ultra - cool '' wd in the solar neighborhood by a factor of about four .
converting this limit to a temperature depends somewhat on the assumed mass and composition , but we believe an effective temperature limit of @xmath108k is a robust upper limit .
for such an object to not be older than the milky way requires that it have already entered the faster debye cooling regime , i.e. , that it already crystallized ( also see @xcite ) . future searches , if they can detect the companion to , will be a unique probe of the very late stages of wd evolution , with a well - determined mass and radius that are not usually available for studies of such objects .
we thank an anonymous referee for useful suggestions , and t. tauris , m. van kerkwijk , i. stairs , p. bergeron and r. oshaughnessy for helpful discussions .
is supported by the national science foundation grant ast-1312822 .
m.a.m . and d.r.l .
are supported by wvepscor , the nsf pire program , and the research corporation for scientific advancement .
jrb acknowledges support from wvepscor , the national radio astronomy observatory , the national science foundation ( ast 0907967 ) , and the smithsonian astrophysical observatory ( chandra proposal 12400736 ) .
a.t.d . was supported by an nwo veni fellowship .
some of 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 authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community .
we are most fortunate to have the opportunity to conduct observations from this mountain . based on observations obtained at the southern astrophysical research ( soar ) telescope , which is a joint project of the ministrio da cincia , tecnologia , e inovao ( mcti ) da repblica federativa do brasil , the u.s
. national optical astronomy observatory ( noao ) , the university of north carolina at chapel hill ( unc ) , and michigan state university ( msu ) . funding for sdss - iii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , and the u.s .
department of energy office of science .
the sdss - iii web site is http://www.sdss3.org/. we made extensive use of simbad , ads , and astropy ( http://www.astropy.org ; @xcite ) . ,
i. h. 2010 , in iau symp .
261 , relativity in fundamental astronomy : dynamics , reference frames , and data analysis , ed .
s. a. klioner , p. k. seidelmann , & m. h. soffel , ( cambridge : cambridge univ . press ) , 218
|
the recycled pulsar is one of the closest known neutron stars , with a parallax distance of @xmath0pc and an edge - on orbit .
we measure the shapiro delay in the system through pulsar timing with the green bank telescope , deriving a low pulsar mass ( @xmath1 ) and a high companion mass ( @xmath2 ) consistent with either a low - mass neutron star or a high - mass white dwarf .
we can largely reject the neutron star hypothesis on the basis of the system s extremely low eccentricity ( @xmath3)too low to have been the product of two supernovae under normal circumstances .
however , despite deep optical and near - infrared searches with soar and the keck telescopes we have not discovered the optical counterpart of the system .
this is consistent with the white dwarf hypothesis only if the effective temperature is @xmath4k , a limit that is robust to distance , mass , and atmosphere uncertainties .
this would make the companion to one of the coolest white dwarfs ever observed .
for the implied age to be consistent with the age of the milky way requires the white dwarf to have already crystallized and entered the faster debye - cooling regime .
|
the evolution of a galaxy is determined by how efficiently and rapidly it can turn interstellar gas into stars .
several models have recently been proposed to explain star formation rates and the interstellar gas cycle in galaxies by self - regulating feedback @xcite .
however , these models typically only include a single dominant feedback mechanism among radiation pressure , thermal / energetic input via an hii region , or mechanical input from winds and supernovae .
all such models require more detailed constraints from observations that resolve molecular clouds near stellar clusters and associations that are energetically affecting the interstellar medium ( ism ) . in galaxies with sub - solar metal abundances ,
less efficient molecular line and dust cooling may allow clumps to remain warmer and resist gravitational collapse .
reduced metallicity massive stars emit harder radiation , and that radiation penetrates more deeply into molecular clouds because of reduced dust abundances and possibly a clumpier structure ( e.g. * ? ? ?
* ) . to understand radiative transfer in molecular clouds , and the effect of radiative feedback on star formation
, one must resolve the self - sheilded clumps from more diffuse inter - clump gas , in which co is more easily dissociated than h@xmath7 @xcite .
most star formation in a molecular cloud occurs in dense sub - parsec sized clumps , which only constitute a fraction of the cloud s volume .
detailed analysis of star formation physics thus requires sub - parsec resolution and accurate column density and mass measurements , which before 2012 was very challenging outside of the milky way .
however , a much more extreme range of physical conditions is accessible outside the milky way than within it : more massive stellar clusters with more energetic feedback , metal abundances significantly different from solar , and different galaxy - scale dynamics .
the most extreme extragalactic molecular cloud and star formation environments in the local universe are super - star clusters ( sscs ) .
sscs form thousands of massive stars in a cubic parsec ; young sscs in the present - day universe offer a unique laboratory in which to study the impact that this extreme mode of star formation has on its environment , including subsequent star formation .
sscs are also likely a dominant mode during the era of galaxy assembly z@xmath02 @xcite , so they are important to understand galaxy evolution .
a premier laboratory for studying feedback and star formation is 30 doradus in the large magellanic cloud ( lmc ) . at 50kpc @xcite ,
r136 in 30 doradus is the nearest ssc by an order of magnitude .
r136 has @xmath11 stars within a few parsecs , allowing investigation of molecular cloud physics at reduced metallicity ( @xmath01/2 solar ; * ? ? ?
* ; * ? ? ?
* ) bathed in a strong radiation field .
star formation in 30 doradus has been going on in multiple generations for @xmath020myr @xcite . by studying how new stars are forming in this environment
we can determine if the efficiency ( stellar divided by molecular mass ) or stellar initial mass functions differ in ssc environments .
co has been studied previously in 30 doradus , as part of lmc surveys , notably @xmath2co@xmath31 - 0 with nanten ( * ? ? ?
* 40pc ) and @xmath2co and @xmath4co@xmath31 - 0 with mopra ( 33= 8pc , * ? ? ?
* ; * ? ? ?
30 doradus was among selected regions observed with the swedish - eso submillimetre telescope ( sest * ? ? ?
* and references therein ) ; @xcite present @xmath2co and @xmath4co@xmath31 - 0 at 45 resolution , and @xmath2co and @xmath4co@xmath32 - 1 at 23 resolution .
before the observations presented here , the highest resolution observed in extragalactic co was 4pc ( @xmath2co@xmath33 - 2 in the lmc with sest ; * ? ? ? * ) .
spatial resolution limitations have resulted in a traditional divide between extragalactic and galactic studies of molecular gas .
extragalactic studies usually consider giant molecular clouds ( gmcs ) as entities 10 - 50@xmath3pc in size , using a small number of abundant molecules , and usually only the strongest optically thick transitions of @xmath2co .
populations of gmcs follow power - law scaling relations in size , linewidth , and @xmath2co luminosity @xcite . by contrast ,
molecular gas emission from milky way molecular clouds are routinely resolved with better than 1pc resolution , and often better than 0.1pc .
proximity makes it possible to observe many transitions of multiple species .
in particular , for optically thick molecules like @xmath2co , a minor isotopologue can be mapped to accurately determine column densities .
clumps on these smaller scales deviate from the larger - scale scaling relations @xcite .
the recently inaugurated atacama large ( sub)millimeter array ( alma ) now allows extragalactic molecular clouds to be studied with similarly high resolution and sensitivity as galactic surveys , revealing the detailed structure and physics .
this paper presents the first sub - parsec extragalactic investigation of molecular emission , in 30 doradus . with alma
we can address how molecular cloud structure is affected by radiative feedback , and whether the dense star - forming structures in clouds are different in such extreme environments .
the new alma observations are described in [ obs ] as well as new apex observations and archival optical to mid - infrared data .
the co@xmath32 - 1 intensity distribution is analyzed in
[ overview ] , and the relationship between co emission and sources in the region is discussed in [ ysos ] . in particular , for the first time we can isolate the denser parts of the molecular cloud directly involved with star formation from the lower density diffuse gas , and associate molecular clumps with young stellar and protostellar clusters .
section [ mass ] describes how temperature and mass are calculated from co and continuum emission .
analysis of dense co clumps is presented in [ clumps ] ; comparison of clump masses calculated from co , dust , and the virial theorem constrains the molecular and dust abundances , the amount of unobserved h@xmath7 not traced by co , and the dynamical state of the clumps .
key results are summarized in [ conclusion ] .
30 doraduswas observed with alma @xcite during early science cycle 0 , using the 1.3@xmath3 mm band 6 recievers @xcite .
the spectral setup had 4@xmath11ghz fdm ( frequency division mode ) spectral windows centered at 218.5 , 219.267 , 230.767 , and 231.688ghz with 244khz ( @xmath00.33km@xmath3s@xmath12 ) resolution .
at the 250 km@xmath3s@xmath12 lsrk velocity of 30 doradus , these simultaneously cover @xmath2co@xmath32 - 1 ( 230.538ghz rest frequency ) , @xmath4co@xmath32 - 1 ( 220.399ghz ) , the h30@xmath6 radio recombination line ( 231.90093ghz ) , c@xmath5o@xmath32 - 1 ( 219.560ghz ) , h@xmath7co@xmath13 and @xmath14 ( 218.47563 , 218.76007ghz ) , hc@xmath15n@xmath324 - 23 ( 218.32472ghz ) , and @xmath4cs@xmath35 - 4 ( 231.2210ghz ) .
observing dates and weather conditions are listed in table [ sb ] . in all observing sessions ,
the quasar j0538 - 440 was used as bandpass calibrator ( flux density @xmath03jy at the time of observation ) and j0635 - 7516 as phase calibrator ( flux density @xmath01jy ) .
callisto was observed before the first observation on 2011 - 12 - 31 , to set the absolute flux scale with estimated uncertainty of @xmath16 using the butler - jpl horizons-2010 model .
the flux scale was then transferred to the other datasets by requiring the two quasars to have the same flux densities .
the data were processed in the common astrononomy software applications ( casa ) package ( http://casa.nrao.edu ) , and visibilities imaged and deconvolved with 0.5@xmath3km@xmath3s@xmath12 channels .
the expected rms per channel is 8@xmath3mjy@xmath3bm@xmath12 , and 810@xmath3mjy@xmath3bm@xmath12 @xmath170.15@xmath3k is achieved away from strong lines ( 510 times worse sensitivity near bright lines because the data there are dynamic range limited , i.e. bright emission creates artifacts in the fainter parts of the image ) . the synthesized beam is @[email protected]= [email protected] for @xmath2co@xmath32 - 1 and scales approximately with frequency for other transitions . the array configuration is sensitive to spatial scales from 1.1to 20(0.28 to 5@xmath18pc ) .
llllll uid://a002/x35edea / x228 & 13 & 2011 - 12 - 31 01:58:12 ut & 43@xmath19 & 3.5@xmath3 mm & 27.7min + uid://a002/x35edea / x5ad & 13 & 2011 - 12 - 31 05:00:33 ut & 40@xmath19 & 4.0@xmath3 mm & 27.7min + uid://a002/x35edea / x77d & 12 & 2011 - 12 - 31 06:31:23 ut & 33@xmath19 & 3.9@xmath3 mm & 27.8min + uid://a002/x36a70f / x895 & 15 & 2012 - 01 - 11 03:49:04 ut & 41@xmath19 & 2.2@xmath3 mm & 27.7min + observations with a single - dish telescope are necessary to recover the total flux on large spatial scales , which is resolved out by an interferometer .
our single - dish observations were made in 2011 october 2127 , 29 , 30 and december 3 , 4 , 12 , and 13 with the 12 m atacama pathfinder experiment telescope , apex @xcite .
apex has a 23 beam at 230@xmath3ghz .
unfortunately , the alma compact array was not yet available at the time of these observations .
the region was mapped twice using on - the - fly observations with the apex-1 single - sideband receiver that has a 2@xmath3ghz spectral window ( consist of two 1@xmath3ghz bands ) and a full resolution of 30.5@xmath3khz ( @xmath20 km@xmath3s@xmath12 at 230 ghz ) .
the mapping , centered on ( 05:38:47.8 , -69:04:50.3 ) with a size of 50@xmath160 , was first done in @xmath4co@xmath32 - 1 for 10h and for another 22h with the spectral window covering @xmath2co@xmath32 - 1 and h30@xmath6 .
an emission - free position at ( 05:40:19.40,-69:01:50.6 ) was used as the off - source position and observed at the end of each scanning leg .
the calibration was performed every @xmath0 10 min with cold and ambient temperature loads .
the alma correlator is more capable and permitted simultaneous observation of additional lines ( see below ) , but these are the only three that we observed with apex .
the data were calibrated with the apex online calibrator ( including application of t@xmath21 temperature scale ) and then further reduced in class ( http://www.iram.fr/iramfr/gildas ) for baseline subtraction on individual spectra and gridding .
the apex data was converted from main beam temperature to jy@xmath3bm@xmath12 by multiplying by 39 ( * ? ? ?
* http://www.apex-telescope.org/telescope/efficiency/ ) , and the resulting cube was exported to casa for combination with the alma data .
rms noise was 0.5jy / bm or 0.03k per 0.5@xmath3km@xmath3s@xmath12channel .
peak intensities measured were 180@xmath3jy beam@xmath12= 4.7k and 21@xmath3jy beam@xmath12= 0.55k in @xmath2co@xmath32 - 1 and @xmath4co@xmath32 - 1 , respectively .
@xmath2co@xmath32 - 1 is significantly brighter ( @xmath13 ) than measured with sest at similar resolution ( although more than an order of magnitude worse sensitivity ) to our apex data @xcite .
alma absolute flux calibration is better than 10% , and agrees with the apex absolute flux calibration to 20% .
it is possible that the quoted sest values are not from exactly the same region since sest does not have a data archive , it is not possible to verify their findings .
the apex image cube was combined with a cleaned alma cube using the casa task feather , which fourier transforms the two images , multiplies the single dish image by the fourier transform of the single dish beam , the interferometer image by the complement of that transfer function , adds the two filtered fourier cubes , and transforms back .
this feathered image is a simple way to combine the two datasets but can be sensitive to the shape of the tapering function and to relative calibration uncertainties . to mitigate these effects we used the feathered image as a model in clean deconvolution , which allows the cleaning process to correct any potential issues with the feathered image which might be inconsisent with the interferometric visibilities .
most ( 80% ) of the apex flux is recovered by alma , so the properties of small clumps do not depend strongly on how the datasets are combined .
( in this paper `` clump '' refers to the structures containing local maxima and @xmath00.5 - 2pc in size , @xmath0100 - 1000@xmath3m@xmath10 in mass , defined by the decomposition described in
[ clumpdecomp ] ) .
we experimented with small variations in the relative flux calibration between the two telescopes , and conclude that the image combination process results in less than 10% uncertainty of @xmath2co clump properties and less than 5% uncertainty of @xmath4co clump properties .
multiwavelength data was retrieved from various archives as listed in table [ archival ] .
we processed eso data through the gasgano data pipeline ( http://www.eso.org/sci/software/pipelines/ ) ; other telescopes provide calibrated reduced archival images .
astrometric registration was performed iteratively starting with 2mass and _ spitzer_/sage , then ground - based images , followed by hst , which has the smallest field of view .
final image registration and distortions are estimated to be @xmath80.4@xmath22 over the alma observed region .
near - infrared ( nir ) data were cross - calibrated to the hst flux scale , then background - matched and mosaicked using montage ( http://montage.ipac.caltech.edu/ ) .
complete coverage at the highest resolution is not available in all filters ; in particular for the nir images discussed below , the f160w and f110w images are from hst , whereas the f205w image is largely from ground - based sofi , with nicmos data inset where available .
lllllll wfc3 & f110w , f160w @xmath0j , h & 0.15 & 12939 & sabbi & 2012 & + nicmos & f205w @xmath0k@xmath23 & 0.2 & 7819 & walborn & 1997 - 8 & @xcite + sofi & j , k@xmath23 & 0.75 & 63.h-0683(a ) & schmutz & 1998 - 9 & @xcite + spitzer / sage & irac 3.5 - 8@xmath24 m & 2 & 20203 & meixner & 2005 - 6 & @xcite + atca & 3cm,6 cm & 2 & c109,c234 & lazendic , dickey & 1993,98,99 & @xcite +
for the first time we can study the detailed subparsec - scale clump mass distribution in extragalactic clouds . in cycle 0 , alma had 1/4 of its expected final sensitivity ( and more than 30 times coarser spatial resolution ) , so the observations presented here were designed as a pilot program . at @xmath2510pc resolution , there are two gmcs within 20pc of the r136 cluster ; we chose to observe the northern , more massive cloud 30dor-10 ( nomenclature from * ? ? ?
* ) . in the following subsections we present the 3-dimensional co emission distribution , describe other detected lines and the continuum .
subsequent sections will connect the co emission to current star formation detected with infrared and maser emission , and quantitatively analyze the molecular mass distribution and clump properties .
figure [ overviewfig ] shows integrated @xmath2co@xmath32 - 1 emission together with mid - infrared continuum ( from the _ spitzer_/irac sage program , * ? ? ?
for the first time , the molecular gas is imaged at the same resolution as _ spitzer_/irac .
@xmath2co is detected across most of the mapped region .
the r136 cluster is 11 pc in projection southwest from the bottom of 30dor-10 .
the edge of the bubble evacuated by r136 is clear in both infrared and co emission the high emission measure edge is enhanced at 4.5@xmath24 m ( green ) because of bright br@xmath6 recombination line emission in that irac band .
deeper in the molecular cloud , knots of co emission correlate to some extent with mir continuum from protostars ; detailed comparison will be discussed in [ ysos ] .
the @xmath2co@xmath32 - 1 integrated maps using alma and apex alone are compared to the combined map in figure [ apexcombine ] .
figure [ int13co ] shows integrated @xmath4co@xmath32 - 1 emission , noting the numbers of the most prominent clumps ( clump decomposition will be described in section [ clumpdecomp ] ) .
there is not an overwhelming amount of large - scale or extended @xmath2co emission in 30dor-10 .
as already noted , 80% of the apex flux is recovered by alma .
the intensity distribution is further quantified in figure [ intenshist ] which compares the cumulative brightness distribution in 30dor-10 and in the galactic massive star formation region w3 @xcite .
the w3 @xmath2co@xmath32 - 1 cube has physical resolution of 0.3pc ( at 2kpc distance ) , 0.15k rms noise per 0.5@xmath3km@xmath3s@xmath12channel , and peak antenna temperature of 54k , compared to the same quantities in 30dor-10 of 0.5pc , 0.15k , and 62k , respectively ( [ obs ] ) .
there is more diffuse @xmath2co emission in a galactic region like w3 than in 30dor-10 .
a reasonable interpretation of this result is that at subsolar metallicity and increased radiation density , the less - dense `` interclump '' co is being selectively photodissociated as expected from pdr models @xcite .
models of low - spatial - resolution molecular line emission derive volume and area filling fractions in 30dor-10 using escape probability techniques .
derived area filling fractions using co@xmath31 - 0 to co@xmath33 - 2 include 0.05 - 0.1 @xcite , 0.05 @xcite , 0.015 @xcite . in our integrated intensity maps , 2/3 of the @xmath2co@xmath32 - 1 flux
is contained in 15% of the spatial area , and half of the flux in 8.9% of the spatial area . for @xmath4co@xmath32 - 1 , 2/3 of the flux
is contained in 9.8% of the area , and half of the flux in 5.6% of the area .
thus most of the flux in our map is contained within 510% of the area , consistent with the escape probability - derived filling fractions .
@xcite derived 0.06 from co@xmath34 - 3 and co@xmath37 - 6 these higher j transitions should be less sensitive to optical depth effects and a better tracer of the dense gas , but also may not be the same structures emitting co@xmath32 - 1 .
figure [ moment0 ] shows integrated @xmath4co emission as a function of right ascension ( r.a . ) and velocity , and as a function of declination and velocity , i.e. the cube collapsed along each spatial axis .
the cloud is highly structured in position and velocity - in particular the brightest clump 52near the center of the map may be at the intersection of two converging flows , as expected from simulations of turbulent clouds ( e.g. * ? ? ?
* ; * ? ? ?
the overall velocity structure of the cloud has a moderate east / west gradient , most evident in @xmath2co emission , although the much more dominant features are the relative velocities of individual clumps , typically separated by 3 - 10@xmath3km@xmath3s@xmath12 .
figure [ moments ] shows the first and second velocity moments .
clump linewidths are 23@xmath3km@xmath3s@xmath12fwhm ( see analysis in section [ size - linewidth ] below ) .
the second velocity moment ( velocity dispersion ) shows a few high peaks ( figure [ moments ] ) , but most of those are merely the spatial superposition of two clumps at different velocities .
clump 72truly does have a very large linewidth ( 10.5@xmath3km@xmath3s@xmath12 in @xmath2co@xmath32 - 1 ) .
that clump contains an embedded mid - ir protostar or protocluster ( see section [ ysos ] below ) , and the large linewidth likely results from massive outflows .
several other lines were observed simultaneously with @xmath2co@xmath32 - 1 and @xmath4co@xmath31 - 0 .
c@xmath5o is detected in the brightest central clump 52 , the tip of the pillar / clump 4 , as well as several other locations in the cloud ( figure [ c18o ] ) .
h@xmath7co 3 - 2 , hc@xmath15n 24 - 23 , and @xmath4cs 5 - 4 are not convincingly detected , although there are some suggestive spectral features associated with the brightest @xmath4co peaks at the @xmath02@xmath26 level .
nondetection is consistent with previous single - dish observations , but h@xmath7co nondetection may be a selection effect , since due to correlator restrictions we observed the weakest of three 3 - 2 lines , and astrochemical models predict fairly strong h@xmath7co @xcite .
we also detected the h30@xmath6 radio recombination line ( rrl ; @xmath27 = 231.90093@xmath3ghz ) in 30 doradus .
the rrl emission is mostly coincident with peaks in the 3 cm radio continuum emission in @xcite and young massive stars in @xcite , and not coincident with molecular gas peaks .
analysis of this data will be presented in kepley et al .
( 2014 , in prep ) .
hcn@xmath31 - 0 , hco@xmath281 - 0 , cs@xmath32 - 1 , and h40@xmath6 at @xmath090ghz were also observed as part of this project and will be presented in brogan et al .
( 2013 , in prep ) .
note that we do not have single - dish data for any of these lines except for h30@xmath6 , which is not detected with apex .
the 1.3@xmath3 mm continuum is also imaged with high fidelity throughout the region .
note that unlike the co cubes analyzed throughout this paper , we do not have single - dish continuum observations , and some large - scale diffuse emission may be resolved out ( the expected largest angular size from the uv distribution of the data is 20 ) . figure [ cont ] compares the alma 1.3@xmath3 mm continuum to 6@xmath3 cm continuum imaged at similar resolution with the australia telescope compact array @xcite .
those data have a synthesized beamsize of [email protected] , and no single dish data included .
the 1.3@xmath3 mm continuum traces the rim on the southern part of the cloud , and likely includes a significant contribution from free - free emission at that location , but further north is morphologically distinct from tracers of ionized gas , and is thus likely mostly dust emission .
in particular the central and eastern dense gas clumps 52and 72are bright in the continuum and both contain an embedded protostar or protocluster ( see section [ ysos ] ) .
the free - free contribution to the 1.3@xmath3 mm continuum can be subtracted by assuming that the 6@xmath3 cm continuum is 100% optically thin free - free emission , and scaling s@xmath29 by @xmath30 . to properly scale and subtract the 6@xmath3 cm image from the 1.3@xmath3 mm image
requires matching the _ uv _ coverage of the two interferometers .
although the restoring beams are not drastically different , a simple scaling by the beam areas ( the images have native units of jy@xmath3bm@xmath12 ) and @xmath30 factor results in negative values over significant parts of the `` dust emission '' map . to better perform the subtraction , we spatiallly filtered the 6@xmath3 cm image to match it to alma s spatial sensitivity , and then convolved both maps to [email protected] .
a total of 39% of the total flux in the 1.3@xmath3 mm image is removed as free - free contamination .
this may still represent oversubtraction of the free - free contribution , but we choose to be conservative in what we maintain as dust emission in subsequent analysis .
the millimeter flux densities of clumps in the body of the molecular cloud are insensitive to this scaling , since their free - free contribution is much smaller than the dust ( the cm flux density is much less than the mm flux density ) .
those on or near the photodissociated rim including clump 4 , 21 , and 11 , have a larger free - free contribution .
we also performed the free - free subtraction using the 3@xmath3 cm image from @xcite , with quantitatively similar results .
figure [ nir_ysos ] shows subarcsecond resolution near - infrared ( nir ) images ( mostly from hst ; see [ archivalsection ] ) overlaid with @xmath2co@xmath32 - 1 contours and with previously known sources marked as well as notable co clumps .
of particular interest is the relation between dense molecular clumps and star formation .
we discuss previously identified star formation from most to least evolved , namely nir protostellar clusters , mid - infrared massive ysos , and masers .
the compact cluster k1 ( original designation from @xcite , studied in more detail by @xcite ) is located at the tip of pillar / clump 4 , in the classic `` pillars of creation '' arrangement seen throughout the milky way @xcite .
the stars are associated with diffuse nir emission , likely brackett@xmath31 emission from the ionized gas being photoablated from the pillar tip . to the east
are two very red young stellar objects only marginally resolved at hst resolution ( irsn122 and 126 in * ? ? ?
both ysos are also located in a pillar - like structure ( clumps 21and 11 ) .
these three regions are morphologically suggestive of triggered star formation via radiatively driven implosion @xcite , but _ proving _ triggering is complex , and these may simply be _ revealed _ clusters that formed in the relatively denser molecular clumps , now left behind after less - dense outer layers are ablated by radiation from r136 to the south ( e.g. * ? ? ?
cluster k1 is unresolved by _
irac , and was designated a massive young stellar object ( myso ) candidate by @xcite , based primarily on its [email protected]@xmath24 m color and infrared morphology . it is somewhat extended in irac images , and thus not included in the more conservative sage point source catalog @xcite or myso list derived from it @xcite .
it is not clear that the cluster resolved by hst contains any intermediate or massive embedded ysos .
this example does highlight a general issue with _ spitzer _
point sources , which we and the identifying authors all call mysos : these are almost certainly small clusters or multiples .
the size and mass scales ( @xmath010@xmath9m@xmath10 , @xmath321pc ) probed by the current alma data are not small enough to be associated with individual protostars , but on the other hand it is unlikely that any massive star formation is _ not _ associated with a clump described here .
co clump 4immediately behind ( in projection , from the point of view of r136 ) k1 is however a likely candidate to contain a very young myso .
clump 4has @xmath2co and @xmath4co@xmath32 - 1 peak brightness temperatures of 61k and 13k respectively and is very compact , suggesting that it may be internally heated ( although no irac source can be separated from k1 and other diffuse emission in the area ) .
oh maser emission at 1720mhz was detected coincident with clump 4and cluster k1 ( circle in fig .
[ nir_ysos ] ) by @xcite and @xcite .
1720mhz oh masers are typically found toward both star formation regions and supernova remnants ; in this case the maser velocity 243@xmath3km@xmath3s@xmath12 is offset from the clump velocity 248@xmath3km@xmath3s@xmath12 , suggesting that the emission is _ not _ associated with the star formation , but on the other hand , previous searches for supernova remnants within 30 doradushave also been unsuccessful or inconclusive @xcite .
the cluster k2 in the center of the field is not cospatial with a co clump the nearest ( clump 52 ) is adjacent .
presumably , the cluster has largely dispersed the dense gas from which it formed , although there are remnants visible in the irregular filamentary structure extending from co clump 52to nir cluster k2 .
a greater spread in age and/or reddening in the optical color - magnitude diagram of k2 presented by @xcite supports the notion that it may be more evolved than k1 .
the cluster has also created its own small hii region , evident in the cm continuum ( [ continuumdata ] ) .
a high / ratio @xcite indicates that this hii region is more highly ionized ( higher ionization parameter and/or harder ionizing field ) than the ionized gas on the rim of the bubble around r136 ( the ridge at the bottom of our field ) .
@xcite classify the brightest source in the cluster , parker 1429 , as o2v((f*))z ; this very young and hot star may be the source of relatively hard ionizing radiation .
the co clumps on either side of cluster k2 each contain much younger mysos .
clump 39to the west contains a 22.235ghz h@xmath7o 6@xmath33 - 5@xmath34 maser at the same velocity as the co emission ( pink @xmath35 on figure [ nir_ysos ] * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the milky way
, water masers are commonly associated with massive young stellar objects ( e.g. * ? ? ?
* ) , and they are beginning to be used in more distant galaxies as signposts of massive star formation as well ( e.g. * ? ? ?
* ; * ? ? ?
the masers require densities @xmath3610@xmath37@xmath38 and warm gas @xmath36400@xmath3k , and are formed in outflows or compressed circumstellar material around mysos . in the galaxy ,
even mysos that are sufficiently embedded that they do not have detectable nir emission are sometimes associated with 22ghz h@xmath7o masers ; there is no detectable mir emission in clump 39 , but the maser indicates a likely very young internal source .
clump 52to the east contains a nir - red source unresolved by hst , which is a bright mir source ( fig .
[ overviewfig ] ) , identified as a myso by @xcite .
it also is associated with 22ghz maser emission at the same velocity as the co , but the co emission is extended in the north / south direction , and the maser position is 2.0 north of the nir source @xcite . the mid - ir myso is discussed as `` s9 '' in @xcite . clump 72contains a 22ghz h@xmath7o maser @xcite and a very red nir source unresolved by hst . the nir source is a bright mir @xcite myso .
furthermore , this clump has the broadest linewidth of any in our data , probably due to outflows from the embedded myso(s ) .
cluster p4 @xcite to the north is unassociated with any co emission peak , so we suggest that it is also relatively evolved and has largely cleared its birth material ( although not completely , as evidenced by the filament seen in absorption on the northwest side of the stellar cluster ) .
@xcite designate the brightest mir source in the cluster `` s8 '' .
column density was calculated from @xmath2co and @xmath4co@xmath32 - 1 using common assumptions @xcite , described in this section along with their caveats .
if @xmath2co@xmath32 - 1 is very optically thick , the excitation temperature t@xmath39 is only a function of the brightness temperature .
@xmath40 where i@xmath41 is the @xmath2co@xmath32 - 1 intensity in k. if it is further assumed that @xmath4co has the same excitation temperature as @xmath2co , the relative abundance is constant , and that co is thermalized ( the level population is described by the boltzmann distribution evaluated at the excitation temperature ) , then the column density per velocity bin can be calculated : @xmath42 \\
n(^{13}{co } ) & = & 1.5\times 10^{14 } { { t_{ex } e^{5.3/t_{ex}}\int \tau^{13}_v dv}\over{1-e^{-10.6/t_{ex } } } } \label{nco}\end{aligned}\ ] ] figure [ col ] shows the calculated excitation temperature and column densities under these assumptions ( the validity of which we consider below ) .
the peak @xmath4co column density is @xmath05@xmath110@xmath43@xmath44 in the central clump 52 .
numerous clumps exceed a @xmath4co column density of 10@xmath43@xmath44 ( h@xmath7 column density of 10@xmath45@xmath44 with the abundances discussed below ) .
models of low - spatial - resolution co emission ( [ introduction ] ) agree on a low filling factor of dense gas , @xmath815% and as low as 1.5% in the lvg models of @xcite .
this can now be compared to our high - resolution data .
the fraction of our map with n(@xmath4co)@xmath465@xmath110@xmath43@xmath44 is less than 1% , quite consistent with the low - resolution lvg parameters .
the assumptions of this `` molecular column density '' or `` lte mass '' calculation have been considered by numerous authors in the past ( e.g. * ? ? ?
* ; * ? ? ?
the higher abundance of @xmath2co relative to @xmath4co means that its effective critical density ( including radiative trapping ) is lower . for densities
n(h@xmath7)@xmath81000@xmath18@xmath38 ( depending on column density ) , @xmath2co will be largely thermalized , whereas @xmath4co remains subthermally excited , lowering the calculated @xmath4co column density ( these statements are true for 1 - 0 and 2 - 1 as one can verify with radex ; * ? ? ?
however , when subthermally excited , the partition function overestimates the population of higher rotational levels , and the calculated total @xmath4co column density could be incorrectly high .
@xcite illustrate this effect for co@xmath31 - 0 in relatively dark clouds , showing that the lte method can underestimate the column in low - density ( n(h@xmath47500@xmath38 ) envelopes by a factor of two and overestimate it for 10@xmath48n(h@xmath7)@xmath810@xmath49@xmath38 .
@xcite concluded from numerical simulations that lte mass can underestimate the true mass by a factor of a few , but the effect is largest at @xmath4co column densities @xmath810@xmath50@xmath44 and h@xmath7 volume densities @xmath810@xmath51@xmath38 as we will determine in [ clumps ] , most of the dense clumps in 30 doradus have column densities @xmath3610@xmath50@xmath44 , and volume densities 10@xmath52n(h@xmath7)@xmath5310@xmath54@xmath38 .
we expect the analysis to be robust , but that individual masses may have uncertainties of a factor of two .
observations of co@xmath32 - 1 on large ( 100 - 1000pc ) scales indicates that the factor of two higher critical density makes it a moderately less robust tracer of bulk molecular gas mass than co@xmath31 - 0 @xcite .
however this occurs because most of a gmc has relatively _ low _ column and volume density
the clumps and dense warm molecular gas observed at much smaller scales here are well - traced by co@xmath32 - 1 .
analysis of co@xmath32 - 1 by @xcite finds that the lte analysis is appropriate for n(h@xmath7)@xmath3610@xmath49 and @xmath2co optical depth @xmath361 , i.e. brightness temperature ratio t@xmath55(@xmath2co@xmath32 - 1)/t@xmath55(@xmath4co@xmath32 - 1 ) @xmath3620 .
the latter condition is violated in only a vanishingly small fraction of our @xmath4co cube ( @xmath530.5% ) , on the edges of clumps .
since we primarily compare clumps in 30dor-10 with those calculated with the same method in other clouds , we note these possible systematic errors but do not expect them to affect our conclusions .
another effect , of particular importance in regions of high radiation intensity , is that of differential abundance , or isotopic fractionation . in diffuse gas ( a@xmath56 1 )
lower @xmath4co column density results in lower self - shielding and a reduced abundance relative to @xmath2co ( e.g. * ? ? ?
* ) . in dense pdrs and clouds ,
this effect is counteracted by ion - molecule fractionation @xmath57 .
selective photodissociation only dominates in optically thin diffuse molecular gas , whereas for the conditions in 30dor-10 , @xmath4co is chemically enhanced at the @xmath581 surface and fractionation is minimal deeper in the cloud where most of the observed photons are emitted @xcite .
next , one must consider that @xmath2co becomes optically thick near the surface of a pdr ( within n(h@xmath7)@xmath010@xmath59 of the transition from c@xmath60 to c@xmath61 to co , typically at @xmath622 for the milky way gas to dust ratio ) , whereas @xmath4co remains at most moderately optically thick throughout a dense clump .
even if the lower @xmath63 levels are thermalized , the kinetic temperature is quite high on the surface of a pdr ( several 100k ) .
one may question whether the @xmath2co excitation temperature is thus representative of most of the @xmath4co - emitting gas .
we estimate from pdr models that the kinetic temperature at the @xmath2co photosphere is no more than @xmath01.5 times the bulk of @xmath4co emitting gas @xcite , resulting in less than @xmath010% error in the calculated co column densities , unless the center of the clump is very cold @xmath5310k .
converting the measured column densities to h@xmath7 column densities requires two abundance ratios : @xmath2co/@xmath4co , and h@xmath7/@xmath2co .
the @xmath2co/@xmath4co ratio is 70 in the solar neighborhood ; atomic isotopic ratios @xmath2c/@xmath4c and @xmath43o/@xmath5o increase with decreasing metallicity in the outer milky way @xcite , but the molecular ratio is less well constrained .
previous measurements of @xmath2co/@xmath4co in 30dor-10 range from 38 @xcite to 50100 @xcite .
the h@xmath7/co abundance depends on the cloud temperature , density , and radiation field .
studies at parsec scales of outer milky way molecular clouds ( at similar 1/2 - 1/3 metallicity to the lmc ) use h@xmath7/@xmath4co of 10@xmath64 @xcite to 3@xmath110@xmath64 @xcite .
the value should increase in photodissociation regions because of selective photodissociation of co relative to h@xmath7 pdr models suggest h@xmath7/@xmath4co @xmath3610@xmath65 even at solar metallicity @xcite , but our analysis shows below that these clumps are better described as dense molecular condensations and not pdrs .
we will use h@xmath7/@xmath4co=5@xmath110@xmath64 , which as shown in section [ massmass ] yields good agreement between several methods of calculating h@xmath7 mass .
once the contribution from thermal bremsstrahlung has been removed , the remaining millimeterwave continuum emission arises from dust , which provides an independent measure of the gas column density and mass .
the intensity i@xmath29 of dust emission is given by : @xmath66 where @xmath67 is the gas mass surface density , @xmath68 the dust emissivity per unit dust mass , and @xmath69 the dust - to - gas ratio by mass .
we adopt @[email protected]@xmath3@xmath72g@xmath12 , midway between 0.72@xmath3@xmath72g@xmath12 used by @xcite in the smc , and 0.52@xmath3@xmath72g@xmath12 used by @xcite in the lmc .
the value is 2 times higher than @xcite , so our dust masses could be low . scaling the gas - to - dust ratio by the lmc metallicity of 1/2 solar @xcite yields @xmath69=1/339 @xcite .
it is not well known the extent to which the gas - to - dust ratio is uniform at such small scales @xcite found significant variations of the fe / h abundance on parsec scales in 30 doradus which might be due to dust destruction . proceeding with this caveat ,
the gas surface density in @xmath72g@xmath12 is @xmath73 if we assume that the co excitation temperature calculated above from the @xmath2co@xmath32 - 1 brightness temperature equals the dust temperature , we can easily make a column density map .
note that the dust temperature derived from _
data by gordon et al .
( 2014 ) in prep . is 55k in the single resolution element containing 30dor-10 , so at least in spatial average , the dust temperature is similar to the gas excitation temperature shown in figure [ col ] . furthermore , since h(@xmath74,cont)/k=10.8@xmath3k is very close to h(@xmath74,@xmath2co@xmath32 - 1)/k = 11.1@xmath3k , we can replace the exponential term with 11.1/t@xmath75 to an accuracy of better than 98% , and @xmath76cm@xmath9g@xmath12i@xmath78/t@xmath75 .
the total mass in the map derived from dust emission is 6@xmath711@xmath110@xmath79m@xmath10 , consistent with 3.6@xmath110@xmath79m@xmath10 and 3.8@xmath110@xmath80m@xmath10 calculated by @xcite from @xmath2co@xmath31 - 0 luminosity and the virial theorem , respectively ( their measurements apply to the entire 30dor-10 cloud , and they do nt use the optically thin isotopologue to calculate molecular mass , so the estimates should only be expected to agree to a factor of several ) .
masses of individual clumps in our data are analyzed next .
emission from complex molecular clouds is often analyzed by decomposition into clouds or clumps , which by construction isolates a particular size scale a few times the spatial beam . in this section
we describe the decomposition , and then how 30dor-10 clumps compare to previously studied molecular clouds in size - linewidth - surface density parameter space .
higher resolution will likely break these clumps up into smaller substructures , but we can gain insight by comparing physical parameters derived from these data with clumps in other clouds decomposed at similar physical resolution .
different decomposition algorithms have been explored , for example gaussclumps @xcite , in which gaussians are fit to the brightest peaks , subtracted , and the process iterated .
such fitting allows the separation of blended structures , but assumes a functional form for the intensity ( gaussians ) . in extragalactic studies , methods which do not presuppose a functional form
have been more common , in particular clumpfind @xcite and cprops @xcite .
both methods select local maxima in the data cube and then assign pixels to those local maxima .
the clumpfind program assigns all pixels with flux above some specified floor to the nearest local maximum , whereas cprops only assigns pixels within an isosurface which contains a single local maximum .
in crowded regions with complex emission , clumpfind tends to produce a `` patchwork quilt '' assignment cube dividing up the low level emission , whereas cprops only assigns the peaks , leaving fairly bright emission unassigned .
we use a hybrid approach in which local maxima are chosen using the cprops criteria , based on their contrast above shared emission , and signal - to - noise
. then emission between local maxima is assigned , down to a flux level containing several local maxima , but not all the way down to the same noise floor across the entire region .
all of the trends discussed below are calculated for a range of decomposition parameters , both with and without the apportionment of emission using the hybrid approach .
although some of the fainter individual clumps properties change , all of the trends below are robust to the decomposition method .
clump properties are measured from the intensity - weighted moments in space and velocity of the pixels assigned to each clump .
these moments are deconvolved by the synthesized beam and velocity channel width , respectively .
the spatial moment must be converted to a radius , which depends in the intensity profile .
the radius of a gaussian beam is often defined as r=2@xmath81 , the 1/e@xmath82 point , or an area containing 68% of the flux .
we adopt the commonly used convention for a radial density profile @xmath83r@xmath12 which yields r=1.91@xmath81 @xcite .
clump properties are listed in table [ properties ] .
radii range from unresolved up to @xmath00.6pc , with a median for nominally resolved clumps of 0.3pc .
the minimum fittable clump size depends on the signal - to - noise as described in @xcite . for the faint clumps in our sample ( 0.75k = 5@xmath26 for @xmath4co@xmath32 - 1 ) ,
the uncertainty in radius if a gaussian were fitted to the clump is 0.35= 0.09@xmath3pc , decreasing linearly with increasing brighness .
we do not use any deconvolved radii less than 0.1@xmath3pc .
velocity dispersions @xmath84 range between 0.5 and 3@xmath3km@xmath3s@xmath12 , with a median value of 1.1@xmath3km@xmath3s@xmath12(fwhm 2.6@xmath3km@xmath3s@xmath12 ) .
approximate volume densities m/(4/3@xmath85r@xmath86 ) range from 10@xmath86 to 10@xmath80@xmath38 , with a median value of 10@xmath79@xmath38 .
lllllllllllllllllll 1 & 5:38:40.53 & -69:04:53.5 & 262.4 & 0.11 & 0.03 & 0.89 & 0.21 & 2.5 & 1.5 & 0.3 & 10.6 & 5.6 & 0.6 & 84 & 19 & 90 & 35 & + 2 & 5:38:40.77 & -69:04:32.7 & 263.8 & 0.11 & 0.05 & 0.85 & 0.59 & 1.7 & 0.6 & 0.6 & 8.6 & 1.3 & 0.3 & 34 & 37 & 83 & 67 & + 3 & 5:38:44.08 & -69:05:12.9 & 250.7 & 0.18 & 0.04 & 0.85 & 0.17 & 5.5 & 4.6 & 0.1 & 30.9 & 30.3 & 0.1 & 307 & 7 & 136 & 39 & + 4 & 5:38:45.12 & -69:05:06.6 & 248.3 & 0.29 & 0.04 & 1.36 & 0.08 & 13.9 & 22.8 & 0.1 & 61.8 & 166.7 & 0.1 & 2256 & 13 & 559 & 91 & pillar , cluster k1 + 5 & 5:38:45.65 & -69:05:04.9 & 251.1 & 0.22 & 0.04 & 0.42 & 0.08 & 3.2 & 3.8 & 0.1 & 27.3 & 33.5 & 0.1 & 260 & 8 & 40 & 11 & + 6 & 5:38:45.68 & -69:05:03.3 & 247.9 & 0.11 & 0.03 & 0.68 & 0.13 & 2.1 & 0.8 & 0.4 & 20.1 & 11.5 & 0.4 & 51 & 25 & 52 & 17 & + 7 & 5:38:45.65 & -69:05:08.5 & 251.1 & 0.11 & 0.02 & 0.72 & 0.13 & 3.8 & 1.9 & 0.2 & 19.0 & 12.4 & 0.2 & 159 & 17 & 59 & 14 & + 8 & 5:38:45.81 & -69:05:08.7 & 246.3 & 0.13 & 0.06 & 0.93 & 0.22 & 1.4 & 1.1 & 0.2 & 15.1 & 10.7 & 0.2 & 85 & 11 & 116 & 57 & + 9 & 5:38:46.04 & -69:05:06.6 & 245.3 & 0.11 & 0.01 & 0.57 & 0.08 & 1.7 & 0.4 & 0.2 & 14.1 & 3.8 & 0.2 & 29 & 14 & 37 & 7 & + 10 & 5:38:46.32 & -69:05:07.7 & 245.6 & 0.11 & 0.04 & 0.92 & 0.08 & 2.2 & 1.4 & 0.2 & 16.3 & 10.6 & 0.1 & 107 & 13 & 97 & 33 & + 11 & 5:38:46.88 & -69:05:04.9 & 248.2 & 0.16 & 0.02 & 1.25 & 0.12 & 6.6 & 7.5 & 0.1 & 33.8 & 52.5 & 0.1 & 540 & 8 & 260 & 40 & pillar , nir yso + 12 & 5:38:46.58 & -69:05:04.1 & 245.6 & 0.19 & 0.04 & 0.59 & 0.23 & 3.2 & 1.3 & 0.2 & 20.2 & 10.3 & 0.2 & 90 & 13 & 68 & 30 & + 13 & 5:38:46.33 & -69:05:02.3 & 248.1 & 0.18 & 0.04 & 1.00 & 0.13 & 4.9 & 3.0 & 0.2 & 32.8 & 25.1 & 0.1 & 239 & 13 & 185 & 49 & + 14 & 5:38:46.17 & -69:05:01.7 & 253.4 & 0.11 & 0.04 & 0.75 & 0.23 & 3.8 & 1.9 & 0.2 & 18.9 & 14.2 & 0.2 & 162 & 18 & 63 & 29 & + 15 & 5:38:44.69 & -69:04:57.9 & 234.5 & 0.11 & 0.04 & 0.43 & 0.13 & 0.9 & 0.2 & 0.7 & 3.3 & 0.6 & 0.3 & 19 & 61 & 21 & 9 & + 16 & 5:38:44.52 & -69:04:57.7 & 240.0 & 0.11 & 0.07 & 1.59 & 0.30 & 0.8 & 0.4 & 0.9 & 4.2 & 1.5 & 0.2 & 34 & 72 & 300 & 202 & + 17 & 5:38:44.23 & -69:04:55.6 & 251.9 & 0.11 & 0.05 & 0.81 & 0.21 & 0.7 & 0.2 & 1.0 & 2.1 & 0.5 & 0.4 & 14 & 57 & 74 & 41 & + 18 & 5:38:45.45 & -69:04:56.3 & 242.5 & 0.15 & 0.04 & 0.65 & 0.06 & 5.0 & 2.7 & 0.1 & 22.5 & 14.7 & 0.1 & 194 & 10 & 65 & 17 & + 19 & 5:38:45.21 & -69:04:52.0 & 243.1 & 0.27 & 0.06 & 0.55 & 0.04 & 4.3 & 2.5 & 0.1 & 21.4 & 13.8 & 0.2 & 169 & 9 & 86 & 21 & + 20 & 5:38:45.80 & -69:04:50.5 & 243.8 & 0.22 & 0.06 & 0.62 & 0.08 & 2.0 & 1.2 & 0.2 & 12.8 & 5.6 & 0.2 & 61 & 11 & 88 & 26 & + 21 & 5:38:47.05 & -69:05:00.3 & 250.1 & 0.41 & 0.08 & 0.93 & 0.06 & 16.5 & 20.4 & 0.1 & 55.7 & 128.8 & 0.1 & 2000 & 12 & 375 & 73 & pillar , nir yso + 22 & 5:38:47.60 & -69:04:57.9 & 251.5 & 0.25 & 0.08 & 0.85 & 0.13 & 7.3 & 5.8 & 0.1 & 45.8 & 41.4 & 0.1 & 519 & 12 & 185 & 69 & + 23 & 5:38:47.09 & -69:04:56.7 & 251.2 & 0.11 & 0.06 & 1.04 & 0.04 & 5.1 & 3.9 & 0.1 & 32.3 & 33.8 & 0.1 & 288 & 9 & 123 & 65 & + 24 & 5:38:47.53 & -69:05:00.4 & 254.1 & 0.11 & 0.05 & 1.27 & 0.30 & 1.3 & 0.9 & 0.2 & 12.7 & 8.3 & 0.2 & 76 & 13 & 185 & 95 & + 25 & 5:38:47.76 & -69:05:02.0 & 243.7 & 0.11 & 0.05 & 0.85 & 0.04 & 2.7 & 1.5 & 0.2 & 19.6 & 11.2 & 0.1 & 106 & 13 & 82 & 39 & + 26 & 5:38:48.16 & -69:04:58.1 & 243.3 & 0.28 & 0.04 & 0.79 & 0.04 & 3.8 & 4.0 & 0.1 & 35.3 & 45.1 & 0.1 & 315 & 10 & 180 & 29 & + 27 & 5:38:46.47 & -69:04:56.6 & 247.5 & 0.11 & 0.03 & 1.27 & 0.25 & 1.9 & 1.4 & 0.3 & 12.2 & 6.8 & 0.3 & 70 & 15 & 185 & 69 & + 28 & 5:38:46.90 & -69:04:56.4 & 243.5 & 0.10 & 0.04 & 0.88 & 0.04 & 4.7 & 3.1 & 0.1 & 20.6 & 17.7 & 0.1 & 200 & 8 & 89 & 30 & + 29 & 5:38:48.36 & -69:05:00.6 & 254.9 & 0.14 & 0.08 & 2.35 & 0.85 & 0.7 & 0.5 & 1.1 & 6.0 & 2.9 & 0.3 & 41 & 98 & 810 & 525 & + 30 & 5:38:47.15 & -69:04:53.1 & 241.6 & 0.11 & 0.04 & 0.51 & 0.04 & 2.1 & 0.8 & 0.4 & 12.6 & 5.5 & 0.3 & 51 & 25 & 29 & 9 & + 31 & 5:38:47.65 & -69:04:54.6 & 248.8 & 0.17 & 0.08 & 1.27 & 0.30 & 2.6 & 1.8 & 0.1 & 24.5 & 18.2 & 0.1 & 131 & 10 & 280 & 146 & + 32 & 5:38:47.75 & -69:04:54.0 & 252.1 & 0.33 & 0.14 & 1.10 & 0.30 & 6.5 & 5.3 & 0.2 & 37.8 & 39.1 & 0.3 & 418 & 13 & 418 & 214 & + 33 & 5:38:48.49 & -69:04:53.8 & 245.8 & 0.36 & 0.16 & 1.10 & 0.21 & 5.2 & 7.9 & 0.1 & 25.9 & 65.7 & 0.1 & 610 & 8 & 457 & 222 & + 34 & 5:38:47.73 & -69:04:50.9 & 250.2 & 0.21 & 0.07 & 1.15 & 0.15 & 4.1 & 3.6 & 0.1 & 27.1 & 27.2 & 0.1 & 274 & 10 & 290 & 108 & + 35 & 5:38:48.35 & -69:04:47.8 & 251.0 & 0.25 & 0.14 & 1.10 & 0.21 & 6.5 & 8.7 & 0.1 & 32.6 & 51.4 & 0.1 & 588 & 6 & 315 & 182 & + 36 & 5:38:47.28 & -69:04:47.5 & 241.0 & 0.11 & 0.03 & 0.85 & 0.42 & 1.1 & 0.3 & 0.5 & 8.8 & 3.1 & 0.2 & 27 & 48 & 82 & 46 & + 37 & 5:38:46.38 & -69:04:48.7 & 241.1 & 0.19 & 0.06 & 1.21 & 0.08 & 1.2 & 1.1 & 0.2 & 7.7 & 5.2 & 0.4 & 61 & 14 & 286 & 91 & + 38 & 5:38:46.18 & -69:04:47.2 & 252.5 & 0.11 & 0.04 & 1.27 & 0.21 & 4.8 & 4.0 & 0.3 & 25.2 & 31.9 & 0.1 & 309 & 22 & 185 & 68 & + 39 & 5:38:46.71 & -69:04:45.0 & 254.5 & 0.28 & 0.09 & 1.21 & 0.06 & 12.7 & 18.0 & 0.3 & 43.2 & 118.1 & 0.1 & 1652 & 24 & 432 & 134 & h@xmath7o maser + 40 & 5:38:47.13 & -69:04:39.5 & 253.3 & 0.32 & 0.10 & 1.78 & 0.17 & 7.4 & 14.4 & 0.1 & 32.2 & 114.7 & 0.1 & 1168 & 7 & 1048 & 360 & + 41 & 5:38:46.49 & -69:04:40.2 & 252.4 & 0.26 & 0.06 & 0.98 & 0.11 & 2.4 & 2.4 & 0.2 & 22.9 & 31.0 & 0.1 & 152 & 12 & 259 & 69 & + 42 & 5:38:46.76 & -69:04:41.2 & 248.0 & 0.33 & 0.09 & 1.70 & 0.42 & 2.5 & 4.2 & 0.1 & 25.3 & 52.7 & 0.1 & 303 & 6 & 1003 & 373 & + 43 & 5:38:47.17 & -69:04:42.1 & 250.6 & 0.11 & 0.04 & 1.91 & 0.42 & 3.0 & 3.6 & 0.1 & 27.9 & 39.2 & 0.1 & 254 & 9 & 417 & 180 & + 44 & 5:38:44.95 & -69:04:38.9 & 251.5 & 0.38 & 0.08 & 1.10 & 0.08 & 10.4 & 16.3 & 0.1 & 48.6 & 97.0 & 0.1 & 1509 & 11 & 481 & 107 & + 45 & 5:38:45.39 & -69:04:40.6 & 250.0 & 0.22 & 0.06 & 1.06 & 0.30 & 8.4 & 7.3 & 0.2 & 42.6 & 63.1 & 0.1 & 646 & 17 & 259 & 103 & + 46 & 5:38:45.43 & -69:04:39.7 & 252.4 & 0.35 & 0.10 & 1.06 & 0.30 & 8.2 & 8.8 & 0.1 & 47.1 & 57.8 & 0.1 & 795 & 10 & 415 & 166 & + 47 & 5:38:45.85 & -69:04:38.3 & 253.3 & 0.13 & 0.07 & 1.02 & 0.17 & 5.6 & 2.4 & 0.2 & 35.5 & 18.5 & 0.2 & 153 & 11 & 144 & 84 & + 48 & 5:38:45.60 & -69:04:33.3 & 250.6 & 0.35 & 0.07 & 0.98 & 0.13 & 6.9 & 7.5 & 0.1 & 29.5 & 62.0 & 0.1 & 515 & 8 & 345 & 83 & + 49 & 5:38:47.78 & -69:04:41.4 & 257.7 & 0.24 & 0.06 & 1.49 & 0.17 & 4.5 & 3.2 & 0.1 & 30.6 & 26.0 & 0.1 & 265 & 11 & 546 & 156 & + 50 & 5:38:48.23 & -69:04:41.6 & 255.1 & 0.39 & 0.08 & 1.15 & 0.13 & 7.3 & 8.2 & 0.1 & 44.8 & 67.3 & 0.2 & 655 & 11 & 529 & 122 & + 51 & 5:38:48.74 & -69:04:42.6 & 253.4 & 0.40 & 0.07 & 1.15 & 0.17 & 6.1 & 9.0 & 0.1 & 45.0 & 71.2 & 0.3 & 709 & 10 & 543 & 129 & + 52 & 5:38:49.35 & -69:04:42.4 & 250.5 & 0.53 & 0.09 & 1.70 & 0.21 & 23.9 & 81.5 & 0.1 & 57.8 & 371.1 & 0.1 & 8528 & 8 & 1585 & 340 & _ spitzer _ myso , h@xmath7o maser + 53 & 5:38:49.41 & -69:04:46.8 & 249.7 & 0.39 & 0.10 & 1.70 & 0.21 & 11.6 & 21.6 & 0.1 & 51.9 & 136.0 & 0.1 & 1934 & 8 & 1165 & 345 & + 54 & 5:38:49.00 & -69:04:38.6 & 249.3 & 0.41 & 0.10 & 1.91 & 0.08 & 8.8 & 13.7 & 0.1 & 36.7 & 118.0 & 0.1 & 1102 & 6 & 1539 & 387 & + 55 & 5:38:48.95 & -69:04:47.9 & 253.1 & 0.36 & 0.10 & 1.59 & 0.08 & 7.3 & 9.8 & 0.1 & 34.2 & 61.3 & 0.2 & 654 & 7 & 936 & 268 & + 56 & 5:38:51.13 & -69:04:51.1 & 250.9 & 0.11 & 0.08 & 1.49 & 0.64 & 0.9 & 0.3 & 1.6 & 4.4 & 2.3 & 0.3 & 36 & 167 & 252 & 219 & + 57 & 5:38:50.14 & -69:04:45.7 & 245.3 & 0.11 & 0.03 & 0.89 & 0.08 & 1.9 & 1.3 & 0.4 & 11.8 & 4.5 & 0.6 & 56 & 17 & 90 & 29 & + 58 & 5:38:50.49 & -69:04:44.5 & 253.8 & 0.11 & 0.03 & 0.54 & 0.06 & 3.4 & 1.6 & 0.2 & 25.3 & 11.4 & 0.4 & 74 & 9 & 32 & 10 & + 59 & 5:38:47.43 & -69:04:34.8 & 253.6 & 0.39 & 0.08 & 1.91 & 0.17 & 3.2 & 7.3 & 0.1 & 25.2 & 64.8 & 0.1 & 472 & 6 & 1468 & 327 & + 60 & 5:38:48.13 & -69:04:31.2 & 249.1 & 0.30 & 0.15 & 2.76 & 0.85 & 1.7 & 3.5 & 0.1 & 19.5 & 72.6 & 0.1 & 248 & 5 & 2363 & 1421 & + 61 & 5:38:47.86 & -69:04:30.8 & 243.6 & 0.30 & 0.04 & 2.76 & 0.85 & 2.0 & 4.6 & 0.1 & 25.8 & 88.8 & 0.1 & 341 & 5 & 2385 & 813 & + 62 & 5:38:46.63 & -69:04:30.5 & 239.1 & 0.08 & 0.07 & 0.64 & 0.42 & 0.6 & 0.3 & 0.8 & 8.6 & 2.0 & 0.3 & 15 & 42 & 46 & 43 & + 63 & 5:38:47.48 & -69:04:29.6 & 239.3 & 0.24 & 0.04 & 1.91 & 0.21 & 2.5 & 4.5 & 0.1 & 22.5 & 46.3 & 0.1 & 229 & 4 & 926 & 192 & + 64 & 5:38:48.59 & -69:04:30.9 & 259.6 & 0.11 & 0.04 & 0.98 & 0.13 & 1.2 & 0.9 & 0.2 & 11.9 & 8.0 & 0.2 & 55 & 15 & 109 & 40 & + 65 & 5:38:52.23 & -69:04:50.8 & 255.4 & 0.33 & 0.11 & 2.59 & 0.17 & 1.1 & 1.4 & 0.1 & 3.7 & 10.5 & 0.2 & 222 & 15 & 2306 & 813 & + 66 & 5:38:54.85 & -69:05:02.2 & 262.2 & 0.23 & 0.09 & 0.66 & 0.06 & 2.1 & 1.6 & 0.6 & 15.2 & 4.6 & 0.2 & 55 & 22 & 104 & 41 & + 67 & 5:38:54.71 & -69:04:59.4 & 260.2 & 0.35 & 0.07 & 1.47 & 0.06 & 1.8 & 1.2 & 2.0 & 9.0 & 6.1 & 0.3 & 95 & 154 & 789 & 168 & + 68 & 5:38:54.29 & -69:04:54.2 & 260.2 & 0.11 & 0.05 & 0.65 & 0.04 & 5.6 & 3.4 & 0.1 & 22.6 & 16.8 & 0.1 & 249 & 10 & 48 & 24 & + 69 & 5:38:54.85 & -69:04:45.8 & 256.4 & 0.16 & 0.10 & 0.85 & 0.21 & 1.6 & 0.8 & 0.6 & 7.8 & 3.4 & 0.2 & 97 & 72 & 119 & 77 & + 70 & 5:38:53.88 & -69:04:42.4 & 263.9 & 0.16 & 0.15 & 0.55 & 0.13 & 1.2 & 0.7 & 0.2 & 8.4 & 4.7 & 0.3 & 68 & 15 & 50 & 48 & + 71 & 5:38:53.42 & -69:04:40.5 & 264.5 & 0.11 & 0.10 & 1.15 & 0.17 & 1.2 & 0.7 & 0.2 & 7.4 & 5.1 & 0.1 & 70 & 15 & 150 & 143 & + 72 & 5:38:52.94 & -69:04:37.4 & 254.1 & 0.49 & 0.12 & 2.89 & 0.08 & 15.3 & 86.1 & 0.1 & 43.4 & 409.9 & 0.0 & 7375 & 4 & 4262 & 1061 & _ spitzer _ myso , h@xmath7o maser + 73 & 5:38:53.72 & -69:04:35.6 & 256.3 & 0.65 & 0.13 & 1.94 & 0.08 & 12.6 & 48.9 & 0.1 & 37.0 & 227.1 & 0.1 & 3564 & 4 & 2525 & 517 & + 74 & 5:38:52.62 & -69:04:34.3 & 260.1 & 0.61 & 0.18 & 2.55 & 0.85 & 2.7 & 5.4 & 0.2 & 24.6 & 43.8 & 0.1 & 326 & 11 & 4123 & 1816 & + 75 & 5:38:53.80 & -69:04:32.5 & 254.7 & 0.07 & 0.04 & 3.18 & 0.85 & 4.6 & 5.0 & 0.4 & 28.0 & 46.5 & 0.2 & 273 & 20 & 1160 & 507 & + 76 & 5:38:51.85 & -69:04:39.5 & 257.3 & 0.11 & 0.05 & 0.62 & 0.08 & 1.7 & 0.6 & 0.2 & 20.4 & 10.8 & 0.2 & 47 & 17 & 43 & 19 & + 77 & 5:38:51.85 & -69:04:38.3 & 253.7 & 0.11 & 0.05 & 0.85 & 0.34 & 2.7 & 1.4 & 0.2 & 19.7 & 11.8 & 0.2 & 92 & 13 & 82 & 49 & + 78 & 5:38:51.55 & -69:04:38.9 & 250.1 & 0.27 & 0.06 & 1.19 & 0.17 & 7.4 & 8.1 & 0.1 & 30.3 & 52.1 & 0.1 & 675 & 11 & 395 & 110 & + 79 & 5:38:51.38 & -69:04:36.6 & 253.1 & 0.47 & 0.10 & 0.89 & 0.04 & 7.8 & 12.3 & 0.1 & 31.6 & 73.7 & 0.1 & 1047 & 12 & 392 & 83 & + 80 & 5:38:50.82 & -69:04:34.7 & 257.9 & 0.37 & 0.08 & 1.17 & 0.08 & 2.4 & 3.3 & 0.2 & 16.6 & 29.5 & 0.1 & 224 & 10 & 517 & 117 & + 81 & 5:38:50.08 & -69:04:33.4 & 259.5 & 0.34 & 0.07 & 1.74 & 0.06 & 2.7 & 6.0 & 0.1 & 24.4 & 55.9 & 0.1 & 323 & 4 & 1080 & 224 & + 82 & 5:38:49.34 & -69:04:27.8 & 246.1 & 0.11 & 0.04 & 2.63 & 0.30 & 4.8 & 5.9 & 0.1 & 28.9 & 55.3 & 0.1 & 415 & 8 & 792 & 282 & + 83 & 5:38:49.66 & -69:04:25.8 & 243.8 & 0.11 & 0.05 & 2.89 & 0.30 & 1.8 & 0.8 & 0.2 & 21.1 & 10.8 & 0.1 & 56 & 11 & 953 & 404 & + 84 & 5:38:48.88 & -69:04:26.5 & 243.3 & 0.11 & 0.10 & 1.40 & 0.08 & 1.1 & 0.9 & 0.6 & 15.4 & 18.9 & 0.3 & 53 & 36 & 224 & 208 & + 85 & 5:38:49.91 & -69:04:27.2 & 248.1 & 0.44 & 0.11 & 1.26 & 0.04 & 8.1 & 17.5 & 0.1 & 36.1 & 132.7 & 0.1 & 1391 & 7 & 723 & 191 & + 86 & 5:38:49.43 & -69:04:23.2 & 247.7 & 0.35 & 0.10 & 1.36 & 0.13 & 1.4 & 1.5 & 0.2 & 14.8 & 26.4 & 0.2 & 98 & 15 & 678 & 196 & + 87 & 5:38:50.26 & -69:04:22.5 & 243.7 & 0.28 & 0.11 & 2.25 & 0.21 & 5.7 & 10.9 & 0.2 & 34.1 & 104.3 & 0.1 & 907 & 18 & 1469 & 589 & + 88 & 5:38:50.69 & -69:04:22.5 & 245.0 & 0.11 & 0.09 & 2.32 & 0.21 & 6.0 & 6.9 & 0.1 & 32.0 & 64.0 & 0.1 & 577 & 10 & 617 & 496 & + 89 & 5:38:50.93 & -69:04:23.7 & 247.9 & 0.28 & 0.11 & 1.70 & 0.17 & 7.8 & 12.4 & 0.1 & 31.4 & 93.6 & 0.1 & 1041 & 10 & 831 & 343 & + 90 & 5:38:50.61 & -69:04:20.2 & 247.2 & 0.14 & 0.09 & 2.87 & 0.21 & 2.4 & 1.5 & 0.1 & 22.6 & 22.1 & 0.2 & 106 & 10 & 1178 & 806 & + 91 & 5:38:50.81 & -69:04:19.1 & 249.0 & 0.11 & 0.05 & 1.87 & 0.13 & 2.3 & 1.4 & 0.2 & 22.5 & 17.6 & 0.1 & 106 & 12 & 399 & 174 & + 92 & 5:38:50.95 & -69:04:18.7 & 245.3 & 0.11 & 0.05 & 0.55 & 0.13 & 1.5 & 0.6 & 0.2 & 21.3 & 9.5 & 0.2 & 44 & 16 & 34 & 16 & + 93 & 5:38:51.60 & -69:04:21.6 & 245.6 & 0.11 & 0.03 & 1.23 & 0.13 & 2.0 & 1.4 & 0.1 & 24.5 & 26.8 & 0.2 & 107 & 10 & 173 & 55 & + 94 & 5:38:51.89 & -69:04:16.9 & 246.8 & 0.17 & 0.06 & 1.70 & 0.06 & 3.4 & 4.0 & 0.4 & 25.4 & 27.7 & 0.2 & 154 & 14 & 509 & 181 & + 95 & 5:38:51.62 & -69:04:12.1 & 247.3 & 0.11 & 0.07 & 0.79 & 0.06 & 5.6 & 2.3 & 0.2 & 22.0 & 6.9 & 0.2 & 150 & 14 & 72 & 48 & + 96 & 5:38:49.76 & -69:04:17.0 & 240.9 & 0.11 & 0.10 & 1.06 & 0.21 & 1.2 & 0.5 & 0.8 & 10.8 & 4.5 & 0.2 & 23 & 38 & 128 & 116 & + 97 & 5:38:48.92 & -69:04:17.6 & 240.9 & 0.11 & 0.07 & 2.12 & 0.25 & 0.9 & 0.8 & 0.5 & 13.5 & 16.1 & 0.2 & 48 & 31 & 515 & 341 & + 98 & 5:38:48.17 & -69:04:14.5 & 248.7 & 0.52 & 0.12 & 1.08 & 0.06 & 4.5 & 6.6 & 0.2 & 23.7 & 34.0 & 0.1 & 364 & 10 & 628 & 147 & near cluster p4 + 99 & 5:38:47.28 & -69:04:11.7 & 248.9 & 0.32 & 0.11 & 0.68 & 0.11 & 7.9 & 7.4 & 0.1 & 24.7 & 28.8 & 0.1 & 509 & 10 & 152 & 59 & + 100 & 5:38:47.21 & -69:04:11.5 & 251.6 & 0.11 & 0.04 & 0.64 & 0.13 & 2.2 & 0.9 & 0.7 & 14.9 & 4.0 & 0.8 & 41 & 30 & 46 & 18 & + 101 & 5:38:48.92 & -69:04:08.0 & 251.2 & 0.11 & 0.04 & 0.76 & 0.04 & 4.3 & 2.3 & 0.2 & 32.0 & 16.8 & 0.2 & 133 & 11 & 66 & 21 & + 102 & 5:38:47.35 & -69:04:10.1 & 241.6 & 0.11 & 0.04 & 0.56 & 0.00 & 1.8 & 0.7 & 1.0 & 9.5 & 1.7 & 0.2 & 44 & 68 & 36 & 11 & + 103 & 5:38:47.76 & -69:04:13.7 & 253.0 & 0.11 & 0.03 & 0.76 & 0.04 & 1.8 & 0.8 & 0.4 & 18.5 & 8.8 & 0.4 & 41 & 22 & 66 & 19 & + observations of molecular clouds on @xmath3610pc scales follow a power - law scaling relation between the size @xmath87 and velocity dispersion @xmath88 .
the index @xmath89 was 0.38 in the original measurement @xcite , and subsequently measured between 1/3 and 1/2 with high scatter .
the relation can be obtained from incompressible turbulence ( @xmath89=1/3 ) , compressible shock dominated turbulence ( @xmath89=1/2 ) , quasi - static gravitational equilibrium ( @xmath89=1/2 ) , or free - fall collapse ( @xmath89=1/2 ) ( e.g. * ? ? ?
* ) . in this section
we analyze the 30dor-10 clump distribution as a function of @xmath87 , @xmath84 , and surface density @xmath90 .
these scaling relations are clearest when combining multiple datasets at different spatial resolutions and using tracers sensitive to different critical densities ; our approach is to compare the clumps in 30dor-10 to the trends derived from such multi - dataset analyses , and to directly compare clumps in our data to clumps observed with similar tracers and physical resolution in the milky way .
the size - linewidth relation for @xmath4co clumps in 30dor-10 is compared to other locations in figure [ sizeline ] .
milky way clouds measured using @xmath2co and @xmath4co@xmath31 - 0 are taken from heyer et al .
2009 , who revisited the seminal study of ( * ? ? ?
* hereafter srby ) .
the plotted relation @xmath84=0.72r@xmath91 is the fit to the data reported by srby . at somewhat smaller spatial scales ,
we show structures in the orion molecular cloud that we obtained using the same cloud decomposition as in 30dor-10 , applied to the @xmath2co@xmath31 - 0 data of @xcite .
both sets of orion clouds are consistent with the srby milky way @xmath84-@xmath92 relation .
massive clumps in orion are also shown , from ( * ? ? ?
* @xmath4co@xmath31 - 0 ) .
the data in their table were converted from velocity fwhm to dispersion , and half - width - half - max radii 1.18@xmath93 to the radius convention used here of 1.91@xmath93 .
finally , clouds in the central molecular zone ( cmz ) of the milky way , measusing @xmath2co@xmath31 - 0 by @xcite are plotted .
clumps in 30dor-10 , and clouds in the cmz , have high linewidths compared to `` ordinary '' molecular cloud structure . for completeness
, we verify that the differences are not due to tracer : in the cmz , cloud properties measured in dense gas tracers ( n@xmath7h@xmath60 , * ? ? ?
* ) fall in the same part of parameter space as those measured by @xcite using @xmath2co@xmath31 - 0 . in 30dor-10 , @xmath2co@xmath32 - 1 radii are [email protected] times the measured @xmath4co@xmath32 - 1 radii , and linewidths are [email protected] times higher neither shift affects the result shown here .
one might also consider the thermal component of the line width parsec and smaller clump linewidths are usually analyzed with the thermal width removed , to test the hypothesis of hierarchical turbulent ( nonthermal ) structure .
however , even at the elevated temperatures in 30dor-10 ( t@xmath9440k , [ ltemass ] ) , @xmath95 is only 0.4@xmath3km@xmath3s@xmath12 , and a minor part of the linewidth .
higher velocity dispersions than predicted by larson s relation can result if clouds are in virial equilibrium , but with external pressure ( e.g. * ? ? ? * ) , in which case the virial equation is @xmath96 the pressure can be high inter - cloud thermal pressure , as invoked by @xcite to explain the milky way central molecular zone . however , the physical scales being measured in 30 doradus are significantly smaller . a larger dispersion and a deviation of @xmath5310pc clumps above the size - linewidth relation has been noted in many galactic clouds with different tracers ( e.g. * ? ? ? * ; * ? ? ? * ) .
this deviation can be explained either by clumps in quasi - equilibrium with interclump pressure , i.e. the `` weight '' of the entire cloud acting on clumps within it @xcite , or equally well by clouds in chaotic gravitational collapse , in which gravity drives turbulent motions and the gravitational and kinetic energies remain balanced @xcite .
it appears that the clumps in this extreme environment do not have dramatically different properties from clumps of similar size in milky way molecular clouds , or at least occupy the same part of parameter space with similar dispersion .
the dynamic state of clumps and clouds can be further assessed by directly comparing the gravitational and kinetic energies , or the virial masses m@xmath97=5@xmath98/g and masses determined from co ( m@xmath99 ) ; in 30dor-10 these are in rough agreement ( figure [ mvirial ] ) . as discussed above ,
the dominant uncertainty in the molecular mass calculated from co is the @xmath4co abundance . calculating the excitation temperature t@xmath39 from @xmath2co emission , assuming that @xmath2co and @xmath4co have the same t@xmath39 , and assuming that the @xmath4co level population is described by a boltzmann distribution at t@xmath39 introduce another factor of @xmath02 uncertainty ( [ ltemass ] ) .
the @xmath4co@xmath32 - 1 sensitivity can be used to calculate the expected mass sensitivity , although it is sensitive to the extitation temperature ( equation [ nco ] ) . for a low @xmath2co@xmath32 - 1 brightness temperature of 5k ,
the excitation temperature is 9.7k ( eqn .
[ tex ] ) . a clump with 5@xmath26 peak @xmath4co@xmath32 - 1 brightness of 0.75k , linewidth of 0.5@xmath3km@xmath3s@xmath12 ( equal to the velocity resolution ) , and a physical size equal to the minimum deconvolvable size at that peak brightness ( see above and * ? ? ?
* ) , has a mass of 20@xmath3m@xmath10 . a similarly bright clump the size of the beam has mass of 65@xmath3m@xmath10 , consistent with the lowest points in subsequent figures .
the dominant uncertainties in the virial mass are the assignment of a `` radius '' and use of the spherical virial theorem to a non - spherical clump , and the assumption of virial equilibrium in the first place virial masses should be regarded as having at least a factor of 2 uncertainty .
furthermore , in 30dor-10 , the selective dissociation of co relative to h@xmath7 may make radii measured from co emission , and derived virial masses , _ systematically _ too small .
there is a trend for m@xmath97/m@xmath99 to decrease with increasing m@xmath99 .
the formal best - fit line is marked , m@xmath97/m@xmath99 = 35@xmath3m@xmath100m@xmath101 .
@xcite derive m@xmath97/m@xmath102m@xmath103 for pressure - confined and magnetized clumps .
the same decreasing trend m@xmath97/m@xmath104m@xmath105 is seen in dense clumps in many galactic clouds , albeit often with a shallower slope : orion , @xmath89=-0.33 @xcite , cygnus @xcite , cepheus and casseopeia , @xmath89=-0.38 @xcite .
if the mass ratios in 30dor-10 are due to pressure confinement , the pressures required are high p / k@xmath3610@xmath64 . the gravitational pressure from the weight of the entire molecular cloud ( inter - clump h@xmath7 and h ) can be estimated as p@xmath106/k @xmath171.5@xmath3@xmath107k@xmath3(m@xmath108)@xmath109(r / pc)@xmath110 @xcite . for 30dor-10 p@xmath106/k = 1.5(9@xmath110@xmath49)@xmath82(8)@xmath110 = 3@xmath110@xmath64 , with large uncertainty due to the radius estimate , so the concept is not unreasonable .
analysis of the clump volume densities and inter - clump gas physical conditions using _ herschel _ spectroscopy of cii , oi , oiii , and nii emission will help reveal the degree to which these clumps are pressure - confined ( chevance et al .
2014 , in prep ) .
heyer et al .
2009 discuss the size - linewidth relation in detail and find that the milky way relation is not merely a power - law but that linewidth also depends on the cloud surface density .
figure [ sizelinesig ] is a reproduction of their figure 7 , showing clouds and clumps in size - linewidth - surface density parameter space .
the line marks the locus of gravitational stability or virial equilibrium for spheres , namely @xmath111 .
clouds following the larson s scaling relations as well as having a universal surface density would occupy a single point on this plot .
clumps in 30dor-10 overlap in parameter space with milky way clouds , but extend to higher surface densities for the same linewidth - size parameter .
one explanation for this population that extends to higher surface density are that these clumps are not supported by ( thermal or nonthermal ) motions against gravitational collapse . the most extreme ( high surface density )
clumps include those half - dozen that have clearly associated protostars in near and mid - infrared images .
however , many high - surface density clumps also do not obviously contain star formation in the infrared .
it is possible that these clumps are supported by magnetic fields , perhaps enhanced by the material having been swept up by the massive stellar winds and radiation .
dense sub - parsec - sized clumps in infrared dark clouds in the milky way display the same extension to larger surface densities , as shown by the overplotted data on clumps in galactic ir - dark clouds @xcite .
comparison of the masses of clumps calculated from dust ( section [ continuum ] ) , using @xmath2co@xmath32 - 1 and @xmath4co@xmath32 - 1 ( section [ ltemass ] ) , and from the virial theorem allows one to determine the relative abundances of co , dust , and h@xmath7 , and infer physical conditions in the clumps .
co - derived and virial masses were compared in the previous section ; figure [ dustmass ] compares these to the dust - derived masses , showing that the dust masses are @xmath02@xmath1 lower than the virial or molecular masses .
this suggests that either the gas - to - dust ratio in 30dor-10 is twice the lmc average , or that the dust temperatures are half the excitation temperatures measured from @xmath2co .
note that both the gas - to - dust ratio and the co abundance h@xmath7/@xmath4co could be increased or decreased by the same factor without changing the agreement between dust and co - derived masses , but a change by more than a factor @xmath02 would then require explaining a difference between dust - derived and virial mass .
we do not find what @xcite found in the smc , virial masses exceeding dust - derived masses and requiring additional cloud support .
the conversion `` x '' factor from co emission to h@xmath7 mass or column density is used widely to measure molecular masses in galaxies at all redshifts .
conversion of optically thick emission ( as most low - j transitions of @xmath2co are ) into mass requires that the molecular clouds have relatively universal structure in space and velocity . even assuming that this is true , the conversion factor is expected to depend on metallicity because h@xmath7 is more effective at self - shielding itself from photodissociation than co. at reduced metallicity and consequently reduced dust abundance
, ultraviolet radiation penetrates more deeply into molecular clouds , and should result in more of this selective photodissociation , and a larger ratio of @xmath112 mass to co emission . the milky way x - factor has been calculated with virial masses , dust emission , gamma - ray emission , extinction , and optically thin isotopologue emission ; @xcite recommend [email protected]@xmath3m@xmath100pc@xmath113(k@xmath3km@xmath3s@xmath12)@xmath12 or [email protected]@xmath110@xmath114@xmath115(k@xmath3km@xmath3s@xmath12)@xmath12 from the compilation .
analysis of the @xmath040pc resolution nanten @xmath2co@xmath31 - 0 survey of the lmc suggested that the x - factor was higher than in the milky way : 19@xmath3m@xmath100pc@xmath116(k@xmath3km@xmath3s@xmath12)@xmath12 @xcite , and this was corroborated by studies of selected regions with sest at @xmath015pc resolution ( 28@xmath714@xmath3m@xmath100pc@xmath116(k@xmath3km@xmath3s@xmath12)@xmath12 * ? ? ? * ) .
however , more recent analysis of the largely complete magma survey ( @xmath015pc resolution ) finds no strong evidence for a higher x - factor in the lmc than in the milky way @xcite .
it is reasonable to expect the x - factor to be different in regions of especially high radiation intensity , such as 30 doradus .
@xcite investigated this with the same data as used by @xcite , but targeting specifically the @xmath02kpc molecular cloud complex extending south from 30 doradus .
they found no variation of x - factor along that molecular ridge .
we compare the calculated molecular mass ( [ ltemass ] ) to the @xmath2co@xmath32 - 1 intensity and calculate m@xmath99/f(@xmath2co@xmath32 - 1 ) = 12@xmath714@xmath3m@xmath100pc@xmath116(k@xmath3km@xmath3s@xmath12)@xmath12 ( fig .
[ xfactor ] ) . applying the same ratio f(@xmath2co@xmath31 - 0)/f(@xmath2co@xmath32 - 1)=0.7 as @xcite yields an x - factor of 8.4@xmath713@xmath3m@xmath100pc@xmath116(@xmath2co@xmath31 - 0@xmath3k@xmath3km@xmath3s@xmath12)@xmath12 _ for clumps _ in 30dor-10 .
this is remarkably almost exactly the milky way value divided by the lmc metallicity .
there is a possible trend of lower x - factor at higher mass , namely m@xmath99/f(@xmath2co@xmath32 - 1 ) = 12(m/1300@xmath3m@xmath10)@xmath117 .
this could be explained by the more massive clumps having a higher beam filling fraction .
alternately , as discussed in [ ltemass ] , the excitation temperature derived from @xmath2co may be higher than the actual @xmath4co excitation temperature in bright clumps , and our calculated mass too high for those clumps .
figure [ xfactor ] only considers co emission associated with dense clumps , but we can also perform the calculation for the entire map : the total calculated h@xmath7 mass _ traced by co _ in 30dor-10 is [email protected]@xmath110@xmath79m@xmath10 , and the total @xmath2co@xmath32 - 1 flux 9400@xmath3k@xmath18km@xmath18s@xmath118pc@xmath82 , which translates into an overall m@xmath99/f(@xmath2co@xmath32 - 1 ) 9.5@xmath3m@xmath100pc@xmath116(k@xmath3km@xmath3s@xmath12)@xmath12 consistent with the factor derived from clumps and with the fact that most @xmath2co@xmath32 - 1 emission is associated with clumps .
this method of calculating a mass using an optically thin isotopologue only measures h@xmath7 associated with co , and not `` co - dark '' h@xmath7 .
the x - factor including any h@xmath7 that may exist between and not associated with clumps is expected to be higher , but we can not measure that directly .
one could in principle calculate the total gas mass from the total dust emission , and derive a total x - factor from that , but unfortunately we do not have the single - dish data to correct for large - scale dust emission resolved out by the interferometer . the most direct comparison to our data would be to an x - factor derived for resolved molecular clouds using the optically thin isotopologue method ; indeed @xcite perform this calculation for the milky way ( albeit with poorer physical resolution than our data in 30dor-10 ) , and derive approximately a factor of two lower x - factor ; @xcite perform a similar calculation with different cloud definitions and find a value consistent with [email protected]@xmath3m@xmath100pc@xmath113(k@xmath3km@xmath3s@xmath12)@xmath12 .
we conclude that the x - factor in dense clumps in 30dor-10 is consistent with merely being scaled by metallicity , may be up to a factor of two higher depending on the milky way comparison value chosen , and that a measurement of any inter - clump `` co - dark '' h@xmath7 is likely to yield a significantly higher value .
we conclude our discussion of clump masses with the clump mass distribution ( fig .
[ mhist ] ) .
the clump decomposition tends to select regions similar to the beamsize , and such mass histograms are fraught with caveats @xcite .
nevertheless , the clumps in 30dor-10 exhibit similar behaviour to those in galactic clumps on similar scales ( e.g. orion ; * ? ? ?
* ) . the best - fit slopes
dn / d(log m ) @xmath102 m@xmath119 for m@xmath25500m@xmath10 are @[email protected] and [email protected] for molecular and virial masses , respectively .
galactic clump distributions have slopes @xmath17 - 1 ( -1.1 in orion ) . as noted above , m@xmath120/m@xmath99 decreases with increasing mass , causing the slope of the virial mass distribution to be steeper .
although higher resolution observations will break up the clumps observed here , the slope of the mass distribution is not dramatically sensitive to that , as shown by varying - resolution simulations in @xcite .
if the intense radiation field from r136 is affecting the molecular cloud , one might expect to see trends of clump properties with increasing distance from r136 , going deeper into the cloud ( clump projected distances from r136 are @xmath01330pc . ) figure [ trends ] shows that this is not obviously the case , except for a tendency for the clumps to be more widely separated deeper in the cloud
i.e. the volume fraction of the cloud that is in dense clumps is somewhat less deeper into the cloud , perhaps due to a different spatial spectrum of turbulence or less fragmentation .
the uncertainty in separation plotted here is the difference between the nearest neighbor distance , and 1/@xmath121 times the distance to the second nearest neighbor .
there is also a weak trend for the size of the resolved clumps to increase with distance away from r136 , although there is no lack of unresolved clumps further back in the cloud as well .
the lack of strong trends in clump properties is consistent with the overall conclusion that dense clumps in this molecular cloud are not significantly different from dense clumps in other environments , but that the larger - scale cloud physics may be affected . in this case , the cloud may be compressed overall closer to the pdr front , bringing the dense clumps closer together , but once dense clumps form , their properties do not apparently vary strongly with depth into the cloud .
if this result holds generally , it would support a `` normal '' stellar initial mass function in super - star clusters : in 30 doradus the super - star cluster is already formed , and current star formation will not amount to another super - star cluster , but we do have evidence that lower - mass stars and small clusters form `` normally '' in dense clumps even in the vicinity of many massive stars .
it is not well - understood whether low - mass stars form before or after massive ones in ssc formation , but our data supports the robustness of the process .
it is also important to remember the 3-dimensional nature of the cloud .
the wispy blue nir emission ( fig . [ nir_ysos ] ) coincident with some of the co features suggests that radiation from r136 may reach structures that are in projection towards the top of our mapped area , but in fact are in front or behind the bulk of molecular material .
if there are trends of clump properties with distance into the cloud , the 3d geometry likely disguises them .
the giant molecular cloud 30dor-10 , north of r136 in 30 doradus , was observed with alma during early science cycle 0 and with apex .
this cloud is strongly irradiated and has lower than solar metal abundances , so is an ideal laboratory in which to test whether either of metallicity or incident radiation field can affect the physical conditions of molecular gas .
the primary resulting data are image cubes of @xmath2co@xmath32 - 1 , @xmath4co@xmath32 - 1 , c@xmath5o@xmath32 - 1 , and an image of the 1.3@xmath3 mm continuum .
analysis of these data leads to the conclusions about clumpy molecular cloud structure listed below . 1
. co emission in 30 doradus is highly structured the molecular cloud is composed of dense clumps and filaments , very similar to galactic molecular clouds .
the relative contribution of lower column density inter - clump regions to the total @xmath2co flux is however smaller than in galactic clouds like w3 .
this suggests that there is significant photodissociation of co by radiation penetrating between the dense clumps .
for the first time in an extreme environment like 30 doradus , we observe directly the decrease in @xmath2co emission relative to total gas mass that manifests as an increased x - factor in more distant unresolved molecular clouds .
we decompose these data into discrete structures or clumps , to analyze trends in physical properties .
this procedure is sensitive to only a modest range of physical scales , and higher resolution observations will reveal substructure , but it is instructive to compare to milky way clouds decomposed at similar physical resolution .
dense clumps measured for the first time on sub - parsec scales outside of our galaxy have properties not unlike those in galactic giant molecular clouds .
the clumps have somewhat larger velocity dispersions than predicted by the size - linewidth relation for milky way clouds .
one interpretation is that the clumps in 30dor-10 are pressure - confined ; the ratio of virial mass ( m@xmath97 ) to mass calculated from co emission ( m@xmath99 ) decreases with increasing m@xmath99 as predicted for pressure - confined clumps .
however , such a decrease is seen in many galactic clouds on small scales , and the interpretation is not unique .
selective photodissiciation of co relative to h@xmath7 may make the measured sizes of clumps measured from co emission , and calculated virial masses , too small .
3 . in size - linewidth - surface density space ,
clumps in 30dor-10 overlap with the locus populated by @xmath2510pc galactic clouds but extend to higher surface densities at the same size and linewidth .
this trend agrees with sub - parsec clumps in galactic infrared - dark clouds ( irdcs ) .
masses calculated from dust m@xmath122 are approximately half of those calculated from co. this could be explained by a gas - to - dust ratio half of the lmc average , or could be a result of assuming that the dust temperature equals the co excitation temperature calculated from the @xmath2co@xmath32 - 1 brightness temperature .
@xmath2co@xmath32 - 1 is quite optically thick , thus its excitation temperature in the emitting clump surface may be warmer than the temperature of most of the dust ; assuming too high of a dust temperature would result in low dust masses .
however , the dust masses are also a factor of two lower than virial masses , which suggests that the reduced dust - to - gas ratio may be the more likely explanation . one can increase or decrease both the assumed gas - to - dust and h@xmath7/@xmath4co ratios and not change the ratio of calculated dust mass to mass calculated from co , but changing either of those abundances dramatically would result in inconsistency with the virial masses .
we suggest that the the co abundance _ in dense clumps _ in 30 doradus is similar to the outer milky way h@xmath7/@xmath4co=5@xmath110@xmath64 , and the gas - to - dust ratio is @xmath0600 .
average abundances for the entire molecular cloud including inter - clump regions , may be different .
use of optically thick @xmath2co emission to trace mass _ within molecular clouds _ is not generally advised , but we calculate that for dense clumps @xmath6 = m@xmath99/f(@xmath2co@xmath32 - 1 ) = 12@xmath714@xmath3m@xmath100pc@xmath116(k@xmath3km@xmath3s@xmath12)@xmath12 .
the data suggest a decrease in @xmath6 in brighter , more massive clumps .
the measured x - factor is 2 times the average value in the milky way and other solar - metallicity galaxies , again merely the scaling expected for the lmc metallicity . since most of the @xmath2co emission is associated with dense clumps , the ratio of total cloud mass to @xmath2co emission may be higher when inter - clump h@xmath7 is included . 6 .
the high end of the dense clump mass function has a power - law slope of [email protected] , similar to sub - parsec sized clumps in galactic massive star formation regions
. clump physical properties are not a strong function of distance from the dominant source of radiation , the r136 cluster core . in particular , we measure no trend in the x - factor or clump mass with distance from r136 . there is a weak trend for the clumps to be larger and further from each other as one travels back into the molecular cloud away from r136 . in other words
the fraction of the molecular cloud filled with dense clumps is larger closer to the pdr edge where radiative feedback is strongest .
alma is beginning to dissect star formation regions and star - forming molecular gas on previously inaccessible size scales .
for the first time we can study molecular cloud physics resolving structures smaller than 1pc outside the milky way . studying physics like this in a wide range of environments promises to revolutionize our understanding of star - forming molecular clouds .
andrews , b. h. , & thompson , t. a. 2011 , , 727 , 97 ballesteros - paredes , j. , hartmann , l. w. , vzquez - semadeni , e. , heitsch , f. , & zamora - avils , m. a. 2011 , , 411 , 65 bally , j. , & langer , w. d. 1982 , , 255 , 143 bertoldi , f. 1989 , , 346 , 735 bertoldi , f. , & mckee , c. f. 1992 , , 395 , 140 bieging , j. h. , & peters , w. l. 2011 , , 196 , 18 bloemen , j. b. g. m. 1989 , the physics and chemistry of interstellar molecular clouds - mm and sub - mm observations in astrophysics , 331 , 22 bolatto , a. d. , leroy , a. k. , rosolowsky , e. , walter , f. , & blitz , l. 2008 , , 686 , 948 bolatto , a. d. , wolfire , m. , & leroy , a. k. 2013 , arxiv:1301.3498 bot , c. , boulanger , f. , rubio , m. , & rantakyro , f. 2007 , , 471 , 103 brand , j. , & wouterloot , j. g. a. 1995 , , 303 , 851 brandner , w. , grebel , e. k. , barb , r. h. , walborn , n. r. , & moneti , a. 2001 , , 122 , 858 brogan , c. l. , goss , w. m. , lazendic , j. s. , & green , a. j. 2004 , , 128 , 700 brogan , c. , johnson , k. , & darling , j. 2010 , , 716 , l51 caselli , p. , & myers , p. c. 1995 , , 446 , 665 churchwell , e. 1990 , , 2 , 79 chu , y .- h . , gruendl , r. a. , chen , c .- h .
r. , lazendic , j. s. , & dickel , j. r. 2004 , , 615 , 727 cohen , r. s. , dame , t. m. , garay , g. , et al .
1988 , , 331 , l95 condon , j. j. 1997 , , 109 , 166 crowther , p. a. , & dessart , l. 1998 , , 296 , 622 crowther , p. a. , dessart , l. , hillier , d. j. , abbott , j. b. , & fullerton , a. w. 2002 , , 392 , 653 dale , j. e. , clark , p. c. , & bonnell , i. a. 2007 , , 377 , 535 dame , t. m. , hartmann , d. , & thaddeus , p. 2001
, , 547 , 792 darling , j. , brogan , c. , & johnson , k. 2008 , , 685 , l39 de marchi , g. , paresce , f. , panagia , n. , et al . 2011 , , 739 , 27 dickey , j. m. , duesterberg , c. , mebold , u. , et al .
1996 , , 28 , 900 dobashi , k. , bernard , j .-
p . , & fukui , y. 1996 , , 466 , 282 draine , b. t. , & lee , h. m. 1984 , , 285 , 89 draine , b. t. , & li , a. 2007 , , 657 , 810 ediss , g. a. , carter , m. , cheng , j. , et al .
2004 , fifteenth international symposium on space terahertz technology , 181 fall , s. m. , & zhang , q. 2001 , , 561 , 751 field , g. b. , blackman , e. g. , & keto , e. r. 2011 , , 416 , 710 fukui , y. , mizuno , n. , yamaguchi , r. , et al . 1999 , ,
51 , 745 fukui , y. , kawamura , a. , minamidani , t. , et al .
2008 , , 178 , 56 galliano , f. , hony , s. , bernard , j .-
, et al .
2011 , , 536 , a88 gibson , d. , plume , r. , bergin , e. , ragan , s. , & evans , n. 2009 , , 705 , 123 gong , h. , & ostriker , e. c. 2011 , , 729 , 120 gruendl , r. a. , & chu , y .- h .
2009 , , 184 , 172 gsten , r. , nyman , l. . ,
schilke , p. , et al .
2006 , , 454 , l13 heikkil , a. , johansson , l. e. b. , & olofsson , h. 1999 , , 344 , 817 heitsch , f. , naab , t. , & walch , s. 2011 , , 415 , 271 hennebelle , p. , & falgarone , e. 2012 , , 20 , 55 hester , j. j. , scowen , p. a. , sankrit , r. , et al .
1996 , , 111 , 2349 heyer , m. h. , carpenter , j. m. , & snell , r. l. 2001 , , 551 , 852 heyer , m. , krawczyk , c. , duval , j. , & jackson , j. m. 2009 , , 699 , 1092 hills , r. e. , kurz , r. j. , & peck , a. b. 2010 , , 7733 , 34 .
hopkins , p. f. , quataert , e. , & murray , n. 2011 , , 417 , 950 huggins , p. j. , gillespie , a. r. , phillips , t. g. , gardner , f. , & knowles , s. 1975 , , 173 , 69p hughes , a. , wong , t. , ott , j. , et al .
2010 , , 406 , 2065 hyland , a. r. , straw , s. , jones , t. j. , & gatley , i. 1992 , , 257 , 391 ikeda , n. , kitamura , y. , & sunada , k. 2009 , , 691 , 1560 imai , h. , katayama , y. , ellingsen , s. p. , & hagiwara , y. 2013 , arxiv:1302.5160 indebetouw , r. , de messires , g. e. , madden , s. , et al .
2009 , , 694 , 84 israel , f. p. 1997
, , 328 , 471 israel , f. p. , johansson , l. e. b. , rubio , m. , et al .
2003 , , 406 , 817 johansson , l. e. b. , greve , a. , booth , r. s. , et al .
1998 , , 331 , 857 johnson , k. e. , hunt , l. k. , & reines , a. e. 2009 , , 137 , 3788 kaufman , m. j. , wolfire , m. g. , hollenbach , d. j. , & luhman , m. l. 1999 , , 527 , 795 kerr , a. r. , pan , s .- k . , lauria , e. f. , et al . 2004 , fifteenth international symposium on space terahertz technology , 55 koda , j. , scoville , n. , hasegawa , t. , et al .
2012 , , 761 , 41 koeppen , j. , & kegel , w. h. 1980 , , 42 , 59 krumholz , m. r. , mckee , c. f. , & tumlinson , j. 2009 , , 699 , 850 larson , r. b. 1981 , , 194 , 809 lazendic , j. s. , whiteoak , j. b. , klamer , i. , harbison , p. d. , & kuiper , t. b. h. 2002 , , 331 , 969 lazendic , j. s. , dickel , j. r. , & jones , p. a. 2003 , , 596 , 287 lebouteiller , v. , bernard - salas , j. , brandl , b. , et al .
2008 , , 680 , 398 liszt , h. s. , & lucas , r. 1998 , , 339 , 561 mckee , c. f. , & ostriker , e. c. 2007 , , 45 , 565 meixner , m. , gordon , k. d. , indebetouw , r. , et al .
2006 , , 132 , 2268 meixner , m. , galliano , f. , hony , s. , et al .
2010 , , 518 , l71 milam , s. n. , savage , c. , brewster , m. a. , ziurys , l. m. , & wyckoff , s. 2005 , , 634 , 1126 millar , t. j. , & herbst , e. 1990 , , 242 , 92 minamidani , t. , mizuno , n. , mizuno , y. , et al .
2008 , , 175 , 485 nikoli , s. , garay , g. , rubio , m. , & johansson , l. e. b. 2007 , , 471 , 561 oka , t. , hasegawa , t. , sato , f. , et al .
2001 , , 562 , 348 oliveira , j. m. , van loon , j. t. , stanimirovi , s. , & zijlstra , a. a. 2006 , , 372 , 1509 ostriker , e. c. , mckee , c. f. , & leroy , a. k. 2010 , , 721 , 975 padoan , p. , juvela , m. , bally , j. , & nordlund , .
2000 , , 529 , 259 pagel , b. e. j. 2003 , cno in the universe , 304 , 187 parker , j. w. 1993 , , 106 , 560 pineda , j. l. , ott , j. , klein , u. , et al .
2009 , , 703 , 736 pineda , j. e. , rosolowsky , e. w. , & goodman , a. a. 2009 , , 699 , l134 pineda , j. l. , mizuno , n. , rllig , m. , et al .
2012 , , 544 , a84 poglitsch , a. , krabbe , a. , madden , s. c. , et al .
1995 , , 454 , 293 roberts , d. a. , & yusef - zadeh , f. 2005 , , 129 , 805 rolleston , w. r. j. , trundle , c. , & dufton , p. l. 2002 , , 396 , 53 roman - duval , j. , jackson , j. m. , heyer , m. , rathborne , j. , & simon , r. 2010 , , 723 , 492 rosolowsky , e. , & leroy , a. 2006 , , 118 , 590 sandford , m. t. , ii , whitaker , r. w. , & klein , r. i. 1982 , , 260 , 183 shetty , r. , beaumont , c. n. , burton , m. g. , kelly , b. c. , & klessen , r. s. 2012 , , 425 , 720 solomon , p. m. , rivolo , a. r. , barrett , j. , & yahil , a. 1987 , , 319 , 730 reid , m. a. , wadsley , j. , petitclerc , n. , & sills , a. 2010 , , 719 , 561 rohlfs , k. , & wilson , t. l. 2004 , tools of radio astronomy , 4th rev . and enl . ed .
, by k. rohlfs and t.l .
wilson .
berlin : springer , 2004 rllig , m. , & ossenkopf , v. 2013 , , 550 , a56 rubio , m. , barb , r. h. , walborn , n. r. , et al .
1998 , , 116 , 1708 sakamoto , s. , hasegawa , t. , handa , t. , hayashi , m. , & oka , t. 1997 , , 486 , 276 sandstrom , k. m. , leroy , a. k. , walter , f. , et al .
2012 , arxiv:1212.1208 sawada , t. , hasegawa , t. , handa , t. , et al .
2001 , , 136 , 189 selman , f. , melnick , j. , bosch , g. , & terlevich , r. 1999 , , 341 , 98 skrutskie , m. f. , cutri , r. m. , stiening , r. , et al .
2006 , , 131 , 1163 stutzki , j. , & guesten , r. 1990 , , 356 , 513 tielens , a. g. g. m. , & hollenbach , d. 1985 , , 291 , 722 townsley , l. k. , broos , p. s. , feigelson , e. d. , et al .
2006 , , 131 , 2140 van der tak , f. f. s. , black , j. h. , schier , f. l. , jansen , d. j. , & van dishoeck , e. f. 2007 , , 468 , 627 van dishoeck , e. f. , & black , j. h. 1987 , nato asic proc . 210 : physical processes in interstellar clouds , 241 van dishoeck , e. f. , & black , j. h. 1988 , , 334 , 771 vzquez - semadeni , e. , gmez , g. c. , jappsen , a. k. , et al . 2007 , , 657 , 870 visser , r. , van dishoeck , e. f. , & black , j. h. 2009 , , 503 , 323 walborn , n. r. , & blades , j. c. 1987 , , 323 , l65 walborn , n. r. , barb , r. h. , brandner , w. , et al . 1999 , , 117 , 225 walborn , n. r. , maz - apellniz , j. , & barb , r. h. 2002 , , 124 , 1601 walborn , n. r. , barb , r. h. , & sewio , m. m. 2013 , , 145 , 98 walker , a. , 2011 , , 746 , whiteoak , j. b. , wellington , k. j. , jauncey , d. l. , et al .
1983 , , 205 , 275 whitney , b. a. , sewilo , m. , indebetouw , r. , et al . 2008 , , 136 , 18 williams , j. p. , de geus , e. j. , & blitz , l. 1994 , , 428 , 693 wilson , t. l. , & rood , r. 1994 , , 32 , 191 wolfire , m. g. , hollenbach , d. , & mckee , c. f. 2010 , , 716 , 1191 wong , t. , hughes , a. , ott , j. , et al .
2011 , , 197 , 16 yonekura , y. , dobashi , k. , mizuno , a. , ogawa , h. , & fukui , y. 1997 , , 110 , 21
|
we present alma observations of 30 doradus the highest resolution view of molecular gas in an extragalactic star formation region to date ( @[email protected] ) .
the 30dor-10 cloud north of r136 was mapped in @xmath2co@xmath32 - 1 , @xmath4co@xmath32 - 1 , c@xmath5o@xmath32 - 1 , 1.3@xmath3 mm continuum , the h30@xmath6 recombination line , and two h@xmath7co@xmath33 - 2 transitions .
most @xmath2co emission is associated with small filaments and clumps ( @xmath81pc , @xmath010@xmath9m@xmath10 at the current resolution ) .
some clumps are associated with protostars , including `` pillars of creation '' photoablated by intense radiation from r136 .
emission from molecular clouds is often analyzed by decomposition into approximately beam - sized clumps .
such clumps in 30 doradus follow similar trends in size , linewidth , and surface density to milky way clumps .
the 30 doradus clumps have somewhat larger linewidths for a given size than predicted by larson s scaling relation , consistent with pressure confinement .
they extend to higher surface density at a given size and linewidth compared to clouds studied at 10pc resolution .
these trends are also true of clumps in galactic infrared - dark clouds ; higher resolution observations of both environments are required .
consistency of clump masses calculated from dust continuum , co , and the virial theorem reveals that the co abundance in 30 doradus clumps is not significantly different from the lmc mean , but the dust abundance may be reduced by @xmath02 .
there are no strong trends in clump properties with distance from r136 ; dense clumps are not strongly affected by the external radiation field , but there is a modest trend towards lower dense clump filling fraction deeper in the cloud .
|
charmonium spectroscopy is of great interest to both the experimentalists and the theorists since its first discovery more than 30 years ago . with the successful running of the two @xmath14-factories at slac and kek , charmonium physics
was also revitalized as more and more charmonium and charmonium - like states were observed in @xmath14 decays , in initial state radiation ( @xmath15 ) processes , in double - charmonium productions , and in two - photon processes .
the observation of the @xmath16 @xcite , @xmath8 @xcite , @xmath17 @xcite and so on may suggest the existence of new type of hadronic states besides the conventional mesons ( @xmath18 ) and baryons ( @xmath19 ) in quark model . among these many newly observed states ,
those with @xmath20 are becoming very puzzling since there are too many of them between 4 - 5 gev/@xmath4 than expected from the potential models @xcite .
in addition to the excited @xmath21 states observed in the inclusive hadronic cross section @xcite ( the @xmath22 , @xmath23 , @xmath24 ) , there are four new structures observed in @xmath15 processes ( the @xmath25 @xcite , @xmath8 @xcite , @xmath1 @xcite , @xmath2 @xcite ) .
the overpopulation of the vector states in this mass range may suggest at least one of them are non - conventional charmonium state @xcite . in the measurement of the @xmath9
cross section via @xmath15 at babar @xcite , a broad resonance like structure was observed with mass @xmath26 mev/@xmath4 and width @xmath27 mev/@xmath4 ; while with more luminosity , the belle experiment found it is indeed due to two narrow resonances with masses of 4361 mev/@xmath4 and 4664 mev/@xmath4 @xcite .
a close examination of the babar observation shows that there is one bin with high cross section at around the second resonance observed in belle experiment although it is not very significant . in this paper
, we try to perform a combined fit to @xmath0 cross sections measured by the babar and belle experiments with two resonances , to obtain a better estimation of the resonant parameters of the @xmath1 and @xmath2 .
we also study the possible production of the @xmath8 in @xmath9 mode .
both babar and belle experiments reported cross sections of @xmath0 for center - of - mass energy ranges from threshold to 5.5 gev .
the integrated luminosity of the babar data sample is @xmath28 while that of the belle data sample is @xmath29 , with @xmath30 of the data were collected at the @xmath31 resonance ( @xmath32 gev ) , while the rest were taken off the @xmath31 peak .
figure [ fig1 ] shows the data , good agreement between babar and belle results is observed , and the two structures are evident .
the cross sections of @xmath0 measured at babar ( stars with error bars ) and belle ( dots with error bars ) .
( a ) shows all the data , while ( b ) shows the region with the @xmath13 resonances.,width=340 ]
the breit - wigner form of a single resonance used in this analysis is the same as in refs .
@xcite , i.e. , @xmath33 where @xmath34 is the mass of the resonance , @xmath35 and @xmath36 are the total width and partial width to @xmath37 respectively , @xmath38 is the branching fraction of @xmath39 decays into final state @xmath40 , and @xmath41 is the three - body decay phase space factor . in fitting to the data , we assume all the cross sections are due to the @xmath1 and @xmath2 resonances , and they are added coherently , that is , @xmath42 where @xmath43 and @xmath44 represent the two resonances and @xmath45 is the relative phase between them .
we fit the data using a binned maximum likelihood method with minuit in the cern program library @xcite . to take the poisson distribution of the small number of events in each @xmath9 mass bin into consideration , we start from the observed number of events in each mass bin instead of from the measured cross section . for each @xmath9 mass bin , the probability of observing @xmath46 events when the expected number of events is @xmath47 is @xmath48 . here
@xmath49 where @xmath50 is the mean cross section in the @xmath51-th bin ( with bin width @xmath52 ) , which is calculated as @xmath53 the number of background events in each bin , @xmath54 for a 25 mev/@xmath4 mass bin for belle , and @xmath55 is neglected for babar .
@xmath56 for each bin is obtained through @xmath57 from the @xmath9 invariant mass distributions and cross sections presented in refs .
@xcite .
the likelihood is defined as the product of @xmath58 over all the @xmath9 mass bins , that is , @xmath59 . in reality
, @xmath60 is minimized to get the best estimation of the parameters .
we fit the belle data on @xmath9 between threshold and 5.5 gev ( in 60 bins ) and the babar data in the same energy range ( in 30 bins ) simultaneously .
figure [ fig2 ] shows the fit results .
two equally good solutions are found with the two amplitudes interfere with each other differently .
the masses and the widths of the two resonances are identical but the partial widths to @xmath37 and relative phases are different in these two solutions , as shown in table [ table1 ] .
the statistical significance of the @xmath2 is calculated by comparing the likelihood of the fit with and without it , and we found it is @xmath61 in the combined fit , while that quoted in belle experiment is @xmath62 @xcite .
the results of the binned maximum likelihood fit to @xmath63 data from belle and babar .
the curves show the best fit with two coherent breit - wigners and the contribution from each component .
the interference between the two amplitudes is not shown .
the two dashed curves at each peak show the two solutions ( see text).,width=340 ] .[table1 ] fit results to the combined babar and belle data on @xmath63 .
the errors are statistical only .
@xmath34 , @xmath35 , and @xmath64 are the mass ( in mev/@xmath65 ) , total width ( in mev/@xmath65 ) , and product of the branching fraction to @xmath66 and the @xmath37 partial width ( in ev/@xmath65 ) , respectively .
@xmath45 is the relative phase ( in degrees ) . [
cols="^,^,^",options="header " , ] the other sources of the systematic errors on the mass and width measurements are not very different from those listed by the belle experiment @xcite , since it is mainly due to the parametrization of the resonance . taking directly the error from belle and adding in quadrature with the uncertainty due to the binning , one gets the total systematic errors of 9 mev/@xmath4 , 11 mev/@xmath4 , 6 mev/@xmath4 , and 6 mev/@xmath4 , for @xmath67 , @xmath68 , @xmath69 , and @xmath70 , respectively .
the @xmath8 was observed in @xmath71 mode @xcite , and its production in @xmath9 is not forbidden by any selection rule , except that the phase space is a bit small due to large @xmath72 mass .
we add the @xmath8 amplitude in eq .
[ eq2 ] coherently with mass ( 4247 mev/@xmath4 ) and width ( 108 mev/@xmath4 ) fixed to the belle measurement @xcite to measure its production rate . figure [ fig_y4260 ] shows the fit result with three resonances , the change of @xmath73 is 5.2 for two more free parameters , this corresponds to a statistical significance of the @xmath8 of @xmath74 .
that is , the production of the @xmath8 in @xmath9 is not significant .
there are two pairs of solutions for @xmath11 , one pair interferes with the @xmath1 constructively , with @xmath75 ev/@xmath4 , and we set the upper limit as 4.3 ev/@xmath4 at the 90% confidence level ; the other pair interferes with the @xmath1 destructively , and @xmath76 ev/@xmath4 . the results of the fit to @xmath63 data from belle and babar .
the solid curve show the best fit with three coherent breit - wigners : the @xmath8 , @xmath1 , and @xmath2 , and the dashed curve is the signal shape of the @xmath8.,width=340 ]
based on combined belle and babar data on @xmath0 , we perform a maximum likelihood fit to get the resonant parameters of the @xmath1 and @xmath2 .
we obtain @xmath77 , @xmath78 , @xmath79 , and @xmath80 .
these results give the best measurement of the @xmath1 and @xmath2 .
there is a faint evidence of @xmath81 , with @xmath82 ev/@xmath4 at the 90% confidence level , or @xmath83
ev/@xmath4 , depending on the interference between the @xmath8 and @xmath1 .
these numbers should be compared with the couplings of the @xmath8 to @xmath71 in ref .
@xcite and to @xmath84 in ref .
@xcite .
we thank prof .
s. olsen for suggesting extend the study to the @xmath8 .
this work is supported in part by the 100 talents program of cas under contract no .
u-25 and by national natural science foundation of china under contract no .
10491303 .
s. k. choi _ et al . _ [ belle collaboration ] , phys .
lett . * 91 * , 262001 ( 2003 ) [ arxiv : hep - ex/0309032 ] .
b. aubert _ et al . _ [ babar collaboration ] , phys . rev
. lett . * 95 * , 142001 ( 2005 ) [ arxiv : hep - ex/0506081 ] . c. z. yuan _ et al .
_ [ belle collaboration ] , phys .
* 99 * , 182004 ( 2007 ) [ arxiv:0707.2541 [ hep - ex ] ] .
s. k. choi _ et al . _
[ belle collaboration ] , phys .
* 100 * , 142001 ( 2008 ) [ arxiv:0708.1790 [ hep - ex ] ] .
e. eichten , k. gottfried , t. kinoshita , k. d. lane and t. m. yan , phys .
d * 21 * , 203 ( 1980 ) .
s. godfrey and n. isgur , phys .
d * 32 * , 189 ( 1985 ) .
m. ablikim _ et al . _
[ bes collaboration ] , phys .
b * 660 * , 315 ( 2008 ) [ arxiv:0705.4500 [ hep - ex ] ] .
w. m. yao _ et al .
_ [ particle data group ] , j. phys .
g * 33 * , 1 ( 2006 ) .
x. l. wang _ et al . _ [ belle collaboration ] , phys .
* 99 * , 142002 ( 2007 ) [ arxiv:0707.3699 [ hep - ex ] ] .
b. aubert _ et al . _ [ babar collaboration ] , phys .
lett . * 98 * , 212001 ( 2007 ) [ arxiv : hep - ex/0610057 ] . for recent reviews ,
see for example , e. s. swanson , phys .
* 429 * , 243 ( 2006 ) ; s. godfrey and s. l. olsen , arxiv:0801.3867 [ hep - ph ] .
f. james , minuit , cern program library writeup d506 .
|
a combined fit is performed to the babar and belle measurements of the @xmath0 cross sections for center - of - mass energy between threshold and 5.5 gev .
the resonant parameters of the @xmath1 and @xmath2 are determined .
the mass is @xmath3 mev/@xmath4 and the width is @xmath5
mev/@xmath4 for the @xmath1 , and the mass is @xmath6 mev/@xmath4 and the width is @xmath7
mev/@xmath4 for the @xmath2 .
the production of the @xmath8 in @xmath9 mode is found to be at @xmath10 level , and @xmath11 is found to be less than 4.3 ev/@xmath4 at the 90% confidence level , or equal to @xmath12 ev/@xmath4 depending on it interferes with the @xmath1 constructively or destructively .
these information will shed light on the understanding of the nature of the @xmath13 states observed in initial state radiation processes .
|
the search for magnetism in graphene - based materials is an attractive research field promising for spintronics applications @xcite . according to theoretical predictions , @xmath2-electron magnetic semiconductors might also have much higher curie temperatures than the conventional ones @xcite .
activities in this direction stimulate synthesis and magnetic measurements of graphene with vacancies and different types of adsorbates @xcite . despite considerable efforts ,
the available experimental data on magnetism in graphene - based systems is still limited . from the theoretical perspective numerous first - principles studies@xcite
allow to provide a microscopic picture on the electronic structure of magnetic graphene in its ground state .
the treatment of excited states constitutes the next important step in the description of the systems in question .
this requires the construction and solution of the model electronic or spin hamiltonians , which have been only marginally addressed in the literature @xcite .
considerable nonlocal coulomb correlations typical to graphene@xcite and spin - orbit coupling@xcite complicate the consideration significantly .
finally , practically important aspects such as the role of temperature and external fields in the evolution of magnetic states also remain unclear .
single - side semifluorinated graphene ( c@xmath1f ) is of special interest because such a system was recently realized in the experiment.@xcite it opens a way for verification and correction of the theoretical models for this material , whose magnetic properties represent a complex interplay between different physical mechanisms . according to the lieb s theorem@xcite formulated for bipartite lattice , c@xmath1f should be ferromagnetic since the fluorine atoms adsorb at the same sublattice of carbon atoms . such model ,
however , does not take into account considerable modification of the electronic structure upon fluorination .
more reliable description of the magnetism in c@xmath1f has been recently proposed using first - principles dft calculations @xcite of the isotropic exchange interactions , predicting a frustrated ground state in c@xmath1f .
such approach yet ignores many - body and relativistic effects , which might be crucial in the formation of magnetism . in this paper
we perform a systematic theoretical characterization of the c@xmath1f and c@xmath1h systems by constructing the low - energy models with spin - orbit coupling . in both cases
the spin orbitals described by magnetic wannier functions form a triangular lattice with short - range nearest - neighbour interactions in the case of c@xmath1f and long - range interactions in the case of c@xmath1h .
the estimation of the on - site and inter - site coulomb interactions suggest significant spatial charge correlations in both systems that are an order of magnitude larger than those in transition - metal compounds .
remarkably , we find a strong ferromagnetic _
direct _ interaction between the neighbouring spin orbitals . due to a delicate balance between the direct ferromagnetic exchange interaction and kinetic anderson s superexchange , our hartree - fock simulations
show that c@xmath1f can demonstrate both the 120@xmath3 nel and ferromagnetic states having comparable energies . at the same time
a robust ferromagnetic solution within the mean - field approximation is found for c@xmath1h , contrary to previous predictions @xcite .
the analysis of the anisotropic superexchange couplings reveals a strong dzyaloshinskii - moriya interaction ( dmi ) of about 1 mev between the nearest spins in c@xmath1f . according to the monte carlo simulations
, dmi can lead to a skyrmion state at finite temperatures and magnetic fields . f and c@xmath4h calculated by using lda+so method and their comparison with the one - band model .
( b ) band splitting @xmath5 due to spin - orbit coupling calculated for the one - band model.,scaledwidth=47.0% ]
the electronic structure calculations were performed within the plane - wave pseudopotential method as implemented in the quantum - espresso simulation package.@xcite exchange and correlation effects were taken into account using the local density approximation ( lda).@xcite spin - orbit ( so ) coupling was included on the basis of fully relativistic pseudopotentials .
we employed an energy cutoff of 50 ry for the plane - wave basis and 400 ry for the charge density , as well as a ( 64@xmath664 ) * k*-point mesh .
the surface layers were separated by a vacuum region of 40 and fully relaxed . to construct the hamiltonian in the ( spinor ) wannier function basis we used the maximally localized wannier function procedure@xcite as implemented in the wannier90 package.@xcite the calculated lda+so band structures and densities of states are presented in fig .
[ lda](a ) .
one can see that in the case of c@xmath1h there are two well - separated bands at the fermi level [ with a small band splitting due to spin - orbit coupling , fig .
[ lda](b ) ] .
these bands demonstrate the maxima between @xmath7 and @xmath8 high - symmetry points . as we will show below , such a band behavior results in the long - range hopping integrals .
however , it is not the case for the c@xmath1f system , where the bands at the fermi level slightly overlap with other bands at the @xmath9 point .
nevertheless , construction of the minimal low - energy model for the relevant bands at the fermi level is confirmed by comparison with the many - orbital tight - binding model described in ref . .
_ wannier functions .
_ to parametrize the lda+so spectra we have constructed the maximally localized wannier functions for the bands located close to the fermi level .
they are visualized in fig .
[ ris : wannier_functions ] . in agreement with the results of ref .
the wannier functions are centered at the non - bonded carbon atoms .
they are strongly delocalized in real space , that should be taken into account when analyzing the experimental data of the magnetic measurements.@xcite the corresponding band splitting due spin - orbit coupling for the one - band model is shown in fig .
[ lda](b ) .
as it is seen , it is an order of magnitude larger in the case of c@xmath1f.@xcite
to describe electronic and magnetic properties of c@xmath4h and c@xmath4f we use the following tight - binding hamiltonian taking into account spin - orbit coupling : @xmath10 where @xmath11(@xmath12 ) are the creation ( annihilation ) operators .
@xmath13 , @xmath14 and @xmath15 are local coulomb , non - local coulomb and non - local ( @xmath16 ) exchange interactions , respectively .
@xmath17 is the element of the spin - resolved hopping matrix .
the model hamiltonian given by eq .
( [ ham ] ) can be solved by static ( hartree - fock)@xcite or dynamical ( dmft)@xcite mean - field numerical techniques to reproduce experimentally observable spectra of electronic and magnetic excitations .
this model is not only widely used in the simulations of physical properties of 3@xmath18 and 5@xmath18 metal compounds , but can also be applied to the systems demonstrating @xmath2-type magnetism.@xcite the hopping matrix @xmath17 was determined using a wannier - parametrization of the first - principles lda+so hamiltonian .
the obtained hopping parameters are presented in table [ tab : hoppings ] and schematically visualized in fig .
[ lattice ] .
one can see that in the case of the c@xmath1f system the nearest - neighbor hopping ( @xmath19 ) provides the dominant contribution , that results in the realization of an almost ideal triangular geometry for hopping integrals . for c@xmath1h ,
the hopping matrix is less trivial with the largest parameters corresponding to the second- and third - nearest - neighbor interactions .
such a difference in hopping integrals of the semifluorinated and semihydrogenated graphene is related to the various inter - atomic distances .
the equilibrium carbon - carbon and carbon - adatom distances @xcite in c@xmath1h are smaller by 4% and 20% , respectively , than those in c@xmath1f .
importantly , there are non - diagonal and imaginary contributions to the hopping integrals between nearest neighbors .
as one can see , such contributions are an order of magnitude larger in c@xmath1f compared to c@xmath1h .
they originate from the spin - orbit coupling and , as we will show below , responsible for a non - zero dzyaloshinskii - moriya interaction . [ cols="^,^,^",options="header " , ] _ dzyaloshinskii - moriya interaction .
_ anisotropic exchange parameters can be derived by extending the theory of superexchange interaction in the case of spin - orbit coupling .
they have the following form:@xcite @xmath20,\ ] ] where @xmath21 are the pauli matrices .
such an approach gives reliable results for the low - dimensional copper oxides.@xcite for the nearest neighbour bonds in c@xmath1f with the radius vectors @xmath22 = ( 1 , 0 , 0 ) , @xmath23 = ( @xmath24 , -@xmath25,0 ) and @xmath26 = ( @xmath24 , @xmath25,0 ) we obtain @xmath27 = ( 0 , -0.93 , -0.28 ) , @xmath28 = ( -0.81 , -0.47 , 0.28 ) and @xmath29 = ( 0.81 , -0.47 , 0.28 ) mev , respectively .
in contrast to c@xmath1f , the magnitude of the dzyaloshoinskii - moriya interaction in the semihydrogenated graphene is much smaller , @xmath27 = ( 0 , -0.017 , 0 ) , @xmath28 = ( -0.015 , -0.008 , 0 ) and @xmath29 = ( 0.015 , -0.008 , 0 ) mev .
higher orders in spin - orbit coupling , such as the symmetric anisotropic exchange interaction @xcite , @xmath30 are small . their magnitude is about 10@xmath31 mev ( c@xmath4f ) and 10@xmath32 mev ( c@xmath1h ) , and thus the dzyaloshinskii - moirya interaction is the main source of the magnetic anisotropy in semifluorinated and semihydrogenated graphene
. symmetry of the spin hamiltonian is consistent with the @xmath33 point group of the c@xmath4f and c@xmath4h systems ( given by the vertical reflection planes going through the nearest - neighbor functionalized and non - functionalized carbon atoms and by @xmath34 rotations around adatom - carbon bonds ) .
however , the resulting wannier functions reside on the non - functionalized carbon sites and form a triangular lattice without reflection of the original lattice ( fig .
[ ris : wannier_functions ] ) . according to moriya
s rules@xcite , since the reflection planes pass through the middle of the bonds between two wannier functions , the corresponding anisotropic exchange parameters lie in the reflection planes and are perpendicular to their bonds .
their directions are given arbitrarily with respect to the mirror planes , and as a result @xmath35 components of the anisotropic exchange parameters can alternate within the coordination sphere under @xmath34 rotations .
thus , the resulting symmetry of the effective model is @xmath34 .
the electronic hamiltonians constructed in the wannier function basis also give us an opportunity to compare the spin - orbit coupling strength in the c@xmath1h and c@xmath4f systems .
the previous first - principles studies @xcite have demonstrated an enhancement of spin - orbit coupling in c@xmath1f compared to c@xmath1h . here
, we confirm this finding by calculating the hopping integrals with spin - orbit coupling and estimating the dzyaloshinskii - moriya interactions .
f. light and dark red arrows denote the dzyaloshinskii - moriya vectors with positive and negative @xmath35 components , respectively .
the corresponding directions of the bonds are denoted with black arrows . ]
we would like to stress that there is a conceptual difference in the origin of magnetic anisotropy in two - dimensional ( 2d ) materials with @xmath2 electrons that we consider and 2d materials with localized @xmath18 electrons .
@xcite while the magnetic anisotropy in @xmath36 , @xmath37 and @xmath38 systems originates from the spin - orbit coupling of individual metallic atoms , it is not the case in 2d materials with @xmath2 electrons . as it was shown in our study
there is a strong delocalization of the magnetic moments in c@xmath1f and c@xmath1h materials .
thus , the magnetic anisotropy in these @xmath2 materials is a collective multi - atomic effect , which was also demonstrated by authors of ref . in their analysis of spin - orbit coupling in the c@xmath1f and c@xmath1h systems . in this situation
the construction of the simple and transparent models for a system in question is a non - trivial task . by using the formalism of the wannier functions we propose an elegant solution of the problem for c@xmath1h and c@xmath1f materials .
the resulting electronic models for c@xmath1h and c@xmath1f are one - band hamiltonians with spin - orbit coupling .
_ estimation of the @xmath39-factor_. to characterize orbital magnetism induced by fluorine atoms we have estimated the value of the @xmath39-factor .
moriya @xcite used the gyromagnetic ratio that can be found from magnetic experiments to estimate a magnitude of the intersite anisotropic exchange interaction . in our case
we are to solve an inverse problem .
having calculated the dzyaloshinskii - moriya interaction we can estimate the @xmath39-factor value for future magnetic experiments on c@xmath1f .
for that the quantity of interest is @xmath40 which is proportional to @xmath41 , where @xmath42 is the deviation of the @xmath39-factor from the value for a free electron . for the semifluorinated graphene
the estimated value of @xmath39 is about 2.025 . as we will show below this information is important for estimating critical magnetic fields at which a skyrmion state is formed .
to solve the electronic models given by eq .
( [ ham ] ) at zero temperature , we have employed the mean - field hartree - fock approximation : @xmath43 where @xmath44 and @xmath45 are the fourier transforms of the hopping parameters and hartree - fock potential , respectively . @xmath46 and @xmath47 are the corresponding eigenvalues and eigenvectors in the wannier function basis ; eq .
( [ hf ] ) is solved self - consistently with respect to the density matrix : @xmath48 and the resulting magnetic state is further defined as @xmath49 . to find more detail on the computation scheme ,
we refer the reader to refs . .
nel ( right ) solutions obtained within the hartree - fock approximation . ]
the results obtained for the semihydrogenated graphene with unit cells of different size do not reveal any signature of the spiral spin ordering , and the ferromagnetic state is stabilized [ fig . [ magn_config](left ) ] .
the energy of the system does not depend on the direction of the total magnetization . from the analysis of the hopping integrals
we conclude that spin - orbit coupling in c@xmath1h is weak and does not produce magnetic anisotropy . since the role of magnetic anisotropy is decisive in the formation of a long - range magnetic order in 2d materials @xcite , no long - range ordering is expected in c@xmath1h at finite temperatures according to the mermin - wagner theorem @xcite . in the case of the semifluorinated graphene
the situation is different .
depending on the choice of the direct exchange interaction @xmath50 , one obtains either the 120@xmath3 nel for @xmath51 40 mev [ fig .
[ magn_config](right ) ] or ferromagnetic states for @xmath52 40 mev [ fig .
[ magn_config](left ) ] .
there is a solution with zero isotropic magnetic interactions at @xmath50 = 40 mev , when the ferromagnetic direct exchange interaction exactly compensates the antiferromagnetic anderson s superexchange .
thus we conclude that c@xmath1f can be considered to be at the threshold of the antiferromagnetic - ferromagnetic instability .
our investigation of the semifluorinated and semihydrogenated graphene revealed that these systems are physical realizations of different spin models on the triangular lattice .
such a lattice is of special interest due to the effects of magnetic frustration and possibility to form topologically protected spin textures , skyrmions at finite temperatures and magnetic fields .
for instance , it was recently shown that skyrmionic states can be stabilized in the @xmath53 model with the ferromagnetic nearest neighbor and antiferromagnetic next - nearest neighbour exchange interactions.@xcite a similar model is derived in our study for c@xmath1h .
however , the ratio @xmath54 is much larger than that proposed in ref . to stabilize a skyrmion state .
another important example known from the literature is the antiferromagnetic ( or ferromagnetic ) triangular lattice with dzyaloshinskii - moriya interactions that favour formation of the antiferromagnetic skyrmion lattice state@xcite ( or a skyrmion lattice state with the nel - type domain wall alignment @xcite ) .
a similar scenario can be realized in the c@xmath1f system . to check whether it is possible to stabilize a distinct spin texture in the semifluorinated graphene we have performed classical monte carlo simulations based on the single - spin metropolis update scheme and the heat bath method combined with overrelaxation for the spin models obtained with different values of @xmath50 . in our calculations supercells of various size from @xmath55 to @xmath56 spins with periodic boundary conditions were used , and a single run contained @xmath57 monte - carlo steps .
while different states can be identified from a real space spin configuration , we have also computed the static spin structure factors : @xmath58 and @xmath59 where @xmath60 means the monte - carlo averaged configuration , as well as the so - called local chirality @xmath61 that is regarded as an order parameter of the corresponding magnetic state . in the case of the c@xmath1f spin models obtained with @xmath51 40 mev ,
the antiferromagnetic skyrmion lattice state is destroyed by a weak @xmath35 component of the dzyaloshinskii - moriya interaction , instead the 120@xmath3 nel state is observed .
the situation is different for the c@xmath1f spin models with @xmath52 40 mev . in this case
a nel type skyrmion state can be realized .
an example shown in fig .
[ montecarloski](a ) was obtained with the exchange interactions @xmath62 = 1.9 mev and @xmath27 = ( 0 , -0.93 , -0.28 ) at the temperature @xmath63= 0.02 and the magnetic field @xmath64 = 0.1 .
fig.[montecarloski](b ) gives the intensity of the spin structure factor for the obtained texture .
there is a superposition of three spirals with @xmath65 pairs of the wave vectors , which is a clear indication of the skyrmion lattice state .
taking into account the estimated value of the @xmath39-factor @xmath39=2.025 , the critical magnetic field needed to stabilize the skyrmion lattice can be defined as 1.62 t. for this set of parameters the size of the individual skyrmion can be estimated to be about 25 .
generally , it is controlled by the ratio @xmath66 and the magnetic field .
our theoretical results indicate that the recent experimental realization of c@xmath1f @xcite opens a way for exploiting truly two - dimensional one - band model materials demonstrating a rich variety of physical properties , such as the strong dzyaloshinskii - moriya interaction , spatial charge correlations , magnetic frustration , skyrmion state and others .
specifically , we found that the wannier functions describing magnetic moments in the c@xmath1f and c@xmath1h systems form a triangular lattice with completely different hopping paths .
the overlap of the neighbouring wannier orbitals results in a strong direct ferromagnetic exchange interaction for both systems in question that can partially or fully compensate the kinetic superexchange interaction between nearest magnetic orbitals . for the semihydrogenated graphene
we predict a short - range ferromagnetic order .
since our hartree - fock calculations do not reveal any magnetic anisotropy in c@xmath1h , the long - range magnetic order in this system is unstable as follows from the mermin - wagner theorem @xcite for two - dimensional systems .
the calculations of the anisotropic exchange interactions suggest that c@xmath1f can be considered as a physical realization of the heisenberg model with dmi on the triangular lattice .
the variation of @xmath15 from the fully screened to bare limits leads to either the 120@xmath3 nel or skyrmion states as it was predicted by our hartree - fock and monte carlo simulations .
the key parameter that is responsible for a variety of magnetic states in c@xmath1f and c@xmath1h is the direct exchange interaction , @xmath67 between the neighbouring wannier functions .
experimentally , one can find different mechanisms to control and tune this interaction .
for instance , in this paper we demonstrate that @xmath67 is sensitive to the screening by the background .
the latter can be changed by coupling with a substrate . @xcite
another attractive control parameter is strain that strongly affects non - local coulomb interactions in graphene systems as it was shown in ref . .
we acknowledge fruitful communications with igor solovyev .
the hospitality of the institute of theoretical physics of hamburg university and radboud university of nijmegen is gratefully acknowledged .
the work is supported by the ministry of education and science of the russian federation , project no .
16.1751.2014/k and the grant of the president of russian federation md-6458.2016.2 .
a.i.l . acknowledges the support of dfg priority programme 1459 .
a.n.r . and m.i.k .
acknowledge support from the european union , horizon 2020 research and innovation programme under grant agreement no . 696656 , graphenecore1 .
s. roche , j. kerman , b. beschoten , j .- c .
charlier , m. chshiev , s. p. dash , b. dlubak , j. fabian , a. fert , m. guimares , f. guinea , i. grigorieva , c. schnenberger , p. seneor , c. stampfer , s. o. valenzuela , x. waintal and b. van wees , _ graphene spintronics : the european flagship perspective _ , 2d materials * 2 * , 030202 ( 2015 ) .
r. r. nair , m. sepioni , i - ling tsai , o. lehtinen , j. keinonen , a. v. krasheninnikov , t. thomson , a. k. geim , and i. v. grigorieva , _ spin - half paramagnetism in graphene induced by point defects _ , nat .
* 8 * , 199 ( 2012 ) .
r. r. nair , i - ling tsai , m. sepioni , o. lehtinen , j. keinonen , a. v. krasheninnikov , a. h. castro - neto , m. i. katsnelson , a. k. geim , and i. v. grigorieva , _ dual origin of defect magnetism in graphene and its reversible switching by molecular doping _
commun . * 4 * , 2010 ( 2013 ) .
h. gonzlez - herrero , j. m. gmez - rodrguez , p. mallet , m. moaied , j. j. palacios , c. salgado , m. m. ugeda , j .- y .
veuillen , f. yndurain , and i. brihuega , _ atomic - scale control of graphene magnetism by using hydrogen atoms _ , science * 352 * , 437 ( 2016 ) .
d. w. boukhvalov , m. i. katsnelson , and a. i. lichtenstein , _ hydrogen on graphene : electronic structure , total energy , structural distortions and magnetism from first - principles calculations _
b * 77 * , 035427 ( 2008 ) . a. n. rudenko , f. j. keil , m. i. katsnelson , and a. i. lichtenstein , _ exchange interactions and frustrated magnetism in single - side hydrogenated and fluorinated graphene _
* 88 * , 081405(r ) ( 2013 ) . t. o. wehling , e. aiolu , c. friedrich , a. i. lichtenstein , m. i. katsnelson , and s. blgel , _ strength of effective coulomb interactions in graphene and graphite _ , phys . rev .
106 * , 236805 ( 2011 ) .
r. j. kashtiban , m. a. dyson , r. r. nair , r. zan , s. l. wong , q. ramasse , a. k. geim , u. bangert , j. sloan , _ atomically resolved imaging of highly ordered alternating fluorinated graphene _ , nat .
commun . * 5 * , 4902 ( 2014 ) .
p. giannozzi , s. baroni , n. bonini , m. calandra , r. car , c. cavazzoni , d. ceresoli , g. l. chiarotti , m. cococcioni , i. dabo , a. dal corso , s. de gironcoli , s. fabris , g. fratesi , r. gebauer , u. gerstmann , c. gougoussis , a. kokalj , m. lazzeri , l. martin - samos , n. marzari , f. mauri , r. mazzarello , s. paolini , a. pasquarello , l. paulatto , c. sbraccia , s. scandolo , g. sclauzero , a. p. seitsonen , a. smogunov , p. umari , and r. m. wentzcovitch , _ quantum espresso : a modular and open - source software project for quantum simulations of materials _ , j. phys . : condens
. matter , * 21 * , 395502 ( 2009 ) .
s. a. nikolaev , v. v. mazurenko , a. a. tsirlin , and v. g. mazurenko , _ first - principles study of the magnetic ground state in kagome francisites cu@xmath70bi(seo@xmath70)@xmath1o@xmath1x ( x = cl , br ) _ ,
arxiv:1603.05192 .
a. i. liechtenstein , m. i. katsnelson , v. p. antropov , and v. a. gubanov , _ local spin density functional approach to the theory of exchange interactions in ferromagnetic metals and alloys _
, j. magn .
magn . mater . *
67 * , 65 ( 1987 ) .
m. schler , m. rsner , t. o. wehling , a. i. lichtenstein , and m. i. katsnelson , _ optimal hubbard models for materials with nonlocal coulomb interactions : graphene , silicene , and benzene _ , phys .
lett . * 111 * , 036601 ( 2013 ) .
i. beljakov , v. meded , f. symalla , k. fink , s. shallcross , m. ruben and w. wenzel , _ spin - crossover and massive anisotropy switching of 5d transition metal atoms on graphene nanoflakes _ , nano lett .
* 14 * , 3364 ( 2014 ) .
d. i. badrtdinov , o. s. volkova , a. a. tsirlin , i. v. solovyev , a. n. vasiliev , and v. v. mazurenko , _ hybridization and spin - orbit coupling effects in quasi - one - dimensional spin @xmath72 magnet ba@xmath70cu@xmath70sc@xmath73o@xmath74 _ , arxiv:1604.03333 .
i. kzsmrki , s. bordcs , p. milde , e. neuber , l. m. eng , j. s. white , h. m. rnnow , c. d. dewhurst , m. mochizuki , k. yanai , h. nakamura , d. ehlers , v. tsurkan , and a. loidl , _ nel - type skyrmion lattice with confined orientation in the polar magnetic semiconductor gav@xmath73s@xmath75 _ , nat . mater . * 14 * , 1116 ( 2015 )
. a. n. rudenko , f. j. keil , m. i. katsnelson , and a. i. lichtenstein , _ adsorption of cobalt on graphene : electron correlation effects from a quantum chemical perspective _ , phys .
b * 86 * , 075422 ( 2012 ) .
|
according to the lieb s theorem the ferromagnetic interaction in graphene - based materials with bipartite lattice is a result of disbalance between the number of sites available for @xmath0 electrons in different sublattices . here
, we report on another mechanism of the ferromagnetism in functionalized graphene that is the direct exchange interaction between spin orbitals . by the example of the single - side semihydrogenated ( c@xmath1h ) and
semifluorinated ( c@xmath1f ) graphene we show that such a coupling can partially or even fully compensate antiferromagnetic character of indirect exchange interactions reported earlier [ phys .
rev .
b * 88 * , 081405(r ) ( 2013 ) ] . as a result , c@xmath1h is found to be a two - dimensional material with the isotropic ferromagnetic interaction and negligibly small magnetic anisotropy , which prevents the formation of the long - range magnetic order at finite temperature in accordance with the mermin - wagner theorem .
this gives a rare example of a system where direct exchange interactions play a crucial role in determining a magnetic structure . in turn
, c@xmath1f is found to be at the threshold of the antiferromagnetic - ferromagnetic instability , which in combination with the dzyaloshinskii - moriya interaction can lead to a skyrmion state .
|
transition metal oxides attract a lot of attention due to a great variety of physical phenomena , most of which go along with the ordering of some microscopic degrees of freedom as a function of , e.g. , temperature , pressure , or doping .
prominent examples are the striking metal - insulator transitions of the vanadates @xcite , high-@xmath4 superconductivity in the cuprates , or the colossal magnetoresistance observed in the manganates @xcite .
cobaltates have aroused much interest due to the occurrence of different spin states @xcite .
in addition , they are promising materials for thermoelectric applications @xcite . known since 1873 , when friedel discovered the mineral @xmath5 , the delafossites @xmath6 keep on generating a strong and ever increasing interest @xcite , especially after kawazoe _ et al .
_ evidenced simultaneous transparency and p - type conductivity @xcite , which laid ground for the development of transparent optoelectronic devices .
furthermore , the quasi two - dimensionality of the lattice and the triangular coordination of atoms gave rise to such exciting physical properties as strong anisotropies of the electrical conductivity and magnetic frustration effects .
the delafossite structure has the space group @xmath7 and results from a stacking of monoatomic triangular layers , see fig .
[ fig1 ] @xcite . .
palladium , cobalt , and oxygen atoms are shown as green , blue , and red ( light , big dark , and small dark ) spheres , respectively . ] in particular , the b - atoms are at the centers of distorted oxygen octahedra , which share edges and form the characteristic @xmath8 sandwich layers . these trilayers are interlinked by linear o a o bonds , resulting in a twofold coordination of the a - atoms
. however , the latter have , in addition , six in - plane nearest neighbour a - atoms .
for this reason , the structure may be likewise regarded as formed from single a - atom layers , which are intertwined by the octahedral sandwiches . for @xmath0 and @xmath1
, we will find this latter point of view particularly useful .
finally , the oxygen atoms are tetrahedrally coordinated by one a - atom and three b - atoms .
pressure studies reveal an increase of the structural anisotropy on compression indicating the high mechanical stability of both the octahedral sandwich layers and the o pd
o ( o pt o ) dumbbells @xcite . in general , interest in the delafossite - type compounds has concentrated quite much on the triangular arrangement of the transition - metal atoms and the resulting possible frustration effects , which arise once localized magnetic moments are established . while most of these oxides have been found to be antiferromagnetic semiconductors , other class members like @xmath9 , @xmath0 , @xmath10 , and @xmath1 attracted interest due to their rather high metallic conductivity .
in particular , @xmath0 has been shown to possess one of the lowest electric resistivities of normal - state oxides , even lower than that of pd metal at room temperature @xcite . yet , the conductivity is strongly anisotropic @xcite . in particular ,
the ratio of the resistivities parallel and perpendicular to the @xmath11 axis can be as large as @xmath12 in @xmath0 @xcite .
photoemission data indicate that the density of states at the fermi energy can be exclusively attributed to the pd @xmath13 states @xcite . from the combination of photoemission spectroscopy and inverse photoemission spectroscopy several authors
concluded that the fermi energy is located at a shallow minimum of the density of states and doping might thus lead to rather high values of the thermoelectric power @xcite . despite their simple chemical formulae the delafossites may be regarded as prototypical superlattices where the composition of both the a and b layers can be used to strongly influence the behavior of the whole system . for instance , in @xmath14 , the fermi energy falls into the cr @xmath15 band , but since the cr layers order magnetically this compound is a magnetic semiconductor . in contrast ,
as will be shown below , in @xmath0 , the co layers only act a charge reservoirs , and conduction takes place almost exclusively in the pd layers . as a matter of fact , quite
a few electronic structure calculations for delafossite compounds have been reported in the literature @xcite . for @xmath0 , there exist linear muffin - tin orbitals calculations by seshadri _
et al . _ as well as by okabe _ et al . _
@xcite . while according to the former authors , who also investigated @xmath1 , the density of states at @xmath16 is mainly due to the pd @xmath13 states with only small contributions from the co @xmath15 and o @xmath17 orbitals , the results by okabe _
_ are not very specific in this respect . for this reason ,
the role of these orbitals for the metallic conductivity is not yet completely clear . in order to resolve the issue and to make closer connection with the photoemission data , we apply in the present work the new full - potential augmented spherical wave method to study the electronic properties of the title compounds .
we concentrate especially on the strong anisotropies and on the influence of the different species and orbitals on the electronic properties .
the calculations are based on density - functional theory and the generalized gradient approximation ( gga ) @xcite with the local - density approximation parametrized according to perdew and wang @xcite .
they were performed using the scalar - relativistic implementation of the augmented spherical wave ( asw ) method ( see refs . and references therein ) .
in the asw method , the wave function is expanded in atom - centered augmented spherical waves , which are hankel functions and numerical solutions of schrdinger s equation , respectively , outside and inside the so - called augmentation spheres . in order to optimize the basis set
, additional augmented spherical waves were placed at carefully selected interstitial sites .
the choice of these sites as well as the augmentation radii were automatically determined using the sphere - geometry optimization algorithm @xcite .
self - consistency was achieved by a highly efficient algorithm for convergence acceleration @xcite .
the brillouin zone integrations were performed using the linear tetrahedron method with up to 1469 * k*-points within the irreducible wedge @xcite . in the present work
, we used a new full - potential version of the asw method , which was implemented only very recently @xcite . in this version ,
the electron density and related quantities are given by a spherical - harmonics expansion inside the muffin - tin spheres .
in the remaining interstitial region , a representation in terms of atom - centered hankel functions is used @xcite . however , in contrast to previous related implementations , we here get away without needing a so - called multiple-@xmath18 basis set , which fact allows for a very high computational speed of the resulting scheme .
while the previous calculations were based on the crystal structure data by prewitt _
et al . _
@xcite , we here calculated these parameters from a minimization of the total energy . to this end , in a first step for each compound the lattice was relaxed and after that the oxygen parameter was optimized .
the results for both compounds are summarized in tab .
[ tab1 ] .
.experimental and calculated lattice parameters ( in ) and atomic positions . [ cols="<,<,^,^,^",options="header " , ] [ tab1 ] note that the deviation of the calculated structural parameters from the measured ones is 2.4% at most , which is very well within the known limits of the gga and an excellent proof of the validity of the new full - potential asw method . in order to discuss the electronic properties , we first concentrate on @xmath0 and only after that discuss the changes coming with the substitution of pt for pd .
the electronic bands along selected high - symmetry lines of the first brillouin zone of the hexagonal lattice , fig .
[ fig2 ] , are displayed in fig .
[ fig3 ] . .
] the corresponding partial densities of states ( dos ) are shown in fig .
[ fig4 ] . .
selection of co @xmath15 orbitals is relative to the local rotated reference frame , see text . ]
the rather complicated structure of both the electronic bands and the dos results from the energetical overlap of the relevant orbitals in the energy interval shown . while the lower part of the spectrum is dominated by o @xmath17 states , the transition metal @xmath2 states
lead to rather sharp peaks in the interval from @xmath19ev to @xmath20ev . in particular , we recognize the @xmath21 and @xmath22 manifolds of the co @xmath15 states as resulting from the octahedral coordination . in representing these partial dos
we have used a local rotated coordinate system with the cartesian axes pointing towards the oxygen atoms .
@xmath23-type overlap of the o @xmath17 states with the co @xmath15 @xmath22 orbitals leads to the rather sharp contribution of the latter near @xmath24ev .
in contrast , due to the much weaker @xmath25-type overlap of the o @xmath17 states with the @xmath21 orbitals , these states give rise to sharp peaks in the interval from @xmath26ev to @xmath16 .
since the fermi energy falls right between the @xmath21 and @xmath22 manifolds , co turns out to be in a @xmath27 low - spin state .
our results thus provide further support to the picture that co is trivalent while pd is in a monovalent @xmath28 configuration @xcite .
furthermore , the co and o states give only a tiny contribution to the electrical conductivity , which is maintained almost exclusively by the pd @xmath13 states .
the latter are further analyzed in fig .
[ fig5 ] , dos of @xmath0 . ] which displays the five pd @xmath13 partial dos . since pd
is linearly coordinated by two oxygen atoms parallel to the @xmath11 axis and has six pd neighbours in the @xmath29-@xmath30 plane we used the global coordinate system to represent these partial dos . with this choice , contributions from the @xmath31 and @xmath32 as well as from the @xmath33 and
@xmath34 states are identical . the pd @xmath13 partial dos are strongly influenced by the aforementioned coordination .
in particular , the six peaks of the @xmath35 states near @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 , and @xmath41ev are complemented by contributions of the o @xmath17 partial dos of similar shape reflecting the strong @xmath23-type @xmath2-@xmath42 overlap along the @xmath11 axis . in contrast , the short in - plane pd - pd distances of about 2.83 ( experimental value ) , which are only by 3% longer than those in metallic pd , lead to the broad pd @xmath43 bands visible in the interval from @xmath44 to @xmath45ev in fig .
in addition , these bands give rise to the largest contribution at @xmath16 , whereas the @xmath46 states do not contribute at all .
in contrast , the contribution from the @xmath35 orbitals is rather similar to that of the @xmath43 orbitals .
however , according to a @xmath47-resolved analysis , their weights vary on the fermi surface . at about @xmath48ev , the contribution of the @xmath43 orbitals is strongly suppressed , and the small dispersion along the line m - k causes the sharp @xmath35 peak seen in fig .
[ fig5 ] .
the strong quasi two - dimensionality of the electronic states is reflected by the fermi surface depicted in fig .
[ fig8 ] . .
] apart from the very small bending parallel to the @xmath11 direction , it gives rise to completely in - plane fermi velocities and explains the strong anisotropy in electric conductivity .
however , note that , according to the electronic bands shown in fig . [ fig3 ] , the dispersion parallel to the direction @xmath49-a in general is not negligible . in passing ,
we mention the negligible dispersion along m - k , which is typical of tight - binding bands in a triangular lattice .
these bands give rise to the sharp peak at about @xmath39ev in fig .
[ fig5 ] and can thus be attributed to the pd @xmath35 states .
overall , our results are in good agreement with the previous calculations by seshadri _
et al . _ and by okabe
_ et al . _
in particular , seshadri _
et al . _
obtained a distribution of states at @xmath16 , which is similar to ours .
furthermore , our results are in agreement with the photoemission data by tanaka _
_ and higuchi _ et al . _
@xcite , who attribute the metallic conductivity almost exclusively to the pd @xmath13 states .
it has been argued by several authors , that the dominant contribution is due to the pd @xmath35 orbitals , which hybridize with the pd @xmath50 states .
while we find indeed a @xmath50 contribution of the order of @xmath51states / ev , hence , of the order of 10% of the @xmath35 partial dos at @xmath16 , we still point to the in - plane @xmath31 and @xmath32 states , which play an even greater role at @xmath16 than the @xmath35 states . as a consequence ,
our results demonstrate that the metallic conductivity is maintained by the in - plane @xmath31 and @xmath32 orbitals and the in - plane part of the @xmath35 orbitals to a similar degree with a somewhat greater influence of the former .
this finding is confirmed by the calculations for @xmath1 , where the dos at @xmath16 is again almost exclusively due to the in - plane states . in general , the partial dos obtained for @xmath1 , which are displayed in fig .
[ fig9 ] , .
selection of co @xmath15 orbitals is relative to the local rotated reference frame , see text . ]
resemble those of the palladium system .
however , the larger extent of the pt @xmath52 orbitals as compared to the pd @xmath13 states leads to an increased width mainly of the in - plane @xmath2 bands as well as an increased overlap of these states with the o @xmath17 states . as a consequence
, the anisotropy of the electrical conductivity will be somewhat reduced in @xmath1 as is indeed observed @xcite . in passing ,
we mention additional gga+u calculations for @xmath0 , which however , confirmed the gga results without any noticeable changes .
in summary , we have shown that the strongly anisotropic metallic conductivity of @xmath0 and @xmath1 is almost exclusively due to the pd ( pt ) layers .
in contrast , the octahedrally coordinated @xmath3 sandwiches are insulating .
as co was found in a low - spin @xmath27 configuration , these @xmath3 complexes strongly suppress the electronic coupling between the pd ( pt ) metallic layers , and the @xmath53 layers merely act as charge reservoirs . in addition , they pronounce the quasi two - dimensionality of the system . it may be speculated how the introduction of impurities into the cobalt layers , i.e. replacing , e.g. fe or ni for co , alters the electronic properties locally and might thus be used to design nano - structured materials .
work along this line is in progress . in this context
one might ask to what extent the structural changes induced by the introduction of impurities with different covalent radii as compared to co might be damped by the intermediate oxygen layers .
this would provide a situation , where chemical substitution only acts on the electronic degrees of freedom .
we gratefully acknowledge many fruitful discussions with t. kopp , c. martin , and w. c. sheets .
this work was supported by the deutsche forschungsgemeinschaft through sfb 484 .
[ fig1 ] and [ fig8 ] were generated using the xcrysden software ( ref . ) .
r. d. shannon , d. b. rogers , and c. t. prewitt , inorg .
chem . * 10 * , 713 ( 1971 ) ; c. t. prewitt , r. d. shannon , and d. b. rogers , _ ibid .
_ * 10 * , 719 ( 1971 ) ; d. b. rogers , r. d. shannon , c. t. prewitt , and j. l. gillson , _ ibid . _ * 10 * , 723 ( 1971 ) .
|
the origin of the quasi two - dimensional behavior of @xmath0 and @xmath1 is investigated by means of electronic structure calculations .
they are performed using density functional theory in the generalized gradient approximation as well as the new full - potential augmented spherical wave method .
we show that the electric conductivity is carried almost exclusively by the in - plane pd ( pt ) @xmath2 orbitals .
in contrast , the insulating @xmath3 sandwich layers of octahedrally coordinated co atoms may be regarded as charge carrier reservoirs .
this leads to a weak electronic coupling of the pd ( pt ) layers .
the obtained nearly cylindrical fermi surface causes the strong anisotropy of the electric conductivity .
|
micro - electro - mechanical systems ( mems ) are increasingly finding practical application .
friction from solid - solid contacts in bearings and hinges can lead to device failure or breakage , so practical devices have primarily used flexing components , such as cantilevers , rather than rotating shafts and gears , which are common in macro - scale devices .
frictional concerns are one of the primary obstacles to using rotating components for mems.@xcite in macroscopic devices , air bearings with an actively injected lubricating gas layer or journal bearings with a dynamically maintained lubricant film are used in high - performance applications , but engineering such intricate structures is challenging in mems.@xcite the use of surface tension to maintain a liquid lubricant layer is an alternative that is most effective in mems , since surface tension is very significant in microscopic devices . furthermore , surface tension and fluid pressure at these scales can actually be used to passively support a mems bearing , eliminating all solid - solid contact . in such bearings
, a liquid , usually water or an aqueous solution , is used between the two solid surfaces .
variations in surface wetting ( e.g. hydrophobicity ) are used to pin the liquid in position .
the liquid - solid contact is low friction compared to solid - solid contacts , allowing parts to slide with less stiction and lower wear.@xcite a variety of geometries for surface tension supported liquid bearings have been investigated . in the simplest case , a single drop of water trapped between identical hydrophilic pads in the center of the bearing ( fig .
[ fig : bearing_types]a ) supports the rotating part ( rotor ) above the stationary substrate ( stator).@xcite wear is reduced due to the lack of solid contact and concentricity of the pads is well maintained by surface tension , but the tilt stability is poor , as there is little energetic cost to tipping the rotor even to the point of a collision between the rotor and stator . using a ring of water reduces the hydrodynamic drag on the surfaces somewhat by reducing the wetted surface area ( fig .
[ fig : bearing_types]b ) , and has increased tilt stiffness.@xcite the tilt stiffness can be greatly increased by breaking up the wetting ring into discrete drops and making the rotor superhydrophobic , as in fig .
[ fig : bearing_types]c.@xcite a superhydophobic surface is a structured surface with a water contact angle greater than 150@xmath0.@xcite since the drops do not wet to the rotor surface , a central drop , which wets to both sides , is added to the center to maintain the relative positions of the surfaces .
the decreased contact area between the water and solid can reduce the hydrodynamic drag , and the superhydrophobic contacts provide high lifting forces and reduce titling , but energy loss due to the remaining contact angle hysteresis of the superhydrophobic surface may dominate the sources of drag.@xcite the ideal surface tension supported liquid bearing would provide high stiffness in all directions while minimizing all sources of drag . in this paper , we investigate a new ring and axle bearing design ( fig .
[ fig : bearing_types]d and fig .
[ fig : bearing ] ) , which combines several of the best features of other designs .
the design uses a central drop , which wets hydrophilic areas on both the rotor and the stator , as an axle .
the axle holds the rotor down and keeps the rotor concentric with the stator .
a ring of water is wetted to a hydrophilic annulus on the stator , while the surface of the rotor in contact with the ring is superhydrophobic .
the ring provides a vertical force to balance that of the axle , and stiffens the bearing with respect to tilting .
since the rotor is superhydrophobic where it contacts the ring of liquid , there may be less hydrodynamic drag than in the ring bearing , where the rotor surface in contact with the liquid is hydrophilic.@xcite furthermore , since there is no wetting and dewetting in the system as the rotor spins , there are no hysteretic losses and thus the drag is expected to be reduced compared to the drop and axle bearing design .
this new ring and axle liquid bearing design is not unconditionally stable , e.g. the ring will actually push the rotor to tilt if it is overfilled with water . here , we examine this system and determine the parametric region of stability . the axle and ring each generally provide two contributions to the forces acting on the rotor : the force due to laplace pressure , which is the result of the pressure difference across the liquid surface , and the force directly due to surface tension of the water acting on the rotor . even in the simplest cases ,
calculating the stability and stiffness of the bearing analytically is not trivial , and numerical solution is more practical .
we examine the bearings via numerical modeling using energy minimization.@xcite
we performed simulations of our bearings using surface evolver@xcite , an open source finite element analysis system .
surface evolver minimizes the surface energy of the liquid subject to constraints , which include the size and contact angles of the wetting and non - wetting surfaces , and the volume of liquid in each region .
the effect of gravity is also included , using experimental values for the mass of the rotor . only the liquid surface was directly modeled ; the rotor and stator were represented by boundary conditions and constraints .
typically , the distance between the rotor and stator is allowed to vary , but the angle of the rotor is fixed in any one calculation .
the stiffness of the bearing and the equilibrium tilt angle were calculated by displacing the rotor vertically or tilting the rotor while adjusting the liquid - air interface to minimize energy .
the tilt with the minimum energy was recorded as the equilibrium tilt , while the curvature of the energy versus height or tilt was used to calculate the stiffness .
an example of a generated liquid surface mesh can be seen in fig .
[ fig : compare ] . not all combinations of constraints resulted in physically realizable bearing configurations . in the model , a configuration
was considered unstable if either the center drop separated or the contact of the ring of water on the upper surface vanished at any point during the calculation . for a range of axle and ring volumes ,
the model was used to determine if the configuration would be stable and , if stable , the equilibrium rotor height , tilt , and tilt stiffness .
to verify the validity of the model , the bearing design was also experimentally tested .
we fabricated bearings with the same parameters as the model starting with @xmath1% pure aluminum and using single point diamond turning ( spdt ) to cut the aluminum to a flat surface ( @xmath2 nm rms roughness ) .
where needed , surfaces were made superhydrophobic by a modified porous aluminum oxide ( paa ) growth technique @xcite and fluoropolymeric coating composed of spin coated hexamethlydisalizane ( hmds)@xcite and solvey plastics hyflon ad-60@xcite , as reported previously.@xcite the stator ring and axle hydrophilic areas were patterned by coating a superhydrophobic surface with photoresist and sputter - depositing ti , then performing lift off , to form a @xmath3 mm radius circular hydrophilic region for the axle and an annular hydrophilic region with an average @xmath4 mm radius and a width of @xmath5 mm .
spdt was then used to cut out @xmath6 cm diameter rotors from @xmath7 mm thick discs with a superhydrophobic surface produced by the same method .
sptd was also used to remove the superhydrophobic structure in a @xmath3 mm radius circle at the center of the rotor to create a matching hydrophilic circle for the axle water to wet .
the rotor hydrophilic area was defined by spdt instead of photolithography to guarantee concentricity of the rotor edge and the axle to within the accuracy of the ultra - precision lathe used for spdt ( @xmath8 nm ) .
a rotor and stator pair are shown in fig .
[ fig : bearing ] .
experimentally , the bearings were photographed in profile over a range of center and ring liquid volumes ( examples shown in fig . [ fig : failure ] ) .
the parts were allowed to completely dry between tests to minimize volume uncertainties , and the ring and axle hydrophilic regions were filled with water just before placement to minimize evaporation . to achieve full wetting of both axle pads
reliably with such small volumes of water , half of the volume had to be placed on each side .
rotors were handled by vacuum tweezers from the back side to ensure that the superhydrophobic surfaces were not damaged .
after placement of the rotor , the entire bearing was carefully rotated to ensure that the image was taken in profile perpendicular to the tilt .
image analysis was then used to find the height and angle of tilt . in both experiments and models
, we find that there are several possible sources of instability .
[ fig : compare ] shows the results of experiments and models with the same range of axle and ring volumes .
if the axle has insufficient liquid relative to the ring , the liquid on the rotor and stator will not join to form an axle or will re - separate due to the lift from the ring with most of the water on one side , and the rotor will not remain centered ( fig . [
fig : failure]a ) ; this leads to the loss of stability seen in the lower right area in each plot in fig .
[ fig : compare ] .
if the liquid in the ring is insufficient relative to the axle , then the rotor is lifted partially off of the ring and tips to one side , since balancing on the center drop alone is unstable ( fig .
[ fig : failure]b ) ; this leads to the loss of stability seen in the upper left area in each graph in fig .
[ fig : compare ] . finally , if the ring has too much water in it , it bulges asymmetrically , again resulting in a tilted ( and often off - center ) rotor ( fig .
[ fig : failure]c ) ; this requires a volume in the rings such that they have more than a semi - circular crossection , which is larger than the maximum ring volume shown in fig .
[ fig : compare ] .
there is an exact match between the predicted and measured regions of stability .
this shows that a simple model can predict the stable range of parameters for the bearing .
the experimentally measured and numerically modeled rotor heights show reasonable agreement .
both show increasing height of the rotor with increasing volume in either the axle or the ring .
quantitatively , the results agree to within @xmath9% with no free parameters in the model .
the results for the rotor tilt show less agreement between the experiments and model .
the model predicts effectively no tilt in the stable region , while experimentally measurable tilt is observed , particularly near the high ring volume and low axle volume boundary of the stable region .
this experimental tilt seems to fluctuate without any clear trend , and is not highly repeatable .
it is likely that experimentally this results from the assembly process of the bearing , where the bearing is not guaranteed to start off with zero tilt ; i.e. near the edge of the stable region , there may be some hysteresis in the tilt .
with some evidence that the model is accurate , we can then predict the stiffness of the bearing with respect to tilt ( vertical stiffness is not generally difficult to achieve , and thus is not analyzed ) .
the results , calculated from the dependence of the energy of the system on tilt , are shown in fig .
[ fig : compare ] .
using the same calculation method , for bearings with the same thickness ( 0.38 mm ) , a model drop bearing ( fig [ fig : bearing_types]a ) has a spring constant of 0.32 nj / degree@xmath10 ( drop covers the same area as the full ring and axle bearing ) , a ring bearing ( fig .
[ fig : bearing_types]b ) has a spring constant of 0.34 nj / degree@xmath10 ( compared to the full ring and axle bearing , the simulated ring bearing has an inner radius equal to the radius of the axle and an outer radius equal to the outer radius of the ring ) , and the new ring and axle bearing ( fig .
[ fig : bearing_types]d ) has a spring constant of 2.55 nj / degree@xmath10 , which represents a significant improvement .
the tilt stiffness of ring and axle bearings is calculated to generally increase with decreasing volume of liquid in either the axle or ring .
it is likely that having less liquid in the ring directly increases the stiffness , while less liquid in the axle pulls the rotor closer to the stator and compresses the ring , also increasing the bearing stiffness .
the surface tension supported ring and axle bearings reported have significant potential for use in mems due to their low frictional losses and complete lack of solid on solid contact .
these results demonstrate that these devices are stable over a large range of ring and axle volumes , and that this region can be modeled using energy minimization , simplifying design and experimental work with these devices .
it also shows that these devices have a higher tilt stiffness for a given size than alternative designs , which is critical since the the tilt stiffness is typically low in surface tension supported bearings .
furthermore , due to the superhydrophobic rotor surface and absence of contact angle hysteresis , ring and axle bearings could have lower drag than previous designs .
a common objection to water - based liquid bearings is that , while promising from a mechanics standpoint , the evaporation of the liquid water makes them of limited practical use .
although the testing presented here was performed with pure water , we have also tested longer term use of a saturated water - cacl@xmath11 solution on superhydrophobic surfaces with patterned hydrophilic regions .
this solution has similar surface tension to pure water and since cacl@xmath11 is deliquescent , it can actually pull moisture from the air , rather than allowing the water to evaporate . in practice , this solution was stable over a period of at least 2 years under ambient conditions , and shows no signs of evaporation or corrosion of the underlying surface .
we would like to thank dr .
kenneth brakke at the mathematics department of susquehanna university , the creator of surface evolver , for his assistance with understanding and optimizing our simulation .
m. l. chana , b. yoxalla , h. parka , z. kangb , i. izyuminb , j. choub , m. m. megensb , m. c. wub , b. e. boserb , and d. a. horsleya , `` design and characterization of mems micromotor supported on low friction liquid bearing , '' _ sensors and actuators a : physical _ , vol .
177 , pp . 19 , 2012 .
s. chu , k. wada , s. inoue , m. isogai , y. katsuta , and a. yaumori , `` large - scale fabrication of ordered nanoporous alumina films with arbitrary pore intervals by critical - potential anodization , '' _ journal of the electrochemical society _ , vol .
b384b391 , 2006 .
n. tasaltin , d. sanli , a. jon , a. kiraz , and c. erkey , `` preparation and characterization of superhydrophobic surfaces based on hexamethyldisilazane - modified nanoporous alumina , '' _ nanoscale research letters _ , vol . 6 , no . 487 , 2011 .
|
friction between contacting solid surfaces is a dominant force on the micro - scale and a major consideration in the design of mems .
non - contact fluid bearings have been investigated as a way to mitigate this issue . here
we discuss a new design for surface tension - supported thrust bearings utilizing patterned superhydrophobic surfaces to achieve improved drag reduction .
we examine sources of instability in the design , and demonstrate that it can be simply modeled and has superior stiffness as compared to other designs .
superhydrophobicity , porous anodized aluminum , bearings , non - contact
|
given a finite set @xmath7 of @xmath0 points , the diameter of @xmath7 denoted by @xmath8 is the maximum distance between two points of @xmath7 . computing the diameter of a set of points has a large history and it may be required in various fields such as database , data mining and vision . a trivial algorithm for
this problem is as follows : compute the distance between each pair of points and then choose the maximum distance .
since , computing the distance takes constant time , this algorithm takes @xmath9 time .
but , this is too slow for large - scale data sets that occur in the fields .
hence , we need a faster algorithm which may be exact or approximation . by reduction to set disjointness problem
, it can be shown that computing the diameter of @xmath0 points in @xmath10 requires @xmath11 operations in algebraic computation - tree model @xcite .
it is shown by yao that , it is possible to compute the diameter in sub - quadratic time in each dimension @xcite . in two and three dimensions ,
there are well - known solutions . in the plane
, this problem can be computed in optimal time @xmath12 , but , in three dimensions , it is hard .
clarkson and shor @xcite present an @xmath13-time randomized algorithm .
their algorithm needs to compute the intersection of @xmath0 balls ( with the same radius ) in @xmath14 .
it may be slower than the brute - force algorithm for the most practical data sets .
moreover , it is not an efficient method for higher dimensions , because the intersection of @xmath0 balls with the same radius has a large size .
some deterministic algorithms with running time @xmath15 @xcite and @xmath16 @xcite are found for solving this problem .
finally , ramos @xcite found an optimal deterministic @xmath13-time algorithm in @xmath14 .
cheong et.al @xcite present an @xmath17 randomized algorithm that computes the all - pairs farthest neighbors for @xmath0 points on the convex position in @xmath14 . in the absence of fast algorithms , many attempts have been done to approximate the diameter in low and high dimensions .
a 2-approximation algorithm with @xmath18 time in @xmath1 dimensions can be found easily by selecting a point @xmath19 of @xmath7 and then finding the farthest point @xmath20 from it by brute - force manner @xcite .
the first non - trivial approximation algorithm for this problem is presented by egeciogln and kalantari @xcite that approximates the diameter with factor @xmath21 and operations cost @xmath18 in @xmath1 dimensions .
also , they present an iteration algorithm with @xmath22 iterations and the cost @xmath18 for each iteration that has approximate factor @xmath23 . agarwal et .
al @xcite present a @xmath2-approximation algorithm in @xmath10 with @xmath24 running time by projecting to directions .
barequet and har peled @xcite present a @xmath25-approximate diameter method with @xmath18 time , since @xmath1 is fixed , it can be denoted by @xmath4 .
they also describe a @xmath2-approximate approach for computing the diameter @xmath26 such that @xmath27 with @xmath28 time in @xmath10 .
they show that , the running time can be improved to @xmath29 .
similarly , chan @xcite observes that a combination of two approaches in @xcite and @xcite yields an @xmath2-approximate algorithm with @xmath30 and an @xmath31-approximate algorithm with @xmath32 time .
he also introduces a core - set theorem , and shows that using this theorem , an @xmath31-approximation for the diameter in @xmath33 time can be found @xcite . also , har peled @xcite present an approach which for the most inputs , it is able to compute the exact diameter or an approximation very fast .
although , in the worst case , the algorithm running time is still quadratic , and is sensitive to the hardness of the input . his algorithm is able to return a pair of points @xmath34 and @xmath35 such that @xmath36 , for each value @xmath37 in each dimension , with @xmath38 running time .
he shows that with a complicated analysis , this running time can be reduced to @xmath39 .
simultaneously , maladain and boissonnat @xcite present an exact algorithm for the diameter by computing the double normals in each dimension .
but , their algorithm is not worst - case optimal .
they also show , with having double normals , a @xmath21-approximation of the diameter in each dimension is provided .
moreover , finocchiaro and pellegrini @xcite describe an algorithm that finds in @xmath40 time with high probability a @xmath2-approximation of the diameter of a set of @xmath0 points in @xmath1-dimensional euclidean space .
as previously mentioned , some @xmath31-approximation algorithms have been presented in literatures .
the first result is an @xmath2-approximate with running time @xmath41 due to agarwal et .
al @xcite which is based on projection to directions
. other methods which present approximate approaches with running time @xmath42 , where they have usually a hidden constant @xmath43 with exponential dependency on @xmath1 .
they include barequet and har peled @xcite method , chan s @xcite results which are a combination of two approaches in @xcite and @xcite , and finally , har peled @xcite approach .
however , most of these methods use a similar idea ( except the last one ) .
they first compute a 2-approximation @xmath44 of the diameter of the point set @xmath45 in @xmath18 time by choosing a point @xmath19 of @xmath45 , and then finding its farthest point @xmath20 , and use @xmath44 for constructing a grid . then , round each point to its nearest grid - point and find the diameter in this new point set .
but , using this idea has the following properties .
firstly , it leads to a rounded point set with @xmath46 points ( balls ) which needs @xmath47 time to compute its diameter .
secondly , it uses a 2-approximation , which choices a point randomly .
this means that , if we run these algorithms twice on an input set , they might result different value for the approximate diameter .
thus , the precisely running time of barequet and har peled @xcite method is @xmath48 @xcite . since , @xmath1 is fixed , they assume that @xmath49 and then result the running time @xmath50 .
this condition also holds for chan s results @xcite . finally , har peled @xcite applied a different approach which is used a fair - split tree .
his algorithms running time precisely is @xmath51 . by assuming that @xmath52 , he obtains the running time @xmath38 . in this paper
, we propose a new @xmath2-approximation algorithm for computing the diameter of a set @xmath7 of @xmath0 points in @xmath10 with @xmath3 time and @xmath4 space , where @xmath5 .
our proposed algorithm obtains new results by applying a different approach in compare with the previous methods .
firstly , we use a constant value @xmath53 for constructing our grids instead of using the 2-approximate diameter @xmath44 .
the value @xmath53 is the largest side of the axis - parallel bounding box of the point set .
so , our algorithm produces same results , if we run it twice or more , because , it depends on a constant value @xmath53 .
secondly , we round each points to its central - cell points in a grid which helps us to get a better running time by decreasing the search domain .
moreover , we propose an approximate algorithm with @xmath6 time .
so , we obtain some approximation algorithms which have some improvements in the running time in compare with previous algorithms for computing the diameter in fixed dimensions .
the reminder of this paper is organized as follows : in section 2 , we describe our proposed algorithm .
subsection 2.1 includes our analysis over the algorithm .
section 3 is conclusions .
in this subsection , we describe our new @xmath2-approximate algorithm to compute the diameter of a set @xmath7 of @xmath0 points in @xmath10 . in this problem , we want find a diametrical pair @xmath54 and @xmath55 such that @xmath56 . to this end , we first find extreme points in each coordinate and compute the axis - parallel bounding box of @xmath7 , which is denoted by @xmath57 .
we use the largest length side of the @xmath57 to impose two grids on it .
in fact , we decompose @xmath57 to a grid of regular hypercubes with side length @xmath58 , where @xmath59 .
we call each hypercube as a cell . then , each point of @xmath7 is rounded to its corresponding central cell - point .
fig [ fig : round ] shows an example of the rounding process for a point set in @xmath60 . and an @xmath58-gird .
initial points are shown by blue points and their corresponding central cell - points are shown by circle points .
( b ) rounded point set @xmath61.,title="fig:",height=4 ] and an @xmath58-gird .
initial points are shown by blue points and their corresponding central cell - points are shown by circle points .
( b ) rounded point set @xmath61.,title="fig:",height=4 ] [ fig : round ] as can be seen in fig [ fig : round ] , many points may be rounded to a point in @xmath61 .
but , for improving the running time of the algorithm , in this stage , we impose an @xmath62-grid to @xmath57 , for @xmath63 and round each point of the new point set @xmath61 to its central cell - point in this new grid .
this results a new rounded point set @xmath64 .
note that , we only need @xmath18 time to round points to its central - cell points .
let , @xmath65 be a hypercube with side length @xmath62 and central - point @xmath34 .
this step reduces regions which may be include the diametrical pairs to two hypercubes @xmath66 and @xmath67 , and helps us to give a better running time .
let , two point sets @xmath68 and @xmath69 include points of the rounded point set @xmath61 , which are inside two hypercubes @xmath66 and @xmath67 , respectively .
it is sufficient to find diameter between points of @xmath61 , which are in two sets @xmath68 and @xmath69 . in this subsection
, we analysis the proposed algorithm .
[ enumi ] theorem * * _ _ the proposed approximate algorithm computes the diameter of a set @xmath7 of @xmath0 points in @xmath10 in @xmath70 time and @xmath4 space , where @xmath71 . finding the extreme points in all
coordinate and finding the largest side of @xmath57 can be done in @xmath18 time .
rounding step takes @xmath72 time for each point and for all of them takes @xmath18 time .
since , @xmath1 is fixed , this can be shown by @xmath4 .
but , for computing the approximate diameter over the rounded point set @xmath64 , we need to know the number of points in the set @xmath64 . we know that the largest side of the bounding box @xmath57 has length @xmath53 and
the side length of each cell is @xmath73 . on the other hand ,
the volume of a hypercube of side length @xmath74 in @xmath1-dimensional space is @xmath75 , therefore , the number of central cell - points in the bounding box @xmath57 is at most @xmath76 .
so , the number of points in @xmath64 is at most @xmath77 .
hence , by brute - force we need @xmath78 time for computing the diameter of rounded point set @xmath64 . in the next step of the algorithm , we need to compute two sets @xmath68 and @xmath69 , which include points of @xmath61 are in two hypercubes @xmath66 and @xmath67 , respectively .
clearly , this work takes @xmath18 time .
then , for finding the diameter between two sets @xmath68 and @xmath69 , we need to know number of points on each of them . on the other hand ,
maximum number of points in two sets @xmath68 and @xmath69 is @xmath79 .
so , for computing @xmath80 , we need to @xmath81 time .
but , we may be have more than one diametrical pairs @xmath68 and @xmath69 . since , we may be have at most @xmath82 different diametrical pairs @xmath68 and @xmath69 , this step takes at most @xmath83 time .
so , total complexity of our algorithm is as follows : @xmath84 since , @xmath1 is fixed , we have : @xmath85 .
we can reduce running time of the proposed algorithm by discarding some central cell - points which do not have any potential to be the diametrical pairs in rounded point set @xmath64 , and similarly , in two point sets @xmath68 and @xmath69 . if we consider all the points which are same in their first @xmath86 coordinates , and keep only highest and lowest . then
, number of points in @xmath64 , and two point sets @xmath68 and @xmath69 can be reduced to @xmath87 .
so , using the brute - force quadratic algorithm we need @xmath88 time to find the diametrical pair .
hence , this gives us the total running time @xmath89 .
about required space , we only need to @xmath4 space for storing required points sets .
so , this completes the proof .
our last result is based on a combination of two approaches which are used in @xcite and @xcite .
let , we round initial point set and obtain a reduced point set with @xmath87 points in @xmath4 time .
after rounding step , in the agrawal et . al @xcite method instead of projecting the points on @xmath90 different directions which need many projection operations , we project points on some hyperplanes and compute the approximate diameter in these hyperplanes .
this results in an improved approximation that we have presented in the following corollary .
[ enumi ] corollary * * _ _ an @xmath31-approximate for the diameter of a set of n points in @xmath1-dimensional euclidean space , can be computed in @xmath91 time , where @xmath92 .
now , back to the details of the approximation . in the next lemma
, we show that the proposed algorithm computes an approximate diameter @xmath26 with error range @xmath93 from the true diameter @xmath94 , where @xmath92 .
[ enumi ] lemma * * _ _ algorithm 1 computes an approximate diameter @xmath26 such that : @xmath95 , where @xmath92 .
let @xmath96 and @xmath97 be two central cell - points that are used by the algorithm for computing the approximate diameter @xmath98 .
then , we have two corresponding exact points @xmath34 and @xmath35 for these two approximate points .
fig [ fig : exam ] illustrates a rounded point set and its diameter .
is computed over rounded point set @xmath61 in @xmath60 .
two central cell - points @xmath96 and @xmath97 are used to compute @xmath98 .
their corresponding initial points are @xmath34 and @xmath35 . ]
so , we have : @xmath99 now , we can simplify ( 2 ) as following : @xmath100 by setting @xmath101 , we have : @xmath102 since @xmath103 , and by setting @xmath104 , we can result : @xmath105 therefore , the lemma is resulted .
we have presented an efficient @xmath2-approximation algorithm to compute the diameter of a point set @xmath7 in @xmath10 , for a fixed dimension @xmath1 in @xmath3 time , when @xmath106 .
the proposed algorithm used new approach that results a method which are simple in implementation . moreover , we show that the running time of the proposed algorithm can be improved to an @xmath107-approximate with @xmath6 time .
thus , these results provide some improvements in the running time of this problem .
cheong , o , shin , c. s. , vigneron a. : computing farthest neighbors on a convex polytope . in proceedings of the 7th annual international computational and combinatoric conference ,
volume 2108 of lncs , pp .
159 - 169 , ( 2001 )
|
in this paper , we study the problem of computing the diameter of a set of @xmath0 points in @xmath1-dimensional euclidean space , for a fixed dimension @xmath1 , and propose a new @xmath2-approximation algorithm with @xmath3 time and @xmath4 space , where @xmath5 .
the proposed algorithm is simple in implementation and does not need complicated data structures .
we also improve the proposed algorithm to another algorithm with @xmath6 running time .
diameter , point sets , approximation algorithm , higher dimensions .
|
in order to understand formation and evolution of galaxies , systematic searches for galaxies in the early universe and detailed studies of the detected galaxies are necessary . in this decade , a number of galaxies detected at high redshift ( @xmath9 ) has been rapidly increasing .
the use of photometric redshift selection from deep broad - band images , especially so - called `` lyman break method '' ( e.g. , @xcite ; steidel , pettini & hamilton 1995 ) gives the largest sample of galaxies at @xmath7 ( lyman break galaxies ; lbgs , e.g. , * ? ? ?
extensive studies of them have been revealing individual and statistical nature of star - forming galaxies at @xmath7 ( e.g. , @xcite ; papovich , dickinson , & ferguson 2001 ) . at the same time , their optical follow - up spectroscopy has also been made extensively ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and revealed rest - frame uv spectral features of lbgs in addition to their redshift information . using about 800 spectra of lbgs at @xmath7 , @xcite found that about three fourth of them show significant ly@xmath3 emission line and the remainders show only ly@xmath3 absorption .
the lbgs with weaker ly@xmath3 emission have stronger low - ionization interstellar metal absorption lines and redder uv continua .
these trend suggest that the lbgs at @xmath7 with weak ly@xmath3 emission are more metal enriched and dusty , i.e. , chemically evolved than those with strong ly@xmath3 emission .
how about properties of lbgs at higher redshift ?
are there any signs of evolution of lbgs compared with lbgs at @xmath7 ? to answer the question and obtain clues to understand formation and evolution of galaxies in the early universe , we made a systematic search for lbgs at @xmath0 ( @xmath101gyr earlier to @xmath11 ) .
we carried out wide ( effectively @xmath12 arcmin@xmath13 ) and deep broad - band ( @xmath14 and @xmath15 ) imaging observations toward an area centered on the hubble deep field - north ( hdf - n ; * ? ? ?
* ) with suprime - cam @xcite attached to the subaru telescope .
thanks to plenty of redshift information of galaxies in and around the hdf - n , we could set suitable color criteria on the two - color ( @xmath16 and @xmath17 ) diagram to effectively select galaxies at @xmath18 avoiding foreground contamination , and obtained @xmath19 lbg candidates at @xmath0 with @xmath20 mag @xmath21 mag @xcite .
this is one of the largest lbg samples at @xmath0 systematically surveyed . using this sample
, we statistically derived a rest - frame uv luminosity function ( uvlf ) of lbgs at @xmath0 , which shows only a slight change from @xmath0 to @xmath7 .
we also found that a spatial distribution of lbgs at @xmath0 shows a large void - like structure and a larger clustering amplitude than that at @xmath7 ( in preparation ) .
the presence of a similar void - like structure is seen in @xcite , and a large clustering amplitude of lbgs at @xmath0 may also suggest the presence of the larger bias in galaxy formation at @xmath0 than at @xmath7 @xcite . in order to confirm redshifts of the lbg candidates and to study spectral features and relations with other properties of them , we started follow - up optical spectroscopy of our lbg candidates at @xmath0 .
although similar systematic searches for @xmath22 galaxies have been carried out ( e.g. , @xcite ; yan , windhorst , & cohen 2003 ; stanway , bunker , & mcmahon 2003 ) , a number of spectra of lbgs at @xmath0 is still extremely small .
moreover most of the spectra obtained so far show only ly@xmath3 emission , and the continuum spectral feature and relations with other properties observed in lbgs at @xmath7 are still quite unknown . in this paper , we present initial results of the optical spectroscopy of a part of our sample .
observations and data reduction are described in section 2 , and the results are presented in section 3 . in section 4 , we compare the results with spectral features obtained for lbgs at @xmath7 , and discuss possible origins of the similarity and difference seen in spectral features .
the magnitude system is based on ab magnitude unless otherwise noted .
we made optical spectroscopy for a subset of our sample using multi - object - spectroscopy ( mos ) mode of the faint object camera and spectrograph ( focas ; * ? ? ?
* ) attached to the 8.2-m subaru telescope .
main spectroscopic targets are our lbg candidates brighter than @xmath23 mag .
since a mask plate of focas mos covers a 6@xmath24 aperture diameter field of view , we designed mos masks to contain as many main targets as possible on each mos field .
the number of targets thus selected is 17 .
in addition to them , we selected 7 objects with @xmath25 mag which lie out of the color selection window but near the border line of the window in order to examine our color selection criteria .
we also included fainter lbg candidates ( @xmath26 mag ) as many as possible in each mask .
spectroscopic observations were made on 2003 february 24 - 26 under clear sky conditions .
we used the grism of 300 lines / mm blazed at 7500 and the so58 order cut filter .
the wavelength coverage was from 5800 to 10000 ( depending on a slit position on a mask ) with a pixel scale of 1.34 .
the slit lengths were typically 10@xmath27 , and the slit widths were fixed to be 0.@xmath278 , giving a spectral resolution of r@xmath10500 .
one ccd pixel covered 0.@xmath281 and a spatial sampling was 0.@xmath273 pixel@xmath29 by on - chip three pixel binning .
an exposure time of each frame was 0.5 hours , and a total exposure time was 5.5 hours for two masks ( mask name f01 and f02 ) , and 5.0 hours for one mask ( mask name f08 ) .
we nodded the telescope with @xmath101@xmath28 along the slit length for each exposure .
spectrophotometric standard stars feige 34 and feige 66 were observed with long - slit mode ( a slit width of 2@xmath28 ) for sensitivity correction .
seeing during the observing runs was 0.@xmath308 .
the data were reduced with standard procedure using iraf .
bias was subtracted by using the overscan region and bias frames , and flat - fielding was made by normalizing averaged dome flat images .
the spectra for each target were then carefully aligned and combined for each night .
wavelength calibration was made using night sky emission lines exposed in the object frame with rms errors of @xmath31 .
sky emission was subtracted by using background task of iraf .
five pixels were binned for a wavelength direction to improve s / n .
one - dimensional spectrum of each object was extracted by using apall task of iraf . since our targets are faint , an aperture for the extraction was determined for each object by eyes to trace the object well .
after sensitivity correction ( including correction for atmospheric a and b bands by tracing the absorption features in the spectra of the standard stars ) was applied to combined spectra for each night , we obtained final spectra by combining them .
figure 1 shows spectra of eight objects identified to be at @xmath0 .
object no.3 ( f01 - 03 ) has already been identified with an agn at @xmath32 by @xcite in a follow - up spectroscopy of 1ms _ chandra _ observation at the hdf - n .
a significant continuum break at the redshifted ly@xmath3 is seen in all the spectra .
the average of the depression factor @xmath33 @xcite for these objects is @xmath34 which is very close to those for qsos at @xmath0 ( e.g. , * ? ? ? * ) and the estimation by @xcite .
the spectra of seven lbgs ( excluding object no.3 ) also show low ionization interstellar ( lis ) metal absorption lines such as @xmath51260 , + @xmath51303 , and @xmath51334 which are prominent features in uv spectra of nearby starburst galaxies ( e.g. , * ? ? ?
* ) and in spectra of lbgs at @xmath7 ( e.g. , * ? ? ?
* ; * ? ? ?
thus they are securely identified to be at @xmath0 .
we determined their redshifts as an average of these two or three lis absorption lines and listed the results in table 1 as well as their coordinates and photometric data .
the obtained redshift range from 4.5 to 5.2 is consistent with that expected from the color selection , figure 7 of @xcite .
we also constructed a composite rest - frame spectrum of the seven lbgs by scaling the continuum level of each spectrum and show the result in figure 2 , together with the composite spectrum of 811 lbgs at @xmath7 @xcite .
intriguingly , spectra of the seven lbgs in figure 1 show no or weak ly@xmath3 emission km@xmath29 ) and strong ( ew@xmath35 ) ly@xmath3 emission . ] .
four objects show no significant ly@xmath3 emission line , and the other three objects , no.1 , no.6 , and no.7 show a ly@xmath3 emission line , but their rest - frame equivalent widths are rather weak ; ew@xmath36 , @xmath37 , and @xmath38 , respectively , with an error of 2030% .
the average rest - frame equivalent width of ly@xmath3 of the seven lbgs is 2.5 , and that derived from the composite spectrum is 4.5 line , which seem to result in the slightly larger ly@xmath3 equivalent width for the composite spectrum . ] .
the position of ly@xmath3 for some of them locates in the wavelength region of the atmospheric b band absorption .
we estimate the uncertainty in the correction for the absorption is at most @xmath38% . a mean rest - frame equivalent width of three lis absorption lines ( @xmath51260 , + @xmath51303 , and @xmath51334 ) for each of the seven lbgs is measured to be @xmath39 and
is listed in table 1 .
the average rest - frame equivalent widths of @xmath51260 , + @xmath51303 , and @xmath51334 of the seven lbgs are @xmath40 , @xmath41 , and @xmath42 , respectively , which agree with those measured in the composite spectrum ; @xmath43 , @xmath44 , and @xmath42 , respectively% . ] . some of the lbgs show a high ionization absorption feature of @xmath451393 , 1402 which are also detected in local starburst galaxies and lbgs at @xmath7 .
most of them lie in the wavelength region where the sky emission is strong or an atmospheric absorption is severe , and we did not measure an equivalent width for each object .
the @xmath461393 , 1402 feature can be seen in the composite spectrum , thanks to the improvement of the s / n .
peaks of the ly@xmath3 emission lines are redshifted with respect to the lis lines by @xmath47 km s@xmath29 , @xmath48 km s@xmath29 , and @xmath49 km s@xmath29 , for no.1 , no.6 , and no.7 , respectively . in the composite spectrum , the offset is @xmath50 km s@xmath29 .
the value is almost the same as that of the lbgs at @xmath7 with weak or no ly@xmath3 emission lines ( 630 km s@xmath29 for group 2 by @xcite ) and also fits a relation between the outflow velocity and the equivalent widths of lis for lbgs at @xmath51 by frye , broadhurst , & benitez ( 2002 ) .
we also measured a continuum slope @xmath52 ( @xmath53 ) of the composite spectrum to be @xmath54 with a fitting uncertainty of @xmath55 .
however , since the value is determined in a short wavelength coverage ( @xmath10250 ) and depends on an applied binning scale and a clipping threshold in making the composite spectrum , the uncertainty of @xmath52 would be much larger .
thus we do not discuss the result further . among the remainders of the spectroscopic sample in the color selection window with @xmath25 mag ,
two objects show hints of the presence of a continuum break at the wavelength corresponding to a redshifted ly@xmath3 at @xmath0 and of one or two lis absorption lines which could be identified to be at @xmath0 . however , the s / ns of their spectra are poor and we do not discuss these objects in this paper . for the remaining targets
, we could not find any significant features due to the poor s /
n . a roughly estimated upper limit on the rest - frame ly@xmath3 equivalent width for these objects is @xmath56 .
for the targets with @xmath57 mag , we could not find any significant features due to the poor s /
n . the s / n of the spectrum depends not only on the @xmath58 magnitude , but also on its surface brightness .
the objects identified to be at @xmath0 tend to have the sizes ( fwhm ) comparable to the seeing size and thus have the brighter peaks than those of the other objects having larger sizes .
figure 3 shows the positions of the objects identified to be at @xmath0 ( filled circles ) in the two color diagram ( @xmath59 and @xmath60 ) as well as the unidentified objects with @xmath61 mag ( open triangles ) .
one of the seven lbgs shown in figure 1 is undetected in @xmath62 band ( @xmath63 mag ) , while the remaining six are detected in @xmath62 band within an aperture of @xmath64 .
four among seven targets ( @xmath61 mag ) outside of the selection window are identified to be galactic m stars ( filled pentagons ) .
these results suggest that our selection criteria for objects at @xmath0 is reasonable .
@xcite divided spectroscopically confirmed @xmath10800 lbgs at @xmath7 into four subgroups ( each group consists of @xmath10200 lbgs ) based on rest - frame ly@xmath3 equivalent width , and made composite spectra for each group .
they found that about three fourth of the @xmath10 800 lbgs at @xmath7 show ly@xmath3 emission lines , and about two third of them show strong ly@xmath3 emission ( ew@xmath65 ) .
line by summing up both emission and absorption . in our case , we measured the equivalent width of ly@xmath3 only for the emission part mainly because it is hard to distinguish intrinsic absorption and intergalactic absorption .
thus the value of rest - frame equivalent width of ly@xmath3 by @xcite is the lower limit when it is compared with our equivalent width .
] they also found that lbgs with weaker ly@xmath3 emission have stronger lis absorption lines , redder uv continuum slopes , and larger @xmath66 values .
the weak ly@xmath3 emission , strong lis absorption lines , the red uv color , and the large @xmath66 are considered to originate in dusty environment . in the lbgs with no or weak ly@xmath3 emission
, star formation may occur earlier and may be chemically more evolved than those with strong ly@xmath3 emission . although shapley et al .
( 2001 ) pointed out a possibility that the lbgs with weak ly@xmath3 emission are younger than those with strong emission from the sed fitting , this might be reconciled with the chemically evolved nature if the ages derived by sed fitting are affected by the most recent star formation occurred in the lbgs .
the spectra of our seven lbgs ( expect for one agn ) at @xmath0 and thus the composite spectrum of them is significantly different from the composite spectrum of lbgs at @xmath7 as shown in figure 2 . the ly@xmath3 emission is much weaker , and the continuum depression in the wavelength region shorter than the redshifted ly@xmath3 emission is much larger than those at @xmath7 .
the measured equivalent width of ly@xmath3 emission is @xmath67 at @xmath0 while @xmath68 at @xmath7 .
the lis absorption lines are stronger in the spectrum for @xmath0 than in that for @xmath7 ; measured equivalent widths of @xmath51260 , + @xmath51303 , and @xmath51334 are @xmath43 , @xmath41 , and @xmath42 , respectively for @xmath0 , while @xmath69 , @xmath41 , and @xmath70 , respectively for @xmath7 . however , the spectra of the seven lbgs at @xmath0 fairly resemble to subpopulations of lbgs at @xmath7 ; the composite spectra of lbgs at @xmath7 with no or weak ly@xmath3 emission ( group 1 and group 2 by * ? ? ?
* ) are quite similar to the spectra of our lbgs at @xmath0 .
the average rest - frame equivalent widths of the three lis absorption lines of the seven lbgs at @xmath0 is @xmath43 being very close to @xmath71 for group 1 ( no ly@xmath3 emission ) .
the average value of ew@xmath72(lis ) corresponds to metallicity of 12@xmath73log(o / h)@xmath74 , if we assume that the relation obtained in the local universe by @xcite can be applied to the high redshift lbgs , which is not certain at this moment .
these results suggest that these lbgs at @xmath0 are chemically evolved to some degree .
all our lbgs confirmed to be at @xmath0 show no or weak ly@xmath3 emission with relatively strong lis absorption lines ; a fraction of lbgs with strong ly@xmath3 emission is very small , though the sample size is still small .
we may be witnessing some sign of evolution of lbgs from @xmath0 to @xmath7 .
the lack of strong ly@xmath3 emission as well as the presence of strong lis absorption at @xmath0 are likely to be due to their dusty and chemically evolved environment ( though the escape of ly@xmath3 photons may not be related in a simple way to the metallicity of the galaxy ( e.g. , * ? ? ?
* ) ) and to the presence of more neutral hydrogen in and/or around a galaxy than that at @xmath7
. however , it is worth emphasizing that the lbgs we observed are relatively brighter ones among those at @xmath0 ( @xmath75 mag corresponds to @xmath76 of uvlf at @xmath0 @xcite . ) .
there is a possibility that the strength of ly@xmath3 emission depends on the magnitude ( i.e. , uv continuum ) .
@xcite found for lbgs at @xmath7 that the average uv magnitude is fainter for the lbgs with stronger ly@xmath3 emission .
they also found that the average rest - frame equivalent width of the ly@xmath3 emission line of faint lbgs is larger than that of bright lbgs among the subgroup with strong ly@xmath3 emission of ew@xmath77 20 .
this trend could also be the case at @xmath0 .
in fact , @xcite , who made a similar search for lbgs at @xmath0 using @xmath78 , @xmath79 , and @xmath80 band deep images , found that all of their spectroscopically confirmed six objects to be at @xmath0 show very strong ( ew@xmath81 ) ly@xmath3 emission lines .
the @xmath79 magnitudes of them are @xmath82 magnitude fainter than those of our seven lbgs at @xmath0 .
in addition , @xcite found that the number density of ly@xmath3 emitters ( laes ) at @xmath83 against to lbgs at @xmath0 rapidly decreases with increasing uv continuum light ( @xmath84 mag ) .
these observational results suggest that the brighter lbgs tend to show the weaker ly@xmath3 emission also at @xmath0 .
it is also worth noting here that the lbgs with secure redshifts tend to show rather compact morphology .
thus the weakness of ly@xmath3 emission and the strong lis absorption lines may also relate to their morphological property , which may also link to an evolutionary stage of galaxies . if we interpret that the weakness of the ly@xmath3 emission is caused by the dusty environment in the lbgs , it might be possible that the brightest lbgs are the most chemically evolved ones at the epoch .
it is known that brighter lbgs have a larger correlation length at @xmath7 @xcite , suggesting that they are associated with more massive dark halos and star formation may occur with biased manner in the earlier epoch as compared with less clustered fainter lbgs .
the similar result has been obtained for lbgs at @xmath85 by @xcite .
we also found the same clustering segregation with magnitude in our sample of lbgs at @xmath0 ( in preparation ) .
thus the brighter lbgs at @xmath0 may be associated with even more massive dark halos and have experienced more biased star formation , resulting in more dusty environment as compared with many of lbgs at @xmath7 with a strong ly@xmath3 emission .
another possible reason of the difference of ly@xmath3 appearance is the effect of a strong clustering of laes . @xcite and @xcite found that there is an overdensity region ( @xmath86 ) of laes ( ew@xmath87 ) at @xmath88 from their wide - field deep narrow - band observations .
the observed field ( 44 arcmin@xmath13 ) of @xcite may happen to fall on an overdensity region of ly @xmath3 emitters ; their rest - frame equivalent widths of ly@xmath3 emission ( @xmath89 ) is larger than the detection threshold of the lae selection . while our field of view for the spectroscopy ( total of @xmath1085 arcmin@xmath13 ) may happen to point to a low density region of laes and we could not observe lbgs with strong ly@xmath3 emission .
however , since we selected spectroscopic targets from regions where the surface density of lbg candidates is relatively high , this may be unlikely provided that the distribution of lbgs broadly coincides with that of laes at the same epoch . to summarize
, the results presented here may show some sign of evolution in spectroscopic feature from @xmath0 to @xmath90 , or the presence of luminosity dependence of nature .
however , our sample size is too small to reach any significant conclusions .
further spectroscopic observations of lbgs at @xmath0 over wider field and magnitude range are necessary to reveal spectroscopic nature and discuss relationship with evolution of lbg population .
we are grateful to the focas team , especially support astronomer dr .
youichi ohyama , for their advises and supports for our observations .
we thank all staffs of subaru telescope .
we also appreciate the referee for the comments which improved this paper .
ii is supported by a research fellowship of the japan society for the promotion of science for young scientists and by a grant - in - aid for the 21st century center of excellence of japan `` center for diversity and universality in physics '' .
this work is partly supported by grants - in - aid for scientific research ( 15540233 ) from the ministry of education , culture , sports , science and technology of japan .
barger , a. j. , et al .
2002 , , 124 , 1839 bertin , e. , & arnouts , s. 1996 , , 117 , 393 bouwens , r. j. , et al .
2003 , , 595 , 589 bremer , m. n. , lehnert , m. d. , waddington , i. , hardcastle , m. j. , boyce , p. j. , & phillipps , s. 2004 , , 347 , l7 coleman , g. d. , wu , c .- c . , & weedman , d. w. 1980 , , 43 , 393 frye , b. , broadhurst , t. , & benitez , n. 2002 , , 568 , 558 giavalisco , m. , & dickinson , m. 2001 , , 550 , 177 heckman , t. m. , robert , c. , leitherer , c. , garnett , d. r. , & van der rydt , f. 1998 , , 503 , 646 iwata , i. , ohta , k. , tamura , n. , ando , m. , wada , s. , watanabe , c. , akiyama , m. , & aoki , k. 2003 , , 55 , 415 kashikawa , n. , et al .
2002 , , 54 , 819 kunth , d. , leitherer , c. , mas - hesse , j. m. , stlin , g. , & petrosian , a. 2003 , , 597 , 263 lehnert , m. d. , & bremer , m. 2003 , , 593 , 630 madau p. 1995
, , 441 , 18 miyazaki , s. , et al .
2002 , , 54 , 833 oke , j. b. , & korycansky , d. g. 1982 , , 255 , 11 ouchi , m. , et al .
2003 , , 582 , 600 ouchi , m. , et al .
2004a , preprint ( astro - ph/0309655 ) ouchi , m. , et al .
2004b , preprint ( astro - ph/0309657 ) papovich , c. , dickinson , m. , & ferguson , h. c. 2001 , , 559 , 620 pettini , m. , shapley , a. e. , steidel , c. c. , cuby , j. g. , dickinson , m. , moorwood , a. f. m. , adelberger , k. l. , & giavalisco , m. 2001 , , 554 , 981 pickles , a.j .
1998 , , 110 , 863 shapley , a. e. , steidel , c. c. , adelberger , k. l. , dickinson , m. , giavalisco , m. , & pettini , m. 2001 , , 562 , 95 shapley , a. e. , steidel , c. c. , pettini , m. , & adelberger , k. l. 2003 , , 588 , 65 shimasaku , k. , et al .
2003 , , 586 , l111 songaila , a. , & cowie , l. l. 2002 , , 123 , 2183 stanway , e. r. , bunker , a. j. , & mcmahon , r. g. 2003 , , 342 , 439 steidel , c. c. , & hamilton , d. 1992 , , 104 , 941 steidel , c. c. , pettini , m. , & hamilton , d. 1995 , , 110 , 2519 steidel , c. c. , giavalisco , m. , dickinson , m. , & adelberger , k. l. 1996a , , 112 , 352 steidel , c. c. , giavalisco , m. , pettini , m. , dickinson , m. , & adelberger , k. l. 1996b , , 462 , l17 steidel , c. c. , adelberger , k. l. , dickinson , m. , giavalisco , m. , pettini , m. , & kellogg , m. 1998 , , 492 , 428 steidel , c. c. , adelberger , k. l. , giavalisco , m. , dickinson , m. , & pettini , m. 1999 , , 519 , 1 steidel , c. c. , adelberger , k. l. , shapley , a. e. , pettini , m. , dickinson , m. , & giavalisco , m. 2003 , , 592 , 728 williams , r. e. , et al .
1996 , aj , 112 , 1335 yan , h. , windhorst , r. a. , & cohen , s. 2003 , , 585 , l93 ccccccccc 1 ( f8 - 2 ) & 12 38 11.2 & + 62 09 19.3 & 24.03 & 2.22 & 0.11 & 4.517 & 1.4 & @xmath91 + 2 ( f2 - 2 ) & 12 37 57.5 & + 62 17 19.2 & 24.16 & 2.40 & 0.07 & 4.695 & 0 & @xmath91 + 3 ( f1 - 3 ) & 12 36 47.9 & + 62 09 41.3 & 24.17 & 3.31 & -0.10 & 5.186 & 44.4 & @xmath92 + 4 ( f1 - 1 ) & 12 37 05.7 & + 62 07 43.1 & 24.62 & 2.26 & 0.15 & 4.650 & 0 & @xmath93 + 5 ( f2 - 6 ) & 12 38 29.0 & + 62 16 18.8 & 24.63 & 1.96 & 0.20 & 4.667 & 0 & @xmath94 + 6 ( f2 - 3 ) & 12 38 25.5 & + 62 18 19.7 & 24.74 & 3.02 & 0.14 & 4.857 & 6.3 & @xmath95 + 7 ( f2 - 8 ) & 12 38 04.4 & + 62 14 19.8 & 24.80 & 3.28 & 0.21 & 5.183 & 9.8 & @xmath44 + 8 ( f2 - 7 ) & 12 38 16.7 & + 62 18 05.5 & 24.83 & @xmath963.19 & 0.42 & 4.615 & 0 & @xmath97 +
|
we report initial results for spectroscopic observations of candidates of lyman break galaxies ( lbgs ) at @xmath0 in a region centered on the hubble deep field - north by using the faint object camera and spectrograph attached to the subaru telescope .
eight objects with @xmath1 mag , including one agn , are confirmed to be at @xmath2 .
the rest - frame uv spectra of seven lbgs commonly show no or weak ly@xmath3 emission line ( rest - frame equivalent width of @xmath4 ) and relatively strong low - ionization interstellar metal absorption lines of @xmath51260 , + @xmath51303 , and @xmath51334 ( mean rest - frame equivalent widths of them are @xmath6 ) .
these properties are significantly different from those of the mean rest - frame uv spectrum of lbgs at @xmath7 , but are quite similar to those of subgroups of lbgs at @xmath7 with no or weak ly@xmath3 emission .
the weakness of ly@xmath3 emission and strong low - ionization interstellar metal absorption lines may indicate that these lbgs at @xmath0 are chemically evolved to some degree and have a dusty environment .
since the fraction of such lbgs at @xmath0 in our sample is larger than that at @xmath7 , we may witness some sign of evolution of lbgs from @xmath0 to @xmath7 , though the present sample size is very small .
it is also possible , however , that the brighter lbgs tend to show no or weak ly@xmath3 emission , because our spectroscopic sample is bright ( brighter than @xmath8 ) among lbgs at @xmath0 .
more observations are required to establish spectroscopic nature of lbgs at @xmath0 .
|
the random field ising model ( rfim ) has been investigated extensively both theoretically and experimentally in last years.@xcite @xmath0 @xcite some questions such as the lower critical dimension and the existence of a static phase transition have been already solved from the theoretical point of view , but others questions as the existence of the tricritical point are still open .
however , it is now understood that rfim describes the essential physics of a rich class of experimentally accessible disordered systems .
these include structural phase transitions in random alloys , commensurate charge - density - wave systems with impurity pinning , binary fluid mixtures in random porous media , and the melting of intercalates i m layered compounds such as tis@xmath1 .
the behavior of the rfim is governed by the competing tendencies of the spins to align ferromagnetically under impetus of the exchange or to follow local fields and so be uncorrelated .
the rfim has been studied with different methods .
the renormalization - group theory ( rg ) of phase transition and critical phenomena represent one of most powerful tool of theoretical physics in last decades .
until then ( early 1970s ) all calculations of critical exponents were made either by exactly or by series expansions .
therefore , employing the renormalization - group arguments ( which can be applied for weak fields ) , the proof of the equivalence between an ising ferromagnet in a random field and a dilute antiferromagnet in a uniform field has been demonstrated .
these question have been very explored ( see [ figueiredo ] and references therein ) . in the rg scheme , such as , the effective field rg ( efrg)@xcite@xmath2 @xcite and the mean - field rg ( mfrg ) @xcite@xmath2 @xcite methods have been successfully employed in spin systems , where the results obtained are in accordance with other more effective approach ( series expansion , monte carlo simulation among others ) . those approaches ( mfrg and efrg )
are based on comparison of two clusters of different sizes , each of them simulating infinite systems . on the other hand ,
the mean field approximation ( mfa ) has been extensively applied in the study of rfim .
the mfa provides very interesting results about rfim .
for example , a bimodal distribution for random field gives rise to a tricritical point and consequently first order phase transitions , at very low temperatures for certain values of field strength , but does not occur in the case of a gaussian distribution.@xcite@xmath0@xcite in addition , it is possible to establish a complete mapping between the parameters of the ising ferromagnet in a random field and the dilute ising antiferromagnet in a uniform field . by employing the same strategy of efrg ,
the effective field theory ( eft ) has been used to study cooperative phenomena and phase transitions in various systems and giving useful qualitative and quantitative insights for the critical behavior of a wide variety of classical and quantum lattice spin systems.@xcite@xmath0@xcite the approach of eft uses , as a starting point , the rigorous callen - suzuki spin identities ( see refs . [ calen63 ] and [ susuki65 ] ) and the effects of the surrounding spins on each cluster are taken into account by using a convenient differential operator technique introduced by honmura and kaneyoshi.@xcite in this procedure , all relevant self - spin correlations are taken exactly into account in the eft equation of state.@xcite therefore , the eft approach is superior to the standard mfa . in this work ,
we study the behavior of phase diagram ( in @xmath3 plane ) for heisenberg classical model of spin-1/2 on a square and simple cubic lattice by employing a bimodal probability distribution for random fields in a cluster with two spins by using the result recently obtained on diluted systems.@xcite the outline of the remainder of this paper is as following : in section * 1 * , the formalism and calculations are developed and the results and conclusions are presented in section * 2*.
the system to be studied is the @xmath4-vectorial model in a random field described by the hamiltonian @xmath5 where the summation is performed over all pairs of the nearest - neighboring sites @xmath6 .
the quantities @xmath7 are isotropically interacting @xmath4 -dimensional classical spins of magnitude @xmath8 localized on site @xmath9 and the cartesian components of @xmath7 obeys the normalization condition,@xcite
@xmath11 [ @xmath12 , @xmath13 is the boltzmann constant and t the temperature ] is the exchange interaction between the spins , @xmath14 [ @xmath15 , where @xmath16 is the bohr magneton and @xmath17 is the random magnetic field ] is the reduced random magnetic field at site @xmath9 with probability distribution @xmath18 .
\label{eqp1}\ ] ] hamiltonian ( [ eq1 ] ) reduces to the @xmath19 ising , planar ( xy ) , heisenberg and spherical models for @xmath20 and @xmath21 , respectively . in this work ,
we follow the eft procedure ( see ref . [ rica_a ] ) to study the critical properties of the hamiltonian described by eq.([eq1 ] ) by employing the axial approximation.@xcite since the hamiltonian for a cluster with two spins can be written as @xmath22 where @xmath23 , ( @xmath24 ) and @xmath25 is the lattice coordination number.the average magnetization per - spin @xmath26 for cluster with two spins is given by ( see refs .
[ doug0a ] and [ doug0b ] ) @xmath27 @xmath28 , @xmath29 , @xmath30 @xmath31 .
@xmath32 ( @xmath33 ) is differential operator which satisfy the mathematical relation @xmath34 where @xmath35 is given by @xmath36 and @xmath37 here @xmath38 is a modified bessel function of the first kind .
( [ meq2 ] ) is exact and will be applied here as a basis of our formalism , since it yields the cluster magnetization @xmath39 and the corresponding multi - spin correlation functions associated with various sites for the cluster under consideration . here
we apply the eft approximation in both sides of eq .
( [ meq2 ] ) , i.e. , the thermal and random average ( denoted by @xmath40 ) , along with the decoupling procedure which ignores all high - order spin correlations , namely @xmath41 , with @xmath42 based on this approximation and replacing each boundary configurational spin average by the symmetry breaking mean - field parameters @xmath43 for all @xmath9 , one set up the equation of state for @xmath44 . by using the properties of differential operator and assuming translational invariance , we obtain for the square ( sq ) and simple cubic lattice ( sc ) , respectivelly , @xmath45 the coefficients @xmath46 [ @xmath47 even ] are zero to satisfy the time reversal symmetry of the ising model , as well the properties of differential operator technique .
in this section we apply the conditions for determining of the second - order transition line and the tricritical point ( tcp ) in the @xmath48 vector model in a random field .
we will focus on the case @xmath49 , which corresponds to the heisenberg classical model of spin-@xmath50 . in the vicinity of the second - order phase transition , the eqs .
( [ meq3 ] - [ meq3a ] ) is given by @xmath51 since the magnetization ( @xmath52 ) goes to zero continuously , to obtain the second order phase transition line , in the @xmath3 plane , we need solve the equations @xmath53 on the other hand , the r.h.s . of eq .
( [ meq5 ] ) must be positive ; otherwise , the transition is interpreted to be of first - order .
unfortunately the first order transition line can not to be analyzed in the frame of the differential operator technique , therefore we have confined our calculations only to the second order transitions , including the tcp .
we obtain the tcp by solving the equations @xmath54 the numerical solution of eq .
( [ meq4 ] ) provide the second - order phase transition line which is shown in figure 1 .
when the random field is zero , we obtain the well - known critical temperature of heisenberg classical model of spin-@xmath50 , @xmath55 and @xmath56 . in figure 1 , the solid line indicate the second - order phase transitions ; the black diamonds denote the position of the tcp at which phase transition changes from second to first order .
the numerical solution of equations ( [ meq6 ] ) provide the values for the tricritical points @xmath57 and @xmath58 for the square and simple cubic lattice , respectivelly .
we might conclude saying that eft approximation is able to predict the presence of tricritical behavior for the heisenberg classical model of spin-@xmath50 on a square and simple cubic lattice in a random magnetic field which takes on the random values @xmath59 with equal probabilities . as can be seen ,
our results show the existence of the tcp for square lattice ( @xmath60 ) in accordance with mean - field@xcite and betthe - peierls@xcite approximations .
this occur because to obtain the first - order phase transition we must have at least the @xmath61 term in the expansion of eqs.([meq3 ] - [ meq3a ] ) .
therefore , in the presente approach we can not observe behavior tricritical for lattice with coordination numbers @xmath62 .
|
by using the differential operator technique and the effective field theory scheme we study the tricritical behavior of heisenberg classical model of spin-1/2 in a random field .
the phase diagram in the _ t - h _ plane on a square and simple cubic lattice for a cluster with two spins is obtained when the random field is bimodal distributed .
71.28+d , 71.30+h , 72.10-d . + _ keywords : _ heisenberg model ; effective field ; random field .
|
the protection of photosensitive optical components from high incident radiation has applications to areas as diverse as microwave and optical communications to optical sensing @xcite . as a result
, a considerable research effort has focused on developing novel protection schemes and materials that provide control of high - level optical and microwave radiation and prevent damages of optical sensors ( including the human eye ) and microwave antennas @xcite .
optical limiters constitute an important class of such protection devices . they are supposed to transmit low - level radiation , while blocking light pulses with high level of radiation .
a typical passive optical limiter absorbs most of the high - level radiation , which can cause its destruction via overheating .
the most common set - up of a passive optical limiter consists of a single protective layer with complex permittivity @xmath0 , where the imaginary part @xmath1 increases sharply with the radiation level .
for low - level radiation , the absorption is negligible , and the protective layer is transparent .
an increase in the radiation level results in an increase in @xmath1 , which renders the protective layer opaque . as a consequence ,
most of the high - level radiation will be absorbed by the limiter , which can cause its overheating and destruction .
it turns out that if the same protective layer is incorporated into a certain photonic layered structure , the entire multilayer can become highly reflective for high - level radiation , while remaining transmissive at certain frequencies if the radiation level is low .
such a reflective limiter can be immune to overheating and destruction by high - level laser radiation , which is our main objective . the physical reasons for the sharp increase in @xmath1 with the radiation level can be different . for instance , it can be photoconductivity , heating , two - photon absorption , or any combination of the these mechanisms . in our previous publication
@xcite we considered the particular case of a strong non - linear dependence of @xmath2 of the protective layer on light intensity .
this can be attributed , for instance , to a two - photon absorption .
we showed that incorporation of such a nonlinear layer in a properly designed low - loss layered structure makes the entire assembly act as a reflective power limiter . in this paper , we consider a more practical particular case where the increase in @xmath1 is due to heating of the protective layer .
we show that , depending on the pulse duration as compared to the thermal relaxation time , the properly design layered structure incorporating such a protective layer can act as a reflective energy limiter , or as a reflective power limiter .
specifically , for short pulses , such a layered structure acts as an energy limiter , reflecting light pulses carrying excessively high energy . by comparison , for sufficiently long pulses ,
the same structure will act as a power limiter . in either case
, most of the incident radiation will be reflected back to space , even though a stand - alone protective layer would act as an absorptive optical limiter .
the proposed architecture consists of a ( protective ) defect layer embedded in a low - loss bragg grating . in contrast to the reflective power limiter introduced in @xcite , the defect layer does not have to be nonlinear , but it must display strong temperature dependence @xmath3 of the imaginary part of its permittivity . if the total energy carried by the pulse is low , @xmath3 remains small enough to support a localized mode and the resonant transmittance associated with this mode . if , on the other hand , the energy carried by the pulse exceeds certain level , the defect layer becomes lossy enough to suppress the localized mode , along with the resonant transmittance .
the entire stack turns highly reflective , which is consistent with our goal .
we refer to this limiter as a _ reflective energy limiter _ in order to distinguish it from the nonlinear reflective power limiter introduced in @xcite .
finally , if the pulse duration significantly exceeds the thermal relaxation time of the defect layer , the entire layered structure will again act as a reflective power limiter with the cut - off light intensity determined by the thermal relaxation time of the defect layer not by the nonlinearity in @xmath1 , as was the case in @xcite .
the organization of the paper is as follows . in sec .
[ model ] , a conceptual design for the reflective energy limiter is presented , along with the mathematical formalism used in our calculations . in sec .
[ theory ] , we analyze the role of thermal conductivity .
the latter plays an important role if the pulse duration is comparable or exceeds the thermal relaxation time of the defect layer . in the defect layer is an increasing function of temperature . ( a ) field distribution at the frequency of resonance transmission for an incident pulse with low energy
the field amplitude at the location of the defect layer is exponentially higher than that of the incident wave .
( b ) transmittance vs. light wavelength for low incident light energy .
( c ) field distribution at the frequency of maximum transmittance for an incident pulse with high energy the amplitude of the suppressed localized mode is lower than that of the incident wave . ( d ) transmittance vs. wavelength for an incident pulse with high energy . ]
we consider two identical losses bragg reflectors consisting of two alternating layers .
each mirror consists of forty layers which are placed at @xmath4 and @xmath5 . for the sake of the discussion
we assume that the layers consist of al@xmath6o@xmath7 and sio@xmath6 with corresponding permittivities @xmath8 and @xmath9 .
these values are typical for these materials at wavelengths @xmath10 . the width of layers is assumed to be @xmath11 and @xmath12 respectively . at @xmath13 we introduce a defect lossy layer with complex permittivity @xmath14 .
we further assume that the imaginary part of the permittivity of the defect layer depends on the temperature @xmath15 i.e. @xmath16 . for simplicity , we assume linear dependence i.e. @xmath17 where @xmath18 are some characteristic constants of the defect . below we assume that @xmath19 ( which is a typical value for , say gaas , at near infrared ) , @xmath20 and @xmath21 while the width of the defect layer is taken to be @xmath22 .
the transmittance @xmath23 , reflectance @xmath24 and absorption @xmath25 of our set up , and the field profile at any frequency can be calculated via the transfer matrix approach .
specifically , a monochromatic electric field of frequency @xmath26 satisfies the helmholtz equation : @xmath27 at each layer inside the grating , eq .
( [ helm ] ) admits the solution @xmath28 , where @xmath29 is the refraction index of the @xmath30-th layer and @xmath31 is the wave vector @xmath32 ( @xmath33 is the speed of light in the vacuum and @xmath34 is the refractive index of air ) . imposing continuity of the field and its derivative at each layer interface , as well as taking into consideration the free propagation in each layer , we get the following iteration relation @xmath35 where @xmath36 at the same time the field outside the layered structured can be written as @xmath37 for @xmath38 and @xmath39 for @xmath40
. the amplitudes of forward and backward propagating waves on the left @xmath38 and right @xmath40 domains are related via the total transfer matrix @xmath41 ( where @xmath42 is the number of layers on each grating and @xmath43 ) : @xmath44 the transmission and reflection coefficients and the field profile , say for a left incident wave , can be obtained by iterating backwards eqs .
( [ transfer0],[transfer ] ) together with the boundary conditions @xmath45 and @xmath46 ( due to the linearity of the equations , one can always impose a value for the outgoing field and calculate via a backward iteration of the transfer matrices the input field to which corresponds @xcite ) .
specifically we have @xmath47 ; @xmath48 .
the absorption coefficient @xmath25 can then be evaluated in terms of transmittances and reflectances as @xmath49 .
in the case that the permittivity of the defect layer is replaced by @xmath50 , the whole structure is periodic and displays a typical dispersion relation consisting of transparent frequency windows ( bands ) where light is transmitted with near - unity transmittance alternated with frequency windows ( gaps ) where the incident light is experiencing almost complete reflection . when the defect is included in the middle of the grating , for zero temperature @xmath51 corresponding to permittivity @xmath52 , the layered structure supports a localized resonant defect mode ( see fig .
[ fig1]a ) with a frequency lying in a photonic band gap of the bragg grating ( see figs .
[ fig1]b ) . for the specific set up that we consider here , we find that a resonant mode is located in the middle of the gap at wavelength @xmath53 . this defect mode is localized in the vicinity of the defect layer and decays exponentially away from the defect ( see fig . [ fig1]a ) . in the vicinity of the localized mode frequency @xmath54
, the entire layered structure displays a strong resonant transmission due to the excitation of the localized mode ( see fig .
[ fig1]b ) .
in other words , the transmittance is @xmath55 while the reflectance and the absorption in the absence of any losses are @xmath56 and @xmath57 respectively .
this picture is still applicable even in the presence of small ( but non - zero ) dissipative permittivity @xmath58 ( see fig .
[ fig1]a , b ) .
an alternative expression for the absorption coefficient @xmath25 can be given in terms of the permittivity and the field intensity @xmath59 inside the defect layer .
the resulting expression is derived by subtracting the product of eq .
( [ helm ] ) with @xmath60 from its complex conjugate form and then integrating the outcome over the interval @xmath61 .
we get @xmath62 substituting in eq .
( [ a_wf ] ) the expressions of the electric field at @xmath63 and @xmath64 respectively we get @xmath65 furthermore we assume that @xmath66 is zero everywhere inside the layered structure apart from the interval @xmath13 where the defect layer is placed . in this interval
it takes a uniform value @xmath67 .
these simplifications allow us to express the absorption coefficient of eq .
( [ asfin ] ) in the following form @xmath68 where @xmath69 is the ratio of the integral of light intensity @xmath70 at the lossy layer and the incident light intensity .
it is obvious from eq .
( [ afin ] ) that @xmath71 depends on both the dissipative part of the permittivity and the value of the electric field inside the defect layer .
although the former increases monotonically with the temperature @xmath15 and thus with the duration time of the incident pulse , this is not true for @xmath72 .
the latter , which is a unique function of the permittivity , remains approximately constant up to some value of @xmath73 above which it decreases , leading eventually to a total decrease of the absorption coefficient together with a simultaneous increase of the reflectivity of the structure .
this is related to the fact that the increase of @xmath73 spoils the resonant localized mode ( see fig .
[ fig1]c ) which is responsible for high transmittance .
specifically , when the losses due to @xmath73 overrun the losses due to leakage from the boundaries of the structure , the resonant mode cease to exist ( see fig .
[ fig1]c ) and the structure becomes reflective , i.e. @xmath74 , and @xmath75 , see fig .
[ fig1]d . as a consequence
we have that @xmath76 and the system does not absorb the high incident energy of the incoming light source but rather reflects it back in space .
in fact , the non - monotonic shape of the envelope of the scattering field in fig .
[ fig1]c is a direct consequence of the fact that the structure becomes reflective @xmath77 .
one has to realize that in the case that both bragg gratings on the left and right of the defect layer are finite , the field inside each half - space is written as a linear combination of two evanescent contributions with exponentially decreasing and exponentially increasing amplitudes .
their relative weight is determined by the boundary conditions @xmath78 and @xmath79 at the two outer interfaces of the layered structure
. in the case of reflective structures these boundary conditions lead to the relation @xmath80 and @xmath81 . it can be shown rigorously that in this case , the field on the left half - space of the structure is dominated originally by the exponentially decaying component while after some turning point @xmath82 the exponentially increasing component becomes dominant up to the defect layer .
after that the field decays exponentially as in the resonant case .
similar scattering field profiles have been found in cases of active ( gain ) defects @xcite .
one can use a simple qualitative argument that allows to estimate the condition under which @xmath71 continues to increase . as we discuss previously
, we assume that the electromagnetic energy losses occur in the lossy defect layer .
the dissipated power can be estimated from eq .
( [ afin ] ) to be @xmath83 .
due to the energy conservation , the rate of energy dissipation can not exceed the energy supply provided by the incident wave .
the latter is @xmath84 . taking this constraint into account we get the following upper limit on the field intensity at the defect layer location @xmath85 above
we have made the additional approximation that @xmath86 , where @xmath87 is a typical value of the field inside the defect layer .
next we recall that a resonant mode with a frequency @xmath26 inside the band - gap has a bloch wave number which is imaginary @xmath88 .
the electric field inside the layered structure , can be expressed as a pair of evanescent modes , one of which is decaying with the distance @xmath89 and another one which is growing i.e. @xmath90 . to the left of the defect ( @xmath91 ) , the electric field
is dominated by the rising evanescent mode @xmath92 while to the right of the defect ( @xmath93 ) , the dominant contribution is provided by the decaying mode @xmath94 @xcite .
the field @xmath87 at the location of the defect layer is provided by the rising evanescent mode evaluated at @xmath95 i.e. @xmath96 .
therefore , the value of this evanescent mode at the left stack boundary at @xmath63 is @xmath97 comparing ( [ ineq ] ) and ( [ equal ] ) we can conclude that if @xmath98 then the amplitude of rising evanescent mode @xmath99 at the left stack boundary is much less than amplitude of the incident wave @xmath100 the latter condition eq .
( [ inequal ] ) , implies that the energy density inside the left grating is much smaller than the energy density of the incident wave , hence , only a small portion of the incident light energy @xmath101 will cross the stack boundary at @xmath63 . in other word , the condition eq .
( [ ineq2 ] ) automatically implies high reflectivity at the stack interface
. the condition eq .
( [ ineq2 ] ) for high stack reflectivity ( and hence low transmittance and absorption ) will always be satisfied if the loss tangent @xmath102 of the defect layer is large enough and/or if the number of layers in the bragg grating is large enough .
of permittivity as a function of pulse duration @xmath103 .
the solid line corresponds to the layered structure in fig .
[ fig1 ] , while the dashed line corresponds to the stand - alone lossy layer .
( b ) the absorption @xmath104 ( black solid line ) , reflectance @xmath105 ( red solid lines ) and transmission @xmath106 ( blue solid line ) of the layered structure in fig .
[ fig1 ] vs. pulse duration . for longer pulse duration ( and larger cumulative energy of the pulse ) , the absorption @xmath25 is suppressed and the set - up becomes highly reflective ( @xmath107 ) .
the dashed lines show the respective values for the stand - alone lossy layer , in which case , the absorption for pulses with longer duration ( and larger cumulative energy ) is much higher , while the reflectivity is much lower than those of the layered structure in fig .
[ fig1 ] . ] but now in the presence of thermal exchange between the defect layer and its surroundings ( @xmath108 ) . for longer pulse duration , a steady state regime
is reached , which corresponds to a crossover from energy limiting regime to a power limiting regime . ]
next , we want to quantify the above arguments . to this end
, we calculate explicitly the transport characteristics of our grating structure for an incident laser pulse .
although the analysis can be generalized for any incident pulse shape , in our numerical simulations below , we have assumed for simplicity that the incident laser pulse has a train - form @xcite @xmath109 we want to calculate the total energy transmitted , reflected , and absorbed during the duration of the pulse .
these can be expressed in terms of the time - dependent transmittance @xmath110 , reflectance @xmath111 and absorption @xmath112 which are the main quantities that we analyze below .
all other observables can be easily deduced from them .
for example , the integrated ( over the period of the pulse ) absorption @xmath113 can be defined as @xmath114 while similar expressions can be used for calculating the total ( over the period of the pulse ) transmission @xmath115 , and reflection @xmath116 .
our starting point is the `` rate '' equation @xmath117 that describes the heating rate of the defect layer . above
, @xmath118 is the heat capacity , @xmath119 is the incident light intensity , and @xmath120 is the thermal conductance of the defect layer . the first term in eq .
( [ heat1 ] ) describes the heating process of the lossy layer while the second one corresponds to heat dissipation from the defect layer to the mirror ( if any ) or to the air . to further simplify our calculations , we assume that the temperature changes are within a domain where both thermal conductance and heat capacity are constants and independent of temperature changes .
substitution of the absorption coefficient from eq .
( [ afin ] ) into eq .
( [ heat1 ] ) leads us to the following equation @xmath121 which expresses the temporal behavior of the temperature @xmath122 in terms of the given profile @xmath123 of the incident pulse .
everything else , e.g. @xmath124 , @xmath112 , @xmath110 and @xmath111 , can be directly and explicitly expressed in terms of @xmath122 . in case that @xmath125 , one can further show that the outcomes can be written in terms of the total incident energy @xmath126 .
furthermore , using eq .
( [ heat4 ] ) we get that @xmath127 .
the associated total absorption is @xmath128 , while similar expressions can be derived for the other transport characteristics . in fig .
[ fig2 ] we report the outcomes of a direct integration of eq .
( [ heat4 ] ) for the case of @xmath125 . in fig .
[ fig2]a we report the temporal behavior of permittivity @xmath129 as a function of the pulse duration @xmath130 . notice that for train pulses the pulse duration @xmath130 is directly analogous of the total incident energy @xmath131 .
we will therefore alternate , in our presentation below , the dependence of @xmath132 , @xmath133 from the pulse duration with the ( more natural parameter for an energy limiter ) total incident energy of the pulse .
originally @xmath129 is essentially unaffected by the incident energy and the same is true for the resonance mechanism ( via the defect mode ) that is responsible for high transmittance in the absence of losses . in this domain @xmath134 ,
@xmath56 while there is a slow increase of the absorption @xmath25 , as it can be seen from fig .
[ fig2]b ( solid lines ) .
once the incident energy ( pulse duration time ) exceeds some critical value , there is a rather abrupt increase in @xmath135 which results to the destruction of the resonance mode .
subsequently , the incident energy does not resonate into the structure , leading to a decaying absorption @xmath57 , while the same is true for the transmittance @xmath136 . at the same time
, there is a noticeable growth of the reflectance which becomes approximately equal to unity @xmath107 . for comparison we also plot at the same figure the results of the stand - alone layer .
we find that for large incident energies ( pulse durations
@xmath130 ) the absorption @xmath112 is higher by more than two orders of magnitude as compared to the case of reflective energy limiter .
we have also performed the same analysis for the case where the thermal conductance @xmath137 is different from zero . in fig .
[ fig3 ] we report the results of the numerical integration of eq .
( [ heat4 ] ) in the presence of thermal conductivity . for long pulse duration
we find a steady state behavior of the transport characteristics of the reflective energy limiter .
the physical nature of the steady - state regime is quite obvious .
it corresponds to the situation when the heat released in the defect layer is completely carried away by thermal conductivity . at this point ,
the temperature of the defect layer stabilizes and the time derivative @xmath138 in eqs .
( [ heat1],[heat4 ] ) vanishes .
the latter condition determines the steady - state values of the defect layer temperature as a function of the incident light amplitude . in this limiting case
our structure acts as a power limiter . for comparison ,
the results of the stand - alone lossy layer are also reported in this figure .
we find that in the steady - state regime our structure performs superbly resulting in absorption values which are more than two orders of magnitude smaller than the onces achieved by the stand - alone lossy layer .
at infrared and optical frequencies , the reflectivity of known uniform materials is well below 90% , especially so when the incident light intensity is dangerously high .
so , if we want to build a highly reflective optical limiter , we have to rely on photonic structures which would support some kind of low - intensity resonant transmission via slow or localized modes at photonic band - gap frequencies .
if the incident light intensity increases , the respective localized mode must disappear , and the entire photonic structure will behave as a simple bragg reflector . here
we considered the so - called `` dissipative '' mechanism of the localized mode suppression . at first glance
, it seems counterintuitive , because the high reflectivity and low absorption are caused by the increase in the loss tangent of the defect layer in fig .
a qualitative explanation for such a phenomenon is that the large value of @xmath1 in the defect layer results in decoupling of the left and the right bragg reflectors in fig .
of course , there might be other ways to suppress resonant transmittance when the incident light intensity , or the total energy of the pulse , grow dangerously high .
still , the presented `` dissipative '' mechanism seems simple and practical .
we stress that in the case that both bragg gratings on the left and right of the defect layer are finite , both evanescent contributions are present in either half - space , although , only one if them is dominant on either side .
furthermore one can show that the presence of both evanescent contributions on either side of the defect layer can provide an energy flux and , hence , a non - zero transmittance .
|
optical limiters transmit low - level radiation while blocking electromagnetic pulses with excessively high energy ( energy limiters ) or with excessively high peak intensity ( power limiters ) .
a typical optical limiter absorbs most of the high - level radiation which can cause its destruction via overheating .
here we introduce the novel concept of a reflective energy limiter which blocks electromagnetic pulses with excessively high total energy by reflecting them back to space , rather than absorbing them . the idea is to use a defect layer with temperature dependent loss tangent embedded in a low - loss photonic structure .
the low energy pulses with central frequency close to that of the localized defect mode will pass through .
but if the cumulative energy carried by the pulse exceeds certain level , the entire photonic structure reflects the incident light ( and does not absorb it ! ) for a broad frequency window .
the underlying physical mechanism is based on self - regulated impedance mismatch which increases dramatically with the cumulative energy carried by the pulse .
|
shape is a fundamental property of an object that influences its interaction with the environment and often determines the object s functional capabilities .
understanding how to generate and control shape by modifying the environmental conditions is of primary importance in designing systems that respond to external clues .
we show here that electrostatic interactions can be used to change the equilibrium shape of soft , nanometer - sized shells .
we find that a uniformly - charged , spherical shell undergoes shape changes , transforming into ellipsoids , discs , and bowls , as the electrolyte concentration in the environment is decreased .
this electrostatics - based shape design mechanism , regulated by varying properties external to the shell , can be used to build efficient nanocontainers for various medical and technological applications .
+ iological matter in cells is often compartmentalized by elastic membranes that take various shapes such as blood cell membranes , organelles and viral capsids .
these biomembranes are highly optimized to perform specific functions .
a key focus of current biomedical technologies is to engineer synthetic materials that can match the performance and structural sophistication displayed by natural entities .
mimicking key physical features of biomembranes , including shape , size , and flexibility , is a crucial step towards the design of such synthetic biomaterials @xcite .
recent findings also indicate that the shape of a drug - carrier nanoparticle directly influences the amount and efficiency of drug delivery @xcite .
the shape and deformability of soft materials such as colloids , emulsions , hydrogels or micelles play an important role in determining their usefulness in various technological applications as well @xcite .
for example , colloidal self - assembly is governed to a large extent by the shape of individual colloids @xcite .
similarly , controlling the shape and size of reverse micelles is of key importance in their use as solvent extraction systems for removing rare - earth metals from aqueous solutions or as templates for nanoparticle synthesis @xcite .
shape transformations in materials are engineered via chemically - induced modifications @xcite or using techniques such as photoswitching of membrane properties @xcite and controlled evaporation of the enclosed solvent @xcite . however , generating desired material shapes with precision and manipulating them with relative ease at the nanoscale has been a challenge @xcite . from the theoretical standpoint ,
much attention has been focused on finding the low - energy conformations of flexible materials , modeled often as soft elastic membranes , in the hope of suggesting superior experimental systems that can enable the design of nanostructures @xcite .
examples include the exploration of shape transitions driven by topological defects @xcite or compression @xcite , and the study of low - energy conformations of multicomponent shells @xcite .
changing the shape of an elastic shell entails bending and stretching it and the associated energy costs form the components of the elastic free energy of the shell @xcite .
however , when the shell is charged , it is possible to compensate for the increase in elastic energy associated with the shape deformation if the latter is accompanied with a significant lowering of the electrostatic free energy @xcite .
previous studies on charged , soft membranes mainly focused on mapping a charged elastic shell to an uncharged elastic shell with charge - renormalized elastic parameters @xcite . in the case of charged nanoshells , electrostatic screening length is comparable to the shell dimensions and the surface charge density can assume high values . as a result , shell models where coulomb interactions are included explicitly are needed @xcite .
using such models , it has been shown that an ionic shell , where positive and negative charges populate the surface , lowers its energy by taking an icosahedral shape with the same surface area @xcite . in this work
, we find that a uniformly - charged , spherical elastic shell , when constrained to maintain the enclosed volume , can lower its free energy by deforming into smooth structures such as ellipsoids , disks , and bowls ( see fig .
[ fig1 ] ) .
we show that the transition to these nonspherical shapes can be driven by varying environmental properties such as the electrolyte concentration in the surrounding solvent . in order to include the non - linear coupling between the shape of the shell and its electrostatic response self - consistently
, we study the soft , charged nanoshells numerically .
we model the charged shell by a set of discrete points placed on a spherical membrane , forming a mesh consisting of vertices , edges and faces , recognizing that in the limit of large number of vertices , the discretized elastic membrane recovers the physics of the associated continuum model ( see _ materials and methods _ for details ) .
the uniform surface charge density is simulated by assigning every vertex with the same charge .
we work with elastic parameters such that the uncharged shell assumes a spherical shape at equilibrium .
we allow only the deformations that preserve the shell s total volume , the latter being chosen to be that of the uncharged conformation .
our model is applicable to monolayers , such as emulsions or reverse micelles where nanodroplets of oil or water are surrounded by properly polymerized charged surfactant molecules , and also to incompressible bilayer systems and nanocontainers that do not exchange material with their environment . in the following sections , we provide evidence that this minimum model reproduces various shapes observed experimentally .
furthermore , we test the validity of this electrostatic model and associated simulation results by providing analytical solutions in limiting cases , namely by computing the electrostatic energy of oblate spheroidal shells and comparing it to that of a sphere of the same volume in salt - free conditions
. effects of ion condensation are then included via a two - state model to derive the renormalized charge on the spherical and spheroidal shells in order to test the robustness of our results . using the discretization of the continuum expression for the elastic energy introduced in ref .
@xcite , we write the free energy @xmath0 associated with the discretized shell as @xmath1 & = & \frac{\kappa}{2 } \sum_{l\in \rm{e } } |\mathbf{n}_{l,1 } - \mathbf{n}_{l,2}|^{2 } + \frac{k}{2r^2 } \sum_{l\in \rm{e } } \left(\left|\mathbf{r}_{l,1 } - \mathbf{r}_{l,2}\right| - a_l\right)^{2 } \nonumber\\ & & + \ ,
\frac{l_{b } z^2}{2 } \sum_{i , j\in \rm{v } } \frac{e^{-|\mathbf{r}_{i } - \mathbf{r}_{j}| /\lambda_{\rm
d}}}{|\mathbf{r}_{i } - \mathbf{r}_{j}|},\end{aligned}\ ] ] where @xmath0 is measured in units of @xmath2 . here
@xmath3 is the room temperature and @xmath4 is the boltzmann constant .
we make the free energy dimensionless by defining @xmath5 , where @xmath6 is proportional to the bending rigidity @xmath7 of the continuum model , and @xmath8 , with @xmath9 being proportional to the 2d young s modulus
@xmath10 of the continuous elastic membrane , and @xmath11 is the spherical shell radius .
we employ the dimensionless bending rigidity @xmath12 and the spring constant @xmath13 as the scale for bending and stretching energies respectively . in eq .
, @xmath14 and @xmath15 denote the set of all edges and vertices respectively , and @xmath16 is the position vector of the @xmath17 vertex .
the first term on the right - hand side is the bending energy with @xmath18 and @xmath19 being the normal vectors to the faces adjacent to edge @xmath20 .
the second term is the stretching energy with @xmath21 and @xmath22 being the position vectors of the vertices corresponding to the edge @xmath20 , and @xmath23 is the rest length of edge @xmath20 .
the last term is the ( dimensionless ) electrostatic energy of the model membrane .
we consider an aqueous environment inhabiting electrolyte whose presence is taken into consideration implicitly , leading to screened coulomb interactions between each vertex pair . here
, @xmath24 denotes the bjerrum length in water , @xmath25 is the debye length and @xmath26 is a dimensionless charge associated with each vertex .
we assume a uniform dielectric in order to simplify the computations , thus ignoring any induced charge effects .
as is evident from eq . , the free energy @xmath0 is a function of the set of vertex position vectors @xmath27 which also parametrizes the shape of the shell .
the equilibrium shape of the shell is the one that corresponds to the minimum of @xmath0 subject to constraint of fixed enclosed volume .
we perform this constrained free - energy minimization using a molecular dynamics ( md ) based simulated annealing procedure , details of which are provided in _ materials and methods_.
the uncharged elastic shell conformation in all our simulations is a sphere of radius @xmath28 nm .
we discretize the sphere with @xmath29 points , generating a nearly - uniform distribution with an average edge length of @xmath30 nm . we fix the elastic spring constant @xmath31 in all simulations .
this value corresponds to @xmath32 @xmath33 per @xmath34 which , for the bending rigidities under investigation , leads to shells characterized by a fppl - von krmn number ( @xmath35 ) in the range @xmath36 .
we consider a monovalent electrolyte with concentration @xmath37 . the debye length @xmath25 is known via the relation @xmath38 nm @xcite .
the concentration @xmath37 thus parametrizes the spatial range of coulomb interactions . in our simulations
, we tune @xmath37 such that this range varies from @xmath39 , in which case the shell mimics the behavior of an uncharged elastic shell , to @xmath40 , which corresponds to the case where most charges feel each other .
snapshots of minimum - energy conformations of charged elastic nanoshells for three different bending rigidities @xmath41 1 , 5 , 10 ( columns , left to right ) . in each column the electrolyte concentration @xmath37 ( m ) decreases ( top to bottom ) as @xmath42 .
different colors suggest different concentration values , with red being the highest @xmath37 under study and purple corresponding to the lowest @xmath37 .
as the concentration is lowered , the range of electrostatic interactions is increased , leading to the variation in the shape of the nanoshell .
we find that for the concentration - range under investigation , softer shells tend to form bowl - like structures , while more rigid vesicles form ellipsoidal and disk - like shapes .
all the above nanostructures have the same total surface charge and volume , fixed to values associated with the spherical conformation . ] in fig . [ fig1 ]
, we show the change in the shape corresponding to the minimum of the shell free energy @xmath0 as @xmath37 is varied . here , we set @xmath43 , which is equivalent to a shell surface potential of @xmath44 mv .
each column represents the shapes obtained for a fixed value of @xmath12 , with the latter increasing from left to right assuming the values @xmath41 1 , 5 , 10 . within each column
, @xmath37 decreases from top to bottom as @xmath45 1 , 0.1 , 0.05 , 0.005 m. this range of concentration covers most biological and synthetic conditions .
we see that the top row ( @xmath46 m ) is comprised of spherical shapes . at @xmath46 m
, the screening length is very small ( @xmath47 ) and hence the electrostatic forces only come into effect at extremely short distances , resulting in a nearly vanishing contribution to the overall free energy .
this leads to conformations that resemble the shape of the uncharged elastic shell which is spherical .
however , as @xmath37 is lowered , transitions to a variety of nonspherical shapes are observed . in case of the most flexible charged shell ( fig .
[ fig1 ] , left column ) , increasing the range of the electrostatic interactions leads to the formation of concave structures , hereafter referred to as bowls .
the opening of the bowl widens with decreasing @xmath37 . for a shell with a higher bending rigidity ( middle column ) , as @xmath37 is lowered ,
the shell first assumes a convex , ellipsoidal shape , then a bi - concave , disc - like structure , and finally the shell deforms into a bowl .
we note the similarity between the bi - concave discs we obtain and the shape of synthetic red blood cells @xcite , despite the differences in their respective physical origins and sizes .
the rightmost column shows the results for the most rigid membrane under study . due to the high energy penalty associated with bending ,
the shell remains spherical even at @xmath48 m. however , upon further lowering of @xmath37 , we first witness an ellipsoidal shape and then a flattened disc - like structure at @xmath49 m. it is worth noting that the discs and bowls we obtain , closely resemble the shapes of elastic structures in ref .
@xcite that are synthesized using light as a tool to engineer shape .
shell shapes that minimize free energy @xmath0 for fixed @xmath50 and @xmath51 m as a function of increasing @xmath52 0.3 , 0.6 , 1 ( from left to right ) . as @xmath26 increases , the strength of the electrostatic interactions increase and the shell transforms from a convex , ellipsoidal form to a dimpled disk and finally to a concave bowl - like structure .
all shapes correspond to the same total volume .
see text for the meaning of symbols . ]
next , we study the effects of modulating the strength of coulomb interactions on shell shape . in fig .
[ fig2 ] we show snapshots of minimum - energy shell conformations when we vary the parameter @xmath26 keeping the flexibility of the shell and the salt concentration in the environment constant ( @xmath50 and @xmath51 m )
. changing @xmath26 corresponds to simulating shells with different total charge on the surface .
the shapes from left to right correspond to the values of @xmath52 0.3 , 0.6 , 1 .
we find that at @xmath53 , the shell assumes a convex ellipsoidal shape .
as @xmath26 is increased to 0.6 , the ellipsoid deforms into a dimpled disk and finally at @xmath54 , the bowl structure is obtained . the transition to nonspherical shapes is accompanied by a decrease in the electrostatic energy . in fig .
[ fig3 ] ( top half ) , we plot @xmath55 , the total coulomb energy of the final structure relative to that of the spherical shell with identical parameters .
the data for @xmath55 is shown as a function of @xmath37 for various values of @xmath56 0.3 , 0.6 , 1 and @xmath57 1 , 10 . in all cases
, @xmath55 is negative . for convex shapes ( spheres and ellipsoids ) , represented by black symbols
, @xmath55 is small . on the other hand , for discs and bowls , represented by blue and red symbols
respectively , the reduction in electrostatic energy is more pronounced . in general , as the concentration @xmath37 is lowered , the behavior of @xmath55 suggests that the spherical shell deforms to an ellipsoid , then to a disk and finally to a bowl .
we find that the nonspherical shapes have a larger surface area relative to the spherical conformation ( see fig .
[ fig3 ] , bottom half ) .
we expect this to be the case as for a given fixed volume , sphere has the lowest surface area .
we find in some cases , the minimum - energy structure has twice the surface area of a sphere with same volume .
though a more general model of the elastic shell would include an energy penalty associated with increasing the surface area , we expect the shape changes to occur in situations where the surface energy increase due to the rise in area is compensated by the adsorption of molecules ( such as neutral surfactants ) to the membrane , thereby reducing its surface tension . using the data in fig .
[ fig3 ] we estimate that the shell surface tension should be low , @xmath58 dyne @xmath59 , for the aforementioned predicted shapes to be realized . in fig .
[ fig4 ] , we show the distribution of local electrostatic and elastic energies on the disc ( top two rows ) and bowl ( bottom two rows ) .
the disc corresponds to the case of @xmath43 , @xmath60 , @xmath49 m and the bowl shape is characterized by @xmath43 , @xmath61 , @xmath48 m. the electrostatic energy at a vertex is computed by summing over the screened coulomb interactions of the charge at that vertex with all other charges on the shell . as the scalebars on the right point out , the electrostatic energy is the dominant of the two energies and drives the shape formation , with the elastic energy adapting locally to conform to the new shape . for both disc and bowl
, the local elastic energy ( second and fourth row ) has large spatial variations and tends to be higher on the more bent regions of the nanoshell . for the disc shape ,
the coulomb energy ( first row ) is higher near the center .
this is , in part , due to the enhanced repulsion resulting from the proximity of the opposite faces which are at a distance less than the debye length associated with this system .
electrostatic contribution to the energy of the shell ( top plot ) and shell s surface area ( bottom plot ) vs salt concentration @xmath37 for different lowest - energy structures .
we plot the electrostatic energy , @xmath62 , which is measured relative to that of a spherical shell with identical parameters .
similarly , the area @xmath63 of the shell is normalized by the area of a sphere with the same volume .
black symbols are spheres or ellipsoids ; blue symbols are discs ; red symbols are bowl - shaped structures .
the inset shows the legend for the symbols used in the plot .
the large ( negative ) changes in coulomb energy help drive the shape transitions . ]
increasing the range or strength of electrostatic interactions enhances the coulomb repulsion between any two charged vertices , making them move apart .
however , the resulting extension in edge lengths is penalized by the rise in the stretching energy . in addition
, the bending energy term penalizes any sharp changes in curvature , thus favoring transitions to smooth shapes .
this competition between the electrostatic and elastic energies sets an effective area for the nanomembrane which in conjunction with the fixed - volume constraint determines the eventual shape of the nanoshell .
varying the screening length or the total charge on the shell changes this effective area , leading to variations in the shell shape .
spatial distribution of electrostatic and elastic energies ( in units of @xmath64 , where @xmath3 is the room temperature ) on the surface of the disc ( top two rows ) and bowl ( bottom two rows ) . left column : front view , middle column : angle view , right column : side view .
for either shapes , the elastic energy ( second and fourth row ) is concentrated in the edges .
the electrostatic energy on the disc ( first row ) is higher in the center where the opposite faces are nearby ( first row ) .
the five - coordinated vertices , which are visible as spots in the electrostatic energy distribution , lead to small fluctuations in the energy . ] to substantiate the above explanation , we focus on the sphere - to - ellipsoid - to - disc part of the observed shape transitions and perform analytical calculations .
judging by the simulation snapshots ( see the images in fig .
[ fig1 ] , rightmost column ) , these shapes can be approximated as oblate spheroids with different degrees of eccentricity @xmath65 and major semiaxis lengths @xmath66 .
since the volume of the shell is fixed , the oblate spheroidal shell can be characterized by a single parameter @xmath65 . for @xmath67 ,
one obtains sphere - like shapes and @xmath68 leads to disc - like conformations .
the competition between elastic and electrostatic energies can now be considered as determining the eccentricity @xmath65 for the oblate spheroid .
the concentration @xmath37 is seen as the control over @xmath65 such that the lowering of @xmath37 can be understood as an increase in @xmath65 .
thus , we can verify the order of shape transitions observed in our simulations by examining the change in the electrostatic energy of a uniformly - charged shell as its eccentricity is increased . for simplicity
, we consider unscreened coulomb interactions in the following calculations .
we evaluate the electrostatic energy @xmath69 of a uniformly - charged oblate spheroidal shell with total surface charge @xmath70 and with volume constrained to @xmath71 ( derivation in _ si text _ ) .
we obtain : @xmath72 where @xmath73 is an even integer , @xmath74 and @xmath75 are legendre polynomials of first and second kind , @xmath76 , and @xmath69 is evaluated relative to the thermal energy at room temperature .
we define @xmath77 as the electrostatic energy of the oblate spheroid relative to the electrostatic energy of the sphere with identical parameters .
we examine the variation of @xmath77 vs @xmath65 for the parameter set associated with the transition recorded in the open blue circles of fig . 3 and find that the coulomb energy of an oblate spheroidal shell subject to the constraint of constant volume decreases with increasing its eccentricity ( see fig . s2 ) .
in other words , a disc - shaped shell has lower coulomb energy than a sphere of the same volume .
the order of shape transitions observed in our simulations is thus backed by the above analytical result .
next , we examine the spatial distribution of the local electrostatic energy @xmath78 on the surface of the shell ( see _ si text _ for details ) .
we find that for a spherical shell , @xmath78 is constant everywhere .
however , as the eccentricity increases , the surface distribution of electrostatic energy becomes increasingly inhomogeneous .
in particular , for @xmath79 , which corresponds to a disc - like shape , we find @xmath78 varies significantly on the disc surface , assuming higher values near the disc center and low magnitude near the edge of the disc ( see fig .
it is evident from the top row of fig .
[ fig4 ] that we observe this trend in our simulation results as well .
we obtain more insight into our results by exploring the low - energy conformations of a very flexible uniformly - charged shell where the elastic energy can be neglected in comparison with the coulomb energy .
equilibrium shapes of such a shell will correspond to the minimum of the total electrostatic energy . in eq .
, taking the limit @xmath80 gives @xmath81 , which is the lowest possible value for the coulomb energy of a uniformly - charged shell .
this limit corresponds to a disc - like spheroidal shell whose area approaches infinity .
further , we check that when the enclosed volume is held fixed , the coulomb energy of a prolate spheroidal shell vanishes as well when the shell is stretched into a long and thin wire - like shape .
thus , we obtain ( at least ) two distinct shell shapes that correspond to the state of lowest electrostatic energy .
this result suggests that in our original model system , electrostatic interactions drive the transformation in the shell shape by favoring the deformation of sphere towards disc - like shapes , while the elastic energies compete with the coulomb energy to generate oblate - shaped ( ellipsoidal , disc - like ) structures of various eccentricities .
it also appears that the elastic energy component of the free energy favors the formation of oblate shapes to prolate ones .
the constraint of fixed enclosed volume is critical to the low - energy shell conformations obtained in our simulations . if instead of the volume
, the shell surface area is fixed , we expect the gallery of lowest free - energy conformations to look different than fig .
we check that under the constraint of fixed area , the coulomb energy of an oblate spheroidal shell is higher when its eccentricity increases , and the spherical shape corresponds to the conformation with the lowest coulomb energy among all oblate shapes . however , sphere is _ not _ the configuration that minimizes the shell electrostatic energy when prolate - shaped deformations are considered .
we find that prolate spheroids of high eccentricities have lower coulomb energy than the sphere and the lowest - energy conformation for the area - constrained system is a prolate spheroidal shell with its major - axis length stretched to infinity .
hence , for the area - constrained problem , we expect the competition between coulomb and elastic energies to give rise to different nonspherical shapes as ground - state solutions . in our charged shell model , we assume that the counterions remain in the bulk and do not condense on the shell surface .
however , in an experiment it is possible that a fraction of the counterions do condense and in that event it becomes important to analyze their effect on the observed shape transitions .
we measure this effect qualitatively in the salt - free limit for the sphere - disc transition by employing the expression for the electrostatic energy of a uniformly - charged oblate shell in a two - state model of free and condensed counterions @xcite .
we consider a spherically - shaped wigner - seitz ( ws ) cell of volume @xmath82 with a single shell of volume @xmath83 and surface charge @xmath70 placed at its center .
we define the quantity @xmath84 as the shell volume fraction .
the cell also contains @xmath85 counterions , each of charge @xmath86 to neutralize the shell charge .
we separate the counterions into two distinct groups : free ions and condensed ions .
free ions occupy the available space in the ws cell which in the dilute limit becomes the volume of the cell .
the condensed counterions are restricted to have translational motion in a thin layer of volume @xmath87 surrounding the shell , where @xmath88 is the surface area of the oblate shell and @xmath89 is the gouy - chapman length that is chosen as the condensed - layer width . here
, @xmath90 is the unrenormalized surface charge density .
when a shell has a higher @xmath90 or the system is characterized by a longer @xmath91 , we expect the condensed - layer width to shrink owing to the enhanced counterion - shell attraction . our choice of @xmath92 as the layer thickness correctly reflects this behavior .
as @xmath92 is a characteristic of the charged planar surface , our analysis is limited to the regime where @xmath92 is much smaller than the lengths of the major and minor semiaxis of the shell .
we write the free energy ( in units of @xmath93 ) associated with the shell in the event of ion condensation as : @xmath94 where @xmath95 is the fraction of counterions that condense and @xmath96 is the thermal de broglie wavelength . here
, the first term is the electrostatic energy of the shell obtained from eq . by replacing @xmath70 with the reduced charge @xmath97 , the next two terms stem from the entropic contribution of the @xmath98 condensed ions , and the last two terms correspond to the entropy of @xmath99 free counterions .
@xmath100 can be considered as a function of two variables : eccentricity @xmath65 , which characterizes the shape of the shell , and condensate fraction @xmath95 , which measures the renormalized charge on the shell . for a given @xmath65
, we find the condensate fraction that extremizes the above free energy @xmath100 .
using @xmath95 , we evaluate the equilibrium free energy difference , @xmath101 , between the free energy of the oblate shell and that of the sphere of the same volume ( see _ si text _ for details ) .
we compute @xmath101 for the parameters associated with the transition recorded in the open blue circles of fig .
[ fig3 ] and find that for all values of the volume fraction @xmath102 , @xmath101 becomes increasingly more negative as the eccentricity @xmath65 is raised , implying that the shape transitions from sphere to oblate spheroids are favored ( see fig .
s4 ) . additionally , we find that the condensate fraction @xmath95 decreases with increasing @xmath65 for all values of @xmath102 . for low @xmath102
, we obtain @xmath103 while for large @xmath102 , we find the condensate fraction to be @xmath104 .
regardless of the amount of condensation , we find the shell with higher eccentricity is preferred energetically .
we next examine the variation of the renormalized electrostatic energy @xmath77 with @xmath65 for different @xmath102 values . for low and high values of the volume fraction , we find that @xmath77 is negative and decreases , just like @xmath101 , upon the increase of the eccentricity ( see fig .
s5 ) . however , for some intermediate @xmath102 values , we observe that @xmath105 , that is the electrostatic energy increases as @xmath65 is raised , in sharp contrast to the free energy associated with the shell .
this suggests that for some values of shell volume fractions , the shape transitions are expected to occur despite an increase in the electrostatic energy .
we attribute the feasibility of such transitions to the gain in entropy by the ions as less number of ions condense when the shape is deformed from a sphere to an oblate .
the main conclusions reached above remain unchanged when we repeat the two - state model analysis assuming that the shell is an equipotential surface . for all values of @xmath102
, @xmath101 increasingly becomes more negative as the eccentricity @xmath65 is raised , implying that the shape transitions from sphere to oblate spheroids are favored ( see fig .
thus , judging by the variation of @xmath101 determined by the above two - state model analysis , we conclude that the shape transitions from sphere to oblates of increasing eccentricity should be feasible in the event of ion condensation .
however , due to the renormalization of the charge on the shell surface , it is likely that the specific parameter values ( for example , concentration strength , bending rigidity ) for which the shape transitions occur , will change .
quantitative results can be obtained by including counterions explicitly in the simulations and taking into account the induced polarization charges on the shell surface in analyzing changes in shape .
we note that our md - based simulation algorithm provides an ideal platform to include these effects via its coupling with recently introduced energy - functional - based approaches of treating dielectric inhomogeneities @xcite .
we investigate the prospects of electrostatics - based generation and control of shapes in materials at the nanoscale . we find that by increasing the strength or the range of coulomb interacting potential , a uniformly - charged spherical shell , constrained to maintain its volume , deforms to structures of lower symmetry , resulting in ellipsoids , discs , and bowls .
this symmetry breaking is accompanied with a reduction in the overall electrostatic energy of the shell and a significant spatial variation in the local elastic energy on the shell surface . to support our simulation findings ,
we show analytically that a uniformly - charged disc - like spheroidal shell has a lower coulomb energy than a spherical shell of the same volume . in order to evaluate the renormalization of shell charge due to non - linear effects
, we use a two - state model of free and condensed ions .
we find that the shape transitions are feasible in the event of ion condensation for a wide range of shell volume fractions .
shape changes in our model membrane are triggered by changing the attributes of the environment external to the membrane , such as the electrolyte concentration in the surrounding solvent .
this is in contrast with transitions brought about by patterning the shell surface with defects @xcite or by introducing elastic inhomogeneities on the shell surface @xcite . in comparison with ionic shells @xcite , where the primary experimental challenge is to synthesize membranes with desired stoichiometric ratios @xcite ,
our base shell surface is uniformly charged and elastically homogeneous , which is relatively simple to design .
we envision that the electrostatics - driven shell design mechanism proposed here can function as a useful template for synthesizing nanoparticle - based drug delivery carriers of desired shapes @xcite .
our results can also prove useful in the analysis of shape changes in charged emulsions or reverse micelle systems that form during the metal - extraction processes involved in the recovery of scarce rare - earth elements or cleaning of nuclear waste @xcite .
in addition , our findings can aid in the development of theories explaining the properties of stretchable electronic materials such as dielectric elastomers where electrostatic field and deformation are intimately coupled @xcite .
we generate the triangulation on the shell via the caspar and klug construction @xcite , which produces a lattice where each point has six neighbors with the exception of 12 five - coordinated vertices ( defects ) . due to the presence of these defects
, the lattice has a non - vanishing initial stretching energy .
we remove this residual strain by appropriately choosing rest lengths of the edges @xcite , leading to a vanishing stretching energy for the initial mesh .
the defects , however , lead to slight variations in the surface charge density and local elastic and electrostatic energies . by choosing a large number of lattice points , the effect of these small deviations on the resulting shape transformations
is minimized . in order to make sure that our results are independent of the particular triangulation
, we perform simulations employing sufficient number of lattice points generated via different choices of caspar and klug constructions , obtaining similar results for all runs . our lattice maintains the initial connectivity throughout the shape evolution .
since each vertex carries a charge of the same sign , our discretized membrane is characterized with an inbuilt self - avoidance due to the mutual electrostatic repulsion between any pair of vertices .
however , to ensure complete stability , we include an additional short - range , purely repulsive lennard - jones potential between two vertices , where each vertex is modeled as a hard sphere with radius chosen to be a fraction of the average edge length @xmath66 associated with the triangular lattice .
we use molecular dynamics ( md ) method to minimize the free energy @xmath106 which requires the analytical expressions for the gradients of @xmath106 with respect to the vertex positions . evaluating the gradient of the bending energy term
is relatively difficult and we show this calculation in the _
si text_. our simulations start from a spherical shell with a nearly homogeneous surface distribution of local elastic and electrostatic energies .
slight deviations in the energies arise from the presence of five - coordinated vertices which are the result of using the aforementioned triangulation of the shell surface .
we assign the vertices a kinetic energy @xmath107 , and direct their motion according to the forces derived from @xmath0 , where the latter plays the role of the potential energy .
we thus obtain the lagrangian @xmath108 , from which we derive the equations of motion for the vertices : @xmath109 . here , @xmath110 is a mass term associated with the vertices which determines the choice of the simulation timestep .
these equations of motion , which form the basis of the md simulation of the vertices , are appropriately augmented to preserve the constraint of fixed total volume .
we achieve this via the shake - rattle routine @xcite of implementing constraints that guarantees the conservation of shell volume at each simulation step . finally , in order to arrive at the shape that corresponds to the minimum of the energy landscape we couple the md scheme with simulated annealing .
we associate a ( fictitious ) temperature with the kinetic energy of the vertices and employ a nose - hoover thermostat to set it .
this temperature is not the physical temperature , it is merely a parameter we employ to control the annealing process .
we reduce this temperature at periodic intervals so as to arrive at the lowest point of the potential energy associated with the md lagrangian , thus reaching the minimum of the free energy @xmath106 .
we thank r. sknepnek , z. yao , g. vernizzi , and j. zwanikken for many insightful discussions .
the model was developed with the financial support of the office of basic energy sciences within the department of energy ( doe ) grant number de - fg02 - 08er46539 .
the computational work was funded by the office of the director of defense research and engineering ( ddr@xmath111e ) and the air force office of scientific research ( afosr ) under award no .
fa9550 - 10 - 1 - 0167 .
shrestha , l. k , sato , t , & aramaki , k. ( 2009 ) intrinsic parameters for structural variation of reverse micelles in nonionic surfactant ( glycerol [ small alpha]-monolaurate)/oil systems : a saxs study . , 42514259 .
vander hoogerstraete , t , wellens , s , verachtert , k , & binnemans , k. ( 2013 ) removal of transition metals from rare earths by solvent extraction with an undiluted phosphonium ionic liquid : separations relevant to rare - earth magnet recycling .
, 919927 .
hamada , t , sugimoto , r , vestergaard , m. c , nagasaki , t , & takagi , m. ( 2010 ) membrane disk and sphere : controllable mesoscopic structures for the capture and release of a targeted object .
, 1052810532 .
|
manipulating the shape of nanoscale objects in a controllable fashion is at the heart of designing materials that act as building blocks for self - assembly or serve as targeted drug delivery carriers .
inducing shape deformations by controlling external parameters is also an important way of designing biomimetic membranes . in this article
, we demonstrate that electrostatics can be used as a tool to manipulate the shape of soft , closed membranes by tuning environmental conditions such as the electrolyte concentration in the medium . using a molecular - dynamics - based simulated annealing procedure ,
we investigate charged elastic shells that do not exchange material with their environment , such as elastic membranes formed in emulsions or synthetic nanocontainers .
we find that by decreasing the salt concentration or increasing the total charge on the shell s surface , the spherical symmetry is broken , leading to the formation of ellipsoids , discs and bowls .
shape changes are accompanied with a significant lowering of the electrostatic energy and a rise in the surface area of the shell . to substantiate our simulation findings ,
we show analytically that a uniformly - charged disc has a lower coulomb energy than a sphere of the same volume .
further , we test the robustness of our results by including the effects of charge renormalization in the analysis of the shape transitions and find the latter to be feasible for a wide range of shell volume fractions .
|
the three classical white dwarfs were , at the beginning of the twentieth century , , , and the . although @xmath3 eri b had been discovered by herschel ( 1785 ) and sirius b had been predicted by bessel ( 1844 ) , it was not until the 1920s when other great astronomers noticed their oddness ( luyten 1922 ) and popularized the term `` white dwarf '' ( eddington 1924 ) .
the extreme physical conditions by which the white dwarfs are supported against gravitational collapse could not be understood until the quantum mechanic was properly developed . since the 1930s ,
the number of known white dwarfs has exponentially increased , from 18 in 1939 ( schatzman 1958 ) and over a hundred in 1950 ( luyten 1950 ) to a few thousands at present ( mccook & sion 1999 ; eisenstein et al . 2006 ) .
see liebert ( 1980 ) , koester ( 2002 ) and hansen & liebert ( 2003 ) for extensive reviews of the general properties of white dwarfs .
the vast majority of the known white dwarfs are very faint , with typical magnitudes in the optical from @xmath4 = 15 to 20mag , or even fainter .
only a few very bright white dwarfs ( @xmath5 12mag ) , including the three classical white dwarfs and , are known .
many of them , especially those that are not in double degenerate systems , are extensively used as spectrophotometric stars ( see , e.g. , the recent catalogue by landolt & uomoto 2007 ) . except for rare exceptions , as the very hot dwarfs and cataclysmic variables found in the _ rosat _ all - sky survey of extreme - ultraviolet sources by pounds et al .
( 1993 ) , all the very bright white dwarfs were discovered during photometric and astrometric surveys before the early 70s ( e.g. kuiper 1941 ; luyten 1949 ; thackeray 1961 ; eggen & greenstein 1965 ; giclas , burnham & thomas 1965 ; schwartz 1972 ) .
afterwards , and in particular with the advent of the sloan digital sky survey , the detection of new white dwarfs and blue subdwarfs has been biased towards magnitudes fainter than @xmath4 = 12mag .
because of that , the photometry - based discovery of a very bright white dwarf candidate 35 years later would be outstanding .
if confirmed , it would yield doubts on the real knowledge that we have of the solar neighbourhood and on the completeness of previous and current surveys for white dwarfs . in this work we present , a previously - unknown very bright ( @xmath6 = [email protected] ) white dwarf candidate .
its finding chart is provided in fig .
[ findingchart ] .
albus 1 was serendipitously discovered during an optical - near infrared photometric study by caballero & solano ( 2007 ) , devoted to characterize the young stars and brown dwarfs surrounding ( @xmath7 ori ) and ( @xmath8 ori ) . as part of this study
, they made a correlation between the tycho-2 ( hg et al .
2000 ) and the 2mass ( cutri et al . 2003 ) catalogues in ten 45arcmin - radius comparison fields at the same galactic latitude of the brightest stars of the young association ( the orion belt ; @xmath9 17.5deg ) .
the total investigated area was only 17.7deg@xmath10 ( @xmath110.04% of the whole sky ) .
albus 1 has a @xmath2 color that clearly deviates from those of the other 1275 investigated sources ( see fig . [ colormagnitude ] ) .
in particular , while the bluest remaining sources have @xmath12 0.3mag , albus 1 has a color @xmath2 = [email protected] .
the tycho-2 @xmath13 and 2mass @xmath14 photometry shows that the object is extremely blue at all the wavelenghts from 0.4 to 2.2@xmath15 m .
given the extreme blueing of albus 1 , we decided to investigate it in detail .
in this work we have taken advantage of the tools offered by the virtual observatory ( vo ; http://www.ivoa.net ) , which is an international , community - based initiative to provide seamless access to the data available from astronomical archive and services .
the vo also aims to provide state - of - the - art tools for the efficient analysis of this huge amount of information .
in particular , we have used aladin ( http://aladin.u-strasbg.fr/aladin.gml ) , a vo - compliant interactive sky atlas developed by cds that allows the user to visualize and analyze astronomical images , spectra and catalogues available from the vo services .
albus 1 has an appreciable tycho-2 proper motion of 19masa@xmath16 . since the comparison fields are relatively close to the antapex ( the point the sun is moving away from ) ,
the foreground objects in this region have not very large proper motions except for some very nearby stars with large tangential velocities .
indeed , the proper motion of albus 1 is in the percentile 23% of the investigated sources ( i.e. 77% of the investigated tycho-2 stars have @xmath17 19masa@xmath16 ) .
this makes albus 1 to be a nearby galactic object .
apart from the coordinates , proper motion and @xmath13 magnitudes from tycho-2 and the @xmath14 magnitudes from 2mass , taken from caballero & solano ( 2007 ) , we have also collected additional photometric and astrometric data from other catalogues : supercosmos science archive ( hambly et al .
2001 ) , usno - b1 ( monet et al . 2003 ) , nomad1 ( zacharias et al . 2005 ) and denis ( denis consortium 2005 ) .
they nicely match between them and tycho-2 and 2mass , except for the fact that for very blue objects the tycho-2 @xmath0 and photographic @xmath18 photometry are not comparable . to avoid superfluous , repetitive information in table [ albus1 ] , we provide only the tycho-2 and 2mass information and the @xmath19 photometric data from usno - b1 .
the search with aladin concluded that there is no radio ( nrao vla ) , mid - infrared ( @xmath20 ) , ultraviolet ( @xmath21 ) , x - ray ( @xmath22 ) source , or object discussed in the literature ( simbad ) at less than 4arcmin to albus 1 .
neither spectroscopic information exists nor photometry in the johnson passbands has been obtained yet .
to ascertain the real nature of albus 1 , we must compare its photometry with that of other very blue objects .
there is a limited number of galactic objects with @xmath2 colors as blue as those of albus 1 : white dwarfs , hot subdwarfs , and early - type main sequence , blue horizontal branch and population ii stars .
[ colorcolor ] compares the optical - near infrared colors of albus 1 with those of dwarf and giant stars in the direction to alnilam and mintaka .
our blue source is even bluer than the late o- and early b - type stars . besides
, such luminous stars are located at long heliocentric distances ( @xmath23 0.4kpc ) , which implies very low proper motions , in contrast with what we have measured for albus 1 .
population ii stars are common in the bulge near the centre of the galaxy and in the galactic halo .
some of the latters cross the solar neighbourhood , but only a few of them display extremely blue colors ( see , e.g. , the recent photometric study of horizontal - branch and metal - poor candidates by beers et al .
therefore , albus 1 is an early - type hot subdwarf or a white dwarf .
the extreme blueing of albus 1 prevents from transforming the @xmath13 magnitudes to johnson @xmath24 magnitudes using standard relations and , therefore , to compare its colors with other white dwarfs tabulated in exhaustive works such as in bergeron , leggett & ruiz ( 2001 ) .
a new comparison may come , instead , from the available tycho-2 and 2mass data . in table
[ wds ] we have compiled the basic data of the brightest white dwarfs and blue subdwarfs identified in the tycho-2 catalogue .
we used the white dwarf lists by mccook & sion ( 1999 ) and holberg , oswalt & sion ( 2002 ) , and looked for the tycho-2 counterparts of the white dwarfs brighter than @xmath4 = 13.0mag .
fainter objects were not considered due to their poor photometric accuracy .
this list surpasses the sample of white dwarf observed by the @xmath25 satellite in vauclair et al .
there are three evident absences : sirius b , @xmath3 eri b , and procyon b ( the three brightest known white dwarfs ) , which are too close to other bright stars and were , thus , not identified by tycho-2 . among the tabulated objects , there is only one hot subdwarf with blue colors , gj 3435 , which indicates its rarity .
data in table [ wds ] is represented in fig .
[ colorcolor ] .
albus 1 is located in the color - color diagram very close to the well known white dwarfs g 191b2b ( da1 ) and gj 433.1 ( da3 ) , at @xmath26 0.8mag , @xmath27 0.2mag .
the resemblance between the spectral energy distributions of albus 1 and g 191b2b ( `` the best studied of all hot white dwarfs '' ; barstow et al .
2003 ) , shown in fig .
[ sed ] , is evident .
both of them have the same @xmath28 magnitude within the error bars ( albus 1 : @xmath28 = [email protected] ; g 191b2b : @xmath28 = [email protected] ) , but g 191b2b is [email protected] brighter in @xmath0 .
it leads to tentatively classify albus 1 as an early da white dwarf slightly cooler than g 191b2b and , therefore , slightly closer to the sun . the stellar common proper - motion companion of g 191b2b has an accurate parallax determination by _ hipparcos _ at @xmath29 = 46@xmath14pc .
hence , albus 1 could be located at about 40pc , which would explain its appreciable proper - motion .
the probability of albus 1 being a more distant blue subdwarf is smaller ( see table [ wds ] ) . from the blue @xmath30 color in fig .
[ colorcolor ] , it is deduced , besides , that albus 1 has no main sequence close companion or forms part of a cataclysmic variable system .
as shown by salim & gould ( 2002 ) , an optical - infrared reduced proper motion digram ( e.g. @xmath31 vs. @xmath32 ) can be used to classify stars even if no parallax information is available . in particular , white dwarfs and subdwarf stars
are easily distinguished from main sequence stars as they are several magnitudes dimmer at the same color .
the position of albus 1 in the reduced proper motion diagram in fig . 4 in gould & morgan ( 2003 ; @xmath33 [email protected] , @xmath34 = [email protected] ) agrees with this requirement .
of the 30 white dwarfs and blue subdwarfs listed in table [ wds ] , only thirteen have tycho-2 @xmath6 magnitudes brighter than 12.0mag .
albus 1 , with @xmath6 = [email protected] , is included in this group .
accounting for the three brightest known white dwarfs not in the table , and discarding the close binary systems bl psc ab , v841 ara ab , v3885 sgr ab , and bd+28 4211 ab , whose spectral energy distributions are affected by the main sequence close companions , then albus 1 is the 12th brightest white dwarf yet known . since six of the white dwarfs brighter than it are in multiple systems ( the binary status of gj 127.1 ab claimed by gill & kaptein 1896 is , however , not confirmed ) , then albus 1 would be the sixth brightest isolated white dwarf , after the long - time known feige 34 , l 145141 , bd07 3632 , and hd 340611 ( luyten 1949 ; eggen & greenstein 1965 ) and the very hot white dwarf and extreme ultraviolet source re j221449 ( holberg et al . 1993 ) .
albus 1 , although located in the southern hemisphere , is visible from the most important northern observatories .
this fact , together with its brightness , makes our blue source an appropiate candidate spectrophotometric standard provided that its white dwarf or hot subdwarf nature is spectroscopically confirmed .
our serendipitous detection has also shown that the @xmath2 color is a good and simple discriminator to look for very blue , relatively bright objects .
a search for new very bright white dwarf candidates using tycho-2 and 2mass is currently ongoing .
this research has made use of the spanish virtual observatory supported from the spanish ministerio de educacin y ciencia through grants aya200504286 and aya200524102e of the plan nacional de astronoma y astrofsica .
barstow m. a. , good s. a. , holberg j. b. , hubeny i. , bannister n. p. , bruhweiler f. c. , burleigh m. r. & napiwotzki r. 2003 , mnras , 341 , 870 beers t. c. , flynn c. , rossi s. et al .
2007 , apjs , 168 , 128 bergeron p. , leggett s. k. & ruiz m. t. 2001 , apjs , 133 , 413 bessel f. w. 1844 , mnras , 6r , 136 caballero j. a. & solano e. 2007 , a&a , to be submitted cutri r. m. , skrutskie m. f. , van dyk s. et al .
2003 , vizier on - line data catalog : ii/246 .
originally published in : university of massachusetts and infrared processing and analysis center ( ipac / california institute of technology ) denis consortium 2005 , vizier on - line data catalog : b / denis .
originally published in : the denis consortium ( 2005 ) desidera s. & barbieri m. 2007 , a&a , 462 , 345 eddington a. s. 1924 , mnras , 84 , 308 eggen o. j. & greenstein j. l. 1965 , apj , 141 , 83 eisenstein d. j. , liebert j. , harris h. c. et al .
2006 , apjs , 167 , 40 giclas h. l. , burnham r. & thomas n. g. 1965 , bulletin lowell observatory , 6 , 155 gill d. & kapteyn j. c. 1896 , the cape photographic durchmusterung for the equinox 1875 , london , printed for h. m. stationery off . , by darling & son , ltd .
, 18961900 gould a. & morgan c. w. 2003 , apj , 585 , 1056 hambly n. c. , macgillivray h. t. , read m. a. et al . 2001 , mnras , 326 , 1279 hansen b. m. s. & liebert j. 2003 , ara&a , 41 , 465 herschel w. 1785 , philosophical transactions of the royal society of london , 75 , 40 hg e. , fabricius c. , makarov v. v. , urban s. , corbin t. , wycoff g. , bastian u. , schwekendiek p. & wicenec a. 2000 ,
a&a , 355 , l27 holberg j. b. , barstow m. a. & buckley d. a. h. 1993 , apj , 416 , 806 holberg j. b. , oswalt t. d. & sion e. m. 2002 , apj , 571 , 512 koester d. 2002 , a&arv , 11 , 33 kuiper g. k. 1941 , pasp , 53 , 248 landolt a. u. & uomoto a. k. 2007 , aj , 133 , 768 liebert j. 1980 , ara&a , 18 , 363 luyten w. j. 1922 , pasp , 34 , 156 luyten w. j. 1949 , apj , 109 , 528 luyten w. j. 1950 , aj , 55 , 86 mcalister h. a. , mason b. d. , hartkopf w. i. , roberts l. c. jr . &
shara m. m. 1996 , aj , 112 , 1169 mccook g. p. & sion e. m. 1999 ,
apjs , 121 , 1 monet d. g. , levine s. e. , casian b. et al .
2003 , aj , 125 , 984 pounds k. a. , allan d. j. , barber c. et al .
1993 , mnras , 260 , 77 ribeiro f. m. a. & diaz m. p. 2007 , aj , 133 , 2659 salim s. & gould a. 2002 , apj , 585 , l83 schatzman e. 1958 , white dwarfs , amsterdam , north - holland pub .
co. ; new york , interscience publishers , 1958 .
schwartz r. d. 1972 , pasp , 84 , 28 thackeray a. d. 1961 , mnssa , 20 , 40 vauclair g. , schmidt h. , koester d. & allard n. 1997 , a&a , 325 , 1055 zacharias n. , monet d. g. , levine s. e. , urban s. e. , gaume r. & wycoff g. l. 2004 , american astronomical society meeting 205 , # 48.15 ; bulletin of the american astronomical society , vol . 36 , p. 1418
lccc name & albus 1 & & 1 + wd number & wd 0604203 & & 1 + @xmath35 ( j2000 ) & 06 06 13.39 & & 2 + @xmath8 ( j2000 ) & 20 21 07.3 & & 2 + @xmath36 & + 7@xmath13 & masa@xmath16 & 2 + @xmath37 & 18@xmath13 & masa@xmath16 & 2 + @xmath0 & [email protected] & mag & 2 + @xmath6 & [email protected] & mag & 2 + @xmath38 & 11.84 & mag & 3 + @xmath39 & 11.90 & mag & 3 + @xmath40 & [email protected] & mag & 4 + @xmath41 & [email protected] & mag & 4 + @xmath28 & [email protected] & mag & 4 + sp .
type & da ? & & 1 + @xmath29 & @xmath1140 ? & pc & 1 + llcccc & ab & da2+k0iv & [email protected] & [email protected] & [email protected] + & & dz7 & [email protected] & [email protected] & [email protected] + & & da3.5 & [email protected] & [email protected] & [email protected] + & , feige 22 & da2.5 & [email protected] & [email protected] & [email protected] + & ab , feige 24 & da+m1v & [email protected] & [email protected] & [email protected] + & @xmath42 & da3 & [email protected] & [email protected] & [email protected] + & @xmath43 & dq7 & [email protected] & [email protected] & [email protected] + & & da1 & [email protected] & [email protected] & [email protected] + * wd 0604203 & * albus 1 & * da ? & * [email protected] & * [email protected] & * [email protected] + & & da1 & [email protected] & [email protected] & [email protected] + & , he 3 & da2.5 & [email protected] & [email protected] & [email protected] + & , gd 294 & sdb & [email protected] & [email protected] & [email protected] + & , feige 34 & da0 : & [email protected] & [email protected] & [email protected] + & & da3 & [email protected] & [email protected] & [email protected] + & , l 145141 & dq6 & [email protected] & [email protected] & [email protected] + & ab@xmath44 & da1+m3.5ve & [email protected] & [email protected] & [email protected] + & @xmath45 & da4 & [email protected] & [email protected] & [email protected] + & & da3 & [email protected] & [email protected] & [email protected] + & @xmath46 & da2 & [email protected] & [email protected] & [email protected] + & ab@xmath47 & doz1+k0v & [email protected] & [email protected] & [email protected] + & , g 22629@xmath48 & dav4.7 & [email protected] & [email protected] & [email protected] + & @xmath49 & dbqa5 & [email protected] & [email protected] & [email protected] + & ab@xmath50 & db : p+m : v & [email protected] & [email protected] & [email protected] + & & da4 & [email protected] & [email protected] & [email protected] + & & da2.5 & [email protected] & [email protected] & [email protected] + & & da2.5 & [email protected] & [email protected] & [email protected] + & & da3.5 & [email protected] & [email protected] & [email protected] + & ab@xmath51 & sdo : p+g : & [email protected] & [email protected] & [email protected] + & & da3 & [email protected] & [email protected] & [email protected] + & & da.76 & [email protected] & [email protected] & [email protected] + * * * * * *
|
we have serendipitously discovered a previously - unknown , bright source ( @xmath0 = [email protected] ) with a very blue @xmath2 color , to which we have named albus 1 .
a photometric and astrometric study using virtual observatory tools has shown that it possesses an appreciable proper motion and magnitudes and colors very similar to those of the well known white dwarf g 191b2b .
we consider albus 1 as a da - type white dwarf located at about 40pc .
if confirmed its nature , albus 1 would be the sixth brightest isolated white dwarf in the sky , which would make it an excellent spectrophotometric standard .
|
photosynthesis is the process used by plants and bacteria to convert the energy of sunlight into chemical energy in order to fuel the organism s activities . in the initial steps of the photosynthetic process
, pigment - protein complexes complete the light - energy transfer and charge separation with a near unity quantum efficiency .
light - harvesting molecular complexes achieve the energy transfer by using an array of light - harvesting pigments that absorb the energy to form excitons and funnel the excitation to the reaction center @xcite . to investigate the energy transfer dynamics in the first steps of photosynthesis on the fs time scale , ultrafast spectroscopic tools are available . among the techniques used , two - dimensional ( 2d )
photon echo spectroscopy is a powerful tool which allows for direct mapping of the excitation energy pathways as a function of absorption and emission wavelength @xcite .
it is particularly useful in examining photosynthetic systems in which the manifold of electronic states is closely spaced and broadening through static disorder yields highly congested spectra .
recent experimental 2d electronic spectroscopic studies of the fenna - matthews - olson ( fmo ) complex observed coherent beating signals and , thus , raised interest in the interplay between energy transfer , long - lived quantum coherence in photosynthetic processes @xcite , and low - frequency vibrations of the molecular back bone . also in photoactive marine cryptophyte algae @xcite , the light - harvesting complex lh2 @xcite of _ rhodobacter sphaeroides _ , and in the reaction center @xcite of the photosystem ii
, long - lived oscillations have been experimentally observed at low ( 77 k ) and room temperatures ( 300 k ) , indicating a strong vibronic coupling in these systems . to analyze the experimental findings in such large and complex photoactive molecular complexes ,
a thorough comparison with theoretical calculations is essential , in order to arrive at a reliable interpretation of the measured 2d spectra . since it is a difficult and computationally demanding task to determine 2d optical spectra , often only the population dynamics of exciton states is calculated . for the fmo complex , a rather small light - harvesting complex ,
the hierarchical equations of motion @xcite were applied and quantum oscillations were observed on the time scale of the 2d experiments employing an environmental debye model spectral density with rather small reorganization energy @xcite . employing the numerically exact quasiadiabatic propagator path integral ( quapi ) allowed to use a more realistic measured environmental spectral density . however , this resulted in a decay of the decoherence faster than experimentally observed @xcite .
this spectral density could , more recently , be employed to calculate the 2d spectra of fmo with the hierarchy equation @xcite and a reasonable agreement between theory and experiment could be achieved .
the calculations of quapi and the hierarchical equations of motion treated the coupling of the complex to environmental fluctuations numerically exactly . however , the computational effort is immense , which makes the simulation of larger light - harvesting molecular complexes ( which contain , typically , dozens to hundreds of excitonic subunits ) virtually impossible .
the need for a highly efficient numerical tool to calculate 2d optical spectra of large molecular complexes with a reasonable numerical effort and a satisfactory accuracy still exists and it is expected to continue to increase . given their complex molecular structures , for the calculation of 2d spectra of large light - harvestors , approximate schemes are usually inavoidable .
standard redfield equations @xcite , which invoke a lowest - order born and a markovian approximation , are good at weak system - bath coupling , but fail for strong coupling .
the regime of intermediate system - bath coupling as present for the exciton dynamics in photoactive complexes @xcite is typically also not properly treated within redfield equations @xcite .
thus , the _ modified _ redfield theory ( mrt ) has been widely used for the description of energy transfer processes of large molecules @xcite . in this approach ,
that contribution of the system - bath coupling hamiltonian , which is diagonal in the eigenbasis of the system , is included fully , while a second - order perturbative approximation is used for the off - diagonal coupling terms .
the equation of motion of the mrt includes a population transfer within the reduced density matrix , but the accompanying population - transfer induced dephasing is neglected .
the accuracy of the mrt in view of the dynamics of the reduced density matrix has been analyzed in detail @xcite .
moreover , mrt has been shown to have a somewhat wider range of applicability when compared to both the original redfield and frster theory @xcite .
also , linear absorption spectra for an ensemble of b850 rings have been determined which shows that mrt includes non - markovian effects which clearly show up in the high - energy part of the static absorption lineshapes @xcite .
different energy transfer components of lhcii trimer and phycoerythrin 545 have been revealed using mrt by simultaneous quantitative fits of the absorption , linear dichroism , steady - states fluorescence spectra , and transient absorption kinetics upon excitation at different wavelengths @xcite .
a more refined description of the quantum dissipative exciton dynamics is achieved in this work upon observing that the population - transfer induced electronic dephasing can be efficiently included in the quantum master equation .
the off - diagonal terms in the quantum master equation now include the decoherence of excited states and electronic dephasing between ground and excited states by exploiting the relation @xmath0 to estimate the different contributions to the dephasing rate , where @xmath1 is the transverse relaxation time and @xmath2 , @xmath3 are the longitudinal relaxation time and pure dephasing time , respectively .
while working out the details with the results reported in this paper , this extended quantum master equation has also been independently put forward very recently in ref .
@xcite and has been named the coherent modified redfield theory ( cmrt ) .
to avoid confusion , we use to this nomenclature also here . for calculating 2d photon - echo spectra ,
essentially two different approaches are available . on the one hand ,
the response to the sequence of applied laser pulses can be calculated by evaluating the third - order optical response function @xcite .
modified redfield theory was successfully applied to simulate the 2d spectra of the double - ring lh2 aggregate of purple bacteria including both the b800 and the b850 ring @xcite .
comparing experimentally measured and theoretically calculated results of 2d spectra revealed that excitation energy transfer through the lhcii happens on three time scales : sub-100fs relaxation through spatially overlapping states , several hundred femtosecond transfer between nearby chlorophylls , and picosecond energy transfer steps between layers of pigments @xcite .
more recently , 2d spectra of the reaction center in the photosystem ii were calculated with mrt at low temperature , and the charge separation process was investigated by mrt including charge transfer states in the model @xcite . an alternative approach to calculate 2d optical spectra , which is especially useful when finite durations of the laser pulses as well as pulse overlap effects are taken into account , is the equation of motion - phase matching approach ( pma ) @xcite . using the pma in combination with the conventional redfield equations , 2d spectra of a single fmo subunit were studied , and the signature of energy transfer was revealed by well - resolved peaks in the simulation with adjustable pure dephasing of exciton states @xcite .
although mrt is used to tackle many different problems in the study energy transport in photosynthetic complexes , no investigation of its reliability in calculating nonlinear and , specifically , 2d optical spectra is at hand . in this work ,
we first verify the cmrt approach by comparing the population dynamics of fmo exciton states with numerically exact results of the quapi approach .
in addition , we combine the cmrt with the pma to calculate 2d photon - echo spectra for a simple dimer model . again
, the results of cmrt+pma are benchmarked against numerically exact results of the quapi approach .
for the long - time steady state dynamics , the cmrt+pma and quapi simulations show differences for intermediate and strong system - bath coupling .
however , for intermediate coupling , as it is typical in photosynthetic complexes , the short time dynamics including dephasing times and coherent beating frequencies are well described by cmrt+pma .
hence , an efficient numerical scheme to calculate 2d photon - echo spectra with a reasonable computational effort is available .
the remainder of this paper is organized in the following way . in the next section [ sec.model ]
, we briefly introduce the model of the fmo complex , for we compare the performance of the the cmrt+pma in calculating a non - trivial population dynamics . additionally , the dimer model is introduced for which we compare results of 2d photon echo spectra .
a brief description of the cmrt+pma and quapi for calculating the reduced density matrix and 2d spectra is given in section [ sec.tech ] . in section [ sec.res ] ,
we give the results of the comparison , and a thorough discussion is appended in section [ sec.disc ] , before we finish with a conclusion .
in the framework of open quantum systems the hamiltonian of the complete system @xmath4 can be decomposed as four parts @xmath5 where @xmath6 is the on - site transition energy and @xmath7 is the intermolecular coupling .
@xmath8 is the total number of monomers .
@xmath9 is the number of bath modes coupled to molecule @xmath10 , which we will take to be infinity . @xmath11 and @xmath12 are the mass weighted position and momentum of the @xmath13th harmonic oscillator bath mode with frequency @xmath14 . the interaction term @xmath15 induces the coupling between system and bath .
it is assumed to be separable such that @xmath16 only acts on the system subspace and @xmath17 only on the bath degrees of freedom . in the following we further assume a linear relation between bath coordinates and the system .
the system - bath interaction is then given as @xmath18 and we furthermore restrict our considerations to pure electronic dephasing only , i.e. @xmath19 . the renormalization term is @xmath20 this term compensates for artificial shifts of the system frequencies due to the system - bath interaction .
the influence of the bath is fully described by its bath spectral density [ specdens ] j_m ( ) = _ j ( - _ mj ) = 2 with reorganization energy @xmath21 and high - frequency cut off @xmath22 .
we assume the bath at each monomer to be independent ( no cross - correlation between baths @xcite ) but with identical spectra , i.e. @xmath23 . laser pulses acting on the exciton complex result in the addition of a system - field interaction , i.e. @xmath24 , which is defined within the dipole approximation according to @xmath25 with the electric field of the laser pulse @xmath26 and the electronic transition dipole operator @xmath27 of the exciton system @xmath28 @xmath29 determines the dipole strength and direction of the @xmath10th monomer .
the dimer is modeled by two monomers with site energies @xmath30 @xmath31 and a coupling @xmath32 @xmath33 .
the two dipole moments are considered to be perpendicular to each other , i.e. @xmath34 . for the bath spectral density as specified in eq .
( [ specdens ] ) , we choose @xmath35 @xmath33 and @xmath36 @xmath33 , and set @xmath37 @xmath33 .
we model a monomer subunit of the fmo trimer by including seven bacteriochlorophyll _ a _ molecular sites .
the single excitation subspace is described by a hamiltonian [ eq : hamiltonian ] h_f = 240 & -87.7 & 5.5 & -5.9 & 6.7 & -13.7 & -9.9 + & 315 & 30.8 & 8.2 & 0.7 & 11.8 & 4.3 + & & 0 & -53.5 & -2.2 & -9.6 & 6.0 + & & & 130 & -70.7 & -17.0 & -63.3 + & & & & 285 & 81.1 & -1.3 + & & & & & 435 & 39.7 + & & & & & & 245 in units of @xmath33 , where we use the site energies and dipolar couplings determined by adolphs and renger @xcite for the fmo complex of _ chlorobium tepidum_. bath parameters are chosen as @xmath21=35 @xmath33 and @xmath22=53 @xmath33 following ref.@xcite .
the cmrt can be derived from the nakajima - zwanzig equation @xcite using a scheme for the separation of the total hamiltonian which does not treat the whole system - bath interaction term @xmath38 perturbatively @xcite@xcite . instead , the hamiltonian is separated according to @xmath39 where @xmath40 are eigenstates of @xmath41 and @xmath42 is the off - diagonal term of the system - bath interaction part in the exciton basis . in this basis , @xmath43 is diagonal and the matrix elements read @xmath44 where @xmath45 is the @xmath46th excitonic level of the system hamiltonian and @xmath47 is the weighted reorganization energy .
moreover , @xmath48 \nonumber \\ \ ] ] describes a bath of harmonic oscillators with mass @xmath49 , frequency @xmath50 and momentum @xmath51 , shifted due to the coupling with the exciton state @xmath40 .
in addition to the redefinition of the system and the bath hamiltonian , one has to define a different type of projection operator which only projects on the diagonal part of the system density matrix in the eigenstate basis .
this is achieved by @xmath52 where @xmath53 is the projector onto the @xmath46th excitonic state and @xmath54 is the equilibrium density matrix of the bath when the system is in the excitonic state @xmath40 . here , @xmath55 with @xmath56 and @xmath57 being the temperature . inserting these definitions into the nakajima - zwanzig equation , determining @xmath42 up to second order and invoking the time - dependent population transfer rate , one obtains an equation of motion for the population transfer terms in the form @xmath58 , \nonumber \\\end{aligned}\ ] ] with the population transfer rates @xcite @xmath59\\ & \times \{\ddot{g}_{\mu\nu\nu\mu}(\tau)-[\dot{g}_{\nu\mu\nu\nu}(\tau)-\dot{g } _ { \nu\mu\mu\mu}(\tau)+2i\lambda_{\nu\mu\nu\nu}]\\ & [ \dot{g}_{\nu\nu\mu\nu}(\tau)-\dot{g}_{\mu\mu\mu\nu}(\tau)+2i\lambda _ { \mu\nu\nu\nu}]\}. \end{aligned}\end{aligned}\ ] ] here , @xmath60 . the lineshape function @xmath61 can be written as the two - time integral of the bath correlation function according to @xmath62 to obtain eq .
( [ eq : mrt tensor ] ) , we have used the cumulant expansion up to second order in the system - bath coupling and have taken the independent bath model into account .
the absorption lineshape within the cmrt is given by @xmath63 .
\end{aligned}\end{aligned}\ ] ] as detailed in ref .
@xcite .
up to this point , eq .
( [ eq : population ] ) constitutes the modified redfield theory , as developed and applied in refs .
@xcite . based on the population transfer term in eq .
( [ eq : population ] ) , we extend the quantum master equation by including also the coherence ( or , off - diagonal ) terms of the reduced density matrix . the resulting coherent modified redfield quantum master equation now reads @xmath64-\re\{\rho(t)\}\ , , \end{aligned}\ ] ] where @xmath65 is the time - dependent system - field interaction term .
the relaxation and dephasing operator @xmath66 now also includes diagonal and off - diagonal terms .
the diagonal part of the relaxation operator , which was desribed in ref .
@xcite , reads @xmath67\ , .
\end{aligned}\end{aligned}\ ] ] the off - diagonal terms @xmath68 are now included in order to describe decoherenece of excited states and electronic dephasing between the ground and excited states . here
, we use an efficient way to obtain the associated rates by exploiting the relation @xmath0 to estimate the different contributions to the dephasing rate .
@xmath1 is the transverse relaxation time , @xmath2 , @xmath3 are the longitudinal relaxation time and pure dephasing time , respectively @xcite . in detail , @xmath69 and @xmath70
is given by the first derivative of lineshape function @xmath71 .
therefore , the off - diagonal terms of the excited states and between the ground and excited states can be written as @xmath72\rho_{\mu\nu}(t),\\ \re\{\rho(t)\}_{\mu 0}=&\{\frac{1}{2}(\sum_{m\neq n}r_{m m n
n}(t)+\sum_{n\neq m}r_{n n m m}(t))+\\ & \dot{g}_{\mu\mu\mu\mu}(t)\}\rho_{\mu 0}(t ) .
\end{aligned}\end{aligned}\ ] ] this extended quantum master equation has also been independently put forward very recently in ref .
@xcite and has been named the coherent modified redfield theory ( cmrt ) .
it is an efficient , but approximate way to take into account population transfer and dephasing on the same footing .
the quasiadiabatic propagator path integral @xcite is a numerically exact approach to determine the influence of environmental fluctuations on the system dynamics within a open quantum systems approach .
specifically , quapi determines the time dependent reduced statistical operator @xmath73 of the system .
it is well established in the literature and we only briefly summarize the central features in the following . the algorithm is based on a symmetric trotter splitting of the short - time propagator @xmath74 for the full hamiltonian into two parts , one depending on the system hamiltonian , and one involving the bath and the coupling term .
the short - time propagator determines the time evolution over a trotter time slice @xmath75 .
the discrete time evolution becomes exact in the limit @xmath76 . for any finite @xmath75 ,
a finite trotter error occurs which has to be eliminated by choosing @xmath75 small enough to achieve convergence . on the other side ,
the environmental degrees of freedom generate correlations which are non - local in time . for any finite temperature , these correlations decay on a time scale denoted as the memory time scale .
the quapi scheme defines an augmented reduced density tensor , which lives on this full memory time window .
then , an iteration scheme is established in order to extract the time evolution of this object .
all correlations are completely included over the finite memory time @xmath77 but are neglected for times beyond @xmath78 .
one increases the memory parameter @xmath79 until convergence is found .
the two strategies to achieve convergence , i.e. , minimize @xmath75 but maximize @xmath77 , are naturally counter - current , but nevertheless convergent results can be obtained in a wide range of parameters , including the cases presented in this work . in the equation of motion - phasing matching approach ( pma ) ,
the polarization in the photon - echo direction is calculated by simultaneously propagating three auxiliary density matrices @xmath80 and @xmath81 , thereby employing also the rotating wave approximation @xcite .
the time evolution equations are given by @xmath82\nonumber \\&-\re\{\rho_{1}(t)\},\\ \frac{\partial}{\partial t}\rho_{2}(t)&=-i[h - v_{1}(t , t_{1})-v^{\dagger}_{2}(t , t_{2}),\rho_{2}(t ) ] -\re\{\rho_{2}(t)\},\nonumber\\ \frac{\partial}{\partial t}\rho_{3}(t)&=-i[h - v_{1}(t , t_{1})-v^{\dagger}_{3}(t , t_{3}),\rho_{3}(t ) ] -\re\{\rho_{3}(t)\}.\nonumber \ ] ] where @xmath83 and @xmath84 , @xmath85 is the pulse duration . to obtain the third - order 2d signal , the polarization in the phase matching direction
is evaluated as @xmath86 where the bracket @xmath87 denotes the trace .
experimentally , in the limit of ideal detection , the heterodyne photo echo signal is proportional to the polarization @xmath88 , where @xmath89 is the detection time .
therefore , the ideal total 2d signal can be expressed as @xmath90 where @xmath91 , @xmath57 and @xmath89 denote coherence time , population ( waiting ) time and detection time , respectively , @xmath92 , @xmath93 .
the coherence time corresponds to a period in which the system is coherently evolving after the first interaction with the optical field .
the second interaction with the field creates population states and the third interaction recovers the coherence again .
the fourier transform in eq .
( [ eq:2d fourier transform ] ) is always performed over the coherence time @xmath91 and the detection time @xmath89 .
the corresponding frequencies @xmath94 , @xmath95 are often referred to as absorption and emission frequencies , respectively .
in addition , gaussian laser pulses have been assumed for a realistic detection scheme , which have the form @xmath96 where @xmath97 , @xmath98 , @xmath99 , and @xmath100 are the amplitude , envelop central time , wave vector and frequency of the pulses and @xmath101 characterizes the pulse duration .
note that all the pulses are assumed to have the same lineshape , carrier frequencies and durations in this paper .
in order to verify the reliability of the cmrt , we present the population dynamics of the fmo complex calculated by cmrt and compare the results to those obtained by the numerically exact quapi method . in fig . [
fig : fmo ] , the population dynamics of selected fmo sites is shown for @xmath102 k for two different initial conditions . in fig . [
fig : fmo]a ) , we assume the energy transfer to start from site 1 .
we monitor then the full transfer which involves all seven fmo sites . for simplicity ,
we only show the population dynamics of the sites 1 , 2 , and 3 .
alternatively , the energy transfer may be assumed to start from site 6 , see fig .
[ fig : fmo]b ) .
there , we depict the population dynamics of the relevant sites 3 , 5 , and 6 .
we observe that the oscillatory behavior of the populations is captured by both approaches .
both also yield the same decay rates and periods of oscillations .
however , a phase shift of the oscillations occurs between the cmrt and quapi results .
energy transfer is believed to be related to the population of the fmo site 3 ( green symbols and lines ) which has the lowest energy in the fmo monomer . in our comparison ,
cmrt slightly overestimates the population transfer efficiency towards site 3 .
all in all , the cmrt results for the fmo exciton population dynamics are in good agreement to numerically exact quapi results . since the system - bath coupling parameters of the fmo complex are typical for natural photosynthetic units , we conclude that cmrt is a useful tool to study their exciton dynamics .
population dynamics of selected fmo sites . in a ) , the population of sites @xmath103 ( black ) , @xmath104 ( red ) and @xmath105 ( green ) with the initial condition @xmath106 is shown , while in ( b ) the populations of sites @xmath105 ( green ) , @xmath107 ( blue ) and @xmath108 ( magenta ) with the initial condition @xmath109 is depicted ( symbols : quapi , full lines : cmrt ) for the parameters as given in the text.,width=283 ] to obtain 2d spectra , we combine the cmrt next with the pma .
this constitutes a very efficient approximate numerical tool whose reliability is assessed by a comparison with 2d spectra obtained by quapi . since 2d spectra
involve extended numerical calculations , quapi results are available only for small model systems with present day hardware technology .
for such a comparison , we present the calculated results for the dimer model .
it allows us to study energy transfer and dephasing ( homogeneous broadening ) as building blocks of the exciton dynamics in larger molecular compounds .
it can still be treated by quapi with reasonable numerical effort .
[ fig:2d dimer ] ( left ) shows 2d spectra of the dimer calculated by cmrt+pma for @xmath21=50 @xmath33 and the other parameters as indicated above .
they are compared to quapi results ( right column in fig .
[ fig:2d dimer ] ) for waiting times @xmath110 fs , @xmath111 fs , @xmath36 fs and @xmath112 fs .
these 2d spectra show two diagonal peaks ( labeled a , b ) which correspond to the two exciton states .
moreover , two cross peaks ( labeled c and d ) arise due to the excitonic coupling between them . for the sake of simplicity and clarity of the comparison , inhomogeneous broadening and the rotational averaging for different laser polarizations and molecular orientations
is not performed here .
although this would be important to describe realistic experimental situations , the averaged results generally show smaller discrepancies ( not shown ) . at @xmath110
fs , the two results show the same profile for diagonal and cross peaks and , indeed , the agreement is excellent .
this shows that the cmrt correctly models the coherence times and the system - bath correlations created during the simulation . with increasing waiting time
, the same coherent dynamics is found for both the diagonal and the cross peaks and even can be inspected by eye .
however , some disagreement is observed at long waiting time @xmath113 fs .
the diagonal peak b in left figure ( cmrt+pma ) shows a somewhat reduced amplitude as compared to the right figure ( quapi ) . for a more refined comparison ,
the amplitudes of the diagonal and cross peaks ( a , b , c and d ) are plotted against the waiting time in fig .
[ fig : diagonal_dynamics ] and fig .
[ fig : anti - diagonal_dynamics ] . in fig .
[ fig : diagonal_dynamics ] , the population dynamics of the diagonal peaks a ( top ) and b ( bottom ) calculated by cmrt+pma from @xmath114 to @xmath115 fs is shown and compared to the quapi result .
we find that the cmrt+pma provides reasonably accurate results for the population transfer and the oscillation period .
however , the amplitude of peak b decays slightly faster in the approximate results as compared to the quapi data .
moreover , both yield different stationary states . in addition , the phase of the oscillations is slightly shifted . for the comparison of the cross peaks , the oscillatory behavior of peaks c and d
is plotted in fig .
[ fig : anti - diagonal_dynamics ] .
cross peak c shows a similar oscillatory behavior but the two approaches yield different stationary states .
peak d shows a only slightly shifted phase of the oscillatory behavior .
such a phase shift was also observed in the population dynamics of the fmo complex shown above .
the phase shift might be due to the neglect of imaginary parts in the redfield relaxation tensor . in order to further assess the reliability of the cmrt+pma
, we have repeated the calculations for a larger reorganization energy , i.e. , for @xmath21=100 @xmath33 ( with @xmath22=100 @xmath33 kept unchanged ) .
2d spectra were again calculated by both approaches and the amplitude of the labeled peaks were extracted .
their time - dependence is plotted in figs .
[ fig : diagonal_dynamics2 ] and [ fig : anti - diagonal_dynamics2 ] .
cmrt+pma still yields quantitative agreement with the quapi result except for the behavior of the damping . the stronger system - bath coupling results in faster damping ( diagonal peak a ) and also in an increased difference between quapi and cmrt+pma as compared to the weaker coupling with @xmath21=50 @xmath33 .
two - dimensional photon - echo spectra of a dimer model calculated by cmrt+pma ( left ) and quapi ( right ) for different waiting times as indicated .
the debye spectral density was used for the calculation with the parameters @xmath116 @xmath33 , @xmath117 @xmath33 and the temperature was set to @xmath102 k.,width=321 ] population dynamics of the labeled diagonal peaks ( a , b ) extracted from the underlying sequence of 2d maps .
the two approaches yield the same oscillation period .
the diagonal peak b obatained from cmrt+pma decays faster as compared to the quapi result .
the oscillation periods can be extracted by data fitting and are : cmrt+pma : 110fs , quapi : 99fs.,width=245 ] coherent oscillations of labeled cross peaks ( c , d ) extracted from 2d maps .
cross peak c obained by cmrt+pma shows the same oscillatory behavior , but with a somewhat smaller amplitude.,width=245 ]
from the above comparison of the results obtained by both approaches , we observe that the discrepancies found in the 2d calculations are more pronounced than in the dynamics of the populations . put differently , nonlinear 2d spectra are more sensitive to assess the performance and reliability of approximate theoretical approaches . in order to understand this
, we point out two fundamental differences between 2d spectra and the population dynamics .
first , entanglement between the system and the bath leads to initial correlations at the beginning of the waiting time window , which are absent in the calculation of the population dynamics .
second , two - exciton states contribute to the 2d spectra during the detection time , and interference between positive and negative peaks changes the observed amplitudes .
this shows that one can not understand the reliability of a method to simulate correct 2d spectra by calculating population dynamics alone . our current framework , in which we use the combined cmrt+pma and compare the results with quapi , is well suited to show the performance of these methods in understanding 2d spectra directly . in more detail ,
we have observed three noticeable discrepancies of the cmrt+pma as compared to quapi : i ) shifted oscillation phase of peak intensities , ii ) a slightly faster decay , and , iii ) a different amplitude of peaks b and c for long waiting times . for the explanation of the shifted oscillation
observed in 2d simulations of the cmrt+pma , we need to notice that eqs .
[ eq : diagonal population ] , [ eq : off - diagonal population ] provide the analytic result for a monomer ( two - level system ) , and that this has been proven by comparing to quapi @xcite . however , cmrt yields a shifted period for the dimer model .
the mismatch is mainly caused by the population transfer term @xmath118 since there is no population transfer term in the monomer model . in this paper , the population transfer rate was calculated by the cumulant expansion in eq .
[ eq : mrt tensor ] @xcite and we only took the real part .
it is well known that the imaginary part dominates the phase of the oscillations @xcite .
so , most likely , the shifted oscillation is mainly caused by the real - value approximation of the population transfer rates .
then , the population transfer term is also derived based on the second - order perturbation approximation , which is one of the reasons for the explanation of the slightly too fast decay of the oscillations found in cmrt calculations .
furthermore , the secular approximation was used to separate the population dynamics and the dephasing process in eqs .
[ eq : diagonal population ] , [ eq : off - diagonal population ] .
this also contributes to the discrepancy in the decay rate , since it neglects the interference between population transfer and coherence dephasing .
a relatively small amplitude of peak b and c was found in fig .
[ fig : diagonal_dynamics ] and fig .
[ fig : anti - diagonal_dynamics ] and it also can be observed by eye in the 2d map for the long waiting time @xmath113 fs .
we observe that peaks b and c are mainly formed by one positive ( red ) peak and overlap with a negative ( blue ) peak in the 2d spectrum ( @xmath113 fs ) .
therefore , the amplitude of those peaks mainly depends on the overlap of two peaks . in the quapi result ,
the two peaks are clearly separated with a larger spectral distance than in the cmrt result and this leads to the larger amplitude of peaks b and c in the 2d spectrum calculated with quapi .
it indicates that , besides the shifted oscillation and the faster decay of the oscillation , cmrt does not properly account for the reorganization energy by the heat bath ( diagonal peaks show slightly different positions in the 2d map : @xmath119 @xmath33 and @xmath120 @xmath33 for cmrt and @xmath121 @xmath33 and @xmath122 @xmath33 for quapi ) . in the cmrt ,
the reorganization energy is included in the diagonal part of the hamiltonian by eq .
[ eq : renormalization ] , where it just brings in a shift of the excitonic transition frequency @xmath123 by the renormalization term eq . [ eq : renormalization ] and does not affect the dynamics of the off - diagonal terms in the hamiltonian in eq .
( [ eq : hamiltonian of mrt ] ) .
amplitude of the diagonal peaks a and b for a stronger system - bath coupling @xmath21=100 @xmath33 ( with @xmath22=100 @xmath33 unchanged ) .
cmrt+pma calculations yield a faster decay ( a ) as compared to the quapi result ( decay rate extracted from a fit : cmrt+pma : 81 fs , quapi : 146 fs).,width=245 ] oscillations of the cross peaks ( c , d ) with for a stronger system - bath interaction ( same parameters as in fig .
[ fig : diagonal_dynamics2 ] ) .
the cross - peak c calculated by cmrt+pma yields a faster decay and a smaller amplitude ( decay rates as obtained from a fit : cmrt+pma : 62 fs , quapi : 177 fs).,width=245 ] on the basis of a clear physical meaning ( population transfer and dephasing terms ) and for the purpose of an efficient and fast calculation , the secular approximation and the second - order perturbation theory were applied to construct the cmrt . on the one hand , the secular approximation leads to a separation of the population dynamics and dephasing process and avoids any complicated interaction terms between diagonal and off - diagonal parts in the equation of motion . on the other hand ,
the second - order perturbation theory simplifies the population transfers .
it is possible to improve the equation by including higher orders .
however , this renders the equation considerably more complicated and requires more computational resources for the simulation and it is a priori unclear how much this improves the accuracy .
in this paper we present the cmrt and compare it in more detail to the quapi method by calculating the population dynamics of selected fmo exciton sites and the 2d - spectrum of a model dimer .
we found that cmrt provides numerical reliable results as compared to numerically exact quapi calculations for both the population dynamics and 2d spectra , as long as the reorganization energy is not too large compared to the typical energy gap of the system .
most importantly , it requires smaller computational efforts and orders of magnitude shorter calculation times .
it provides us with an efficient approach to study the energy transfer in super - large molecular complexes and to perform complicated 2d simulations .
we found that the 2d profile calculated from cmrt+pma agrees well with the corresponding quapi results .
for a quantitative comparison , the amplitudes of diagonal and cross peaks were extracted from 2d maps and compared to those calculated by quapi .
quantitative agreement was found .
we observe some discrepancies .
in particular , oscillations are shifted , they decay slightly faster , and positions of peaks are slightly shifted .
this becomes more serious if the reorganization energy is increased , which is also the case in ref .
@xcite .
the simulation protocol developed here can be used for arbitrary forms of the spectral density .
one can envision an approach where the cmrt method is used to simulate super - large complexes , while numerically exact methods such as quapi play a role in benchmarking the accuracy of the simulations of smaller systems .
we acknowledge financial support by the joachim - herz - stiftung , hamburg within the pier fellowship program ( hgd ) and by the excellence cluster `` the hamburg center for ultrafast imaging - structure , dynamics and control of matter at the atomic scale '' of the deutsche forschungsgemeinschaft .
agd was supported by a marie curie international incoming fellowship within the 7th european community framework programme .
pn acknowledges financial support by the dfg project na394/2 - 1 .
m. b. plenio , j. almeida , and s. f. huelga , j. chem . phys . *
139 * , 235102 ( 2013 ) .
p. nalbach , d. braun and m. thorwart , phys .
e * 84 * , 041926 ( 2011 ) .
m. thorwart , j. eckel , j. h. rina , p. nalbach and s. weiss , chem .
. lett . * 478 * , 234 ( 2009 ) . c. kreisbeck , and t. kramer , j. phys .
lett . * 3 * , 2828 ( 2012 ) .
w. m. zhang , t. meier , v. chernyak and s. mukamel , j. chem .
* 108 * , 7763 ( 1998 ) .
m. yang , g. r. fleming , chem . phys .
* 275 * , 355 ( 2002 ) .
a. gelzinis , l. valkunas , f. d. fuller , j. p. ogilvie , s. mukamel and d. abramavicius , new j. phys . * 15 * , 075013 ( 2013 ) . v. i. novoderezhkin , a. g. yakovlev , r. van grondelle , and v. a. shuvalov , j. phys .
b * 108 * , 7445 ( 2004 ) .
q. ai , t. -c .
yen , b. -y .
jin and y. -c .
cheng , j. phys .
lett . * 4 * , 2577 ( 2013 ) .
m. cho et al
. j. phys .
b * 109 * , 10542 ( 2005 ) .
k. l. m. lewis , f. d. fuller , j. a. myers , c. f. yocum , s. mukamel , d. abramavicius and j. p. ogilvie , j. phys .
a * 117 * , 34 ( 2013 ) .
y. chang , and y .- c .
cheng , j. chem .
* 142 * , 034109 ( 2015 ) .
m. schrder , u. kleinekathfer and m. schreiber , j. chem . phys . * 124 * , 084903 ( 2006 ) .
r. van grondelle and v. i. novoderezhkin , phys .
chem . phys .
* 8 * , 793 ( 2006 ) .
h . hwang - fu , w. chen , and y .- c .
cheng , chem .
phys . * 447 * , 46 ( 2015 ) .
s. mukamel , principles of nonlinear optical spectroscopy ( oxford university press , new york , 1995 ) .
o. rancova , and d. abramavicius , j. phys .
b * 118 * , 7533 ( 2014 ) .
g. s. schlau - cohen , t. r. calhoun , n. s. ginsberg , e. l. read , m. ballottari , r. bassi , r. van grondelle and g. r. fleming , j. phys .
b * 113 * , 15352 ( 2009 ) .
d. abramavicius , s. mukamel , j. chem . phys . * 134 * , 174504 ( 2011 ) . m. f. gelin , d. egorova and w. domcke , j. chem
. phys . * 123 * , 164112 ( 2005 ) .
l. z. sharp , d. egorova , and w. domcke , j. chem .
phys . * 132 * , 014501 ( 2010 ) .
p. nalbach , j. eckel and m. thorwart , new j. of phys .
* 12 * , 065043 ( 2010 ) .
j. adolphs and t. renger , biophys .
j. * 91 * , 2778 ( 2006 ) . c. meier and d. j. tannor , j. chem .
phys . * 111 * , 3365 ( 1999 ) .
m. schrder , m. schreiber , u. kleinekathfer , j. luminescence , * 125 * , 126 ( 2007 ) .
v. may , o. khn , charge and energy transfer dynamics in molecular systems , ( wiley - vch press , weinheim , 2011 ) . n. makri and d. e. makarov , j. chem . phys . *
102 * , 4600 ( 1995 ) .
we compare the absorption lineshape calculated by cmrt and quapi .
they show exactly same result .
a. ishizaki , and g. r. fleming , j. chem . phys . *
130 * , 234111 ( 2009 ) . j. ma , j. m. moix , and j. s. cao , j. chem .
* 142 * , 094107 ( 2015 ) .
j. m. moix , j. ma , and j. s. cao , j. chem . phys . *
142 * , 094108 ( 2015 ) .
|
we combine the coherent modified redfield theory ( cmrt ) with the equation of motion - phase matching approach ( pma ) to calculate two - dimensional photon echo spectra for photoactive molecular complexes with an intermediate strength of the coupling to their environment .
both techniques are highly efficient , yet they involve approximations at different levels . by explicitly comparing with the numerically exact quasi - adiabatic path integral approach , we show for the fenna - matthews - olson complex that the cmrt describes the decay rates in the population dynamics well , but final stationary populations and the oscillation frequencies differ slightly .
in addition , we use the combined cmrt+pma to calculate two - dimensional photon - echo spectra for a simple dimer model .
we find excellent agreement with the exact path integral calculations at short waiting times where the dynamics is still coherent . for long waiting times
, differences occur due to different final stationary states , specifically for strong system - bath coupling . for
weak to intermediate system - bath couplings , which is most important for natural photosynthetic complexes , the combined cmrt+pma gives reasonable results with acceptable computational efforts .
|
the tully - fisher relation ( tfr ) relates the intrinsic luminosity to the maximum rotation velocity of the gas , @xmath1 , a proxy for the circular velocity , of late - type galaxies @xcite .
it reflects the equilibrium state of late - type galaxies but , unlike the fundamental plane of elliptical galaxies , which is a three - parameters relation @xcite , has only two parameters , implying additional relations between the observational characteristics .
still for gas - rich late - type galaxies , the tfr has been generalised to a relation between @xmath1 and stellar mass ( the stellar - mass tfr or stfr ) and between @xmath1 and hi gas+stellar mass ( the baryonic tfr or gstfr in our notation ) @xcite .
these relations are the subject of very active theoretical and observational work .
e.g. , @xcite present simulations of the evolution of the tfr of massive late - types ( @xmath2 km / s ) and show that while the tfr undergoes strong luminosity evolution , the stfr and gstfr have remained constant since @xmath3 . to test for any possible environmental influences , these authors switched off star - formation in a disk galaxy by removing all its halo gas .
after some fading and reddening , this galaxy ends up slightly below the b - band tfr but remains on the stfr and gstfr . @xcite
have simulated the evolution of the stfr and gstfr of dwarf galaxies ( @xmath4 km / s ) .
again , little evolution of these relations with redshift is found .
these authors predict a steepening of the stfr slope below a stellar mass of @xmath5 while the gstfr is expected to have a constant slope .
if instead of the observed rotation we consider the circular velocity characterizing the gravitational potential , these relations can be extended to any type of galaxies , in particular to giant and dwarf ellipticals ( des ) which are not dominated by rotation .
this generalization would allow to probe further the similarities of the dark - matter distribution of early- and late - type galaxies that were already investigated by @xcite .
the rotation curves of ellipticals can not readily be observed since they contain little or no hi @xcite .
@xcite and @xcite determined the circular velocities of bright ellipticals using dynamical models .
@xcite measured stellar rotation curves for a sample of 16 flattened virgo des .
these authors found des to adhere closely to the tfr of gas - rich dwarf and spiral galaxies .
however , they did not correct for asymmetric drift , which , for des , can be as large as the velocity dispersion . in order to obtain more reliable rotation curves of des
, we constructed dynamical models for 13 des from the fornax cluster , nearby southern groups , and the local group , to stellar kinematics out to @xmath6 r@xmath7 .
the local group des were observed with the ohp 1.93-m telescope @xcite .
the other des were observed in the course of eso large programme 165.n 0115 ( see e.g. @xcite ) .
we use the observed surface brightness distribution , the mean velocity , the velocity dispersion , and , if available , the central fourth order moment of the line - of - sight velocity distributions calculated from the kinematic parameters up to @xmath8 @xcite , as data .
a detailed account of the modeling method can be found in @xcite , @xcite , @xcite , and @xcite . in @xcite , the model for fs373
is discussed ; in @xcite , we present the models for ngc147 , ngc185 , and ngc205 ( including a technical description of the modeling method , a presentation of the data , and a comparison of the models with the data ) . we can define the range of models , and hence mass distributions , that are consistent with the data and determine the best fitting model .
the strong dependence of the model mass profile , and consequently the corresponding circular velocity curve , on the velocity dispersion profiles makes estimates of @xmath9 based on dynamical models much less sensitive to the unknown inclination than @xmath9 estimates based on direct measurements of stellar rotation curves .
there remains the caveat that @xmath9 estimates derived from dynamical models are by construction to some extent model dependent and are the result of the non - trivial conversion of stellar kinematics into a dark - matter density profile .
the new data for the 13 des are presented in table [ tab1 ] .
@xmath10 is the maximum circular velocity of the best fitting model ; @xmath11 and @xmath12 are the lowest and highest maximum circular velocities of models that are consistent with the data at the 90% confidence level .
all velocities are expressed in km / s . @xmath13 and @xmath14 are the b and 2mass ks band absolute magnitudes , respectively .
the maximum extent of the kinematic data in units of the half - light radius is indicated by @xmath15 .
the des with `` fcc '' designations are taken from the @xcite fornax cluster catalog ; `` fs '' refers to the @xcite catalog of southern groups ; ngc5898_dw1 and ngc5898_dw2 are two des in the ngc5898 group @xcite . throughout this paper ,
we use @xmath16 km s@xmath17 mpc@xmath17 .
the b - band tfrs traced by early and late - type galaxies are plotted in the left panel of fig . [ tf ] .
the black spiral symbols represent late - type galaxies taken from @xcite ( tp00 ) , @xcite ( c00 ) , @xcite ( m05 ) , and @xcite ( g06 ) .
if galaxies appear in more than one data set , we use only the tp00 data . based on 115 galaxies ,
tp00 find @xmath18 , with @xmath19 expressed in solar b - band luminosities and @xmath9 in km s@xmath17 .
we fitted a straight line to the combined tp00 and m05 data sets , taking into account the errors on the luminosities @xmath20 and on the circular velocities @xmath9 , using the routine fitexy of @xcite .
the diagonal elements of the covariance matrix are used as approximations of the variances of the regression coefficients .
going back to the original papers from which the m05 data set is compiled , the average error on @xmath9 is @xmath21 km s@xmath17 .
we adopt a 15% error on the total luminosity , roughly accounting for the various sources of statistical and systematic errors . limiting ourselves to the 128 galaxies brighter than @xmath22 , or @xmath23 mag , we find the relation @xmath24 at lower luminosities ,
the situation becomes very unclear .
the m05 late - types fall systematically below the tfr whereas the g06 galaxies lie above it . this may be due to an increased scatter about the tfr at low luminosities .
@xcite note that at @xmath25 , turbulent gas motions start dominating the ordered rotation , causing the scatter about the tfr to increase dramatically .
they also suggest that at that point the notion of a thin , well - aligned gas disk might break down . on the other hand ,
even if low - mass galaxies are supported by turbulence rather than by rotation , one would expect some kind of tfr to persist , even though the underlying equilibrium of these fainter systems might be different .
we fitted a straight line to the data of the early - type galaxies , taking into account the errors on the luminosities @xmath20 and on the circular velocities @xmath9 .
this data - set consists of the luminosities and the circular velocities of bright ellipticals , estimated by @xcite ( k00 ) and @xcite ( mb01 ) from spherical dynamical models , and des from @xcite and @xcite ( d06 ) . in case where galaxies appear in both the k00 and mb01 data sets
, we opted to use the k00 data because these models allow for a radially varying anisotropy .
we note that the k00 and mb01 @xmath9 estimates of overlapping galaxies are in good agreement . for the luminosities , as for the late - type galaxies , we assume a 15% error
; for the circular velocities , we use the 90% confidence level uncertainties given by the various authors .
we find that the tfr of the early - type galaxies can be well represented by a single power - law over 3 decades in luminosity : @xmath26 within the error bars , the b - band tfrs of early - type and late - type galaxies have the same slope . in the de - regime , at about @xmath27 , or @xmath28 mag , late and early - type dwarfs essentially overlap in a @xmath20 versus @xmath29 diagram . in the b - band , ellipticals ,
in the regime defined by @xmath30 , or @xmath31 to @xmath32 mag , are about a factor of @xmath33 , or about 1.5 mag , fainter than spiral galaxies with the same @xmath9 .
the k - band tfr of early and late - type galaxies is plotted in the right panel of fig .
[ tf ] , using 2mass ks - band magnitudes for the dwarf and giant early - type galaxies .
tp00 find @xmath34 for the k - band tfr of late - type galaxies .
our fit to the k - band tfr of the d06 , k00 , and mb01 galaxies yields @xmath35 in the k - band , ellipticals are roughly a factor of 3 , or @xmath36 mag , fainter than late - types with the same @xmath9 .
going from the well - known tfr to the stellar mass tully - fisher relation ( stfr ) requires the conversion of luminosities , and , if available , colours , to stellar masses , @xmath37 . @xcite
have fitted a suite of spectrophotometric disk evolution models to a set of observed properties of late - type galaxies .
the acceptable models produce a tight correlation between the stellar mass - to - light ratio ( @xmath38 ) and colour .
alternatively , one can use the stellar @xmath38 that gives the best mond fit to the rotation curves or use the maximum disk @xmath38 @xcite .
these @xmath38 estimates generally agree to within of a factor of 2 . in short ,
the @xmath38s and stellar masses of late - type galaxies can be estimated straightforwardly from their colours or rotation curves . here
, we use the @xmath38 based on the properties of the stellar population ( colours , ages , metallicities ) .
the uncertainty on @xmath37 , which is the combined uncertainty on @xmath19 and @xmath38 , can be quite substantial and we estimate it to be of the order of 100% , on average . no easy - to - use tool for estimating @xmath38s of early - type galaxies exists .
however , the mass - metallicity relation of early - type galaxies is observationally well constrained , from the faintest dwarfs up to the brightest giants , using either luminosity @xcite or velocity dispersion @xcite as a substitute for galaxy mass .
the thoretical predictions for this relation and the observations are in reasonably good agreement @xcite .
we tried different methods to calculate @xmath38 . _
( i ) _ we fit a 4@xmath39 order polynomial to the empirical luminosity - metallicity relation of @xcite in order to estimate the metallicities of the galaxies in the k00 , mb01 , and d06 data sets . plugging this metallicity , along with an average age of 10 gyr @xcite , in the ssp - models of @xcite or @xcite yields the b - band @xmath38 . _
( ii ) _ for the giant ellipticals , one can use the relation @xmath40 of @xcite between stellar mass @xmath37 , expressed in @xmath41 , and velocity dispersion @xmath42 , expressed in km / s , or , alternatively , _
( iii ) _ the empirical metallicity and age relations , @xmath43 } = -1.06 + 0.55 \times \log(\sigma)$ ] and @xmath44 , of @xcite in combination with the @xcite or @xcite models .
we found all methods to be in excellent agreement .
they have systematic offsets much smaller than the errorbars on the datapoints and yield stfr slopes that agree to within the parameter uncertainties ( see below ) .
we also converted the 2mass ks - magnitudes into @xmath37 using the @xcite models .
this gave results that were entirely consistent with the stfr based on the b - band data .
for the remainder , we adopt approach _ ( iii ) _ for the giant ellipticals and approach _ ( i ) _ for the des . the stfrs of early and late - type galaxies are plotted in the left panel of fig .
[ mass ] . both early and late - type galaxies trace a single yet curved stfr over 3.5 orders of magnitude in stellar mass .
for all galaxies in the range @xmath45 we find @xmath46 using the @xmath47 relation of @xcite yields a stfr slope of @xmath48 , consistent with eq .
( [ stfr_all1 ] ) .
this is in good agreement with theoretical predictions @xcite .
the curvature of the stfr is at least partially responsible for @xcite concluding that early type galaxies have lower stellar masses than late - type galaxies at the same @xmath10 if the stfr of the late - types is extrapolated .
the mass of the gaseous component in late - type galaxies follows from 21 cm observations .
we denote the sum of the stellar and hi gas mass by @xmath49 .
since early - type galaxies do not contain a significant interstellar medium ( see @xcite and references therein , @xcite , @xcite ) , the hi gas+stellar mass tully - fisher relation ( gstfr ) of early - types to a good approximation equals their stfr . for the late - types
, we use the data of @xcite .
the gstfr of the early and late - type galaxies , presented in the right panel of fig .
[ mass ] , is less curved than the stfr , and , for the whole range @xmath50 , can be fitted by the linear relation @xmath51 this can be compared with the gstfr for giant and dwarf late - type galaxies constructed by @xcite , who find a slope @xmath52 .
our result agrees much better with the slope of 3 which one would expect from the virial theorem , assuming a constant virial overdensity and a constant baryon - to - total mass ratio @xcite .
late - types show a vertical scatter of 0.2 dex in @xmath53 about the gstfr , giant early - types have a slightly larger scatter of 0.3 dex .
des are offset downwards by 0.4 dex , probably due to them having lost part of their baryons by galactic winds @xcite or ram - pressure stripping @xcite .
early - type and late - type galaxies trace different yet approximately parallel tfrs , with early - types being roughly 1.5 mag fainter in the b - band than late - types for the same @xmath9 .
surprisingly , all galaxies trace the same stfr and gstfr over a range of 3.5 decades in stellar or hi gas+stellar mass .
des lie slightly below the general gstfr .
this seems to indicate that early - type dwarfs , which , contrary to the late - types , reside in high - density environments have lost their gas due to environmental influences , e.g. by galactic winds or ram - pressure stripping .
this also shows that the environment only plays a secondary role in shaping these relations , making them a `` clean '' cosmological tool .
@xmath0cdm simulations are able to account for the observed slopes of these relations .
wwz acknowledges the support of the austrian science fund ( project p14753 ) .
sdr is a postdoctoral fellow of the fund for scientific research - flanders ( belgium)(f.w.o . ) .
this publication makes use of 2mass data , a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation , the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration , and the leda database ( http://leda.univ-lyon1.fr ) .
we thank the anonymous referee for comments that helped to significantly improve this letter .
bell , e. f. , de jong , r. s. , 2001 , , 550 , 212 bender , r. , burstein , d. , faber , s. m. , 1993 , apj , 411 , 153 bertola , f. , pizzella , a. , persic , m. , salucci , p. , 1993
, , 416 , l45 bruzual , g. & charlot , s. , 2003 , , 344 , 1000 buyle , p. , de rijcke , s. , michielsen , d. , baes , m. , dejonghe , h. , 2005 , , 360 , 853 conselice , c. j. , oneil , k. , gallagher , j. s. , wyse , r. f. g. , 2003 , , 591 , 167 cot , s. , carignan , c. , freeman , k. , 2000 , , 120 , 3027 ( c00 ) dejonghe , h. , & de zeeuw , t. , 1988 , , 329 , 720 dejonghe , h. , de bruyne , v. , vauterin , p. , zeilinger , w. w. , 1996 , , 306 , 363 de lucia , g. , springel , v. , white , s. d. m. , croton , d. , kauffmann , g. , 2006 , , 366 , 499 de rijcke , s. , dejonghe , h. , zeilinger , w. w. , hau , g. k. t. , 2001 , , 559 , l21 de rijcke , s. , dejonghe , h. , zeilinger , w. w. , hau , g. k. t. , 2004 , , 426 , 53 de rijcke , s. , michielsen , d. , dejonghe , h. , zeilinger , w. w. , hau , g. k. t. , 2005 , , 438 , 491 de rijcke , p. prugniel , f. simien , h. dejonghe , 2006 , , 369 , 1321 djorgovski , s. & davis , m. , 1987 , , 313 , 59 ferguson , h. c. , 1989 , aj , 98 , 367 ferguson , h. c. & sandage , a. , 1990 , aj , 101 , 765 geha , m. , guhathakurta , p. , van der marel , r. p. , 2003
, , 126 , 1794 geha , m. , blanton , m. r. , masjedi , m. , west , a. a. , accepted by apj , astro - ph/0608295 ( g06 ) gerhard , o. , kronawitter , a. , saglia , r. p. , bender , r. , 2001 , , 121 , 1936 grebel , e. k. , gallagher , j. s. , iii , harbeck , d. , 2003 , aj , 125 , 1926 kronawitter , a. , saglia , r. p. , gerhard , o. , bender , r. , 2000 , a&as , 144 , 53 ( k01 ) mac low , m. & ferrara , a. , 1999 , , 513 , 142 magorrian , j. & ballantyne , d. , 2001 , , 322 , 702 ( mb01 ) mateo , m. l. 1998 , ara&a , 36 , 435 mcgaugh , s. , schombert , j. m. , bothun , g. d. , de blok , w. j. g. , 2000 , , 533 , l99 mcgaugh , s. , 2005 , , 632 , 859 ( m05 ) michielsen , d. , de rijcke , s. , zeilinger , w. w. , prugniel , p. , dejonghe , h. , roberts , s , 2004 , , 353 , 1293 mori , m. & burkert , a. , 2000 , , 538 , 559 nagashima , m. & yoshii , y. , 2004 , , 610 , 23 osullivan , e. , ponman , t. j. , collins , r. s. , 2003 , , 340 , 1375 portinari , l. & sommer - larsen , j. , 2006 , submitted to , astro - ph/0606531 press , w. h. , teutolsky , s. a. , vetterling , w. t. , flannery , b. p. , 1992
, numerical recipes in c , cambridge university press , new york , us proctor , r. n. , forbes , d. a. , hau , g. k. t. , beasley , m. a. , de silva , g. m. , contreras , r. , terlevich , a. i. , 2004 , mnras , 349 , 1381 prugniel , p. & simien , f. , 1996 , , 309 , 749 rakos , k. , schombert , j. , maitzen , h. m. , prugovecki , s , .
odell , a. , 2001 , 121 , 1974 sanders , r. h. & mcgaugh , s. s. 2002 , ara&a , 40 , 263 simien , f. & prugniel , ph .
, 2002 , , 384 , 371 tassis , k. kravtsov , a. v. , gnedin , n. y. , submitted to , astro - ph/0609763 thomas , d. , maraston , c. , bender , r. , mendes de oliviera , c. , 2005 , , 621 , 673 tully , r. b. & fisher , j. r. , 1977 , , 54 , 661 tully , r. b. & pierce , m. j. , , 533 , 744 ( tp00 ) van der marel , r. p. & franx , m. , 1993 , , 407 , 525 van zee , l. , skillman , e. d. , haynes , p. , 2004
, , 128 , 121 vazdekis , a. , casuso , e. , peletier , r. f. , beckman , j. e. , , 106 , 307 @xmath9 & @xmath12 & @xmath13 & @xmath14 & @xmath54 + fcc046 & 68 & 83 & 108 & -15.48 & -18.51 & 1.5 + fcc150 & 65 & 102 & 113 & -15.77 & -18.44 & 1.2 + fcc207 & 74 & 85 & 113 & -15.20 & -18.50 & 1.2 + fcc245 & 50 & 65 & 80 & -15.47 & / & 0.5 + fcc266 & 70 & 85 & 100 & -15.62 & -18.47 & 1.0 + fcc288 & 79 & 105 & 112 & -15.89 & -18.76 & 2.0 + fs029 & 105 & 112 & 126 & -17.31 & -20.95 & 2.0 + fs373 & 104 & 118 & 142 & -17.50 & -20.82 & 1.6 + ngc5898_dw1 & 61 & 71 & 81 & -16.75 & -19.95 & 2.0 + ngc5898_dw2 & 64 & 91 & 105 & -16.31 & -18.35 & 2.0 + ngc147 & 25 & 41 & 57 & -14.44 & -16.95 & 1.2 + ngc185 & 42 & 49 & 54 & -14.67 & -17.39 & 1.1 + ngc205 & 56 & 68 & 79 & -15.79 & -18.99 & 2.3 +
|
we study the locus of dwarf and giant early and late - type galaxies on the tully - fisher relation ( tfr ) , the stellar mass tully - fisher relation ( stfr ) and the so - called baryonic or hi gas+stellar mass tully - fisher relation ( gstfr ) .
we show that early - type and late - type galaxies , from dwarfs to giants , trace different yet approximately parallel tfrs .
surprisingly , early - type and late - type galaxies trace a single yet curved stfr over a range of 3.5 orders of magnitude in stellar mass .
moreover , _ all _ galaxies trace a single , linear gstfr , over 3.5 orders of magnitude in hi gas+stellar mass .
dwarf ellipticals , however , lie slightly below the gstfr .
this may indicate that early - type dwarfs , contrary to the late - types , have lost their gas , e.g. by galactic winds or ram - pressure stripping .
overall , environment only plays a secondary role in shaping these relations , making them a rather `` clean '' cosmological tool .
@xmath0cdm simulations predict roughly the correct slopes for these relations .
|
applications of slow and stored light based on electromagnetically - induced transparency ( eit ) , in both quantum information processing and optical communication , will benefit from improved efficiency and delay - bandwidth product . in quantum information ,
stored light has emerged as a promising technique for applications such as single - photon generation on demand @xcite and quantum memories @xcite and repeaters @xcite . however
, practical applications will require significant improvements in the efficiency of writing , storing , and retrieving an input photon state beyond values achieved to date @xcite . in classical communications ,
optical buffers require adjustable delay time ( i.e. , group index ) for input signal pulses with minimal pulse distortion and loss @xcite and high compression of the input pulse for high data density inside the delay medium .
however , high efficiency and large delay - bandwidth product have not fully been realized in all - optical systems .
here we present a preliminary study of the optimization of slow and stored light using eit in warm rb vapor as a function of optical depth , laser power , and input pulse bandwidth . in eit , a weak signal pulse propagates with very slow group velocity through
an otherwise absorbing medium @xcite , which is established by a strong control field
. the reduced group velocity with which a signal propagates under eit is the result of the transfer of photonic excitation in the pulse to the collective spin coherence of the atoms .
the group velocity is proportional to the control field intensity , and can be reduced to zero by shutting off the control light . in this case the photonic information of the pulse is completely transferred to the atoms .
this coherent transfer is reversible , and thus after some storage time the pulse may be retrieved , converting it back into photonic excitation when the control light is turned back on .
optimization of slow and stored light requires reducing the absorption of signal light and spin decoherence , as well as increasing optical depth to achieve a large delay - bandwidth product @xcite . in warm atom vapor cells ,
signal light absorption is typically driven by incomplete polarization of the atomic medium due to such mechanisms as radiation trapping and competing nonlinear processes .
atomic decoherence arises from mechanisms such as collisions with buffer gas atoms and cell walls , diffusion out of the laser beam , and residual magnetic field gradients .
modeling the atomic medium as a collection of stationary three - level @xmath0-systems ( fig .
[ levelsandsetup]a ) , we find that slow and stored light properties improve with increasing rubidium density or optical depth .
the eit bandwidth , @xmath1 , for moderate to large optical depths is found to be @xcite @xmath2 where @xmath3 is the rabi frequency of the control field , @xmath4 the excited - state line width , @xmath5 the total number of three - level atoms , @xmath6 the length of the medium ( the optical depth , @xmath7 ) , @xmath8 the light - atom coupling coefficient , and @xmath9 the speed of light . similarly , the group velocity @xmath10 resulting from the steep resonant dispersion for the signal pulse in the slow light medium is @xcite @xmath11 the absolute pulse delay , @xmath12 , is then given by the length of the medium divided by the group velocity and we find that @xmath13 , that is the delay is proportional to the optical depth .
a careful theoretical study of storage efficiency @xmath14 ( defined as the ratio of the number of output photons to the number of input photons ) , optimized for a given optical depth @xcite , finds that the optimal temporal pulse width @xmath15 , and finds that @xmath14 scales as @xmath16 at high optical depth , and increases more rapidly at lower @xmath17 .
this treatment implies continued efficiency improvement with increasing optical depth . in practice
, experiments have found that slow and stored light efficiency improves with increasing optical depth for modest optical depths . however , at larger optical depths , slow and stored light efficiency reach a maximum and then degrade .
note that a procedure for optimizing stored light efficiency for a given @xmath17 , using the time - reversed output pulse to determine the shape of the input pulse , has been developed theoretically @xcite and directly implemented experimentally @xcite . in the present study
we report preliminary experimental investigations of the behavior of slow and stored light over a wide range of optical depths , by varying the temperature of rb in a vapor cell .
measuring coherence loss rates and including them in the efficiency optimization model yields good agreement between theory and experiment as @xmath14 falls off with increasing atomic density .
we focus on the role played by optical depth in coherence loss , in particular the effect of radiation trapping , where incoherent photons are repeatedly re - absorbed by the medium before exiting @xcite .
we also report promising preliminary efficiency results for a new cell geometry designed to reduce coherence loss due to radiation trapping .
we measured eit and slow and stored light , using the @xmath18 transition in a warm @xmath19rb vapor .
the relevant energy levels and coupling fields are shown in fig .
[ levelsandsetup]a : a strong control field is resonant with the @xmath20 transition and a weak signal field is resonant with @xmath21 .
the optical fields were generated by an external cavity diode laser tuned to @xmath22 nm , amplified by a tapered amplifier , and spatially filtered by a pinhole ( fig .
[ levelsandsetup]b ) .
an electro - optic modulator ( eom ) phase modulated the laser at the ground state hyperfine splitting ( @xmath23 ghz ) ; the @xmath24 sideband acted as the signal field , with a maximum signal to control intensity ratio of @xmath25 ; the sideband amplitude was varied to shape the desired signal pulse .
the frequency could be varied to change the two - photon detuning @xmath26 of the eit transition for measuring its line shape .
the overall intensity was regulated with an acousto - optic modulator ( aom ) , which also shifted the frequency by @xmath27 mhz . a quarter - wave plate ( @xmath28 ) before the cell converted the beam to circular polarization .
the collimated beam entering the cell had a diameter of approximately @xmath29 mm .
the output fields were measured by sending the laser into an optical fiber , then into a fast photodetector ( pd ) along with a reference beam picked off prior to the eom and aom . the control field , signal field , or the unused @xmath30 sideband
could thus be measured by beat note detection with a spectrum analyzer , at @xmath31 mhz , @xmath32 ghz @xmath33 mhz , and @xmath32 ghz@xmath34 mhz respectively .
the transmission of the @xmath30 sideband ( off eit resonance ) , which is @xmath35 ghz away from any rb resonance , was used as a reference for determining slow light delay or storage efficiency , where @xmath36 was the difference between pulse peak times and efficiency @xmath37 ( signal pulse area)/(@xmath30 sideband area ) .
the cell was housed in a plastic oven , which was heated by blown warm air .
the cell temperature was varied between @xmath38 and @xmath39c , or atomic number density between @xmath40 and @xmath41 @xmath42 .
three layers of cylindrical high - permeability shielding surrounded the oven to screen out stray laboratory magnetic fields , and a solenoid inside was used to cancel any small constant background field .
three rb vapor cells were used in our experiments ; all three contained isotopically enriched @xmath19rb and were filled with buffer gas to confine atoms and extend their coherence life times . the cell used for all measurements except where specifically noted had length @xmath43 cm and diameter @xmath44 cm , and was filled with @xmath38 torr of ne .
we studied slow and stored light as a function of optical depth by measuring the delay - bandwidth product and efficiency of `` optimized '' pulses as the vapor cell temperature was varied .
such pulses were constructed through an iterative optimization procedure in which a signal pulse was stored and retrieved , its output shape measured and a new pulse stored with the shape of the output pulse .
this process is repeated until convergence .
this procedure has been shown both theoretically @xcite and experimentally @xcite to produce slow - light pulse shapes that optimize storage efficiency for the given optical depth and laser power . in effect
this procedure selects a pulse bandwidth that balances fractional delay and absorption to yield the largest output signal pulse .
measured eit bandwidths and slow light delays scaled as expected with optical depth .
figure [ eitanddelays]a shows eit linewidth ( fwhm , as extracted from a lorentzian fit ) vs. atomic density , which follows the @xmath45 trending expected from eq . [ e.eitwidth ] .
figure [ eitanddelays]b shows that absolute pulse delay @xmath46 ; and figure [ eitanddelays]c shows that the optimal pulse width @xmath47 .
we measured line widths and delays at two control field powers , @xmath48 mw and @xmath49 mw , corresponding to respective intensities of @xmath50 and @xmath51 mw/@xmath52 and rabi frequencies of @xmath53 and @xmath50 mhz .
slow light delays in figs .
[ eitanddelays]b and [ eitanddelays]c are tailored for optimized stored light , obtaining neither the maximum absolute delay ( achieved for a temporally long pulse experiencing only the steepest part of the dispersion , but having a small fraction of the pulse in the medium at one time ) nor the maximum fractional delay ( achieved for a temporally short pulse which suffers large absorption due to frequency components outside the eit bandwidth ) . comparing the optimized pulse bandwidth ( where @xmath54 is the temporal pulse width ) to the measured eit bandwidth ( fig .
[ eitanddelays]d ) , we find that @xmath55 , over a large range of fractional delays ( from @xmath56 to @xmath57 ) , a somewhat greater pulse bandwidth than expected theoretically @xcite , where @xmath58 or even @xmath59 .
note that based on the scalings of @xmath60 and @xmath15 , we would expect the optimal frequency bandwidth to be quadratic in the eit line width which is not clearly observed in fig .
[ eitanddelays]d .
c and uncertainty in the laser frequency of @xmath61 mhz .
dashed lines are fits of data to the simplified three - level model described in text . ]
stored light efficiency was measured for a range of rubidium densities between @xmath62 @xmath42 ( @xmath63c ) and @xmath64 @xmath42 ( @xmath39c ) and storage intervals from @xmath65s to @xmath66 ms and plotted for @xmath67s in fig .
[ theoryexp ] .
the efficiencies peak at temperatures between @xmath68 and @xmath69c , falling at higher optical depth .
the @xmath70 coherence lifetimes during storage ( fig .
[ oneoveretimes ] ) also fall at high density .
this points to possible high - density or high - temperature phenomena adversely affecting slow and stored light .
increased loss during storage and readout at high optical depth agrees with predictions from theoretical simulations when the density dependence of decoherence is included ( fig .
[ theoryexp ] ) .
the simulations , described fully in earlier work @xcite , consist of iterated solutions to the stored light dynamics equations for an ensemble @xmath0-system . the dashed curve in fig .
[ theoryexp ] uses the measured rates of decoherence during storage and readout taken from fig .
[ oneoveretimes ] as an input parameter for the spin - wave decay rate , and most closely matches the experimental results . in figure [ accounting ]
we verify that the loss mechanisms for stored light are all self - consistent . a slow light pulse that is not trapped in the medium ( the control field
is left on ) has loss from imperfect eit transmission and from pulse bandwidth lying outside the region of maximum transmission ( after normalization to a reference pulse , this is the `` slow - light efficiency '' ) .
a stored and retrieved pulse will experience that same loss , and additional loss from two sources : ( i ) part of the pulse escapes the medium before the pulse is stored ( fractional delay less than unity implies significant losses from this mechanism ) ; and ( ii ) atomic coherence decay during the storage and readout periods .
figure [ accounting]a shows the directly measured slow light efficiency @xmath71 , the area ratio of the output pulse to the reference pulse , and an inferred @xmath72 from the storage measurement . using the measured decoherence rate from fig .
[ oneoveretimes ] and the storage readout areas at @xmath67s , we can back out the storage readout at @xmath73 ; combining this with the area of the not - stored pulse leakage we infer @xmath71 ( see fig .
[ accounting]b ) : @xmath74 where @xmath75 is the leaked pulse area divided by the reference pulse area , and @xmath76 is the atomic coherence lifetime from fig .
[ oneoveretimes ] .
the measured and inferred slow light efficiencies match very well , only deviating significantly at the highest atomic density , where the term with exp@xmath77s@xmath78 becomes very sensitive to the short coherence lifetime measurement , which is more error prone due to the low snr in the signal transmission . as indicated from both the storage efficiency and the comparison between slow and stored light , at high rb densities the effectiveness of the medium
is reduced .
we believe that several mechanisms are responsible for this .
first , at higher densities , temperature or density - dependent losses such as rb - rb spin exchange become more significant , contributing to decoherence even during the storage interval ( see fig .
[ oneoveretimes ] ) .
second , in the presence of light at high rb densities , radiation trapping becomes significant : light absorbed and then incoherently radiated by one rubidium atom is reabsorbed by another rubidium atom , and this process leads to decoherence and loss of efficiency .
in the next sections , we present preliminary investigations of high - density losses .
direct measurements of large optical depths are difficult due to the large corresponding absorption of the optical field .
to characterize the optical depth for comparison with the optimization models described above , we used optical absorption line width as a proxy for the optical depth .
the signal field was detuned from two - photon resonance by changing the frequency driving the eom and the signal intensity monitored as the one - photon ( laser ) detuning was swept through resonance .
the line width of the signal field absorption serves as a proxy for the optical depth because the logarithm of the measured absorption is proportional to the product of the detuning and the optical depth .
figure [ radtrap]a shows a comparison of the measured one - photon linewidth and the expected linewidth from a simple numerical model based on an optically thick sample of three - level @xmath0-systems . at high densities ( and optical depths ) , the measured linewidth stops growing while the model shows continued line broadening .
we attribute this to radiation trapping : photons scattered inside the medium depolarize the rubidium ground state .
fewer atoms are polarized in the signal field channel , leading to bothdecoherence and absorption of the signal field .
we also directly measured the presence of radiation trapping in our vapor cells , with a simplified apparatus .
light from an extended cavity diode laser was focused through an aom with a small spot size to allow for rapid rise times in the aom .
this aom output was turned on and off at a rate of @xmath79 mhz with a rise time of @xmath80 ns .
the beam diameter was then expanded to approximately @xmath81 mm and sent into a rb vapor cell housed inside a small plastic oven , unshielded from stray laboratory magnetic fields , and warmed with resistive heaters to temperatures between @xmath82 and @xmath83c .
output light from the cell was measured with two photodetectors : one for transmission , and the other for fluorescence mounted to the side of the cell oven where a lens close to the cell focused scattered light onto the detector . when light was turned on via aom modulation , the rise time of the signal was observed in both detectors .
evidence for radiation trapping is shown in fig . [ radtrap]b . at high atomic densities , the rise time in the fluorescence
detection signal increases significantly .
the greater the transverse optical depth , the greater the number of absorption and re - emission cycles that occur before absorbed photons exit the medium , so the rise time seen in fluorescence detection should grow with the transverse optical depth . in fig .
[ radtrap]b we see rise times up to @xmath84 ns , whereas the lifetime for the excited state of the @xmath85 state of rb is @xmath86 ns , indicating many scattering events for each absorbed photon .
in order to minimize the decoherence resulting from radiation trapping in our vapor cells , we recently developed a cell designed to minimize its effects .
a 15 cm long , 1.2 cm diameter cell with n@xmath87 buffer gas was designed to reduce radiation trapping in three ways : ( i ) the aspect ratio is four times larger than previous cells , allowing fluorescence to escape the cell in the transverse direction with fewer depolarizing interactions with rb atoms ; ( ii ) the absolute length of the cell is twice that of previous vapor cells allowing the equivalent optical depth to be reached at lower rb density , minimizing density - dependent effects , including spin exchange ; ( iii ) n@xmath87 buffer gas acts to quench radiation trapping by collisionally de - exciting rb atoms before the atoms fluoresce , preventing the emission of unpolarized photons that could destroy the coherence of nearby atoms .
initial results of stored light efficiency in this cell appear promising . for a laser power of @xmath88 mw and a temperature of @xmath89c ( the same longitudinal optical depth as the density @xmath90 @xmath42 data in fig .
[ theoryexp ] ) , efficiency of @xmath91 was achieved .
this is three times greater than the @xmath14 for @xmath92 mw and twice as great as for @xmath93 mw at the same optical depth .
future measurements will further explore this promising avenue towards high - efficiency light storage in rb vapor cells .
we reported a preliminary experimental study of slow and stored light at a variety of rb densities ( i.e. , optical depths ) , and as a function of pulse duration , delay and storage time , and control field intensity .
experimental results are in good qualitative agreement with theoretical calculations based on a simplified three - level model at moderate densities ; and indicate that radiation trapping is an important limitation at high atomic density .
we have developed a new cell geometry which we expect to reduce losses and improve efficiency through reduced radiation trapping and density - dependent decoherence .
99 c. w. chou , s. v. polyakov , a. kuzmich , and h. j. kimble , phys .
92 * , 213601 ( 2004 ) .
d. n. matsukevich , and a. kuzmich , science * 306 * , 663 ( 2004 )
. m. d. eisaman , l. childress , a. andr , f. massou , a. s. zibrov , and m. d. lukin , phys . rev
. lett . *
93 * , 233602 ( 2004 ) .
d. n. matsukevich , t. chanelire , s. d. jenkins , s .- y .
lan , t. a. b. kennedy , and a. kuzmich , phys .
lett . * 97 * , 013601 ( 2006 ) .
s. chen , y .- a .
chen , t. strassel , z .- s .
yuan , b. zhao , j. schmiedmayer , and j .- w .
pan , phys .
lett . * 97 * , 173004 ( 2006 ) . c. w. chou , h. de riedmatten , d. felinto , s. v. polyakov , s. j. van enk , and h. j. kimble , nature * 438 * , 828 ( 2005 ) .
t. chanelire , d. n. matsukevich , s. d. jenkins , s .- y .
lan , t. a. b. kennedy , and a. kuzmich , nature * 438 * , 833 ( 2005 ) .
m. d. eisaman , a. andr , f. massou , m. fleischhauer , a. s. zibrov , and m. d. lukin , nature * 438 * , 837 ( 2005 ) .
t. chanelire , d. n. matsukevich , s. d. jenkins , t. a. b. kennedy , m. s. chapman , and a. kuzmich , phys .
lett . * 96 * , 093604 ( 2006 ) .
l. m. duan , m. d. lukin , j. i. cirac , and p. zoller , nature * 414 * , 413 ( 2001 ) .
d. felinto , c. w. chou , j. laurat , e. w. schomburg , h. de riedmatten , and h. j. kimble , nature phys .
* 2 * , 844 ( 2006 ) .
i. novikova , m. klein , d. f. phillips , and r. l. walsworth , proc .
spie * 5735 * , 87 ( 2005 ) .
d. felinto , c. w. chou , h. de riedmatten , s. v. polyakov , and h. j. kimble , phys . rev .
a * 72 * , 053809 ( 2005 ) .
j. laurat , h. de riedmatten , d. felinto , c .- w .
chou , e. w. schomburg , and h. j. kimble , opt
* 14 * , 6912 ( 2006 ) .
j. k. thompson , j. simon , h. loh , and v. vuleti , science * 313 * , 74 ( 2006 ) .
r. s. tucker , p. s. ku , and c. j. chang - hasnain , j. lightwave technol .
* 23 * , 4046 ( 2005 ) .
l. v. hau , s. e. harris , z. dutton , and c. h. behroozi , nature * 397 * , 594 ( 1999 ) .
m. d. lukin , rev .
phys . * 75 * , 457 ( 2003 ) .
a. v. gorshkov , a. andre , m fleischhauer , a. s. sorensen , and m. d. lukin , phys .
lett . * 98 * , 123601 ( 2007 ) ; a. v. gorshkov , a. andre , m. d. lukin , and a. s. sorensen , phys .
a * 76 * , 033805 ( 2007 ) .
i. novikova , a. v. gorshkov , d. f. phillips , a. s. sorensen , m. d. lukin , and r. l. walsworth phys .
98 * , 243602 ( 2007 ) ; i. novikova , a. v. gorshkov , d. f. phillips , y. xiao , m. klein , and r. l. walsworth , proc .
spie * 6482 * , 64820 m ( 2007 ) .
a. f. molisch and b. p. oehry , _ radiation trapping in atomic vapors _ ( oxford university press , usa , 1999 ) ; a. b. matsko , i. novikova , m. o. scully , and g. r. welch , phys .
lett . * 87 * , 133601 ( 2001 ) ; a. b. matsko , i. novikova , and g. r. welch , j. mod .
opt . * 49 * , 367 ( 2002 ) .
|
we present a preliminary experimental study of the dependence on optical depth of slow and stored light pulses in rb vapor .
in particular , we characterize the efficiency of slow and stored light as a function of rb density ; pulse duration , delay and storage time ; and control field intensity .
experimental results are in good qualitative agreement with theoretical calculations based on a simplified three - level model at moderate densities .
|
the measurement of the angular power spectrum of the cmb anisotropies , @xmath0s , has become one of the most important tools in modern cosmology . as long as they remain in the linear regime , the fluctuations predicted by most inflationary scenarii @xcite lead to gaussian anisotropies on the cmb .
thus the angular power spectra in temperature and polarization contain all the cosmological information on the cmb sky .
cosmological parameters and other physical quantities of interest in the early universe can be directly derived from them . in parallel to the explosion of cmb datasets both in size and quality ( wmap @xcite , archeops @xcite , boomerang @xcite , maxima @xcite , dasi @xcite , vsa @xcite , cbi @xcite , acbar @xcite ) , fast codes have been developed to estimate the cmb angular power spectrum ( cmbfast @xcite , camb @xcite ) allowing us to compare fast and efficiently theory and observations using powerful statistical tests ( cmbeasy @xcite , cosmomc @xcite , @xcite ) .
furthermore , huge efforts are undertaken to ease the estimation of the angular power spectrum from input cmb maps in order to cope with larger , deeper and more complex sky surveys in a reasonable amount of computing time .
excluding very specific methods for example those which are under study for the planck satellite mission and which take advantage of the planck ring scanning strategy @xcite most cmb power spectrum estimators can be grouped into two categories : maximum likelihood and ` pseudo'-@xmath0 estimators . a complete review and comparison between the two methods
can be found in @xcite , here we just discuss the key points of each of them .
maximum likelihood methods for temperature anisotropies @xcite are based on the maximization of the quadratic likelihood .
the method estimates the sky angular power spectrum from the angular correlation function of the data .
error bars for the power spectrum are generally computed directly from the likelihood function which is either fully sampled in the range of interest or approximated by a quadratic form .
dealing with an inhomogeneous coverage of the sky involves a great computational complexity where @xmath1 is the number of pixels of the input map .
therefore , these methods are very cpu time consuming for current large datasets like wmap and probably not well adapted @xcite for future satellite missions like planck which will produce maps of the sky of more than @xmath2 pixels . a generalization of these methods to the analysis of cmb polarization is discussed in @xcite . alternatively
, the so - called
pseudo-@xmath0 s estimators compute directly the _ ` pseudo ' _ angular power spectrum from the data .
then , they correct it for the sky coverage , beam smoothing , data filtering , pixel weighting and noise biases .
a comprehensive description of this method was first given by @xcite and an application to the angular clustering of galaxies can be found in @xcite .
more recently , several approaches to this method have been developed . among them , spice @xcite and its extension to polarization @xcite compute in the real space first the correlation function to correct for the sky coverage bias and then the power spectrum from the latter .
a pseudo-@xmath0 estimator in the spherical harmonic space applied to cmb experiments is given in @xcite and @xcite ( master ) .
they computes directly the power spectrum before correct it for the different biases .
an approach applied to apodised regions of the sky is presented in @xcite and extended to polarization in @xcite .
these estimators can be evaluated using fast spherical harmonic transforms @xmath3 and therefore provide fast and accurate estimates of the @xmath0s .
however , they require an accurate knowledge of the instrumental setup and noise in order to correct them for the biases discussed previously .
in fact , they use an estimation of the power spectrum of the noise in the map , generally computed via monte - carlo simulations , which is subtracted from the original power spectrum .
this is also used to estimate the error bars in the power spectrum by calculating the variance of the @xmath0s over the set of simulations . in this paper , we describe a method to estimate the @xmath0s by computing the cross - power spectra between a collection of input maps coming either from multiple detectors of the same experiment or from different instruments .
the ` pseudo ' cross - power spectra are explicitly corrected for incomplete sky coverage , beam smoothing , filtering and pixelization . assuming no correlation between the noise contribution from two different maps , each of the corrected cross - power spectra is an unbiased estimate of the @xmath0s .
analytical error bars are derived for each of them .
the cross - power spectra , that do not include the classical _ auto_-power spectra , are then combined using a gaussian approximation of the likelihood function . in the same way
, we can also compute the estimate of the common angular power spectrum , @xmath4 , of sky maps from different experiments . +
a similar method also based on the combination of a set of cross - power spectra has been first used to obtain recent results from the first year wmap data @xcite .
the main difference between the method presented is this paper and the wmap one is the determination of the cross - correlation matrix ( see sect.[correlation ] ) of the corrected cross - power spectra used both for the combination of these into a single power spectra and for the estimation of the error bars on the latter .
the wmap team estimates the cross - correlation matrix from a model of the data .
this includes specific terms related to the wmap data such as the contribution from point sources and the uncertainties on the beam window functions as well as a term related to the cmb anisotropies which is estimated from a fiducial model .
the wmap cross - correlation matrix , used for the combination of the cross - power spectra , does not incorporate the effects of mode coupling . in a further step , they account for the mode coupling and the dependence on the fiducial cmb model for the computation of the uncertainties on the final power spectrum .
by contrast , the method presented here computes the cross - correlation matrix directly from the cross - power spectra estimated from the data .
this allows us to include naturally the mode coupling in this matrix .
further , this permits the computation of analytical error bars ( as described in sect .
[ correlation ] ) which are very compatible with those obtained from simulations ( see sect . [ archeops ] ) .
because of the above , this method can be applied without modification to the estimation of the power spectrum of the correlated signal between a set of maps of the sky coming from multiple instruments with potentially different sky coverages .
for example , we have used this method on archeops data for the estimation of the cmb angular power spectrum and the contribution from foregrounds to this one @xcite .
we have also used it for the estimation of the foreground emission at the sub - millimeter and millimeter wavelength by cross - correlating the archeops data with foreground dust templates . in sect .
[ cross ] , we remind to the reader the computation of the cross - power spectra from ` pseudo ' cross - power spectra . in sect .
[ correlation ] , we specify the correlation between cross - power spectra and between multipoles . analytical expressions for the error bars and the covariance matrix for each cross - power spectra are derived .
section [ combination ] discusses the combination of the cross - spectra from either a single full data set ( sect .
[ lincomb ] ) or several independent experiments ( sect .
[ common ] ) .
finally , xspect is applied to simulations of the archeops balloon - borne experiment in sect .
[ archeops ] .
under the assumption of uncorrelated noise between detectors , the cross - power spectrum computed from sky maps is a non noise - biased estimate of the angular power spectrum . in general , when computing the @xmath0s , other instrumental effects as beam smoothing , incomplete sky coverage and time ordered data filtering need to be taken into account .
these effects and the way to correct them have been deeply covered in the literature for classical ` pseudo-@xmath0 ' estimators @xcite which are noise biased as they use directly the auto power spectrum of each detector map .
as shown in the following , these corrections can be extended to the case of cross - power spectra .
the cmb temperature anisotropies , @xmath5 , over the full - sky can be decomposed into spherical harmonics as follows , @xmath6 where the coefficient @xmath7 are given by @xmath8 the cmb temperature field @xmath5 predicted by most inflationary models @xcite is in general gaussian distributed so that the ensemble average of the @xmath7 coefficients are @xmath9 an unbiased estimate of the cmb temperature power spectrum @xmath10 is therefore given by @xmath11 ground - based and balloon - borne cmb experiments present an inhomogeneous sky coverage .
on the other hand , satellite experiments like cobe , wmap and planck , although they provide full - sky maps , residuals of foreground contamination in the galactic plane and point sources contamination make impossible to use the complete maps when computing the cmb angular power spectrum .
furthermore , for most cmb experiments , the noise properties vary considerably due to different redundancies between pixels of the same map . so obtaining an estimate of the power spectrum
requires a differential weighting of pixels within the same map which translates into an effective inhomogeneous sky coverage .
the decomposition in spherical harmonics of the observed temperature anisotropies including weighting can be written for a single detector as follows @xmath12 where @xmath13 for equal pixels ( as in healpix pixelization , @xcite ) and @xmath14 is the mask applied to the input sky temperature @xmath15 .
this temperature can be decomposed into signal and noise , @xmath16 which are assumed to be uncorrelated , _
@xmath17 the effect of a non - homogeneous coverage of the sky can be described in spherical harmonics @xcite by a mode - mode coupling matrix @xmath18 which depends only on the angular power spectrum of the weighting scheme ( hereafter weighting mask ) applied to the sky to account for the incomplete coverage , the removal of the galactic plane and the inhomogeneous noise properties of the detector .
thus , the ` pseudo-@xmath0 ' estimator @xmath19 is defined as follows @xmath20 the relation that links the ` pseudo ' power spectrum , directly measured on the sky , and the power spectrum @xmath0 of the cmb anisotropies is given by @xmath21 where @xmath22 is the beam transfer function describing the beam smoothing effect ; @xmath23 is the transfer function of the pixelization scheme of the map describing the effect of smoothing due to the finite pixel size and geometry ; @xmath24 is an effective function that represents any filtering applied to the time ordered data ; and @xmath25 is the noise power spectrum .
an unbiased estimate of the cross - power spectrum @xmath26 between the full sky maps of two independent and perfect detectors @xmath27 and @xmath28 can be obtained from @xmath29 where @xmath30 and @xmath31 are the coefficients of the spherical harmonic decomposition of maps @xmath27 and @xmath28 respectively . in the same way
, we can compute the ` pseudo ' cross - power spectrum @xmath32 between any two detectors @xmath27 and @xmath28 by generalizing eq .
[ pseudo_cl ] @xmath33 each of the terms in eq .
[ pseudo_cross ] is described in more details in the following subsections .
the main advantage of using cross - power spectra is that the noise is generally uncorrelated between different detectors @xmath35 this assumption will be maintained throughout this paper .
thus , the cross - power spectra are straightforward estimates of the angular power spectrum on the sky and for two different detectors ( _ i.e. _ @xmath36 ) , the ` pseudo ' cross - power spectrum reads @xmath37 the coupling kernel matrix @xmath18 , introduced in eq .
[ pseudo_cl ] and described in details in @xcite , reads @xmath39 where @xmath40 is the power spectrum of the mask .
it takes into account the mask applied to the data where mask represents both the sky coverage and the weighting scheme .
equation [ mll_master ] can be easily generalized to the case of two different masks applied respectively to each map of the two detectors involved in the cross - power spectrum calculation .
replacing the quadratic terms @xmath41 by @xmath42 in the computation of eq .
[ mll_master ] leads to @xmath43 where @xmath44 , the cross - power spectrum of the masks .
this property allows us to deal with independent masks representing different sky coverages and to apply an appropriate specific weighting scheme to each detector map .
note that the correction in the multipole space discussed here is fully equivalent to an appropriate normalization of the cross correlation between the two sky masked maps in real space .
this analogy is important as it helps to understand why no fully overlapping masks for the input maps can be considered .
the filter function @xmath24 accounts for the filtering of the time ordered data which is generally needed in most cmb experiments either to avoid systematic effects or to reduce correlated low frequency noise . the time domain filtering is performed along a preferred direction on
the sky ( scanning direction ) and so leads commonly to an anisotropic sky even if the assumption of initial isotropic temperature fluctuations holds .
in this case , the estimates of the angular power spectrum provided by eq . [ pseudo_cl ] and eq .
[ pseudo ] are not exact any more and should be corrected for a function both in @xmath46 and @xmath47 , @xmath48 . obtaining accurate estimates of such
a correction is particularly complex and for most cases , as proposed by @xcite , the correction for an effective @xmath49 is good enough for the accuracy required in the reconstruction of the cmb power spectrum
. the @xmath49 function can be , for example , computed via monte - carlo simulations of the sky from which mock time ordered data are produced for each of the detectors involved and then filtered . from an initial theoretical cmb power spectrum , we compute a large number of realizations of the sky using the healpix software _ synfast _
@xcite and compute mock time ordered data from the scanning strategy of each detector .
maps are then computed with and without filtering before re - projection .
the @xmath24 function is obtained from the mean ratio of the ` pseudo ' power spectra of the filtered and not filtered maps .
the latter are obtained using the healpix software _
anafast_. in the case of the cross - power spectra and considering the previous approximation , an effective filter function @xmath50 will be considered in the following .
as defined , the effective filtering function allows us to consider detectors for which the time domain filtering is different .
note that , for nearly white noise and all - sky surveys such as wmap or planck missions , filtering may not be required and thus @xmath51 .
the beam window function @xmath22 describes the smoothing effect of the main instrumental beam under the hypothesis of circularity .
the latter does not hold in general as for most experiments the main beam pattern is asymmetric . as beam uncertainties have become the most important source of systematic errors , taking into account the asymmetry of the beam pattern is necessary .
several solutions have been proposed either circularizing the beam @xcite or assuming an elliptical gaussian beam @xcite .
the work of @xcite presents how to convolve _ exactly _ two band limited but otherwise arbitrary functions on the sphere - which can be the 4-@xmath53 beam pattern and the sky descriptions . an analytic framework for studying the leading order effects of a non - circular beam on the cmb power spectrum estimation
is proposed in @xcite .
the authors of this paper present _ asymfast _ @xcite , a general method to estimate an effective @xmath22 function taking into account the asymmetry of the main beam and the scanning strategy .
_ asymfast _ is based on a decomposition of the main beam pattern into a linear combination of gaussians which permits fast convolution in the spherical harmonic space along the scanning strategy .
the cross - power spectrum @xmath54 can be obtained from the ` pseudo ' cross - power spectrum @xmath55 by resolving eq .
[ pseudo_cross ] which leads to invert the coupling kernel matrix @xmath56 . in general for complex sky coverage and weighting schemes ,
this matrix is singular and can not be inverted directly . to avoid this problem , @xcite proposed the binning of the coupling kernel matrix which reduces considerably the complex correlation pattern in the @xmath0s introduced by the applied mask .
the binning is obtained by applying the operators @xmath57 and @xmath58 as follows @xmath59 the solution of eq .
[ pseudo ] in the new base reads @xmath60 with @xmath61 as the correlation between multipoles @xmath46 depends very much on the instrumental setup , the sky coverage and the weighting scheme , the binning has to be defined for each experiment .
it is a compromise between a good multipole sampling and low correlations between adjacent bins .
hereafter to avoid confusion and makes the notation simpler all equations will be written in @xmath46 instead of @xmath62 .
from @xmath63 input maps we can obtain @xmath64 cross - power spectra @xmath65 ( @xmath66 ) which are unbiased estimates of the angular power spectrum but which are obviously not independent . in this section
, we describe the estimation of the cross - correlation matrix between cross - spectra and between multipoles .
we show how the error bars and the covariance matrix in multipole space can be deduced for each cross - power spectra . given a sky map from a detector @xmath27 ,
it can be combined to each of the other detector maps to form @xmath68 cross - power spectra which will therefore be highly correlated .
furthermore , due to the masking we also expect that each cross - power spectra will be correlated for adjacent multipoles and thus , correlations between adjacent multipoles will be also present between different cross - power spectra . to describe this complexity we define the cross - correlation matrix @xmath69 of the cross - power spectra @xmath70 ( @xmath36 ) and @xmath71 ( @xmath72 ) which can be fully computed as shown in the following . from eq .
[ pseudo ] we can express the ` pseudo ' cross - power spectrum between detectors @xmath27 and @xmath28 as follows @xmath73 where @xmath74 and @xmath75 and therefore the corrected cross - power spectrum for detectors @xmath27 and @xmath28 is given by @xmath76 using the above expression and following the previous definition , the cross - correlation matrix @xmath77 reads @xmath78 the above equation as it stands can not be used in practice for @xmath46 above @xmath79 as the calculation of @xmath80 described in appendix [ appendix_xi ] , is numerically unstable @xcite .
however , for high multipoles and sufficiently large sky coverage ( as it is the case for satellite missions as wmap and planck ) it can be simplified @xcite by replacing @xmath81 and @xmath82 by @xmath83 and @xmath84 respectively and then applying the completeness relation for spherical harmonics @xcite .
this is because the @xmath85 , which is diagonal for full sky coverage , is quasi - diagonal for large sky coverage .
+ under the above hypothesis and following appendix [ appendix_xi ] , the cross - correlation matrix reads @xmath86 ( { \cal m}_{\ell ' \ell_2}^{cd\,-1})^{t } \label{xi}\end{aligned}\ ] ] where @xmath87 and @xmath88 is the quadratic coupling kernel matrix @xmath89 associated to the cross - power spectrum of the product of the masks @xmath90 .
the @xmath91 represents the spherical harmonic coefficients for the product of the masks associated to detectors @xmath27 and @xmath28 .
+ equation [ xi ] can be further simplified by assuming uniform weighting and the same sky coverage for all detectors , as well as a diagonal dominated coupling kernel matrix , @xmath92 \label{xi_approx}\ ] ] in this case , the effect of a non - homogeneous sky coverage is represented by a simple function @xmath93 which can be associated to the effective number of degrees of freedom in the @xmath94 distribution of the @xmath0s over the sky ( see @xcite ) @xmath95 where @xmath96 is the @xmath97-th moment of the mask @xmath98 in eq .
[ xi ] and [ xi_approx ] , the @xmath0 can be either cross - power spectra or auto - power spectra depending on the combination of @xmath27 , @xmath28 , @xmath99 and @xmath100 ( the only condition is @xmath66 and @xmath101 ) .
in one hand , noise terms , as included in the auto - power spectra , appear in the analytical form of the correlation matrix @xmath67 . on the other hand , for a set of 4 independant detectors ( @xmath102 ) ,
the correlation matrix is the variance of the signal .
it is important to notice that the cross - correlation matrix contains all the needed information to perform error bars and covariance matrix for each single cross - power spectrum . from eq .
[ xi ] ( or eq . [ xi_approx ] for the approximated form ) , the covariance matrix providing the correlation between adjacent multipoles is given by @xmath103 \label{covariance}\end{aligned}\ ] ] in the same way , we can also write the variance of each cross - power spectra which corresponds to the error bars @xmath104 associated to each cross - power spectra @xmath105 \label{error_bar}\end{aligned}\ ] ] in these cases , instrumental noise appears in the auto - power spectra @xmath106 ( for which @xmath107 ) . as it is computed directly from the data , no estimation of the noise is needed .
this section describes the estimation of the final cmb angular power spectrum and the error bars associated to it using the corrected cross - power spectra described in sect .
[ cross ] in addition to the cross - correlation matrix computed in sect .
[ correlation ] .
we propose a simple but efficient way to combine them by maximizing a quadratic likelihood function and to deduce the final error bars from the cross - correlation matrix . for this paper
we have considered a very general form for the likelihood function limited to a simple cmb plus noise model .
however , the likelihood function can be adapted more specifically to the data , for example to include point sources as it is the case for the wmap results presented in @xcite .
more generally , it could be extended to algorithms of spectral matching decomposition ( _ e.g. _ @xcite ) . in the following ,
we present how this method can be used to provide both an estimate of the @xmath0s from a full data set ( sect .
[ lincomb ] ) and of the angular power spectrum of common sky fluctuations between maps coming from two or more independent surveys ( sect .
[ common ] ) .
once all possible cross - power spectra have been computed from a data set , we dispose of @xmath64 different but not independent measurements of the angular power spectrum , @xmath108s . to combine them and obtain the best estimate of the power spectrum @xmath109
, we maximize the gaussian approximated likelihood function @xmath110 \label{likelihood}\ ] ] where @xmath111 is the cross - correlation matrix of the cross - power spectra described before ( @xmath97 and @xmath112 @xmath113 ) . the auto - power spectra are not considered . from this and neglecting the correlation between adjacent multipoles ,
it is straightforward to show that the estimate of the angular power spectrum is @xmath114 } { \sum_{ij } |\xi^{-1}|_{\ell\ell}^{ij } } \label{bestcell}\ ] ] the final covariance matrix can be obtained from eq . [ likelihood ] , @xmath115 and the final error bars are given by @xmath116 depending on the cross - correlation between cross - power spectra , the instrumental noise variance is reduced by a factor comprised between n , the number of independant detectors , and @xmath64 , the number of cross - power spectra . for the noise dominated case ,
the correlation between different cross - power spectra can be neglected and the cross - correlation matrix @xmath67 becomes diagonal ( see eq . [ xi ] and [ xi_approx ] ) .
the values in the diagonal are the @xmath64 products of 2 auto - power spectra ( including noise ) .
the variance of final power spectrum is then proportional to @xmath117 . in any case , when combining cross - power spectra , the upper limit for the variance comes from the combination of @xmath63 independent detectors ( proportional to @xmath118 ) .
the xspect formalism allows us to compare two or more different sets of sky maps coming from two or more independent experiments . in this respect ,
the quantity of interest is the power spectrum of the common fluctuations on the sky with the same physical origin .
we will call the latter common angular power spectrum @xmath119 for simplicity .
in fact , if we compare for example two sets of cmb maps at low and high frequencies , the foreground contamination will be different and we can expect to obtain a better estimation of the cmb power spectrum . in the same way , for two different experiments ,
systematic effects will be decorrelated and will not contribute to the common angular power spectrum calculated from them .
in addition , template maps of foreground emission can be correlated to the cmb experiments maps to monitor and to subtract foreground residuals .
+ to simplify the notation we will only consider two sets of sky maps @xmath27 and @xmath28 , corresponding to @xmath120 and @xmath121 detectors respectively . following the previous paragraph considerations , by cross - correlating each detector map from @xmath27 which each detector map from @xmath28 and correcting the ` pseudo ' cross - power spectra as described before
, we can form @xmath122 cross - power spectra which are unbiased estimates of the common power spectrum but not fully independent .
the cross - power spectra formed in this way can be noted @xmath123 and their cross - correlation matrix @xmath124 reads @xmath125 which can be computed using previous approximations .
+ finally , equations [ bestcell ] and [ errorbarbestcell ] are used to obtain the estimate of the common angular power spectrum @xmath126 and the error bars associated to it .
this method allows us to deal naturally with very different experimental configurations such as for example completely different scanning strategies , different timeline filtering and different beam patterns for each of the detectors involved .
furthermore , it can be easily generalized to work with non cmb data such as maps of the sunyaev - zeldovich effect in clusters of galaxies or maps of mass fluctuations from weak lensing observations .
xspect was mainly developed to measure the cmb temperature angular power spectrum from the archeops data as well as to compare these data to other observations of the sky as for example those from the wmap satellite .
archeops @xcite is a balloon - borne experiment conceived as a precursor of the planck high frequency instrument .
it consists of a 1.5 m off - axis gregorian telescope and of an array of 21 photometers cooled down to @xmath127 mk and which operates in 4 frequency bands : 143 and 217 ghz ( cmb channels ) and 353 and 545 ghz ( galactic channels ) .
for the latest flight campaign , the entire archeops data cover about @xmath128% of the sky , including the galactic plane . as a first step in the application of xspect to the archeops data set , we have produced 500 simulations of the sky observed by archeops including both the cmb signal and the detector noise for the six most sensitive archeops detectors at 143 and 217 ghz .
these simulations are computed from realizations of the cmb sky at @xmath129 ( corresponding to a pixel size of @xmath130 arcmin ) for the archeops best - fit cmb model presented in @xcite .
the sky maps are convolved by the main beam pattern of each of the archeops detectors using the beam transfer function and then deprojected following the archeops scanning strategy to produce mock archeops timelines .
the beam transfer functions were computed individually for each detector from jupiter s crossings in the data using the _ asymfast _ method presented in @xcite .
simulations of archeops observations of the cmb sky ( see text for details ) .
we use the archeops best - fit model @xcite to produce cmb maps which are smeared out using the beam pattern of each of the six most sensitive archeops detectors and deprojected into mock archeops timelines .
noise is then added up for each detector and the resulting timelines projected into sky maps are analyzed using xspect .
the error bars shown here are computed analytically as described in previous sections .
the xspect estimate of the angular power spectrum ( in red ) is an unbiased estimate of the input cmb model ( in yellow ) . ]
the detector noise is computed from the time power spectrum of each of the detectors and added to the mock signal timelines .
the method used for the estimation of the time noise power spectra is described in details in @xcite . to suppress @xmath131-noise , remaining atmospheric and galactic contamination as well as non - stationary high frequency noise ,
the archeops time streams are fourier filtered using a bandpass filter between 0.1 and 38 hz .
we apply the same filtering to the simulated timelines before projection .
the corresponding multipole filtering function @xmath49 was independently calculated by specific simulations .
the filtered timelines are then reprojected using the archeops scanning strategy to produce @xmath132 simulated co - added maps for each simulation run .
a galactic mask deduced from a sfd iras map @xcite , extrapolated to 353 ghz , using a cut in amplitude ( greater than 0.5 mjy.str ) is applied to the simulated maps .
this reduces the total archeops coverage to @xmath133 of the sky .
we have used two different weighting schemes on the maps .
the first one is an uniform weighting and the second one , a @xmath134 weighting per pixel and per detector , where @xmath135 is the noise variance for the given pixel .
then we apply the xspect method to each simulations - run .
the @xmath136 cross - power spectra are computed with their associated cross - correlation matrices and combined as described in sect .
[ lincomb ] using the gaussian approximation of the likelihood function .
in addition to the above simulations , we produced , in the same way , 500 non noisy simulations from which we can extract for example the sample variance for the archeops coverage .
figure [ cl_simuarcheops ] shows the mean angular power spectrum computed from the angular power spectrum estimates obtained for each of the 500-simulations runs .
the error bars are analytically computed .
we observe that the gaussian approximated combination of cross - power spectra is an unbiased estimate of the input cmb model for the angular power spectrum .
this mean angular power spectrum is a combination of two weighting schemes .
up to @xmath137 , we use uniform weighting whereas , for larger @xmath46 , a mask inversely proportional to the noise in each of the pixels has been applied .
this allows us to choose the smaller error bars in each region .
error bars for the archeops data computed analytically , from eq .
[ xi_approx ] ( red solid line ) and from eq .
[ xi ] ( yellow solid line ) .
they are compared to the standard deviation of @xmath138 simulations ( black dashed line ) .
differences essentially originate from the extremely inhomogeneous archeops sky coverage ( see text for details ) . ]
figure [ error_simuarcheops ] shows analytic estimates of the error bars in the final angular power spectrum compared to the dispersion for each multipole bin over the 500 monte - carlo simulations .
the archeops sky coverage is very inhomogeneous due to the particular choice of the scanning strategy which tries to maximize the area of the sky observed ( @xmath128% ) in a very reduced amount of observation time ( less than 12 hours ) .
archeops performs large circles on the sky that leads to a quasi ring - like coverage with a large uncovered region at the center of the ring as shown on fig .
[ archeops_cover ] .
this implies that the approximation used to obtain eq .
[ xi_approx ] is not valid , especially at low multipoles , and therefore the error bars obtained analytically ( red solid line ) are underestimated up to @xmath133 at very low @xmath46 with a mean of @xmath139 . by computing @xmath140 as given by eq .
[ xi ] , the analytic error bars ( yellow solid line ) increase as expected and fit much better the dispersion in the simulations ( black dashed line ) .
the agreement is then within @xmath141 over the full range of multipoles with a mean of @xmath142 .
equivalent results are obtained using non noisy simulations . sky coverage and
weighting scheme applied to the sky map ( @xmath143 ) obtained from the archeops most sensitive detector at 143 ghz .
we observe a ring - like structure with a large uncovered area in the center and highly redundant areas on the edges of the ring . ]
xspect has been applied to the data from the last archeops flight campaign using the six most sensitive bolometers as in the simulations presented above .
the results of this analysis is presented in a collaboration paper @xcite .
cross - correlation with wmap maps is also under study in order to assess the electromagnetic spectrum of the cmb anisotropies from 40 to 217 ghz .
in this paper , we have presented a method , xspect , for the obtention of the cmb angular power spectrum with analytical error bars based on four main steps : 1 .
given @xmath63 independent input sky maps from different detectors either from a single experiment or from multiple ones , we estimate @xmath64 ` pseudo ' cross - power spectra as well as @xmath63 ` pseudo ' auto - power spectra using the healpix package @xcite .
cross - power spectra are obtained by correcting the ` pseudo ' power spectra for weighting scheme , beam smoothing and filtering as discussed in section [ cross ] . in the same way
, we compute auto - power spectra with noise .
xspect can deal with different and complex weighting schemes for each of the detectors involved .
the coupling matrix , beam and filtering transfer functions are precomputed for each pair of detectors .
we compute the full cross - correlation matrix , @xmath144 , between cross - power spectra @xmath70 ( @xmath36 ) and @xmath71 ( @xmath145 ) ( section [ correlation_matrix ] ) from which we can extract the covariance matrix and the error bars for each cross - power spectra ( sect [ covpercross ] ) .
4 . finally
, the corrected cross - power spectra are combined into a single angular power spectrum using their cross - correlation matrices which are assumed to be diagonal in multipole space .
the gaussian approximation of the likelihood function is fully justified in the large multipole range . in addition ,
analytical estimates for the error bars and for the covariance matrix are computed ( section [ combination ] ) .
the covariance matrix can be used to check the degree of correlation between multipoles . as
this method estimates the angular power spectrum using ` pseudo ' cross - power spectra , we can obtain a non noise - biased power spectrum avoiding the estimation of the noise power spectra which requires heavy monte - carlo simulations .
equally , this permits the estimation of analytical error bars which are compatible to those computed from simulations .
furthermore , xspect allows us to obtain the angular power spectrum of common sky fluctuations between two or more experiments .
associated to other surveys used as templates , it can provide estimations of systematic residuals or astrophysical contaminations .
nevertheless , xspect computes only the common structures between maps , assuming a single physical component .
this means that , for cmb purposes , foregrounds and systematics must be subtracted beforehand .
xspect has been successfully applied to simulations of the archeops experiment which presents an inhomogeneous sky coverage and detectors with unequal sensitivities . even for a balloon - borne experiment that can be compared to satellite missions neither in noise level nor in sky coverage , the analytical estimates of the errors computed with xspect are accurate within few percents .
this property is of great interest when producing monte carlo simulations in the process of testing and improving the estimation of the @xmath0s or when checking the robustness of the data analysis with respect to the various choices of mask , filtering or binning .
the application of xspect to the archeops data set will be published by the archeops collaboration shortly @xcite . for the wmap satellite , which covers the full sky in a very homogeneous way , it was also interesting to check the analytical estimates of the error bars .
xspect analytical errors are equivalent to those provided by the wmap team within 15% with a mean of 1.8% .
this agreement is satisfactory as our analysis was more basic than the one used to derive the first year wmap results presented in @xcite .
in particular all maps at all multipoles were used without any point - source specific treatment and only two weighting schemes were applied .
the extension of xspect to cmb polarization maps is under development . as for the temperature power spectrum ,
the @xmath146 and @xmath147 power spectra obtained directly from a single set of i , q and u maps via a pseudo power spectrum estimator are noise biased .
a pseudo cross - power spectrum estimator adapted to polarization can solve this problem by using independent sets of i , q and u maps .
authors are grateful to j - ch . hamilton for discussions in the initial stage of this work in our team .
we also want to thank f.x dsert for useful discussions on the method and for corrections of this paper .
the healpix package @xcite was used extensively .
amblard a. , hamilton j. c. , 2004 , , * 417 * , 1189 ansari r. _ et al .
_ , 2003 , , * 343 * , 552 bennett c. l. _ et al . _ , 2003 , , * 148 * , 97 benot a. _ et al . _ , 2003 , , * 399 * , l19 benot a. _ et al . _ , 2003 , , * 399 * , l25 bond j.r . ,
jaffe a.h . , knox l. , 1998 , , * 57 * , 2117 .
borrill j. , 1999a , , * 59 * , 27302 .
borrill j. , 1999b , in ec - tmr conf .
476 , 3k cosmology , eds .
l. maiani , f. melchiorri and n. vittorio , woodbury aip , 277 .
challinor a.d .
_ et al . _ , 2002 , , * 331 * , 994
. chon g. _ et al .
_ , 2004 , , , * 350 * , 914 delabrouille j. , cardoso j .- f . , patanchon g. , 2003 , , * 346 * , 1089 doran m. , 2003 , arxiv : astro - ph/0302138 . douspis m. , bartlett j.g . , blanchard a. and le dour m. , 2001 , , * 368 * , 1 efstathiou g. , 2004 , , * 349 * , 603 fosalba p.
_ et al . _ , 2002 , , * 65 * , 63003 gorski k.m .
, hivon e. & wandelt b. d. , 1998 , proceed . of the mpa / eso conf . on evolution of large - scale structure : from recombination to garching , 2 - 7 august 1998 , eds .
banday , r.k .
sheth and l. da costa , astro - ph/9812350 , + http://www.eso.org/science/healpix grainge k. _ et al .
_ , 2003 , , * 341 * , l23 halverson n.w . _ et al .
_ , 2002 , , * 568 * , 38 hansen f. , grski k. m. , hivon e. , 2002 , , * 336 * , 1304 hansen f. , grski k. m. , 2003 , , * 343 * , 559 hinshaw g. _ et al . _ , 2003 , , * 148 * , 135 hivon e. _ et al . _ , 2002 , , * 567 * , 2 hu w. , spergel d.n . , white m. , 1997 , , * 55 * , 3288 jaffe a.h
_ et al . _ , 2003 , new astronomy review ,
* 47 * , 727 kuo c.l .
_ et al . _ , 2004 , ,
* 600 * , 32 lewis a. , bridle s. , 2002 , , * 66 * , 103511 lewis a. , challinor a. , lasenby a. , 2000 , , * 538 * , 473 liddle a.r . ,
lyth d.h . , 2000 , _
cosmological inflation and large - scale structure _ , cambridge university press , cambridge .
linde a. , sasaki m. , tanaka t. , 1999 , , * 59 * , 123522 mitra s. , sengupta a. s. , souradeep t. , 2004 , , accepted , astro - ph/0405406 .
page l. _ et al .
_ , 2003 , apjs , * 148 * , 39 peebles p.j.e . , 1973 , , * 185 * , 431 peebles p.j.e . ,
hauser m.g . , 1974 , , * 28 * , 19 . ruhl j.e .
_ et al . _ , 2003 , , * 599 * , 786 schlegel d. , finkbeiner d. , davis m. , 1998 , , * 500 * , 525 seljak u. , zaldarriaga m. , 1996 , , * 469 * , 437 sievers j.l .
_ et al . _ , 2003 , , * 591 * , 599 souradeep t. , ratra b. , 2001 , , * 560 * , 28 szapudi i. _ et al .
_ , 2001 , , * 548 * , l115 tegmark m. , 1997 , , * 55 * , 5895 . tegmark m. , de oliveira - costa a. , 2001 , , * 64 * , 63001 . tristram m. , hamilton j - ch . , macas - perz j.f
. , renault c. , 2004 , , 69 , 123008 tristram m. , patanchon g. , macas - perz j.f .
_ et al . _ , 2005 , , submitted , astro - ph/0411633 van leeuwen f. _ et al . _ , 2002 , , * 331 * , 975 .
varshalovich d.a .
, moskalev a.n .
, khersonoskii v.k . , 1988 , _ quantum theory of angular momentum _ , world scientific , singapore .
verde l. _ et al .
_ , 2003 , , * 148 * , 195 wandelt b. d. , hivon e. , gorski k. m. , 2001 , , * 64 * , 083003 wandelt b. d. , gorski k. m. , 2001 , , * 63 * , 123002 wu j.h.p .
_ et al . _ , 2000 ,
apjs , * 132 * , 1 zaldarriaga m. , seljak u. , bertschinger , e. , 1998 , , * 494 * , 491 zaldarriaga m. , seljak u. , 2000 , , * 129 * , 431
in this appendix we describe in details the calculation of the cross - correlation matrix @xmath77 under the hypothesis of large sky coverage which leads to where @xmath153 are the coefficients of the spherical harmonic decomposition of the masked sky map such that @xmath154 with @xmath155 @xmath156 is the area of pixel @xmath97 .
@xmath96 represents the mask applied to the sky map and @xmath157 the coefficients of the spherical harmonic decomposition of the mask @xmath158 where @xmath88 is the quadratic coupling kernel matrix @xmath167 associated to the cross - power spectrum of the products of the masks @xmath168 for each of the cross - power spectra @xmath70 et @xmath71 , @xmath169
|
we present xspect , a method to obtain estimates of the angular power spectrum of the cosmic microwave background ( cmb ) temperature anisotropies including analytical error bars developed for the archeops experiment .
cross - power spectra are computed from a set of maps and each of them is in itself an unbiased estimate of the power spectrum as long as the detector noises are uncorrelated .
then , the cross - power spectra are combined into a final temperature power spectrum with error bars analytically derived from the cross - correlation matrix .
this method presents three main useful properties : ( 1 ) no estimation of the noise power spectrum is needed , ( 2 ) complex weighting schemes including sky covering and map noise properties can be easily taken into account , and corrected for , for each input map , ( 3 ) error bars are quickly computed analytically from the data themselves with no monte - carlo simulations involved .
xspect also permits the study of common fluctuations between maps from different sky surveys such as cmb , sunyaev - zeldovich effect or mass fluctuations from weak lensing observations .
cosmic microwave background
cosmology : observations methods : data analysis
|
99 for reviews , see , e.g. , the special issue on complex systems in 1999 _ science _ * 284 * 79 ; watts d j , _ small worlds _ ( princeton univ . press , princeton , 1999 ) ; newman m e j 2000 _ j. stat . phys . _ * 101 * 819 ; strogatz s h 2001 _ nature ( london ) _ * 410 * 268 ; dorogovtsev s n and mendes j f f , _ adv
. phys . _ * 51 * , 1079 ( 2002 ) ; albert r and barabsi a - l 2002 _ rev .
phys . _ * 74 * 47 .
|
we consider atoms interacting each other through the topological structure of a complex network and investigate lattice vibrations of the system , the quanta of which we call _ netons _ for convenience .
the density of neton levels , obtained numerically , reveals that unlike a local regular lattice , the system develops a gap of a finite width , manifesting extreme rigidity of the network structure at low energies .
two different network models , the small - world network and the scale - free network , are compared : the characteristic structure of the former is described by an additional peak in the level density whereas a power - law tail is observed in the latter , indicating excitability of netons at arbitrarily high energies .
the gap width is also found to vanish in the small - world network when the connection range @xmath0 .
a variety of network systems such as computer networks , neuronal networks , biochemical networks and social networks , possesses complex topological structure , which can be described neither by regular networks nor by completely random networks @xcite . among various models for generating such complex networks , the watts - strogatz ( ws ) model @xcite and the barabsi - albert ( ba ) model @xcite provide the two most representative ones .
both model networks are characterized by very small network diameters proportional to the logarithm of the network size , which is dubbed `` small - world behavior '' and commonly observed in various real networks .
other important properties of many real network systems include high clustering @xcite and scale - free degree distributions @xcite .
while most existing studies of complex networks have focused on the geometrical and topological characterization , there now emerges an increasing number of studies paying attention to the dynamics defined on the networks .
lattice vibrations and associated phonon excitations on local regular networks , e.g. , one - dimensional ( 1d ) chains or two - dimensional square lattices , have been a textbook example of the introductory solid - state physics . in the case of such regular networks , atoms located on vertices interact only with nearest neighboring atoms , resulting in the standard linear phonon dispersion . on the other hand , when the interaction between atoms takes the form of the connection topology of a complex network , ingredients of long - range nature enter into the system .
for example , the ws network has both local edges and long - range shortcuts @xcite , which in this work are interpreted as the coexistence of the local and the long - range interactions .
existing studies on phonon excitations have mostly been performed on the local regular networks such as @xmath1-dimensional hypercubic lattices or on the fractal geometry @xcite .
recently , the spectral properties of lattice vibrations have been studied in a small - world network similar to the ws network @xcite .
however , effects of the local interaction range and the dependence of the spectral properties on the rewiring probability have not been investigated in detail .
the main purpose of this work is to study in detail the vibration spectra of both the ws network and the ba network , in comparison with those of regular lattices .
we thus consider atoms located on vertices of a complex network and interacting with each other connected via edges of the network , and investigate lattice vibrations and phonon excitations , with emphasis on the effects of the network topology .
the results of such study can have implications in the mechanical property of a system of long chains bent in a complicated way .
for example , in a bundle of long flexible polymer chains , some monomers which were separated by a long distance along a chain can make new couplings , building shortcuts . throughout this work
, we refer to phonons in a complex network as _ netons _ , to manifest qualitatively different characteristics due to the fact that the underlying geometric structure is not a usual periodic lattice . by means of numerical diagonalization of the dynamic matrix , neton
excitation spectra are obtained and the corresponding density of neton levels are computed .
it is revealed that the complex network in general has a finite gap in the level density , indicating the absence of low - energy excitations .
the dependence of the gap width on the rewiring probability and the local interaction range is investigated in detail . the existence of the gap is also reflected by the vanishingly small specific heat at low temperatures . in the ws network three singularities
are observed in the neton level density : among them two are shown to be reminiscence of the van hove singularity present in the local regular network , while the third one originates from the complex network structure .
the ba network is also considered and the scale - freeness results in a power - law distribution of the level density , suggesting excitability of netons with arbitrarily high energies .
we build the ws network following ref .
@xcite : first , a 1d regular network with only local connections of range @xmath2 is constructed under the periodic boundary conditions .
next , each local edge ( or link ) is visited once , and with the rewiring probability @xmath3 removed and reconnected to a randomly chosen vertex .
after a whole sweep over the entire network , the average number of shortcuts in the system of size @xmath4 is given by @xmath5 . in the ws network built as above , an atom is put on every vertex whereas an edge connecting two vertices represents the coupling between the two atoms located at the two vertices . for simplicity , we assume that all atoms are identical , each having mass @xmath6 and moving only along the direction of the chain .
the equation of motion for the @xmath7th atom at the position @xmath8 then reads @xmath9 where @xmath10 stands for the set of vertices connected to the @xmath7th vertex ( via either local edges or shortcuts ) and the interaction strength @xmath11 is assumed to be the same for any pairs of the interacting atoms . in the absence of shortcuts , eq .
( [ eq : motion ] ) describes a 1d chain of atoms with only local connections up to the @xmath2th nearest neighbors .
the analytic solution in this limit of @xmath12 is easily found to take the form : @xmath13 , where @xmath14 represents the interparticle spacing , @xmath15 is the eigenfrequency and the periodic boundary conditions , @xmath16 , restrict the wavenumber @xmath17 to take the value @xmath18 with integer @xmath19 . for simplicity , we set @xmath20 , and obtain @xmath21 which leads to the dispersion relation @xmath22 for small wave numbers ( @xmath23 .
equation ( [ eq : rsmall ] ) shows that @xmath24 regardless of the local interaction range @xmath2 .
such linear dependence of the frequency @xmath15 on the wavenumber @xmath17 implies that the group velocity equals the phase velocity , both of which are frequency independent .
henceforth we measure the frequency @xmath15 in units of @xmath25 for convenience . in the presence of shortcuts @xmath26
, the plane wave solution fails , owing to the lack of the translational symmetry . substituting @xmath27 in eq .
( [ eq : motion ] ) , we write the equation of motion in the matrix form @xmath28 where @xmath29 is the @xmath4-dimensional column vector with the components @xmath30 @xmath31 .
the @xmath32 dynamic matrix @xmath33 has the elements @xmath34 where @xmath35 denotes the number of vertices connected directly to the @xmath36th vertex via local edges or shortcuts .
it should be noted that @xmath37 lacks the translational symmetry due to the existence of shortcuts . we have performed numerical diagonalization of the dynamic matrix @xmath33 for various network sizes @xmath38 , 200 , 400 , and 800 , only to find insignificant difference between the two largest sizes .
the eigenfrequencies found numerically for given network realization are sorted in the ascending order and an integer label @xmath19 is attached from the lowest eigenfrequency ( @xmath39 for the smallest @xmath15 , @xmath40 for the next smallest @xmath15 , and so on ) . to obtain better statistics ,
we have also considered 1000 - 10000 different network realizations , over which averages are taken .
figure [ fig : wk ] shows the neton frequency @xmath15 versus the _ quasi - wavenumber _
@xmath41 at various values of the rewiring probability @xmath3 for the local interaction range ( a ) @xmath0 and ( b ) @xmath42 .
note that the quasi - wavenumber @xmath43 becomes identical to the true wavenumber @xmath17 only when @xmath12 .
in general , the wavenumber @xmath17 is not appropriate for the system in the presence of shortcuts ( @xmath44 ) : it does not make a good quantum number for @xmath44 , where the translational symmetry is broken . for the range @xmath45 , fig .
[ fig : wk](a ) suggests that the linear dispersion relation @xmath46 , which has been obtained analytically in eq .
( [ eq : rsmall ] ) for @xmath12 , still appears to hold at any value of @xmath3 .
in contrast , for the local interaction range @xmath42 , the dispersion @xmath47 changes its functional form as @xmath3 is raised from zero , reducing to the linear relation @xmath46 only in the limit of @xmath48 [ see fig .
[ fig : wk](b ) ] .
in particular , it is evident in fig .
[ fig : wk](b ) that for @xmath49 there appears a nonzero minimum value of @xmath15 , below which the eigenfrequency does not exist .
we have also examined the cases with the range larger than two ( @xmath50 ) and confirmed that these features remain qualitatively the same . with regard to this , it is of much interest to note that the ws network has a finite clustering coefficient @xcite , only when @xmath51 @xcite , which makes it plausible that the difference in the neton frequency spectrum between the cases @xmath45 and @xmath42 , shown in fig .
[ fig : wk](a ) and ( b ) , respectively , is a manifestation of the clustering effects .
we have also considered numerically the decomposition property of the network and found that while the ws network with @xmath51 does not decompose into distinct parts , for @xmath45 the probability of decomposition is finite . however , it should be noted that those networks with @xmath45 which are not decomposed still display the vanishing gap .
consequently , the vanishing gap in fig .
[ fig : wk](a ) for @xmath45 has its origin not in the decomposition property but presumably in the clustering property of the network .
similar ws networks can also be built in a slightly different manner : instead of rewiring local connections with the total number of edges fixed , one may _
add _ shortcuts to the network without removing local connections . in this case , we have confirmed numerically that there exists a gap even for @xmath45 , which keeps parallel to the observation that the @xmath52 model on the ws network with @xmath45 , constructed in such a way , has a finite - temperature transition , similar to that in the original ws network with @xmath53 @xcite .
we next consider the neton - level density @xmath54 , which is given by @xcite @xmath55 in one dimension .
using the mathematical identity @xmath56 with @xmath57 , we write eq .
( [ eq : density1d ] ) in the form @xmath58 in the regular network ( @xmath12 ) , eq .
( [ eq : gw ] ) , together with eq .
( [ eq : rsmall ] ) , gives the density @xmath54 independent of @xmath15 , i.e. , @xmath59 , in the low - frequency limit @xmath60 .
-0.8 cm in the presence of shortcuts , we use the box counting method to compute the neton - level density from the obtained eigenvalues of the dynamic matrix . the resulting density @xmath54
is plotted in fig .
[ fig : g(w ) ] , which exhibits several characteristic features . first , as already found in fig .
[ fig : wk](b ) , the density @xmath54 develops a gap @xmath61 below which @xmath62 .
figure [ fig : wc ] shows that the gap @xmath61 , well described by the linear relation @xmath63 for small @xmath3 ( see the inset of fig . [ fig : wc ] ) , keeps increasing as the rewiring probability @xmath3 is raised and then appears to saturate for @xmath64 .
such saturation behavior has also been observed in the synchronization phenomena in the ws network @xcite .
the emergence of the gap implies that the ws network is very rigid against lattice vibrations at small energies . in other words
, one can not excite netons at sufficiently low energies .
we also compute the specific heat of the ws network for @xmath65 , and compare it with that of the local regular network ( @xmath12 ) in fig .
[ fig : wscv ] . as expected , the presence of a finite gap in the neton - level density gives rise to exponentially small specific heat at low temperatures .
this is in contrast to the local regular network , where thermal excitations of phonons are gapless and accordingly , the specific heat is described by the well - known debye form with power - law dependence on the temperature , i.e. , @xmath66 in one dimension . the second interesting feature observed in the neton - level density in fig .
[ fig : g(w ) ] is the emergence of sharp peaks at small @xmath3 . figure [ fig : vanhove](a ) shows in detail the density @xmath54 at @xmath67 , displaying three peaks .
it is well known that a local regular network ( @xmath12 ) ( i.e. , a regular lattice ) has van hove singularities whenever @xmath68 vanishes [ see eq .
( [ eq : gw ] ) ] .
for example , in the local regular network with the second - nearest neighbor interaction @xmath69 , the positions of the van hove singularities are easily obtained : the eigenfrequency @xmath70 given by eq .
( [ eq : w2 ] ) , combined with the condition @xmath71 , leads to two singularities at @xmath72 and @xmath73 ; this allows us to conclude that the first two peaks in fig .
[ fig : vanhove](a ) are remnants of the van hove singularities present in the local regular network . on the other hand , the third peak at @xmath74
can not be explained in terms of the singularity of the local regular network . in order to probe the origins of the peaks ,
we measure how the peak heights change as the rewiring probability @xmath3 is varied and plot the results in fig .
[ fig : vanhove](b ) .
the first peak and the second one at @xmath75 and at @xmath76 , respectively , are observed to have heights which decrease with @xmath3 .
this again confirms that these two peaks originate from the local regular network , thus becoming weaker as the network loses its local structure .
the height of the third peak , in contrast , is found to grow as @xmath3 is increased , which clearly indicates that the third peak provides a feature generic to the ws network structure . finally , we examine neton excitations on the ba network @xcite . in more detail , we first grow the ba network , following ref .
@xcite : starting from @xmath6 vertices initially ( at time @xmath77 ) , we attach a vertex with @xmath6 edges at every time step .
once the ba network is built , we then put an atom on each vertex , assuming that the network is embedded in a 1d chain .
the frequency spectrum is obtained in the same way as that of the ws network , with the dynamic matrix still given by eqs .
( [ eq : matrix ] ) and ( [ eq : dij ] ) .
similarly to the ws network , revealed is the emergence of a gap .
as the number @xmath6 becomes larger , the gap grows and the peaks tend to smear out .
another interesting observation in the ba network is the slow decay of the density @xmath54 at large frequencies , which is to be compared with the vanishingly small density at high frequencies found in the ws network .
the log - log plot of the density @xmath54 versus the frequency @xmath15 , displayed in fig .
[ fig : batail ] , reveals that the ba network exhibits power - law decay which is well scaled as @xmath78 .
in contrast , @xmath54 in the ws network shows very rapid decay , which is presumably of the exponential form .
the observed power - law decay of @xmath54 in the ba network implies that netons may be excited with arbitrarily high energies .
it is of interest to compare this with the behavior of the adjacency matrix of the ba network @xcite . although the adjacency matrix @xmath79 , with @xmath80 and @xmath81 for @xmath82 , and the dynamic matrix @xmath33 have different diagonal components , both show similar scale - free distributions at large eigenvalues . in summary
, we have investigated the phonon spectra in the monatomic chain where atoms interact with each other through the edges of complex networks . the density of phonon levels in the complex network , i.e.
, neton levels has been found to develop a gap , which indicates that netons may not be excited up to the critical frequency , and accordingly , a finite amount of energy is required to generate lattice vibrations .
this robustness of a complex network has also been shown to yield exponentially small specific heat at low temperatures , in contrast to the power - law temperature dependence in the local regular network .
the characteristic structure of the ws network has been described by the emergence of a peak in the level density in addition to those originating from the van hove singularities . in the ba network ,
on the other hand , the level density has been found to decay algebraically at high frequencies , which suggests excitability of arbitrarily high - energy netons .
this work was supported in part by the korea science and engineering foundation through grant no .
r14 - 2002 - 062 - 01000 - 0 ( bjk ) and in part by the ministry of education of korea through the bk21 project ( myc ) .
|
it was already in the 60 s when the pioneering works of higgs , kibble , englert and others suggested the existence of a fundamental spinless building - block of nature , whose non - vanishing vacuum expectation value could explain the spontaneous breaking of the @xmath13 gauge group of the electroweak ( ew ) interactions down to the @xmath14 abelian symmetry group of the electromagnetism . indeed , this so - called higgs mechanism is the most fundamental longstanding issue that remains experimentally unsettled in particle physics .
for one thing it is the only known strategy capable of building up a ( perturbative ) renormalizable quantum field theoretical description of the electroweak symmetry breaking ( ewsb ) phenomenon .
it is difficult to overemphasize that , to a great extent , the higgs mechanism embodies the backbone of the sm structure and that in its absence we would have to cope with an entirely different conception of the inner functioning of the sm as a quantum field theory ( qft ) .
the central assumption here is the existence of ( at least ) one elementary spinless field
the so - called higgs boson .
its relevance is particularly evident if we take into account that the presence of one or more such fields in the structure of the model is indispensable for the unitarity of the theory .
if no higgs bosons exist below the tev scale , weak interactions would actually become strong at that scale and e.g. the @xmath15 cross - section would violate the unitarity of the scattering matrix .
moreover , in the absence of higgs bosons the various particle masses could not be generated consistently ( i.e. without spoiling the ultraviolet behavior of the theory at higher orders of perturbation theory ) . in short ,
if no higgs bosons are there to protect the ( presumed ) perturbative structure of the weak interactions , the latter would enter a perilous runaway regime at high energies .
it is , thus , essential to confirm experimentally the existence of one or more higgs boson particles through their explicit production in the colliders . amazingly enough ,
more than 40 years after the higgs mechanism was naturally blended into the conventionally accepted sm landscape of the strong and electroweak interactions@xcite , it remains still thickly curtained and no direct evidence has been found yet of its main testable prediction : the existence of fundamental spinless particles . notwithstanding
the many efforts devoted at lep and at the tevatron in the last decades , all experimental searches for higgs boson signatures have come up empty - handed , and one has to admit that no conclusive physical signature for the existence of these elementary scalar fields not even the single one predicted by the sm has been found for the time being .
thus , we do not know whether the conventional higgs boson exists at all , or if there are no higgs particles of any sort .
on the contrary , it could be that there are extensions of the sm spanning a richer and elusive spectrum of higgs bosons which have escaped all our dedicated experimental searches .
be that as it may , this pressing and highly intriguing enigma will hopefully be unveiled soon , specially with the advent of the new generation of supercolliders , like the brand new lhc , the future international linear collider ( ilc)@xcite and , hopefully , the compact linear collider ( clic ) too@xcite .
therefore , we have to ( and we can ) be well prepared to recognize all kind of hints from higgs boson physics .
let us recall that apart from the single neutral ( @xmath16-even ) higgs boson @xmath17 of the sm , other opportunities arise in model building that could serve equally well the aforementioned purposes .
perhaps the most paradigmatic extension of the sm is the minimal supersymmetric standard model ( mssm)@xcite , whose higgs sector involves two doublets of complex scalar fields .
the physical spectrum consists of two charged states , @xmath18 , two neutral @xmath16-even states @xmath19 ( with masses conventionally chosen as @xmath20 ) and one @xmath16-odd state @xmath21 .
we shall not dwell here on the why s and wherefores of supersymmetry ( susy ) as a most sought - after realization of physics beyond the sm .
it suffices to say that it provides a higgs sector which is stable ( to all orders in perturbation theory ) under embedding of the low - energy structure into a grand unified theory , and in particular it provides the most natural link with gravity@xcite .
but , how to test the structure of the higgs sector ? in the mssm case , such structure is highly constrained by the underlying supersymmetry .
a consequence of it , which is particularly relevant for our considerations , is the fact that the self - interactions of the susy higgs bosons turn out to be largely immaterial from the phenomenological point of view , in the sense that they can not be enhanced ( at the tree - level ) as compared to the ordinary gauge interactions and , therefore , can not provide outstanding signatures of physics beyond the sm .
the bulk of the enhancing capabilities of the mssm lagrangian is concentrated , instead , in the rich structure of yukawa - like couplings between higgs bosons and quarks or between higgs boson and squarks , or even between squarks and chargino - neutralinos , as it was shown long ago in plentiful phenomenological scenarios involving quantum effects on the gauge boson masses and gauge boson and top quark widths , as well as in many other processes and observables ( see e.g. @xcite . for the present state of the art , cf .
@xcite ; and for recent comprehensive reviews , see @xcite , for example .
the two - higgs @xmath22-doublet structure of the higgs sector in the mssm is a trademark prediction of susy invariance .
nonetheless , the very same doublet structure can appear in the form of a non - supersymmetric framework , the so - called general ( unconstrained ) two - higgs - doublet model(2hdm ) , which does also entail a rich phenomenology .
again , the same spectrum of @xmath16-even ( @xmath19 ) and @xmath16-odd ( @xmath21 ) scalar fields arise .
however , we should keep in mind that the most distinctive aspects of the general 2hdm higgs bosons could well be located in sectors of the model quite different from the mssm case@xcite . indeed , even the very couplings involved in the structure of the higgs potential ( viz .
the so - called trilinear and quartic higgs boson self - interactions ) could be the most significant ones as far as the phenomenological implications are concerned , rather than just the enhanced yukawa couplings with the heavy quarks . in such circumstance
, we should be able to detect a very different kind of experimental signatures .
assuming , for instance , that the lhc will find evidence of a neutral higgs boson , a critical issue will be to discern whether the newcomer is compatible with the sm or any of its extensions and , in the latter case , to which of these extensions it most likely belongs . in order to carry out this `` finer '' hunting of the higgs boson(s )
, a @xmath23-range accessible linear collider machine ( such as the aforesaid ilc / clic ) will be needed . in this work
, we will show how useful it could be such collider to discriminate neutral supersymmetric higgs bosons from generic ( non - susy ) 2hdm ones .
the paper is organized as follows . in the next section ,
we remind the reader of some relevant higgs boson production processes in the linear colliders . in section [ sec:2hdm ]
, we describe the structure of the 2hdm and the relevant phenomenological restrictions , including the properties of the trilinear higgs self - interactions and their interplay with the notion of unitarity and vacuum stability . in section [ sec :
renorm ] , we develop in detail the renormalization program for the higgs sector of the general 2hdm . the theoretical setup for the one - loop computation of the higgs boson cross - sections is elaborated in section [ sec : general ] , leaving section [ sec : results ] to present a comprehensive numerical analysis of our results .
finally , section [ sec : conclusions ] is devoted to a thorough general discussion and to present the conclusions of our work .
the leading higgs boson production mechanisms at the lhc have been studied thoroughly for the last twenty years , they are well under control both in the sm@xcite and in the mssm@xcite and there are also some studies in the 2hdm@xcite .
hopefully , they may soon help revealing some clues on higgs boson physics at the lhc for a review see e.g.@xcite and references therein . however , it is not obvious that it will be possible to easily disentangle the nature of the potentially produced higgs boson(s ) .
the next generation of tev - class linear colliders ( based on both @xmath24 and @xmath25 collisions ) , such as the ilc and the clic projects @xcite , will be of paramount importance in finally settling the experimental basis of the `` higgs issue '' as the most fundamental theoretical construct of the sm of electroweak interactions .
thanks to its extremely clean environment ( in contrast to the large qcd background inherent to a hadronic machine such as the lhc ) , a linear collider ( _ linac _ for short ) should allow for a precise measurement of the higgs boson parameters , such as : i ) the higgs boson mass ( or masses , if more than one ) ; ii ) the couplings of the higgs bosons to quarks , leptons and gauge bosons ; iii ) the higgs boson self - couplings mentioned above , i.e. the trilinear ( 3h ) and quartic ( 4h ) higgs boson self - interactions . at the end of the day
, a linac should allow us to dig deeper than ever into the structure of the ewsb and , hopefully , even to reconstruct the higgs potential itself .
a detailed road map of predictions for higgs - boson observables at the linear colliders is called for , with a special emphasis on those signatures which may be characteristic of the different extensions of the sm .
for example , as indicated above , triple higgs boson self - interactions ( 3h ) may play a cardinal role in this endeavor because , in favorable circumstances , they could easily distinguish between the mssm and the general 2hdm higgs sectors .
such 3h couplings can mediate a plethora of interesting processes .
the key point here is the potentially large enhancements that the 3h couplings may accommodate in the general 2hdm case , in contrast to the supersymmetric extensions .
actually , even for the susy case there is a large number of works attempting to extract vestiges of non - standard dynamics in these couplings , mainly through the radiative corrections that they can undergo .
for instance , the 3h - couplings have been investigated in @xcite , and in some cases considered for possible phenomenological applications in tev - class linear colliders through the double - higgs strahlung process @xmath26 or the double - higgs @xmath27-fusion mechanism @xmath28 .
these processes , which include vertices like @xmath29 , @xmath30 , @xmath31 , @xmath32 and @xmath33 , are possible both in the sm and its extensions , such as the mssm and the general 2hdm .
unfortunately , the cross - section turns out to be rather small both in the sm and in the mssm , being of order @xmath34 pb at most , i.e. equal or less than @xmath35 fb @xcite .
even worse is the situation regarding the triple higgs boson production in the mssm , in which except in the case of some particular resonant configuration the typical cross - sections just border the line of @xmath36fb or less @xcite . in the latter reference , for instance , it has been shown that if the double and triple higgs production cross - sections would yield sufficiently high signal rates , the system of couplings could in principle be solved for all trilinear higgs self - couplings up to discrete ambiguities using only these processes . but this is perhaps a bit too optimistic since in practice these cross - sections are manifestly too small to be all measurable in a comfortable way . in stark contrast to this meager panorama within the sm and the mssm , 3h couplings have been tested for tree - level processes in the context of the unconstrained 2hdm .
for example , in ref.@xcite the tree - level production of triple higgs boson final states was considered in the ilc .
there are three classes of processes of this kind compatible with @xmath37-conservation , namely @xmath38 where , in class 2 ) , we assume that the two neutral higgs bosons @xmath39 must be the same , i.e. the allowed final states are @xmath40 , @xmath41 and @xmath42 . the cross - sections in the 2hdm were shown to reach up to @xmath43@xcite , i.e. several orders of magnitude over the corresponding mssm predictions@xcite .
besides the exclusive production of three higgs bosons in the final state , also the inclusive pairwise production of higgs bosons may be critically sensitive to the 3h self - interactions . sizable production rates , again in the range of @xmath44 have recently been reported in ref.@xcite , whose focus was on the inclusive higgs boson - pair production at order @xmath4 through the mechanism of gauge - boson fusion @xmath45 similar effects have also been recently computed on 2h - strahlung processes of the guise @xmath46 @xcite . in a complementary way
, 3h couplings can also be probed in loop - induced processes .
for instance , ref .
@xcite presents a computation of the single higgs boson production rate through the scattering process of i ) two real photons , using the @xmath25 mode of a linear collider , i.e. @xmath47 ; and ii ) the more traditional mechanism of virtual photon fusion in @xmath24 colliders , @xmath48 . in either case , the obtained cross sections within the general 2hdm ( up to @xmath49 for the first mechanism , and @xmath50 for the latter ) are 1 to 2 orders of magnitude larger than the expected sm yields and also well above the mssm results ( see @xcite and references therein ) .
promising signatures have also been reported within the general 2hdm using loop - induced production of two neutral higgs bosons through real @xmath25 collisions @xcite , although in this case the cross - sections are smaller than in the primary mechanism @xmath47 mentioned before . a crucial observation concerning our aim here is the following : all the processes described above are directly sensitive to the 3h self - couplings already at the leading order . in this paper
, we continue exploiting the properties of these couplings in the general 2hdm , but we concentrate now on their impact at the level of indirect effects through radiative corrections .
the latter may significantly affect processes that are kinematically more favored ( e.g. because they have a smaller number of particles in the final state ) , but which are nevertheless totally insensitive to the trilinear higgs boson couplings at the lowest order .
specifically , we wish to concentrate on identifying the largest quantum effects that the 3h couplings can stamp on the cross - sections for two - body neutral higgs boson final states : @xmath51 surprisingly enough , very little attention has been paid to these basic production processes and to the calculation of the corresponding radiative corrections within the 2hdm . in contrast
, a lot of work has been invested on them from the point of view of the mssm@xcite see also @xcite and references therein . to the best of our knowledge , only refs .
@xcite have addressed this topic in the general 2hdm , although they are restricted to the production of charged higgs pairs @xmath52 . in the present paper
, we shall be concerned exclusively on computing the quantum effects involved on the neutral higgs boson channels ( [ 2h ] ) at order @xmath53 ( and eventually also the leading @xmath54 terms ) . as we will see , very large quantum effects ( of order @xmath55 )
may arise on the cross - sections of the two - body higgs boson channels ( [ 2h ] ) as a result of the enhancement capabilities of the triple ( and , to a lesser extent , also the quartic ) higgs boson self - interactions .
these effects are completely unmatched within the mssm and should therefore be highly characteristic of non - supersymmetric higgs boson physics .
the two - higgs - doublet model ( 2hdm ) is defined upon the canonical extension of the sm higgs sector with a second @xmath56 doublet with weak hypercharge @xmath57 , so that it contains @xmath58 complex scalar fields arranged as follows : @xmath59 with the help of these two weak - isospin doublets , we can write down the most general structure of the higgs potential fulfilling the conditions of @xmath16-conservation , gauge invariance and renormalizability . moreover , a discrete @xmath60 symmetry @xmath61 which will be exact up to soft - breaking terms of dimension 2 is usually imposed as a sufficient condition to guarantee the proper suppression of the flavor - changing neutral - current ( fcnc ) effects that would otherwise arise within the quark yukawa sector @xcite symmetry , see e.g. @xcite and references therein . ]
( we shall return to this point later ) ) below . ] .
all in all , one arrives at the following expression for the tree - level potential : @xmath62 ^ 2 \nonumber \\ & & + \lambda_4\,\left[(\phi_1^\dagger \phi_1)(\phi^\dagger_2\phi_2 ) - ( \phi^\dagger_1\phi_2)(\phi^\dagger_2\phi_1)\right ] \nonumber \\ & & + \lambda_5\ , \left[\re\,e\,(\phi_1^\dagger \phi_2 ) -\ , \frac{v_1\,v_2}{2}\right]^2\nonumber\\ & & + \lambda_6\ , \left[\im\,m\,(\phi_1^\dagger \phi_2 ) \right]^2\ , , \label{eq : potential}\end{aligned}\ ] ] where @xmath63 are dimensionless real parameters and @xmath64 stand for the non - vanishing vev s that the neutral component of each doublet acquires , normalized as follows : @xmath65 .
the complex degrees of freedom encoded within each higgs doublet in eq.([eq : h1h2 ] ) are conveniently split into real ones in the following way : @xmath66 upon diagonalization of the higgs potential ( [ eq : potential ] ) we may obtain the physical higgs eigenstates in terms of the gauge ( weak - eigenstate ) basis : @xmath67 the parameter @xmath68 is given by the ratio of the two vev s giving masses to the up- and down - like quarks : @xmath69 incidentally , let us note that this parameter could be ideally measured in @xmath70 colliders@xcite since the charged higgs event rates with mixed decays , @xmath71 , do not involve the mixing angle @xmath72 .
the physical content of the 2hdm embodies one pair of @xmath16-even higgs bosons ( @xmath73 , @xmath74 ) ; a single @xmath16-odd higgs boson ( @xmath21 ) ; and a couple of charged higgs bosons ( @xmath75 ) .
the free parameters of the general 2hdm are usually chosen to be : the masses of the physical higgs particles ; the aforementioned parameter @xmath68 ; the mixing angle @xmath72 between the two @xmath16-even states ; and finally the coupling @xmath6 , which can not be absorbed in the previous quantities and becomes tied to the structure of the higgs self - couplings .
to summarize , the vector of free inputs reads @xmath76 therefore , we are left with @xmath77 free parameters , which indeed correspond to the original @xmath78 couplings @xmath79 and the two vev s @xmath80 the latter being submitted to the constraint @xmath81 , which is valid order by order in perturbation theory , where @xmath82 is the @xmath83 mass and @xmath84 is the @xmath22 gauge coupling constant . the dimension 2 term that softly breaks the @xmath60 symmetry can be written as @xmath85 , with @xmath86 the second equality being valid at the tree - level . given @xmath68 one may trade the parameter @xmath6 for @xmath87 through this formula , if desired .
since we are going to adopt the on - shell renormalization scheme@xcite for the one - loop calculation of the cross - sections ( cf .
section [ sec : renorm ] ) , it is convenient to introduce the electromagnetic fine structure constant @xmath88 , which is one of the fundamental inputs in this scheme ( together with the physical @xmath89 mass , @xmath90 , and the higgs and fermion particle masses ) .
the electron charge @xmath91 is connected to the original @xmath22 gauge coupling and weak mixing angle @xmath92 through the well - known relation @xmath93 , which is also preserved order by order in perturbation theory .
then the @xmath83 mass can be related to @xmath94 and the fermi constant @xmath95 in the standard way in the on - shell scheme , namely @xmath96where the parameter @xmath97@xcite vanishes at the tree - level , but is affected by the radiative corrections to @xmath98-decay both from standard as well as from new physics , for instance from the mssm@xcite and also from the 2hdm ( see below ) .
notice that , in the above formula , @xmath82 also enters implicitly through the relation @xmath99 , which is valid order by order in the on - shell scheme .
since @xmath95 can be accurately determined from the @xmath98-decay , and @xmath90 has been measured with high precision at lep , it is natural to take them both as experimental inputs .
then , with the help of the non - trivial relation ( [ deltar ] ) , the @xmath83 mass can be accurately predicted in a modified on - shell scheme where @xmath95 , rather than @xmath82 , enters as a physical input .
we shall not dwell here on the detailed structure of the yukawa couplings of the higgs boson to fermions ( see for an exhaustive treatment ) , since their contribution to the processes @xmath100 under study at the one - loop level is largely subdominant in the main regions of parameter space .
however , it is precisely the coupling pattern to fermions that motivates the distinction between the different types of 2hdm s , so let us briefly describe them .
assuming natural flavor conservation@xcite , we must impose that at most one higgs doublet can couple to any particular fermion type . as a result , the coupling of the 2hdm higgs bosons with fermions ( say the quarks ) can be implemented in essentially two manners which insure the absence of potentially dangerous ( tree - level ) fcnc s , to wit ( cf .
table [ typeiandii])-violation there are four manners to have natural flavor conservation , giving rise to four different kinds of allowed 2hdm s that do not lead to tree - level fcnc s , see @xcite , although we will restrict ourselves here to the two canonical ones@xcite . ] : i ) type - i 2hdm , wherein the higgs doublet ( @xmath101 ) couples to all of quarks , whereas the other one ( @xmath102 ) does not couple to them at all ; or ii ) type - ii 2hdm , in which the doublet @xmath102(resp .
@xmath101 ) couples only to down - like ( resp .
up - like ) right - handed quarks . in the latter case , an additional discrete symmetry involving the chiral components of the fermion sector must be imposed in order to banish the tree - level fcnc processes , e.g. @xmath103 for the down and up right - handed quarks in the three flavor families @xmath104 . in turn , the higgs self - couplings @xmath79 in the higgs potential can be rewritten in terms of @xmath68 and the physical parameters of the on - shell scheme such as the masses and the electromagnetic fine structure constant @xmath94 . at the tree - level , we have @xmath105\nonumber\ , , \\
\lambda_2 & = & \frac{\lambda_5\,(1 - 1/\tan^2\beta)}{4}\ , + \frac{\alpha_{em}\,\pi\,}{2\,m^2_w\,{\ensuremath{s_w^2}}\,\sin^2\beta}\nonumber \\ & & \times\left[m_{h^0}^2\,\cos^2\alpha + m_{h^0}^2\,\sin^2\alpha\nonumber \right . \\ & & \left .
-\frac12(m^2_{h^0}-m^2_{h^0})\sin 2\alpha\,\tan\beta\right]\nonumber \ , , \\ \nonumber \\ \lambda_3 & = & -\,\frac{\lambda_5}{4}\,+\frac{\alpha_{em}\,\pi\,\sin 2\alpha } { 2\,m^2_w\,{\ensuremath{s_w^2}}\,\sin 2\beta}\,(m^2_{h^0 } - m^2_{h^0})\nonumber\ , , \,\\
\lambda_4 & = & \frac{2\,\alpha_{em}\,\pi}{m^2_w\,{\ensuremath{s_w^2}}}\,m^2_{h^\pm}\nonumber\ , , \\ \nonumber \\ \lambda_6 & = & \frac{2\,\alpha_{em}\,\pi}{m^2_w\,{\ensuremath{s_w^2}}}\,m^2_{a^0}\ , , \label{eq : param2hdm}\end{aligned}\ ] ] where @xmath106 , @xmath107 . from these lagrangian couplings in the higgs potential
, we can derive the `` physical couplings '' , namely those affecting the physical higgs bosons in the mass - eigenstate basis .
we call these couplings the triple ( 3h ) and quartic ( 4h ) higgs couplings .
their behavior and enhancement capabilities are at the very core of our discussion .
indeed , from our analysis it will become clear that they furnish the dominant source of quantum corrections to the higgs - pair production processes ( [ 2h ] ) within the general 2hdm .
the physical 3h and 4h couplings are not explicitly present in the 2hdm potential ( [ eq : potential ] ) .
they are derived from it after spontaneous breaking of the ew symmetry and corresponding diagonalization of the higgs boson mass matrix using the rotation angles @xmath72 and @xmath108 in ( [ eq : tl ] ) . in the particular case of the sm , the trilinear and quartic higgs couplings have fixed values , which depend uniquely on the actual mass of the higgs boson . in the mssm , however , and due to the susy invariance , the higgs boson self - couplings are of pure gauge nature , as we shall revise briefly below .
this is in fact the primary reason for the tiny production rates obtained for the triple higgs boson processes ( [ 3h ] ) within the framework of the mssm@xcite .
in contrast , the general 2hdm accommodates trilinear higgs couplings with great potential enhancement .
the full list is displayed in table [ tab : trilinear ] .
as can be seen , the couplings in this table depend on the @xmath77 free parameters ( [ freep ] ) . in the particular case where @xmath109
, this table reduces to table 1 of ref .
the equality of these couplings takes place automatically e.g. in the mssm , where in addition other simplifications occur , as we discuss below .
let us note , for example , that in the limit @xmath110 the @xmath111-trilinear coupling in table [ tab : trilinear ] reduces exactly to the sm form @xmath112 where we denote by @xmath113 the sm higgs boson .
this situation would correspond to the so - called decoupling limit in the mssm@xcite since it is correlated with @xmath114 , although there is no such correlation in the general 2hdm .
[ typeiandii ] [ cols="^,^ " , ] some more comments are in order concerning important features contained in figs .
[ fig : overs_s2][fig : overs_s4 ] . as we move from fig .
[ fig : overs_s2 ] to [ fig : overs_s3 ] , hence as we transit from a relatively light to a heavier higgs boson spectrum ( cf . table [ tab : masses ] ) , we encounter that the resulting cross sections , both at the leading and at one - loop level , are pulled down by roughly a factor of 2 .
unsurprisingly , this is signaling the exhaustion of the available phase space . by the same token ,
the peak - shaped maximum of the @xmath115 curve gets shifted forward by about @xmath116 , in correspondence to the production threshold of the higgs pair .
moreover , the heavier the higgs bosons , the more suppressed the higgs - mediated loops become , and hence the lower is the influence of 3h self - couplings .
this is precisely what we confirm numerically in fig .
[ fig : overs_s3 ] ( lower panels ) , in which we can appreciate how @xmath117 reaches @xmath118 at most in contrast to @xmath119 that was spotted for set ii ( cf . fig .
[ fig : overs_s2 ] ) . in turn ,
set vi ( in fig .
[ fig : overs_s4 ] ) embodies again a rather light higgs boson spectrum which resembles that of set ii , even though the present choice of masses is obtained upon artificially engineering a susy setup . here ,
owing to the fact that @xmath120 is heavier than in set ii , unitarity conditions settle a more stringent upper bound on @xmath121 , and hence on the maximum enhancement of the triple neutral higgs couplings . as a result ,
radiative corrections in this case are still sizable , but retreat to the @xmath118 level
. of great relevance is the study of the behavior of @xmath122 as a function of the higgs boson masses .
we present these results conveniently in figs .
[ fig : overmh0_s2][fig : overmh0_s4 ] , where again we superimpose both the tree - level and the loop - corrected cross - sections for the process @xmath123 . the dependence with @xmath124
is indicated on the left panels and that with @xmath125 on the right panels .
the numerical results are shown for the sets ii , iii and iv and have been obtained , once more , by setting @xmath126 , @xmath127 , and making allowance for @xmath121 to take on the highest permissible value in each set ( see table [ tab : maxl5 ] ) .
the center - of - mass energy is settled at the fiducial value @xmath128 .
the plots under consideration evince that the raise of the higgs boson masses exerts a twofold effect on @xmath129 , one kinematical and the other dynamical .
the former simply means that owing to the reduction of the available phase space the cross - section obviously falls down .
we can check it in the figures , where we see that both the tree - level @xmath130 and the loop - corrected @xmath131 indeed decrease monotonically with growing @xmath124 and @xmath125 .
as far as the quantum corrections themselves are concerned , the situation is a bit more subtle .
although heavier higgs bosons imply larger propagator suppressions of the higgs - mediated one - loop diagrams , as we have just described in the previous paragraph , the 3h self - couplings can partially offset this situation since they get invigorated in such circumstances .
this counterbalance feature is manifest in the explicit expressions for the 3h couplings in table [ tab : trilinear ] the uttermost limitation to their growing being the unitarity constraints . in particular , 3h couplings involving neutral fields ( which can be shown to carry the dominant effect on @xmath123 ) turn out to grow either with @xmath124 or @xmath125 .
although both effects partially balance each other , we are ultimately left with the expected decoupling behavior in the limit of large higgs boson masses , with the proviso that such decoupling takes place in a softer way than that of the born - level cross section .
in fact , in figs [ fig : overmh0_s2][fig : overmh0_s4 ] we can verify that the one - loop corrected cross - section remains most of the time on top of the tree - level result when we increase the higgs boson masses , and moreover the slope of the former is clearly milder , particularly for @xmath132 .
this translates into a growing trend of @xmath117 up to values of @xmath133 .
let us also mention in passing that the sudden spike in the @xmath134 curve at @xmath135 ( which is also barely visible for @xmath129 ) corresponds to the @xmath136 pseudo - threshold in the one - loop vertex amplitude ( cf .
the first row of diagrams in fig .
[ fig : vert11 ] ) .
analogous trends are encountered for sets iii and iv of higgs boson masses , the only reportable difference being that the decoupling behavior of @xmath131 with @xmath137 is more pronounced , signaling that the increase of these masses does not boost the 3h self - couplings as efficiently as in the case of set ii .
the shaded regions in fig .
[ fig : overmh0_s3]-[fig : overmh0_s4 ] correspond to mass ranges already excluded by experimental measurements of the @xmath138 parameter ( cf .
section [ sec:2hdm ] ) . on general grounds , the basic phenomenological features are common to the @xmath140 and @xmath141 channels , with the only proviso that , in the latter case , the resulting production rates are slightly smaller simply because @xmath142 .
the final cross - sections are nevertheless comparable and reach up to a few dozen fb . in table
[ tab : xsheavy ] we collect , in a nutshell , the essential results from a dedicated numerical analysis of @xmath143 for all sets of higgs boson masses ( cf .
table [ tab : masses ] ) and different values of the tree - level @xmath144 coupling , again at fixed @xmath126 and using the largest permissible value of @xmath121 in each set ( see table [ tab : maxl5 ] ) .
the relative size of the quantum corrections @xmath117 is also included in table [ tab : xsheavy ] . as in the previous channel , we have used fiducial center - of - mass energies @xmath145 respectively .
the maximum production rates at @xmath128 are achieved for relatively light higgs boson masses ( set i ) and for @xmath146 , in which case the tree - level @xmath144 interaction is not suppressed , see eq .
( [ eq : lag12 ] ) . within this setup ,
one is left with @xmath147 .
indeed the production rates lie at the @xmath148 level in a broad range around @xmath146 for all the mass sets under analysis .
if we nevertheless consider @xmath149 , when the tree - level coupling gets suppressed by @xmath150 , the resulting cross sections are still at the level of a few fb each femtobarn rending 500 events per 500 @xmath151 of integrated luminosity .
radiative corrections thus may leave a solid imprint into the final predicted @xmath131 of the current process .
in particular , for @xmath110 quantum effects rubber - stamp a characteristic @xmath152 boost upon the tree - level cross - section , which could hardly be missed .
other choices of @xmath72 amount to negative one - loop contributions in which @xmath144 is dramatically suppressed .
for instance , in the case of set i for @xmath153 , it renders @xmath154 . for center - of - mass energies from around @xmath155 onwards ,
a powerful destructive interference is seen to be under way between the born and the one - loop amplitudes .
this translates into large ( negative ) quantum corrections which may well reach @xmath156 , thus literally stamping out the signal and pulling it down at the @xmath157 level or even below . to better
illustrate this feature , in figs .
[ fig : overs_s2_heavy][fig : overs_s4_heavy ] we plot the behavior of @xmath158 ( top panels ) and @xmath117 ( bottom panels ) as a function of @xmath159 for the representative sets ii - iv of higgs boson masses of table [ tab : masses ] and for different values of the tree - level coupling ; viz .
@xmath146 ( left panels ) , which corresponds to the maximum of the tree - level @xmath144 coupling , and @xmath160 ( right panels ) , corresponding to the _ fermiophobic _ limit for @xmath74 boson within type - i 2hdm ( cf .
table [ typeiandii ] ) .
again we fix @xmath126 and we vary @xmath6 within the range allowed by unitarity and vacuum stability restrictions . as it was already observed in the analysis of the @xmath140 channel , heavier higgs - boson spectra furnish smaller production rates .
we refrain from extending the discussion , which is entirely similar to that of @xmath123 . to our knowledge ,
no pre - existing study of the one - loop effects on the neutral higgs boson channels ( [ 2h ] ) can be found in the literature .
let us , however , briefly comment on the prospects for the charged higgs - pair channel , @xmath161 .
this was considered some time ago in references @xcite .
these old studies were devoted to the computation of @xmath162 at one - loop in both the mssm and the 2hdm .
in the latter case , they also uncover very large quantum effects ( of order @xmath163 or even higher is mandatory before we can judiciously compare with the results that we have presented here for the neutral higgs boson channels . however , due to the increase of the charged higgs boson mass bounds in the last few years , the maximum cross - section rates should correspondingly fall down , thereby making this process less competitive . ] ) which are attributed to enhanced 3h self - interactions involving charged higgs bosons .
such quantum effects may lie well above the typical mssm counterparts , which are found to render up to @xmath164 , and mostly negative .
let us , however , remark that the @xmath165 channel could operate at sizable cross - sections levels only for type - i models since the type - ii models require a rather heavy mass spectrum .
in this work , we have undertaken a thorough investigation of the pairwise production of neutral higgs - bosons within the context of the general two - higgs - doublet model ( 2hdm ) at the ilc / clic linear colliders ( linac ) through the basic processes @xmath166 . to the best of our knowledge
, this study is the first one existing in the literature which addresses these important processes at the one - loop level . to sum up the main objectives of our work
, we can assert that we have concentrated on the following three tasks : i ) we have , on the one hand , singled out the most favorable regimes for the higgs - pair production processes once the radiative corrections are taken into account , specifically within the framework of the on - shell renormalization scheme with an appropriate definition of @xmath68 and of the @xmath16-even mixing angle @xmath72 at the quantum level ; ii ) next , we have carefully quantified the importance of the leading quantum corrections , and not only so , we have pinpointed also the origin and relative weigh of the various sources ; and , finally , iii ) we have correlated the most powerful source of quantum effects with the strength of the higgs boson self - interactions ( essentially the triple higgs boson self - couplings @xmath167 ) , to which the processes @xmath166 are highly sensitive through higgs - mediated one - loop diagrams . with these aims in mind , we have systematically searched over the 2hdm parameter space by considering different higgs boson mass scenarios and different values of the tree - level @xmath168 and @xmath144 couplings . at the same time , we have carefully explored the very frontiers of the generic 2hdm parameter space , specially the limits that can not be trespassed by the strengths of the various 3h self - couplings . among the various subtle triggers that control the size of these important couplings , one of them stands out with overwhelming supremacy over the others , and this is the @xmath6 coupling .
this is the only one parameter in the list of seven free parameters cf .
( [ freep ] ) conforming the structure of the general ( @xmath16-invariant ) higgs potential of the 2hdm that can not be ultimately traded for a physical mass or a mixing angle .
this parameter is , in addition , the only higgs self - coupling that need not be renormalized at one loop in the current framework . despite its special status of being apparently unfettered and without obvious physical connotations , one quickly discovers that it becomes severely restrained by subtle bounds connected with the quantum field theoretical consistency of the model , namely the stringent constraints imposed by tree - level unitarity and vacuum stability conditions .
these restrictions , together with the limitations dictated by custodial symmetry , and of course also those emerging in the light of the direct collider bounds on the masses of generic higgs boson models , not to mention the indirect phenomenological restraints ( in particular , the upper bound on the charged higgs mass in type - ii models ensuing from radiative @xmath169-meson decays ) , do indeed eventually determine a fairly compact region of parameter space .
in particular , @xmath6 is compelled to be mostly negative and in the range @xmath170 . at the same time
, @xmath68 can not be large when @xmath121 is also large without violating the constraints .
the bottom line is that the value that maximizes the processes under consideration is a rather equitable one : @xmath171 . notwithstanding the sharp limits placed upon the physical region of the parameter space , our explicit calculation of radiative corrections has shown and we feel it is truly remarkable that the general 2hdm models are still able to unleash a good deal of their `` repressed '' power . if only we could catch a smoking gun of this stupendous potential through the accurate measurements that a linear collider should be able to render on fundamental processes like @xmath172 , we might be on the verge of experiencing an intense episode of higgs boson physics beyond the sm . in the following ,
we describe the basic results and strategies that we could follow to this effect .
we have identified three main scenarios of potential phenomenological interest in the corresponding parameter space . in all of them
we observe no remarkable dependence either on the details of the higgs mass spectrum nor on the pattern of higgs - fermion yukawa couplings ( either for type - i or type - ii 2hdm ) , not even on the particular channel @xmath173 under study . in what follows
we describe each one of these scenarios and spell out clearly their differences , in particular we provide the characteristic potential size of the relative quantum correction @xmath117 cf .
eq.([deltar ] ) typically associated to them ( cf . figs.[fig : scan - s2]-[fig : scan - s4 ] ) : * * scenario 1*. to start with , regions of @xmath174 and large @xmath121 ( specifically , @xmath175 ) support the bulk of the radiative corrections .
they may well reach @xmath176 at the largest attainable values of @xmath121 .
furthermore , these effects critically depend on the actual size of @xmath121 and on the center - of - mass energy @xmath159 ( cf .
figs.[fig : overs_s2]-[fig : overs_s4 ] ) , to wit : @xmath117 is positive at `` low '' center - of - mass energies around @xmath177 , while it is negative for @xmath178 and certainly also in the uppermost foreseeable energy segments @xmath179 , usually reserved for an upgraded ilc and specially for clic .
thus , interestingly enough , quantum corrections turn out to drastically invigorate the cross section of the basic processes @xmath180 near the standard booting energy of the ilc , while they tend to suppress the two - body higgs boson signal for a center - of - mass energy some @xmath181 beyond this initial regime and of course for the tev range and above as a result of constructive / destructive interference effects , respectively , between the tree - level and the one - loop amplitudes .
incidentally , we clarify that this optimal scenario can not be iterated for large and positive values of @xmath6 because the combined set of constraints ( mainly the vacuum stability bounds , in this case ) preclude most of the region @xmath182 ; * * scenario 2*. on the other hand , we find a tail of subleading effects within a band of the @xmath183 subspace at moderate @xmath184 and @xmath185 wherein the maximum radiative corrections turn out to stagnate and remain barely at the level of @xmath186 or less .
should the parameters of the model inhabit this region , the practical possibilities to detect these effects would obviously become thinner than in the previous case .
still , @xmath187 effects are not that alien to the high precision standards planned for a linac collider , and hence it should not deter us from the searching task . at this point , the detection strategies associated to the higgs boson decay patterns could be essential ( see below ) ; * * scenario 3*. finally , there is the region of parameter space where the coupling @xmath6 is very small or it just vanishes . here
the quantum corrections can still have a chance provided @xmath188 , and indeed we have found domains of this kind preserving all the basic constraints . while this region is not the most favored one ,
neither from the theoretical point of view nor from the experimental data , specially if @xmath189 , we should keep it in mind as a possibility .
even for moderately small @xmath68 , say for @xmath190 , it can be the source of sizeable quantum effects of order @xmath191 ( and positive ) , which are larger than those in scenario 2 and , in some cases , even comparable to those in scenario 1 .
it is worth noticing that , in this region , there is an interesting collaborative interplay between the higgs boson trilinear self - couplings and the top quark yukawa coupling ( which is the same in type i or type ii 2hdm s ) . only in this scenario
a yukawa coupling could play a significant quantitative role on equal footing with the trilinear higgs boson self - couplings .
obviously , the success of the whole higgs boson search endeavor will also depend on the pattern of distinctive signatures available to these particles in the final state .
for instance , for @xmath192 , we should basically expect back - to - back pairs of highly energetic ( roughly @xmath193 ) and collimated b - quark and/or @xmath194-lepton jets from @xmath195 which , in contrast to the lhc , . and of course the same reasoning applies to the @xmath196 final state . for @xmath197 ,
on the other hand , signatures from @xmath198 and @xmath199 could also play a role ( depending on the particular choice of @xmath72 and @xmath108 , cf .
table [ tab : coup ] ) .
they could ultimately lead to signatures with two ( up to four ) charged leptons against a @xmath200 or @xmath201 pair in the final state ( from @xmath202 and , specially , from @xmath203 ) .
even if the branching ratios of gauge bosons into leptons are very small ( @xmath204),the predicted number of higgs - pair events is large enough so that , in the most favorable scenarios , one could collect @xmath205 events for these complementary signatures , which could be instrumental for a final tagging and identification of the higgs bosons . notice that we use the fact that the @xmath16-odd @xmath21 `` always '' decays into @xmath200 or , to a lesser extent , to @xmath201 ( even for @xmath206 ) as it can not couple to gauge bosons .
so the patterns of signatures are rather characteristic and could not be missed in normal circumstances .
but there is more than meets the eye in the whole strategy game , as we shall see later on . in the most favorable regimes ( scenario 1 )
, the experimental chances could be spectacular , especially in the case of relatively light higgs boson mass spectrum and a large ( negative ) value of @xmath6 , for which the predicted one - loop cross sections @xmath207 at @xmath208 may border the @xmath209 level .
this would translate into barely @xmath210 events per @xmath211 of integrated luminosity at the linac .
while the former situation describes perhaps the most optimistic possibility , cross sections of @xmath148 at @xmath208 should be quite common in a wide patch of the 2hdm parameter space , and correspondingly of @xmath212 at @xmath213 ( see table [ tab : masses ] ) , thus delivering rates that range from a few hundred to a few thousand events respectively .
with this level of statistics ( plus a comparable number of events from the @xmath141 channel ) it should be perfectly possible to insure a comfortable tagging of the higgs bosons in the clean environment of a linac machine .
most important , we have been able to demonstrate that the dominant quantum effects are , as we expected , primarily nurtured by the higgs - mediated one - loop corrections to the @xmath214 vertices , and can be ultimately traced back to the potentially enhanced 3h self - interactions @xmath167 ( cf . table ii ) .
this is indeed a trademark feature of the 2hdm , with no counterpart in the mssm in which susy invariance highly curtails the structure of the higgs boson self - interactions and compels them to be purely gauge - like .
obviously , probing the structure of higgs boson self - interactions in the present case through the analysis of radiative corrections on direct production processes is a most useful strategy to disentangle susy and non - susy higgs physics scenarios . to be specific , for the analysis of the @xmath215 cross sections we have faced the computation of the full set of @xmath3 quantum effects , among them the subset of @xmath216 corrections , and supplemented them with the leading @xmath4 pieces arising from the square of the one - loop diagrams involving higgs boson self - couplings .
these ( finer ) corrections are of @xmath217 ( see , however , below ) .
their inclusion is actually _ a must _ at large center - of - mass energies ( i.e. once @xmath159 has left behind the startup situation , and of course also in the @xmath218 regime ) .
the reason is that the tree - level @xmath219 and the one - loop @xmath3 amplitudes virtually cancel out each other at high energies , owing to the characteristic large and negative quantum effects in this regime .
we can convince ourselves of this fact from the severe depletion exhibited by the one - loop curves
see e.g. figs .
[ fig : overs_s2]-[fig : overs_s4 ] as compared to the tree - level ones .
typically , the effect appears after we increase the center - of - mass energy by @xmath116 beyond the initial @xmath220 , and persists till 1 tev and afterwards . at large energies , the inclusion of the @xmath217 effects is a consistency requirement as it prevents the cross - section from becoming negative .
these terms can actually be relevant also at the initial energy interval @xmath221 at which the linac will first operate ; for example , for the scenarios with maximally enhanced 3h self - couplings , they may lead to an additional @xmath222 contribution to @xmath117 , hence pushing the overall quantum correction even higher . in this respect
, we note that our earlier estimate of these higher order terms , eq.([eq : estimate2 ] ) , was able to hit the right order of magnitude , although it actually underestimates the maximum size revealed by the exact numerical analysis not too surprising from such a rough attempt at guessing their bulk size .
let us clarify that the two - loop diagrams involving the higgs boson - mediated @xmath214 vertex corrections , lead of course also to amplitudes of order @xmath217 .
these are actually finite , and originate from the interference of the tree - level amplitude with the two - loop one .
an obvious concern is then the following : can we safely neglect them ? upon inspection of the higgs - mediated two - loop diagrams , and taking into account power counting and dynamical considerations , one can show that they are actually suppressed , and particularly so for those scenarios where @xmath223 is optimized .
but , most significantly , there is a deeper observation to be adduced at this point , which is well in the spirit of the effective field theory approach to qft .
the ( finite ) two - loop effects induced by the enhanced higgs boson self - couplings can be conveniently reabsorbed in the form of one - loop corrections to the bare higgs self - couplings , which first appear at the one - loop level in the process under consideration .
we can iterate this algorithm at any order in perturbation theory by further reabsorbing the higher order higgs mediated loops into the triple _ and _ quartic higgs self - couplings introduced at one - loop order . in this way we can define a collection of effective 3h and 4h self - couplings that encapsulate all these higher order effects .
the above remark is an interesting one , as it tells us that the application of the stringent tree - level unitarity relations that we have used can be directly reiterated for these effective couplings .
therefore , no matter how big are the quantum corrections induced by the 3h self - couplings in subsequent orders of perturbation theory , their overall effect is constrained by the same set of formal unitarity relations .
if expressed in terms of the original couplings , these relations will be equally stringent so that , on balance , the maximum enhancement capabilities of the 3h self - couplings are basically the same at any order of perturbation theory .
this is the procedure adopted in the present work , and we believe it is perfectly consistent , in the sense that we have retained the contributions to the @xmath224 amplitudes at leading order in the 3h self - couplings while employing the unitarity constraints of ref .
@xcite-@xcite , also derived at leading order in @xmath167 .
how do our results compare with the expectations in the mssm ? as we have stressed in section ii ,
a lot of work on higgs boson production in @xmath24 colliders has been reported in the literature thus far , but mainly within the context of the mssm .
in particular , supersymmetric radiative corrections to @xmath122 have been considered in detail , although mostly in the context of lep and including sometimes a more or less timid incursion into the tev - class colliders , cf .
let us try to see how some of these results compare with ours .
for the sake of definiteness , let us concentrate on the particular case of ref .
@xcite , in which @xmath225 is analyzed in the mssm for a linear collider operating at lep 2 energy and further extended up to @xmath128 . in the most favorable regimes
, the loop - corrected cross sections may reach up to barely @xmath226 at @xmath128 , which falls in the ballpark of the results that we have obtained here within the 2hdm .
this is not surprising because in both cases the leading amplitude ( i.e. the lowest order one ) is purely gauge and hence the order of magnitude of the cross - section is independent of whether we compute it in the mssm or in a generic 2hdm .
quantum effects can be fairly efficient in the mssm too , and they may also boost the tree - level predictions substantially ( @xmath227 ) . however , their size is never as big as the maximal 2hdm contributions ; and , no less important for our conceptual understanding , the fundamental origin of these quantum effects is completely different from that of the 2hdm . as already mentioned in the introduction , the leading quantum effects in the mssm are due to the genuine higgs - quark , higgs - squark and quark - squark - chargino / neutralino yukawa - like supersymmetric couplings , whose contributions carry enhancement factors of the fashion @xmath228 , none of them related to the structure of the higgs potential and hence none of them associated to higgs boson self - couplings . in contradistinction to the mssm ,
the 3h self - couplings in the 2hdm embody the full potential for triggering quantum effects in physical processes , and these effects become thence tied to the very structure of the higgs potential .
even if quantitatively similar in some scenarios , quantum effects in susy and non - susy two - higgs doublet extensions of the sm are , therefore , prompted by intrinsically different mechanisms . in practice ,
the task of distinguishing whether the produced higgs boson particles at a linac are supersymmetric or not can start by addressing the decaying signatures of the higgs bosons into @xmath200 or @xmath201 pairs , as sketched above .
however , in certain regions of parameter space the number of events can be by itself a direct signature of the sort of higgs model that we have behind .
for example , in the mssm , the tree - level coupling @xmath229 undergoes a severe suppression with growing values of @xmath68 and/or @xmath125 as @xmath72 depends on @xmath125 in the mssm .
thereby the predicted cross sections , either at the born or at the one - loop level , are dramatically weakened as a function of @xmath125 .
for instance , @xmath230 at @xmath231 and @xmath232 may end up as tiny as @xmath233 ( see e.g. fig .
9 of ref .
@xcite ) while the 2hdm prediction is of @xmath148 , i.e. one thousand times bigger !
but there is more to say here , as we have announced above .
the identification strategy can be highly facilitated through the interplay of further 2h ( and eventually 3h ) channels , such as ( [ 3h ] ) and ( [ 2hx ] ) , which are also highly distinctive in the 2hdm see @xcite and @xcite .
in fact , for these channels the difference with the mssm can be apparent already at the lowest order of perturbation theory .
let us be a bit more quantitative here . in table
[ tableend ] , we compare the numerical predictions on the cross - sections of these various processes in truly equitable conditions , i.e. for different sets of common parameters for all the processes , and for three realistic values of the center - of - mass energy of the ilc / clic ( the latter operating always at the highest edge of the energy band , although it could nominally reach the 3tev end ) . to be more precise , the cross - sections that we are comparing on equal footing in table [ tableend ] are the following : i ) the one - loop corrected @xmath234 ; ii ) the leading - order @xmath3 cross - sections of the two triple higgs production processes @xmath235 ; and iii ) @xmath236 ) at leading - order @xmath4 .
notice that the bulk of the contribution from processes ii ) and iii ) comes from the @xmath237 and @xmath238 parts , respectively . in all cases
we take @xmath126 , @xmath127 and the maximum allowed value of @xmath121 according to table [ tab : maxl5 ] , that is , the most favorable scenario for @xmath140 pair production .
we use sets i - iii for the higgs boson masses .
the resulting predictions for the different channels turn out to be highly illustrative of the complementary nature of such higgs - boson production signatures . for a thorough discussion of the dynamical features and signatures of the multiparticle processes ii ) and iii ) , we refer the reader to the original references @xcite . here
we have recomputed the cross - sections ii ) and iii ) under the same kinematical conditions and in the very same region of the parameter space as the main processes i ) under consideration , which is of course the indispensable first step that enables us to compare them all meaningfully . in this way , we are well prepared to infer the final strategy that can be designed to efficiently search for 2hdm higgs bosons at the linear colliders .
@xmath239 & @xmath240 + + @xmath140 & 34.13 & 2.89 & 0.70 + @xmath241 & 3.09 & 8.58 & 5.17 + @xmath242 & 6.75 & 19.65 & 12.27 + @xmath243 & 13.29 & 79.00 & 146.08 + + @xmath140 & 26.71 & 4.07 & 1.27 + @xmath241 & 0.02 & 5.03 & 3.55 + @xmath242 & 0.17 & 11.93 & 8.39 + @xmath243 & 1.47 & 17.36 & 38.01 + + @xmath140 & 11.63 & 6.11 & 2.52 + @xmath241 & below threshold & 1.25 & 1.33 + @xmath242 & below threshold & 0.69 & 2.14 + @xmath243 & 0.92 & 9.72 & 23.40 + the strategy that follows from the results encapsulated in table [ tableend ] and the information that we have gathered in the figures of the last section should be by now crystal - clear , and it comes down to the following two - step procedure : * * step 1*. at the vicinity of the startup energy of the ilc ( approximately in the range @xmath244 ) , the exclusive neutral double higgs boson channels @xmath245 eq . ( [ 2h ] ) prove to be the dominant ones as compared to the triple higgs boson production processes @xmath246 and the inclusive double higgs production ones @xmath247 , see equations ( [ 3h ] ) and ( [ 2hx ] ) . therefore , in this lower energy band , the study of potential signatures of new physics , and in particular the identification of the nature of the produced higgs bosons at the ilc , must be conducted through the careful analysis of the quantum corrections affecting the two - body channels @xmath248 . here is precisely where the detailed results of the present work could be most useful , and indeed should be the first analysis to be implemented when the linear colliders are set to work in the future ; * * step 2*. at higher energies , however , say for @xmath249 gev and , for that matter , in the entire tev range ( hence at the highest nominal energy regimes scheduled for the ilc , and specially for clic ) , the influence of the @xmath89-propagator stifles dramatically the cross - section of the exclusive neutral double higgs boson channels @xmath248 .
therefore , in these high energy domains , they can no longer compete with the mechanisms providing multiparticle final states , such as ( [ 3h ] ) and ( [ 2hx ] ) .
the latter process , @xmath250 , becomes indeed the most efficient one at the highest energies since it is not crippled by the s - channel propagator ; quite on the contrary , its cross - section increases steadily with the energy ( see ref.@xcite for details ) .
therefore , in the upgraded phase of the ilc , and certainly for the clic collider , one must concentrate all the search strategy power on looking for an anomalously large number of inclusive higgs boson pairs of the type @xmath251 , @xmath252 , @xmath140 and @xmath253 , which should emerge mostly acollinear and dynamically thrusted along the beam direction , rather than appearing in a simple back - to - back geometry characteristic of the two - body process @xmath248 .
notice furthermore that , in contradistinction to the latter , the fused pairs consist of identical higgs bosons .
unmistakeably , the predicted numbers in table [ tableend ] could not , by any means , be accounted for if they were to be ascribed to a supersymmetric origin .
finally , the cross - correlation of these higher energy effects with the previous ones at lower energies , which we could track very well while upgrading the ilc collider from @xmath254tev up to @xmath255tev and eventually till @xmath256tev ( through clic ) should provide plenty of unambiguous evidence of non - supersymmetric higgs physics gauge boson : @xmath257 ( @xmath258 ) , also called `` higgs - strahlung '' processes ; see @xcite for a thorough study at the quantum level in the general 2hdm . ] .
let us finally introduce a remark that hints at the potentially far reaching implications of the 2hdm dynamics in different sectors of particle physics .
the triple higgs boson self - couplings could also have an indirect impact on the best high precision observables at our disposal , even well before the ilc / clic colliders are commissioned and put effectively to work . for example , they could have a bearing on the famous electroweak precision parameter @xmath259
see eq.([deltar ] ) .
since our main aim here has been to exploit the leading contributions from the higgs boson self - couplings in the arena of the direct production processes ( [ 2h ] ) , the @xmath97 part did not enter our quantum computation ; it would only enter at two - loop level .
however , we suspect@xcite that the influence of the 3h self - couplings on @xmath97 might have a real interest _ per se
_ , for it could induce a correction to @xmath138 and ultimately trigger a shift in the @xmath83 mass , and the result might well compete with the highly accurate calculations that have been performed on this observable within the alternative framework of the mssm@xcite .
to summarize , if the genuine enhancement properties of the 2hdm effectively hold in the real world , the combined analysis of the exclusive double higgs production ( 2h ) processes at the startup energy of the ilc , @xmath220 , and subsequently of the triple ( 3h ) and inclusive double higgs production processes ( @xmath260 ) at the upgraded ilc / clic ( @xmath261 ) , might reveal strong hints of higgs boson physics beyond the sm .
the bottom line of our study is that the physics of the linear colliders can be truly instrumental to unveil the nature of the higgs bosons .
these bosons may , or may not , have been discovered during the lhc era while the linear colliders take the floor , but the real issue at stake here is of highest priority and should strengthen the need for such machines .
high precision higgs boson physics in a linac can indeed provide a keen insight into the most sensitive building blocks of modern gauge theories of weak and electromagnetic interactions , viz . in the very core architecture of the electroweak symmetry breaking .
this work has been supported in part by the eu project rtn mrtn - ct-2006 - 035505 heptools .
dlv thanks an esr position of this network , and also the theory group at the max - planck institut fr physik in munich for the warm hospitality extended to him while part of this project was being carried out .
the work of dlv has also been supported partially by the mec fpu grant ref .
ap2006 - 00357 .
js has been supported in part by mec and feder under project fpa2007 - 66665 , by the spanish consolider - ingenio 2010 program cpan csd2007 - 00042 and by diue / cur generalitat de catalunya under project 2009sgr502 .
js also thanks the mpi in munich for the hospitality and financial support while part of this work was under way .
the authors are grateful to prof .
w. hollik for useful discussions on this topic and for his support .
early discussions with n. bernal and j. guasch are also acknowledged .
s. weinberg , phys .
* 19 * ( 1967 ) 1264 ; a. salam , _ elementary particle theory ; nobel symposium _ , no.8 , ed .
n. svartholm ( almqvist and wiksell , stockholm 1969 ) ; s.l .
glashow , _ nucl .
* 22 * ( 1961 ) 579 .
, , , , * * , ( ) , ; , , , * * , ( ) , ; a. bartl _
et al . _ ,
b378 * ( 1996 ) 167 , arxiv : hep - ph/9511385 ; , , , , , * * , ( ) , ; j.a .
coarasa , ricardo a. jimnez , joan sol , phys . lett .
* b406 * ( 1997 ) 337 , .
a. djouadi , m. spira , p.m. zerwas , phys . lett . * b264 * ( 1991 ) 440 ; m. spira , a. djouadi , d. graudenz , p.m. zerwas , nucl
b453 * ( 1995 ) 17 ; s. dawson , c. kao , y. wang , phys . rev . *
d77 * ( 2008 ) 113005 ; w. hollik , t. plehn , m. rauch , h. rzehak , phys .
* 102 * ( 2009 ) 091802 .
s. kanemura , s. moretti , y. mukai , r. santos , k. yagyu , ; a. arhrib , r. benbrik , c.h .
chen , r. guedes , r. santos , jhep 0908 ( 2009 ) 035 , ; s.s .
bao , y.l .
wu , ; m. beccaria , g. macorini , l. panizzi , f.m .
renard , c. verzegnassi , .
|
the pairwise production of neutral higgs bosons ( @xmath0@xmath1 , @xmath2@xmath1 ) is analyzed in the context of the future linear colliders , such as the ilc and clic , within the general two - higgs - doublet model ( 2hdm ) .
the corresponding cross - sections are computed at the one - loop level , including the full set of contributions at order @xmath3 together with the leading @xmath4 terms , in full compliance with the current phenomenological bounds and the stringent theoretical constraints inherent to the consistency of the model .
we uncover regions across the 2hdm parameter space , mainly for low @xmath5 and moderate and negative values of the relevant @xmath6 parameter , wherein the radiative corrections to the higgs - pair production cross section , @xmath7 , can comfortably reach @xmath8 .
this behavior can be traced back to the enhancement capabilities of the trilinear higgs self - interactions a trademark feature of the 2hdm , with no counterpart in the minimal supersymmetric standard model ( mssm ) .
the corrections are strongly dependent on the actual value of @xmath6 as well as on the higgs mass spectrum .
interestingly enough , the quantum effects are positive for energies around @xmath9 , thereby producing a significant enhancement in the expected number of events precisely around the fiducial startup energy of the ilc
. the higgs - pair production rates can be substantial , typically a few tens of femtobarn , therefore amounting to a few thousand events per @xmath10 of integrated luminosity .
in contrast , the corrections are negative in the highest energy range ( viz . @xmath11 and above ) .
we conclude that a precise measurement of the 2h final states could carry unambiguous footprints of an extended ( non - supersymmetric ) higgs sector . finally , to better assess the scope of these effects , we compare the exclusive pairwise production of higgs bosons with the inclusive gauge boson fusion channels leading to @xmath12 final states , and also with the exclusive triple higgs boson production .
we find that these multiparticle final states can be highly complementary in the overall higgs bosons search strategy .
|
two schools currently exist for the numerical exploration of quantum gravity which are both based on simplicial discretization of spacetime . in the dynamical triangulation approach @xcite
the edge lengths on the discretized lattice are fixed and connectivities are allowed to vary whereas in the regge calculus approach @xcite , at least insofar as practical implementations are concerned , the connectivities are fixed and edge lengths are allowed to vary . in two dimensions
dynamical triangulations are a direct implementation of matrix models and the results obtained are in good agreement with the theory , whereas there is currently some uncertainty as to whether the regge calculus agrees with analytical results @xcite .
indeed , some recent results raise the question of whether two dimensions is a suitable testing ground for discretized quantum gravity at all @xcite . in higher dimensions , however , the regge calculus and dynamical triangulations appear to be in at least qualitative agreement on the phase structure of quantum gravity @xcite . all of the work alluded to above has explored discretized gravity in the euclidean sector and its relation to lorentzian gravity is a moot point , as is also the case in the continuum theory . in principle
the regge approach is not shackled to euclidean space .
the volume of a regge simplex in @xmath0 dimensions is given by @xmath1 where the vectors @xmath2 start at some basis vertex , say @xmath3 , and point to vertex @xmath4 thus spanning the simplex .
in the two dimensional case the volume formula gives the the area of a triangle as @xmath5 where @xmath6 are the squared edge lengths of the triangle .
the triangle can be lorentzian if the determinant is less than zero or euclidean if it is greater than zero . in simulations to date using regge calculus the edge lengths have been generally taken to be continuous variables with triangle inequalities being explicitly enforced to ensure that the triangles stay euclidean .
one interesting alternative scheme that has been proposed is the so - called ising link quantum gravity in which only two squared edge lengths are allowed , @xmath7 , where @xmath8 is an ising variable and @xmath9 is chosen to maintain euclidean triangles ( @xmath10 in two dimensions ) @xcite . in the spirit of the ising link approach , we consider a generalization of dynamical triangulation with both spacelike and timelike edges , @xmath11 .
we can accommodate both euclidean and lorentzian triangles in this scheme : triangles with @xmath12 and @xmath13 edges give euclidean areas from equ.([area ] ) ; whereas triangles with either @xmath14 or @xmath15 edges give lorentzian areas .
it is possible to take the edge lengths squared themselves as ising spins , but instead we will take them to be the product of ising spins at the end of each link , which will allow us to derive some properties of our toy model of quantum gravity in the lorentz sector from already solved ising models in the euclidean sector . if we allow all four possible sorts of triangle and assign fugacities @xmath16 to the euclidean triangles and @xmath17 to the lorentzian triangles we can write down the following partition function for our model @xmath18 where the outer sum runs over ( euclidean ) triangulations and the inner sum is over the different possible ways of soldering the various sorts of triangles together .
this partition function can be retrieved as the free energy of a two - matrix model @xcite whose perturbative expansion generates our triangulated universe of mixed lorentzian and euclidean patches @xmath19 where we use the @xmath20 hermitian matrix @xmath21 to represent a spacelike edge and the @xmath20 hermitian matrix @xmath22 to represent a timelike edge .
the @xmath23 , @xmath24 terms thus represents the euclidean triangles , the @xmath25 , @xmath26 terms the lorentzian triangles and the @xmath27 and @xmath28 terms solder together different triangles .
as it stands this model has not been solved graphs gives an action with identical terms to equ.([action0 ] ) but the coefficients of @xmath23 and @xmath24 are not equal , nor are those in front of @xmath29 and @xmath25 . ]
so we make a further bold simplification and restrict ourselves by fiat to one sort of euclidean triangle , namely @xmath12 , and one sort of lorentzian triangle , @xmath14 which leaves us with the following matrix model @xmath30 this two matrix model is known to represent the ising model on dynamical triangulations @xcite with the spins at the triangle vertices , rather than the more usual formulation with the spins at triangle centres @xcite . in this context
, @xmath21 represents edges with similar spins at each end and @xmath22 represents edges with differing spins at each end . in figures.1 - 4
we denote the @xmath21 edges by bold lines and the @xmath22 edges by dotted lines , as well as indicating the corresponding vertex spin arrangements . after suitable scalings
the ising action on dynamical triangulations can be written as @xmath31 so we see that the zero temperature ( @xmath32 ) limit is the euclidean sector when the model is regarded as ising link quantum gravity , whereas an antiferromagnetic limit ( @xmath33 , @xmath34 with @xmath35 finite ) can be regarded as the purely lorentzian limit . in this
limit the action becomes @xmath36 and the gaussian integration over @xmath21 may be performed , giving @xmath37 which is again a pure gravity action . in geometrical terms
this means that the triangulations generated in the lorentzian limit that gives equ.([actionii ] ) are those which admit a perfect matching in which the triangles are paired off into squares @xcite .
this is clear from equ.([actioniii ] ) where the effect of integrating out the @xmath21 is to give a @xmath38 term that generates squares .
if we take the opposite convention for the lorentzian triangles and keep @xmath12 and @xmath39 triangles only we have the action @xmath40 which is is now gaussian in @xmath22 , allowing it to be integrated out to give @xmath41 this is an action for a universe built out of both triangles and squares .
this action is known to give the yang - lee edge singularity on a random surface for suitably tuned @xmath16 and @xmath17 @xcite , but generically it will be pure gravity . again
the purely lorenztian limit is to take @xmath42 which leaves us with a theory of squares alone ( but with spacelike edges , rather than the timelike edges of equ.([actioniii ] ) ) .
we have glossed over one important respect in which our toy model fails to be realistic : we have implicitly assumed that @xmath16 and @xmath17 are _ real _ to get the standard critical behaviour , ie @xmath43 , @xmath44 , so that for a fixed topology @xmath45 where @xmath0 is the number of vertices in a triangulation . in the continuum
limit the exponential gives the cosmological constant term @xmath46 without a factor of @xmath4 in front , whether we are in the euclidean sector or the dense lorentzian sector .
, that has been suppressed at fixed topology in the above also fails to display a factor of @xmath4 in the continuum limit , giving @xmath47 .
] however , it is known that the action in equ.([actioniii ] ) may be defined for complex @xmath17 @xcite , so we can approach the standard continuum limit along a complex path , which is precisely what is required for obtaining a factor of @xmath4 in the action .
the action for the yang - lee edge singularity in equ([actionv ] ) may also be analytically continued to complex couplings . in any event
, these problems are swept under the carpet in a microcanonical simulation at fixed topology as the `` factors of @xmath4 '' only come into play when the number of vertices or the topology are changed .
the problem re - emerges when one considers applying a scaling analysis to simulations of differently sized lattices .
it is interesting to note in passing that a rather more sophisticated form of regge calculus with more than two quantized edge lengths devised by ponzano and regge for three dimensional gravity @xcite also fails to reproduce the correct factors of @xmath4 .
when the continuum limit is taken in the lorentzian sector such a theory gives no @xmath4 in front of the action @xcite , whereas taking the continuum limit in the euclidean sector does give a factor of @xmath4 .
in a numerical simulation of the model the allowed moves for various sorts of edges can be deduced by looking at the ising interpretation of the model .
one new feature compared with an undecorated euclidean triangulation is that the standard `` flip '' move used in simulations can have the effect of changing the sign of an edge , depending on the local vertex spin configuration , which in turn has the effect of changing the signature of one of the participating triangles .
the model thus incorporates local signature changing moves .
the choice of only @xmath12 triangles is completely trivial in a purely euclidean theory - it is simply a matter of convention . in a purely lorentzian sector
we can see that the choice of @xmath14 triangles is compatible with the dynamical triangulation moves shown in figure.1 , neither of which change the sign of the flipped edge or the nature of the participating triangles .
the further two flip moves shown in figure.2 _ do _ change the sign of the flipped edge and exchange lorentzian and euclidean triangles .
these signature changing moves will be present if we choose a non - zero fugacity @xmath16 for the euclidean triangles .
as we have defined the signature of an edge in terms of the ising spins living at either end it is also possible for a changing spin to influence the local geometry .
an example of this in the purely lorentzian sector is shown in figure.3 where flipping the central spin still keeps only @xmath25 triangles but rotates the bold @xmath21 edges around the central point . in general any vertex with an equal number of alternating @xmath21 and @xmath22 edges emanating from it
allows a similar sort of move , whilst still keeping the restriction to @xmath25 triangles only . in all cases ,
as we might expect given the underlying matrix model of equ.([action ] ) , @xmath48 and @xmath49 triangles are consistently excluded , if the starting configuration does not contain them .
for a lattice of toroidal topology , for instance , a section of a suitable starting configuration might look like figure.4 in which horizontal spacelike lines are joined by the inclined timelike edges .
such a starting configuration gives a foliation of the spacetime manifold by spacelike lines .
a lattice of spherical topology would require in addition `` caps '' of purely timelike edges radiating from the poles .
we conducted a simulation of the model with the action in equ.([actionii ] ) ( `` perfect matching '' ) which is in a purely lorentzian sector according to our definitions and measured the probability @xmath50 for the occurence of vertices with coordination number @xmath0 .
these are listed in the following table for @xmath51 for a toroidal lattice with 1024 vertices and a statistics of 40k _ independent _ configurations obtained via 1 m sweeps .
the statistical error is about @xmath52 percent .
the columns @xmath53 and @xmath54 contain the numerical values for lorentz and euclidean sector respectively , while @xmath55 shows the exact result for the euclidean sector .
[ cols= " < , < , < , > " , ] * table 1 : * measured values of the probability @xmath50 for the occurence of vertices with coordination number @xmath0 .
we can see that the main effect in the lorentz sector is to increase @xmath56 ( and to a much lesser extent @xmath57 ) at the expense of @xmath58 , the @xmath59 s for larger @xmath0 being essentially identical .
this ties in with the intuition that a perfect matching is going to be easier to achieve when there is a larger proportion of vertices with an even number of neighbours .
the simulation indicates that local properties need not be identical to those of pure euclidean gravity , even though we are still dealing with a summation over ( in this case a restricted set of ) triangulations .
this observation is backed up by considering the dual viewpoint of equ.([action ] ) , where one regards the @xmath23 and @xmath25 terms in equ.([action ] ) as representing cubic vertices rather than triangles .
this recasts the model as an @xmath60 version of the @xmath61 loop gas on a random surface , whose critical behaviour has been investigated in some detail @xcite .
viewed in this way the model generates @xmath22 loops floating in a background of @xmath21 edges . in this model , as well as
the pure euclidean gravity phase observed on increasing @xmath16 with @xmath17 fixed , a dense critical phase is encountered on increasing @xmath17 with @xmath16 fixed . in this the entire lattice is densely filled with @xmath22 loops , which translates back to the lorenztian sector of perfectly matched triangulations in our original , non - dual , picture .
the @xmath62 appearing in the @xmath61 model can conveniently be parametrized as @xmath63 looking at the scaling of various operators and demanding consistency with the continuum kpz / ddk @xcite results then gives ( for spherical topology on the underlying euclidean lattice ) that @xmath64 so in our case @xmath60 gives @xmath65 .
we recover the ising model value of @xmath66 at the critical point and we find the ( euclidean ) pure gravity value of @xmath67 throughout the dense phase that we have identified as lorentzian .
although @xmath67 in the dense / lorenztian phase at least some of the local geometrical properties are not the same as the euclidean theory .
watermelon operators on the form @xmath68 measure the expectation values of @xmath69 @xmath22 lines pinned at two points .
their conformal dimensions @xmath70 are known to satisfy @xmath71 as pure gravity is retrieved at the critical point for @xmath72 and ising link gravity still has @xmath73 in the dense phase these scaling dimensions are not the same .
we have suggested that a model combining dynamical triangulations and ising link quantum gravity might capture some of the features of the discretized microgeometry of a 2d gravity theory admitting both lorentzian and euclidean signatures . with suitable restrictions on the allowed triangles
this model is precisely the ising model coupled to 2d gravity .
the purely lorentzian sector appears as an antiferromagnetic limit in the ising model , giving either a summation over perfectly matched triangulations or a mixed triangulation / quadrangulation depending on the choice of convention for the lorentzian triangles .
as the model is being used away from the ising critical point some properties are essentially those of discretized pure euclidean gravity based on triangulations ( ie @xmath74 ) , but non - universal features such as curvature distributions may be different , as witnessed by the simulation and the watermelon dimensions .
we have not really resolved the issue of `` factors of @xmath4 '' in this paper .
as noted , this does not play a role in microcanonical simulations but would have to be faced in extracting scaling information from the results .
however , it is reassuring that the models of equ.([actioniii ] ) and equ.([actionv ] ) representing the extreme lorentzian limit for both possible choices of convention _ can _ be defined for complex couplings , so the standard scaling limit can be approached through complex coupling values , which is precisely what is required in our context .
modulo this caveat , the results here appear to be rather reassuring for euclidean practitioners , as we are still dealing with a theory of triangulations , albeit decorated , in our toy lorentzian model .
perhaps the most important question we have left unanswered in the current work is whether the microscopic causality structure in our model , which looks very chaotic , becomes tamer at larger scales .
this would be a prerequisite for extracting a sensible continuum theory .
a numerical investigation of the role of signature changing would also be possible within the context of the model .
other ways of grafting a lorentzian stucture onto euclidean lattices also suggest themselves , such as using @xmath75 graphs and taking the edges emanating from a vertex as the lightcone @xcite . these might also be amenable to numerical investigation . in any event
the current work suggests that the ising antiferromagnet , or more generally the complex temperature ising model , coupled to 2d gravity might be worthy of further investigation .
wb is supported in part by fonds zur frderung der wissenschaftlichen forschung under contract p9522-phy .
dj is partially supported by ec hcm grant crrx - ct930343 .
wb and dj would also like to acknowledge a british council / arc exploratory travel grant and the ( excellent ! ) hospitality of harald markum at the technical university of vienna that made possible the work in this paper .
99 for a review see `` quantization of geometry '' by j. ambjrn , les houches summer school session lxii .
t. regge , nuovo cimento 19 ( 1961 ) 558 .
m. gross and h. hamber , nucl
b364 * ( 1991 ) 703 ; + c. holm and w. janke , phys .
b335 * ( 1994 ) 143 ; + nucl . phys .
* b * ( proc suppl . ) 42 ( 1995 ) 722,725 ; + `` measuring the string susceptibility in 2d simplicial quantum gravity using the regge approach '' , hep - lat/9511029 .
b. berg and w. beirl , nucl .
b452 * ( 1995 ) 415 . j. ambjrn and j. jurkiewicz , nucl .
b451 * ( 1995 ) 643 ; + j. ambjrn and j. jurkiewicz , phys . lett .
* b335 * ( 1995 ) 355 ; + s. catterall , j. kogut and r. renken , phys . lett .
* b328 * ( 1994 ) 277 ; + b. de bakker and j. smit , nucl . phys .
* b454 * ( 1995 ) 343 .
w. beirl , h. markum and j. riedler , phys .
d49 ( 1994 ) 5231 ; + b. berg , phys .
* 55 * ( 1985 ) 904 ; + phys . lett .
* b176 * ( 1986 ) 39 ; + h. hamber , nucl .
b400 * ( 1993 ) 347 .
w. beirl , h. markum and j. riedler , int . j. mod
c5 * ( 1994 ) 359 ; + w. beirl , p. homolka , b. krishnan , h. markum and j. riedler , nucl . phys . * b * ( proc suppl . ) 42 ( 1995 ) 710 ; + t. fleming , m. gross and r. renken , phys . rev .
* d50 * ( 1994 ) 7363 .
s. carroll , m. ortiz and w. taylor iv , `` a geometric approach to free variable loop equations in discretized theories of 2d gravity '' , hep - th/9510199 ; + s. carroll , m. ortiz and w. taylor iv , `` spin / disorder correlations and duality in the @xmath76 string '' , hep - th/9510208 .
d. johnston , phys .
b314 * ( 1993 ) 69 .
v. a. kazakov , phys . lett .
* a119 * ( 1986 ) 140 .
boulatov and v.a .
kazakov , phys . lett .
* b186 * ( 1987 ) 379 .
c. bachas , `` on triangles and squares '' , cern - th.5290/90 .
m. staudacher , nucl . phys .
* b336 * ( 1990 ) 349 .
f. david , nucl . phys .
* b348 * ( 1991 ) 507 . t. ponzano and t. regge , `` semiclassical limit of racah coefficients '' , in spectroscopic and group theoretical methods in physics , ed .
f. bloch ( north holland , amsterdam , 1968 ) .
j. barrett and t. foxon , class .
* 11 * ( 1994 ) 543 .
b. duplantier and i. kostov , phys .
lett . * 61 * ( 1988 ) 1433 ; + i. kostov , mod .
* a4 * ( 1989 ) 217 ; + m. gaudin and i. kostov , phys .
b220 * ( 1989 ) 200 ; + b. duplantier and i. kostov , nucl . phys . * b340 * ( 1990 ) 491 ; + i. kostov and m. staudacher , nucl . phys . *
b384 * ( 1992 ) 459 ; + b. eynard and j. zinn - justin , nucl . phys .
* b386 * ( 1992 ) 558 ; + b. eynard and c. kristjansen , nucl . phys .
* b455 * ( 1995 ) 577 ; + b. eynard and c. kristjansen , `` more on the exact solution of the @xmath77 model on a random lattice and an investigation of the case @xmath78 '' , spht-95/133 .
v. khizhnik , a. polyakov and a. zamalodchikov , mod .
a3 * ( 1988 ) 819 ; + j. distler and h. kawai , nucl . phys .
* b321 * ( 1989 ) 501 ; + f. david , mod .
* a3 * ( 1988 ) 1651 .
w. siegel , `` actions for qcd - like strings '' , hep - th/9601002 .
|
we suggest a generalization of the dynamical triangulation approach to quantum gravity with both timelike and spacelike edges , which can serve as a toy model for quantum gravity in the lorentz sector in two dimensions .
it is possible to consider the model in a purely lorentzian sector or to relax this constraint and allow local signature changing moves .
we show that , with suitable conventions , the model is equivalent to an ising model coupled to 2d euclidean quantum gravity and conduct a preliminary numerical simulation of the lorentz sector .
+ + + + 23.5 cm -1.75 cm 0 cm 0 cm
|
scaling relations between the global cluster parameters such as total mass , x - ray luminosity , and average temperature are important tools for studies of galaxy clusters and their cosmological applications .
simple self - similar theory predicts that these relations have a power - law form , @xmath9 , and @xmath10 ( e.g. , kaiser 1991 ) . deviations of the observed relations from these theoretical expectations have been used to assess the role of non - gravitational processes in the cluster formation ( see voit 2005 for a recent review ) . of prime interest
are relations between the total cluster mass and easily observed quantities , such as the average temperature of the intracluster medium ( icm ) or x - ray luminosity .
such relations allow one to estimate mass functions for large samples of poorly observed clusters and use them for cosmological constraints . to derive
the mass - observable relation requires accurate mass measurements in a representative sample of clusters . establishing an accurate normalization of the @xmath2 and @xmath11 relations
has been a focus of many recent observational and theoretical studies .
x - ray observations of dynamically relaxed clusters can be used to infer the total mass via the hydrostatic equilibrium equation ( e.g. sarazin 1988 ) . however , these are technically challenging observations that require accurate determinations of the cluster temperature profiles . mass measurements at radius @xmath12 are only as accurate as @xmath5 and @xmath13 at that @xmath12 . simplifying assumptions , such as
that the temperature is constant and gas density follows a @xmath14-model , lead to large biases in the mass determination @xcite .
therefore , it is essential to have direct measurements of the icm density and temperature distributions at large radii . long - exposure _ chandra _ and _ xmm - newton _ observations now provide such measurements for samples of low - redshift clusters @xcite , and the resulting @xmath2 relations are in very good agreement with the state - of - the - art cosmological simulations @xcite .
in addition to normalization of the local scaling relations , cosmological applications require experimental constraints on their evolution at high redshifts .
several studies over the past few years have used _
chandra _ and _ xmm - newton _ observations of small samples of distant clusters to study evolution of the @xmath2 relation .
however , these works have used an isothermal @xmath14-model to infer cluster masses , which can bias results on evolution of the scaling relations . in this _ paper _ , we present a spatially - resolved analysis of _ xmm - newton _ observations of a sample of 10 distant clusters spanning a range of temperatures and redshifts , @xmath15 kev , @xmath16 .
the large effective area of _ xmm - newton _ provides good statistical quality for most of the distant cluster data .
the finite angular resolution of the _ xmm - newton _ mirrors ( @xmath17 fwhm or 24 kpc at @xmath18 ) is the main technical challenge in this analysis .
however , this problem can be solved and deconvolved icm temperature and density profiles can be restored for our distant clusters .
we use these measurements to infer the total cluster mass , as well as the x - ray luminosity and temperature excluding the central cool regions , and thus study the evolution of cluster scaling relations at @xmath19 .
all distance - dependent quantities are derived assuming the @xmath20 , @xmath21 cosmology with the hubble constant @xmath22 km s@xmath23 mpc@xmath23 .
statistical uncertainties are quoted at @xmath24 cl .
p4.0cmclclcl cl 0016 + 16 & 20.5 & 0.86 & 25.3 & 0.94 & 22.2 & 0.90 + cl 0024 + 17 & 31.7 & 0.92 & 43.6 & 0.89 & 42.4 & 0.88 + ms0302.5 + 1717 & 05.7 & 1.20 & 09.9 & 1.00 & 09.9 & 0.96 + ms1054.4 - 0321 & 12.5 & 1.07 & 19.8 & 0.97 & 15.8 & 0.97 + rxj1120.1 + 4318 & 11.6 & 0.96 & 16.1 & 0.99 & 14.6 & 0.96 + rxj1334.3 + 5030 & 20.0 & 1.01 & 26.5 & 0.99 & 26.5 & 0.97 + wj1342.8 + 4028 & 17.8 & 1.07 & 27.6 & 1.02 & 25.3 & 1.01 + warpj0152.7 - 1357 & 33.4 & 0.98 & 44.7 & 0.94 & 46.8 & 0.90 + rxj0505.3 + 2849 & 12.2 & 0.99 & 21.5 & 0.93 & 20.8 & 0.90 + cl0939 + 472 & 20.7 & 0.91 & 30.5 & 0.85 & 30.5 & 0.82 our sample was selected from publicly available _ xmm - newton _ observations of distant clusters .
the final goal was to measure the cluster scaling relations at high redshift and we restricted our sample to objects with @xmath25 .
the selected clusters are listed in table [ live ] .
we used the data from all epic cameras ( mos1 , mos2 , and pn ) . for data reduction
, we used the _ xmm - newton _ science analysis system ( sas ) v6.0.0 and the calibration database with all updates available prior to november , 2004 .
all data were reprocessed with these latest versions of gain files .
the initial data screening was applied using the recommended sets of event patterns , 012 and 04 for the mos and pn cameras , respectively , and excluding all known bad ccd pixels .
our science goals required accurate background subtraction .
the epic background is highly variable and only its quiescent component can be accurately modeled . therefore , our next step was to detect and exclude the periods of flaring background .
we extracted the light curves for each camera in the 215 kev band using the data from the entire field of view excluding detectable sources .
the light curves were binned to 200 s time resolution and flares were detected as @xmath26 deviations from the mean .
experiments with different choices of energy bands and flare detection thresholds have shown that our choice was close to optimal .
for example , more flares are typically detected in the 215 kev band than in the frequently used @xmath27 kev band because some flares are soft . in some observations ,
almost the entire exposure was affected by flares , and these observations had to be discarded .
the clean exposure times for each camera are listed in table [ live ] . to account for strong _ xmm _ mirror vignetting , we used an approach proposed by @xcite .
each photon was assigned a weight proportional to inverse vignetting and these weights were then used in computing images and spectra .
this was done using the sas tool _
evigweight_. background modeling in our analysis was implemented following the double - subtraction method of .
the first step of this method is to subtract the particle - induced background component .
this component can be estimated from a set of _ xmm _ observations with the filter wheel closed ( so called `` closed data '' ) .
we compiled the closed dataset from public observations available in the _ xmm _ data archive ; these data were reduced following steps indetical to those of the science observations .
the closed background was adjusted to the cluster observations using the observed flux in the 1015 kev band outside the field of view .
the scaling factors are listed in table [ live ] .
the second step is to determine the cosmic x - ray background ( cxb ) component .
its spatial distribution should be flat because vignetting correction is already applied .
therefore , the cxb spectrum can be measured in the source - free regions of the field of view , typically at @xmath28 from the cluster center and then directly subtracted .
finally , we computed corrections for photons registered during the ccd readouts ( `` out - of - time events '' ) , using the sas tools _ epchain _ and _ emchain _ with the _ withoutoftime = yes _ flag .
cccccccccc cl 0016 + 16 & 0.54 & @xmath29&@xmath30 & 176.4/128 & & @xmath31&@xmath32&@xmath33 & 158.1/127 + cl 0024 + 17 & 0.39 & @xmath34&@xmath35 & 98.7/65 & & @xmath36&@xmath37&@xmath38 & 78.3/64 + ms0302.5 + 1717 & 0.42 & @xmath39&@xmath40 & 74.9/83 & & @xmath41&@xmath42 & @xmath43 & 69.0/82 + ms1054.4 - 0321 & 0.82 & & & & & & & & + rxj1120.1 + 4318 & 0.60 & @xmath44&@xmath45 & 77.6/83 & & @xmath46&@xmath47&@xmath48 & 77.4/82 + rxj1334.3 + 5030 & 0.62 & @xmath49&@xmath50 & 94.5/77 & & @xmath51&@xmath52&@xmath53 & 87.2/76 + wj1342.8 + 4028 & 0.70 & @xmath54&@xmath55 & 40.7/56 & & @xmath56&@xmath57 & @xmath58 & 35.1/55 + warpj0152.7 - 1357 & 0.83 & & & & & & & & + rxj0505.3 + 2849 & 0.51 & @xmath59&@xmath60 & 103.1/96 & & @xmath61&@xmath62 & @xmath63 & 101.6/95 + cl0939 + 472 & 0.41 & & & & & & & & p4cmcccccrccc cl 0016 + 16&@xmath64 & @xmath65 & 1.0 & @xmath66 & @xmath67 & @xmath68 & @xmath69 & @xmath700 + cl 0024 + 17&@xmath71 & @xmath72 & 0.5 & @xmath73 & @xmath74 & @xmath75 & @xmath76 & @xmath77 + ms0302.5 + 1717&@xmath78 & @xmath79 & 0.5 & @xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath84 + ms1054.4 - 0321&@xmath85 & @xmath86 & 0.7 & & & & & @xmath870 + rxj1120.1 + 4318&@xmath88 & @xmath89 & 1.0 & @xmath90 & @xmath91 & @xmath92 & @xmath93 & @xmath940 + rxj1334.3 + 5030&@xmath95 & @xmath96 & 0.8 & @xmath97 & @xmath98 & @xmath99 & @xmath100 & @xmath101 + wj1342.8 + 4028&@xmath102 & @xmath103 & 0.6 & @xmath104 & @xmath105 & @xmath106 & @xmath107 & @xmath108 + warpj0152.7 - 1357&@xmath109 & @xmath110 & 0.8 & & & & & @xmath1110 + rxj0505.3 + 2849&@xmath112 & @xmath113 & 1.0 & @xmath114 & @xmath115 & @xmath116 & @xmath117 & @xmath118 + cl0939 + 472&@xmath119 & @xmath120 & 1.0 & & & & & @xmath121 we excluded all detectable point sources from the data in our spectral and spatial analysis .
the sources were detected separately in the `` optimal '' 0.3 - 3 kev , `` soft '' 0.3 - 0.8 kev , and `` hard '' 2.0 - 6.0 kev energy bands . detected
point sources were masked with circles of 80% psf power radii .
spatial analysis of the cluster emission was performed in the @xmath122 kev energy band .
the images for each camera were corrected for vignetting and out - of - time events , with the particle background component subtracted as described in 2 , and analyzed independently .
we extracted the azimuthally averaged surface brightness profiles centered on the x - ray surface brightness peak , even for clusters with irregular morphology ( see below ) , excluding the ccd gaps and circles around the point sources .
the obtained profiles were used to derive the parameters of the spatial distribution of the icm , the cluster fluxes , and the background levels .
the cluster surface brightness profiles are often modeled with the so - called @xmath14-model , @xmath123 or @xmath124 @xcite
. however , this model poorly describes clusters with sharply peaked surface bigness profiles related to the radiative cooling of the icm in the cluster centers .
the simple modification of the @xmath14-model @xcite @xmath125 allows adequately describing these cooling regions in the low redshift clusters . for @xmath126 ,
this @xmath127-@xmath14 model is identical to the usual @xmath14-model . the model for the observed surface brightness profiles
can be obtained by numerical integration of eq.([eq : alpha - beta ] ) along the line of sight and convolution of the result with the xmm psf . to represent the uniform sky x - ray background
, we added a constant component to the model and treated it as a free parameter .
the values of @xmath127 , @xmath14 , and @xmath128 were derived from the joint fit to the observed profiles in the mos1,mos2 , and pn cameras , with the overall normalizations and background levels fitted independently for each camera . for comparison
, we also fitted the standard @xmath14-model by setting @xmath126 .
the obtained parameters for the @xmath127-@xmath14 model and the standard @xmath14 model fits are summarized in table [ main1 ] .
the best - fit icm model was used to derive the total flux and the total mass from the hydrostatic equilibrium equation ( see 6 below ) . strictly speaking , eq.([eq : alpha - beta ] )
can not be applied to the icm distribution in clusters with irregular morphology . in these cases ,
the model fit was used only to measure the image background , and we do not list the values of @xmath127 , @xmath14 , and @xmath128 in table [ main1 ] .
, and solid line shows the best - fit relation , @xmath129 . ] the cluster spectra were extracted in annuli and circles within radii listed in table [ main2 ] .
the response matrices and effective area files were generated by the standard sas tasks . because the data were previously vignetting corrected , the effective area files were created for the on - axis position using the routine _
the response matrices were generated in the spectrum extraction region via _
rmfgen_. the spectra were binned , so that there were at least 20 counts per bin .
the observed spectra were fitted in the @xmath130 kev energy band .
we checked that there were no prominent residuals near the instrumental lines and that exclusion of the instrumental line regions did not affect final results .
we used the mewe - kaastra - liedahl plasma emission model , with free metallicity and galactic absorption fixed at the value derived from the radio surveys .
the epic - pn , mos1 , and mos2 spectra for each extraction region were fitted jointly , with the spectral parameters tied and relative normalizations free .
the resulting spectral parameters are listed in table [ main2 ] .
our single - temperature fits to wide - beam spectra are in good agreement with the published results ( e.g. , , @xcite ) .
to compare our results with the low - redshift scaling relations we need to correct for any systematic differences between _ asca _ , _ chandra _ , and _ xmm - newton _ measurements .
@xcite verified that there is no systematic difference between the _ chandra _ and the _ asca _ temperature measurements .
therefore we need to cross - calibrate only _ xmm - newton _ and _ chandra_. several of our clusters were in the @xcite sample .
the comparison of _ xmm - newton _ and _ chandra _ temperatures for these clusters is shown in fig .
[ fig : chandra_xmm ] .
there is a good overall agreement , although the _ chandra _ temperature for the hottest cluster , cl0016 + 16 , is marginally higher than our value .
the linear fit gives @xmath131 .
a similar systematic difference is found in the comparison of the _ chandra _ and _ xmm _ temperature profiles of nearby clusters @xcite .
we applied this correction factor to all temperature values because we use the _ asca _ @xmath3 relation as low - redshift reference . for @xmath2 relation ,
this correction is unimportant ( see 7 ) .
comparison of the cluster luminosities shows a good agreement , within @xmath132% , between _ chandra _ and _
the statistical quality of _ xmm - newton _ data allows to reconstruct the temperature profiles from all our clusters .
the main complication is that the xmm psf size is non negligible compared with the angular size of distant clusters .
for example , @xmath133 of the flux from the central 100 kpc region ( @xmath134 at @xmath135=0.6 ) is scattered to larger radii ( @xmath136 of the flux stays within 260 kpc regions ) .
the temperature in these regions is often lower than the cluster average because of radiative cooling , and hence this scattered flux can significantly bias the temperature measurements at large radii .
we corrected for the xmm psf using an approach used for @xmath137 data analysis by . using the best - fit @xmath127-@xmath14 models of the cluster brightness and the xmm psf calibration
, we calculated the redistribution matrix , @xmath138 , of each temperature to each annulus that represents relative contribution of emission from annulus @xmath139 to the observed flux in annulus @xmath140 .
the model spectrum , @xmath141 , is then given by @xmath142 where @xmath143 is the temperature in annulus @xmath139 and @xmath144 is the _ mekal _ spectrum for this temperature .
fitting this model to the observed spectra in all annuli simultaneously and treating all @xmath143 as free parameters gives the deconvolved temperature profile .
the raw and deconvolved temperature profiles are shown in fig.[fig : cl0016][fig : ms0302 ] .
the deconvolved temperature profiles are within 1 @xmath145 of the raw measurements in all cases .
however the psf correction is systematic and results in stronger temperature gradients . thus neglecting this effect
can slightly biases the mass measurements . in many cases ,
we observe a decrease of temperature at large radii , which is qualitatively consistent with the results for low - redshift clusters @xcite . the only cluster that appears approximately isothermal is rxj1120.1 ; our results for this cluster are fully consistent with the analysis by .
we fitted the observed temperature profiles by the function : @xmath146 a similar model describes the temperature profiles for low redshift clusters @xcite . for local clusters ,
@xmath147 scales with the average temperature as @xmath148 mpc ( see fig . 16 in @xcite ) .
the statistical uncertainties in our temperature profiles are insufficient to fit @xmath149 , @xmath147 , and the outer slope @xmath127 .
therefore we fixed @xmath147 at the value suggested by the low - redshift correlation , with an additional scaling , @xmath150 , to account for the redshift dependence of the virial radius for a fixed temperature @xcite . to fit the observed profiles , we projected the 3-d model along the line of sight using the emission measure profile from the best fit @xmath127-@xmath14 model
. projection was based on a weighting method that correctly predicts the best - fit spectral @xmath5 for a mixture of different temperature components @xcite . for those clusters with the central temperature decrements we excluded the innermost bin from the fit .
this procedure is correct because our temperature profiles were corrected for the xmm psf .
the obtained best - fit models are shown along with the deconvolved temperature profiles in fig.[fig : cl0016][fig : ms0302 ] .
assuming hydrostatic equilibrium for the icm , we can use the best fit temperature and density profiles to derive the total cluster masses : @xmath151 the mass was calculated within the radius that corresponds to the mean overdensity @xmath152 relative to the critical density at the cluster redshift .
the uncertainties on the masses were calculated from monte - carlo simulations .
the mass uncertainties are dominated by statistical uncertainties of the temperature profiles , and we neglected all other sources of error .
we used the best - fit temperature profile as a template , applied gaussian scatter with the _ rms _ equal to the statistical uncertainties , and fitted the simulated profile and derived the mass .
the mass uncertainty was estimated as an _
scatter in 1000 simulations . for each cluster we calculated the following temperature averages : @xmath153
@xmath154 weighted with @xmath155 .
@xmath153 is needed to self - consistently compare the high - z @xmath3 relation of our sample with the low - redshift result of @xcite .
@xmath156 is obtained by integrating a combination of @xmath154 and @xmath155 as described in @xcite .
@xmath156 is needed for comparison with the low - redshift @xmath2 relation from @xcite .
@xmath156 and @xmath153 were averaged in the radial range 70 kpc @xmath157 r @xmath157 @xmath1 .
the uncertainties on @xmath153 and @xmath156 were also calculated from monte - carlo simulations .
the obtained values of @xmath153 , @xmath156 , @xmath158 , and the corresponding overdensity radius @xmath1 are listed in table [ main2 ] .
the @xmath1 radii are also shown by vertical dotted lines in fig.[fig : cl0016][fig : ms0302 ] . to test how background subtraction can affect our mass result , we checked two sources of uncertainties . first , we varied normalization of the article - induced component by @xmath159 .
normalization of cxb component was varied by @xmath160 ( quoted xmm vignetting uncertainties ) .
we refitted all spectra with renormalized background and repeated the mass analysis .
the resulting variations in @xmath158 were well within our statistical uncertainties .
self - similar theory ( e.g , @xcite ) predicts that the relation between cluster mass and temperature is a power law that evolves as @xmath161 , where @xmath162 for the adopted cosmology .
our measurements for distant clusters can be used to test these predictions .
the reference low - redshift @xmath2 relation was adopted from a _
sample of @xcite , which is also close to the _ xmm _ results of @xcite ) .
sample contains 13 nearby clusters with exposures sufficient to measure temperature profiles to @xmath163 .
the temperatures of these clusters , 210 kev , match well the temperature range in our distant sample .
@xcite derived the mass - temperature relation , @xmath164 , where @xmath165 .
the derived slope is consistent with the self - similar expectation , @xmath166 .
figure [ fig : mt]a shows the quantity @xmath167 as a function of @xmath135 for our clusters .
this quantity represents the normalization of the @xmath2 relation as constrained by each cluster . for non - evolving @xmath2 relation , all values of @xmath168 would be consistent with 1
however , the observed values of @xmath168 clearly indicate evolution . to quantify the observed evolution , we fitted the data from fig.[fig : mt]a with a power law of @xmath169 , @xmath170 .
the best - fit index is @xmath7 , where the error bar includes the uncertainties in low-@xmath135 normalization of the @xmath2 relation and our high-@xmath135 mass measurements .
the derived rate of evolution is consistent with the theoretically expected one , @xmath8 .
therefore we can assume that the normalization of the @xmath2 relation evolves exactly as @xmath171 .
now we can derive slope and normalization of the @xmath2 relation defined by our distant clusters .
for this , we corrected the mass measurements for evolution by multiplying them by @xmath169 .
figure [ fig : mt]b shows the corrected cluster masses as a function of temperature .
the power - law fit , @xmath172 , to distant clusters only gives @xmath173 .
the obtained value of the slope,@xmath174 is statistically consistent with the value , @xmath175 , predicted by the self - similar theory .
finally , we note that the results on @xmath2 relation are insensitive to the absolute calibration of the _ xmm _ effective area .
the main effect of small calibration errors is to bias the derived temperatures by a constant factor .
it follows from eq.([eq : def1_m ] ) that the normalization of the @xmath2 relation as given by individual clusters , @xmath176 , is @xmath177 where the last two terms are evaluated at the overdensity radius , @xmath178 .
small calibration errors lead to small changes in estimated @xmath178 ( @xmath179 ) but the normalization of @xmath2 relation is not affected because both density and temperature are nearly power - law functions of @xmath12 in the interesting range of radii .
our results can also be used to test the evolution in the cluster x - ray luminosity versus temperature relation .
we performed at spatially - resolved analysis of the temperature and surface brightness profiles , including the psf deconvolution .
therefore , we can directly exclude the contribution of the central cooling regions to both temperature and luminosity .
this significantly reduces the scatter in the @xmath3 relation @xcite and thus makes any evolution more prominent .
the x - ray bolometric luminosities were calculated using the mewe - kaastra - liedahl plasma emission model .
we used the measured best - fit temperatures for three objects with irregular x - ray morphology ( ms1054.4 - 0321 , warpj0152.7 - 1357,and cl0939 + 472 ) and the obtained emission - weighted temperatures for all other clusters .
all temperatures were renormalized by @xmath180 to account for _ chandra _ versus _ xmm - newton _ cross - calibration _ asca _
measurements by @xcite . ] .
for the three objects with irregular x - ray morphology , we use the observed @xmath122 kev counts rates within @xmath181 mpc as the normalizing fluxes . for all other clusters , the normalizing fluxes were calculated by subtracting from the observed @xmath122 kev counts rates within @xmath181 mpc the fluxes calculated from the best - fit @xmath127-@xmath14 models within @xmath182 kpc and multiplying the result by a factor of 1.06 to account for the flux within @xmath182 kpc in a typical @xmath14-model cluster ( see @xcite ) .
the resulting luminosities are listed in table [ main2 ] .
figure [ fig : lt2 ] shows the resulting @xmath3 relation .
the solid line shows the relation for low - redshift clusters @xcite .
clearly , distant clusters have higher luminosities for the given temperature . parametrizing the evolution as @xmath183
, we obtain @xmath184 for @xmath127 fixed at the value for low - redshift sample , @xmath185 ( fig .
[ fig : lt2]b ) . the observed rate of evolution is consistent with the results of @xcite obtained from an independent _ chandra _ sample .
we presented a spatially - resolved analysis of the icm density and temperature radial profiles in a sample of 10 @xmath19 clusters observed with _
xmm - newton_. the main results can be summarized as follows .
we use the spatially - resolved measurements to study evolution of the @xmath3 relation excluding the cluster cooling cores .
the observed evolution of the bolometric luminosity for a fixed temperature is @xmath186 in the redshift range @xmath0 , fully consistent with earlier _ chandra _ results @xcite and other _ xmm - newton _ studies , but in apparent conflict with conclusions of . most of our clusters have non - constant temperature profiles .
there is a temperature decline at large radii in most objects , which is qualitatively consistent with the temperature profiles in low - redshift clusters @xcite . using the derived temperature and density profiles we determine the total cluster mass within the radius @xmath1 without using the usual assumption of an isothermal @xmath14-model
this allows a direct comparison of the high - redshift @xmath2 relation with the recent high - quality _ chandra _ and _ xmm - newton _
measurements for low - redshift clusters ( @xcite ; @xcite ) .
the observed @xmath2 relation for distant clusters is @xmath173 .
the derived slope , @xmath174 , is statistically consistent with both @xmath187 measured by @xcite for low-@xmath135 clusters and @xmath175 predicted by the self - similar theory .
we would like to thank m. markevitch and b. maughan for useful discussions .
this work was supported by nasa grant nag5 - 9217 and contract nas8 - 39073 .
o. k. thanks sao for hospitality during the course of this research .
akritas , m.g . , & bershady , m.a .
, 1996 , apj,470,706 arnaud , m. , neumann , d.m . ,
aghanim , n. , et al .
, 2001,a&a , 365 , l80 arnaud , m. , majerowicz , s. , lumb , d. , et al . , 2002 , a&a , 390 , 27a arnaud , m. , pointecouteau , e. , pratt , g. w. 2005 , in press , ( astro - ph/0502210 ) borgani , s. , et al . 2004 , , 348 , 1078 bryan , g. l. , & norman , m. l. 1998 , , 495 , 80 cavaliere , a.,fusco - femiano , r. , 1976 , a@xmath188a , 49 , 137 de grandi , s. , & molendi , s. 2002 , , 567 , 163 dickey , j. m. & lockman , f. j. 1990 , , 28 , 215 ettori , s. , tozzi , p. , borgani , s. , & rosati , p. 2004
, , 417 , 13 evrard , a. e. , metzler , c. a. , & navarro , j. f. 1996 , , 469 , 494 finoguenov , a. , reiprich , t. h. , bhringer , h. 2001 , , 368 , 749 kaiser , n. 1991 , , 383 , 104 kirsch m. , 07/28/2004 , xmm - soc - cal - tn-0018 lumb , d. h. , et al .
2004 , , 420 , 853 markevitch , m. , & vikhlinin , a. 1997 , , 491 , 467 markevitch , m. , 1998 , , 504,27 markevitch , m. , forman , w. r. , sarazin , c. l. , & vikhlinin , a. 1998 , , 503 , 77 mathiesen , b. f. , & evrard , a. e. 2001 , , 546 , 100 maughan , b. j. , jones , l. r. , ebeling , h. , & scharf , c. 2005 , apj , submitted .
mazzotta , p. , rasia , e. , moscardini , l. , & tormen , g. 2004 , , 354 , 10 mewe , r. , gronenschild , e. h. b. m. , & van den oord , g. h. j. 1985 , , 62 , 197 nevalainen , j. , markevitch , m. , & forman , w. 2000 , , 532 , 694 pointecouteau , e. , arnaud , m. , kaastra , j. , & de plaa , j. 2004 , , 423 , 33 pratt , g.w . , arnaud , m. , 2002 , a&a , 394 , 375 reiprich , t. h. , bohringer , h. 2002 , , 567 , 716 sanderson , a. j. r. , & ponman , t. j. 2003 , , 345 , 1241 sarazin , c. l. 1988 , x - ray emission from clusters of galaxies ( cambridge : cambridge university press ) vikhlinin , a. , vanspeybroeck , l. , markevich , m. et al . , 2002 ,
apj 578 , l107 vikhlinin , a. , 2005 , apj , submitted , ( astro - ph/0504098 ) vikhlinin , a. , markevitch , m. , murray , s. s. , jones , c. , forman , w. , & van speybroeck , l. 2005a , , 628 , 655 vikhlinin , a. , kravtsov a. , forman , w. , jones , c. , markevitch , m. , murray , s. s. , & van speybroeck 2005b , apj submitted , ( astro - ph/0507092 ) voit , m. 2005 , rev .
, in press ( astro - ph/0410173 ) worrall , d. m. , & birkinshaw , m. 2003 , , 340 , 1261
|
we present a spatially - resolved analysis of the temperature and gas density profiles of galaxy clusters at @xmath0 observed with _ xmm - newton_. these data are used to derive the total cluster mass within the radius @xmath1 without assuming isothermality , and also to measure the average temperature and total x - ray luminosity excluding the cooling cores .
we derive the high - redshift @xmath2 and @xmath3 relations and compare them with the local measurements .
the high - redshift @xmath3 relation has low scatter and evolves as @xmath4 for a fixed @xmath5 , in agreement with several previous _
chandra _ and _ xmm - newton _ studies ( vikhlinin et al . , lumb et al . , and maughan et al . ) .
the observed evolution of the @xmath2 relation follows @xmath6 , where we measure @xmath7 .
this is in agreement with predictions of the self - similar theory , @xmath8 .
|
in this report i will be concerned mainly with the frequentist ( classical ) theory of statistical inference , but i think that it is interesting and useful that i express my opinion on the war between frequentists and bayesians . to the question
`` are you frequentist or bayesian '' ?
i answer `` i like statistics . ''
i think that if one likes statistics , one can appreciate the beauty of both frequentist and bayesian theories and the subtleties involved in their formulation and application .
i think that both approaches are valid from a statistical as well as physical point of view .
their difference arises from different definitions of probability and their results answer different statistical questions .
one can like more one of the two theories , but i think that it is unreasonable to claim that only one of them is correct , as some partisans of that theory claim .
these partisans often produce examples in which the other approach is shown to yield misleading or paradoxical results .
i think that each theory should be appreciated and used in its limited range of validity , in order to answer the appropriate questions .
finding some example in which one approach fails does not disprove its correctness in many other cases that lie in its range of validity .
my impression is that the bayesian theory ( see , for example , @xcite ) has a wider range of validity because it can be applied to cases in which the experiment can be done only once or a few times ( for example , our thoughts in everyday decisions and judgments seem to follow an approximate bayesian method ) . in these cases
the bayesian definition of probability as _ degree of believe _ seems to me the only one that makes sense and is able to provide meaningful results .
let me remind that since galileo an accepted basis of scientific research is the _ repeatability of experiments_. this assumption justifies the frequentist definition of probability as ratio of the number of positive cases and total number of trials in a large ensemble .
the concept of _ coverage _ follows immediately : a @xmath0 _ confidence interval _ for a physical quantity @xmath1 is an interval that contains ( covers ) the unknown true value of that quantity with a frequentist probability @xmath2 . in other words ,
a @xmath0 confidence interval for @xmath1 belongs to a set of confidence intervals that can be obtained with a large ensemble of experiments , @xmath0 of which contain the true value of @xmath1 .
i think that in order to fully appreciate the meaning and usefulness of frequentist confidence intervals obtained with neyman s method @xcite , it is important to understand that the experiments in the ensemble do not need to be identical , as often stated , or even similar , but can be real , different experiments @xcite .
one can understand this property in a simple way @xcite by considering , for example , two different experiments that measure the same physical quantity @xmath1 .
the @xmath0 classical confidence interval obtained from the results of each experiment belongs by construction to a set of confidence intervals which can be obtained with an ensemble of identical experiments and contain the true value of @xmath1 with probability @xmath2 .
it is clear that the sum of these two sets of confidence intervals , containing the two confidence intervals obtained in the two different experiments , is still a set of confidence intervals that contain the true value of @xmath1 with probability @xmath2 .
moreover , for the same reasons it is clear that _ the results of different experiments can also be analyzed with different frequentist methods _
@xcite , _ i.e. _ methods with correct coverage but different prescriptions for the construction of the confidence belt .
this for me is amazing and beautiful : _ whatever method you choose you get a result that can be compared meaningfully with the results obtained by different experiments using different methods _ !
it is important to realize , however , that the choice of the frequentist method must be done independently of the knowledge of the data ( before looking at the data ) , otherwise the property of coverage is lost , as in the `` flip - flop '' example in ref .
@xcite .
this property allow us to solve an apparent paradox that follows from the recent proliferation of proposed frequentist methods .
this proliferation seems to introduce a large degree of subjectivity in the frequentist approach , supposed to be objective , due to the need to choose one specific prescription for the construction of the confidence belt , among several available with similar properties . from the property above , we see that whatever frequentist method one chooses , if implemented correctly , the resulting confidence interval can be compared statistically with the confidence intervals of other experiments obtained with other frequentist methods .
therefore , _ the subjective choice of a specific frequentist method does not have any effect from a statistical point of view _ !
then you should ask me : why are you proposing a specific frequentist method ?
the answer lies in _ physics _
, not statistics .
it is well known that the statistical analysis of the same data with different frequentist methods produce different confidence intervals .
this difference is sometimes crucial for the physical interpretation of the result of the experiment ( see , for example , @xcite ) .
hence , the physical significance of the confidence intervals obtained with different frequentist methods is sometimes crucially different .
in other words , _ the frequentist method suffers from a degree of subjectivity from a physical , not statistical , point of view_.
the possibility to apply successfully frequentist statistics to problematic cases in frontier research has received a fundamental contribution with the proposal of the unified approach by feldman and cousins @xcite .
the unified approach consists in a clever prescription for the construction of a classical confidence belt which unifies the treatment of upper confidence limits for null results and two - sided confidence intervals for non - null results " . in the following
i will consider the case of a poisson process with signal @xmath1 and known background @xmath3 .
the probability to observe @xmath4 events is @xmath5 the unified approach is based on the construction of the acceptance intervals @xmath6 $ ] ordering the @xmath4 s through their rank given by the relative magnitude of the likelihood ratio @xmath7 where @xmath8 is the maximum likelihood estimate of @xmath1 , @xmath9 \ , .
\label{mu - best}\ ] ] as a result of this construction the confidence intervals are two - sided ( _ i.e. _ @xmath10 $ ] with @xmath11 ) for @xmath12 , whereas for @xmath13 they are upper limits ( _ i.e. _ @xmath14 ) . the fact that the confidence intervals are two - sided for @xmath12 can be understood by considering @xmath15 , that gives @xmath16 . in this case
the likelihood ratio ( [ lr ] ) is given by @xmath17 - ( \mu+b ) \right\ } \stackrel{n\to\infty}{\longrightarrow } 0 \ , .
\label{lr-2}\ ] ] this implies that the rank of high values of @xmath4 is very low and they are excluded form the confidence belt .
therefore , the acceptance intervals @xmath6 $ ] are always bounded , _
i.e. _ @xmath18 is finite , and the confidence intervals are two - sided for @xmath12 , as illustrated in fig .
[ ualim ] , where the solid lines show the borders of the confidence belt for a background @xmath19 and a confidence level @xmath20 .
the fact that the confidence intervals are upper limits for @xmath13 can be understood by considering @xmath21 , for which we have @xmath22 and the likelihood ratio that determines the ordering of the @xmath4 s in the acceptance intervals is given by @xmath23 considering now the acceptance interval for @xmath24 , we have @xmath25 .
therefore , all @xmath21 for @xmath24 have highest rank and are guaranteed to lie in the confidence belt .
this is illustrated in fig .
[ ualim ] , where the thick solid segment shows the @xmath21 part of the acceptance interval for @xmath26 , that must lie in the confidence belt . since @xmath1 is a continuous parameter , also for small values of @xmath1
the @xmath21 have rank close to the highest one and lie in the confidence belt .
indeed , for @xmath27 , the likelihood ratio ( [ lr ] ) increases for @xmath4 going form zero to the largest integer smaller or equal to @xmath3 and decreases for larger values of @xmath4 .
hence , the largest integer @xmath28 such that @xmath29 has highest rank . if @xmath1 is sufficiently small all @xmath21 have rank close to maximum and are included in the confidence belt if the confidence level is large enough , @xmath30 .
for example , @xmath31 for @xmath32 .
therefore , the left edge of the confidence belt must change its slope for @xmath13 and intercept the @xmath1-axis at a positive value of @xmath1 , as illustrated in fig .
[ ualim ] . the value of @xmath1 at which the left edge of the confidence belt intercepts the @xmath1-axis , that corresponds to @xmath33 , depends on the value of the background @xmath3 and on the value of the confidence level @xmath2 .
however is devoid of any practical interest .
] , for small values of @xmath2 the unified approach gives zero - width confidence intervals for @xmath34 , as illustrated in fig .
[ uabelt ] , where i have chosen @xmath19 and @xmath35 .
one can see that the segment @xmath21 is enclosed in the confidence belt for @xmath26 , but for any value of @xmath27 the sum of the probabilities of the @xmath4 s close to @xmath36 is enough to reach the confidence level and low values of @xmath4 are not included in the confidence belt .
hence , in this case the unified approach gives zero - width confidence intervals for @xmath37 .
the unification of the treatments of upper confidence limits for null results and two - sided confidence intervals for non - null results obtained with the unified approach is wonderful , but it has been noticed that the upper limits obtained with the unified approach for @xmath38 are too stringent ( meaningless ) from a physical point of view @xcite . in other words , although these limits are statistically correct from a frequentist point of view , they can not be taken as reliable upper bounds to be used in physical applications .
this problem is illustrated in fig .
[ higback]a , where i plotted the 90% cl upper limit @xmath39 as a function of @xmath3 for @xmath40 . the solid part of each line shows where @xmath41 .
one can see that for a given @xmath4 , @xmath39 decreases rather steeply when @xmath3 is increased , until a minimum value close to one is reached .
the curves have jumps because @xmath4 is an integer and generally the desired confidence level can not be obtained exactly , but with some unavoidable overcoverage .
let me emphasize that the problem of obtaining too stringent upper limits for @xmath38 is very serious for a scientist that wants to obtain reliable information from experiment and use this information for other purposes ( as input for a theory or another experiment ) . in the past , researchers bearing the same physical point of view refrained to report empty confidence intervals or very stringent upper limits when @xmath38 was measured .
these confidence intervals are correct from a statistical point of view , but useless from a physical point of view .
furthermore , the same reasoning lead to prefer the unified approach to central confidence intervals or upper limits , because the non - empty confidence interval obtained when @xmath38 is measured is certainly more significant , from a physical point of view , than an empty one , although they are statistically equivalent , as shown in section [ significance ] .
in the unified approach @xmath8 is positive and equal to zero for @xmath21 .
if @xmath8 is forced to be always bigger than zero , the @xmath4 s smaller than @xmath3 have rank higher than in the unified approach . as a consequence , the decrease of the upper limit @xmath39 as @xmath3 increases is weakened .
this is illustrated by a _
`` brutally modified unified approach '' _ ( bmua ) in which we take @xmath42 \ , , \label{mu - best - bmua}\ ] ] where @xmath43 is a positive real number . in fig .
[ aubelt ] i plotted the confidence belts for @xmath44 ( solid lines ) , that corresponds to the unified approach , @xmath45 ( dashed lines ) and @xmath46 ( dotted lines ) , for @xmath47 . one can see that in the bmua the upper limits of the confidence intervals are considerably higher than in the unified approach . the behavior of @xmath39 as a function of @xmath3 for @xmath48 is shown in fig .
[ auhigback ] , from which it is clear that the decrease of @xmath39 when @xmath3 increases is much weaker in the bmua ( dashed and dotted lines ) than in the unified approach ( solid line ) and it is almost absent for @xmath49 .
let me emphasize that 1 .
the bmua is a statistically correct frequentist method and coverage is satisfied .
2 . in the bmua one obtains upper limits for @xmath13 and central confidence intervals for @xmath12 , as in the unified approach we have @xmath50 and the likelihood ratio ( [ lr ] ) becomes @xmath51 for @xmath52 , we have @xmath53 and @xmath54 decreases with increasing @xmath4 .
let us consider now @xmath55 , for which @xmath16 and the likelihood ratio ( [ lr ] ) is given by the expression in eq .
( [ lr-2 ] ) .
this expression has a maximum for @xmath4 equal to one of the two integers closest to @xmath36 . for @xmath52 ,
this integer is the first one in the considered range ( @xmath55 ) .
therefore , for sufficiently low values of @xmath1 , @xmath52 , the likelihood ratio ( [ lr ] ) decreases monotonically as @xmath4 increases . in this case ,
low values of @xmath4 have highest ranks and are guaranteed to lie in the confidence belt and the left edge of the confidence belt must change its slope for @xmath56 and intercept the @xmath1-axis at a positive value of @xmath1 , as illustrated in fig .
[ aubelt ] . ] .
the bmua method is not general ( although it can be extended in an obvious way at least to the case of a gaussian variable with a physical boundary ) .
_ i am not proposing the bmua _ !
( but those that think that the upper limit for @xmath48 should not depend on @xmath3 may consider the possibility of using the bmua with @xmath46 instead of resorting to more complicated methods that may even jeopardize the property of coverage . ) as shown in fig .
[ aubelt ] , the right edge of the confidence belt in the bmua is not very different from the one in the unified approach .
this is due to the fact that adding small values of @xmath4 with low probability to the acceptance intervals has little effect .
moreover , it is clear that the acceptance interval for @xmath26 is equal for all frequentist methods with correct coverage that unify the treatment of upper confidence limits and two - sided confidence intervals .
an elegant , natural and general way to obtain automatically @xmath57 is given by the _
bayesian ordering _
method @xcite , in which @xmath8 is replaced by the bayesian expectation value for @xmath1 , @xmath58 . choosing a natural flat prior , the bayesian expectation value for @xmath1 in a poisson process with background is given by @xmath59 the obvious inequality @xmath60 implies that @xmath61 .
therefore , the reference value for @xmath1 in the likelihood ratio @xmath62 that determines the construction of the acceptance intervals as in the unified approach , is bigger or equal than one . as a consequence ,
the decrease of the upper confidence limit @xmath39 for a given @xmath4 when the expected background @xmath3 increases is significantly weaker than in the unified approach , as illustrated in fig .
[ higback]b .
figure [ higback]c shows @xmath39 as a function of @xmath3 in the bayesian theory with a flat prior and shortest credibility intervals is @xmath63 and the probability ( degree of believe ) that the true value of @xmath1 lies in the range @xmath64 $ ] is given by @xmath65|n , b ) = \left ( e^{-\mu_1 } \sum_{k=0}^{n } \frac{(b+\mu_1)^k}{k ! } - e^{-\mu_2 } \sum_{k=0}^{n } \frac{(b+\mu_2)^k}{k ! }
\right ) \left ( \displaystyle \sum_{k=0}^{n } \frac{b^k}{k ! } \right)^{-1 } \ , .
\label{integral - probability}\ ] ] the shortest @xmath0 credibility intervals @xmath10 $ ] are obtained by choosing @xmath66 and @xmath39 such that @xmath67|n , b)=\alpha$ ] and @xmath68 if possible ( with @xmath69 ) , or @xmath70 . ] .
one can see that the behavior of @xmath39 obtained with the bayesian ordering method is intermediate between those in the unified approach and in the bayesian theory . although one must always remember that the statistical meaning of @xmath39 is different in the two frequentist methods ( unified approach and bayesian ordering ) and in the bayesian theory , for scientists using these upper limits it is often irrelevant how they have been obtained .
hence , i think that an approximate agreement between frequentist and bayesian results is desirable . from eq .
( [ bev ] ) one can see that @xmath71 therefore , for @xmath72 the confidence belt obtained with the bayesian ordering method is similar to that obtained with the unified approach .
the difference between the two methods show up only for @xmath13 .
this is illustrated in figs .
[ bolim ] and [ bobelt ] , that must be confronted with the corresponding figures [ ualim ] and [ uabelt ] in the unified approach .
notice that , as shown in fig .
[ bobelt ] , contrary to the unified approach , the bayesian ordering method gives physically significant ( non - zero - width ) confidence intervals even for low values of the confidence level @xmath2 .
* criticism : * _ bayesian ordering is a mixture of frequentism and bayesianism .
the uncompromising frequentist can not accept it .
_ no ! it is a frequentist method .
bayesian theory is only used for the _ choice of ordering _ in the construction of the acceptance intervals , that in any case is subjective and beyond frequentism ( as , for example , the central interval prescription or the unified approach method ) .
the bayesian method for such a subjective choice is quite natural .
if you belong to the frequentist orthodoxy ( sort of religion ! ) and the word `` bayesian '' gives you the creeps , you can change the name `` bayesian ordering '' into whatever you like and use its prescription for the construction of the acceptance intervals as a successful recipe .
* criticism : * _ in the unified approach ( and maybe bayesian ordering ? )
the upper limit on @xmath1 goes to zero for every @xmath4 as @xmath3 goes to infinity , so that a low fluctuation of the background entitles to claim a very stringent limit on the signal .
_ this is not true !
one can see it is given by the expression in eq .
( [ lr-1 ] ) , that tends to @xmath73 for @xmath74 and small @xmath1 . for @xmath75 , @xmath76 and all @xmath34
have rank close to maximum . for @xmath15
the likelihood ratio is given by the expression in eq .
( [ lr-2 ] ) . for large values of @xmath3 , taking into account that @xmath15 , we have @xmath77 and @xmath78 , which imply that @xmath79 .
so the rank drops rapidly for @xmath80 .
therefore , for small values of @xmath1 the @xmath4 s much smaller than @xmath3 have highest rank . since they have also very small probability , they all lie comfortably in the confidence belt , if the confidence level @xmath2 is sufficiently large ( @xmath30 ) . ] doing a calculation of the upper limit for @xmath1 as a function of @xmath3 for large values of @xmath3 .
the result of such a calculation in the unified approach is shown in fig .
[ higback - asy]a , where the 90% cl upper limit @xmath39 is plotted as a function of @xmath3 in the interval @xmath81 for @xmath48 ( solid line ) , @xmath82 ( dashed line ) and @xmath83 ( dotted line ) .
one can see that initially @xmath39 decreases with increasing @xmath3 , but it stabilizes to about 0.8 for @xmath74 , with fluctuations due to the discreteness of @xmath4 . figure [ higback - asy]b shows the same plot obtained with the bayesian ordering .
one can see that initially @xmath39 decreases with increasing @xmath3 , but less steeply than in the unified approach , and it stabilizes to about 1.8 . for comparison , in fig .
[ higback - asy]c i plotted @xmath39 as a function of @xmath3 in the bayesian theory with a flat prior and shortest credibility intervals .
one can see that the behavior of @xmath39 in the three methods considered in fig .
[ higback - asy ] is rather similar .
* criticism : * _ for @xmath48 the upper limit @xmath39 should be independent of the background @xmath3 . _
but for @xmath84 the upper limit @xmath39 always decreases with increasing @xmath3 ! it is true that for @xmath48 one is sure that no background event as well as no signal has been observed .
but this is just the effect of a low fluctuation of the background that _ is present _ !
should we built a special theory for @xmath48 ?
i think that this is not interesting in the frequentist framework , because i guess that it leads necessarily to a violation of coverage ( that could be tolerated , but not welcomed , only if it is overcoverage ) .
i think that if one is so interested in having an upper limit @xmath39 independent of the background @xmath3 for @xmath48 , one better embrace the bayesian theory ( see fig .
[ higback]c , fig . [ higback - asy]c and ref .
@xcite ) , which , by the way , may present many other attractive qualities ( see , for example , @xcite ) . *
criticism : * _ a ( worse ) experiment with larger background @xmath3 should not give a smaller upper limit @xmath39 for the same number @xmath4 of observed events . _ but , as shown in fig .
[ higback ] , this always happens ! notice that it happens both for @xmath15 ( dotted part of lines ) and for @xmath21 ( solid part of lines ) , in frequentist methods as well as in the bayesian theory ( for @xmath84 ) .
as far as i know , nobody questions the decrease of @xmath39 as @xmath3 is increased if @xmath15 .
so why should we question the same behavior when @xmath21 ?
the reason for this behavior is simple : the observation of a given number @xmath4 of observed events has the same probability if the background is small and the signal is large or the background is large and the signal is small .
i think that it is physically desirable that and experiment with a larger background do not give a _ much smaller _ upper limit for the same number of observed events , but a _ smaller _ upper limit is allowed by _
statistical fluctuation_. indeed , _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ upper limits ( as confidence intervals , etc . )
are statistical quantities that _ must fluctuate ! _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ i think that the current race of experiments to find the most stringent upper limit is bad , because it induces people to think that limits are fixed and certain . instead , everybody should understand that _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ a better experiment can sometimes give a worse upper limit because of statistical fluctuations and there is nothing wrong about it !
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
in this report i have shown that the necessity to choose a specific frequentist method , among several available , does not introduce any degree of subjectivity from a statistical point of view ( section [ significance ] ) @xcite .
in other words , all frequentist methods are statistically equivalent . however , the physical significance of confidence intervals obtained with different methods is different and scientists interested in obtaining reliable and useful information on the characteristics of the real world must worry about this problem . obtaining empty or very small confidence intervals for a physical quantity as a result of
a statistical procedure is useless .
sometimes it is even dangerous to present such results , that lead non - experts in statistics ( and sometimes experts too ) to false believes . in section [ beauty ] i have discussed some virtues and shortcomings of the unified approach @xcite .
these shortcomings are ameliorated in the bayesian ordering method @xcite , discussed in section [ ordering ] , that is natural , relatively easy , and leads to more reliable upper limits . in conclusion
, i would like to emphasize the following considerations : * one must always remember that , in order to have coverage , the choice of a specific frequentist method must be done independently of the knowledge of the data . * finding some examples in which a method fails does not imply that it should not be adopted in the cases in which it performs well . * since all frequentist methods are statistically equivalent , + there is no need of a general frequentist method !
+ in each case one can choose the method that works better ( basing the judgment on easiness , meaningfulness of limits , etc . ) .
complicated methods with a wider range of applicability are theoretically interesting , but not attractive in practice . *
somebody thinks that the physics community should agree on a standard statistical method ( see , for example , @xcite ) . in that case , it is clear that this method must be always applicable .
but this is not the case , for example , of the unified approach , as shown in @xcite .
although the bayesian ordering method has not been submitted to a similar thorough examination , i doubt that it is generally applicable .
+ i do not see why experiments that explore different physics and use different experimental techniques should all use the same statistical method ( except a possible ignorance of statistics and blind believe to `` authorities '' ) .
+ i would recommend that + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ instead of wasting time on useless characteristics as generality , _ the physics community should worry about the usefulness and credibility of experimental results_. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
i would like to thank marco laveder for fruitful collaboration and many stimulating discussions .
|
it is shown that all the frequentist methods are equivalent from a statistical point of view , but the physical significance of the confidence intervals depends on the method . the bayesian ordering method is presented and confronted with the unified approach in the case of a poisson process with background . some criticisms to both methods are answered .
it is also argued that a general frequentist method is not needed .
|
we consider classical knot and link exteriors as a context in which to introduce techniques which determine the geometric structure of non - simple @xmath1-generator @xmath0-manifolds from purely algebraic data . here
the application of these techniques are to a pair of related conjectures on the heegaard realization of generators .
these are the so called two generator is tunnel one " and meridional generator " conjectures .
these two conjectures arise from two of the problems in r. kirby s famous 1978 list [ k ] .
the first is a problem attributed to l. moser which asks for a geometric characterization of those knots and links whose fundamental group may be presented using two group elements .
the two generator is tunnel one " conjecture purports to be the solution to this particular problem .
this conjecture is usually attributed to m. scharlemann , who attributes it to a. casson , who notes it is implicit in a question asked in 1967 by f. waldhausen [ w ] .
the second is a problem attributed to s. cappell and j. shaneson which asks if generators which can be represented ( under a suitable choice of basepoint ) by a simple closed curve on the boundary ( the so - called _ meridional _ generators ) correspond to the bridges of a bridge decomposition of the knot exterior .
the meridional generator conjecture asserts an affirmative answer to cappell and shanneson s question although considered in the broader context of a generalized notion of bridge decomposition .
little is known on either of these conjectures .
the only published result on the first is due to s. bleiler [ b ] where the conjecture is established for cable knots . a single published result exists for the second as well .
this is due to m. boileau and b. zimmermann [ bz ] , who use thurston s orbifold technology to show that a knot or link exterior whose group is generated by a pair of meridional elements is in fact two bridge . the principal result of this paper , theorem 3.1 , classifies those satellites with a two generator presentation in which one of the generators is meridional .
note that this includes all known two generator satellites .
the two generator is tunnel one and meridional generator conjecture for these knots follows as a corollary to the classification .
it also follows from this classification that a two generator satellite knot out in this class is a counter - example to the two generator is tunnel one conjecture .
the key to the classification is an algebraic theorem given in 3.10 .
the paper is organized as follows .
the first and second sections respectively contain the topological and algebraic backgound necessary for the classification theorem which is stated and proved in section three .
the techniques developed here apply more generally to the classification of non - simple @xmath0-manifolds .
the details of this will appear in a subsequent publication .
* 1.1 . * begin by recalling b. clark s construction of a knot whose fundamental group is at most @xmath2-generator [ c ] .
start with a genus @xmath2 heegaard splitting of the @xmath0-sphere and consider one of the handlebodies as the union of a single @xmath3-handle and @xmath2 @xmath4-handles .
form the exterior of a knot by removing the @xmath3-handle and one of the @xmath4-handles .
the cores of the remaining @xmath4-handles form a set of _ unknotting tunnels _ for the knot @xmath5 .
the minimal cardinality of a set of unknotting tunnels for @xmath5 is called the _ tunnel number _ of @xmath5 .
it follows that the rank of the fundamental group of a tunnel number @xmath6 knot is at most @xmath7 .
it is a consequence of dehn s lemma that the rank of the group of a tunnel number one knot is exactly two .
clark s construction can be generalized to form complements of @xmath6 component links by decomposing the genus @xmath2 handlebody into a collection of @xmath6 @xmath3-handles and @xmath8 @xmath4-handles and deleting a collection of @xmath6 @xmath3-handles and @xmath6 @xmath4-handles .
it is occasionally useful to consider this construction in the context of general 3-manifolds .
next , recall h. doll s generalization of schubert s notion of bridge number [ d ] .
a properly embedded arc @xmath9 in a handlebody is said to be _ trivial _ if there exists an arc @xmath10 in the boundary of the handlebody whose endpoints agree with those of @xmath9 such that @xmath11 bounds a disc in the handlebody .
a knot in a @xmath0-manifold @xmath12 has a _
@xmath13-decomposition _ or alternately , is said to be in _ @xmath14-bridge position with respect to a heegaard surface @xmath15 of genus @xmath2 _ if the knot intersects the closure of each complementary handlebody to @xmath15 in @xmath14 trivial arcs .
one now defines the _
genus @xmath2 bridge number _ for a link @xmath16 to be the minimal number @xmath14 for which the link has a @xmath13 decomposition with respect to a genus @xmath2 splitting of the ambient manifold @xmath12 . note that schubert s original definition of bridge number is the above definition applied to @xmath17 with @xmath18 .
also note that when @xmath19 , by stabilizing the heegaard splitting one can always change bridges to genus , that is , a link with a @xmath13 decomposition always has a @xmath20 decomposition .
that the converse is false is demonstrated by the @xmath21 pretzel knot which has no @xmath22 decomposition ( i.e. is not two bridge ) , but as the closure of the 7-string braid @xmath23 has a @xmath24 decomposition
. see figure 1 . also note from figure 1 that as a pretzel knot
, the @xmath25 pretzel also has a @xmath26 decomposition .
for convenience , we refer to a knot with a @xmath24 decomposition as _
one bridge with respect to a torus_. note that such knots are naturally tunnel one , see [ g ] .
* we now develop the terminology that describes how these geometric concepts appear in the algebraic data .
an element of the fundamental group of a link exterior is said to be _ meridional _ if there is a choice of basepoint so that the element is represented by a simple closed curve in the boundary .
note that an element can be meridional and yet not lie in the fundamental group of the boundary of the knot , as exemplified by the two bridge knot and link exteriors .
the fundamental group of a two bridge link can be generated by two meridional elements , both of which can not lie in the fundamental group of the boundary .
a link exterior is said to have a _
@xmath13 presentation _ if its fundamental group has a presentation with @xmath27 generators , @xmath14 of which are meridional .
note that if @xmath19 that a knot with a @xmath13 decomposition has a @xmath13 presentation .
the converse is : * meridional generator conjecture . * _ a knot with a @xmath13 presentation has a @xmath13 decomposition . _
remark : that a knot with a @xmath28 decomposition need not have a @xmath28 presentation is exemplified by the torus knots .
however , as of this writing , no counterexample is known to the conjecture that a knot with a @xmath28 presentation has a @xmath28 decomposition .
remark : cappell and shaneson s original problem that n - meridional generators implies @xmath6-bridge is just the special case @xmath18 . * 1.4 .
* the tunnel one satellite knots in @xmath17 were classified in 1988 by k.morimoto and h. sakuma in [ ms ] and the tunnel one satellite knots in general @xmath0-manifolds were classified in 1995 by j. neil [ n ] . per morimoto and sakuma s classification ,
a satellite knot @xmath5 is tunnel one when it satisfies the following three conditions .
first , the exterior of @xmath5 must be the union of a simple link exterior @xmath29 and a torus knot exterior @xmath30 .
second , the link exterior @xmath29 must be attached to @xmath30 in a manner that takes a meridian of @xmath29 to the seifert fibre of @xmath30 .
finally , the link exterior @xmath29 must embed in @xmath17 as the exterior of a @xmath1-bridge link other than the un- or hopf links .
as can be seen from the construction , all the knots obtained in this manner are one bridge with respect to the torus knot s torus . see figure 2 where this is illustrated for the rolfsen - bailey knot , i.e. the one obtained by attaching the exterior of the two bridge link @xmath31 to the exterior of the left hand trefoil , as described above .
it is interesting to note that m. eudave - munoz was able to recover this classification by first proving that a tunnel one satellite in @xmath17 has a @xmath24-decomposition , see [ e ] .
* 2.1 . * here we review some of the properties of groups needed in future discussion . let @xmath32 denote a group . a subgroup @xmath33 of @xmath32
is called _ proper _ if @xmath33 is nontrivial and there exist elements of @xmath32 outside of @xmath33 . a proper subgroup @xmath33 is said to be _ malnormal _ in @xmath32 if the intersection of subgroups @xmath34 and @xmath33 is trivial if and only if @xmath35 is an element of @xmath32 outside of @xmath33 .
thus @xmath33 intersects each of its nontrivial conjugates trivially .
an element @xmath36 is said to be _ central _ if @xmath35 commutes with each element of @xmath32 .
a _ torsion _ element of @xmath32 is an element @xmath35 satisfying the relation @xmath37 for some nonzero integer @xmath6 .
the _ rank _ of a finitely presentable group is the minimal number of generators needed to present the group .
a group @xmath32 is called an _
n - generator _ _ group _ if @xmath32 is known to have rank @xmath6 .
the groups of rank one are the finite cyclic groups together with the infinite cyclic group .
the groups of higher rank form a much more complicated class . indeed , determining the rank of a group is in general a difficult problem .
the following concepts are presented in greater detail in magnus , karrass , and solitar [ mks ] and are recalled here for the convenience of the reader .
let @xmath38 denote the free product with amalgamation of the non - trivial groups @xmath32 and @xmath39 along the group @xmath40 , that is , @xmath40 is isomorphic to a proper subgroup @xmath41 of @xmath32 and also isomorphic to a proper subgroup @xmath42 of @xmath39 .
the subgroups @xmath41 and @xmath42 are identified with the group @xmath40 via the respective isomorphisms .
when the context is clear , we simply refer to the subgroup @xmath41 of @xmath32 by @xmath40 , and similarly the subgroup @xmath42 of @xmath39 will be referred to by @xmath40 .
note that group @xmath43 contains each of @xmath32 , @xmath39 and @xmath40 as proper subgroups .
the groups @xmath32 and @xmath39 are called the _ free _ _ factors _ for the group @xmath38 and the group @xmath40 is called the _ amalgamated _ _
subgroup_. any element @xmath44 of @xmath38 may be uniquely expressed in _
_ form _ as @xmath45 where @xmath46 is an element of @xmath40 and the @xmath47 are alternately elements of a fixed set @xmath48 of right transversals for the ( right ) cosets of @xmath40 in @xmath32 and a corresponding fixed set @xmath49 of right transversals for the ( right ) cosets of @xmath40 in @xmath39 .
the coset @xmath40 in either of @xmath32 or @xmath39 always receives the transversal represented by 1 and the transversal representative 1 never appears in the normal form word for @xmath44 . if @xmath44 is expressed in the above normal form , then @xmath44 is said to have length @xmath50 , and one writes @xmath51 .
if @xmath52 is an element of @xmath32 , then @xmath44 is said to begin in @xmath32 .
if @xmath53 is an element of @xmath32 , then @xmath44 is said to end in @xmath32 .
similarly , if @xmath47 is an element of @xmath39 for @xmath54 or @xmath55 , then @xmath44 is said to begin or end respectively in @xmath39 .
if @xmath44 begins in @xmath32 and ends in @xmath39 , then @xmath44 can be written concretely in the normal form @xmath56 .
if @xmath44 can be written in this form , then @xmath44 is said to be a word of length @xmath50 beginning in @xmath32 and ending in @xmath39 , and similarly for the other cases .
two words @xmath44 and @xmath57 written in normal form are multiplied by concatenation .
the product is subsequently converted to normal form by moving the elements of @xmath40 to the right through the word @xmath57 . if the word @xmath58 ends in the same group that the word @xmath59 begins in , then the product @xmath60 is an element of one of @xmath32 or @xmath39 .
if @xmath60 is an element of the complement of @xmath40 in that factor , then @xmath61 is said to have _ amalgamation _ and @xmath62 . if the product @xmath63 , an element of @xmath40 , then the word @xmath61 has _ cancellation _ and the product @xmath64 is then examined for additional cancellation and/or amalgamation and @xmath65 .
if @xmath44 ends in a different group than @xmath57 begins in , there is no amalgamation or cancellation in @xmath61 and @xmath66 .
now recall the results of [ bj ] which will be relevent to the discussion at hand . *
lemma [ bj , 1.5 ] . * _ suppose that the group @xmath38 is generated by @xmath6 elements @xmath67 , @xmath68 ,
the set @xmath70 consisting of those transversals of @xmath48 appearing in any of the @xmath71 together with all of the elements of @xmath40 generates the group @xmath32 .
_ * 2.5 .
* by symmetry , the group @xmath39 may be generated by the set @xmath72 consisting of all transversal elements of @xmath49 that appear in at least one of the generators @xmath71 together with the elements of @xmath40 .
theorem [ bj 3.1 ] . *
_ let @xmath73 be the free product with amalgamation of the torsion free group @xmath32 and the group @xmath39 .
if the amalgamated subgroup @xmath40 is of rank at least two and malnormal in @xmath32 and if @xmath43 has rank two , then either @xmath39 contains torsion elements or @xmath43 has a generating pair of one of four types ( listed here as in [ bj , 5.1 ] ) _ : @xmath74 @xmath75 @xmath76 @xmath77 * 2.7 .
proposition [ bj 5.1.1 ] .
* _ if @xmath43 has a generating pair of type @xmath78 and @xmath40 is abelian , then there is a minimal positive integer @xmath50 such that the element @xmath79 is an element @xmath80 of @xmath40 which is central in @xmath39 . _
* 3.1 . theorem . * _ the exterior of a @xmath24 presented satellite knot in the @xmath0-sphere decomposes as the union along a torus of a simple link exterior @xmath29 and a torus knot exterior @xmath30 .
moreover , the link exterior @xmath29 embeds in the @xmath0-sphere as the complement of a two bridge link and is attached to @xmath30 so that a meridian for this two bridge link is identified with the seifert fibre of the torus knot exterior . _ * 3.4 .
* let @xmath5 be a @xmath24 presented satellite knot in @xmath17 .
if @xmath5 is a cable knot , then by [ b ] , @xmath5 is the @xmath81 cable on the @xmath82 torus knot . here
, the space @xmath29 embeds in @xmath17 as the complement of the two bridge link corresponding to the even integer @xmath83 .
the conclusion of 3.1 follows . without loss of generality , the knot @xmath5 is not cabled .
denote the exterior of @xmath5 by @xmath84 .
the space @xmath84 has a decomposition by essential tori into simple pieces , see [ js ] or [ j ] .
one of these simple pieces has @xmath85 as one of its boundary components .
denote this piece by @xmath86 .
thurston s geometrization for haken manifolds [ t ] implies that @xmath86 is either a seifert fibred space or its interior admits a hyperbolic metric .
if not hyperbolic , the space @xmath86 contains a non - boundary parallel essential annulus , vertical with respect to the seifert fibration on @xmath86 , with both boundary components lying on the boundary of @xmath84 . as @xmath5 is not a cabled knot , it follows that the knot @xmath5 is composite , contradicting norwood s result [ n ] that two - generator knots are prime .
hence the space @xmath86 is hyperbolic .
the space @xmath86 has at least two boundary tori .
let the torus @xmath87 be a boundary torus that is different from @xmath85 .
this torus @xmath87 separates the space @xmath84 into two pieces .
let @xmath29 be the component of @xmath88 containing @xmath86 and @xmath30 the complementary component .
since t is a separating torus , by the seifert - van kampen theorem , the group @xmath89 splits as the free product @xmath90 where @xmath32 is the group @xmath91 , @xmath39 is the group @xmath92 and @xmath40 is the group of the torus @xmath87 .
( _ proof of 3.5 .
_ ) let @xmath2 be an element of the group @xmath32 such that @xmath93 for elements @xmath94 .
hence there exists a proper , @xmath95-injective map @xmath96 of the annulus @xmath97 into @xmath29 .
put the image of @xmath96 into general position with respect to the decomposing tori for @xmath29 inherited from the torus decomposition of @xmath84 .
while the intersection of the image of @xmath96 with one of these tori may be hideous on the torus , the intersection on the annulus may , by standard techniques , be reduced to a family @xmath98 of parallel essential simple closed curves on the annulus .
the restricition of @xmath96 to a complementary component of @xmath98 which contains a component of @xmath99 is a proper , @xmath95-injective map of an annulus into the hyperbolic manifold @xmath86 , hence the image of this restricted @xmath96 can be isotoped into the torus @xmath87 .
induction then shows that the entire image @xmath96 can be isotoped into @xmath87 and hence that @xmath100 .
@xmath101 * 3.8 . *
( _ proof of 3.7 .
_ ) as the group @xmath43 is @xmath24-presented , the group @xmath104 has a generating pair @xmath105 . without loss of generality ,
the generator @xmath67 is meridional and hence it has normal form @xmath106 .
thus there are exactly five possibilites for the normal forms of the generating set @xmath105 .
these are : by inverting the second element , we can consider a pair of type @xmath112 to be of type @xmath113 .
but by [ bj 4.4 ] , a pair of type @xmath113 can not generate @xmath43 and by [ bj 4.5 ] neither can a pair of type @xmath114 .
so if @xmath115 is not of type @xmath78 , it must be of type @xmath116 .
note that by an argument similar to that in 3.6 , the element @xmath117 is never an element of the subgroup @xmath40 .
this is essentially the argument that a hyperbolic manifold contains no essential annuli . thus conjugating a pair of type @xmath116 by the element @xmath118 produces a generating set of type @xmath112 or @xmath114 .
thus @xmath119 is not of type @xmath116 .
the desired result follows .
@xmath101 * 3.9 . * from 3.7 and 2.7 , the element @xmath120 is central in @xmath39 .
it follows from g. burde and h. zieschang [ bz ] that @xmath30 is the complement of a torus knot and that the element @xmath80 is a power of the class of the torus knot s regular seifert fibre .
there is a choice of basis @xmath121 for the group @xmath40 so that @xmath44 is the class of a meridian for the torus knot and @xmath57 coincides with @xmath80 if @xmath80 is a primitive element of @xmath40 and is a primitive root of @xmath80 otherwise . *
* _ the elements @xmath122 and @xmath80 generate the group @xmath32 .
hence @xmath32 has a @xmath22 presentation .
in particular , the space @xmath29 embeds in @xmath17 as the exterior of a two bridge link with meridians representing @xmath122 and a primitive root of @xmath80 . _
( _ proof of 3.11 .
_ ) assume not .
for any integer @xmath125 , an element of form @xmath126 is an element of @xmath40 outside of the cyclic subgroup generated by @xmath80 , and as an element of @xmath43 can be expressed as a word @xmath127 in the generating pair @xmath103 .
the word @xmath127 can be written in the following form : the integers @xmath78 and @xmath129 are allowed to be zero , but all other integers @xmath130 are nonzero . as @xmath126 is an element of @xmath40 the word @xmath127 admits amalgamations and cancellations that reduce its length to zero .
we now begin a series of rewritings of the word @xmath127 .
note that if @xmath131 is a multiple of the integer @xmath50 , then we may replace the phrase @xmath132 by a power of @xmath80 .
do this throughout the word @xmath127 and combine maximal length phrases within @xmath127 that are words in the elements @xmath122 and @xmath80 . the word @xmath127 now has the following appearance : each word @xmath134 is a word in the elements @xmath122 and @xmath135 . with the exception that @xmath78 and @xmath136 are allowed to be zero ,
no power @xmath131 is a multiple of the integer @xmath50 , else it is combined into a neighboring @xmath137 or @xmath138 .
if @xmath127 has form @xmath139 , then we have established 3.11 .
hence we assume that each phrase @xmath137 , if not in @xmath140 is a power of @xmath80 .
since @xmath80 and @xmath141 commute , all pwers of @xmath141 my be moved right through the word @xmath127 .
if @xmath137 is a power of @xmath80 , any occurence of the phrase @xmath142 may be rewritten as @xmath143 , and then reindexed as necessary . in this way ,
the word @xmath127 has been rewritten in the form : each word @xmath134 is a word in @xmath140 . with the exception that @xmath78 is allowed to be zero and @xmath136 could be an arbitrary integer , no other power @xmath131 is a multiple of the integer @xmath50 .
notice that @xmath127 has now been written in a form admitting no further cancellation or amalgamation .
thus if @xmath127 is to have length zero , it must have form @xmath145 where @xmath125 is a multiple of @xmath50 or @xmath139 .
but either of these options produce a power of @xmath80 , a contradiction . @xmath101 * 3.13 . *
( _ proof of 3.10 .
_ ) by 3.11 , the elements @xmath122 and @xmath80 generate an element of form @xmath146 . if @xmath147 , then by 2.4 , the elements @xmath122 and @xmath80 generate the entire group @xmath32 .
assume that @xmath122 and @xmath80 do not generate the element @xmath44 and let @xmath148 be the minimal positive integer for which the elements @xmath122 and @xmath80 generate the element @xmath149 .
it follows from the argument in 3.12 that since the pair @xmath150 generate the group @xmath43 , that the elements @xmath141 and @xmath149 must generate an element of form @xmath151 where @xmath152 .
thus , theorem 3.10 follows from the proposition below . *
( _ proof of 3.14 .
_ ) as noted in 3.9 , @xmath39 is the group of a torus knot .
hence the group @xmath39 decomposes as the free product with amalgamation of form @xmath153 , where each of the groups @xmath154 , @xmath155 and @xmath15 are infinite cyclic .
geometrically , this corresponds to the familiar picture of constructing a torus knot exterior by glueing together two solid tori along an annulus in the boundary of each .
each infinite cyclic group @xmath155 and @xmath15 corresponds to the group of a solid torus .
let @xmath156 be the homomorphism defined by quotienting the group @xmath39 by the infinite cyclic subgroup generated by the class of the regular seifert fibre .
by abusive notation we both @xmath157 and @xmath158 will be denoted by the symbol @xmath157 .
similarly , both @xmath159 and @xmath160 will be denoted by the symbol @xmath159 , the context making the meaning clear .
note for later use that as @xmath125 and @xmath124 are the indices of the exceptional fibres in the seifert fibration of a torus knot exterior , they are relatively prime .
thus one of these integers must be odd and so one of @xmath161 or @xmath162 has no element of order two . as the meridian for the torus knot ,
the element @xmath44 has normal form @xmath163 , and so the element @xmath164 has normal form @xmath165 in the free product @xmath166 . note that the normal form for the element @xmath149 is then @xmath167 , in particular has length @xmath168 . since @xmath169 ,
the element @xmath170 lies in either a free factor of @xmath171 or a conjugate thereof .
each of these cases is examined in turn .
* suppose the element @xmath170 lies in a free factor of @xmath166 .
without loss of generality , @xmath172 .
let @xmath127 be a word in the elements @xmath170 and @xmath173 .
then we may assume that @xmath127 appears as follows : with the possible exception of @xmath78 and @xmath129 , all integers @xmath130 are nonzero and such that no power of @xmath175 is trivial . if @xmath127 is a nontrivial power of the element @xmath175 , note that @xmath176 in @xmath166 is one .
if @xmath127 is a power of @xmath173 , say @xmath177 , then @xmath178 .
we now assume that @xmath127 has neither of these forms and compute the length of @xmath127 . before computing lengths , some preliminaries
are required : given a fixed word @xmath127 as above , associate a series of integers to @xmath127 as in [ bj ] . in particular ,
let @xmath179 denote the associated sequence @xmath180 .
the integer @xmath181 denotes the number of times that the sequence @xmath179 changes algebraic sign .
for example , if @xmath127 is the word @xmath182 , then the sequence @xmath179 is @xmath183 and the associated integer @xmath184 is @xmath1 .
note that the relation @xmath185 holds .
further , for each integer @xmath186 satisfying @xmath187 , let @xmath188 denote the length of the phrase @xmath189 .
also note that with the possible exception of @xmath78 and @xmath129 all these integers are one . without amalgamation or cancellation ,
the uncancelled " length of the word @xmath127 is given by : in what follows , assume that the integer @xmath148 is at least two .
note that as the sequence @xmath179 proceeds from @xmath191 to @xmath192 , the phrase @xmath193 may admit cancellation or amalgamation as outlined below : * @xmath194 * if both @xmath191 and @xmath192 are positive , the phrase @xmath195 may have cancellations and amalgamation . by 3.15 ,
there are at most two cancellations and a single amalgamation yielding a maximal length reduction of five .
note that in all cases the elements @xmath175 and @xmath173 can not generate any elements of length less than @xmath205 other than those elements of form @xmath206 .
for instance , in 3.15.1.a or in 3.15.1.b , then the length of the word @xmath127 is at least @xmath207 .
note if j = 0 , the resulting word is of form @xmath206 .
so @xmath208 and if @xmath148 is at least three , then the inequality @xmath209 implies that @xmath210 , a contradicition .
but if @xmath211 , then the inequality reduces to @xmath212 , another contraction .
the statement of 3.14 follows .
let @xmath127 be a word in the elements @xmath170 and @xmath173 . recall that @xmath173 has normal form @xmath167 .
the normal form for the element @xmath173 can be written out and depends upon which of the four normal forms above the element @xmath2 has .
since @xmath161 is an abelian group , we may assume for the purposes of working with @xmath170 that @xmath2 has normal form of type ii or type iv .
if not the trivial element , the element @xmath224 has length at least three . by conjugating , if necessary
, we may assume that the element @xmath225 beginning the word @xmath2 in case ii is not the inverse of the element @xmath226 that begins the word @xmath167 .
similarly , we may assume that the element @xmath227 beginning the word @xmath2 in case iv is not the inverse of the element @xmath228 , ending the word @xmath167 .
the word @xmath127 has the following form : with the possible exception of @xmath78 and @xmath129 , all integers @xmath130 are nonzero and no power of @xmath170 is trivial . for each of cases ii and iv , the phrase @xmath230 is examined for cancellation and or amalgamation : * 3.15.2.1 . *
if @xmath125 is a positive integer , writing out the phrase using @xmath231 instead of @xmath224 in case ii shows that there is an amalgamation in the elements @xmath232 , and no cancellation since the elements @xmath225 and @xmath226 are not inverses . in case iv
, there is no cancellation or amalgamation of any kind .
if @xmath125 is a negative integer , there is no amalgamation or cancellation in case ii . in case iv
, there is amalgamation , but no cancellation since the elements @xmath227 and @xmath228 are not inverse . for the variable @xmath249
, shows that if @xmath127 is the representative of a word of length less than @xmath205 , then @xmath127 has form @xmath224 .
but by 3.15.2 , such an element has odd length . as the length of the element @xmath250 is the even number @xmath251 , the elements @xmath170 and @xmath252 can not generate an element of form @xmath250 for @xmath152 .
theorem 3.14 , and hence theorem 3.10 follows .
|
techniques are introduced which determine the geometric structure of non - simple two - generator @xmath0-manifolds from purely algebraic data . as an application
, the satellite knots in the @xmath0-sphere with a two - generator presentation in which at least one generator is represented by a meridian for the knot are classified .
|
sinyukov , s.v .
akkelin , y. hama , phys .
rev . lett . * 89 * , 052301 ( 2002 ) .
akkelin , y. hama , iu.a .
karpenko , yu.m .
sinyukov , in preparation .
t. hirano , phys .
c * 65 * , 011901 ( 2001 ) ; t. hirano , k. tsuda , phys .
c * 66 * , 054905 ( 2002 ) ; t. hirano , private communication m. laine , y. schrder , phys .
d * 73 * , 085009 ( 2006 ) .
t. lappi , phys .
b * 643 * , 11 ( 2006 ) .
|
we describe rhic pion data in central a+a collisions and make predictions for lhc based on hydro - kinetic model , describing continuous 4d particle emission , and initial conditions taken from color glass condensate ( cgc ) model .
hydro - kinetic approach to heavy ion collisions proposed in ref .
@xcite accounts for continuous particle emission from 4d volume of hydrodynamically expanding fireball as well as back reaction of the emission on the fluid dynamics .
the approach is based on the generalized relaxation time approximation for relativistic finite expanding systems , @xmath0 where @xmath1 is phase - space distribution function ( df ) , @xmath2 is local equilibrium distribution and @xmath3 is relaxation time , @xmath4 as well as @xmath5 are functional of hydrodynamic variables .
complete algorithm described in detail in ref .
@xcite includes : solution of equations of ideal hydro ; calculation of a non local equilibrium df and emission function in the first approximation ; solution of equations for ideal hydro with non - zero right - hand - side that accounts for conservation laws at the particle emission during expansion ; calculation of `` improved '' df and emission function ; evaluation of spectra and bose - einstein correlations . here
we present our results for the pion momentum spectra and interferometry radii calculated for rhic and lhc energies in the first approximation of the hydro - kinetic approach . for simulations we utilize ideal fluid model @xcite and realistic equation of state ( eos ) that combines high temperature eos with crossover transition
@xcite adjusted to the qcd lattice data and eos of hadron resonance gas with partial chemical equilibrium @xcite .
the gradual disappearance of pions during the crossover transition to deconfinement and different intensity of interactions of pions in pure hadronic and `` mixed '' phases are taken into account in the hydro - kinetic model ( hkm ) , but resonance contribution to pion spectra and interferometry radii is not taken into account in the present version of the hkm .
we assume the following initial conditions at proper time @xmath6 fm / c for hkm calculations : boost - invariance of a system in longitudinal direction and cylindrical symmetry with gaussian energy density profile in transverse plane . the maximal energy densities at rhic , @xmath7 gev/@xmath8 and at lhc , @xmath9 gev/@xmath8 , were calculated from ref .
@xcite in approximation of bjorken expansion of free ultrarelativistic partons till @xmath10 and adjusted for transverse gaussian density profile .
the ( pre - equilibrium ) initial transverse flows at @xmath10 were estimated assuming again a free - streaming of partons , with transverse modes distributed according to cgc picture , from proper time @xmath11 0.1 fm / c till @xmath6 fm / c . finally , we approximate the transverse velocity profile by @xmath12 where @xmath13 both for rhic and lhc energies and we suppose the fitting gaussian radius for rhic top energy , @xmath14 , to be the same for lhc energy .
our results for rhic and predictions for lhc are presented in fig . @xmath15 .
the relatively small increase of the interferometry radii with energy in hkm calculations is determined by early ( as compare to sharp freeze - out prescription ) emission of hadrons , and also by increase of transverse flow at lhc caused by longer time of expansion .
it is noteworthy that in the case of eos related to first order phase transition , the satisfactory fitting of the rhic hbt data requires non - realistic high initial transverse flows at @xmath6 fm / c : @xmath16 .
-0.0in-0.0 in , @xmath17 , @xmath18 radii measured by the phenix collaboration for au+au central collisions ( hbt radii data were recalculated for @xmath19 centrality ) at rhic with the hkm calculations , and hkm predictions for pb+pb central collisions at lhc . for the sake of convenience
the calculated one - particle spectra are enhanced in @xmath20 times.,title="fig:",width=307 ] , @xmath17 , @xmath18 radii measured by the phenix collaboration for au+au central collisions ( hbt radii data were recalculated for @xmath19 centrality ) at rhic with the hkm calculations , and hkm predictions for pb+pb central collisions at lhc . for the sake of convenience
the calculated one - particle spectra are enhanced in @xmath20 times.,title="fig:",width=307 ] , @xmath17 , @xmath18 radii measured by the phenix collaboration for au+au central collisions ( hbt radii data were recalculated for @xmath19 centrality ) at rhic with the hkm calculations , and hkm predictions for pb+pb central collisions at lhc .
for the sake of convenience the calculated one - particle spectra are enhanced in @xmath20 times.,title="fig:",width=307 ] , @xmath17 , @xmath18 radii measured by the phenix collaboration for au+au central collisions ( hbt radii data were recalculated for @xmath19 centrality ) at rhic with the hkm calculations , and hkm predictions for pb+pb central collisions at lhc . for the sake of convenience
the calculated one - particle spectra are enhanced in @xmath20 times.,title="fig:",width=307 ] -0.3 cm [ fig ] -0.8 cm
|
as we know the light rays or photons would be deviated from their straight way when they pass close to the massive object such as black holes .
this deflection of light rays is known as gravitational lensing .
this gravitational lensing is one of the applications and results of general relativity @xcite and is used as an instrument in astrophysics , because it can help us to extract the information about stars . in 1924
chwolson pointed out that when a star(source ) , a deflector(lens ) and an observer are perfectly aligned , a ring - shape image of the star appears which is called einstein ring. other studies have been led by klimov , liebes , refsdal and bourassa and kantowski @xcite .
klimov investigated the lensing of galaxies by galaxies @xcite , but liebes studied the lensing of stars by stars and also stars by clusters in our galaxy @xcite .
refsdal showed that the geometrical optics can be used for investigating the gravitational lenses properties and time delay resulting from it @xcite .
the gravitational lensing has been presented in details in @xcite and reviewed by some papers ( see for examples @xcite-@xcite ) . at this stage ,
the gravitational lensing is developed for weak field limit and could not describe some phenomena such as looping of light rays near the massive objects .
hence , scientists started to study these phenomena from another point of view and they proposed gravitational lensing in a strong field limit . when the source is highly aligned with lens and the observer , one set of infinitive relativistic `` ghost '' images would be produce on each side of black hole .
these images are produced when the light rays that pass very close to black hole , wind one or several times around the black hole before reaching to observer . at first
, this phenomenon was proposed by darwin @xcite and revived in refs .
darwin proposed a surprisingly easy formula for the positions of the relativistic images generated by a schwarzschild black hole .
afterward several studies of null geodesics in strong gravitational fields have been led in literatures : a semi - analytical investigation about geodesics in kerr geometry has been made in @xcite , also the appearance of a black hole in front of a uniform background was studied in refs .
recently , virbhadra and ellis formulated lensing in the `` strong field limit '' and obtain the position and magnification of these images for the schwarzschild black hole @xcite . in ref @xcite , by an alternative formulation , frittelli , kiling and newman attained an exact lens equation , integral expressions for its solutions , and compared their result with virbhadra and ellis .
afterwards , the new method was proposed by bozza et al .
in which they revisited the schwarzschild black hole lensing by retaining the first two leading order terms @xcite .
this technic was used by eiroa , romero and torres to study a reissner - nordstrom black hole @xcite and petters to calculate relativistic effects on microlensing events @xcite .
finally , the generalization of bozza s method for spherically symetric metric was developed in @xcite .
bozza compared the image patterns for several interesting backgrounds and showed that by the separation of the first two relativistic images we can distinguish two different collapsed objects . further development for other black holes can be found in @xcite-@xcite .
several interesting studeies are devoted to lensing by naked singularities @xcite , janis - newman - winicour metric @xcite and role of scaler field in gravitational lensing @xcite . in ref @xcite , virbhadra et al . have considered a static and circularly symmetric lens characterized by mass and scalar charge parameter and investigated the lensing for different values of charge parameter .
+ the gravitational lenses are important tools for probing the universe .
narasimha and chitre predicted that the gravitational lening of dark matter can give the useful data about of position of dark matter in the universe @xcite .
also in some papers gravitational lens is used to detect exotic objects in the universe , such as cosmic strings @xcite-@xcite .
+ recently , the idea of large extra dimensions has attracted much attention to construct theories in which gravity is unified with other forces @xcite .
one of the most interesting problems is the verification of extra dimensions by physical phenomena . for this purpose
higherdimensional black holes in accelerators @xcite and in cosmic rays @xcite-@xcite and gravitational waves from higher - dimensional black holes @xcite are studied . the five - dimensional einstein - maxwell theory with a chern- simons term predicted five - dimensional charged black holes @xcite .
such a higher - dimensional black holes would reside in a spacetime that is approximately isotropic in the vicinity of the black holes , but effectively four - dimensional far from the black holes @xcite .
these higher dimensional black holes are called kaluza - klein black holes .
the presence of extra dimension is tested by quasinormal modes from the perturbation around the higher dimensional black hole @xcite-@xcite and the spectrum of hawking radiation @xcite-@xcite .
the gravitational lensing is another method to investigate the extra dimension .
thus , the study of strong gravitational lensing by higher dimensional black hole can help us to extract information about the extra dimension in astronomical observations in the future .
the kaluza - klein black holes with squashed horizon @xcite is one of the extra dimensional black holes and it s hawking radiation and quasinormal modes have been investigated in some papers @xcite-@xcite . also , the gravitational lensing of these black holes is studied in several papers .
liu et al . have studied the gravitational lensing by squashed kaluza - klein black holes in refs @xcite and sadeghi et al .
investigated the charged type of this black hole @xcite .
+ one the other hands we know that our universe is rotational and it is reasonable to consider gdel background for our universe .
an exact solution for rotative universe was obtained by gdel .
he solved einstein equation with pressureless matter and negative cosmological constant @xcite .
the solutions representing the generalization of the gdel universe in the minimal five dimensional gauged supergravity are considered in many studies @xcite-@xcite .
the properties of various black holes in the gdel background are investigated in many works @xcite .
the strong gravitational lensing in a squashed kaluza - klein black hole in a gdel universe is investigated in ref .
+ in this paper , we tudy the strong gravitational lensing in a charged squashed kaluza - klein gdel black hole . in that case
, we see the effects of the scale of the extra dimension , charge of black hole and gdel parameter on the coefficients and observables of strong gravitational lensing .
+ so , this paper is organized as follows : section 2 is briefly devoted to charged squashed kaluza - klein gdel black hole background . in section 3
we use the bozza s method @xcite to obtain the deflection angle and other parameters of strong gravitational lensing as well as variation of them with extra dimension , gdel parameter and charge of black hole . in section 4
, we suppose that the supermassive object at the center of our galaxy can be considered by the metric of charged squashed kaluza - klein gdel black hole .
then , we evaluate the numerical results for the coefficients and observables in the strong gravitational lensing . in the last section
, we present a summary of our work .
the charged squashed kaluza- klein gdel black hole spacetime is given by @xcite , @xmath2,\ ] ] where @xmath3 @xmath4 and @xmath5 , @xmath6 , @xmath7 and @xmath8 . here
@xmath9 and @xmath10 are the mass and charge of the black hole respectively and @xmath11 is the parameter of gdel background .
the killing horizon of the black hole is given by equation @xmath12 , where @xmath13[m - q-8j^2(m+q)^2]}.\end{aligned}\ ] ] + we see that the black hole has two horizons .
as @xmath14 the horizon of the squashed kaluza- klein gdel black hole is recovered @xcite and when @xmath10 and @xmath11 tend to zero , we have @xmath15 , which is the horizon of five - dimensional schwarzschild black hole . here
we note that the argument of square root constraints the mass , charge and gdel parameter values .
when @xmath16 , we have @xmath17 , which means that the squashing effect disappears and the five - dimensional charged black hole is recovered . + by using the transformations , @xmath18 , @xmath19 and @xmath20 , the metric ( [ metric1 ] ) can be written in the following form , @xmath21 @xmath22+\rho_0[\rho_0(\rho_m+6\rho_q)+7\rho_m\rho_q]\right\},\end{aligned}\ ] ] with @xmath23 where @xmath24 and @xmath25 denote the outer and inner horizons of the black hole in the new coordinate and @xmath26 is a scale of transition from five - dimensional spacetime to an effective four - dimensional one . here @xmath27 , so that @xmath28 .
the komar mass of black hole is related to @xmath29 with @xmath30 , where @xmath31 is the four dimensional gravitational constant . by using relations ( [ horizon1 ] ) and ( [ tranformations ] )
we can obtain @xmath32 in the following coupled equations , @xmath33={a}(\rho_{h+},\rho_{h-}),\nonumber\\ & & 2\left[\rho_0(\rho_{h+}-\rho_{h-})\right]={b}(\rho_{h+},\rho_{h-}),\end{aligned}\ ] ] where @xmath34 when @xmath35 , the horizon of black hole becomes @xmath36 , as obtained in @xcite . in case of @xmath35 and @xmath37
, we have @xmath38 which is consistent with neutral squashed kaluza - klein black hole @xcite .
you note that the square root in relation ( [ ab ] ) constrains the values of @xmath26 , @xmath39 and @xmath11 . in the case
@xmath40 , the permissive regime is shown in ref.@xcite . for any value of @xmath39 there
is allowed rang for @xmath26 .
hence these parameters can not select any value and when @xmath11 increases from zero , permissive regime for @xmath39 and @xmath26 becomes more confined .
we solved the above coupled equations numerically and results are shown in figure 1 .
we see that the outer horizon of black hole increases with the size of extra dimension , @xmath26 and decreases with @xmath11 and @xmath39 .
note that two horizons of black hole coincide in especial values of parameters , which in that case we have an extremal black hole .
in this section , we are going to investigate the deflection angle of light rays when they pass close to a charged squashed kaluza - klein gdel black hole .
we also study the effect of the charge parameter @xmath39 , the scale of extra dimension @xmath26 and gdel parameter on the deflection angle and it s coefficients in the equatorial plane ( @xmath41 ) .
+ in this plane , the squashed kaluza - klein gdel metric reduces to @xmath42 where @xmath43 the null geodesic equations are , @xmath44 where @xmath45 where @xmath46 is the tangent vector to the null geodesics and the dote denotes derivative with respect to affine parameter .
we use equation ( [ geodesic ] ) and obtain the following equations , @xmath47 @xmath48.\ ] ] where @xmath49 , @xmath50 and @xmath51 are constants of motion .
also , the @xmath52-component of equation ( [ geodesic ] ) in equatorial plane @xmath41 , is given by , @xmath53=0.\end{aligned}\ ] ] if @xmath54 , then deflection angle will be zero and this is illegal , so we set @xmath55 . by using equation ( [ rhogeo ] )
one can obtain following expression for the impact parameter , @xmath56 and the minimum of impact parameter takes place in photon sphere radius @xmath57 , that is given by the root of following equation @xcite , @xmath58\mathcal{c}^\prime(\rho_s)- \mathcal{c}(\rho_s)\,\left[\mathcal{d}(\rho_s)^2\mathcal{f}^\prime(\rho_s)+ 2\mathcal{d}(\rho_s)h(\rho_s)h^\prime(\rho_s)-h(\rho_s)^2\mathcal{d}^\prime(\rho_s)\right]=0.\ ] ] here @xmath59 is the closet approach for light ray and the prime is derivative with respect to @xmath59 .
the analytical solution for the above equation is very complicated , so we calculated the equation ( [ phequ ] ) numerically .
variations of r photon sphere radius are plotted with respect to the charge @xmath39 , the scale of extra dimension @xmath26 and gdel parameter in the figure 2 .
also figure 3 shows variations of impact parameter in it s minimum value ( at radius of photon sphere ) .
these figures show that by adding the charge to the black hole , the behavior of the photon sphere radius and minimum of impact parameter is different compare with the neutral black hole @xcite .
as @xmath26 approaches to it s minimum values the radius of photon sphere and impact parameter become divergent .
+ by using the chain derivative and equation ( [ rhogeo ] ) , the deflection angle in the charged squashed kaluza - klein gdel black hole can be written as , @xmath60 @xmath61 @xmath62 with @xmath63 when we decrease the @xmath59 ( and consequently @xmath64 ) the deflection angle increases . at some points , the deflection angle exceeds from @xmath65 so that the light ray will make a complete loop around the compact object before reaching at the observer . by decreasing @xmath59
further , the photon will wind several times around the black hole before emerging .
finally , for @xmath66 the deflection angle diverges and the photon is captured by the black hole .
moreover , from equations ( [ iphi ] ) and ( [ ipsi ] ) , we can find that in the charged squashed kaluza - klein gdel black hole , both of the deflection angles depend on the parameters @xmath11 , @xmath26 and @xmath39 , which implies that we could detect the rotation of universe , the extra dimension and charge of black hole in theory by gravitational lens .
note that the @xmath67 depend on @xmath68 , not @xmath11 .
it shows that the deflection angle is independent of the direction of rotation of universe .
but , from equation ( [ ipsi ] ) we find that the integral @xmath69 contains the factor @xmath11 , then the deflection angle @xmath70 for the photon traveling around the lens in two opposite directions is different .the main reason is that the equatorial plan is parallel with gdel rotation plan @xcite .
when @xmath11 vanishes , the deflection angle of @xmath71 tends zero @xcite .
we focus on the deflection angle in the @xmath72 direction , + so we can rewrite the equation ( [ iphi ] ) as , @xmath73 with @xmath74 and @xmath75 where @xmath76 .
the function @xmath77 is regular for all values of @xmath78 and @xmath59 , while @xmath79 diverges as @xmath78 approaches to zero .
therefore , we can split the integral ( [ iz ] ) in two parts , the divergent part @xmath80 and the regular one @xmath81 , which are given by , @xmath82 @xmath83\,dz.\ ] ] here , we expand the argument of the square root in @xmath79 up to the second order in @xmath78 @xcite , @xmath84 where @xmath85,\ ] ] @xmath86.\ ] ] for @xmath87 , @xmath88 is nonzero and the leading order of the divergence in @xmath89 is @xmath90 , which have a finite result .
as @xmath91 , @xmath88 approaches zero and divergence is of order @xmath92 , that makes the integral divergent .
therefor , the deflection angle can be approximated in the following form @xcite , @xmath93 where @xmath94}{u_{ps}\sqrt{\mathcal{a}^3(\rho_{ps})\mathcal{c}(\rho_{ps})}}\,,\nonumber\\ & & b_r = i_r(\rho_{ps}),\,\,\,\,\,\,u_{ps}=\sqrt{\frac{\mathcal{c}(\rho_{ps})}{\mathcal{a}(\rho_{ps})}}\,.\end{aligned}\ ] ] by using ( [ deflection ] ) and ( [ a ] ) , we can investigate the properties of strong gravitational lensing in the charged squashed kaluza- klein gdel black hole .
in this case , variations of the coefficients @xmath0 and @xmath1 , and the deflection angle @xmath95 have been plotted with respect to the extra dimension @xmath26 , charge of the black hole @xmath39 , and gdel parameter @xmath11 in figures 4 - 6 .
+ as @xmath11 tends to zero , these quantities reduce to charged squashed kaluza - klein black hole @xcite and with @xmath96 the squashed kaluza - klein black hole recovers @xcite .
one can see that the deflection angle increases with extra dimension and decreases with @xmath39 . by comparing these parameters with those in four - dimensional schwarzschild and reissner - nordstrm black holes
, we could extract information about the size of extra dimension as well as the charge of the black hole by using strong field gravitational lensing . [ cols="^,^,^,^,^,^,^,^,^,^,^",options="header " , ]
in the previous section , we investigated the strong gravitational lensing by using a simple and reliable logarithmic formula for deflection angle , which was obtain by bozza et al .
now , by using relations between the parameters of the strong gravitational lensing and observables , estimat the position and magnification of the relativistic images . by comparing these observables with the data from the astronomical observation
, we could detect properties of an massive object .
we suppose that the spacetime of the supermassive object at the galaxy center of milky way can be considered as a charged squashed kaluza - klein gdel black hole , then we can estimate the numerical values for observables .
+ we can write the lens equation in strong gravitational lensing , as the source , lens , and observer are highly aligned as follows @xcite , @xmath97 + where @xmath98 is the distance between the lens and source .
@xmath99 is the distance between the observer and the source so that , @xmath100 .
@xmath101 and @xmath52 are the angular position of the source and the image with respect to lens , respectively .
@xmath102 is the offset of deflection angle with integer @xmath103 which indicates the @xmath103-th image .
+ the @xmath103-th image position @xmath104 and the @xmath103-th image magnification @xmath105 can be approximated as follows @xcite , @xmath106 @xmath107 @xmath108 is the angular position of @xmath109 . in the limit @xmath110
, the relation between the minimum of impact parameter @xmath111 and asymptotic position of a set of images @xmath112 can be expressed by @xmath113 . in order to obtain the coefficients @xmath0 and @xmath1 , in the simplest case , we separate the outermost image @xmath114 and all the remaining ones which are packed together at @xmath112 , as done in refs @xcite . thus @xmath115 is considered as the angular separation between the first image and other ones and the ratio of the flux of them is given by , @xmath116 we can simplify the observables and rewrite them in the following form @xcite , @xmath117 thus , by measuring the @xmath118 , @xmath119 and @xmath112 , one can obtain the values of the coefficients @xmath0 , @xmath1 and @xmath120 . if we compare these values by those obtained in the previous section , we could detect the size of the extra dimension , charge of black hole and rotation of universe .
another observable for gravitational lensing is relative magnification of the outermost relativistic image with the other ones .
this observable is shown by @xmath121 which is related to @xmath122 by , @xmath123 using @xmath124 and equations ( [ a ] ) , ( [ sr ] ) and ( [ rm ] ) we can estimate the values of the observable in the strong field gravitational lensing .
the variation of the observables @xmath112 , @xmath118 and @xmath121 are plotted in figures 7 - 9 .
note that the mass of the central object of our galaxy is estimated to be @xmath125 and the distance between the sun and the center of galaxy is @xmath126 @xcite . + for different @xmath26 , @xmath39 and @xmath11 , the numerical values for the observables are listed in table 1 .
one can see that our results reduce to those in the four - dimensional schwarzschild black hole as @xmath127 .
also our results are in agreement with the results of ref .
@xcite in the limit @xmath35 and in the limit @xmath37 , the results of ref .
@xcite are recovered .
the light rays can be deviated from the straight way in the gravitational field as predicted by general relativity .
this deflection of light rays is known as gravitational lensing . in the strong field limit
, the deflection anglethe of the light rays which pass very close to the black hole , becomes so large that , it winds several times around the black hole before appearing at the observer .
therefore the observer would detect two infinite set of faint relativistic images produced on each side of the black hole . on the other hand
the extra dimension is one of the important predictions in the string theory which is believed to be a promising candidate for the unified theory .
also it is reasonable to consider a rotative universe with global rotation .
hence the five - dimensional einstein - maxwell theory with a chern- simons term in string theory predicted five - dimensional charged black holes in the gdel background . in our study , we considered the charged squashed kaluza - klein gdel black hole spacetime and investigated the strong gravitational lensing by this metric .
we obtained theoretically the deflection angle and other parameters of strong gravitational lensing .
finally , we suppose that the supermassive black hole at the galaxy center of milky way can be considered by this spacetime and we estimated numerically the values of observables that are realated to the lensing parameters .
theses observable parameters are @xmath112 , @xmath118 and @xmath128 , where @xmath112 is the position of relativistic images , @xmath118 angular separation between the first image @xmath114 and other ones @xmath112 and @xmath128 is the ratio of the flux from the first image and those from all the other images . our results are presented in figures 1 - 9 and table 1 . by comparatione observable parameters with observational data measured by the astronomical instruments in the future , we can discuss the properties of the massive object in the center of our galaxy .
11 a. einstein , _ science _ , * 84 * , 506 ( 1936 ) .
r. r. bourassa and r. kantowski , _ astrophys .
j_. * 195 * , 13 ( 1975 ) .
y. g. klimov , _ sov .
. doklady _ * 8 * , 119 ( 1964 ) .
s. liebes , jr . , _ phys .
_ * 133 * , b835 ( 1964 ) .
s. refsdal , _ mon . not .
r. astron . soc_. * 128 * , 295 ( 1964 ) .
s. refsdal , _ mon . not .
r. astron . soc_.
* 128 * , 307 ( 1964 ) .
p. schneider , j. ehlers , and e. e. falco `` gravitational lenses '' _ springer - verlag , berlin _ , ( 1992 ) .
r. d. blandford and r. narayan , _ annu .
astron . astrophys_. * 30 * , 311 ( 1992 ) .
s. refsdal and j. surdej , _ rep .
prog . phys_. * 56 * , 117 ( 1994 ) .
r. narayan and m. bartelmann , `` lectures on gravitational lensing '' , ed .
a. dekel and j. p. ostriker ; cambridge university press ( 1995 ) .
j. wambsganss , `` gravitational lensing in astronomy '' , _ living rev .
relativity _ * 1 * , 12 ( 1998 ) . c. darwin , _ proc . of the royal soc . of london
_ * 249 * , 180 ( 1959 ) .
s. chandrasekhar , `` the mathematical theory of black holes '' , _ oxford university press _ , ( 1983 ) .
ohanian , _ amer .
_ 55 , 428 ( 1987 ) .
luminet , _ a and a _ 75 , 228 ( 1979 ) .
viergutz , _ a and a _ * 272 * , 355 ( 1993 ) . j. m. bardeen , `` black holes '' , ed .
c. dewitt and b. s. dewitt , _ gordon and breach _ , * 215 * ( 1973 ) .
h. falcke , f. melia and e. agol , _ apj letters _ * 528 * l13 ( 1999 ) . k. s. virbhadra and g. f. r. ellis , _ phys .
d _ * 62 * , 084003 ( 2000 ) .
k. s. virbhadra , _ phys .
d _ * 79 * , 083004 ( 2009 ) .
s. frittelli , t. p. kling and e. t. newman , _ phys .
d _ * 61 * , 064021 ( 2000 ) .
v. bozza , s. capozziello , g. lovane and g. scarpetta , _ gen .
grav_. * 33 * , 1535 ( 2001 ) .
e. f. eiroa , g. e. romero and d. f. torres , _ phys .
d _ * 66 * , 024010 ( 2002 ) ; e. f. eiroa , _ phys .
d _ * 71 * , 083010 ( 2005 ) ; e. f. eiroa , _ phys .
d _ * 73 * , 043002 ( 2006 ) .
a. o. petters , _ mnras _ * 338 * , 457 ( 2003 ) .
v. bozza , _ phys .
d _ * 66 * , 103001 ( 2002 ) . v. bozza , f. de luca , g. scarpetta and m. sereno , _ phys .
d _ * 72 * , 08300 ( 2005 ) ; v. bozza , _ phys .
d _ * 67 * , 103006 ( 2003 ) .
a. bhadra , _ phys .
d _ * 67 * , 103009 ( 2003 ) .
galin n. gyulchev and ivan zh .
stefanov , _ phys .
d _ 87 , 063005 ( 2013 ) .
h. saadat , _ int j theor phys _ , [ arxiv:1306.0601v1 [ gr - qc ] ] . k. s. virbhadra and g. f. r. ellis , _ phys .
d _ * 65 * , 103004 ( 2002 ) .
k. s. virbhadra and c. r. keeton , _ phys .
* 77 * , 124014 ( 2008 ) .
k. s. virbhadra , d. narasimha and s. m. chitre , _ astron .
astrophys_. * 337 * , 1 - 8 ( 1998 ) . c. j. hogan , r. narayan , _ mnras _ * 2111 * , 575 ( 1984 ) .
a. vilenkin , _ apjl _ * 51 * , 282 ( 1984 ) .
a. vilenkin , _ nat _ * 322 * , 613 ( 1986 ) .
r. emparan and h. s. reall , _ living rev . relativity _ * 11 * , 6 ( 2008 ) .
s. dimopoulos and g. landsberg , _ phys .
_ * 87 * , 161602 ( 2001 ) .
s. b. giddings and s. thomas , _ phys .
. d _ * 65 * , 056010 ( 2002 ) .
n. arkani - hamed , s. dimopoulos and g. r. dvali , _ phys .
b _ * 429 * , 263 ( 1998 ) . p. c. argyres , s. dimopoulos , and j. march - russell , _ phys .
b _ * 441 * , 96 ( 1998 ) .
j.l . feng and a.d .
shapere , _ phys .
rev . lett_. * 88 * , 021303 ( 2002 ) .
l. a. anchordoqui , j.l .
feng , h. goldberg , and a.d .
shapere , _ phys .
d _ * 65 * , 124027 ( 2002 ) .
s. s. seahra , c. clarkson and r. maartens , _ phys .
rev . lett_. * 94 * , 121302 ( 2005 ) . m. gunaydin , g. sierra and p. k. townsend , _ nucl .
phys . b _ * 253 * , 573 ( 1985 ) .
r. c. myers , _ phys .
d _ * 35 * , 455 ( 1987 ) .
j. y. shen , b. wang and r. k. su , _ phys .
d _ * 74 * , 044036 ( 2006 ) .
e. abdalla , b. cuadros - melgar , a. b. pavan and c. molina , _ nucl .
b _ * 752 * , 40 ( 2006 ) . s. b. chen , b. wang and r. k. su , _ phys . lett .
b _ * 647 * , 282 ( 2007 ) ; p. kanti , r. a. konoplya and a. zhidenko , _ phys . rev
. d _ * 74 * , 064008 ( 2006 ) .
p. kanti , r. a. konoplya , _ phys .
d _ * 73 * , 044002 ( 2006 ) ; d. k. park , _ phys . lett .
b _ * 633 * , 613 ( 2006 ) .
j. sadeghi , a. banijamali , m. r. setare and h. vaez , _ j. int
* 26 * , 2783 ( 2011 ) .
c. m. harris and p. kanti , _ jhep _ * 0310 * 014 ( 2003 ) .
p. kanti , _ j. int .
* 19 * 4899 ( 2004 ) .
argyres , s. dimopoulos and j. march - russell , _ phys .
b _ * 441 * 96 ( 1998 ) .
r. emparan , g. t. horowitz and r. c. myers , _ phys .
rev . lett_.
* 85 * 499 ( 2000 ) .
h. ishihara and k. matsuno , prog .
116 , 417 ( 2006 ) .
h. ishihara and j. soda , _ phys .
d _ * 76 * , 064022 ( 2007 ) .
s. chen , b. wang and r. su , _ phys .
* 77 * , 024039 ( 2008 ) .
x. he , b. wang and s. chen , _ phys .
d _ * 79 * , 084005 ( 2009 ) ; x. he , s. chen , b. wang , r. g. cai and c. lin , _ phys .
b _ * 665 * , 392 ( 2008 ) . h. ishihara , m. kimura , r. a. konoplya , k. murata , j. soda and a. zhidenko , _ phys
d _ * 77 * , 084019 ( 2008 ) .
y. liu , s. chen and j. jing , _ phys .
d _ * 81 * , 124017 ( 2010 ) .
s. chen , y. liu and j. jing , _ phys .
d _ * 83 * , 124019 ( 2011 ) .
j. sadeghi , a. banijamali and h. vaez , _ astrophys . space .
sci _ * 343 * , 559 ( 2013 ) .
gdel , _ rev .
mod . phys_.
* 21 * , 447 ( 1949 ) .
j. p. gauntlett , j. b. gutowski , c. m. hull , s. pakis , and h. s. reall , classical quantum gravity 20 , 4587 ( 2003 ) .
s. q. wu , _ phys .
lett _ * 100 * , 121301 ( 2008 ) .
e. g. gimon and a. hashimoto , _ phys .
rev . lett_. * 91 * , 021601 ( 2003 ) .
m. cvetic , h. lu , and c. n. pope , _ phys .
b _ * 598 * , 273 ( 2004 ) . c. a. r. herdeiro , nucl , phys .
b 665 , 189 ( 2003);c .
a. r. herdeiro , classical quantum gravity 20 , 4891 ( 2003 ) .
r. kerner and r. v. mann , _ phys .
d _ * 75 * , 084022 ( 2007 ) .
j. c. breckenridge , r. c. myers , a. w. peet and c. vafa , _ phys .
b _ * 391 * , 93 ( 1997 ) .
h. k. kunduri and j. lucietti , _ nucl
b _ * 724 * , 343 clarissa - marie claudel , k. s. virbhadra , g. f. r. ellis , _
j.math.phys _ * 42 * , 818 ( 2001 ) .
|
in this paper we investigate the strong gravitational lansing in a charged squashed kaluza - klein gdel black hole .
the deflection angle is considered by the logarithmic term proposed by bozza et al .
then we study the variation of deflection angle and its parameters @xmath0 and @xmath1 .
we suppose that the supermassive black hole in the galaxy center can be considered by a charged squashed kaluza - klein black hole in a gdel background and by relation between lensing parameters and observables , we estimate the observables for different values of charge , extra dimension and gdel parameters .
+ + * pacs numbers : * 95.30.sf , 04.70.-s , 98.62.sb + * keywords : * gravitational lensing ; charged squashed kaluza - klein gdel black hole +
|
the radius of the proton - proton interaction can be measured in two different ways : as the @xmath5-slope of the elastic @xmath0-scattering or via the width of the peak in identical particle correlations . at first sight
, these two methods do not seem related to each other . in this paper
, we use a regge approach to trace , and to exploit , the connection .
the _ universal _ shrinkage of the diffraction cone in elastic @xmath0- , @xmath6- and @xmath7-scattering , driven by the pomeron , was already discovered at fixed - target cern - sps energies , starting with the cern - gatchina experiment @xcite .
the behaviour of elastic scattering at high collider energies is well described by a parametrization in which the elastic slope logarithmically grows with energy b_el(s ) = b_0 + 2^_p(s / s_0 ) . for multiparticle production in proton - proton collisions one might expect that the size of the emission source , @xmath8 , in high - energy inelastic interactions also increases with the collision energy . moreover , within regge theory ( pomeron approach ) we expect _ universal _ behaviour of this radius . to measure the value of @xmath8 one may use the bose - einstein correlations ( bec ) between two identical pions .
the high statistics of the lhc allows the space - time structure of high - energy interactions to be studied in detail .
the standard way to determine the size of the interaction region is to measure the width of the peak at low @xmath9 in bec @xcite of two identical pions emitted with four momenta @xmath10 and @xmath11 , where @xmath12 .
that is , the lorentz - invariant quantity @xmath9 gives a measure of the proximity of the two identical particles in phase space .
the first lhc data showing the bec enhancements at low @xmath9 have been reported by cms for @xmath0 collisions @xcite and by alice for @xmath0 and for pb - pb collisions @xcite .
the data are presented as the ratio of the @xmath9 distribution for pairs of identical particles to a reference distribution for pairs of particles which are expected to have no bec effect r(q ) = ( dn / dq)/(dn_ref / dq ) .
the peak in the @xmath13 distribution is then fitted to the exponential @xmath14 ( in cms case ) or to the gaussian @xmath15 ( alice ) parametric form , where the size of the pion - pair emission region is characterised by an effective radius @xmath8 .
here , we discuss the bec effect in terms of the microscopic interaction , rather than in terms of collective thermodynamic variables . for high - energy @xmath0 interactions , the total and elastic cross sections are well described in terms of regge theory by an effective pomeron pole an universal object with its own internal structure which does not depend on the energy , @xmath16 , of the colliding protons .
thus in the one - pomeron exchange approach we do not expect the spectra and density of secondaries in the central region , @xmath17 to depend on @xmath18 .
however , this naive expectation contradicts the recent lhc data .
in particular , the height of the central plateau in @xmath19 is observed to grow by more than 60@xmath20 in going from a collider energy of 0.9 tev to 7 tev @xcite . within regge field theory this growth
is explained by mutli - pomeron exchange contributions .
besides one - pomeron , there may be contributions from two or more pomeron exchanges .
the mean number of pomeron exchanges increases with energy , since at a higher energy we have a larger total cross section and , thus , a larger probability of an additional rescattering , or additional multiple interactions , described by the exchange of additional pomerons .
this leads to the growth of the multiplicity @xmath21 . in the regge approach
each soft pomeron can be regarded as a multiperipheral ladder , see fig .
[ fig : abc ] .
the ladder shown in ( a ) corresponds to the elastic @xmath0 scattering amplitude , while on cutting the pomeron , as in ( b ) , we obtain the cross section for multiparticle production .
cutting @xmath22 pomerons in a multi - pomeron exchange diagram , ( c ) , gives a final density of secondaries in the central region that is @xmath22 times larger than that for one - pomeron exchange , see , for example , @xcite for more details .
it is observed that the mean radius @xmath23 , measured by identical particle correlations without multiplicity selection , grows with energy .
this is naturally explained within the multi - pomeron picture .
indeed , if we select events with low multiplicity then we deal with a process described by one - pomeron exchange . in this case , the width of the correlation peak measures the size of the pion source originating from one individual pomeron . on the contrary , in high multiplicity events ,
combinatorial factors mean that the two identical pions are dominantly emitted from different pomerons .
then the value of @xmath23 reflects the distance between the pomerons .
thus we expect @xmath23 will increase with the charged particle multiplicity , @xmath24 , while the energy behaviour is explained by the growth of the number of exchanged pomerons ( and thus the number of secondaries ) with @xmath16 .
indeed , the data @xcite show this kind of behaviour .
we emphasize that the values of @xmath23 measured at quite different collider energies , 0.9 tev to 7 tev , are , to good accuracy , the same for the same value of @xmath24 , that is , for the contribution from the same number of pomeron exchanges .
such ` scaling ' behaviour ( in which the value of @xmath23 does not depend on the proton energy , but on @xmath24 only ) confirms the interpretation based on one , or more , pomerons exchanged .
thus the smallest radius should be considered as the radius of one individual pomeron ( or , to be more precise , as the size of the source formed by a single pomeron ) . in other words ,
identical pion correlations at low @xmath24 allow the measurement of the size of an individual pomeron .
the same effect has been observed in heavy - ion collisions @xcite .
the value of @xmath23 measured in events with a given particle density , @xmath21 , turns out to be practically constant starting from very low fixed - target energies right up to the lhc energy @xmath25=2.76 tev of the colliding constituent nucleons . in the case of heavy - ion collisions
the final multiplicity depends on the centrality of the events . in other words , @xmath21 is proportional to the number of @xmath26 collisions . for a very peripheral interaction , only one pair of nucleons
collides , whereas for a central collision the discs of the heavy ions overlap completely , leading to the simultaneous interactions of many pairs of nucleons .
let us return to @xmath0 collisions , and recall the partonic structure of the proton . in a @xmath0 collision
the parton - parton interaction is analogous to a @xmath26 collision in heavy - ion collisions .
thus we expect @xmath23 to saturate at the value of @xmath24 when the partonic discs of the colliding protons overlap , that is at a value of @xmath23 characteristic of the @xmath0 interaction .
let us compare the value of @xmath27 fm observed for large @xmath24 with that obtained from the @xmath5-slope of the elastic @xmath0 cross section .
the slope expected at 7 tev is about @xmath28 .
now the relation between the slope and the radius is - slope , @xmath29 , of the elastic amplitude , we deal with a two - dimensional transverse vector .
therefore we use @xmath30 instead of @xmath31 . ]
b_el/2 = r^2 /4 gives r^2 ~1.55 fm^2 .
however , the above relation is obtained with a gaussian form @xmath32 , while the correlation is observed @xcite to be better described by the exponential form @xmath33 .
the relation between @xmath23 for the two forms is _ exp = _ gaussian .
thus we obtain @xmath34 fm close to the value found in the bec data @xcite .
that is , we have correspondence between the value of @xmath23 at large @xmath24 and the interaction radius determined via the elastic slope @xmath35 . since @xmath35 increases with energy , we therefore expect that the ` saturated ' bec value of @xmath23 will also increase a little . to conclude our discussion so far , we see that the bec effects at the lhc offer the opportunity to confirm the universal small - size structure expected for the pomeron , is much smaller than the slope @xmath36 observed for the trajectories of the secondary reggeons , like @xmath37 ; recall that the value of @xmath38 is proportional to the square of the transverse size of the corresponding reggeon .
next , the success of the additive quark model ( @xmath39 ) is easy to explain when the radius of quark - quark interaction is much smaller than the hadron ( pion , proton ) radius .
however now we have the first _
direct _ measurement of the size of the pomeron .
this result should not be affected by possible pomeron - pomeron interactions .
the triple - regge analysis indicates a very small radius of multi - pomeron vertices , so the multi - pomeron interactions do not noticeably change the radius of the pion source . ] and to measure both the size of individual pomeron ( at comparatively small @xmath24 ) and the radius of the proton - proton interaction ( at large @xmath24 ) .
we sketch the expected behaviour of the average radius of the identical particle emission region , @xmath23 with @xmath24 in fig .
[ fig : r ] . at low @xmath24
we measure the size of the pion source created by one pomeron , while at larger @xmath24 the result is mainly driven by the spatial separation between two pomeron exchanges .
an important result of the observation of bec is that at low @xmath24 the size of the source does not depend on energy , within the experimental accuracy .
this confirms the universal structure of the pomeron pole . at large @xmath24
we expect saturation of @xmath23 at a value corresponding to the radius , @xmath40 , of the @xmath0 interaction .
the corresponding plateau in the @xmath24-plot should therefore increase slowly with energy , as @xmath41 . as a function of the multiplicity of charged particles , @xmath24 .
the size @xmath23 of the source of identical pions originating from different mechanisms is indicated .
the continuous and dashed curves correspond to , say , @xmath42 7 and 0.9 tev respectively . at low @xmath24
we expect @xmath23 to be independent of the collider energy @xmath16 , while for the plateau we expect @xmath43 to increase very slowly with energy .
very large multiplicities are expected to arise from high-@xmath44 events originating from a single ladder .
the scale for the @xmath24 axis depends on the experimental acceptance domain . here
, we indicate an approximate typical scale for @xmath24 corresponding to a relevant experimental @xmath45 interval , @xmath46 , and @xmath47 cut . ,
height=226 ] there is the possibility that very large multiplicity events will arise dominantly from high @xmath44 production , see point ( iii ) of section [ sec : a5 ] . a high-@xmath44 process will mainly originate from one - ladder exchange , that is from one pomeron cross sections decrease steeply with increasing energy , and it is better not to distribute the energy among a few pomeron exchanges . ] .
thus at very large @xmath24 we expect that the value of @xmath23 will decrease back towards the value due to the size of one pomeron ( or one jet ) ; that is , close to the value of @xmath23 at low @xmath24 .
so far our discussion has been idealistic .
the actual situation is more complex .
there are background sources which may complicate the interpretation of the bec measurements .
we divide these into three classes .
\(i ) let us start with @xmath23 at small @xmath24 .
in the ideal case , this is just the radius of an individual pomeron .
unfortunately here the radius is not well defined . in terms of the quark and gluon degrees of freedom
the pomeron , may be an almost point - like object with @xmath48
. however we need space to form a hadron ( pion ) , which has its own size .
in particular , in @xmath49 annihilations , where the interaction starts from the point - like production of a @xmath50 pair , an analogous radius is observed to be @xmath51 fm , depending on the form of the fit @xcite .
this is comparable to the radius , @xmath52 , of the pion .
thus , even for a point - like pomeron , we expect hadronization will give a non - zero @xmath53 for small @xmath24 .
\(ii ) the next problem is that an appreciable fraction of pions is produced via resonance production and decay , rather than directly . for a long living resonance ( with a small width @xmath54 ) , the distance between the creation and the decay of the resonance , @xmath55 , may be rather large leading to a larger @xmath23 .
the contribution , via resonances , is smaller for pions of larger transverse momentum @xmath56 .
indeed , there is a tendency for @xmath23 to decrease with @xmath56 @xcite .
\(iii ) on the other hand for large @xmath56 , we have another ` background ' problem to consider . actually the value of @xmath24 measured at the lhc is not for the whole event . instead , it is the number of tracks observed in the restricted region covered by the central tracking detectors .
so by including the exchange of an additional pomeron we only enlarge @xmath24 by some fixed amount .
as we have mentioned in the previous section , high @xmath44-jet production is another possible origin of large @xmath24 events .
now jet multiplicity grows exponentially @xcite , @xmath57 , and the resulting particles are dominantly observed in the central detectors .
thus for very large @xmath24 , there may be competition between contributions from high-@xmath44 jets and soft multi - pomeron interactions .
this background can be quantified by selecting events with one high @xmath44 particle and looking for bec between two soft pions coming from the underlying event . in this way we may check whether the soft underlying event , which accompanies the hard process , has the same properties as normal soft multiparticle production .
if so , we may correct the data by eliminating the jet events .
it looks , however , that the observed saturation effect already occurs at not too large values of @xmath24 , where there is little jet contamination , see fig .
[ fig : r ] .
thanks to the large luminosity and enormous statistics available at the lhc , we now have the possibility to study the size of the interaction volume , which produces the identical pion pair , under different kinematic conditions .
in particular , we may measure : * @xmath23 via the correlations of pions with , say , @xmath58 as a function of the multiplicity observed in the other hemisphere with @xmath59 . in this way we may check whether we deal with one source of secondaries or whether there are a few independent ` fireballs ' ( that is , preconfinement clusters ) .
* @xmath23 in the events with a high @xmath44 jet ( or one high-@xmath47 hadron ) .
does high @xmath44 select the small - size component of the incoming proton ?
* @xmath23 in the events with a large rapidity gap .
a small gap survival probability @xmath60 pushes the amplitude to the periphery , where there is a smaller probability of rescattering ( and of an extra pomeron exchange ) ; for a recent review see @xcite . *
@xmath23 as a function of the rapidity and/or the transverse momentum of the pion pair .
* @xmath23 in the correlations of kaons ( @xmath61 ) ; and to compare with radius of the pion source .
mgr would like to thank the ippp at the university of durham for hospitality ; and the federal program of the russian state rsgss-65751.2010.2 for support of grant rfbr 11 - 02 - 00120-a .
r. hanbury - brown and r.w .
twiss , phil . mag . * 45 * , 663 ( 1954 ) ; proc .
242a * , 300 ( 1957 ) ; _ ibid _ * 243a * , 291 ( 1957 ) .
g. goldhaber , w.b .
fowler , s. goldhaber et al .
rev . lett . * 3 * , 181 ( 1959 ) .
kopylov and m.i .
podgoretskii , sov . j. nucl
15 * , 219 ( 1972 ) ; * 18 * , 336 ( 1973 ) .
g. alexander , rep .
. phys . * 66 * , 481 ( 2003 ) .
cms collaboration : v. khachatryan et al .
, arxiv:1101.3518 .
alice collaboration : k. aamodt et al .
, arxiv:1101.3665 .
alice collaboration : k. aamodt et al .
, arxiv:1007.0516 .
alice collaboration : k. aamodt et al .
, arxiv:1012.4035 .
cms collaboration : v. khachatryan et al .
lett . * 105 * , 022002 ( 2010 ) .
atlas collaboration : georges aad et al .
, arxiv:1012.0791 .
ryskin , a.d .
martin , v.a .
khoze and a.g .
shuvaev , j. phys .
g. * 36 * , 093001 ( 2009 ) .
see , for example , p. abreu et al .
b471 * , 460 ( 2000 ) .
dokshitzer , v.a .
khoze , a.h .
mueller and s.i .
troian , ` basics of perturbative qcd ' , _ gif - sur - yvette , france : ed .
frontieres ( 1991 ) 274 p. _ a.d .
martin , m.g .
ryskin and v.a .
khoze , acta phys .
* b40 * , 1841 ( 2009 ) [ arxiv:0903.2980 ] .
|
bose - einstein correlations of identical pions produced in high - energy @xmath0 collisions at the lhc allow a probe of the pomeron exchange mechanism .
the size of the domain which emits the pions depends on the multiplicity of events , but not on the collider energy .
this confirms the universal structure of pomeron exchange .
the data at relatively low multiplicities indicate that the size of the source created by one - pomeron exchange is much less than the size of the proton .
ippp/11/05 + dcpt/11/10 + + * pomeron universality from identical pion correlations * + * at the lhc * v.a .
schegelsky@xmath1 , a.d .
martin@xmath2 , m.g .
ryskin@xmath3 and v.a .
khoze@xmath3 + @xmath4 petersburg nuclear physics institute , gatchina , st . petersburg , 188300 , russia + @xmath2 institute for particle physics phenomenology , university of durham , durham , dh1 3le +
|
lattice monte carlo ( lmc ) computer simulations are often used to study diffusion problems when it is not possible to solve the diffusion equation .
if the lattice mesh size is small enough , lmc simulations provide results that are in principle arbitrarily close to the numerical solution of the diffusion equation . in lmc simulations
, a particle is essentially making an unbiased random - walk on connected lattice sites , and those moves that collide with obstacles are rejected @xcite .
the allowed monte carlo moves are usually displacements by one lattice site along one of the @xmath0 spatial directions . in the presence of an external field
, one must bias the possible lattice jumps in order to also reproduce the net velocity of the particle .
however , this is not as easy as it looks because one must also make sure that the diffusion coefficient is correctly modelled along each of the @xmath0 spatial directions . using a metropolis weighting factor @xcite does not work because in the limit of large driving fields , all the jumps along the field axis are in the same direction and hence the the velocity saturates and the diffusion coefficient in this direction vanishes
this approach is thus limited to weak fields , at best .
a better approach is to solve the local diffusion problem ( i.e. , inside each lattice cell ) using a first - passage problem ( fpp ) @xcite approach , and to use the corresponding probabilities and mean jumping times for the coarser grained lmc moves . in this case
, the mean jumping times are shorter along the field axis , but one can easily renormalize the jumping probabilities to use a single time step . in a recent paper @xcite , we demonstrated that although this method does give the correct drift velocity for arbitrary values of the driving field , it fails to give the correct diffusion coefficient .
the problem is due to the often neglected fact that the variance of the jumping time affects the diffusion process in the presence of a net drift @xcite .
lmc models do not generally include these temporal fluctuations of the jumping time , at least not in an explicit way . in the same article @xcite , we showed how to modify a one - dimensional lmc algorithm with the addition of a stochastic jumping time @xmath1 , where the appropriate value of the standard - deviation @xmath2 was again obtained from the resolution of the local fpp . for simulations in higher spatial dimensions @xmath3 ,
it is possible to use our one - dimensional algorithm with the proper method to alternate between the dimensions as long as the monte carlo clock advances only when the particle moves along the field direction @xcite .
lmc simulations of diffusion processes actually use stochastic methods to resolve a discrete problem that can be written in terms of coupled linear equations .
several years ago , we proposed a way to compute the exact solution of the lmc simulations via matrix methods , thus bypassing the need for actual simulations .
this alternative method is valid only in the limit of vanishingly weak driving fields , but it produces numerical results with arbitrarily high precision . the crucial requirement of the method is a set of lmc moves that have a common jumping time .
dorfman @xcite suggested a slightly different but still exact numerical method , and the two agree perfectly at zero - field .
more recently @xcite , we extended our numerical method to cases with driving fields of arbitrary magnitudes ; in order to do that , we used lmc moves that possess a single jumping time for all spatial directions , but this forced us to neglect the temporal fluctuations discussed above . as a consequence , our numerical method generates exact velocities but fails to provide reliable diffusion coefficients . again , dorfman s alternate method also give the same velocities , but because the lmc moves do not include the proper temporal fluctuations , neither method can be used to compute the diffusion coefficient along the field axis . in summary , a fixed - time lmc algorithm can be used with exact numerical methods to compute the net velocity , but temporal fluctuations ( and hence computer simulations ) must be used to compute the diffusion coefficient .
we recently solved the problem of defining a lmc algorithm with both a fixed time step and the proper temporal fluctuations @xcite .
this required the addition of a probability to stay put on the current lattice site during a given time step ( of course , this change also implies a renormalization of the jumping probabilities ) .
this probability of non - motion has a direct effect on the real time elapsed between two displacements of the brownian walker , and this effect can be adjusted in order to reproduce the exact temporal fluctuations of the local fpp .
we showed that this new lmc algorithm can be used with dorfman s exact numerical method to compute the exact field - dependence of both the velocity and the diffusion coefficient of a particle on a lattice in the presence of obstacles .
as far as we know , this is the first biased lattice random - walk model that gives the right diffusion coefficient for arbitrary values of the external field .
other models , such as the repton model @xcite , are restricted to weak fields .
several other articles ( see , e.g. @xcite ) report simulations of diffusive processes , but all of them appear to be limited to small biases .
unfortunately , our lmc algorithm @xcite has a fatal flaw : for dimensions @xmath4 , some of the jumping probabilities turn out to be negative .
this failure suggests that there is a fundamental problem with this class of models , or more precisely with standard lmc moves ( however , note that it is still possible to use computer simulations and fluctuating jumping times @xmath1 , as explained above ) . in other words , it is impossible to get both the right velocity and the right diffusion coefficient in all spatial directions ( if @xmath4 ) when the lmc jumps are made along a single axis at each step . in this article
, we examine an alternative to the standard lmc moves in order to derive a valid lmc algorithm with a common time step for spatial dimensions @xmath4 .
we suggest that a valid set of lmc moves should respect the fact that motion along the different spatial directions is actually simultaneous and not sequential .
as we will show , this resolves the problem and allows us to design a powerful new lmc algorithm that can be used both with exact numerical methods and stochastic computer simulations .
as mentioned above , metropolis - like algorithms are not reliable if one wants to study diffusion via the dynamics of biased random - walkers on a lattice @xcite .
the discretization of such continuous diffusion processes should be done by first solving the fpp of a particle between two absorbing walls ( the distance between these arbitrary walls is the step size @xmath5 of the lattice ) . indeed
, completion of a lmc jump is identical to the the first passage at a distance @xmath5 from the origin . in one dimension , this fpp has an exact algebraic solution , and the resulting transition probabilities ( noted @xmath6 for parallel and antiparallel to the external force @xmath7 ) are @xcite : @xmath8 where @xmath9 is the ( scaled ) external field intensity , @xmath10 is boltzmann s constant and @xmath11 is the temperature .
the time duration of these fpp jumps is @xcite : @xmath12 where @xmath13 , the time duration @xmath14 of a jump when no external field is applied , is called the brownian time .
although eqs .
[ e : p1d ] and [ e : tau1d ] can be used to simulate one - dimensional drift problems ( the net velocity is then correct ) , they erroneously generate a field - dependent diffusion coefficient for a free particle , which is wrong .
this failure is due to the lack of temporal fluctuations in such a lmc algorithm ( at each step , the particle would jump either forward ( @xmath15 ) or backward ( @xmath16 ) , and all jumps would take the same time @xmath17 ) .
as mentioned above , it is possible to fix this problem @xcite with a stochastic time step like @xmath18 where @xmath19 can also be calculated exactly within the framework of fpp s @xcite : @xmath20 however , the resulting algorithm can only be used in monte carlo computer simulations because exact resolution methods @xcite require a common time step for all jumps .
alternatively , temporal fluctuations can be introduced using a probability @xmath21 to remain on the same lattice site during the duration of a fixed time step @xmath22 @xcite .
not moving has for effect to create a dispersion of the time elapsed between two actual jumps . in order to obtain the right free - solution diffusion coefficient
, we must have @xcite : @xmath23 this modification also forces us to renormalize the other elements of the lmc algorithm : @xmath24 @xmath25 equations [ e : ss ] to [ e : tt ] define a lmc algorithm that can be used with monte carlo simulations ( or exact numerical methods ) to study one - dimensional drift and diffusion problems .
one can easily verify @xcite that it leads to the proper free - solution velocity ( @xmath26 ) and diffusion coefficient ( @xmath27 ) , while satisfying the nernst - einstein relation @xmath28 .
these equations will thus be the starting point of our new multidimensional lmc algorithm .
in principle , we can build a simple model for @xmath3 dimensions using the elements of a one - dimensional biased random walk for the field axis and those of an unbiased random - walk for each of the @xmath29 transverse axes .
indeed , it is possible to fully decouple the motion along the different spatial directions if the field is along a cartesian axis .
such an algorithm is divided into three steps : 1 .
first , we must select the jump axis , keeping in mind that the particle should share its walking time equally between the @xmath0 spatial directions . the probability to choose a given axis
should thus be inversely proportional to the mean time duration of a jump in this direction ( note that the time duration of a jump is shorter in the field direction ) .
2 . secondly , the direction ( @xmath6 ) of the jump must be selected .
3 . finally , the time duration of the jump must be computed and the monte carlo clock must be advanced .
there are several ways to implement these steps .
the easiest way is to use eqs .
[ e : p1d ] to [ e : dtau ] ; in this case , the lmc clock must advance by a stochastic increment @xmath30 each time a jump is made along the field axis ( in order to obtain the proper temporal fluctuations , the clock does not advance otherwise ) .
a slightly more complicated way would be to use eqs .
[ e : ss ] to [ e : tt ] ; again , the clock advances only when the jump is along the field axis , but this choice has the advantage of not needing a stochastic time increment . although both of these implementations can easily be used with computer simulations , they would not function with exact numerical methods because of the way the clock is handled . for exact numerical methods ,
an algorithm with a common time step and a common clock for all spatial directions is required .
we showed that it is indeed possible to do this if we renormalize eqs .
[ e : p1d ] and [ e : tau1d ] properly @xcite ; this approach works for any dimension @xmath3 , but it can only be used to compute the exact velocity of the particle since it neglects the temporal fluctuations . in order to also include these fluctuations , one must start from eqs .
[ e : ss ] to [ e : tt ] instead .
unfortunately , this can be done only in two dimensions since the renormalization process gives negative probabilities when @xmath4 @xcite .
clearly , in order to derive a multi - dimensional lmc algorithm with a fixed time - step , a common clock and the proper temporal fluctuations , we need a major change to the basic assumptions of the lmc methodology . in the next section ,
we propose to allow simultaneous jumps in all spatial directions .
this is a natural choice since lmc methods do indeed assume that the motion of the particle is made of @xmath0 entirely decoupled random - walks .
current lmc methods assume this decoupling to be valid , but force the jumps to be sequential and not simultaneous .
in our multi - dimensional algorithm @xcite , the lmc moves were the standard unit jumps along one of the cartesian axes , and a probability to stay put was used to generate temporal fluctuations . since moving along a given axis actually contributes to temporal fluctuations along all the other axes @xcite , the method fails for @xmath4 because the transverse axes then provide an excess of temporal fluctuations .
this strongly suggests that the traditional sequential lmc moves are the culprit .
sequential lmc moves are used solely for the sake of simplicity , but they are a poor representation of the fact that real particles move in all spatial directions at the same time .
this weakness is insignificant for unbiased diffusion , but it becomes a roadblock in the presence of strong driving fields . in order to resolve this problem , we thus suggest to employ a set of moves that respect the simultaneous nature of the dynamics along each of the @xmath0 axes . to generate a lmc algorithm for this new set of moves
, we will use our exact solution of the one - dimensional problem for each of the @xmath0 directions .
our new lmc moves will include one jump attempt along each of @xmath0 spatial directions .
the list will thus consist of @xmath31 different moves since we must allow for all possible permutations of the three fundamental jumps ( of length @xmath32 and @xmath33 ) used by the exact one - dimensional model that we will be using for each axis .
note that the external field must be parallel to one the cartesian axes ( we choose the @xmath34-axis here ) .
the dynamics is governed by @xmath35 , @xmath21 and @xmath22 in the @xmath34-direction ( eqs .
[ e : ss ] to [ e : tt ] ) , whereas we can in principle use @xmath36 and @xmath13 for the transverse directions because there is no need to model the temporal fluctuations when there is no net drift in the given direction @xcite .
the optimal time step for our new moves is @xmath37 , the duration of the fastest unit process .
we thus have to rescale the transverse probability @xmath38 accordingly : @xmath39 this generates an arbitrary probability to stay put in the transverse directions : @xmath40 in the zero - field limit , this probability gives : @xmath41 therefore , the probability to stay put is the same in all the directions in this limit , as it should . in the opposite limit @xmath42 , we have : @xmath43 and the jumps in the transverse directions become extremely rare , as expected .
equations [ e : ss ] to [ e : sy ] are sufficient to build the table of multi - dimensional moves and their different probabilities since the @xmath0 directions are independent .
figure [ f : schema ] illustrates the new lmc moves for the @xmath44 and @xmath45 cases in the absence of obstacles .
the moves , all of duration @xmath22 , combine @xmath0 simultaneous one - dimensional processes and include net displacements along lattice diagonals .
the @xmath46 paths are further defined in table [ t : free]a ; such a description of the trajectories will be essential later to determine the dynamics in the presence of obstacles .
it is straightforward to extend this approach to higher dimensions ( @xmath4 ) .
we can easily verify that this new set of lmc moves gives the right free - solution velocity and diffusion coefficients for all dimensions @xmath47 . if the field is pointing along the @xmath34-axis ,
the average displacement per time step is @xmath48 , while the average square displacement is @xmath49 . using these results
, we can compute the free - solution velocity @xmath50 and diffusion coefficient @xmath51 : @xmath52 and @xmath53 one can also verify that @xmath54 and @xmath55 .
these are precisely the results that we expect .
therefore , the model introduced here does work for all values of the external field @xmath56 and all dimensions @xmath57 in the absence of obstacles .
the problems faced in ref .
@xcite have been resolved by making the @xmath0 directions truly independent from each other and choosing @xmath22 as the fundamental time step of the new lmc moves .
since this new model works fine in free - solution , the next step is to define how to deal with the presence of obstacles .
the rule that we follow in those cases where a move leads to a collision with an obstacle is the same as before , i.e. , such a jump is rejected and the particle remains on the same site . in our algorithm , though , this means that one ( or more ) of the @xmath0 sub - components of a @xmath0-dimensional move is rejected .
therefore , the list of transition probabilities must take into account all of the possible paths that the particle can follow given the local geometry .
a two - dimensional example is illustrated in table [ t : free]b .
we see that the two different trajectories that previously lead to the upper right corner ( site @xmath58 ) now lead to different final positions due to the rejection of one of the two unit jumps that are involved .
the final transition probabilities for this particular case are listed in table [ t : free]b .
of course , all local distributions of obstacles can be studied using the same systematic approach .
in order to test our new set of lmc moves for systems with obstacles , we will compare its predictions to those of our previous two - dimensional algorithm @xcite since we know that both can properly reproduce the velocity and the diffusion coefficient of a particle in the case of an obstacle - free system .
however , the different moves used by these two algorithms means that a true comparison can only be made in the limit of the continuum since the choice of moves always affects the result of a coarse - grained approach if there are obstacles .
the exact numerical method that we developed in collaboration with dorfman @xcite is not limited to the previous set of lmc moves .
it can easily be modified to include other lmc moves , including diagonal moves . combining dorfman s method @xcite and our new lmc moves
, we now have a way to compute the exact velocity and the exact diffusion coefficient of a particle in the presence of arbitrary driving field for any dimension @xmath59 .
we thus studied the system shown in fig .
[ f : continuum]b using both algorithms , and we repeated the calculation for different lattice parameters @xmath60 ( with @xmath61 ) while the obstacle size ( @xmath5 ) remained constant ( the surface concentration of obstacles is thus kept constant at @xmath62 ) .
the limit of the continuum corresponds to @xmath63 .
we compared the velocities and diffusion coefficients along the field - axis obtained with both algorithms over a wide range of @xmath64 .
note that the value of the external scaled field @xmath56 , which is proportional to the lattice parameter ( @xmath65 ) , has to be rescaled by the factor @xmath64 .
figure [ f : continuum]a presents the data for both algorithms for a nominal field intensity @xmath66 .
we clearly see that the two approaches converge perfectly in the @xmath67 limit .
interestingly , the new algorithm converges slightly faster towards the asymptotic continuum value .
this is explained by the fact that the diagonal transitions reduce the number of successive collisions made by a random - walker when it is trapped behind an obstacle at high field .
conventional three - dimensional lmc algorithms can not be used to study both the mean velocity and the diffusion coefficient of a brownian particle if the time step has to be constant ( as required by exact numerical methods ) .
this limitation is due to the fact that these algorithms only allow jumps to be made along one axis at each time step .
such unit jumps make it impossible to obtain the proper temporal fluctuations that are key to getting the right diffusion coefficient .
we propose that lmc moves should actually respect the fact that all of the @xmath0 spatial dimensions are fully independent .
this means that each move should include a component along each of these dimensions .
this complete dimensional decoupling allows us to conserve the proper temporal fluctuations and hence to reproduce the correct diffusion process even in the presence of an external field of arbitrary amplitude .
this approach leads to a slightly more complicated analysis of particle - obstacle collisions , but this is still compatible with the exact numerical methods developed elsewhere @xcite .
the new lmc algorithm presented in this paper opens the door to numerous coarse - grained simulation and numerical studies that were not possible before because previous algorithms were restricted to low field intensities .
this work was supported by a discovery grant from the natural science and engineering research council ( _ nserc _ ) of canada to gws .
mgg was supported by a _
scholarship , an excellence scholarship from the university of ottawa and a strategic areas of development ( sad ) scholarship from the university of ottawa .
( squares ) and diffusion coefficient @xmath68 ( circles ) vs the mesh size @xmath64 for @xmath66 .
these calculations were done using the algorithm presented in ref .
@xcite ( filled symbols ) and the one proposed in this paper ( empty symbols ) .
( b ) the obstacle is of size @xmath69 , the lattice is of size @xmath70 ( with periodic boundary conditions ) , and the particle ( not shown ) is of size @xmath71 . the system is shown for three different values of the mesh size parameter @xmath64 . , width=528 ]
|
we recently demonstrated that standard fixed - time lattice random - walk models can not be modified to properly represent biased diffusion processes in more than two dimensions .
the origin of this fundamental limitation appears to be the fact that traditional monte carlo moves do not allow for simultaneous jumps along each spatial direction .
we thus propose a new algorithm to transform biased diffusion problems into lattice random walks such that we recover the proper dynamics for any number of spatial dimensions and for arbitrary values of the external field .
using a hypercubic lattice , we redefine the basic monte carlo moves , including the transition probabilities and the corresponding time durations , in order to allow for simultaneous jumps along all cartesian axes .
we show that our new algorithm can be used both with computer simulations and with exact numerical methods to obtain the mean velocity and the diffusion coefficient of point - like particles in any dimensions and in the presence of obstacles . , , diffusion coefficient , biased random - walk , monte carlo algorithm .
|
boundary scattering is essential for complete description of nanosystems such as quantum wells , ultrathin films or wires , _ etc . _ due to the large surface - to - volume ratio , boundaries are expected to play a much greater role in determining the overall properties in a nanostructure than in a bulk material .
for example , recent scanning tunneling microscopy ( stm ) data have shown that electron energy spectra can be more strongly correlated to the buried interfacial lattices than to the surface immediately beneath the stm tip @xcite .
these observations clearly indicate that a small lateral variation along the boundary can have a significant long range effect in a semi - ballistic electron system .
thus , a more realistic description of a nanoscale quantized system must go beyond the common perfect geometric boundary and include boundary corrugations .
indeed , random surface roughness of a thin metal film can dominate incoherent scattering and relaxation , and can lead to anomalous quantum size effect such as large oscillatory dependence of the in - plane conductivity on the film thickness @xcite .
the same must be true not only for the quantum well ( film ) walls but also for the interlayer interfaces in multilayer films .
it is well known that the roughness of the interlayer interfaces plays an important role in , for example , giant magnetoresistance ( see review @xcite and references therein ) .
the purpose of this paper is to analyze the effect of irregular corrugation of the interlayer interfaces on the lateral conductivity of quantized multilayer films without magnetic effects .
we will see that the interface scattering can result in unique features of the quantum size effect ( qse ) which are strikingly different from qse with scattering by bulk or wall inhomogeneities .
orbital and spin magnetic effects of the type studied in refs .
@xcite will be studied separately . in ultrathin films
, the motion of electrons across the films can be quantized .
qse in metal films is studied experimentally by measuring conductivity @xcite and susceptibility @xcite of the films or in spectroscopy @xcite and stm @xcite measurements ( for earlier results , see references therein ) . as a result of qse ,
the @xmath0 electron spectrum @xmath1 splits into a set of minibands @xmath2 where @xmath3 is the @xmath4 momentum along the film ( @xmath5plane ) . in the simplest case of a single - layer film approximated by a rectangular quantum well
, the quantized values of the @xmath6component of momentum are @xmath7 ( here and below @xmath8 ) .
if in such quantized metal films the fermi energy @xmath9 is unaffected by the quantization , the fermi surface reduces to a set of @xmath4 curves @xmath10 that correspond to cross - sections of the @xmath0 fermi surface @xmath11 by a set of planes @xmath7 , @xmath12 .
this quantization of motion , which is determined by the film thickness @xmath13 , leads to several types of qse .
first , any change of the film thickness @xmath13 results in the change in size and number of the fermi curves @xmath10 .
this thickness - driven change in number of the fermi curves @xmath14 [ or , what is the same , number of occupied minibands @xmath2 ] leads to a singularity in the density of states .
these singularities are the most obvious manifestations of qse .
these singularities in the density of states , by itself , do not lead to any _ striking _ anomalies in the dependence of the lateral conductivity @xmath16 of the film on the thickness @xmath13 .
the conductivity is more sensitive to electron scattering than to the density of states .
however , the change in the number of the occupied minibands @xmath17 can be accompanied by a change in number of allowed scattering channels that correspond to the scattering - driven electron transitions between minibands @xmath18 . the effect of this step - like change in the number of scattering channels on the conductivity is much stronger than that of the singularities in the density of states @xcite
when all scattering - driven interband transitions are allowed , qse manifests itself as a pronounced saw - like dependence of the conductivity on the film thickness .
this type of qse in quantized films has been predicted both for scattering by impurities and surface inhomogeneities @xcite .
when the main scattering mechanism is the scattering by surface inhomogeneities , many of the interband transitions can often be suppressed .
this happens , for example , when the average size of the surface inhomogeneities @xmath19 is much larger than the the thickness of the film and/or the particle wavelength @xmath20 .
then the usual qse , which is described above , disappears and is replaced by a different kind of the size effect @xcite .
this anomalous qse , which is somewhat reminiscent of the magnetic breakthrough , is completely decoupled from the singularities in the density of states and is associated solely with opening of interband scattering channels for gliding electrons at certain values of the film thickness , @xmath21 .
the main goal of this paper is to analyze qse in _ multilayer _ films with an emphasis on the scattering by the interface between the layers .
we will see that , in addition to the above two types of qse , the multilayer films can exhibit a peculiar `` geometric '' qse with very narrow high peaks in the lateral conductivity .
some of the positions of these spikes in conductivity are universal ; these spikes appear when the ratio of the thicknesses of the film layers is given by simple fractions .
the position of the rest of the spikes depend on the strength of the interlayer interface . in the next section ,
we briefly present the main equations for the conductivity and introduce proper dimensionless variables .
the results are presented in section iii .
section iv contains the summary and brief discussion of the results .
appendix contains auxiliary information on the energy spectrum of multilayer films of the type used in the calculations .
for simplicity , we consider an ultrathin film of thickness @xmath13 consisting of only two layers with the thickness of @xmath22 and @xmath23 .
the interface between the layers is rough with random corrugation .
the exact position of the interface , @xmath24 is described by the random function @xmath25 with the zero average , @xmath26 .
the random interface inhomogeneities @xmath27 are best characterized by the correlation function @xmath28 , @xmath29 where the vector @xmath30 gives the @xmath4 coordinates along the interface and @xmath31 is the averaging area . here
, it is assumed that the correlation properties of the surface do not depend on direction .
two main characteristics of the surface correlation functions @xmath32 are the average amplitude ( height ) and the correlation radius ( size ) of surface inhomogeneities , @xmath33 and @xmath19 . to emphasize the scattering by inhomogeneities of the interlayer interface
, we start from films with ideal outside walls that do not contribute to electron scattering . the combined effect of interface and wall inhomogeneities will be considered elsewhere .
mostly we are interested in the dependence of the lateral conductivity on the film thickness and have in mind the following experimental situation .
the first layer of the film is grown on some ( ideal ) substrate .
the surface is then roughened by adding inhomogeneous adsorbate or by some other means . the growth of the second layer starts from this rough interface and the conductivity is measured at different values of @xmath23 either in the process of growth or after the growth is completed .
an advantage of such setup with a buried interface is that it allows one to measure the conductivity at various values of the film thickness with _ exactly the same _ random rough interface . in this setup ,
the thickness of the first layer , @xmath22 , should be considered as fixed , while the thickness of the second layer , @xmath23 , is variable .
below we are calculate the film conductivity @xmath16 as a function of the film thickness @xmath34 , @xmath35 , assuming that @xmath36 .
the measurements of conductivity can be performed in stationary conditions at different values of @xmath23 or as a function of time , in the process of film growth as in ref .
@xcite .
the second layer can be made of the same or different material as the first .
if the material is different , then the electron potential energy between the layers differs by some @xmath37 .
the the structure of the energy spectrum becomes a complicated function of @xmath37 making the behavior of conductivity highly irregular @xcite .
below we consider both layers to be made of the same material with the interface being the only disruption in the potential relief .
then the simplest model of the interface is the @xmath38functional potential barrier@xmath39 this immediately introduces two new physical parameters into the problem : the strength of the barrier @xmath40 and its ( average ) position @xmath22 . in what follows , we study the dependence of the conductivity on these parameters .
when necessary , instead of the @xmath41-function we will study the corrugated interface with the finite width @xmath42 . in experiment
, the barrier can be a dislocation wall , twin boundary , or an oxide or dielectric layer ( see , _
e.g. , _ ref . @xcite and references therein ) .
the presence of the interface @xmath43 changes the spectrum .
when calculating the changes in the spectrum , one can ignore small corrugation @xmath25 .
the changes in spectrum caused by the @xmath38type barrier @xmath43 are discussed in appendix .
the random corrugation of the interface is responsible for the electron scattering and gives rise to the collision operator in the transport equation .
the scattering by the interface inhomogeneities leads to the transitions between the states @xmath44 .
several ways of calculating the corrugation - driven transition probabilities @xmath45 are described in ref .
the simplest methods are either the direct perturbation approach @xcite or the mapping transformation method @xcite with both giving the same result in most of the parameter range .
the corrugation - driven contribution @xmath46 to the interface potential , eq .
@xmath43 , with small corrugation @xmath47 is@xmath48 the matrix element @xmath49 of this perturbation between the states @xmath50 is @xmath51 , \label{b3}%\end{aligned}\ ] ] where @xmath52 are the quantized wave functions for electron motion across the film .
note that the derivatives @xmath53 for films with a @xmath38type barrier inside are discontinuous at the position of the barrier @xmath54 .
therefore , @xmath55 in eq .
@xmath56 should be understood as @xmath57 /2 $ ] . the corrugation - driven transition probability @xmath58 is given by the square of this matrix element which should be averaged over the random inhomogeneities @xmath47 : @xmath59 ^{2 } , \label{b5}%\end{aligned}\ ] ] where @xmath60 is the fourier image of the correlation function of the interface inhomogeneities @xmath61 .
the coefficients @xmath62 are calculated with the help of the wave functions presented in the appendix .
the explicit form of @xmath62 is given in the next subsection .
the transport equation is a set of equations for the electron distribution functions @xmath63 in minibands @xmath64 and has the standard boltzmann - waldmann - snider form @xcite : @xmath65 \delta\left ( \epsilon_{i\mathbf{q}}-\epsilon_{j\mathbf{q}^{\prime}}\right ) \frac{d^{2}q^{\prime}}{\left ( 2\pi\right ) ^{2}}. \label{e6}%\ ] ] the integration over @xmath66 is done using the @xmath41-function , @xmath67 , where @xmath68 is the solution of the equation @xmath69 and the effective masses @xmath70 . as always in the transport theory , the angular integration is eliminated by using the angular harmonics .
the current is given by the first harmonic of the distribution @xmath71 the equation for which involves only the zeroth and first harmonics @xmath72 of @xmath73 over the angle @xmath74 , @xmath75 , \nonumber\\ w_{ij}^{\left ( 0,1\right ) } & = u_{0}^{2}\zeta^{\left ( 0,1\right ) } \left ( \mathbf{q}_{i}-\mathbf{q}_{j}\right ) g_{ij},\nonumber\end{aligned}\ ] ] where we , to simplify the equations , assume that the effective mass @xmath76 does not depend on its indices , @xmath77 .
the solution of eqs .
@xmath78 provides the @xmath4 conductivity of the film : @xmath79 the problem involves several length scales - particle fermi wavelength @xmath80 the thickness of the layers @xmath22 and @xmath81 ( @xmath82 ) , the correlation radius of the surface inhomogeneities @xmath19 , and the interface thickness @xmath42 .
another length parameter , the amplitude of inhomogeneities @xmath33 , is perturbative and enters conductivity as a coefficient , @xmath83 note , that we consider only the contribution from surface roughness and disregard the bulk scattering . as a result ,
the conductivity @xmath84 diverges in the limit of vanishing inhomogeneities @xmath85 or @xmath86 .
the proper account of bulk scattering @xcite eliminates this divergence .
it is convenient to measure all the length parameters in the units of the fermi wavelength @xmath87 . instead of the interface strength @xmath88
we use interchangeably two equivalent dimensionless parameters @xmath89 and @xmath90,@xmath91 ( @xmath89 is convenient for the calculation of the spectrum while @xmath90 is a proper energy parameter for characterization of the conductivity in our setup ) .
the position of the interface is characterized by the parameter @xmath92@xmath93 in computations , @xmath41 changes from @xmath94 ( no second layer ) to @xmath95 ( the second layer much wider than the first ) .
it is worth repeating that we are looking at the experimental situation when the thickness of the first layer is fixed and the conductivity is measured as a function of the thickness of the second layer ( or the overall film thickness ) . the energy spectrum @xmath96 is described by dimensionless energy units @xmath97@xmath98 where @xmath99 is given by the solution of the @xmath100 schroedinger equation for a quantum well with a @xmath38type barrier inside ( see appendix ) : @xmath101 = 0 , \label{b9}%\ ] ] finally
, the conductivity @xmath35 for the experimental setup which has been described above , will be displayed by the dimensionless function @xmath102@xmath103 for various values of @xmath104 , @xmath105 , @xmath106 and the strength of the barrier @xmath90 .
all the figures below present this dimensionless function @xmath107 .
this function is plotted under the assumption that the experiment is performed at fixed thickness of the first layer . for uniformity ,
the figures for weak interfaces are plotted for @xmath108 , and for strong interface barriers - for @xmath109 .
the simplest energy spectrum corresponds to thin first layers , @xmath110 .
therefore , for transparency of results , the majority of the data are presented for @xmath111 ( for comparison , some of the graphs give the conductivity for larger @xmath22 ) .
the computational results below are presented for the gaussian correlation function of the interface inhomogeneities,@xmath112 the angular harmonics for this correlator , which enter the transition probabilities in eq .
@xmath78 , are equal to @xmath113 e^{-\left ( q - q^{\prime}\right ) ^{2}/2},\label{b22}\\ \zeta^{\left ( 1\right ) } \left ( q_{i},q_{j}\right ) & = 4\pi\ell^{2}% r^{2}\left [ e^{-qq^{\prime}}i_{1}\left ( qq^{\prime}\right ) \right ] e^{-\left ( q - q^{\prime}\right ) ^{2}/2}\nonumber\end{aligned}\ ] ] where @xmath114 , @xmath115
. analysis of qse in ref .
@xcite for ultrathin films with scattering by the film walls demonstrated that the results for all types of correlators are qualitatively the same as for the gaussian one when @xmath116 . for large inhomogeneities , @xmath117 ,
the results for all types of correlators with exponential power spectra are similar to those for the gaussian correlator and are qualitatively different from the power - law correlators .
the results for the power - law correlators are less interesting : such films always exhibit the standard saw - like qse irrespective of the value of @xmath19 because of the wider fluctuations of the inhomogeneity sizes . therefore , in this paper we consider only the exponential correlators with a well - defined size of inhomogeneities .
the standard quantum size effect ( qse ) in films manifests itself by a saw - like dependence of the conductivity @xmath16 on the film thickness @xmath13 @xcite .
the positions of the singularities - the saw teeth - correspond to the values of thickness at which a new energy miniband @xmath118 becomes accessible .
the amplitude of the conductivity drop in such a singular point depends , in the case of scattering by surface inhomogeneities , on the effectiveness of the roughness - driven interband transitions .
if the probability of such transitions @xmath119 , is small in comparison to the rate of the intraband scattering @xmath120 , the singularities in the curves @xmath35 are almost completely suppressed and the standard qse disappears @xcite .
analysis of the roughness - driven transition probabilities for surface scattering in ref.@xcite for different classes of surface roughness showed that , when the average size of inhomogeneities , @xmath19 , is much smaller than the film thickness @xmath13 , the values of the interband transition probabilities @xmath119 are comparable to that for the intraband scattering @xmath120 , all the scattering channels are equally important . in this case , the curves @xmath35 always exhibit the standard qse .
the same should be true for scattering by the interlayer interfaces .
this is illustrated in figures 1 and 2 which show @xmath35 for a weak and strong interface potentials @xmath108 and @xmath109 respectively . in figure 1 the thickness of the first layer is @xmath121 , in figure 2 - @xmath122 . in both figures ,
the size of inhomogeneities is @xmath123 .
both figures exhibit a well - pronounced saw - like structure .
the positions of the singularities for the weak interface are almost equidistant reflecting the fact the energy structure is close to that for a square well without perturbation inside .
the strong interface affects the energy spectrum and , therefore , the positions and the shapes of the saw teeth .
however , at very large film thickness @xmath124 the interface is located very close to the well wall and the spectrum start to recover its unperturbed structure .
this manifests itself in a recovery of the equidistant distribution of the singularities in figure 2 at large @xmath13 . because of a peculiar dependence of the transition probabilities on the interface strength ( see appendix )
, the conductivity grows much faster with increasing film thickness in the case of the weak interface than for the strong interface . .
the saw - like dependence is typical for the standard quantum size effect . the correlation radius of inhomogeneities @xmath125 , the thickness of the first layer @xmath126 , the width of the interface @xmath127 , and the strength of the barrier @xmath128 . ] for strong interface potential , @xmath129 . the correlation radius of inhomogeneities @xmath130 , the thickness of the first layer @xmath131 , the width of the interface @xmath132 . ]
the standard qse of the type described in the previous subsection disappears in the single - layer film when the correlation size of inhomogeneities , @xmath19 , is larger than the film thickness , @xmath117 , _ and _ the correlation function in the momentum space @xmath133 ( the so - called power spectrum of inhomogeneities ) decays exponentially at large wave numbers @xmath3 .
instead , the single - layer films exhibit an anomalous qse @xcite .
the explanation involves the interband transitions .
it seems that at large @xmath19 the off - diagonal @xmath134 are small and the interband transitions are suppressed .
however , at certain values of large @xmath13 , few of the elements @xmath134 , which are close to the main diagonal , could become comparable to @xmath120 even for large @xmath19 .
then the transitions @xmath135 could become noticeable leading to a drop in conductivity . a simple estimate of the peak positions is the following .
scattering by surface inhomogeneities changes the tangential momentum by @xmath136 .
this is sufficient for the interband transition when this @xmath137 . when the number of occupied minibands is large , the lateral fermi momentum for the gliding electrons , _
i.e. , _ electrons from the miniband with a relatively small index @xmath138 @xmath139 . for such electrons , @xmath140 . on the other hand
, the energy conservation law dictates @xmath141 accordingly , with increasing @xmath13 the transition channel @xmath135 opens at @xmath142 .
the opening of a new scattering channel in the points @xmath143 is always accompanied by a drop in conductivity .
the first such drop occurs for the electrons in the lowest miniband @xmath144 with @xmath145 , _
i.e. , _ for the grazing electrons . note , that these particular electrons contribute the most to the conductivity .
since the electrons from the lowest miniband are responsible for the dominant contribution to the conductivity , the conductivity drops almost by half in the point @xmath146 where @xmath147 becomes comparable to @xmath148 and the effective cross - section doubles .
[ in the quasiclassical film without bulk scattering , the current , which is an integral over momenta , diverges when the component of momentum perpendicular to the film goes to zero , _
i.e. _ , for the grazing electrons . without the bulk scattering ,
the conductivity is finite only because of the quantum cut - off at @xmath149 .
the anticipation was that this type of qse should manifest itself also for the interface scattering in multilayer films at @xmath117 for exponentially decaying surface correlators .
indeed , such a picture can be observed in figures 3 and 4 for @xmath150 respectively ( in both figures , @xmath151 ) .
the positions of the peaks in figure 3 for the weak interface are close to eq .
@xmath152 . in the case of the strong interface ,
the shift of the energy levels from those for an `` empty '' square well is much more noticeable and the positions of the peaks in figure 4 deviate from those given by eq
. @xmath153 at large values of @xmath13 , the positions of the peak with strong interface become close to the points in which the thickness of the second layer , @xmath154 , rather than the overall thickness @xmath13 is given by eq .
the amplitude of the anomalous qse oscillations grows with the increasing strength of the interface approaching that for the impenetrable wall . .
the correlation radius of inhomogeneities is large @xmath155 , the thickness of the first layer @xmath126 , the width of the interface @xmath156 , and the strength of the interface barrier @xmath128 . ] . ] of course , for the inhomogeneities of the intermediate size , the picture exhibits the features of both standard and anomalous qse .
as it has already been mentioned , our numerical examples address the experiment in which the size of the inhomogeneities @xmath19 is fixed while the thickness of the film @xmath13 is changing . in general , at the values @xmath157 one should see the smooth anomalous qse oscillations with large period , while at @xmath158 one should , on the same curve , see the reappearance of the standard qse with sharper oscillations with period equal to 1 .
roughly , the transitions between the regimes occurs when the distance between the peaks of the anomalous qse , eq .
@xmath159 , decreases to the value @xmath160 . in principle
, the reappearance of the standard qse should be seen in figures 3 and 4 when the computations are extended to sufficiently large @xmath13 . however , the amplitude of the standard qse oscillations on these curves is very small and the reappearance of the oscillations is barely noticeable on the scale of the curve .
it is much more illustrative to demonstrate the effect at intermediate values of @xmath19 when both anomalous and standard qse oscillations have comparable amplitude .
this is shown in figure 5 for @xmath161 and weak interface @xmath108 .
on the left side of the graph one can clearly see smooth `` new '' oscillations with a relatively large period , while on the right side the oscillations recover the sharp saw - like structure with the period equal to 1 . for the intermediate values of the size of inhomogeneities , @xmath162 . the thickness of the first layer @xmath131 , the width of the interface @xmath156 , and the interface barrier @xmath128 . at small
@xmath13 , the curve exhibits the smooth oscillations of the anomalous qse with a large period , while qse for large @xmath13 recovers the standard saw - like shape with the period equal to 1 . ]
to exhibit the qse oscillations of the previous subsection , figures 3 and 4 were plotted not for the exact @xmath38type interfaces @xmath163 but for a somewhat smeared ( less sharp ) interface@xmath164 .
\label{b11}%\ ] ] the interface width @xmath42 can have two origins
. if its origin is corrugation - related , then the interface width is given by the next term of expansion of the interface barrier in @xmath47 and is characterized by the same parameters @xmath33 and @xmath19 , @xmath165 . in this case , depending on the correlation function , @xmath166 or @xmath167 . on the other hand ,
@xmath42 can originate from some `` internal '' smearing of the interface and can exist even without surface inhomogeneities . in this case
, @xmath42 is a new independent small parameter .
note , that here we are interested in the `` smearing '' of the interface and not in its `` fixed '' width so that the average of the square of the matrix elements of @xmath46 over the interface starts from @xmath168 . in figures 3,4 ,
the interface thickness was chosen as @xmath169 .
if the interface is thinner , the character of the curves changes dramatically .
for example , figure 6 presents the conductivity @xmath35 exactly for the same values of all parameters as in figure 3 except for the interface thickness which is now @xmath170 .
the difference between the two curves is astonishing . .
the fractions near the spikes give the values of the resonance positions of the interface @xmath171 . ] .
the fractions near the spikes give the values of the resonance positions of the interface @xmath171 . ]
the conductivity in figure 6 exhibits two types of spikes .
the explanation for first type of spikes is the following .
the scattering - driven transition probabilities @xmath172 , eq.@xmath173 contain the factor with the correlation function @xmath174 and the geometric coefficients @xmath175 , eq.@xmath176 for exponentially decaying correlators with large @xmath177 the off - diagonal values of the correlation function @xmath178 with @xmath179 are exponentially small in comparison with the diagonal ones , @xmath180 .
then it is sufficient to analyze only the diagonal elements of the matrix @xmath175 , eq.@xmath181 @xmath182 if , accidentally , the @xmath38type interface is positioned in the points in which either @xmath183or @xmath184 , then the coefficient @xmath185 , and , therefore , the transition probability @xmath120 , become zero .
this , in turn makes the conductivity of electrons in the miniband @xmath64 , and , therefore , the overall conductivity , almost infinite . the cut - off is determined by one of three factors : 1 ) exponentially small interband transitions ; 2 ) scattering by other defects such as impurities , inhomogeneities of external walls , _ etc . _ ; 3 ) smearing of the interface @xmath186 that leads to the averaging of @xmath187 eq .
@xmath188 over a finite interval making it non - zero
. in this paper , for obvious reasons , we are interested in the third option . note , that in the case of scattering by external film walls instead of the interlayer interface , the coefficients @xmath189 are never equal to zero and this type of qse does not exist .
the first type of spikes corresponds to @xmath190 .
the `` resonance '' positions of the @xmath38type interface are universal and do not depend on the potential strength .
this is true for all rational points @xmath191 .
of course , the conductivity of the film becomes infinite for this position of the interface _ only _ if the corresponding miniband @xmath64 is occupied .
this means that the integer @xmath192 in the denominator of the corresponding fraction @xmath193 should not exceed the number of the occupied minibands , @xmath194 $ ] .
indeed , for points @xmath195 there is a number of the wave functions @xmath52 of the _ empty _ well that have nodes in the points @xmath54 .
since the unperturbed homogeneous potential barrier has a @xmath38functional form @xmath196 , these wave functions @xmath52 remain the eigenfunctions of the well _ with _ the unperturbed barrier @xmath197 inside and retain their nodes in the points @xmath54 . then the corresponding diagonal coefficients @xmath185 are zero making the diagonal roughness - driven transition probabilities @xmath120 for particles from the miniband @xmath64 equal to zero as well . since the off - diagonal transition probabilities are exponentially small in @xmath198 , the condition @xmath199 makes the conductivity for particles from the miniband @xmath200 and , therefore , the overall conductivity exponentially large in @xmath198 .
the structure of the corresponding resonance spikes becomes more and more complicated with an increase in @xmath22 when the structure of the minibands and their occupancy become more convoluted .
the simplest structure is observed when @xmath22 is between @xmath20 and @xmath201 as in figure 6 . in this case , the observed rational spikes correspond to the rational numbers of the form @xmath202 and are equidistant with the separation @xmath106 .
the first spike corresponds to the film with @xmath203 , the second - to @xmath204 , the third - to @xmath205 , the fourth - to @xmath206 , and so on .
the odd peaks , with the exception of the first one , look wider and consist of bigger and smaller sub - peaks .
the smaller sub - peaks correspond to the geometrical resonance with @xmath202 which is described above .
the bigger and wider sub - peaks have a somewhat different nature and are not universal with respect to the barrier strength .
these sub - peaks will be described later .
note , that the peak @xmath205 is so close to the first peak from the other series that these two peaks are hardly distinguishable .
when @xmath22 becomes bigger , the first few geometric resonances can be observed at much narrower second layers , well before the point @xmath207 , while the density of the resonances become higher .
for example , figure 7 presents the conductivity as a function of thickness for the film with the same parameters as in figure 6 except for the thickness of the first layer which is now @xmath208 .
though the overall distribution of the peaks is now much more complicated , the majority can still be understood as the ones generated by the eigenfunctions of the empty quantum well with the nodes in the positions of the barrier .
the complexity of the peak structure is explained by the fact that at wider first layer @xmath22 more minibands are occupied thus allowing a wider selection of the rational numbers that determine the peak positions @xmath195 .
the geometric resonances can coexist with the anomalous qse of the previous section if the interface is relatively strong as in figure 8 for the same configuration as in figure 6 but with much higher value of @xmath90 , @xmath209 . for weak interfaces ,
the geometric resonances suppress the qse of the previous subsection which gets restored only for bigger values of the interface thickness @xmath210 .
this graduate disappearance of the geometric resonances can be seen when comparing figure 6 for @xmath211 @xmath132 with figure 9 ( @xmath212 ) and figure 10 ( @xmath213 ) .
figure 10 presents conductivity for the same configuration as figure 3 but in logarithmic scale . in this scale ,
one can see both the wide qse oscillations of the previous subsection and the only surviving geometric resonance at @xmath214 .
the rest of the parameters are the same parameters as in figure 6 . for easier comparison with figures 4 and 6
, the insert gives the same data in logarithmic scale.,title="fig : " ] . ] .
data as in figure 3 , but in logarithmic scale . ] above we explained only the narrower , universal geometric resonances at @xmath193 in figures 6 - 9 . the second
, non - universal type of spikes has a similar explanation - zeroes of @xmath185 . according to appendix,@xmath215 ^{2 } , \label{b15}%\end{aligned}\ ] ] where @xmath216 is given by the solution of the @xmath100 schroedinger equation @xmath217 for a quantum well with a @xmath38type barrier inside .
the explicit form of the coefficients @xmath218 is not important .
the factor @xmath219 in eq .
@xmath220 corresponds to @xmath221 ; its zeroes are responsible for the geometric resonances with rational @xmath202 .
there are no other zeroes of @xmath222 . however , eq .
@xmath220 also contains the factor in the square brackets which corresponds to @xmath223 .
the simultaneous solution of eq .
@xmath217 for the spectrum , @xmath224 and equation @xmath225 yield the following equation for the resonance positions of the interface:@xmath226 = 0 , \label{b17}%\ ] ] which is equivalent to @xmath227 with integer @xmath228 .
first , there is a universal solution @xmath214 at @xmath229 . in this case ,
when the interface is exactly in the middle of the film , both @xmath230 and @xmath231 are zero ( the former with an even index , the latter - with the odd ) .
this explains why the geometric resonance with @xmath214 is the most stable one with respect to the smearing of the interface .
the rest of the resonances with @xmath232 are not universal .
these resonances explain the earlier unaccounted for spikes in figures 6 - 8 . since the spectrum @xmath216 is a complicated function of the interface strength and its position , the solution of eq .
@xmath233 for @xmath232 is rather complicated
. we will give the analytic equation for the simplest case of @xmath234 when eq .
@xmath235 is equivalent to@xmath236 ( the only allowed values of @xmath228 are @xmath237 ) . for weak interfaces @xmath238 this equation
can be rewritten as@xmath239 , \
\delta _ { n}\approx\frac{g}{\pi z_{n}}\sin^{2}\left ( \pi n\delta\right ) .
\label{b20}%\ ] ] analysis of eq .
@xmath240 shows that several first of such resonance positions of the interface are indeed close to the odd rational universal resonances as in figure 6,@xmath241 and separate from the universal resonances with increasing integer @xmath242 .
the very first resonance at @xmath214 is , as it is explained above , exactly the same as the first universal resonance .
the reason why these non - universal resonances are wider and stronger than the universal ones described above is still unclear .
in summary , we analyzed qse in conductivity of multilayer films when the main scattering mechanism is the scattering of electrons by random inhomogeneities of the interlayer interface .
three different types of qse are predicted .
the first one is a standard qse with a typical saw - like dependence of the conductivity @xmath16 on the film thickness @xmath13 , @xmath35 .
this effect dominates when the correlation radius ( size ) of the interface inhomogeneities @xmath19 is much smaller than the film thickness , @xmath116 .
this effect should be observed for all types of the correlation functions of the interface roughness .
this effect is easily explained by the singularities in the electron density of states related to the quantization of motion across the film .
the second type of qse is explained not by the quantization - driven singularities in the density of states , but by the anomalies in the cross - section for scattering by interface inhomogeneities .
this scattering - driven qse replaces the standard saw - like qse when the correlation radius ( size ) of the interface inhomogeneities is large , @xmath117 .
this type of qse manifests itself as smooth large - scale oscillations on the dependence @xmath35 and should be observed only when the fourier image of the interface correlation function ( the so - called power spectrum of inhomogeneities ) decays exponentially at large momenta .
the main difference of this qse from a similar effect in scattering by the film walls @xcite is that the observation of this effect in multilayer film requires certain smearing of the interface .
the third type of qse is new and is most unusual .
this effect manifests itself as a set of very narrow and high spikes in @xmath35 and replaces the scattering - induced qse described above when the interface is narrow .
the finite cutoff in the spikes can be ensured either by some other scattering mechanism or by the smearing of the interface .
the spikes are observed only for certain resonance positions of the interface .
the number of spikes is determined by the relation between the thickness of the layers and the fermi wavelength .
the resonance positions of the interface are described .
these positions can be split into two general classes .
some of these positions are universal and do not depend on the amplitude of the interface potential barrier and correspond to the situations when the ratio of the layer widths is given by simple rational fractions .
the integer in the denominator of such fractions does not exceed the number of occupied minibands .
the remaining resonance positions of the interface are non - universal and depend on the strength of the interface potential . in the case of weak interface ,
some of these non - universal positions are close to the universal ones giving the impression of a split in the conductivity spikes .
too small width of the resonance spikes can impede the experimental observation of the geometric resonances .
the width of the resonance spikes increases and their height decreases with increasing smearing of the interface and the resonance spikes gradually disappear .
note , that this disappearance of the resonance spikes is related not to the widening , but to the random smearing of the interface - the widening of the interface , by itself , results just in a shift of the spike positions .
the width of the universal resonances is equal , by the order of magnitude , to the width of the smeared interface @xmath42 , eq .
( [ b11 ] ) , or , in dimensionless variables , @xmath243 .
the width of non - universal resonances is somewhat larger and is less sensitive to @xmath42 ; the reason is still unclear .
note , that the smearing width @xmath42 can be much smaller than the physical thickness of the interface , which in metals is often larger than or of the order of the fermi wavelength @xmath20 .
in contrast to this , the smearing parameter @xmath243 can be very small .
the most stable spike with respect to smearing corresponds to the layers of equal width .
the wide range of possible values of @xmath42 , which are determined either by the roughness with @xmath42 of the order of @xmath33 or @xmath244 or by the `` internal '' smearing of the interface , makes the observation of the geometric resonances possible .
the spikes in the conductivity occur when the scattering probabilities for electrons in one of the quantized minibands become exactly zero .
since scattering probabilities for scattering by different interfaces add up , the spikes in conductivity of multilayer films with many layers can be observed only if the scattering probabilities for electrons from one miniband become zero simultaneously for scattering by all the interfaces .
this can happen only if all the interfaces are located in the universal resonance positions corresponding to the rational fractions from the same series .
otherwise , the scattering by inhomogeneities of the `` non - resonant interface '' will curtail the contributions from the resonant ones .
this imposes a restriction on the number of layers for an observation of this type of qse for a film of fixed overall thickness .
we analyzed the multilayer films under the condition that the disruption in the electron spectrum is caused only by the interface potential while electron potential deep into the layers is the same for all layers .
one can imagine a different physical situation when the electron potential in different layers differ from each other as in ref .
in this situation the resonance spikes in conductivity should be observed when the position of the interface coincides with one of the nodes in the wave function .
it is clear that this occurs at least for certain values of the interlayer potential difference @xmath37 .
the calculations in the paper are aimed primarily at the experimental setup when the lateral conductivity is measured as a function of the film thickness at fixed thickness of the first layer ( fixed position of the buried interface ) .
the main obstacle for the experimental observation of the predicted effect is a rather small width of the conductivity spikes and their sensitivity to the position of the interface . on the other hand , this sensitivity of qse to the position of the interface may open the door for using this effect for precision control of the interface positions in multilayer films
this may be very useful for better quality ultrathin films without short - range surface inhomogeneities @xcite .
recent experiments with controlled ultrathin metal films with buried rough interfaces @xcite indicate that the existing experimental setups are sufficient for the observation of the predicted quantum size effect .
usually , qse in conductivity of semiconductor films is less pronounced than for the metal films .
this is explained by the smoother distribution of electrons in non - degenerate semiconductors . in the absence of sharp drop in the distribution at the fermi energy ,
singular features in the conductivity , which is an integral over the particle distribution , tend to be smeared out .
however , the universal geometric spikes in conductivity , which are described above , are explained by the zeroes in quantized electron wave functions on the interface and have nothing to do with the electron distribution . then these spikes in conductivity can be the only striking common feature for qse in multilayer metal and semiconductor films .
the only obstacle for observation of such spikes in semiconductors could be a relatively large screening radius which may lead to an effective smearing of the interface .
this work was supported by nsf grant dmr-0077266 .
one - dimensional schrdinger equation for a square well with a @xmath245-functional barrier inside has the form@xmath246 where@xmath247 the wave functions can be written as@xmath248 in the opposite case of strong interface @xmath257 , the spectrum decouples into two independent series of levels for each layer:@xmath258 for large , but finite @xmath89 , the corrections to the spectrum @xmath259 can be easily obtained by expansion in @xmath260 or @xmath261:@xmath262 an important restriction for eq .
@xmath263 is that the energy levels in each of the layers , eq .
@xmath264 , should not be very close to each other . in the case of near degeneracy , the two close levels , as usual , repel each other with a resulting gap equal to@xmath265 the above equations should be modified if the interface is very close to one of the external walls of the well , _
i.e. _ , if either @xmath266 or @xmath267 . note , that if the @xmath41-type barrier is located exactly in the node of one the wave functions of the empty well , this wave function remains the eigenfunction of the well with a barrier inside irrespective of the strength of the barrier .
this means that the energy levels that correspond to such wave functions are not shifted by the presence of the barrier .
the matrix elements of the roughness - related perturbation @xmath268 can be calculated with the help of the above functions @xmath269:@xmath270 ^{2}\nonumber\\ & = \left ( \frac{2\pi}{l^{2}}a_{m}a_{n}\right ) ^{2}g_{nm}^{2},\label{ap11}\\ g_{nm } & = g\sin\left ( \pi z_{m}\delta\right ) \sin\left ( \pi z_{n}% \delta\right ) + z_{m}\cos\left ( \pi z_{m}\delta\right ) \sin\left ( \pi z_{n}\delta\right ) \nonumber\\ & + z_{n}\cos\left ( \pi z_{n}\delta\right ) \sin\left ( \pi z_{m}% \delta\right ) .\nonumber\end{aligned}\ ] ] the most important are the diagonal matrix elements@xmath271 \label{ap12}\\ & = \frac{z_{n}\sin\left ( \pi z_{n}\delta\right ) \sin\left [ \pi z_{n}\left ( 1 - 2\delta\right ) \right ] } { \sin\left [ \pi z_{n}\left ( 1-\delta\right ) \right ] } .\nonumber\end{aligned}\ ] ] note that the zeroes of the denominator in eq .
@xmath272 are cancelled out by the zeroes of @xmath273 , eq .
when the interface has a finite width @xmath210 , the matrix elements acquire the following addition:@xmath275 where @xmath276 the total matrix element , @xmath277 , is never zero .
this means that the term with @xmath278 , which originates from the smearing of the interface , provides a natural cutoff for the conductivity .
m. jalochowski , m. hoffmann , and e. bauer , phys.rev.lett . * 76 * , 4227 ( 1996 ) ; phys .
b * 51 * , 7231 ( 1995 ) ; l. a. kuzik , yu .
e. petrov , f. a. pudonin , and v. a. yakovlev , sov.phys .
- jetp * 78 * , 114 ( 1994 ) ; g. m. mikhailov , i. v. malikov , and a. v. chernykh , jetp lett . * 66 * , 725 ( 1997 ) m. jalochowski , e. bauer , h. knoppe , and g. lilienkamp , phys.rev .
b * 45 * , 13607 ( 1992 ) ; m. jalochowski , m. hoffmann , and e. bauer , phys.rev .
b * * 51 * * , 7231 ( 1995 ) ; h. sakaki , t. noda , k. hirakawa , m. tanaka , and t. matsusue , appl .
. lett . * 51 * , 1934 ( 1987 ) ; l .- w .
tu , g. k. wong , and j. b. ketterson , appl .
. lett . * 55 * , 1327 ( 1989 ) j. j. paggel , t. miller , and t. c. chang , science , * 283 * , 1709 ( 1999 ) ; d. a. evans , m. alonso , r. cimino , and k. horn , phys.rev.lett . * 70 * , 3483 ( 1993 ) ; j. e. ortega , f. j. himpsel , g. j. mankey , and r. f. willis , phys.rev . b**47 * * , 1540 ( 1993 ) j. j. paggel , t. miller , and t. c. chiang , phys.rev.lett . * 81 * , 5632 ( 1998 ) ; science * 283 * , 1709 ( 1999 ) ; f. patthey and w .- d .
schneider , phys.rev .
b * 50 * , 17560 ( 1994 ) ; m. schmid , w. hebenstreit , p. varga , and s. crampin , phys .
lett . * 76 * , 2298 ( 1996 )
|
conductivity of quantized multilayer metal films is analyzed with an emphasis on scattering by rough interlayer interfaces .
three different types of quantum size effect ( qse ) in conductivity are predicted .
two of these qse are similar to those in films with scattering by rough walls .
the third type of qse is unique and is observed only for certain positions of the interface .
the corresponding peaks in conductivity are very narrow and high with a finite cutoff which is due only to some other scattering mechanism or the smearing of the interface .
there are two classes of these geometric resonances .
some of the resonance positions of the interface are universal and do not depend on the strength of the interface potential while the others are sensitive to this potential .
this geometric qse gradually disappears with an increase in the width of the interlayer potential barrier .
|
noncommutativity has been intensively studied in past years , in part due to its role in string theory ( @xcite and references therein ) and its application to condensed matter physics ( see , for example , @xcite and references therein )
. it also offers interesting alternatives for gauge theories . from the mathematical point of view
, the development of noncommutative geometry @xcite and the formality theorem in deformation quantization @xcite , have been fundamental to the understanding of noncommutativity .
the case of a constant noncommutativity parameter , which leads to the familiar moyal product , is well understood ( see , for example , @xcite ) . however , space - dependent noncommutativity is much more involved , since there is no general procedure to find an explicit representation for the resulting infinite - dimensional algebra , except when it comes from a poisson structure @xcite .
the whole physical picture corresponding to this situation is , consequently , not completely known , although a lot of progress has been made due to recent efforts .
for example , the connection between curved d - branes and space - dependent noncommutativity has been established in @xcite , and the construction of gauge theories in curved noncommutative spaces has been presented in @xcite .
see also refs.@xcite-@xcite .
besides , an illuminating field - theory interpretation of kontsevich 's construction in terms of a sigma model has been provided in @xcite . the aim of this paper is to show how certain cases of coordinates satisfying a space - dependent commutation relation may be analyzed with the help of a change of variables .
the main idea is that it may be possible to define new variables in terms of which one obtains a constant commutator . in terms of those coordinates
, we can work with the usual moyal representation of the noncommutative product and then easily construct a calculus and study the geometry of the space .
if the relevant objects ( product , derivatives , integrals , ) may at the end be written in terms of the original variables , we can obtain explicit representations for the original algebra , as well as for the derivatives and integrals .
interestingly enough , cases in which this can be achieved are not artificial but arise in a variety of problems , for example related to the construction of noncommutative solitons and instantons @xcite-@xcite .
this paper is organized as follows : in section [ sec : metric ] , we discuss the relation between the metric and a space - dependent commutation relation for the coordinates of a two - dimensional space , based on general considerations . in section [ ssec : particular ] , we consider the particular case of a commutation relation that depends on only one of the coordinates . we show that , for this particular case , it is always possible to perform a change of variables to new coordinates verifying a constant ( i.e. , space - independent ) commutation relation so that the moyal product representation can be used with the resulting simplifications it implies .
we then extend all the previous results to the example of a curved interface dividing two regions that have ( different ) constant values for @xmath0 in section [ sec : interface ] , where we present the concrete example of a scalar field theory .
the possibility of generalizing the method is discussed in section [ sec : general ] . in section [ sec : conc ] we present our conclusions .
coordinate commutators in a non - flat background spatial metric , necessarily imply the introduction of a coordinate - dependent noncommutativity parameter @xmath1 : @xmath2_\star = i\theta^{ij}(x ) \;.\ ] ] now , as it is well known , there are severe constraints that have to be satisfied for a consistent definition of a noncommutative associative @xmath3-product . in general , even when those constraints are satisfied , it is not possible to give a closed and explicit formula for the @xmath3-product between two arbitrary functions , as it can be done for the constant-@xmath0 case ( where one has the standard moyal product ) .
the general conditions under which such a product can be defined are derived in @xcite . a sufficient condition to have an associative noncommutative product
may be stated as follows @xcite : @xmath4 where @xmath5 is the covariant derivative corresponding to a metric connection .
although up to this point we have not introduced any metric into the game , a natural one will appear precisely when solving ( [ suff ] ) in @xmath6 dimensions . that metric will be determined ( up to a constant scalar factor ) by @xmath7 itself
; see ( [ conchap1 ] ) and ( [ conchap2 ] ) below . condition ( [ suff ] ) may be derived by starting from the definition of the poisson bracket @xmath8 for two functions @xmath9 and @xmath10 : @xmath11 and then noting that a noncommutative product , up to order @xmath0 , takes the form are presented in @xcite . ] : @xmath12 the associativity of the @xmath3-product ( [ starr ] ) requires the jacobi identity for the poisson bracket ( [ pb ] ) to hold true , which in turns results in the equation : @xmath13 i.e. , @xmath1 is a poisson structure .
in fact , this implies @xcite that the product will be associative to all orders .
the previous condition can be written covariantly in the form : @xmath14 since the terms in @xmath15 containing the ( symmetric ) connection cancel out .
condition ( [ suff ] ) is a sufficient condition for ( [ eq : jacovar ] ) to be true , namely , for the associativity to be valid to all orders in @xmath0 .
moreover , ( [ suff ] ) is very easy to deal with on a two dimensional space .
indeed , the most general @xmath1 can be written in @xmath16 in the form @xmath17 where @xmath18 is a scalar and @xmath19 is the determinant of the two dimensional metric @xmath20 .
but then , condition ( [ suff ] ) reduces to @xmath21 and hence @xmath22 must be a constant . then , the sufficient condition ( [ suff ] ) is equivalent , in @xmath16 , to the equation : @xmath23 where @xmath22 is a real ( non - vanishing ) constant .
this expression relates the space - dependent associative noncommutativity to a non - trivial background metric @xmath24 .
in fact , it is believed @xcite that quantum gravity is at the origin of noncommutative effects , so that kind of relation is not entirely unexpected .
the @xmath24 tensor can be explicitly written by taking into account that in @xmath6 dimensions every metric is conformally flat : @xmath25 and that the conformal factor @xmath26 is determined by @xmath27 : @xmath28 let us also note that , in the presence of a metric , the natural integration measure , @xmath29 , should be @xmath30 which is , as we shall see , consistent with the definition of a noncommutative product in which the integral acts as a trace .
let us specialize the previous discussion in two dimensional space to the case in which @xmath0 depends on only one coordinate , @xmath32 say .
we know that @xmath0 can then be written as follows : @xmath33 where @xmath34 is a positive definite function .
our aim is to present an explicit formula for a noncommutative associative @xmath3-product such that @xmath35 _ { \star } \;=\ ; x^1 \star x^2 - x^2 \star x^1 \;=\ ; i \ , \theta_0 \ ,
t(x^1 ) \label{11 } \;.\ ] ] an expression for @xmath3 in powers of @xmath22 was already found by kontsevich @xcite , where the first terms are given by ( [ starr ] ) .
however , the procedure to construct it , based on the formality theorem , can be quite involved in practice . rather than following that approach
, we shall show how to arrive to an associative noncommutative product by a different path .
inspired by the change of variables found in @xcite , while studying noncommutative vortices in a curved space , we left and right multiply ( [ 11 ] ) by @xmath36 where the square root is the noncommutative one , i.e. , @xmath37 .
then : @xmath38 here we have used the associativity of the @xmath3-product , which is valid for any positive function @xmath39 .
we then change variables , from @xmath40 to new ones @xmath41 defined by : @xmath42 obtaining : @xmath43 note that ( [ tres ] ) involves the ( well - defined ) square root of a positive element , and besides , the definition of the new variables is not sensitive to the sign ambiguity of the square root , since that square root appears quadratically .
also , the variables @xmath44 and @xmath45 , as defined in ( [ tres ] ) , are hermitian and , because of ( [ cuatro ] ) , the noncommutative @xmath3-product can then be realized as an ordinary moyal product .
we shall use ` @xmath46 ' to denote the constant @xmath0 ( @xmath47 ) moyal product in terms of variables @xmath48 : @xmath49 for functions of @xmath44 and @xmath45 , working with the @xmath46-product , we can use all the standard ( flat - space ) noncommutative geometry tools @xcite .
besides , we can rewrite the moyal formula in terms of the original variables , where it leads to a concrete expression for the @xmath3-product . to that end , we first note that from ( [ tres ] ) one has @xmath50 using ( [ eq : moyal1 ] ) , we obtain @xmath51 where all the noncommutative artifacts have disappeared . then : @xmath52 is an ( exact ) expression for the change of variables , that will allow us to derive explicit expressions for @xmath3 and for the integration measure in the ` physical ' variables @xmath32 , @xmath53 . as an example , consider the @xmath3-product between two functions of the original variables @xmath54 .
it can be defined in the form @xmath55 it is easy to verify that ( [ cinco ] ) provides a consistent associative realization of the algebra ( [ 11 ] ) . the first line in eq.([cinco ] ) may be applied to verify _ a posteriori _ the associativity of the @xmath3-product to all orders . indeed , that property is , in this light , simply _ inherited _ from the ( well - known ) associativity of the @xmath56 product : @xmath57 @xmath58 of course , since we have at our disposal the formula for the change of variables ( [ eq : changecoord ] ) , we may write everything in terms of @xmath32 and @xmath53 on the right hand side of eq.(22 ) .
one has to take into account the formul @xmath59 using ( [ car ] ) , the @xmath3-product of functions of the original variables can be finally written , to all orders in @xmath22 , as @xmath60\!f(x ) g(\tilde x ) \big \vert _
{ \tilde x = x}\nonumber\\\end{aligned}\ ] ] by explicit computation order to order in @xmath22 one can verify the _ inherited _ associativity .
for example , up to order @xmath61 , the product ( [ eq : moyal2 ] ) takes the form @xmath62 where @xmath63 and @xmath64 are partial derivatives with respect to @xmath32 and @xmath53 .
one can explicitly verify that this formula satisfies associativity .
note that kontsevich construction @xcite leads to equivalence classes ( one for each particular function @xmath0 ) of associative products .
the product we have defined is not the representant chosen in @xcite where the procedure to change the representant is given .
we are ready now to consider integrals of products of fields and derivatives , starting from the known expressions in terms of the new variables .
taking into account that : @xmath65 and starting from an integral @xmath66 involving the @xmath56-product of two functions , we can write an equivalent expression in the new variables : @xmath67 where @xmath68 .
this expression guarantees that integrals of quadratic terms coincide with ordinary ones ( when fields satisfy appropriate boundary conditions ) . up to order @xmath69 this can be verified using eq.([eq : moyal3 ] ) .
hence we see that also in this approach , the natural integration measure , in terms of variables @xmath40 is that given in ( [ dmu ] ) , @xmath70 this agrees with the result obtained in @xcite , based on an operatorial description of noncommutativity . let us end the previous discussion by noting that the machinery developed in this section is very helpful in the construction of solitons ( vortices and monopoles ) and instantons in noncommutative gauge theories .
indeed , one can connect an axially symmetric ( in time @xmath71 ) ansatz for instantons in @xmath72-dimensional noncommutative yang - mills theory with noncommutative vortices in two dimensional curved space with variables @xmath73 .
now , the metric in which this connection is realized takes the form @xmath74 .
since it depends on just one variable , one can proceed , as explained above , to change variables in order to work with a standard moyal product . in this way , passing as usual to the fock space framework , explicit vortex , monopole and instanton solutions can be found in a very simple way @xcite-@xcite .
as an application ( and extension ) of the previous results , we shall study now the case in which @xmath75 is a function whose variation is concentrated along a curve @xmath76 , and is approximately constant elsewhere . in particular , we are interested in considering cases where @xmath0 has different ( constant ) values on two different regions which have @xmath77 as a common interface . a simple , and yet non - trivial example of this situation corresponds to a @xmath78 with the following structure : @xmath79 where @xmath80 and @xmath81 are two different real constants , and @xmath82 denotes the heaviside s step function , with integral representation @xmath83 the shape of the interface is determined by the function @xmath84 , which we assume to be smooth .
the zeroes in the argument of @xmath82 correspond to @xmath85 which defines a smooth curve @xmath76 , dividing the plane into two regions with different values of @xmath0 , denoted @xmath80 and @xmath81 ; see figure [ fig : single ] .
( 0,0 ) that takes the values @xmath81 and @xmath80 in two regions which are separated by the curve @xmath76,title="fig : " ] # 1#2#3#4#5 @font ( 3657,2953)(889,-2513 ) ( 2791,-556)(0,0)[lb ] ( 4546,-1141)(0,0)[lb ] ( 2341,344)(0,0)[lb ] ( 1171,-196)(0,0)[lb ] ( 3241,-2266)(0,0)[lb ] ( 3601,-196)(0,0)[lb ] our first step in constructing a noncommutative theory based on ( [ eq : single ] ) consists in smoothing out the shape of the function @xmath75 , to avoid introducing singularities due to the existence of a finite jump on the interface .
that effect may be achieved by replacing @xmath86 by a smooth approximation to it , @xmath87 , such that @xmath88 for @xmath89 .
for example , @xmath90 \;.\ ] ] with this particular approximant , which we shall adopt , @xmath75 may be written more explicitly as follows : @xmath91 where we defined @xmath92 and @xmath93 .
we now proceed to change variables from the original coordinates @xmath40 to new ones @xmath94 which are defined by : @xmath95 so that the new variables verify the commutation relation : @xmath96_\star \;=\ ; i \,\left ( \theta_0 \,+\ , \delta \ , \tanh(\frac{q^1}{\epsilon } ) \right)\;.\ ] ] in what follows we shall assume that the constants @xmath80 and @xmath81 have the same sign ( positive , say ) , since the opposite situation , which necessarily involves a passage of @xmath0 through zero , is qualitatively different , as it will become clear below . with this in mind , we may write ( [ eq : commy ] ) in the form : @xmath97_\star \;=\ ; i \,\theta_0 \ ; t(q^1 ) \;,\ ] ] where : @xmath98 assuming that both @xmath80 and @xmath81 are positive , the function @xmath71 shall always be positive , and we are of course in the same situation we considered in [ ssec : particular ] .
based on those results , we introduce a second set of variables , @xmath99 , such that : @xmath100^{-1/2 } \,\star\ , q^2 \,\star\ , \big [ t(q^1 ) \big]^{-1/2 } \ ; , \end{array } \right.\ ] ] which verify the constant-@xmath0 algebra @xmath101_\star \;=\ ; i \ ; \theta_0 \;,\ ] ] with as the noncommutativity parameter . then , in terms of the original variables , we have : @xmath102^{-1/2 } \,\star\ , x^2 \,\star\ , \big[t(x^1-\varphi(x^2 ) ) \big]^{-1/2}\;\;\;. \end{array } \right.\ ] ] of course , by a similar procedure to the one explained in [ ssec : particular ] , we may write everything in terms of the variables @xmath44 and @xmath45 , which have a much simpler commutation relation than the original coordinates @xmath32 and @xmath53 .
the inverse transformation is easy to find , @xmath103 where all the products are commutative .
therefore , we also have the direct transformation : @xmath104 an important tool in the construction of noncommutative field theories is the definition of derivatives .
since we know two variables @xmath44 and @xmath45 , which verify a ` canonical ' ( i.e. , constant-@xmath0 ) commutation relation , it is natural to construct derivative operators in terms of them : @xmath105\nonumber\\ d_2 & \equiv & \frac{\partial}{\partial y^2}\,=\ , t(x^1 - \varphi(x^2 ) ) \ , \big [ \varphi'(x^2 ) \frac{\partial}{\partial x^1 } + \frac{\partial}{\partial x^2 } \big ] \;.\end{aligned}\ ] ] we can then obtain an explicit expression for the @xmath3-product between @xmath9 and @xmath27 : from ( [ cinco ] ) and ( [ eq : direct ] ) using the derivatives above defined , @xmath106 where : @xmath107 this expression shows explicitly how the interface enters in the star product
. it may be hard to attempt a direct proof of the fact that the operators defined in this way do satisfy leibniz rule ( with the star product ( [ eq : newstar ] ) ) .
however , this can be easily achieved if we note that @xmath108 are in fact _ inner _ derivations .
indeed , @xmath109_*\,=\,\,i \theta_0 [ x^2 \ , t^{-1}(x^1-\varphi(x^2)),f(y(x))]_\star\;,\ ] ] and a similar expression for @xmath110 .
see , for example , @xcite for a discussion of inner derivations in a general setting . one immediate outcome of ( [ eq : direct ] )
is the expression for the integration measure , assumed to be defined for @xmath111 and @xmath112 , in terms of @xmath32 and @xmath53 : @xmath113^{-1 } \;.\ ] ] it is now evident that the case in which @xmath0 changes its sign through the interface is qualitatively different : the change of variables becomes singular on @xmath76 , when @xmath114 , and as a consequence one has to introduce different changes of variables on different charts .
we have at this stage the complete set of tools to define a noncommutative field theory . for a scalar field in euclidean time @xmath115 ,
satisfying a reality condition which we will determine in a moment , we may define the action : @xmath116 \;=\ ; \int\frac{d\tau dx^1 dx^2}{|t(x^1 - \varphi(x^2))| } \ , \big[\frac{1}{2}\big ( \partial_\tau\sigma \star \partial_\tau\sigma + d_j \sigma \star d_j \sigma + m^2 \sigma \star \sigma \big ) + v_\star ( \sigma ) \big]\ ] ] where the @xmath117 s ( @xmath118 ) have been defined in ( [ eq : defder ] ) , and the @xmath3-product is the one of ( [ eq : newstar ] ) .
this expression may of course be converted into its equivalent version in the variables @xmath119 , a procedure that yields the much simpler expression : @xmath120 \;=\ ; \int d\tau dy^1 dy^2 \ , \big[\frac{1}{2}\big ( \partial_\tau{\tilde \sigma } \ast \partial_\tau{\tilde \sigma } + \partial_j{\tilde \sigma } \ast \partial_j{\tilde \sigma}+ m^2 { \tilde \sigma }
\ast { \tilde \sigma}\big ) + v_\ast ( { \tilde \sigma } ) \big]\ ] ] with @xmath121 , and @xmath122 . here
it becomes evident that a positive action requires @xmath123 to be a real function of @xmath124 ; then , noting that the transformation ( [ eq : direct ] ) is real , @xmath125 has to be real .
working in terms of the variables @xmath48 provides another important simplification : from the moyal product ( [ eq : moyal1 ] ) , we see that @xmath126 then the quadratic part of the action coincides with its commutative counterpart .
noncommutative effects appear in @xmath127 and , when we return to the original coordinates @xmath40 , they show up in the existence of nontrivial metric and derivatives .
this procedure of ` pulling back ' the action to the coordinates @xmath119 may be used , for example , to derive the free propagator in terms of the ` physical ' coordinates @xmath128 .
considering for simplicity the massless case , we see that : @xmath129^{-1/2}\;.\end{aligned}\ ] ] this expression allows one to derive , in particular , the propagator for points on the interface .
indeed , considering the propagator for the two points @xmath130 and @xmath131 , where @xmath132 and @xmath133 , we see that : @xmath134^{-1/2}\;,\ ] ] where we used the fact that @xmath135 on @xmath76 .
the result is different for pairs of arguments that are entirely inside each one of the two constant-@xmath0 regions .
for example , for the case @xmath136 , @xmath137 , we have : @xmath138^{-1/2}\;.\ ] ] and a similar expression ( with @xmath139 ) for the other region .
thus the presence of the interface introduces a fundamental change in the propagator , which never becomes equal to the free one , except for points on the interface .
we have mentioned the difficulties that arise when @xmath0 vanishes ; indeed , it is clear now that this implies a singularity in the change of variables . assuming that @xmath81 and @xmath80 are such that @xmath0 vanishes on the interface , we know that @xmath71 will vanish ( with a non - zero normal derivative ) at the interface .
one can , however , still define an action for the region where @xmath140 , say .
it is easy to write that action in terms of the variables @xmath119 , and then pass to the physical variables whenever necessary : @xmath141 \;=\ ; \int_{y_1 > 0 } d\tau dy^1 dy^2 \ , \big[\frac{1}{2 } \big ( \partial_\tau{\tilde \sigma } * \partial_\tau{\tilde \sigma } + \partial_j{\tilde \sigma } * \partial_j{\tilde \sigma}+ m^2 { \tilde \sigma } * { \tilde \sigma}\big ) + v _ * ( { \tilde \sigma } ) \big ] \;,\ ] ] since @xmath142 amounts to @xmath143 .
it is obvious that the existence of a border at @xmath144 requires the introduction of a boundary condition for @xmath145 .
a non - trivial consequence of the non - linear relation between the @xmath146 s and the physical variables is that the neumann condition : @xmath147 becomes too strong in terms of the new variables : @xmath148|_{x^1 \to \varphi(x^2 ) } \;=\ ; 0\;.\ ] ] thus we should have not only @xmath149 , but also @xmath150 at the interface , since @xmath71 vanishes with a non - zero derivative there .
the last condition means that the derivative of @xmath151 in the direction of the curve @xmath76 should vanish , since @xmath152 is tangent to that curve .
we now turn to generalizations of the change of variables approach to deal with the general situation as arising from a nontrivial metric .
moreover , we shall not distinguish between sub and super - indices . ]
@xmath153_\star \,=\,i \ , \theta(x_1 , x_2)\;.\ ] ] one may wonder to what extent the method and results of sections [ sec : metric ] and [ sec : interface ] depend on choosing the moyal product to represent relation ( [ cuatro ] ) , which is equivalent to the weyl ordering prescription to define the map between operators and functions . to show that the change of variables is not restricted to this case
, we will now work in a normal order , which arises naturally from the holomorphic representation of two - dimensional systems .
this framework is particularly appropriate to construct scalar multisolitons , as described in @xcite and extended in @xcite , where the generalization of berezin approach to deformation quantization is applied to the case of arbitrary khler manifolds .
planar systems are characterized by creation and annihilation operators @xmath154 and @xmath155 that satisfy @xmath156\,=\,1 \,,\ ] ] and we shall use that representation to discuss different changes of variables . for future convenience ,
we symbolize them as @xmath157 we represent the algebra over the hilbert space @xmath158 with basis @xmath159 of eigenstates of @xmath160 ( coherent states ) . on this basis , the action of the operators is @xmath161 the scalar product in @xmath158 is defined by @xmath162 with @xmath163 with this scalar product , @xmath164 and @xmath165 are relatively adjoint , and the eigenstates @xmath166 of the hermitean operator @xmath167 form an orthonormal basis . besides
, we can obtain the integral representation of the identity operator : @xmath168 now we represent the algebraic structure of @xmath169 in the space of functions @xmath170 by defining a @xmath3-product .
to each operator @xmath171 we associate the function @xmath172 \equiv \frac{\langle z \vert a(\hat z , \hat{\bar z})\vert \bar z \rangle}{\langle z \vert \bar z \rangle } \;.\ ] ] this corresponds to the normal - ordering prescription in which all the @xmath160 are to the left of all the @xmath173 .
we define the @xmath3-product as @xmath174 \;.\ ] ] then @xmath175 it is immediate to verify that @xmath3-product reproduces the algebraic structure of @xmath169 : @xmath176_\star\,=\,1 \;.\ ] ] some useful relations are : @xmath177 @xmath178 @xmath179 where @xmath180 is the usual derivative . the integral is defined as @xmath181 the usual derivatives @xmath182 , @xmath183 are realized as @xmath3-product commutators , @xmath184_\star \;\;\ ; , \;\;\ ; \partial _ { \bar z } \,=\ , - [ z , \;]_\star \;.\ ] ] returning to eq .
( [ eq : gencommrel ] ) , it is useful to move to the complex plane @xmath185 : @xmath186 thus we have to study the commutation relation @xmath187_\star\,=\,\theta\big(\frac{1}{\sqrt 2}\,(w+\bar w),-\frac{i}{\sqrt 2}\,(w-\bar w)\big ) \;.\ ] ] we want to define new coordinates @xmath188 : @xmath189 such that @xmath190_\star\,=\,1 $ ] ; i.e. , the unknown function @xmath27 must satisfy @xmath191 where @xmath192 is the function @xmath0 expressed in terms of @xmath188 .
if this is the case , then we can still apply the formulae from the previous section . for instance @xmath193 where @xmath194
the integrals will be of the form @xmath195 and comparing with eq.([eq : integrz ] ) , we obtain the integration measure @xmath196 where @xmath197 denotes the jacobian @xmath198 finally , from ( [ eq : derivz ] ) , two possible inner derivations are @xmath199_\star \,=\,\partial_z \tilde a(z , \bar z ) \;,\ ] ] @xmath200_\star \,=\,\partial_{\bar z } \tilde a(z , \bar z ) \;,\ ] ] being @xmath201 the inverse function of @xmath202
. they satisfy leibniz rule and @xmath203 let us consider the special and important case of conformal transformations @xmath204 with @xmath27 an analytic function . applying eqs.([eq : rel1])-([eq : rel3 ] ) to our present situation , we see that @xmath27 must satisfy the condition @xmath205 this is of course much simpler than the general condition ( [ eq : condition1 ] ) , but it will only be useful to deal with some special cases . for instance , we use this to study the rotation - invariant case @xmath206_\star \,=\,i \frac{\theta_0}{2 } ( x_1 ^ 2 + x_2 ^2)\;.\ ] ] in the complex plane , @xmath207_\star\,=\,\theta_0 \ , w \bar w \;.\ ] ] from the condition ( [ eq : condition2 ] ) , @xmath208 has to satisfy @xmath209 this suggests to take @xmath210 where @xmath211 is a parameter with units of length and @xmath212 is a dimensionless constant . indeed , @xmath213 so , @xmath212 must satisfy @xmath214 according to ( [ eq : measure2 ] ) , the integrals will be of the form @xmath215 since @xmath216 the derivations are , using the chain rule , @xmath217 @xmath218 from @xmath219 and eqs.([eq : rel1])-([eq : rel3 ] ) , the fundamental @xmath3-products are @xmath220 ( no summation over repeated indices ) and @xmath221 for @xmath222 .
it is worth noting that with this star product , @xmath223 is also rotation - invariant .
we have shown how , in some special cases , planar noncommutative theories with a space - dependent @xmath0 can be equivalently described in terms of a constant @xmath0 , by performing a suitable change of variables . in this way
, we have been able to find an explicit representation of the noncommutative product , and to deal with its associated calculus ( i.e. , derivatives and integration ) .
we applied that procedure in section [ sec : metric ] to the concrete example of a @xmath224 depending on only one coordinate .
we have also unravelled the connection between space - dependent noncommutative parameter @xmath224 and associative noncommutativity in a non - trivial background metric @xmath24 .
as explained at the end of this section , apart from its interest _ per se _ , this connection can be useful in the solution of selfduality noncommutative soliton equations which can be derived , one from the other if one introduces an appropriate background metric .
the resulting @xmath3-product was later on ( section [ sec : interface ] ) adapted to the example of a smooth interface @xmath225 dividing two regions with different constant @xmath0-values . the resulting noncommutative field theory obtained in this way
could be used , for example , to study the physics of a reduced ( relativistic ) lowest landau level description , when the magnetic field is not homogeneous , but rather has two different constant values in two respective regions ( which have @xmath225 as a boundary ) .
finally , in section [ sec : general ] , we sketched the extension of the change of variables to more general situations , by taking advantage of the holomorphic representation for planar systems .
the simple and yet important case in which the redefinition of variables is realized by a conformal transformation , has been analyzed in [ conformal ] , applying the results to the particular case of a rotation - invariant @xmath0 .
a crucial point in the method is , of course , to find an adequate change of variables .
this , in turn , depends on the explicit form of @xmath224 , in particular on its symmetries .
that is the reason why it does not seem possible to construct an _
explicit _ coordinate transformation to address the general case .
nevertheless , we believe that the present results can contribute to an understanding of the properties of theories with a space - dependent noncommutativity in a simpler way , by mapping them to systems equipped with the standard moyal product .
d. h. c. , c. d. f and g. t. have been supported by a fundacin antorchas grant .
d. c and f. a. s. have been partially supported by unlp , cicba , conicet and anpcyt ( pict grant 03 - 05179 ) .
bib a. connes , m. douglas and a. schwarz , jhep * 9802 * 003 ( 1998 ) .
v. schomerus , jhep * 9906 * 030 ( 1999 ) .
n. seiberg and e. witten , jhep * 9909 * 032 ( 1999 ) .
m. ciccolini , c.d .
fosco and a. lopez .
j.phys.a * 35 * , 10077 ( 2002 ) . c. d. fosco , a. lopez
. j. phys .
* a37 * , 4123 ( 2004 ) .
a. connes , _ noncommutative geometry _ , academic press , 1994 .
m. kontsevich , lett .
phys . * 66 * , 157 ( 2003 ) . m. douglas and n. nekrasov , rev .
phys . 73 : 977 - 1029 ( 2001 ) . l. cornalba and r. schiappa .
phys . * 225*,33 ( 2002 ) .
j. madore , s. schraml , p. schupp and j. wess .
j. c * 16 * , 161 ( 2000 ) .
w. behr and a. sykora , hep - th/0309145 .
a. das and j. frenkel .
d * 69 * , 065017 ( 2004 ) .
x. calmet , hep - ph/0305027 .
r. jackiw and s. y. pi , phys .
* 88 * ( 2002 ) 111603 .
r. jackiw , s. y. pi and a. p. polychronakos , annals phys .
* 301 * ( 2002 ) 157 .
g. amelino - camelia and s. majid , int .
j. mod .
a * 15 * , 4301 ( 2000 ) .
a. cattaneo and g. felder . commun.math.phys .
* 212 * , 591 ( 2000 ) .
d. h. correa , e. f. moreno and f. a. schaposnik , phys .
b * 543 * , 235 ( 2002 ) .
d. h. correa , p. forgacs , e. f. moreno , f. a. schaposnik and g. a. silva , `` noncommutative 3 dimensional soliton from multi - instantons , '' arxiv : hep - th/0404015 , jhep in press . j. madore gr - qc/9906059 . c. d. fosco and g. torroba .
planar field theories with space - dependent noncommutativity .
hep - th/0405090 .
r. gopakumar , s. minwalla and a. strominger , jhep * 0005 * , 020 ( 2000 ) . m. spradlin and a. volovich , jhep * 0203 * , 011 ( 2002 ) .
|
we consider planar noncommutative theories such that the coordinates verify a space - dependent commutation relation .
we show that , in some special cases , new coordinates may be introduced that have a constant commutator , and as a consequence the construction of field theory models may be carried out by an application of the standard moyal approach in terms of the new coordinates .
we apply these ideas to the concrete example of a noncommutative plane with a curved interface .
we also show how to extend this method to more general situations .
|
one possible way to investigate rare ultra energetic collisions of cosmic ray primaries with the air is to investigate the cherenkov light emitted in atmospheric showers . charged daughter particles produced along a hadronic cascade have mostly very high velocities and therefore emit in the air cherenkov photons which in principle carry an information about longitudinal shower development including an information on the characteristics of the primary collision , e.g. particle multiplicity . as optical photons suffer relatively smaller absorption in the atmosphere , their density on the ground is higher than that of charged particles and their lateral distribution is broader .
the cherenkov light highlights differences in the mechanism of hadronic collisions that are expected to occur in the first stage of the shower development and that can propagate further during development of the cascade .
since available experimental data do not contradict to the fact that the abundance of @xmath0fe nuclei in the primary flux may be above @xmath1 @xcite , a non negligible portion of iron nitrogen or iron oxygen nucleus
nucleus collisions can be observed in upper atmosphere .
the number of nucleons participating in such collisions may well grow up to a few tens and a fireball of hot and dense elementary matter of quarks and gluons ( qgp ) can be created .
its thermal equilibrium is achieved in rapid statistical energy distribution among partons . as a result
an almost ideally thermalized system of nucleons , pions and kaons ( heavier baryons and mesons are neglected in our approach ) is produced in the primary collision .
the energy density of the fireball can be estimated from the bjrken formula @xcite , @xmath2 , where @xmath3 is the radius of the smaller nucleus , the formation time was chosen as @xmath4 and @xmath5 is c.m.s .
energy of produced particles in the unit rapidity interval in the vicinity of its c.m.s .
energy densities and corresponding upper bounds of the transition temperatures achieved in our simulations of the primary interaction of the cosmic iron nucleus with the air nitrogen are shown in fig.[f01 ] as functions of the fermi energy variable @xmath6 , where @xmath7 is an invariant mass of a pair of interacting nucleons @xcite .
when the number of wounded nucleons @xmath8 in the first collision is greater than a certain value ( typically @xmath9 ) , we assume that several nucleons melt down into the qgp phase and a certain part of the total mass of the fireball @xmath10 is released .
the model parameter @xmath11 ( @xmath12 ) is assumed to be independent of the collision energy and the system size .
a portion of released energy is channeled to nucleons with momenta generated according to the relativistic boltzmann distribution isotropically with respect to the fireball center of mass system .
the remaining energy is then used to create a number of @xmath13 and @xmath14 mesons with isotropic and relativistic boltzmann distributed momenta in fireball system .
thermalized hadronic species are produced with probabilities @xmath15 and with approximately equal abundances of different charges .
chemical freeze out is controlled by a constant temperature ( @xmath16 ) and zero chemical potentials .
multiplicities of light mesons produced in the initial interaction , as calculated in our model , are shown in fig.[f02 ] .
to achieve the quantitative results , the corsika simulation code @xcite has been adapted to calculate the contributions of the cherenkov photon intensity stemming from the quark
gluon equilibration in primary collisions of the cosmic iron nuclei with the air .
the qgsjet01 model of high energy interactions has been employed ; the geisha procedure has been used to treat hadronic collisions at energies @xmath17 .
it is worth to note that the uncertainty related to the modelling of standard hadronic interactions turns out to be of minor influence , as the present investigation of qgp effects relies upon comparative characteristics of the detectable photons .
we analyze the contributions of cherenkov light to the lateral ground level distribution functions of iron induced showers for a given primary energy ( @xmath18 ) averaged over 100 showers .
the primary zenith angle was fixed at zero degrees ; the altitude of the initial interaction was left free .
the kinetic energy cutoffs for hadrons and muons have been set to @xmath19 ; for electrons / positrons and photons we use cutoffs of @xmath20 and @xmath21 , respectively . in all calculations a thinning level @xmath22 has been adopted . the cherenkov light has been collected by a row of 100 ideal detectors with detection area @xmath23 and spacing @xmath24 located in north south direction at a height of @xmath25 above the sea level ( @xmath26 ) .
in all simulations the standard u.s .
atmosphere has been used ; propagation of the cherenkov light has been corrected for attenuation in the air .
in our simple model , more secondary particles are created in the first interaction when qgp is formed ( qgp shower ) compared to the ordinary fe
n collisions ( fe shower ) .
air showers following qgp production have their maxima higher in the atmosphere and produce noticeably more cherenkov light at higher altitudes ; this effect grows with increasing incident energy . in fig.[f03 ]
the average ground level densities and ratios of total numbers of cherenkov photons produced along the qgp and fe showers are depicted as functions of the slant depth of their birth .
average ratios of the ground level cherenkov photon lateral densities from the qgp and fe air cascades are depicted in fig.[f04 ] .
the photon core densities of the qgp showers are typically by a factor of 1.2 smaller than the core densities of the fe showers of the same primary energy . at very high primary energies , @xmath27 , the relative fraction of photons at outer parts of the qgp shower increases .
thus the qgp effects could be visible comparing the number of cherenkov photons reaching the detector level relatively far from the shower core with all detected photons .
any influence of the transverse momenta of the direct hadrons in the first interaction becomes negligible as in the case of very energetic primary the lateral profile of shower is dominated by interactions of particles at lower energies in the later part of the shower development .
the qgp showers behave similarly as ordinary hadronic showers initiated by the iron primary of somewhat lower energy . however , the lateral spread of cherenkov photons caused by the equilibrated qgp products is larger .
all these effects depend on both the primary energy and the distance between detector and shower maximum ; the latter observable is also energy dependent .
the lateral ground level cherenkov photon densities of the extended air showers can be used to identify showers that follow the qgp transformation in the first interaction .
it is worth noting that similar qgp related effects can be observed in lateral distributions of high momenta muons detected underground @xcite .
99 s.p.swordy et al . ,
18 ( 2003 ) 129 .
m.gadicki , z.phys .
c * 66 * ( 1995 ) 659 .
j.d.bjrken , phys.rev . *
d27 * ( 1982 ) 140 .
d.heck et al . , report * fzka 6019 * ( 1998 ) .
j. ' idk ' y , astropart.phys .
17 ( 2002 ) 355 .
|
a simulation study of the effects resulting from creation of quark gluon plasma in high energy collisions of cosmic iron nuclei with the air has been carried out .
the cherenkov light emission has been found to reveal some features of the dynamics of high energy hadronic interactions .
|
new massive gravity ( nmg ) is a novel parity - preserving , unitary @xcite , power - counting super - renormalizable @xcite , three - dimensional model describing the propagation of a massive positive - energy graviton with two polarization states of helicity @xmath1 in a minkowski vacuum , whose linearized version is equivalent to the pauli - fierz theory for a massive spin-2 field in three dimensions .
the action of nmg consists of a `` wrong sign '' einstein - hilbert term plus a quadratic curvature term given by a precise combination of the ricci tensor and the curvature scalar , which introduces a mass parameter @xcite . as with other massive gravity theories ,
nmg also admits black hole - type solutions with several asymptotics and additional parameters @xcite .
even though this last feature could challenge the usual einstein - hilbert gravity , it is seen that the definition of mass in this new type of black holes is a conserved charge computed from a combination of the black hole parameters , which satisfies the first law of thermodynamics . in this contribution
we present the results of a study on the stability of the three - dimensional lifshitz black hole found in the context of the so - called new massive gravity ( nmg ) @xcite .
these solutions exhibit the anisotropic scale invariance , @xmath2 , @xmath3 , where @xmath4 is the dynamical critical exponent . specifically
, we deal with the solutions found for the particular case of @xmath0 and a precise value of the mass parameter .
we aimed to determine the stability of this kind of black holes by calculating the quasinormal modes ( qnm ) and quasinormal frequencies ( qnf ) of scalar and spinorial matter fields in the probe limit , _
i.e _ , there are no backreaction effects upon the asymptotically lifshitz black hole metric .
qnm do not depend of the particular initial perturbation that excited the black holes , while qnf depend only on the parameters of the black hole .
for this reason qnf are considered the `` footprints '' of this structure .
the study of qnm has motivated the development of numerical and analytical techniques for their computation ( @xcite ) .
qnm are important in astrophysical grounds as a way of possible detection of black holes through the observation of its gravitational wave spectrum .
they are also relevant in the context of gauge / gravity correspondence @xcite , where the hawking temperature of the black hole is related to the temperature of a thermal field theory defined at the boundary , and qnf can be related to the dual field theory relaxation time . in this work we focus in the application of qnm as probes of stability .
moreover , the qnf we obtain are useful to calculate the area spectrum of lifshitz black holes . actually , according to bekenstein the horizon area of a black hole
must be quantized @xcite , so that the area spectrum has the form @xmath5 , with @xmath6 a dimensionless constant to be determined . in this way
, we find this spectrum by using an approach suggested by maggiore @xcite , who claims that qnm should be described as damped harmonic oscillators , and the module of the entire qnf is the proper physical frequency that should be used in the bohr - sommerfeld quantization of the adiabatic invariant @xmath7 . in the following sections we describe the model we dealt with , the equations for the scalar and spinorial perturbations , the numerical analysis , and the calculation of the area spectrum . in the last section we discuss our results and conclude .
the black hole we study along this paper is a solution of the nmg @xmath8-dimensional action , @xmath9\,,\ ] ] where @xmath10 is the so - called `` relative '' mass parameter , and @xmath11 is the three - dimensional cosmological constant .
this theory has a black hole solution with lifshitz scaling @xmath0 given by @xmath12 where @xmath13 and @xmath14 with @xmath15 an integration constant and @xmath16 .
also , the nmg admits as a solution , the well - known baados - teitelboim - zanelli ( btz ) black hole with the dynamical critical exponent @xmath17 .
the kretschmann scalar for the metric ( [ metric ] ) , @xmath18\,,\ ] ] diverges for @xmath19 .
thus , the black hole solution has a genuine spacetime singularity at the origin . from the penrose - carter diagram in fig.[penrose ]
we see that the spacetime singularity at @xmath20 is light - like covered by a regular event horizon at @xmath21 .
is light - like and covered by a regular event horizon @xmath22.,height=302 ]
in order to consider the response of the lifshitz black hole background to a scalar perturbation we consider a scalar field obeying the klein - gordon equation , @xmath23 where @xmath10 is the mass of the field @xmath24 .
after decomposing the angular part of the field this equation takes the following form , @xmath25 although this equation has an analytical solution , which we will show later , it is a useful test for our numerical code to check the numerical results . with this goal
we need to rewrite ( [ kg3 ] ) in a shrdinger - like form , @xmath26 where @xmath27 is the so - called _ tortoise _ coordinate given by @xmath28 \,,\ ] ] and the scalar effective potential can be written as @xmath29 now let us return to eq.([kg3 ] ) and discuss its analytical solution .
first of all we set the time dependence of the scalar field as @xmath30 , and then redefine the radial coordinate as @xmath31 .
the resultant equation has the following solution in terms of the heun confluent functions , @xmath32 where @xmath33 and @xmath34 are integration constants , while @xmath35 , and @xmath36 . after imposing dirichlet condition at infinity and the boundary condition of ingoing waves at the horizon we obtain the quasinormal frequency spectrum given by @xmath37\,,\ ] ] where @xmath38 is a positive integer .
these same frequencies were also obtained in subsequent works@xcite by using equivalent analytical procedures .
the imaginary part of the fundamental frequency ( @xmath39 ) is negative provided that @xmath40 while the asymptotic frequency ( @xmath41 ) is given by @xmath42 thus , since the imaginary part of the quasinormal frequencies is negative provided that the parameters respect the relation ( [ condition ] ) , we can conclude that the model is stable under scalar perturbations .
let us consider a spinorial field perturbing the lifshitz black hole background , which obeys the covariant dirac equation , @xmath43 where greek indices refer to spacetime coordinates @xmath44 , and the latin indices enclosed in parentheses describe the flat tangent space in which the triad basis @xmath45 is defined .
the spinor covariant derivative @xmath46 is given by @xmath47\,,\ ] ] where @xmath48 is the spin connection , @xmath49 @xmath50 are the metric connections , and the gamma matrices we employ along this work are @xmath51 , @xmath52 , and @xmath53 . in order to solve the dirac equation in a numerical way once again we should adapt it to a shrdinger - like form .
first , we redefine the two component spinor as , @xmath54 as @xmath55^{1/4}e^{-i\omega t+im\phi}e^{i\theta/2}r_{+}(r),\nonumber
\\ \psi_{2}&=&\left[a(r)\delta\right]^{1/4}e^{-i\omega t+im\phi}e^{-i\theta/2}r_{-}(r)\,,\end{aligned}\ ] ] where @xmath56 and in addition we make @xmath57 to finally have @xmath58 where @xmath59 are the superpartner potentials , @xmath60 written in terms of the so - called superpotential @xmath61 , @xmath62 in the case of a massless spinor ( @xmath63 ) the superpartner potentials reduce to @xmath64
although in the scalar case we already found the corresponding qnf , the numerical analysis is an interesting tool to verify the applicability of certain numerical methods in asymptotically lifshitz spacetimes .
specifically , here we use the finite difference and the horowitz - hubeny @xcite methods . let us begin with the scalar perturbation . by rewriting eq.([sc - pot ] ) in terms of a new variable @xmath65 , we obtain @xmath66\,,\ ] ] where @xmath67 .
the parable in brackets tends to infinity as long as @xmath68 , which is consistent with the breitenlohner - freedman - type bound for the present case .
the roots of this polynomial potential are given by @xmath69\,.\end{aligned}\ ] ] if @xmath70 , we see that @xmath71 .
thus , going back to the original variable @xmath72 , the roots of the potential are @xmath20 with double multiplicity , @xmath73 and @xmath74 ( @xmath75 and @xmath76 are excluded as @xmath77 ) .
then , since @xmath78 is the biggest root and @xmath79 tends to @xmath80 when @xmath72 tends to @xmath80 , the potential is positive - definite in the region @xmath81 .
therefore , the quasinormal modes for @xmath70 are necessarily stable @xcite . imaginary part of scalar quasinormal frequencies .
we display the results obtained using different approaches.,width=226,height=302 ] imaginary part of scalar quasinormal frequencies .
we display the results obtained using different approaches.,width=226,height=302 ] in figs.[escalar_varm]-[comparacao ] we show the numerical results for the decay of the scalar field and a comparison of the scalar qnf obtained by three different approaches , analytical calculation , finite difference , and horowitz - hubeny methods .
our results reinforce the conclusion already found analytically , the @xmath0 lifshitz black hole is stable under scalar perturbations .
moreover , according to fig.[comparacao ] , the numerical results have a very good agreement with the analytical calculation .
however , the horowitz - hubeny method gives unreliable results .
we discarded explanations related to the not - convergence of the frequencies or the size of the black holes ocurring in other cases @xcite , in spite of that , there is no clear explanation for the limitation of the method . in our case
the asymptotic behavior of the black hole under study might play an important role in the convergence of the method .
nonetheless , a general criteria for the convergence of the horowitz - hubeny method remains an open question .
decay of massless spinor with @xmath82 and black hole mass @xmath83 for different values of the azimuthal parameter @xmath10.,width=226,height=302 ] decay of massless spinor with @xmath82 and black hole mass @xmath83 for different values of the azimuthal parameter @xmath10.,width=226,height=302 ] now turning to the case of spinorial perturbations , we can analogously analyze the superpartner potentials ( [ potential - weyl ] ) and its derivatives , which we rewrite as @xmath84\,,\ ] ] @xmath85\right\ } \,.\ ] ] we see that outside the event horizon @xmath86 is positive - definite if @xmath87 , and @xmath88 if @xmath89 . whereas @xmath90 is positive - definite if @xmath89 , and @xmath91 if @xmath87 .
furthermore , we notice that if @xmath92 , we have a free - particle case .
the decaying behavior of the massless spinor is given in figs.[espinorm10]-[espinorm15 ] .
thus , we conclude that the @xmath0 lifshitz black hole is stable under massless spinorial perturbations .
in this section we use our analytic results for the qnf to find the area spectrum of the lifshitz black hole . in order to apply maggiore s method
we need the asymptotic value of the qnf given in ( [ asymptotic ] ) , so that the proper physical frequency of the damped harmonic oscillator equivalent to the black hole qnm turns out to be @xmath93 now we can calculate the adiabatic invariant @xmath94 where @xmath95 is the adm mass of the lifshitz black hole , and @xmath96 is the change of proper frequency between two neighboring modes , which can be obtained from ( [ physfreq ] ) .
thus , @xmath97 which is quantized via bohr - sommerfeld quantization in the semiclassical limit . then , expressing this equation in terms of the horizon area @xmath98 , with @xmath99 , we finally obtain the spectrum we look for , @xmath100 with @xmath101 an integer number .
as we can see , the horizon area of the lifshitz black hole is quantized and equally spaced , a result not expected from a theory containing higher order curvature corrections . in addition , since this black hole has an entropy proportional to its horizon area @xcite , our result ( [ areasp ] ) also says that the entropy is quantized with an evenly spaced spectrum .
subsequent studies on area spectrum of other kind of lifshitz black holes @xcite seem to confirm this last conclusion .
we have studied the stability of the three - dimensional lifshitz black hole under scalar and spinorial perturbations in the probe limit through the computation of quasinormal modes . in addition , we have found the event horizon area quantization as an application of the results for quasinormal modes using maggiore s prescription .
regarding the scalar perturbations , we found that the qnf are purely imaginary and negative pointing out a stable equilibrium configuration for the lifshitz black hole .
our numerical results perfectly agree with the analytical calculation in the case of the finite difference method , which allows to obtain the temporal behavior of the fields showing all the stages of the decay .
however , in the case of the horowitz - hubeny method the frequencies values fail to be obtained .
an important observation is that apart from the numerical factor , the asymptotic scalar frequency is the same as that calculated in the hydrodynamic limit in the context of gauge / gravity duality @xcite . concerning the spinorial perturbation
, our numerical results also show probe massless spinors decaying , and therefore , the @xmath0 lifshitz black hole is stable under spinorial perturbations too . additionally , using the asymptotical form of the analytical qnf we calculated the area spectrum of the lifshitz black hole using maggiore s method .
our results show that the horizon area and the entropy are quantized and their spectra are equally spaced . at last
, we should stress that the definite answer on the stability of the lifshitz black hole should come from gravitational perturbations , mainly from the tensor part of the metric perturbation , since nmg admits the propagation of gravitational waves contrary to einstein gravity in three dimensions , that has no propagating degrees of freedom .
9 nakasone m and oda i , _ prog . theor .
phys . _ * 121 * , 1389 ( 2009 ) .
deser s , _ phys .
lett . _ * 103 * , 101302 ( 2009 ) .
clement g , _ class .
* 26 * , 105015 ( 2009 ) ; oliva j , tempo d and troncoso r , _ jhep _ * 0907 * , 011 ( 2009 ) ; giribet g , oliva j , tempo d and troncoso r , _ phys .
d _ * 80 * , 124046 ( 2009 ) ; ghodsi a and moghadassi m , _ phys .
b _ * 695 * , 359 ( 2011 ) .
|
we consider scalar and spinorial perturbations on a background described by a @xmath0 three - dimensional lifshitz black hole .
we obtained the corresponding quasinormal modes which perfectly agree with the analytical result for the quasinormal frequency in the scalar case .
the numerical results for the spinorial perturbations reinforce our conclusion on the stability of the model under these perturbations .
we also calculate the area spectrum , which prove to be equally spaced , as an application of our results .
|
period - doubling transition to chaos has been extensively studied in a one - parameter family of one - dimensional ( 1d ) unimodal maps , @xmath0 where @xmath1 is a state variable at a discrete time @xmath2 . as the control parameter @xmath3
is increased , the 1d map undergoes an infinite sequence of period - doubling bifurcations accumulating at a critical point @xmath4 , beyond which chaos sets in . using a renormalization group ( rg ) method , feigenbaum @xcite has discovered universal scaling behavior near the critical point @xmath4 .
here we are interested in the period doublings in a system consisting of two 1d maps with a one - way coupling , @xmath5 where @xmath6 and @xmath7 are state variables of the first and second subsystems , @xmath3 and @xmath8 are control parameters of the subsystems , and @xmath9 is a coupling parameter .
note that the first ( drive ) subsystem acts on the second ( response ) subsystem , while the second subsystem does not influence the first subsystem .
this kind of unidirectionally - coupled 1d maps are used as a model for open flow @xcite . a new kind of non - feigenbaum scaling behavior was found in a numerical and empirical way near a bicritical point @xmath10 where two critical lines of period - doubling transition to chaos in both subsystems meet @xcite .
for this bicritical case , a rg analysis was also developed and the corresponding fixed point , governing the bicritical behavior , was numerically obtained by directly solving the rg fixed - point equation using a polynomial approximation @xcite . in this paper , using two different methods , we also make the rg analysis of the bicriticality , the results of which agree well with those of previous works .
this paper is organized as follows . in sec .
[ sec : sb ] we study the scaling behavior near a bicritical point @xmath10 , corresponding to a border of chaos in both subsystems , by directly following a period - doubling sequence converging to the point @xmath10 for a fixed value of @xmath9 . for this bicritical case , a new type of non - feigenbaum critical behavior appears in the second subsystem , while the first subsystem is in the feigenbaum critical state . employing two different methods
, we make the rg analysis of the bicritical behavior in sec .
[ sec : ra ] . to solve the rg fixed - point equation
, we first use an approximate truncation method @xcite , corresponding to the lowest - order polynomial approximation .
thus we analytically obtain the fixed point , associated with the bicritical behavior , and its relevant eigenvalues .
compared with the previous numerical results @xcite , these analytic results are not bad as the lowest - order approximation . to improve accuracy
, we also employ the `` eigenvalue - matching '' rg method @xcite , equating the stability multipliers of the orbit of level @xmath11 ( period @xmath12 ) to those of the orbit of the next level @xmath13 .
thus we numerically obtain the bicritical point , the parameter and orbital scaling factors , and the critical stability multipliers .
we note that the accuracy is improved remarkably with increasing the level @xmath11 .
finally , a summary is given in sec .
[ sec : sum ] .
in this section we fix the value of the coupling parameter by setting @xmath14 and directly follow a period - doubling sequence converging to the bicritical point @xmath10 , which corresponds to a border of chaos in both subsystems . for this bicritical case ,
the second subsystem exhibits a new type of non - feigenbaum critical behavior , while the first subsystem is in the feigenbaum critical state .
the unidirectionally - coupled 1d maps ( [ eq : ucm ] ) has many attractors for fixed values of the parameters @xcite . for the case @xmath15
, it breaks up into the two uncoupled 1d maps .
if they both have stable orbits of period @xmath16 , then the composite system has @xmath16 different stable states distinguished by the phase shift between the subsystems .
this multistability is preserved when the coupling is introduced , at least while its value is small enough . here
we study only the attractors whose basins include the origin @xmath17 .
such attractors become in - phase when @xmath18 and @xmath15 .
stability of an orbit with period @xmath19 is determined by its stability multipliers , @xmath20 here @xmath21 and @xmath22 determine the stability of the first and second subsystems , respectively .
an orbit becomes stable when the moduli of both multipliers are less than unity , i.e. , @xmath23 for @xmath24 .
figure [ fig : sd ] shows the stability diagram of periodic orbits for @xmath14 . as the parameter @xmath3 is increased , the first subsystem exhibits a sequence of period - doubling bifurcations at the vertical straight lines , where @xmath25 . for small values of the parameter @xmath8
, the period of oscillation in the second subsystem is the same as that in the first subsystem , as in the case of forced oscillation .
as @xmath8 is increased for a fixed value of @xmath3 , a sequence of period - doubling bifurcations occurs in the second subsystem when crossing the non - vertical lines where @xmath26 .
the numbers inside the different regions denote the period of the oscillation in the second subsystem .
we consider a pair of the parameters @xmath27 , at which the periodic orbit of level @xmath11 ( period @xmath12 ) has the stability multipliers @xmath28 . hence , the point @xmath27 corresponds to a threshold of instability in both subsystems .
some of such points are denoted by the open circles in fig .
then such a sequence of @xmath27 converges to the bicritical point @xmath10 , corresponding to a border of chaos in both systems , with increasing the level @xmath11 .
the bicritical point is denoted by the solid circle in fig . 1 . to locate the bicritical point with a satisfactory precision
, we numerically follow the orbits of period @xmath29 up to level @xmath30 in a quadruple precision , and obtain the sequences of both the parameters @xmath27 and the orbit points @xmath31 approaching the origin .
we first note that the sequences of @xmath32 and @xmath33 in the first subsystem are the same as those in the 1d maps @xcite .
hence , only the sequences of @xmath34 and @xmath35 in the second subsystem are given in table [ tab : spo ] .
.sequences of the parameter and the orbit point , @xmath36 and @xmath37 , in the second subsystem . [ cols="^,^,^ " , ]
we have studied the scaling behavior of period doublings near the bicritical point , corresponding to a threshold of chaos in both subsystems . for this bicritical case
, a new type of non - feigenbaum critical behavior appears in the second ( response ) subsystem , while the first ( drive ) subsystem is in the feigenbaum critical state . employing the truncation and eigenvalue - matching methods
, we made the rg analysis of the bicritical behavior .
for the case of the truncation method , we analytically obtained the fixed point , associated with the bicritical behavior , and its relevant eigenvalues .
these analytic rg results are not bad as the lowest - order approximation . to improve the accuracy
, we also employed the numerical eigenvalue - matching rg method , and obtained the bicritical point , the parameter and orbital scaling factors , and the critical stability multipliers .
note that the accuracy in the numerical rg results is improved remarkably with increasing the level @xmath11 .
consequently , these numerical rg results agree well with the results obtained by a direct numerical method .
the results on the bicritical behavior in the abstract system of unidirectionally - coupled 1d maps are also confirmed in the real system of unidirectionally - coupled oscillators @xcite .
this work was supported by the korea research foundation under project no .
1998 - 015-d00065 and by the biomedlab inc .
some part of this manuscript was written during my visit to the center of nonliner studies in the institute of applied physics and computational mathematics , china , supported by the korea science and engineering foundation and the national natural science foundation of china .
i also thank profs . chen and liu and other members for their hospitality during my visit .
m. j. feigenbaum , j. stat .
* 19 * , 25 ( 1978 ) ; * 21 * , 669 ( 1979 ) .
kaneko , phys .
a * 111 * , 321 ( 1985 ) ; i. s. aranson , a. v. gaponov - grekhov and m. i. rabinovich , physica d * 33 * , 1 ( 1988 ) ; k. kaneko , physica d * 68 * , 299 ( 1993 ) ; f. h. willeboordse and k. kaneko , phys . rev . lett . *
73 * , 533 ( 1994 ) ; f. h. willeboordse and k. kaneko , physica d * 86 * , 428 ( 1995 ) ; o. rudzick and a. pikovsky , phys .
e * 54 * , 5107 ( 1996 ) ; j. h. xiao , g. hu and z. qu , phys .
* 77 * , 4162 ( 1996 ) ; y. zhang , g. hu and l. gammaiton , phys .
e * 58 * , 2952 ( 1998 ) .
b. p. bezruchko , v. yu .
gulyaev , s. p. kuznetsov , and e. p. seleznev , sov .
* 31 * , 268 ( 1986 ) .
a. p. kuznetsov , s. p. kuznetsov and i. r. sataev , int .
j. bifurcation and chaos * 1 * , 839 ( 1991 ) .
mao and j. greene , phys .
a * 35 * , 3911 ( 1987 ) ; s .- y . kim and h. kook , phys .
a * 178 * , 258 ( 1993 ) .
b. derrida , a. gervois and y. pomeau , j. phys .
a * 12 * , 269 ( 1979 ) ; b. derrida and y. pomeau , phys .
a * 80 * , 217 ( 1980 ) ; b. hu and j. m. mao , phys . rev .
a * 25 * , 1196 ( 1982 ) ; * 25 * , 3259 ( 1982 ) ; phys .
a * 108 * , 305 ( 1985 ) ; j. m. mao and r. h. g. helleman , phys .
a * 35 * , 1847 ( 1987 ) ; s .- y . kim and b. hu , phys .
a * 41 * , 5431 ( 1990 ) .
b. p. bezruhko , yu . v. gulyaev , o. b. pudovochkin , and e. p. seleznev , sov .
. dokl . * 35 * , 807 ( 1991 ) .
r. s. mackay , ph .
d. thesis , princeton university , 1982 .
see eqs .
3.1.2.12 and 3.1.2.13 .
o. e. r@xmath38ssler , phys .
a * 71 * , 155 ( 1979 ) .
y . kim and w. c. lim ( to be published ) .
|
we study the scaling behavior of period doublings in two unidirectionally - coupled one - dimensional maps near a bicritical point where two critical lines of period - doubling transition to chaos in both subsystems meet .
note that the bicritical point corresponds to a border of chaos in both subsystems . for this bicritical case ,
the second response subsystem exhibits a new type of non - feigenbaum critical behavior , while the first drive subsystem is in the feigenbaum critical state . using two different methods
, we make the renormalization group analysis of the bicritical behavior and find the corresponding fixed point of the renormalization transformation with two relevant eigenvalues .
the scaling factors obtained by the renormalization group analysis agree well with those obtained by a direct numerical method . 2
|
launched respectively in 2004 and 2008 , _ swift _ and _ fermi _ jointly have become the primary workhorses for extensive studies of gamma - ray bursts ( grbs ) . the burst alert telescope ( bat )
onboard _ swift _ , with its 5200 @xmath0 detector area ( @xmath1 @xmath0 effective area ) can detect very faint bursts down to @xmath2 ( barthelmy et al .
2005 ; gehrels et al . 2004 ) in the 15 - 150 kev band . due to rapid slew rate and high resolution instruments ,
_ swift _ has enabled redshift measurement of many grbs .
fermi hosts two instruments , namely gamma ray burst monitor ( gbm ) and large area telescope ( lat ) .
gbm , the primary instrument for studying grb prompt emission , covers a wide energy band ( 8 kev - 40 mev ; meegan et al .
2009 ) . together
_ swift _ and _ fermi _ have provided a wealth of prompt emission data .
swift has revealed that grbs are frequent at very high redshifts ( z ) .
the highest spectroscopic z is 8.2 ( tanvir et al . 2009 ; salvaterra et al . 2009 ) , which corresponds to a very young universe , even less than 5% of its present age . the high z of these objects make them valuable as high - z luminosity indicators .
indeed , correlations of some observables ( e.g. , source frame peak energy @xmath3 , lag @xmath4 , variability etc . ) with the energetics of a grb ( e.g. , isotropic equivalent energy @xmath5 , isotropic peak luminosity @xmath6 , collimation corrected energy @xmath7 ) have been found and discussed in the recent years ( e.g. , fenimore & ramirez - ruiz 2000 ; norris et al .
2000 ; amati et al . 2002
; schaefer 2003 , 2004 ; ghirlanda et al .
2004 ) . in principle , through these correlations an observable can be used to estimate the energetics and hence , the distance of the object .
therefore , it is suggested that grbs can be used as `` standard candles '' in the same way as type ia supernovae ( sne ) .
moreover , the redshift barrier can be pushed to a much higher value compared to sne ia .
these correlations are , however , empirical in nature , and may arise due to instrumental selection biases ( e.g. , see nakar & piran 2005 ; band & preece 2005 ; schaefer & collazzi 2007 ; collazzi et al . 2012 ) .
one way to confirm the reality of these correlations and understand the effect of selection biases is to examine the correlations within a grb .
for example , ghirlanda et al . (
2010 ) , using a set of 9 grbs with known redshifts , have shown that the @xmath3 - @xmath6 correlation , known as the yonetoku correlation ( yonetoku et al .
2004 ) , holds within the time - resolved data .
however , basak & rao ( 2012a ; br12a hereafter ) , using the same set of grbs , have shown that the @xmath3 - @xmath5 correlation ( amati correlation ; amati et al .
2002 ) breaks down in the time - resolved data .
they concluded that the amati correlation is not meaningful in time - resolved study .
however , the fact that a grb is made up of broad pulses ( see e.g. , norris et al .
2005 ) and that the spectrum evolves independently in the individual pulses demand that pulses should be analyzed separately .
br12a used 22 pulses of these grbs and found that the amati correlation not only holds but it is better than the average amati correlation . in this paper , we expand the pulse - wise study of grbs to include 43 pulses of 19 grbs .
our main motivation is to establish that the pulse - wise correlation is suitable to use grbs as standard candles for cosmological studies .
knowledge about the redshift evolution and other biases is very useful in this context .
with a larger sample compared to that used in br12a , we can perform some critical tests .
we first examine whether there is a redshift evolution of the pulse - wise amati correlation by dividing our sample in various redshift bins .
this is very crucial because the correlation has to hold for various redshifts if we want to use grbs as `` standard candle '' at various redshifts .
another interesting feature to be investigated is the dependence of the pulse properties in their sequence of occurrence which might put additional constraints for using a standard luminosity for a given pulse .
it has been suggested that grbs tend to be harder at the beginning than the rest of the burst ( crider et al .
1997 ; ghirlanda et al . 2003 ; kaneko et al . 2003 ; ryde and peer 2009 ) .
this hardness should reflect in the peak energy of the spectrum . for a given energy ,
the peak energy of the first pulses should be biased towards higher values and will tend to give a systematic shift in the peak energy luminosity correlation .
we examine this possibility by dividing our sample into first / single pulses and rest of the pulses , and then studying the pulse - wise amati correlation for them .
following is the plan of this paper . in section 2
, we shall present our sample and the analysis technique .
results are discussed in section 3 .
major conclusions are discussed in section 4 . [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] [ table2 ] ) - log(@xmath5 ) data .
the parallel thin lines are 3@xmath8 scatter of the data .
the other thin lines are linear fit for the data in various redshift bins .
markers are : boxes for z<1 , filled circles for 1<z<2 and open circles for z>2 bins .
[ fig1],width=288 ]
in table [ table1 ] we show our set of grbs in chronological order .
the redshifts are quoted from the web page of jochen greiner . in columns 4 and 5
we specify the time interval for a given pulse . in the next three columns we show the parameters of band model fit .
note that the low energy photon index ( @xmath9 ) sometime crosses the `` synchrotron line of death '' of electron slow cooling regime ( i.e. , the value of @xmath9 is greater than -2/3 ; preece et al .
1998 ) . hence ,
these models are either unphysical or the spectrum is not synchrotron ( zhang & meszaros , 2002 ) . however , the @xmath10 of the fits are acceptable .
we are interested in measuring the average properties of the pulses , namely , peak of the ef(e ) spectrum and the bolometric flux .
looking at the @xmath10 of the fits we can safely say that these values should not vary much for any other model which gives acceptable fit .
the isotropic energy for the individual pulses are shown in the last column . in figure
[ fig1 ] , we show the scatter plot of @xmath3 vs. @xmath5 .
the correlation coefficients are as follows : r=0.86 with @xmath11 , @xmath12=0.86 with @xmath13 .
if we do not use the logarithmic values then the pearson correlation is 0.80 .
we note that there is a significant correlation in pulse - wise amati data .
note that due to larger sample with respect to br12a the significance of the correlation has improved ( previously @xmath14 was @xmath15 ) . to obtain the average amati correlation we use the data for these grbs , published by gruber et al .
the correlation coefficients are r=0.83 with @xmath16 , @xmath12=0.85 with @xmath17 , and r without taking logarithmic values is 0.69 with @xmath18 .
the thick line in figure [ fig1 ] is the linear fit to the logarithmic values of pulse - wise amati data .
the corresponding fit results are shown in table [ table2 ] ( first row ) .
note that the slope ( @xmath19 ) of the linear fit is @xmath20 , which gives an empirical relation as @xmath21 .
the value of the slope ( @xmath19 ) and the intercept ( k ) for 22 pulses , reported by br12a , were @xmath22 and @xmath23 .
the current values are : @xmath24 and @xmath25 .
@xmath19 is within 0.36@xmath26 and k is within 0.36@xmath26 of the previous values .
note that due to larger size of the present data , the values are now determined with better accuracy .
however , the intrinsic scatter ( @xmath8 ) has increased to @xmath27 from @xmath28 , but still within 0.25@xmath26 of the old value . in figure
[ fig1 ] , the lines parallel to the thick line show the 3@xmath8 scatter of the data . in order to investigate the possible evolution with redshift of the pulse - wise @xmath3-@xmath5 correlation , we divide our data set in three redshift bins : ( i ) z<1 ( contains 17 data points ) , ( ii ) 1<z < 2 ( 16 data points ) and ( iii ) z>3 ( 10 data points ) . in figure
[ fig1 ] , we show these different data by different markers .
we fit the individual scattered data by a straight line as before .
the thin lines represent these fits . in table
[ table2 ] ( 2nd , 3rd and 4th row ) , we show the k,@xmath19 , @xmath8 and @xmath10 of these fits .
we note that there are deviations of these lines from the average correlation line . however , the sample size is too low at present to conclude that these variations are significant . in figure
[ fig2 ] , we show the slope ( @xmath19 ; shown by filled boxes ) and the intercept ( k ; shown by filled circles ) of these linear fits as functions of z. as there are only three redshift bins , we can fit at most a straight line to the z-@xmath19 and z - k data .
the dashed lines in figure [ fig2 ] show the linear fits to the data .
the slope of the z-@xmath19 fit is @xmath29 , while that of the z - k fit is @xmath30 .
the errors quoted are nominal 90% errors .
the average values of k and @xmath19 are shown by open circle and open box respectively .
we note that there is an indication of evolution of the correlation . however ,
the errors in @xmath31 and @xmath32 are large , and hence , the values are not statistically significant .
redshift evolution of amati correlation and other prompt emission correlations have been studied by some researchers ( li 2007 ; tsutsui et al .
2008 ; ghirlanda et al . 2008
; azzam 2012 ) .
they have reported mild evolution of the correlation .
note that the evolution of tsutsui et al .
( 2008 ) is monotonic , but figure [ fig2 ] indicates that the there is no systematic evolution of the pulse - wise amati correlation . to further investigate this matter , we divide our sample into low and high redshift bins .
as there are different number of pulses in a given grb , the division is not exactly half .
the lower redshift bin contains 23 pulses upto redshift 1.096 , and the other sample contains the rest .
in figure [ c1 ] , we have shown the @xmath33 contour of k and @xmath19 for these two redshift bins .
the contour levels are 1@xmath26 for single parameter ( i.e. , @xmath34 ) , 1@xmath26 , 2@xmath26 and 3@xmath26 for two parameters
( i.e , @xmath35 , 4.61 and 6.17 , respectively ) .
note that they agree within @xmath34 .
hence , we conclude that the present data is still consistent with no evolution , within statistical limit .
yonetoku et al .
( 2010 ) , for example , have studied both amati and yonetoku correlation in a set of 101 grbs from different detectors .
they found that the yonetoku correlation has a redshift evolution at 2@xmath26 level , while the amati correlation has that at 1@xmath26 level . hence , the evolution of both amati and pulse - wise amati correlations are very weak and insignificant .
we divide our sample into single / first pulses ( 19 data points ) and the rest of the pulses ( 24 data points ) .
if the first / single pulses have harder spectrum , the peak energy for a given luminosity should have a bias towards higher value .
we fit linear curves to the log(@xmath3 ) - log ( @xmath5 ) data for both the cases .
the corresponding fit parameters are reported in table [ table2 ] .
figure [ fig3 ] shows the correlation for first / single ( open boxes ) and the rest of the pulses ( open circles ) .
the average linear fit to the full data is over - plotted as thick line .
note that the intercept ( k ) of the first / single pulses ( @xmath36 ) is higher than the rest of the pulses ( @xmath37 ) .
however , the slope ( @xmath19 ) is lower ( @xmath38 ) than the rest ( @xmath39 ) .
note that @xmath40 , hence the bias is nullified by a combination . to investigate this further
, we draw @xmath33 contour plot of k and @xmath19 for the two cases ( see figure [ c2 ] ) .
the blue contours at the right are those for the single / first pulses .
the contour levels are the same as in figure [ c1 ] .
note that the two sets agree with each other within 1@xmath26 .
hence , we conclude that the correlation is unaffected by the fact that the first / single pulses are harder . in order to compare the pulse - wise amati correlation with the familiar time - integrated amati correlation
, we use the recent results of nava et al .
they have studied the time - integrated spectrum of 136 grbs with known redshift .
we denote the slope as @xmath41 and the normalization by @xmath42 for their value to distinguish them from the present work . from nava
et al . ( 2012 ) we quote the following values . ( a ) for the full set of 136 grbs @xmath43 , @xmath44 , @xmath45 , @xmath46 .
( b ) for a complete set of 46 grbs @xmath47 , @xmath48 , @xmath49 , @xmath50 , ( c ) for the complementary sample of 90 grbs , @xmath51 , @xmath52 , @xmath53 , @xmath50 . in figure
[ compare ] , we have shown together the data points of the pulse - wise analysis and the time - integrated @xmath3-@xmath5 correlation of the complete sample ( nava et al .
we have shown the 3@xmath54 scatter in the time - integrated data by a shaded region .
in our analysis , we have normalized the @xmath3 by 100 kev and @xmath55 by @xmath56 erg . converting this value to be compatible with the nava et al .
( 2012 ) , we obtain @xmath57 , which is comparable those obtained by time - integrated analysis . the slope ( @xmath19 ) of the time - integrated correlation ( @xmath58 ) is also comparable to the pulse - wise correlation ( @xmath24 ) .
to compare these correlations , we have plotted the time - integrated correlation line ( dashed line ) and the pulse - wise correlation line ( thick line ) . in order to find where the time - integrated data of the 19 grbs used for the present study lie
, we also show the time - integrated values ( taken from david gruber private communication ) by open circles .
the linear fit to this time - integrated data is shown by dot - dashed line .
note that , all these lines are comparable . though the pulse - wise amati correlation is comparable to the time - integrated correlation in terms of correlation coefficient , compared to the time - integrated correlation of nava et al .
( 2012 ) pulse - wise amati correlation is tighter ( @xmath59 ) .
this points to the fact that pulse - wise correlation is indeed meaningful .
to summarize , we have enlarged our sample twofold compared to br12a to study the pulse - wise amati correlation .
we confirm that the correlation holds with a good correlation coefficient .
we could divide our sample in various redshifts and study the same correlation in these redshift bins .
we found no statistically significant evolution of the correlation with redshift .
possibly a larger sample is needed to further investigate this matter . to check whether the correlation has a bias due to the fact that the first / single pulse have a tendency for higher peak energy ( @xmath60 )
, we divided our sample into the first / single pulses and the rest of the pulses .
we confirm that the correlation remains unaffected by such a bias .
another interesting fact we notice from close inspection of figure [ compare ] is as follows . the time - integrated and pulse - wise correlation
are in general agreement with each other ( compare the thick line with the dashed line ) .
however , the best fit line ( dot - dashed ) to the time - integrated data of 19 grbs ( present sample ) lies slightly lower than that of the pulse - wise data .
though the slope of these lines are similar , their normalization is slightly different .
the difference in the normalizations of the pulse - wise correlation and the time - integrated correlation can be understood as due to the summing of the individual pulses in the individual grbs .
for example , if a grb has two pulses , the time averaged value of e@xmath61 would be the average value of individual e@xmath62 values of the pulses ( possibly biased towards the brighter pulse ) , whereas the value of @xmath5 would simply be the sum of individual pulse values .
hence we can expect a shift in the data point towards right by about a factor of two .
however , a larger sample is required to draw such conclusion , but at present , there is no significant evidence of different slope , dispersion and normalization between the `` two versions '' of the correlation .
the present work strongly indicates that the amati correlation is a pulse characteristics rather than an average grb property .
the strengthening of the amati correlation when taken pulse - wise is a clear vindication that the correlation is real and not due to any artifact of selection or observational biases
. it would be interesting to investigate all grb prompt emission correlations taking individual pulses so that a very tight multi - parameter correlation could be obtained which will eventually lead to the use of grbs as standard candles for cosmology .
this research has made use of data obtained through the heasarc online service , provided by the nasa / gsfc , in support of nasa high energy astrophysics programs .
rb would like to thank david gruber for providing useful information of _
fermi_/gbm grbs .
a sincere thank to david fanning ( fanning software consulting , inc . ) for online help of idl plots .
we thank the referee for valuable suggestions which improved the quality of the paper .
|
we make a detailed pulse - wise study of gamma - ray bursts ( grbs ) with known redshift detected by _
fermi_/gamma ray burst monitor ( gbm ) .
the sample contains 19 grbs with 43 pulses .
we find that the average peak energy is correlated to the radiated energy ( the amati relation ) for individual pulses with a correlation coefficient of 0.86 , which is slightly better than the correlation for the full grbs .
as the present correlation holds within grbs , it is a strong evidence supporting the reliability of such a correlation .
we investigate several aspects of this correlation .
( i ) we divide our sample into redshift bins and study the evolution of the correlation .
though there is a marginal indication of evolution of the correlation , we can conclude that the present data is consistent with no evolution .
( ii ) we compare the correlation in the first or single pulses of these grbs to that of the rest of the pulses , and confirm that the correlation is unaffected by the fact that first / single pulses are generally harder than the rest .
finally , we conclude that the pulse - wise amati correlation is more robust and it has the potential of refining the correlation so that grb study could be used as a cosmological tool .
[ firstpage ] gamma - rays : bursts methods : observational methods : statistical methods : data analysis gamma - rays : observations
( cosmology : ) early universe .
|
cyg x-1 was the first binary system in which the presence of a black hole was suggested ( bolton 1972 ; webster & murdin 1972 ) . for about a decade , it was the only object of that type .
many , sometimes exotic , models and scenarios were devised to avoid the presence of a black hole and to replace it with a neutron star . with the advent of subsequent black hole candidates ( lmc x-3 , a620@xmath400 ) , the motivation for such efforts substantially diminished . at present , there are no doubts about the presence of a black hole in the system . however , in spite of three decades of the investigations , there is still substantial uncertainty concerning the masses of both components .
the mass function is known rather precisely .
its most recent value was given by gies et al .
( 2003 ) : @xmath5 m@xmath1 .
there are also no doubts that the optical component must be close to filling its roche lobe .
gies & bolton ( 1986a , 1986b ) , analyzing the emission lines of the stellar wind , the rotational broadening of the absorption lines of the optical component and the photometric v - band light curve , found that the fill - out factor must be greater than 0.9 and its best value is in the range 0.95 to 1.the value of the inclination of the orbit is less certain . from the analysis mentioned above ( rotational broadening of the absorption lines of the optical component and the ellipsoidal light variations in the v - band light curve ) , the authors estimated the inclination of the orbit to be @xmath6 .
on the other hand , from polarimetric measurements ( in three colours ) dolan & tapia ( 1989 ) found the inclination @xmath7 @xmath8 , which , while different , is not inconsistent with the value of gies & bolton .
paczyski ( 1974 ) has shown that basing only on the lack of the x - ray eclipses and the fact that the optical component can not be larger than its roche lobe , one can obtain lower limits to the masses of both components as functions of the distance to the system .
paczyski found that the mass of the compact component has to be larger than 3.2 m@xmath1 if the distance is greater than 1.3 kpc .
he noticed also that , basing on stellar evolution consideration , the distance can not be larger than about 2.1 kpc .
the distance to cyg x-1 is not very well established , but , certainly , it is greater than 1.3 kpc .
margon , bowyer & stone ( 1973 ) and bregman et al .
( 1973 ) analyzed interstellar reddening in the direction of hde 226868 and found that its distance is about 2.5 kpc .
wu et al . ( 1982 ) , from a study of the uv colors of hde 226868 , concluded that its distance must be greater than 1.9 kpc .
bregman s value was confirmed by ninkov , walker & young ( 1987 ) , who got the distance of 2.5 @xmath0 0.3 kpc from the equivalent width of the h@xmath9 line .
hde 226868 lies only @xmath10 from the center of ngc 6871 ( the core of the cyg ob3 association ) and seems to be its member .
however , the estimates of the distance to the association ( or its core ) gave mixed results .
crawford , barnes & warren ( 1974 ) used three methods ( ubv photometry , the h - r diagram fitting and the equivalent width of the h@xmath11 line ) and obtained the distance to the association of about 2 kpc .
humphreys ( 1978 ) estimated the photometric distance to cyg ob3 to be about 2.3 kpc .
janes & adler ( 1982 ) quote in their catalogue the distance to ngc 6871 as only 1.8 kpc .
the value of 1.8 kpc as the distance to cyg ob3 was obtained again by garmany & stencel ( 1992 ) , who used the h - r diagram fitting method .
the discrepancy between the estimates of the distances to hde 226868 and to cyg ob3 cast the doubt on the membership of cyg x-1 in ngc 6871 .
fortunately , the extensive photometry and spectroscopy by massey , johnson & degioia - eastwood ( 1995 ) seemed to clarify the situation .
they estimated the distance to ngc 6871 as 2.14 @xmath0 0.07 kpc , which was in reasonable agreement with the independent ( h@xmath9 width ) estimate of the distance to hde 226868 ( ninkov et al .
let us note that a very similar value of the distance ( @xmath12 2.15 kpc ) was derived also by gies & bolton ( 1986a ) as a byproduct of their careful analysis of a large set of observational data for hde 226868/cyg x-1 binary system .
their result ( see the point @xmath13 m@xmath1 and @xmath14 m@xmath1 in their fig .
10 ) was obtained quite independently of any photometric considerations .
we should also note , however , that while herrero et al .
( 1995 ) do not discuss explicitly the value of the distance in their atmosphere modelling paper , their final model implies the value @xmath3 1.8 kpc .
finally , the question of cyg x-1 membership in ngc 6871 was definitely solved in recent paper by mirabel & rodriguez ( 2003 ) . comparing the high precision vlbi astrometry for cyg x-1 and hipparcos astrometry for the members of cyg ob3
, the authors convincingly demonstrated that cyg x-1 shares , quite precisely , common proper motion with cyg ob3 association . the velocity of cyg x-1 with respect to the sun is 70 @xmath0 3 km / s .
the relative space velocity of cyg x-1 with respect to cyg ob3 is only 9 @xmath0 2 km / s , which is typical of the random velocities of stars in expanding associations ( blaauw 1991 ) . after the work by massey et al .
( 1995 ) , several more estimates of distances to both cyg ob3 and hde 226868 were made and the results still exhibit a substantial scatter .
malysheva ( 1997 ) , using different photometry than massey et al .
( 1995 ) , obtained for the distance to cyg ob3 again @xmath12 1.8 kpc .
dambis , melnik & rastorguev ( 2001 ) calculated the trigonometric distance to cyg ob3 as a median value of the hipparcos distances to the 18 individual member stars and found @xmath15 = 2.3 ( + 1.4,@xmath40.6 ) kpc . in the case of cyg
ob3 their trigonometric distance agreed very well with the photometric distance of blaha & humphreys ( 1989 ) , although they found that , typically , hipparcos trigonometric distances to ob associations are by @xmath3 12% smaller than photometric distances of blaha & humphreys . a recent photometric distance estimate made by mikoajewska ( 2002 , private communication ) gave the result @xmath15 = 1.95 @xmath16 kpc .
lestrade et al . ( 1999 ) used the vlbi astrometry to estimate the distance to cyg x-1 and obtained @xmath15 = 1.0 @xmath2 2.3 kpc .
all this extensive effort leads to the conclusion that the distance to hde 226868/cyg x-1 binary system can not be significantly different from 2 kpc .
therefore , i believe that the best choice will be to follow massey et al .
( 1995 ) and use @xmath15 = 2.15 @xmath0 0.07 @xmath17 error ) or @xmath0 0.2 @xmath18 error ) kpc as the distance to cyg x-1 . as we shall see at the end of the paper
, this choice will be , quite independently , supported by evolutionary considerations .
the above value is used throughout the rest of this paper ( except when i assume the distance to be a free parameter ) . in those parts of the discussion where the distance is assumed to be a free parameter
, i shall consider the value 1.8 kpc as a reasonable lower limit to the distance of cyg x-1 .
in this section , i shall repeat paczyski s ( 1974 ) analysis , using the present day data .
this analysis is based only on solid observational facts and is , therefore , model independent .
let us remind that these solid facts include : ( i ) the value of the mass function , ( ii ) the lack of the x - ray eclipses , ( iii ) the obvious fact that the optical component can not be larger than its roche lobe , and ( iiii ) the spectral type and the photometry of the optical component ( used to derive its effective temperature and the bolometric correction ) .
the most recent determination of the mass function is @xmath19 m@xmath1 ( gies et al .
2003 ) , which differs only slightly from the earlier values @xmath20 m@xmath1 ( brocksopp et al . 1999 ) and @xmath21 m@xmath1 ( gies & bolton 1982 ) . the corresponding projected radius of the optical component orbit is @xmath22sin@xmath23 r@xmath1 . the spectral type of hde 226868 was determined by walborn ( 1973 ) as o9.7 iab .
this classification was confirmed by gies & bolton ( 1986a ) , by ninkov et al .
( 1987 ) and by herrero et al .
( 1995 ) and was used in most of the literature . on the other hand , massey et al .
( 1995 ) obtained , from their new massive spectroscopy of ob associations , the somewhat earlier spectral type - on9 ifa+ . according to the calibration of effective temperatures for late o - type supergiants derived by vacca , garmany & shull ( 1996 , hereafter vgs ) , one has @xmath24 = 32740 k for o9 ia star and extrapolation for o9.7 ia gives @xmath24 = 30690 k. on the other hand , herrero et al .
( 1995 ) used atmosphere models to reproduce the observed spectrum of hde 226868 and obtained @xmath25 = 32000 k as the best fit for the effective temperature of this star . both vgs and herrero et al .
calibrations were based on pure h - he models of the atmospheres that neglected the effects of metallic lines blanketing .
it is well known by now , that taking into account the line blanketing leads to the lower effective temperature for a given spectral type ( martins et al .
2002 , repolust et al .
2004 , markova et al .
the most recent calibration of effective temperatures for o - type supergiants ( accounting for the line blanketing effects ) was derived by markova et al .
their scale gives generally lower effective temperatures than vgs scale ( by up to 10 000 k for the earliest spectral types ) .
however , for the spectral type o9.7 i , both temperature scales converge to the value @xmath24 = 30700 k. it seems that , at the present state of knowledge , the most reasonable approach would be to assume that @xmath24 = 30700 k is the best estimate of the effective temperature of hde 226868 . to take into account the uncertainty of this estimate , in further discussion i shall consider a broad interval of @xmath26 k , as the range of the possible values of the effective temperature of hde 226868 . as the effective temperature - bolometric correction relation is not affected by the line blanketing ( martins et al .
2002 ) , it is possible to take this relation from vgs tables . for the effective temperature 30700 k , the appropriate value of the bolometric correction is @xmath27 = @xmath43.06
. the corresponding unreddened colour index should be @xmath28 ( schmidt - kaler 1982 ) .
the most recent photometry of hde 226868 by massey et al .
( 1995 ) gives @xmath29 and @xmath30 .
therefore the reddening is @xmath31 and the unreddened @xmath32 magnitude is @xmath33 ( following massey et al .
, i adopt @xmath34 ) . with the effective temperature and the bolometric correction given above
, we can then express the radius and luminosity of hde 226868 as functions of the distance : @xmath35 @xmath36 following the procedure of paczyski ( 1974 ) , we can now calculate the lower limits to the masses of both components as functions of the distance to the system ( under the assumptions listed at the beginning of this section ) .
the results of these calculations are given in the first part of table 1 .
& & & & + + + + 1.80&19.1&20.4&6.&5.46 + 1.95&20.7&25.4&7.4&5.53 + 2.08&22.0&30.2&8.3&5.59 + 2.15&22.8&33.0&8.7&5.62 + 2.22&23.5&36.0&9.2&5.65 + 2.35&24.9&41.9&10.2&5.69 + 2.50&26.5&49.5&11.3&5.75 + 2.70&28.6&60.9&12.9&5.82 + + + + 1.80&19.7&22.3&6.8&5.33 + 1.95&21.3&27.6&7.8&5.40 + 2.08&22.8&32.9&8.7&5.46 + 2.15&23.5&35.9&9.2&5.49 + 2.22&24.3&39.2&9.&5.51 + 2.35&25.7&45.7&10.8&5.56 + 2.50&27.3&53.9&12.0&5.62 +
2.70&29.5&66.3&13.7&5.68 + + + + 1.80&18.7&19.5&6.3&5.52 + 1.95&20.3&24.1&7.2&5.59 + 2.08&21.6&28.7&8.0&5.64 + 2.15&22.4&31.4&8.5&5.67 + 2.22&23.1&34.2&8.9&5.70 + 2.35&24.4&39.9&9.9&5.75 + 2.50&26.0&47.1&11.0&5.80 + 2.70&28.1&57.9&12.5&5.87 + + let us notice that the changes , comparing with paczyski s table 1 , are mostly due to substantial change in the effective temperature ( paczyski took @xmath24 = 25000 k ) and the corresponding change in the bolometric correction for hde 226868 .
let us also notice that the effects of the changes in the effective temperature and the bolometric correction , to a large degree , cancel each other when we consider the radius of hde 226868 and the lower limits to the masses of both components .
they introduce , however , the dramatic ( by a factor of @xmath37 ) increase of the luminosity of hde 226868 ( for a given distance ) . as a result ,
hde 226868 appears to be a very bright star : its optical luminosity is @xmath3 1 @xmath2 2 @xmath38 10@xmath39 erg / sec ( by two orders of magnitude higher than the typical x - ray luminosity of cyg x-1 ) .
undoubtedly , we know now the calibration of effective temperatures for o - type supergiants much better than 30 years ago .
however , still , there remains some uncertainty , clearly demonstrated by the recent major revision introduced by accounting for the line blanketing effects ( as discussed earlier ) . to investigate the effect of this uncertainty
, i assumed that the effective temperature of hde 226868 lies in the range 28000 @xmath2 32000 k and calculated the corresponding lower limits to the masses of both components ( using the modified versions of eqs .
( 1)@xmath4(2 ) ) .
the results of these calculations are given in the second and the third part of table 1 .
all sets of results are illustrated in fig .
1 . comparing the both sections of the table ( or looking at the fig .
1 ) , we may notice that the uncertainty of the effective temperature does not affect significantly the lower limits to the masses . summarizing the considerations of this section
, we may state that the masses of the compact and the optical components must be greater than @xmath40 m@xmath1 and @xmath41 m@xmath1 , respectively ( if the distance to the system is in the range 2.15 @xmath0 0.07 kpc ) .
let us remind that these values are model independent and therefore very hard to contest . if one wants to be more conservative and uses the @xmath42 error as the uncertainty of the distance estimate , then these lower limits become @xmath43 m@xmath1 and @xmath44 m@xmath1 , respectively . even adopting the smallest value ever claimed for the distance ( @xmath15 = 1.8 kpc )
, one still obtains the limits @xmath45 m@xmath1 and @xmath46 m@xmath1 , respectively .
one should add that our method gives only approximate estimates , since ( as noted by gies 2004 , private communication ) it assumes isotropic optical emission , which is not the case for tidally distorted star .
the resulting inaccuracies should not be , however , significant .
the lower limits , presented in table 1 and fig .
1 , become the actual values if the fill - out factor is equal 1.0 and the inclination of the orbit corresponds to the grazing eclipse orientation .
the fill - out factor can not be much different from 1.0 ( gies & bolton 1986a , b ) . however , the inclination is the source of larger uncertainty .
if the inclination is close to @xmath47 ( as suggested by the polarimetry , dolan & tapia 1989 ) , then we are close to the grazing eclipse situation and the masses are close to the lower limits obtained in this section , i.e. @xmath48 m@xmath1 and @xmath49 m@xmath1 , respectively . if ( as it seems more likely @xmath4 see the discussion in section 5 ) , the inclination is rather close to @xmath50 ( advocated by gies & bolton 1986a and still not excluded by polarimetry ) , then the masses are substantially higher and probably close to the values suggested by gies & bolton ( @xmath51 m@xmath1 and @xmath52 m@xmath1 , respectively ) .
i shall return to this point in further discussion .
quite independent constraints on the masses may be obtained from analysis of the evolutionary status of the optical component . first , let us note that we know relatively well its luminosity
. at the distance of 2.15 @xmath0 0.07 kpc , the luminosity must be @xmath53 ( equivalent to log(@xmath54/l@xmath55 ) = 5.62 @xmath0 0.03 in table 1 ) . including the uncertainty of the effective temperature and assuming @xmath42 error in the distance estimate
, the range for the luminosity becomes @xmath56 ( equivalent to log(@xmath54/l@xmath55 ) = 5.40 @xmath2 5.75 in table 1 ) .
the real range is somewhat wider because one should include also the errors in @xmath57 ( about @xmath0 0.05 ) and @xmath58 ( about @xmath0 0.1 ) estimates .
the final range becomes therefore @xmath59 .
one may note that this range corresponds to @xmath60 , which agrees quite well with @xmath61 obtained by ninkov et al .
( 1987 ) from the equivalent width of the h@xmath9 line .
second , let us consider the evolutionary phase of hde 226868 . even without the specific evolutionary calculations , it is relatively straightforward to argue that it must be a core hydrogen burning configuration .
it comes from the fact that hde 226868 must be in a relatively stable phase of its evolution .
we observe no measurable variations of the orbital period of the system .
the @xmath62 upper limit for the relative change of the orbital period @xmath63d ln @xmath64/d @xmath65 is @xmath66 yr@xmath67 ( gies & bolton 1982 ) .
this implies that the present evolutionary timescale of hde 226868 must be substantially longer than @xmath68 yr .
this excludes the possibility of the post - main sequence evolution , as the expected evolutionary time scale is , in that case , of the order of @xmath69 yr .
zikowski ( 1977 ) , investigating the post - main sequence evolution of a @xmath70 m@xmath1 star ( of similar radius but smaller luminosity than hde 226868 ) approaching its roche lobe , found that its radius was increasing from the fill - out factor equal 0.95 to 1.00 in only @xmath3 600 yr . after filling the roche lobe , it took only @xmath3 200 yr for the mass transfer rate to exceed the eddington limit by 3 orders of magnitude . for hde 226868 , which is more luminous and ,
probably , more massive ( @xmath71 m@xmath1 , as we shall see later ) , the relevant time scales should be even shorter .
if hde 226868 were expanding that fast ( 5% radius increase in less than 600 yr ) , we should observe the noticeable rise of its stellar wind strength and of the x - ray luminosity of the system ( both parameters are sensitive to the fill - out factor if it is close to 1 ) .
moreover , the chance of observing the system during such a short time window ( just few hundred years , before it gets extinguished as an x - ray source due to hyper - eddington accretion ) is very small .
therefore , we have to assume that hde 226868 ( similarly as the optical components of other massive x - ray binaries - see zikowski 1977 ) must still burn hydrogen in its core .
exactly the same conclusion may be obtained directly as a unique outcome of the numerical modelling of the evolution of hde 226868 .
this topic will be discussed in the next section .
i computed evolutionary tracks for core hydrogen burning phase of stars with the initial masses in the range @xmath72 m@xmath1 .
the warsaw evolutionary code developed by bohdan paczyski and maciek kozowski and kept updated by ryszard sienkiewicz was used .
the initial chemical composition @xmath73=0.7 and @xmath74=0.03 is adopted .
the opacity tables incorporating opal opacities ( iglesias & rogers 1996 ) as well as molecular and grain opacities ( alexander & ferguson 1994 ) were used .
the nuclear reaction rates are those of bahcall & pinsonneault ( 1995 ) .
the equation of state used was livermore laboratory opal ( rogers , svenson & iglesias 1996 ) .
i neglected the semiconvective mixing , as it is not important during the evolutionary phase considered ( most of the models of interest had central hydrogen content @xmath75 ) .
similarly any overshooting at the border of the convective core was neglected ( it is even less important ) .
the calculations were carried out under the assumption that the evolution starts from the homogeneous configurations .
it means that the consequences of the fact that some of the matter of the star , possibly dumped from the progenitor of the present black hole , could have somewhat modified chemical composition , were neglected .
it means also , that the consequences of the fact that some nuclear evolution ( hydrogen burning ) could , possibly , take place while the mass of the star was smaller ( prior to the mass transfer ) were neglected as well .
it seems that both simplifications do not alter significantly the outcome of the evolutionary calculations .
i shall return to this point in later discussion .
the most uncertain element of the calculations of the early evolution of massive stars is the mass loss due to stellar wind .
the uncertainty of the estimate of its rate is the single most important factor influencing the outcome of the calculations .
the observations seem to indicate that there is a substantial scatter of the mass loss rates ( typically , by a factor of two , but sometimes this factor can reach up to five ) among the stars of similar luminosities and effective temperatures ( see e.g. de jager , nieuwenhuijzen & van der hucht 1988 ) .
i decided to use the prescription given by nieuwenhuijzen & de jager ( 1990 ) in the form of the formula derived by hurley , pols & tout ( 1999 , hereafter hpt ) .
bearing in mind that the formula gives the mass loss rate estimate with the accuracy that is probably not better than within a factor of two , i introduced the multiplying factor @xmath76 applied to hpt formula .
for each initial mass of the star three evolutionary sequences were calculated with the value of the parameter @xmath76 equal , in sequence , 0.5 , 1 and 2 . in this way ,
the uncertainty of the theory of evolution could be , hopefully , taken into account .
it might be interesting to note , at this point , the astonishingly good agreement between the hpt formula and the observational determination of the mass loss rate for hde 226868 : for the most likely values of the parameters of the star , the hpt value agrees with the observed one ( @xmath77 m@xmath1/yr , gies et al .
2003 , see below ) to an accuracy of about 5% ( this is , definitely , better than the precision of both the hpt formula and of the observational estimate ) .
i should remind at this point that the supergiant components in some high mass x - ray binaries are significantly undermassive for their luminosities ( zikowski 1977 ) . in some systems , like cen x-3 ,
this undermassivness is very serious and requires much stronger mass loss than the normal stellar wind ( zikowski 1978 ) .
however , most likely , it is not the case for hde 226868 .
its parameters may be fully explained by the evolution with the normal stellar wind ( as will be demonstrated in the further discussion ) .
some of the obtained evolutionary tracks in the h - r diagram are shown in fig .
2 . the careful reader might notice that the luminosities of my models are by about 0.1 dex ( @xmath3 25% ) smaller than the evolutionary tracks of schalller al .
this difference should be attributed mainly to the fact that i used later edition of the opacity tables containing higher values of the opacities ( the opacity tables always evolve in the direction of the growing opacity never the other way ) .
the correctness of the new opacities was confirmed by stellar pulsations calculations
see e.g. pamyatnykh ( 1999 ) .
part of the luminosity difference is due to the fact , that i used higher metallicity ( @xmath74 = 0.03 instead of @xmath74= 0.02 ) . and
, finally , part of the luminosity difference ( all factors work in the same direction ) is due to the fact , that my stellar winds are stronger ( again due to higher metallicity ) . as may be seen from a quick look at fig .
2 , the initial ( zero age main sequence ) mass of hde 226868 had to be in the range 35 to 55 m@xmath1 .
the masses of the models corresponding to the present day state of hde 226868 ( inside the @xmath42 parallelogram ) are in the range 32.4 to 50.5 m@xmath1 .
the central hydrogen content in these models is between 0.126 and 0.265 and their evolutionary age ( since the beginning of the central hydrogen burning ) is between 2.7 and 4.3 @xmath78 yr . [ f2 ] [ f3 ] to obtain evolutionary models which satisfactorily reproduce the present day state of hde 226868 , one has to match not only the luminosity and the effective temperature but also the rate of the stellar wind mass loss .
the first observational determination of this parameter for hde 226868 was done by hutchings ( 1976 ) who , analyzing the visual spectrographic data , got the value @xmath79 m@xmath1/yr .
persi et al .
( 1980 ) estimated the rate of the mass outflow from the infrared emission of the expanding circumstellar envelope and got the value @xmath80 m@xmath1/yr .
more recent estimate of herrero et al .
( 1995 ) was based on the fits of the h@xmath81 profile and led to the conclusion that the mass loss rate lies between 2 and 6 @xmath82 m@xmath1/yr .
they noted that the inaccuracy of the fits is , probably , due to the fact that the stellar wind from hde 226868 is focused towards the black hole ( for the discussion of the `` focused '' wind model see gies & bolton 1986 and miller et al .
2002 ) . in their test modelling of the atmosphere of hde 226868 , herrero et al . were using the value @xmath83 m@xmath1/yr and for their final model they chose @xmath84 m@xmath1/yr .
the most recent estimate by gies et al .
( 2003 ) gave the value @xmath85 m@xmath1/yr for the low / hard state ( which is a typical state of cyg x-1 ) and @xmath86 m@xmath1/yr for the high / soft state ( which is less frequent in this source ) .
taking all this into account , i assumed that the observed rate of mass outflow from hde 226868 is @xmath85 m@xmath1/yr with an error by a factor of about 2 , i.e. i assumed that the observed rate of mass outflow is between 1.3 and 5.2 @xmath87 m@xmath1/yr .
let me remind that our theoretical evolutionary models were using mass loss rates in the range 0.5 to 2 times the rate given by hpt .
altogether , it means , that the model for which the mass loss rate calculated with the hpt formula would be by a factor of up to four smaller or by a factor of up to four larger than the nominal observational value ( @xmath88 m@xmath1/yr ) is still considered to be an acceptable match .
this is , probably , more than sufficient allowance for the uncertainty of the mass loss rates .
the parameters of some of these models are given in table 2 .
the positions of the acceptable models in the mass@xmath4luminosity diagram ( for the most likely value of the effective temperature @xmath24 = 30700 k ) are shown in the fig .
since the luminosity of a model of hde 226868 can be ( for an assumed effective temperature ) directly translated into the distance to the system ( see table 1 ) , the fig .
3 can be also considered as the mass@xmath4distance diagram .
the corresponding distance scale is shown on this picture . the line corresponding to the lower mass limit@xmath4distance relation obtained in section 2 ( see table 1 and fig .
1 ) is also shown .
one may notice that the deduced mass of hde 226868 depends mainly on the assumed distance to the binary system , and only very weakly on the assumptions about the mass loss rates .
for the assumed effective temperature ( @xmath24 = 30700 k ) this mass has to be in the range 35 @xmath2 45 m@xmath1 ( if the distance to the system is in the range 1.8 to 2.35 kpc ) .
one may also conclude ( with the help of the lower mass limits derived in section 2 ) that the distance to the system can not be larger than @xmath3 2.5 kpc ( still for the assumed effective temperature ) .
similar diagrams as fig .
3 may be constructed for other possible values of the effective temperature of hde 226868 .
qualitatively , they look similar to the fig .
3 . for the effective temperature
@xmath24 = 28000 k the mass of the star has to be in the range 29 @xmath2 35 m@xmath1 and the upper limit for the distance to the system is @xmath3 2.1 kpc . for the effective temperature @xmath25 = 32000 k
the mass of the star has to be in the range 37 @xmath2 50 m@xmath1 and the upper limit for the distance to the system is @xmath3 2.7 kpc .
the obvious and fairly strong conclusion derived from the evolutionary calculations is that hde 2226868 had to be quite massive in the past and is still very massive at present . to put it very briefly
, it has to be very massive because it is very bright ( while still burning hydrogen in the core ) .
for the most likely range of parameters ( 30000 to 31000 k for the effective temperature and 2.08 to 2.22 kpc for the distance ) the present mass of hde 2268668 has to be in the range 37 to 44 m@xmath1 . increasing the interval of the possible values of the distance to 1.95 to 2.35
kpc increases this range only slightly ( to 33.5 @xmath89 ) . extending the interval of the possible values of the effective temperatures to 28000 to 32000
k leads to the range of possible masses 31 @xmath90 . finally , extending the range of the possible distances down to 1.8 kpc ( the smallest value quoted in the literature ) results in the final range of the possible masses of hde 226868 : 29 @xmath90 .
the value of 29 m@xmath1 seems to be a firm lower limit for the present mass of hde 226868
. one may ask whether it is possible to construct substantially less massive evolutionary models by taking more liberal estimate of the observational uncertainties or accounting for some simplifications of our models .
the answer is negative .
decreasing the effective temperature of hde 226868 would result in lower luminosities and so would lead to less massive configurations .
however , the value @xmath24 = 28000 k , taken as the lower limit to the effective temperature in the above consideration is probably already too low ( the true value is probably in the range 30000 to 31000 k ) . decreasing the minimum distance from 1.95 to 1.8
kpc decreased the minimum possible masses of the models by only @xmath91 , as was mentioned above .
let us now consider the possible consequences of some simplifications of our evolutionary calculations .
i neglected the possible dumping of matter on hde 226868 from the progenitor of cyg x-1 ( and so `` rejuvenation '' of the star ) during the first mass loss / exchange phase .
most likely , there was no significant accretion on hde 226868 because any substantial mass transfer in such a close binary and in so early stage of evolution ( when both stars had no distinct cores ) would lead to the rapid built - up of a common envelope and the coalescence of both components .
so , most likely , the first mass transfer was mainly the mass loss from the system .
if there were no serious accretion , then the nuclear evolution prior to the mass transfer is properly accounted for in our calculations .
if ( which is rather unlikely ) there was a substantial mass dumping on hde 226868 , then the nuclear evolution before the mass transfer is negligible because the substantially less massive star was evolving much more slowly ( large mass gain essentially resets the evolutionary clock to zams ) .
one may wonder whether smaller metallicity ( @xmath74 = 0.02 instead of @xmath74= 0.03 ) would not be a better choice for hde 226868 . to check the possible effects of such alternative choice
, i calculated several evolutionary tracks for stars with @xmath74 = 0.02 .
as might be expected , the changes in the results were small .
for the models reproducing the `` best '' parameters of hde 226868 , the masses decreased from about 41.7 m@xmath1 to about 39.9 m@xmath1 ( only about 4% ) .
there remains a rather remote possibility that the chemical composition of hde 226868 is not normal , i.e. that it contains less hydrogen than a normal population i star .
e.g. herrero et al .
( 1995 ) concluded , from fitting models of the atmosphere , that helium might be overabundant by a factor of about two .
the evolutionary calculations presented here indicate that the stellar wind mass loss from hde 226868 was not severe enough for hydrogen depleted layers to show up on the surface .
however , if a large amount of matter with significantly decreased hydrogen content was dumped on hde 226868 from the progenitor of cyg x-1 , then the star might contain less hydrogen than the assumed value of 0.7 .
the lack of the cno anomalies ( dearborn 1977 ) testifies against this being the case .
however , to check the consequences of such ( rather unlikely ) situation , i calculated several evolutionary sequences for the stars with the initial chemical composition @xmath73=0.5 , @xmath74=0.03 ( this corresponds to an overabundance of helium by a factor of two ) .
as could be easily expected , the models with similar luminosities were now less massive , but the change was not very substantial . for the `` best fit '' models , the masses decreased from about 41.7 m@xmath1 to about 35.3 m@xmath1 ( only about 15% ) .
however , i would like to stress again , that there are no good reasons to expect such abnormal hydrogen content in hde 226868 .
the only way to produce an evolutionary model with the mass @xmath92 m@xmath1 is to decrease drastically the distance to the system .
i constructed one evolutionary track that produced a 19.5 m@xmath1 configuration at the effective temperature 30700 k with the mass loss rate @xmath93 ( the initial mass of the star was 32.5 m@xmath1 ) .
however , the price was very high : the distance had to be decreased to @xmath3 0.9 kpc and the rate of the mass loss during the evolution had to be multiplied by a factor of @xmath3 20 with respect to the hpt formula .
any of these conditions , even taken separately , seems extremely unlikely .
there is a clear conflict between the above considerations and the mass estimate given by herrero et al .
they used atmosphere models to reproduce the observed spectrum of hde 226868 .
then , they used their fit to determine the effective temperature , the surface gravity , the radius and the helium abundance of the star . their mass estimate ( @xmath94 m@xmath1 ) is a consequence of the surface gravity and the radius determination .
i see no solution of this conflict .
i may only repeat that the evolutionary calculations for such an early evolutionary phase are very robust and rather difficult to dispute ( it is , essentially , the mass - luminosity relation for the main sequence stars ) .
i may also add that the evolutionary results presented above are roughly consistent with the model of gies & bolton ( 1986a ) , based on an extensive analysis of the large and diversified collection of the observational data .
in the previous section we concluded that the present mass of hde 226868 has to be in the range of 29 @xmath95 50 m@xmath1 ( assuming the widest possible ranges of the distances and the effective tempeatures ) .
the mass of its companion , black hole cyg x-1 , does not come out directly from the evolutionary calculations .
however , it can be calculated ( with the help of some observational constraints ) , once we select a chosen evolutionary model of hde 226868 .
once the selection is made , we know the mass @xmath96 and the radius @xmath97 of the optical component . we can also calculate the distance , with the help of eq .
( 1 ) or ( 2 ) ( or the corresponding expressions for other effective temperatures ) .
subsequently , we can use two equations to solve for the inclination of the orbit @xmath98 and the mass ratio @xmath99 .
one of these equations makes use of the mass function , @xmath100.\end{aligned}\ ] ] the other relates the radius of the star to the size of the orbit , @xmath101 where @xmath102 is the radius of the roche lobe around hde 226868 , @xmath103 is the fill - out factor ( @xmath104 ) , @xmath105 is the orbital separation of the binary components and @xmath22 is the radius of the orbit of hde 226868 ( see section 2 ) . inserting the observational data , eqs .
( 3)@xmath4(4 ) can be written as : @xmath106 = 0.251,\end{aligned}\ ] ] @xmath107 once a given evolutionary model of hde 226868 is selected from the grid of the acceptable models and a value of the parameter @xmath108 is assumed , the eqs .
( 5)@xmath4(6 ) can be solved for @xmath98 and @xmath109 .
knowing @xmath109 we can immediately calculate also the mass of compact component @xmath110 . in principle , this procedure can be applied to any combination of the evolutionary model and of the value of @xmath103 .
in fact , however , not every evolutionary model of hde 226868 ( acceptable if we consider the optical component alone ) permits the construction of a consistent model of the binary system .
this is because of the observational constraints on the value of the inclination @xmath98 and , especially , because of the strong observational constraints on the value of the fill - out factor @xmath103 .
as demonstrated by gies & bolton ( 1986a , b ) , in order to explain quantitatively the he emission lines produced in the stellar wind from hde 226868 , the fill - out factor @xmath108 has to be larger than 0.9 and , most likely , not smaller than 0.95 ( perhaps the best value would be around 0.98 ) . on the other hand , gies & bolton demonstrated that observed rotational broadening of the photospheric absorption lines of hde 226868 and the observed amplitude of the ellipsoidal light variations impose substantial constraints on the values of the mass ratio , the fill - out factor and the inclination .
the rotational broadening depends mainly on the mass ratio and the measured broadening indicates the mass ratio in the range @xmath111 .
the amplitude of the ellipsoidal light variations is determined mainly by the fill - out factor and the inclination .
for the assumed values of @xmath103 equal 0.9 , 0.95 and 1 , the resulting inclination is @xmath112 , @xmath113 and @xmath114 , respectively .
constructing models of the binary system to be consistent with observations , i assumed that for any adopted value of @xmath108 , the calculated inclination should be within @xmath115 from the corresponding values quoted above .
i started with different models of the optical component , acceptable from the point of view of the stellar evolution , as described in section 4.3 .
then , i assumed the value of @xmath103 equal 0.95 and solved eqs .
( 5)@xmath4(6 ) to find @xmath109 , @xmath98 and @xmath110 .
subsequently , i tried higher values of @xmath103 .
the higher values of @xmath103 produced the solutions with lower ( in many cases unacceptably low ) values of the inclination .
the dependence of @xmath109 , @xmath98 and @xmath110 on the assumed value of @xmath103 may be seen from many sequences of binary models , presented in tab .
2 . for all sequences included in the table ,
i present only two limiting models for the lowest and the highest value of @xmath103 for which an acceptable model could still be obtained .
if there is only one entry for a given sequence ( as is the case for @xmath116 m@xmath1 and @xmath117 m@xmath1 ) , it means that only one value of @xmath103 produced an acceptable solution ( the value of @xmath103 was varied with the increment of 0.01 ) .
i classified , as acceptable , the models satisfying the following criteria : \(1 ) @xmath118 kpc \(2 ) @xmath26 k \(3 ) @xmath119 m@xmath1/yr
\(4 ) @xmath120 \(5 ) @xmath121 ( this corresponds to @xmath122 for @xmath123 and @xmath124 for @xmath125 , as advocated by gies & bolton 1986a )
. parameters of selected acceptable models are given in the second part of table 2 .
for each assumed value of the effective temperature , i present two sequences of models corresponding to the lowest and the highest current mass of the optical component .
the exception to this rule is the sequence of models for @xmath126 k and @xmath127 m@xmath1 ( @xmath15 = 2.15 kpc ) .
the initial evolutionary parameters ( @xmath128 and @xmath76 ) of the optical component model for this sequence were adjusted so as to obtain the perfect fit with the `` best '' values of the observational parameters of hde 226868 ( effective temperature , luminosity and the rate of stellar wind mass loss ) .
all models from this sequence ( for @xmath129 ) are successful from the point of view of our criteria .
there are several conclusions that can be drawn from the collection of the obtained models of the binary system ( only some of these models are shown in table 2 ) . the first concerns the mass of the compact component . assuming the widest possible intervals of the distances and the effective temperatures : 1.8 to 2.35 kpc and 28000 to 32000 k
, the mass of the compact component must be in the range 13.5 @xmath2 28.5 m@xmath1 .
for the most likely intervals of these parameters : 1.95 to 2.35 kpc and 30000 to 31000 k , this range narrows to 15.5 @xmath2 25 m@xmath1 .
the second conclusion is related to the distance to the binary system .
it appears that for a given ( assumed ) effective temperature , consistent models are possible only for some , relatively narrow , interval of distances ( the distances out of this interval would require unacceptably low or unacceptably high values of the inclination ) . for the effective temperatures equal to 28 , 30 , 30.7 , 31 and 32 @xmath130k , the corresponding intervals are [email protected] , [email protected] , [email protected] , [email protected] and [email protected] kpc , respectively .
it is certainly encouraging that the most likely range of the effective temperatures ( 30000 to 31000 k ) requires the distance interval ( [email protected] kpc ) that almost exactly coincides with the independent estimate of the most likely distance range ( [email protected] kpc ) .
the obtained relation between the distance and the effective temperature of hde 226868 means that , if in the future the distance will be known more precisely ( e.g. from the future astrometric space missions ) , then it will be possible to set constraints on the effective temperature .
for example , if it is found that the distance to the system is greater than 2.1 kpc , it would mean that the effective temperature of hde 226868 has to be higher than 30000 k. also the opposite is true . if it is found ( e.g. from the better models of the atmospheres ) that the effective temperature of hde 22868 is greater than 30000 k
, it would mean that the distance to the system must be larger than 2.0 kpc ( and this conclusion would not depend on any photometric or astrometric considerations ) .
finally , the third conclusion confirms the earlier results ( based on the evolutionary calculations alone ) limiting the present mass of hde 226868 to the range @xmath131 m@xmath1 and the initial mass to the range @xmath132 m@xmath1 .
let us note that for the black hole mass in the range 15 @xmath2 25 m@xmath1 , the state transitions occurs in cyg x-1 at the luminosity level equal 0.025 to 0.015 of the eddington luminosity ( assuming @xmath15 = 2.15 kpc and using the flux values given by zdziarski et al .
these values lie in the range 0.007 to 0.03 found for other x - ray binaries ( maccarone 2003 ) . to summarize this section ,
the evolutionary considerations provide much stronger lower limits to the masses of both components than the model independent analysis , presented in section 2 .
in addition , they yield some independent constraints on the distance to the system .
let us speculate a little bit about the possible evolutionary past of our binary system .
it had to start as a very massive system ( the initial primary had to complete its evolution by the time the secondary ( hde 226868 ) reached its present evolutionary state i.e. in less than @xmath133 yr ) .
the initial masses of the components were probably @xmath134 and @xmath135 m@xmath1 .
the massive primary ( the progenitor of a black hole ) was shedding the mass mainly in the form of the stellar wind ( there was probably little , if any , mass transfer to the companion ) .
the mass of the primary decreased , at the end of its evolution ( before the collapse ) , to @xmath136 m@xmath1 .
the low spatial velocity of cyg x-1 with respect to its parental association indicates , as argued by mirabel & rodriguez ( 2003 ) , that the final collapse to a black hole proceeded with very little ( @xmath137 m@xmath1 ) mass ejection .
it could be , even , a prompt collapse with no accompanying supernova explosion at all ( `` formation of a black hole in the dark '' ) .
the secondary ( the progenitor of hde 226868 ) was also losing mass in the form of the stellar wind , although at a lower rate .
this mass loss took place both before and after the collapse of its companion .
this process decreased the mass of the secondary to the present value of @xmath138 m@xmath1 . in this way
, the present day binary consisting of a massive supergiant of @xmath139 m@xmath1 ( with a central hydrogen content of @xmath140 ) and a black hole of @xmath141 m@xmath1 was formed .
the above evolutionary history is only only one of the possible scenarios .
as noted by gies ( 2004 , private communication ) , the above scenario implies that about half of the total initial mass of the binary system was lost in the form of stellar wind .
this implies , in turn , that the initial orbital period was @xmath142 days , which means that the system was very tight .
one can not exclude that the initial masses were smaller , the initial orbital period longer and the system passed through the mass transfer and the common envelope phase .
the calculations modelling the evolution of hde 226868 , under different assumptions about the stellar wind mass loss rate , provide robust limits on the present mass of the star . for the most likely intervals of the values of the distance and of the effective temperature : 1.95 to 2.35 kpc and 30000 to 31000 k ,
the mass of hde22868 is 40 @xmath143 m@xmath1 . extending the intervals of these parameters to 1.8 to 2.35 kpc and 28000 to 32000 k , one obtains the mass of the star in the range @xmath131 m@xmath1 . including the additional constraints resulting from the observed properties of the binary system hde 226868/cyg x-1 , one can estimate the mass of the black hole component . for the most likely values of the parameters mentioned in the item ( 1 ) above this mass
is 20 @xmath143 m@xmath1 .
for the extended intervals of the parameters the mass is in the range of @xmath144 m@xmath1 .
the distance to the binary system has to be in the ranges [email protected] , [email protected] , [email protected] , [email protected] and [email protected] kpc , for the effective temperature of hde 226868 equal to 28 , 30 , 30.7 , 31 and 32 @xmath145 k , correspondingly .
i would like to thank a. zdziarski for careful reading of the manuscript and for many helpful comments and stimulating discussions .
i would like also to thank d. gies who , acting as a referee , made several comments and suggestions which helped to improve this paper .
this work was partially supported by the state committee for scientific research grants no 1 p03d 018 27 and no pbz kbn 054/p03/2001 .
alexander d.r .
, ferguson j.w . , 1994 , apj , 437 , 789 bahcall j.n . ,
pinsonneault m.h . , 1995 , rev .
phys . , 67 , 781 blaauw a. , 1991 , in lada c.j . , kylafis n.d . ,
eds , nato asic proc .
342 , the physics of stars formation and early stellar evolution , kluwer , dordrecht , netherlands , p. 125
blaha c. , humphreys r.m .
, 1989 , aj , 98 , 1598 bolton c.t
. , 1972 , nature , 235 , 271 bregman j. , butler d. , kemper e. , koski a. , kraft r.p . , stone r.p.s .
, 1973 , apj , 185 , l117 brocksopp c. , tarasov a.e . , lyuty v.m . , roche p. , 1999
, a&a , 343 , 861 crawford d.l . ,
barnes j.v . , warren , w.h .
, 1974 , aj , 79 , 623 dambis a.k . , melnik a.m. , rastorguev , a.s . , 2001 ,
lett . , 27 , 58 dearborn d.s.p . , 1977 , astrophysical lett . , 19
, 15 de jager c. , nieuwenhuijzen h. , van der hucht , k.a . , 1988 , a&as , 72 , 259 dolan j.f . , tapia , s. , 1989 , apj , 344 , 830 garmany c.d . ,
stencel , r.e . , 1992 ,
a&a , 94 , 211 gies d.r . , bolton , c.t .
, 1982 , apj , 260 , 240 gies d.r . , bolton , c.t .
, 1986a , apj , 304 , 371 gies d.r . , bolton , c.t .
, 1986b , apj , 304 , 389 gies d.r .
, bolton c.t .
, thomson j.r . , huang w. , mcswain m.v . ,
riddle r.l . , wang z. , wiita p.j . ,
wingert d.w . , csak b. , kiss l.l . , 2003 ,
apj , 583 , 424 herrero a. , kudritzki r.p . , gabler r. , vilchez j.m . , gabler a. , 1995 , a&a , 297 , 556 humphreys r.m . , 1978 ,
apjs , 38 , 309 hurley j.r . , pols o.r . ,
tout , c.a .
, 2000 , mnras , 315 , 543 ( hpt ) hutchings j.b . , 1976 , apj , 203 , 438 iglesias c.a . ,
rogers , f.j . , 1996 ,
apj , 464 , 943 janes k. , adler , d. , 1982 , apjs 49 , 425 lestrade j .- f . ,
preston r.a .
, jones d.j . ,
philips r.b . , rogers a.e.e . ,
titus m.a . , rioja m.j . ,
gabuzda d.c . , 1999 ,
a&a , 344 , 1014 maccarone t.j . , 2003 , a&a , 409 , 697 malysheva l.k . , 1997 ,
, 23 , 585 margon b. , bowyer s. , stone r.p.s .
, 1973 , apj , 185 , l113 markova n. , puls j. , repolust t. , markov h. , 2004 , a&a , 413 , 693 martins f. , schaerer d. , hillier d.j .
, 2002 , a&a , 382 , 999 massey p. , johnson k.e .
, degioia - eastwood k. , 1995 , apj , 454 , 151 miller j.m . , wojdowski p. , schultz n.s .
, marshall h.l . , fabian a.c . ,
remillard r.a . , wijnands r. , lewin w.h.g . , 2002 , astro - ph 0208463 mirabel i.f . , rodriguez i. , 2003 , science , 300 , 1119 nieuwenhuijzen h. , de jager c. , 1990 , a&a , 231 , 134 ninkov z. , walker g.a.h .
, yang s. , 1987 , apj , 321 , 425 paczyski b. , 1974 , a&a , 34 , 161 pamyatnykh a.a . , 1999 ,
acta astr .
, 49,119 persi p. , ferrari - toniolo m. , grasdalen g.l . , spada g. , 1980 , a&a , 92 , 238 repolust t. , puls j. , herrero a. , 2004 , a&a , 415 , 349 rogers f.j .
, svenson f.j . ,
iglesias c.a . , 1996 ,
apj , 456 , 902 schaller g. , schaerer d. , meynet g. , maeder a. , 1992 , a&a suppl . , 96 , 269 schmidt - kaler th . , 1982 , in schaifers k. , vogt h.h .
, eds , landolt - bornstein , numerical data and functional relationships in science and technology , new series , springer - verlag , berlin , group vi , vol .
vacca w.d . , garmany c.d . ,
shull j.m .
, 1996 , apj , 460 , 914 ( vgs ) walborn n.r . , 1973 , apj , 179 , l123 webster b.l . , murdin p. , 1972 ,
nature , 235 , 37 wu c .- c .
, eaton j.a . ,
holm a.v . , milgrom m. , hammerschlag - hensberge g. , 1982 , pasp , 94 , 149 zdziarski a.a . ,
poutanen j. , paciesas w.s . , wen l. , 2002 , apj , 578 , 357 zikowski j. , 1977 , eight texas symposium on relativistic astrophysics , annals new york academy sci .
, 302 , 47 zikowski j. , 1978 , in @xmath146ytkow a.n . , ed . , nonstationary evolution of close binaries , polish scientific publishers , warsaw , p. 29
|
calculations carried out to model the evolution of hde 226868 , under different assumptions about the stellar wind mass loss rate , provide robust limits on the present mass of the star .
it has to be in the range 40 @xmath0 5 m@xmath1 if the distance to the system is in the range 1.95 to 2.35 kpc and the effective temperature of hde 226868 in the range 30000 to 31000 k. extending the possible intervals of these parameters to 1.8 to 2.35 kpc and 28000 to 32000 k , one gets for the mass of the star the range 40 @xmath0 10 m@xmath1 . including into the analysis observational properties such as the profiles of the emission lines , rotational broadening of the absorption lines and the ellipsoidal light variations
, one can estimate also the mass of the compact component .
it has to be in the ranges 20 @xmath0 5 m@xmath1 and 13.5 @xmath2 29 m@xmath1 for the cases described above .
the same analysis ( using the evolutionary models and the observational properties listed above ) yields lower limit to the distance to the system of @xmath3 2.0 kpc , if the effective temperature of hde 226868 is higher than 30000 k. this limit to the distance does not depend on any photometric or astrometric considerations .
[ firstpage ] binaries : general x - ray sources evolution stars : individual ( cyg x-1 ) masses of the components .
|
we shall discuss methods of calculation of propagator diagrams in various theories and kinematics , up to 3 loops . in order to obtain renormalization - group functions ,
it is often enough to calculate propagator diagrams .
massless diagrams are discussed in sect .
[ s : q ] ; they are useful , e.g. , in qcd with light quarks . heavy quark effective theory ( hqet )
is discussed in sect .
[ s : hqet ] , and massive on - shell diagrams in sect .
[ s : m ] .
finally , some mathematical methods , most useful in massless and hqet calculations , are considered in sect .
[ s : hyper ] . throughout these lectures , we shall discuss scalar feynman integrals .
tensor integrals can be expanded in tensor structures , scalar coefficients are found by solving linear systems . their solution can be written as projectors ( tensor , @xmath3-matrix ) applied to the original diagram .
scalar products of momenta in the numerator can be expressed via the denominators . in some classes of problems ,
there are not enough independent denominators to express all scalar products .
then we add the required number of scalar products to the basis , and consider scalar integrals with powers of these scalar products in the numerators .
some diagrams contain insertions into internal lines .
the denominators of the propagators on both sides coincide .
if powers of these denominators are @xmath4 and @xmath5 , we can combine them into a single line with the power @xmath6 ( fig .
[ f : insert ] ) .
( 94,24 ) ( 21,12)(0,0 ) ( 73,12)(0,0 ) ( 47,12)(0,0)@xmath7 ( 6,16)(0,0)[r]@xmath4 ( 36,16)(0,0)[l]@xmath5 ( 86,19)(0,0)[l]@xmath6 throughout these lectures , minkowski notations with the metric @xmath8 are used .
the 1-loop massless propagator diagram ( fig .
[ f : q1 ] ) is @xmath9 it is symmetric with respect to @xmath10 .
( 64,42 ) ( 32,21)(0,0 ) ( 32,0)(0,0)[b]@xmath11 ( 32,42)(0,0)[t]@xmath12 ( 32,8)(0,0)[b]@xmath4 ( 32,34)(0,0)[t]@xmath5 if @xmath4 is integer and @xmath13 , it becomes the vacuum massless diagram ( fig .
[ f : q10 ] ) with the polynomial @xmath14 in the numerator .
it reduces to a sum of powers of @xmath15 multiplied by integrals @xmath16 this integral has the mass dimension @xmath17 ; it contains no dimensionful parameters , and the only possible result for it is @xmath18 ; more careful investigation @xcite shows that the right - hand side contains @xmath19 . ] .
therefore , @xmath20 for non - positive integer @xmath4 ( or @xmath5 ) .
more generally , any massless vacuum diagram with any number of loops is scale - free and hence vanishes .
( 14,11 ) ( 7,5.5)(0,0 ) this diagram is calculated by the fourier transform to @xmath21-space and back : @xmath22 our diagram in @xmath21-space is the product of two propagators ( [ q1:p2x ] ) with the powers @xmath4 and @xmath5 : @xmath23 transforming it back to @xmath24-space ( [ q1:x2p ] ) , we arrive at @xmath25 diagrams are calculated by going to euclidean momentum space : @xmath26 , @xmath27 . using a dimensionless loop momentum @xmath28 , we can rewrite ( [ q1:g ] ) as @xmath29^{n_1 } \left[k^2\right]^{n_2 } } = \pi^{d/2 } g(n_1,n_2)\,,\ ] ] where @xmath30 is a unit euclidean vector ( @xmath31 ) .
we can perform inversion @xmath32 then @xmath33 and we obtain ( fig . [ f : q1i ] ) @xmath34 ( 74,26 ) ( 16,13)(0,0 ) ( 58,13)(0,0 ) ( 37,13)(0,0)@xmath7 ( 16,0)(0,0)[b]@xmath4 ( 16,26)(0,0)[t]@xmath35 ( 58,0)(0,0)[b]@xmath4 ( 58,26)(0,0)[t]@xmath36 there is 1 generic topology of 2-loop massless propagator diagrams ( fig .
[ f : q2 ] ) : @xmath37 all other diagrams ( e.g. , fig . [
f : insert ] ) are particular cases of this one , when some line shrinks to a point , i.e. , its index @xmath38 .
this diagram is symmetric with respect to @xmath39 , and with respect to @xmath40 .
if indices of any two adjacent lines are non - positive , the diagram contains a scale - free vacuum subdiagram , and hence vanishes .
( 64,32 ) ( 32,18)(0,0 ) ( 14,32)(0,0)@xmath41 ( 50,32)(0,0)@xmath42 ( 12,3)(0,0)@xmath43 ( 52,3)(0,0)@xmath44 ( 42,18)(0,0)@xmath45 ( 22,12)(0,0)@xmath4 ( 42,12)(0,0)@xmath5 ( 22,24)(0,0)@xmath46 ( 42,24)(0,0)@xmath47 ( 27,18)(0,0)@xmath48 if @xmath49 , our diagram is just the product of two 1-loop diagrams ( fig .
[ f : q2z5 ] ) : @xmath50 if @xmath51 , the inner loop gives @xmath52 , and ( fig .
[ f : q2z1 ] ) @xmath53 the cases @xmath54 , @xmath55 , @xmath56 are symmetric .
( 52,26 ) ( 26,13)(0,0),title="fig : " ] ( 16,0)(0,0)[b]@xmath4 ( 36,0)(0,0)[b]@xmath5 ( 16,26)(0,0)[t]@xmath46 ( 36,26)(0,0)[t]@xmath47 ( 134,32 ) ( 25,17.5)(0,0),title="fig : " ] ( 76,17.5)(0,0),title="fig : " ] ( 118,17.5)(0,0),title="fig : " ] ( 55,17.5)(0,0)@xmath7 ( 97,17.5)(0,0)@xmath57 ( 25,0)(0,0)[b]@xmath5 ( 15,30)(0,0)@xmath46 ( 33,30)(0,0)@xmath47 ( 24,18.5)(0,0)@xmath48 ( 76,4.5)(0,0)[b]@xmath48 ( 76,30.5)(0,0)[t]@xmath58 ( 118,4.5)(0,0)[b]@xmath5 ( 118,30.5)(0,0)[t]@xmath59 when all @xmath60 , the problem does not immediately reduce to a repeated use of the 1-loop formula ( [ q1:g1 ] ) .
we shall use a powerful method called integration by parts @xcite .
it is based on the simple observation that any integral of @xmath61 ( or @xmath62 ) vanishes ( in dimensional regularization no surface terms can appear ) . from this
, we can obtain recurrence relations which involve @xmath63 with different sets of indices . applying these relations in a carefully chosen order
, we can reduce any @xmath63 to trivial ones , like ( [ q2:z5 ] ) , ( [ q2:z1 ] ) .
the differential operator @xmath64 applied to the integrand of ( [ q2:g ] ) acts as @xmath65 applying @xmath66 to the integrand of ( [ q2:g ] ) , we get a vanishing integral . on the other hand , from ( [ q2:g ] ) , @xmath67 , @xmath68 , @xmath69 ,
we see that this differential operator is equivalent to inserting @xmath70 under the integral sign ( here @xmath71 ) . taking into account the definition ( [ q2:g ] )
, we obtain the recurrence relation @xmath72 g = 0\ , .
\label{q2:tri1}\ ] ] here @xmath73 and similar ones for the other indices .
this is a particular example of the triangle relation .
we differentiate in the loop momentum running along the triangle 254 , and insert the momentum of the line 4 in the numerator .
the differentiation raises the degree of one of the denominators 2 , 5 , 4 . in the case of the line 4 ,
we get @xmath74 in the numerator , giving just @xmath75 . in the case of the line 5
, we get the denominator @xmath76 of the line attached to the vertex 45 of our triangle , minus the denominators @xmath77 and @xmath78 .
the case of the line 2 is similar ; the denominator of the line attached to the vertex 24 of our triangle is just @xmath15 , and it does not influence any index of @xmath79 .
of course , there are three more relations obtained from ( [ q2:tri1 ] ) by the symmetry .
another useful triangle relation is derived by applying the operator @xmath80 : @xmath81 g = 0\ , .
\label{q2:tri2}\ ] ] one more is obtained by the symmetry .
relations of this kind can be written for any diagram having a triangle in it , when at least two vertices of the triangle each have only a single line ( not belonging to the triangle ) attached .
we can obtain a relation from homogeneity of the integral ( [ q2:g ] ) in @xmath24 . applying the operator @xmath82 to the integral ( [ q2:g ] )
, we see that it is equivalent to the factor @xmath83 . on the other hand , explicit differentiation of the integrand
gives @xmath84 .
therefore , @xmath85 i = 0\ , .
\label{q2:hom}\ ] ] this is nothing but the sum of the @xmath66 relation ( [ q2:tri1 ] ) and its mirror - symmetric @xmath86 relation .
another interesting relation is obtained by inserting @xmath87 into the integrand of ( [ q2:g ] ) and taking derivative @xmath88 of the integral . on the one hand ,
the vector integral must be proportional to @xmath89 , and we can make the substitution @xmath90 in the integrand .
taking @xmath88 of this vector integral produces ( [ q2:g ] ) with @xmath91 inserted into the integrand . on the other hand , explicit differentiation in @xmath24 gives @xmath92 therefore , we obtain @xmath93 g = 0\ , .
\label{q2:larin}\ ] ] this relation has been derived by s.a .
larin in his m. sc .
three more relations follow from the symmetries . expressing @xmath63 from ( [ q2:tri2 ] ) : @xmath94 we see that the sum @xmath95 reduces by 1 .
if we start from an integral belonging to the plane @xmath96 in fig .
[ f : red ] , then each of the integrals in the right - hand side belong to the plane @xmath97 ( fig .
[ f : red ] ) .
therefore , applying ( [ q2:tri ] ) sufficiently many times , we can reduce an arbitrary @xmath79 integral with integer indices to a combination of integrals with @xmath49 ( fig .
[ f : q2z5 ] , ( [ q2:z5 ] ) ) , @xmath51 ( fig .
[ f : q2z1 ] , ( [ q2:z1 ] ) ) , @xmath55 ( mirror - symmetric to the previous case ) .
of course , if @xmath98 , it may be more efficient to use the relation mirror - symmetric to ( [ q2:tri2 ] )
. the relation ( [ q2:larin ] ) also can be used instead of ( [ q2:tri2 ] ) .
( 82,82 ) ( 41,41)(0,0),title="fig : " ] ( 33,71)(0,0)[l ] ( 33,61)(0,0)[l ] ( 32,17 ) ( 16,8.5)(0,0 ) ( 104,17 ) ( 11,8.5)(0,0 ) ( 33,8.5)(0,0)[l]@xmath99 ( 68,8.5)(0,0 ) ( 85,8.5)(0,0)[l]@xmath100 ( 44,26 ) ( 22,13)(0,0)-loop massless sunset diagrams , title="fig : " ] ( 22,17)(0,0)@xmath101
let s summarize .
there is 1 generic topology of 2-loop massless propagator diagrams ( fig .
[ f : q2 t ] ) .
the integrals with all integer @xmath102 can be expressed as linear combinations of 2 basis integrals ( fig .
[ f : q2b ] ) ; coefficients are rational functions of @xmath103 . here
@xmath104 are @xmath30-loop massless sunset diagrams ( fig .
[ f : qsun ] ) : @xmath105 using inversion ( [ q1:inv ] ) of both loop momenta and @xmath106 we obtain relation of fig . [ f : q2i ] @xcite ( with 1440 elements ) of the diagram of fig .
[ f : q2 t ] .
if we connect the external vertices by a line having @xmath107 , we obtain a logarithmically divergent 3-loop tetrahedron vacuum diagram it is proportional to our original diagram times the logarithm of the ultraviolet cut - off .
we can cut it at any of its lines , and get a diagram of fig .
[ f : q2 t ] with 5 indices @xmath102 out of 6 .
this gives the tetrahedron symmetry group .
also , we have the inversion relations ( fig . [
f : q2i ] ) .
fourier transform to @xmath21-space gives a new integral of the form of fig .
[ f : q2 t ] with @xmath108 .
there is an additional symmetry following from the star - triangle relation @xcite . ] .
( 98,34 ) ( 12,17)(0,0 ) ( 66,17)(0,0 ) ( 39,17)(0,0)@xmath7 ( 2,6)(0,0)[r]@xmath4 ( 22,6)(0,0)[l]@xmath5 ( 2,28)(0,0)[r]@xmath46 ( 22,28)(0,0)[l]@xmath47 ( 13,17)(0,0)[l]@xmath48 ( 56,6)(0,0)[r]@xmath4 ( 76,6)(0,0)[l]@xmath5 ( 56,30)(0,0)[r]@xmath109 ( 52,25)(0,0)[r]@xmath110 ( 76,30)(0,0)[l]@xmath111 ( 80,25)(0,0)[l]@xmath110 ( 67,17)(0,0)[l]@xmath48 there are 3 generic topologies of 3-loop massless propagator diagrams ( fig . [
f : q3 t ] ) .
each has 8 denominators .
there are 9 scalar products of 3 loop momenta @xmath112 and the external momentum @xmath24 .
therefore , for each topology , all scalar products in the numerator can be expressed via the denominators and one selected scalar product .
integration - by - parts recurrence relations for these diagrams have been investigated in @xcite .
they can be used to reduce all integrals of fig .
[ f : q3 t ] , with arbitrary integer powers of denominators and arbitrary ( non - negative ) powers of the selected scalar product in the numerators , to linear combinations of 6 basis integrals ( fig .
[ f : q3b ] ) .
this algorithm has been implemented in the @xcite package @xcite and later re - implemented @xcite in @xcite .
it has also been implemented in the @xcite package @xcite only the last , non - planar , topology in fig .
[ f : q3 t ] involve the last , non - planar , basis integral in fig .
[ f : q3b ] .
( 94,60 ) ( 47,45)(0,0 ) ( 21,15)(0,0 ) ( 73,15)(0,0 ) ( 129,77 ) ( 21,68.5)(0,0 ) ( 43,68.5)(0,0)[l]@xmath113 ( 83,68.5)(0,0 ) ( 105,68.5)(0,0)[l]@xmath114 ( 21,42.5)(0,0 ) ( 43,42.5)(0,0)[l]@xmath115 ( 83,42.5)(0,0 ) ( 105,42.5)(0,0)[l]@xmath116 ( 21,12.5)(0,0 ) ( 43,12.5)(0,0)[l]@xmath117 ( 108,12.5)(0,0 ) ( 102,70 ) ( 11,52.5)(0,0 ) ( 73,52.5)(0,0 ) ( 37,52.5)(0,0)@xmath7 ( 11,17.5)(0,0 ) ( 73,17.5)(0,0 ) ( 37,17.5)(0,0)@xmath7 ( 1,43)(0,0)[r]@xmath4 ( 21,43)(0,0)[l]@xmath5 ( -2,59)(0,0)[r]@xmath46 ( 24,59)(0,0)[l]@xmath47 ( 12,46)(0,0)[l]@xmath48 ( 11,65)(0,0)[b]@xmath118 ( 7.5,55)(0,0)[r]@xmath119 ( 14.5,55)(0,0)[l]@xmath120 ( 63,43)(0,0)[r]@xmath4 ( 83,43)(0,0)[l]@xmath5 ( 62,62)(0,0)[r]@xmath109 ( 59,58)(0,0)[r]@xmath121 ( 84,62)(0,0)[l]@xmath111 ( 87,58)(0,0)[l]@xmath122 ( 74,46)(0,0)[l]@xmath48 ( 73,65)(0,0)[b]@xmath123 ( 69.5,55)(0,0)[r]@xmath119 ( 76.5,55)(0,0)[l]@xmath120 ( -2,11)(0,0)[r]@xmath4 ( 24,11)(0,0)[l]@xmath5 ( 11,5)(0,0)[t]@xmath46 ( -2,24)(0,0)[r]@xmath47 ( 24,24)(0,0)[l]@xmath48 ( 5,17.5)(0,0)[l]@xmath118 ( 17,17.5)(0,0)[r]@xmath119 ( 11,30)(0,0)[b]@xmath120 ( 60,11)(0,0)[r]@xmath4 ( 86,11)(0,0)[l]@xmath5 ( 73,5)(0,0)[t]@xmath46 ( 62,27)(0,0)[r]@xmath109 ( 59,23)(0,0)[r]@xmath124 ( 84,27)(0,0)[l]@xmath111 ( 87,23)(0,0)[l]@xmath125 ( 67,17.5)(0,0)[l]@xmath118 ( 79,17.5)(0,0)[r]@xmath119 ( 73,30)(0,0)[b]@xmath126 ( 94,25 ) ( 21,12.5)(0,0 ) , @xmath127,title="fig : " ] ( 73,12.5)(0,0 ) , @xmath127,title="fig : " ] ( 47,12.5)(0,0)@xmath7 the first 4 basis integrals are trivial : they are expressed via @xmath104 ( [ q2:sun ] ) ( fig .
[ f : qsun ] ) , and hence via @xmath128 functions .
the 4-th one differs from the 3-rd one ( @xmath129 ) by replacing the 2-loop subdiagram : the second one in fig .
[ f : q2b ] ( @xmath130 ) by the first one ( @xmath131 ) .
therefore , it can be obtained from @xmath129 by multiplying by @xmath132 the 5-th basis integral is proportional to the 2-loop diagram @xmath133 with a non - integer index of the middle line @xmath134 , and will be discussed in sect .
[ s : gn ] .
the 6-th one , non - planar , is truly 3-loop and most difficult ; it will be discussed in sect .
[ s : np ] .
performing inversion ( [ q1:inv ] ) of the loop momenta , we obtain the relations in fig .
[ f : q3i ] .
for example , the ladder diagram with all indices @xmath135 is convergent ; its value at @xmath127 is related to a simpler diagram ( fig .
[ f : q3i2 ] ) by the second inversion relation . the non - planar topology ( fig .
[ f : q3 t ] ) involves lines with sums of 3 momenta ; they do nt transform into anything reasonable under inversion , and there is no inversion relation for it .
this diagram is @xcite @xmath136\ , .
\nonumber\end{gathered}\ ] ] this is a particular case of a more general result @xcite for the diagram with 3 non - integer indices
. the two lines with unit indices must be adjacent ( fig .
[ f : gn ] ) ; all such diagrams are equivalint , due to the tetrahedron symmetry mentioned above .
this result was derived using gegenbauer polynomial technique @xcite ; we shall discuss it for a simpler example in sect .
[ s : in ] . when @xmath30 is integer , the @xmath137 in ( [ gn : f ] )
is expressed via @xmath128 functions ; in order to show equivalence to the standard results ( sect .
[ s : q2 ] ) , one should use @xmath138 and trigonometric identities .
( 152,26 ) ( 22,13)(0,0 ) ( 12,2)(0,0)[r]@xmath4 ( 32,2)(0,0)[l]@xmath5 ( 23,13)(0,0)[l]@xmath48 ( 76,13)(0,0 ) ( 66,2)(0,0)[r]@xmath4 ( 66,24)(0,0)[r]@xmath46 ( 77,13)(0,0)[l]@xmath48 ( 130,13)(0,0 ) ( 120,2)(0,0)[r]@xmath4 ( 140,2)(0,0)[l]@xmath5 ( 120,24)(0,0)[r]@xmath46 similar expressions for these diagrams were also obtained in @xcite functions of unit argument was communicated by d. broadhurst to a. kotikov before @xcite . ] . in our particular case , they reduce to @xmath139 kazakov
@xcite obtained an expression for @xmath133 via two @xmath137 functions of argument @xmath140 much earlier .
recently , @xmath63 for arbitrary indices was calculated @xcite in terms of a double integral or several double series .
these series can be systematically expanded in @xmath1 up to any order , and coefficients are expressed via multiple @xmath2-values ( sect .
[ s : zeta ] ) . in order to calculate the 2-loop diagram of fig .
[ f : q2 t ] with a 1-loop insertion in the middle light line , we need @xmath141 .
it is easy to shift @xmath48 by @xmath142 using the relation @xcite @xmath143 g(1,1,1,1,n_5 ) = 2\,{\mathbf{1}^+ } ( { \mathbf{3}^- } - { \mathbf{2}^-}{\mathbf{5}^+ } ) g(1,1,1,1,n_5)\ , , \label{gn : n5}\ ] ] which follows from integration - by - parts relations ( sect .
[ s : q2 ] ) ( all terms in its right - hand side are trivial ) .
therefore , it is sufficient to find it just for one @xmath30 .
the simplest choice for which the algorithm of @xmath1-expansion ( sect .
[ s : hyper ] ) works for ( [ gn : f ] ) is @xmath144 : @xmath145\ , .
\end{split } \label{gn : f2}\ ] ] expansion of this @xmath137 in @xmath1 is ( sect .
[ s : hyper ] ) @xmath146 similarly , we obtain from ( [ gn : bgk ] ) for @xmath147 @xmath148 where @xmath149 also , the last formula in ( [ gn : bgk ] ) gives for @xmath150 @xmath151\ , , \end{split } \label{gn : bgk0}\ ] ] where @xmath152 it is not difficult to find several additional terms to ( [ gn : fe ] ) , ( [ gn : fe1 ] ) , ( [ gn : fe0 ] ) .
of course , all 4 series ( [ gn : bgk1 ] ) for @xmath153 agree , and those for @xmath154 ( [ gn : f2 ] ) and @xmath155 ( [ gn : bgk0 ] ) agree with the recurrence relation ( [ gn : n5 ] ) . the non - planar basis diagram ( the last one in fig .
[ f : q3b ] ) is finite ; its value at @xmath156 can be obtained without calculations by the method called gluing @xcite .
let s consider the 4-loop vacuum diagram in the middle of fig .
[ f : glue ] .
let all lines have mass @xmath157 , then there are no infrared divergences
. there are no divergent subdiagrams ; the diagram has an overall ultraviolet divergence , and hence a @xmath158 pole .
( 70,82 ) ( 21,69.5)(0,0 ) ( 43,69.5)(0,0)[l]@xmath159 ( 21,41)(0,0 ) ( 43,41)(0,0)[l]@xmath160 ( 21,12.5)(0,0 ) ( 43,12.5)(0,0)[l]@xmath159
we can imagine that the middle vertex consists of 2 vertices connected by a shrunk line ( with power 0 ) , one vertex involves 2 lines on the left , and the other one 2 lines on the right . if we cut this shrunk line , we get the ladder diagram ( the upper one in fig . [ f : glue ] ) .
it is finite .
therefore , the ultraviolet divergence of the vacuum diagram comes from the last integration in @xmath161 of the ladder diagram . at @xmath162
, we may neglect @xmath157 ; by dimensionality , the massless ladder diagram behaves as @xmath163 . reducing it to the basis integrals and using results of sect .
[ s : gn ] , we can obtain @xmath164 ( first line of fig .
[ f : glue ] ) term was found in @xcite using gegenbauer polynomials in @xmath21-space @xcite ; however , it is difficult to obtain further terms of @xmath1 expansion by this method . ] .
it is not difficult to obtain several additional terms .
the ultraviolet divergence of the vacuum diagram comes from @xmath165 where @xmath166 is an infrared cutoff , see the middle line in fig .
[ f : glue ] .
on the other hand , we can imagine that one of the vertices in the middle of the vacuum diagram involves the upper left line and the lower right one , and the other vertex the lower left line and the upper right one .
cutting the shrunk line , we get the non - planar diagram ( the lower one in fig .
[ f : glue ] ) .
it is finite ; at @xmath162 , where @xmath157 can be neglected , it behaves as @xmath167 .
integrating it in @xmath161 , we obtain the same vacuum diagram , with the same @xmath158 pole . therefore , @xmath168 : _ the non - planar basis diagram has the same value at @xmath127 as the ladder one _ @xcite , namely , @xmath169 ( fig .
[ f : glue ] ) .
this method tells us nothing about further terms of expansion in @xmath1 .
the highly non - trivial problem of calculating the @xmath170 term in the non - planar basis diagram was solved in @xcite .
heavy quark effective theory ( hqet , see , e.g. , @xcite ) is an effective field theory constructed to approximate results of qcd for certain problems with a single heavy quark .
let the heavy quark stay approximately at rest ( in some frame ) ; its momentum and energy are @xmath171 , @xmath172 .
momenta and energies of light quarks and gluons are @xmath173 , @xmath174 . here
@xmath175 is some fixed characteristic momentum , and we consider the limit @xmath176 . scattering amplitudes and on - shell matrix elements of operators in qcd , expanded up to some order in @xmath177 , can be reproduced from a simpler theory hqet .
the lowest - energy state of this theory ( `` vacuum '' ) is a single heavy quark at rest , and has energy @xmath157 .
therefore , for any system containing this quark it is natural to measure energy relative this zero level , i.e. , to consider its residual energy @xmath178 .
the free heavy quark has dispersion law ( dependence of energy on momentum , or mass shell ) @xmath179 .
neglecting @xmath180 corrections , we may simplify it to @xmath181 : the heavy - quark energy is zero , and does not depend on its momentum .
the heavy quark at rest is described by a 2-component spinor field @xmath182 , or a 4-component spinor having only upper components : @xmath183 .
the dispersion law @xmath181 follows from the lagrangian @xmath184 . reintroducing the interaction with the gluon field by the requirement of gauge invariance
, we obtain the leading - order ( in @xmath180 ) hqet lagrangian @xmath185 where all light - field parts ( denoted by dots ) are exactly the same as in qcd .
the heavy quark interacts with @xmath186 : creates the coulomb chromoelectric field , and reacts to external chromoelectric fields .
the lagrangian ( [ crash : l ] ) gives the heavy quark propagator @xmath187 in the momentum space it depends only on @xmath188 but not on @xmath189 , because we have neglected the kinetic energy .
therefore , in the coordinate space the heavy quark does not move .
the vertex is @xmath190 , where @xmath191 is the heavy - quark 4-velocity .
heavy - quark loops vanish , because it propagates only forward in time .
hqet is not lorentz - invariant , because it has a preferred frame the heavy - quark rest frame .
however , it can be rewritten in covariant notations .
momentum of any system containing the heavy quark is decomposed as @xmath192 where the residual momentum is small : @xmath193 .
the heavy - quark field obeys @xmath194 .
the lagrangian is @xmath195 it gives the propagator @xmath196 and the vertex @xmath190 .
the qcd propagator at large @xmath157 becomes @xmath197 a vertex @xmath198 sandwiched between two projectors @xmath199 may be replaced by @xmath200 .
therefore , at the tree level all qcd diagrams become the corresponding hqet diagrams , up to @xmath180 corrections . in loop diagrams , momenta can be arbitrarily large , and this correspondence breaks down .
renormalization properties of hqet differ from those of qcd .
the ultraviolet behavior of an hqet diagram is determined by the region of loop momenta much larger than the characteristic momentum scale of the process @xmath201 , but much less than the heavy quark mass @xmath157 ( which tends to infinity from the very beginning ) .
it has nothing to do with the ultraviolet behavior of the corresponding qcd diagram , which is determined by the region of loop momenta much larger than @xmath157 . the 1-loop hqet propagator diagram ( fig .
[ f : h1 ] ) is @xmath202 it vanishes if @xmath13 or @xmath203 .
( 64,27 ) ( 32,13.5)(0,0 ) ( 32,0)(0,0)[b]@xmath204 ( 32,27)(0,0)[t]@xmath12 ( 32,8)(0,0)[b]@xmath4 ( 32,19)(0,0)[t]@xmath5 the fourier transform of the 1-dimensional ( hqet ) propagator is @xmath205 our diagram in @xmath21-space is the product of the heavy propagator ( [ h1:p2x ] ) with the power @xmath4 and the light propagator ( [ q1:p2x ] ) with the power @xmath5 .
the heavy quark stays at rest : @xmath206 .
therefore , @xmath207 in ( [ q1:p2x ] ) ?
the wick rotation to the euclidean @xmath21-space is @xmath208 .
there are no imaginary parts at @xmath209 . ] .
our diagram in the @xmath21-space is @xmath210 transforming it back to @xmath24-space ( [ h1:x2p ] ) , we arrive at @xmath211 there are 2 generic topologies of 2-loop hqet propagator diagrams .
the first one is ( fig .
[ f : h2ti ] ) : @xmath212 this diagram is symmetric with respect to @xmath39 .
it vanishes if indices of any two adjacent lines are non - positive .
( 64,27 ) ( 32,16.5)(0,0),title="fig : " ] ( 16,22)(0,0)@xmath41 ( 48,22)(0,0)@xmath42 ( 22,5)(0,0)@xmath213 ( 42,5)(0,0)@xmath214 ( 41,15)(0,0)@xmath45 ( 22,11.5)(0,0)@xmath4 ( 42,11.5)(0,0)@xmath5 ( 24,19)(0,0)@xmath46 ( 40,19)(0,0)@xmath47 ( 28,15)(0,0)@xmath48 ( 52,19 ) ( 26,9)(0,0),title="fig : " ] ( 16,0)(0,0)[b]@xmath4 ( 36,0)(0,0)[b]@xmath5 ( 16,17)(0,0)[t]@xmath46 ( 36,17)(0,0)[t]@xmath47 ( 134,20 ) ( 25,11.5)(0,0),title="fig : " ] ( 76,11.5)(0,0),title="fig : " ] ( 118,11.5)(0,0),title="fig : " ] ( 55,11.5)(0,0)@xmath7 ( 97,11.5)(0,0)@xmath57 ( 25,0)(0,0)[b]@xmath5 ( 15,18)(0,0)@xmath46 ( 35,18)(0,0)@xmath47 ( 24,9)(0,0)@xmath48 ( 76,0)(0,0)[b]@xmath48 ( 76,24)(0,0)[t]@xmath58 ( 118,2)(0,0)[b]@xmath5 ( 118,21)(0,0)[t]@xmath59 ( 134,20 ) ( 25,11.5)(0,0),title="fig : " ] ( 76,11.5)(0,0),title="fig : " ] ( 118,11.5)(0,0),title="fig : " ] ( 55,11.5)(0,0)@xmath7 ( 97,11.5)(0,0)@xmath57 ( 17,0)(0,0)[b]@xmath4 ( 33,0)(0,0)[b]@xmath5 ( 25,21)(0,0)[t]@xmath47 ( 26,10)(0,0)@xmath48 ( 76,2)(0,0)[b]@xmath4 ( 76,20)(0,0)[t]@xmath48 ( 118,2)(0,0)[b]@xmath215 ( 118,20)(0,0)[t]@xmath47 if @xmath49 , our diagram is just the product of two 1-loop diagrams ( fig .
[ f : h2z5 ] ) : @xmath216 if @xmath51 , the inner loop gives @xmath52 , and ( fig .
[ f : h2z1 ] ) @xmath217 if @xmath55 , the inner loop gives @xmath218 , and ( fig .
[ f : h2z3 ] ) @xmath219 the cases @xmath54 , @xmath56 are symmetric .
when all @xmath60 , we shall use integration by parts @xcite .
the differential operator @xmath64 applied to the integrand of ( [ h2:i ] ) acts as @xmath220 applying @xmath66 , @xmath80 to the integrand of ( [ h2:i ] ) , we get vanishing integrals . on the other hand , using @xmath221 , @xmath69 , we see that these differential operators are equivalent to inserting @xmath222 under the integral sign .
we obtain the recurrence relations @xmath223 i = 0\ , , \label{h2:tri1}\\ \left [ d - n_2-n_4 - 2n_5 + n_2{\mathbf{2}^+}{\mathbf{1}^- } + n_4{\mathbf{4}^+}({\mathbf{3}^-}-{\mathbf{5}^- } ) \right ] i = 0 \label{h2:tri2}\end{gathered}\ ] ] ( two more relations are obtained by @xmath40 ) .
similarly , applying the differential operator @xmath224 is equivalent to inserting @xmath225 and we obtain @xmath226 i = 0\ , .
\label{h2:tri3}\ ] ] ( there is also the symmetric relation , of course ) .
we can obtain a relation from homogeneity of the integral ( [ h2:i ] ) in @xmath227 . applying the operator @xmath228 to @xmath229 ( [ h2:i ] ) , we see that it is equivalent to the factor @xmath230 . on the other hand , explicit differentiation of @xmath231
gives @xmath232 .
therefore , @xmath233 i = 0\ , .
\label{h2:hom}\ ] ] this is nothing but the sum of the @xmath66 relation ( [ h2:tri1 ] ) and its mirror - symmetric @xmath86 relation . a useful relation can be obtained by subtracting @xmath234 shifted ( [ h2:hom ] ) from ( [ h2:tri2 ] ) : @xmath235 i = 0\ , .
\end{split } \label{h2:david}\ ] ] expressing @xmath236 from ( [ h2:tri2 ] ) @xmath237 we see that the sum @xmath95 reduces by 1 ( fig .
[ f : red ] ) .
therefore , applying ( [ q2:tri ] ) sufficiently many times , we can reduce an arbitrary @xmath238 integral with integer indices to a combination of integrals with @xmath49 ( fig .
[ f : h2z5 ] , ( [ h2:z5 ] ) ) , @xmath51 ( fig .
[ f : h2z1 ] , ( [ h2:z1 ] ) ) , @xmath55 ( fig . [ f : h2z3 ] , ( [ h2:z3 ] ) ) .
of course , if @xmath98 , it may be more efficient to use the relation mirror - symmetric to ( [ h2:tri2 ] ) .
the relation ( [ h2:david ] ) also can be used instead of ( [ h2:tri2 ] ) .
the second topology of 2-loop hqet propagator diagrams is ( fig .
[ f : h2tj ] ) : @xmath239 this diagram is symmetric with respect to @xmath240 .
it vanishes if @xmath241 , or @xmath242 , or two adjacent heavy indices are non - positive . if @xmath55 , our diagram is just the product of two 1-loop diagrams ( fig .
[ f : h2z5 ] ) . if @xmath51 , it is also trivial ( fig .
[ f : h2z3 ] ) .
( 72,42 ) ( 36,21)(0,0),title="fig : " ] ( 46,0)(0,0)[b]@xmath42 ( 26,42)(0,0)[t]@xmath41 ( 16,19.5)(0,0)[t]@xmath213 ( 40,19.5)(0,0)[t]@xmath243 ( 56,22.5)(0,0)[b]@xmath214 ( 16,22.5)(0,0)[b]@xmath4 ( 36,22.5)(0,0)[b]@xmath46 ( 59,19.5)(0,0)[t]@xmath244 ( 26,34)(0,0)[t]@xmath47 ( 46,8)(0,0)[b]@xmath48 the denominators in ( [ h2:j ] ) are linearly dependent : @xmath245 . therefore @xcite @xmath246 applying this relation sufficiently many times , we can kill one of the lines 1 , 2 , 3 , and thus reduce any integral @xmath247 with integer indices to trivial cases .
in fact , we have not enough independent denominators to express all scalar products in the numerator .
therefore , we have to consider a more general integral than ( [ h2:j ] ) , containing powers of @xmath248 in the numerator . this wider class of integrals can also be reduced to the same trivial cases @xcite .
let s summarize .
there are 2 generic topologies of 2-loop hqet propagator diagrams ( fig .
[ f : h2 t ] ) .
the integrals with all integer @xmath102 can be expressed as linear combinations of 2 basis integrals ( fig .
[ f : h2b ] ) ; coefficients are rational functions of @xmath103 . here
@xmath249 are @xmath30-loop hqet sunset diagrams ( fig .
[ f : hsun ] ) : @xmath250 ( 84,17 ) ( 16,8.5)(0,0 ) ( 63,8.5)(0,0 ) ( 104,10 ) ( 11,5)(0,0 ) ( 33,5)(0,0)[l]@xmath251 ( 68,5)(0,0 ) ( 85,5)(0,0)[l]@xmath252 ( 42,17 ) ( 21,8.5)(0,0)-loop hqet sunset diagrams , title="fig : " ] ( 21,11)(0,0)@xmath101 there are 10 generic topologies of 3-loop hqet propagator diagrams ( fig .
[ f : h3 t ] ) .
diagrams in the first 2 rows of the figure have one scalar product which can not be expressed via denominators ; those in the 3-rd row have 1 linear relation among heavy denominators , and hence 2 independent scalar products in the numerator ; those in the last row have 2 relations among heavy denominators , and hence 3 independent scalar products in the numerator .
integration - by - parts recurrence relations for these diagrams have been investigated in @xcite .
they can be used to reduce all integrals of fig .
[ f : h3 t ] , with arbitrary integer powers of denominators and arbitrary numerators , to linear combinations of 8 basis integrals ( fig .
[ f : h3b ] ) .
this algorithm has been implemented in the package @xcite .
recently , the heavy - quark propagator and the heavy - light current anomalous dimension have been calculated @xcite at 3 loops , using this package .
( 132,109 ) ( 66,54.5)(0,0 ) ( 116,82 ) ( 21,75)(0,0)[b ] ( 43,77)(0,0)[l]@xmath253 ( 83,75)(0,0)[b ] ( 105,77)(0,0)[l]@xmath254 ( 16,57)(0,0)[b ] ( 33,59)(0,0)[l]@xmath255 ( 16,41)(0,0)[b ] ( 33,43)(0,0)[l]@xmath256 ( 68,41)(0,0)[b ] ( 85,43)(0,0)[l]@xmath257 ( 16,21)(0,0)[b ] ( 33,23)(0,0)[l]@xmath258 ( 21,0)(0,0)[b ] ( 43,9)(0,0)[l]@xmath259 ( 100,21)(0,0)[b ] the first 5 basis integrals are trivial : they are expressed via @xmath249 ( [ h2:sun ] ) ( fig .
[ f : hsun ] ) and @xmath104 ( [ q2:sun ] ) ( fig .
[ f : qsun ] ) , and hence via @xmath128 functions .
the integrals in the 3-rd row differ from the one in the 2-nd row by replacing a two - loop subdiagram .
in the first case , the 2-loop hqet subdiagram ( fig .
[ f : h2 t ] ) is substituted : @xmath260 . in the second case , the 2-loop massless subdiagram ( fig .
[ f : q2 t ] ) is substituted : @xmath261 .
therefore , these diagrams are obtained from @xmath262 by multiplying by and @xmath263 as elements of its basis , not the diagrams in fig . [
f : h3b ] . ]
@xmath264 two basis integrals are proportional to @xmath265 ( sect .
[ s : in ] ) and @xmath266 ( sect .
[ s : jn ] ) , for non - integer @xmath30 .
the last one is truly 3-loop and most difficult ; it will be discussed in sect .
[ s : m3 ] . this diagram ( fig .
[ f : jt ] ) has been calculated in @xcite . in @xmath21-space
, it is is @xmath267 or , going to euclidean space ( @xmath268 ) , @xmath269 where the power of @xmath270 is obvious from counting dimensions . collecting factors from fourier transforms , we have @xmath271 where the dimensionless integral @xmath247 is @xmath272 the substitution @xmath273 gives @xmath274 ( 42,17 ) ( 21,8.5)(0,0 ) in @xmath21-space , title="fig : " ] ( 6,7)(0,0)[t]@xmath18 ( 16,10)(0,0)[b]@xmath275 ( 26,7)(0,0)[t]@xmath276 ( 36,10)(0,0)[b]@xmath277 ( 21,10)(0,0)[b]@xmath30 integrations in @xmath270 and @xmath21 would factor , if not the factor @xmath278 .
let s expand this troublesome factor according to the newton binomial : @xmath279 here the pochhammer symbol is @xmath280 then the integrations become trivial : @xmath281 recalling the definition of the hypergeometric function @xmath282 and taking into account @xmath283 , we can rewrite the result as @xmath284 and hence ( [ jn : jn ] ) @xmath285 this is the easiest diagram calculation in the world having @xmath137 as its result : just a double integral .
it is not more difficult to obtain the general result in @xcite , i ve no idea why : the general case is not more difficult . ]
@xmath286 in order to calculate the 2-loop diagram of fig .
[ f : h2tj ] with a 1-loop insertion in the middle heavy line , we need @xmath287 .
it is easy to shift @xmath46 by @xmath142 using ( [ h2:parfrac ] ) .
therefore , it is sufficient to find it just for one @xmath30 .
the simplest choice for which the algorithm of @xmath1-expansion ( sect .
[ s : hyper ] ) works is @xmath288 .
we obtain @xcite from ( [ jn : jf ] ) @xmath289 expansion of this @xmath137 in @xmath1 is ( sect .
[ s : hyper ] ) @xmath290 it is not difficult to find several additional terms . this diagram ( fig .
[ f : it ] ) has been calculated in @xcite , using gegenbauer polynomials in @xmath21-space @xcite .
this is , probably , the simplest example of applying this useful technique , therefore , we shall consider some details of this calculation . in euclidean @xmath21-space , this diagram is @xmath291^\lambda \left[(x+v(t - t_1))^2\right]^\lambda } = \frac{2\pi^{d/2}}{\gamma(d/2 ) } i t^{2(n - d)+5}\,,\ ] ] where @xmath292 , and the power of @xmath270 is obvious from counting dimensions .
the middle vertex has been chosen as the origin , because the propagator with the non - standard power begins at this vertex ( it is much easier to handle propagators beginning at the origin ) .
collecting factors from fourier transforms , we have @xmath293 the dimensionless integral @xmath238 is @xmath294^\lambda \left[(x+v(1-t))^2\right]^\lambda}\,,\ ] ] where @xmath295 and the angular integration measure @xmath296 is normalized : @xmath297 ( 32,12 ) ( 16,6)(0,0 ) in @xmath21-space , title="fig : " ] ( 16,-5)(0,0)[b]@xmath298 ( 26,-5)(0,0)[b]@xmath299 ( 4,-5)(0,0)[b]@xmath300 ( 16,15)(0,0)[t]@xmath21 this problem is more difficult than the one in sect .
[ s : jn ] : now we have @xmath103-dimensional integration over the coordinates of the light vertex .
it can be simplified by expanding 2 massless propagators ( whose ends are not at the origin ) in series in gegenbauer polynomials @xcite @xmath301^\lambda } = \frac{1}{\max^{2\lambda}(x , y ) } \sum_{k=0}^\infty t^k(x , y ) c^\lambda_k(\hat{x}\cdot\hat{y})\ , , \label{in : gegen}\ ] ] where @xmath302 then the angular integration can be easily done using the orthogonality relation @xmath303 in our particular case , the unit vectors @xmath304 and @xmath305 are @xmath306 and @xmath307 , and the result involves @xmath308 now we have a single sum : @xmath309^k}{\left[\max(x , t)\max(x,1-t)\right]^{d-2}}\,.\end{gathered}\ ] ] the contributions of the regions @xmath310 and @xmath311 are equal , and we may consider the first of them and double it . the radial integral ( in @xmath21 ) has to be calculated in several intervals separately , because of @xmath312 and @xmath313 in ( [ in : gegen ] ) : @xmath314^k}{\left[t(1-t)\right]^{d-2 } } + \int_t^{1-t } dx \frac{x^{2n-1}\left[\frac{t}{x}\frac{x}{1-t}\right]^k}{\left[x(1-t)\right]^{d-2 } } + \int_{1-t}^\infty dx \frac{x^{2n-1}\left[\frac{t}{x}\frac{1-t}{x}\right]^k}{\left[x^2\right]^{d-2}}\,.\ ] ] the result is @xmath315\,.\end{aligned}\ ] ] unfortunately , the integral in @xmath270 can not be immediately evaluated .
the authors of @xcite had to do some more juggling to obtain @xmath316\ , .
\nonumber\end{gathered}\ ] ] this result can also be rewritten @xcite as @xmath317\ , , \end{split } \label{in : f2}\ ] ] this series has larger region of convergence . in order to calculate the 2-loop diagram of fig .
[ f : h2ti ] with a 1-loop insertion in the middle light line , we need @xmath318 .
it is easy to shift @xmath48 by @xmath142 using the relation @xcite @xmath319 i(1,1,1,1,n_5 ) = { } \\ \left[(2d-2n_5 - 7){\mathbf{1}^-}{\mathbf{5}^+ } - { \mathbf{3}^-}{\mathbf{4}^+}{\mathbf{5}^- } + { \mathbf{1}^-}{\mathbf{3}^+ } \right ] i(1,1,1,1,n_5)\ , , \end{split } \label{in : n5}\ ] ] which follows from integration - by - parts relations ( sect .
[ s : h2 ] ) ( all terms in its right - hand side are trivial ) .
therefore , it is sufficient to find it just for one @xmath30 .
the simplest choice for which the algorithm of @xmath1-expansion ( sect .
[ s : hyper ] ) works is to use @xmath320 in ( [ in : f2 ] ) .
we obtain @xcite @xmath321\ , .
\end{split } \label{in : i1}\ ] ] this @xmath137 function has 3 upper indices @xmath322 ; therefore , its expansion ( apart from the leading 1 ) starts from @xmath323 : @xmath324 it is not difficult to find several additional terms .
in this section , we shall consider self - energy diagrams of a massive particle with mass @xmath157 , when the external momentum is on the mass shell : @xmath325 , or @xmath326 .
why this special case is interesting ?
first of all , in order to calculate any @xmath327-matrix elements , one has to consider mass @xcite and wave - function @xcite renormalization in the on - shell scheme .
calculations of form factors and their derivatives at the point @xmath328 belong to this class , e.g. , the anomalous magnetic moment @xcite and electron charge radius @xcite in qed .
finally , such calculations are used for qcd / hqet matching @xcite .
the 1-loop on - shell propagator diagram ( fig .
[ f : m1 ] ) is @xmath329 it vanishes if @xmath13 .
( 64,27 ) ( 32,13.5)(0,0 ) ( 32,0)(0,0)[b]@xmath330 ( 32,27)(0,0)[t]@xmath12 ( 32,8)(0,0)[b]@xmath4 ( 32,19)(0,0)[t]@xmath5 the definition ( [ m1:m ] ) of @xmath331 expressed via the dimensionless euclidean momentum @xmath332 becomes @xmath333 the definition ( [ h1:i ] ) of the 1-loop hqet integral
@xmath334 expressed via the dimensionless euclidean momentum @xmath335 is @xmath336 inversion ( [ q1:inv ] ) transforms the on - shell denominator to the hqet one : @xmath337 ( and vice versa ) .
therefore , the problem of calculating @xmath331 reduces to the previously solved one for @xmath334 ( sect .
[ s : h1 ] ) : @xmath338 ( fig .
[ f : m1i ] ) .
( 74,19 ) ( 16,9)(0,0 ) ( 58,9)(0,0 ) ( 37,9)(0,0)@xmath7 ( 16,0)(0,0)[b]@xmath4 ( 16,19)(0,0)[t]@xmath35 ( 58,0)(0,0)[b]@xmath4 ( 58,19)(0,0)[t]@xmath36 there are 2 generic topologies of 2-loop on - shell propagator diagrams .
the first one is ( fig .
[ f : m2tm ] ) : @xmath339 the second topology is ( fig .
[ f : m2tn ] ) : @xmath340 ( 64,27 ) ( 32,16.5)(0,0),title="fig : " ] ( 16,22)(0,0)@xmath41 ( 48,22)(0,0)@xmath42 ( 22,5)(0,0)@xmath341 ( 42,5)(0,0)@xmath342 ( 41,15)(0,0)@xmath45 ( 22,11.5)(0,0)@xmath4 ( 42,11.5)(0,0)@xmath5 ( 24,19)(0,0)@xmath46 ( 40,19)(0,0)@xmath47 ( 28,15)(0,0)@xmath48 ( 72,42 ) ( 36,21)(0,0),title="fig : " ] ( 46,0)(0,0)[b]@xmath42 ( 26,42)(0,0)[t]@xmath41 ( 16,19.5)(0,0)[t]@xmath341 ( 40,19.5)(0,0)[t]@xmath343 ( 56,22.5)(0,0)[b]@xmath342 ( 16,22.5)(0,0)[b]@xmath4 ( 36,22.5)(0,0)[b]@xmath46 ( 59,19.5)(0,0)[t]@xmath244 ( 26,34)(0,0)[t]@xmath47 ( 46,8)(0,0)[b]@xmath48 ( 112,26 ) ( 21,13)(0,0 ) ( 82,13)(0,0 ) ( 87,17 ) ( 43.5,8.5)(0,0 ) the integrals of both generic topologies ( fig .
[ f : m2 t ] ) with all integer @xmath102 can be reduced , by using integration - by - parts recurrence relations , to linear combinations of 3 basis integrals ( fig . [
f : m2b ] ) ; coefficients are rational functions of @xmath103 .
only the topology @xmath344 involve the last , and the most difficult , basis integral in fig .
[ f : m2b ] .
the reduction algorithm has been constructed and implemented as package in @xcite .
it also has been implemented as package @xcite .
the first 2 basis integrals are trivial .
the @xmath30-loop on - shell sunset @xmath345 ( fig .
[ f : msun ] ) is @xmath346 in the 1-loop case , it can be reduced to the 1-loop massive vacuum diagram @xmath347 ( fig . [
f : msun1 ] ) .
( 42,17 ) ( 21,8.5)(0,0)-loop on - shell sunset diagrams , title="fig : " ] ( 21,11)(0,0)@xmath101 ( 76,11 ) ( 16,5.5)(0,0 ) ( 74,5.5)(0,0 ) ( 47,5.5)(0,0)@xmath348 instead of using @xmath349 ( the last diagram in fig .
[ f : m2b ] , which is divergent ) as an element of the basis , one may use the convergent diagram @xmath350 ( the second one in fig .
[ f : m2 t ] ) .
it has been calculated in @xcite : @xmath351\ , .
\end{split } \label{m2:f}\ ] ] its value at @xmath156 is @xcite @xmath352 d. broadhurst discovered @xcite a symmetry group of a class of @xmath137 functions like those in ( [ m2:f ] ) ; this allowed him to obtain ( [ m2:n1 ] ) and two further terms , up to @xmath353 , by purely algebraic methods .
using inversion of euclidean dimensionless momenta ( [ q1:inv ] ) and ( [ m1:inv ] ) , we obtain @xcite relation of fig . [
f : m2i ] .
( 94,22 ) ( 11,13)(0,0 ) ( 68,13)(0,0 ) ( 37,13)(0,0)@xmath7 ( 4.5,0)(0,0)[b]@xmath4 ( 18.5,0)(0,0)[b]@xmath5 ( 3,14)(0,0)[r]@xmath46 ( 19,14)(0,0)[l]@xmath47 ( 12,9)(0,0)[l]@xmath48 ( 60.5,0)(0,0)[b]@xmath4 ( 75.5,0)(0,0)[b]@xmath5 ( 64,19)(0,0)[r]@xmath109 ( 60,15)(0,0)[r]@xmath110 ( 72,19)(0,0)[l]@xmath111 ( 76,15)(0,0)[l]@xmath110 ( 69,9)(0,0)[l]@xmath48 in realistic theories , there are several massive particles with different masses , as well as massless particles ( qed with @xmath354 , @xmath355 , @xmath356 ; qcd with @xmath357 , @xmath358 , @xmath270 ; the standard model ) .
therefore , diagrams like in fig .
[ f : m22 ] appear . here
the thin line is massless , the thick solid line has mass @xmath157 , and the thick dashed one has mass @xmath359 . combining identical denominators , we obtain the topology shown in fig . [
f : m22 t ] .
( 42,13 ) ( 21,6.5)(0,0 ) ( 32,17 ) ( 16,8.5)(0,0 ) an algorithm for calculating such diagrams based on integration - by - parts recurrence relations has been constructed in @xcite .
the algorithm has been implemented in .
there is one scalar product which can not be expressed via the denominators .
the integrals of fig .
[ f : m22 t ] with arbitrary integer powers of the denominators , and with arbitrary non - negative power of the scalar product in the numerator , can be expressed as linear combinations of 4 basis integrals ( fig .
[ f : m22b ] ) ; coefficients are rational functions of @xmath103 , @xmath157 , and @xmath359 .
( 107,17 ) ( 53.5,8.5)(0,0 ) instead of using the last diagram in fig .
[ f : m22b ] as an element of the basis , one may prefer to use the sunset diagram ( the third one in [ f : m22b ] ) with one of the denominators squared ( either the one with mass @xmath157 or the one with mass @xmath359 ) .
the reduction algorithm for the sunset diagrams with the masses @xmath157 , @xmath359 , @xmath359 and @xmath325 , constructed in @xcite , is just a particular case of the algorithm from @xcite . the first 2 basis integrals in fig .
[ f : m22b ] are just products of 1-loop massive vacuum diagrams , and thus trivial .
the 2 nontrivial ones have been calculated in @xcite using the mellin - barnes representation technique : @xmath360 if @xmath361 , they are not independent @xmath362 can be reduced to @xmath363 ( the last diagram in fig .
[ f : m2b ] ) using integration by parts ( sect .
[ s : m2 ] ) : @xmath364 we can express @xmath350 via @xmath363 or @xmath362 ( [ m2m : i ] ) ; in the second case , for example , we have @xmath365\\ & { } + \frac{1}{(d-3)(d-4 ) } \left [ 2 \frac{\gamma^2(1-\varepsilon)\gamma(1 + 2\varepsilon)\gamma(1 - 4\varepsilon)}{\gamma(1+\varepsilon)\gamma(1 - 2\varepsilon)\gamma(1 - 3\varepsilon ) } - 3 \right ] \biggr\}\ , .
\end{split } \label{m2m : n1}\ ] ] this should be equivalent to ( [ m2:f ] ) .
the basis integrals ( [ m2m : i ] ) , expanded @xcite in @xmath1 up to @xmath366 , @xmath367\\ & { } - m^{\prime2 - 4\varepsilon } \left [ \frac{1}{\varepsilon^2 } + \frac{3}{\varepsilon } - 2 \log r + 6 \right ] + \mathcal{o}(\varepsilon)\,,\\ \frac{i_1}{\gamma^2(1+\varepsilon ) } = { } & m^{-4\varepsilon } \left [ \frac{1}{2\varepsilon^2 } + \frac{5}{2\varepsilon } + 2 ( 1+r)^2 l_+ + 2 ( 1-r)^2 l_- - 2 \log^2 r + \frac{19}{2 } \right]\\ & { } + \mathcal{o}(\varepsilon)\ , , \end{split } \label{m2m : ie}\ ] ] ( where @xmath368 ) , are expressed via dilogarithms : @xmath369 the sunset basis integral @xmath363 ( [ m2m : i ] ) ( the third one in fig . [
f : m22b ] ) was recently expanded @xcite up to @xmath370 ; the coefficients involve harmonic polylogarithms of @xmath371 .
in particular , at @xmath372 , one more term in ( [ m2:n1 ] ) ( @xmath323 ) can be obtained .
there are 11 generic topologies of 3-loop massive on - shell propagator diagrams ( fig .
[ f : m3 t ] ) .
ten of them are the same as in hqet ( fig .
[ f : h3 t ] ) , plus one additional topology with a heavy loop ( in hqet , diagrams with heavy - quark loops vanish ) .
integration - by - parts recurrence relations for these diagrams have been investigated in @xcite .
they can be used to reduce all integrals of fig .
[ f : m3 t ] , with arbitrary integer powers of denominators and arbitrary numerators , to linear combinations of the basis integrals in fig .
[ f : m3b ] .
this algorithm has been implemented in the package @xcite . some of the basis integrals are trivial .
others were found , to some orders in @xmath1 , during many years , in the course of calculation of the anomalous magnetic moment in qed at 3 loops @xcite .
some additional terms of @xmath1 expansions were obtained in @xcite .
( 132,109 ) ( 66,54.5)(0,0 ) ( 160,57 ) ( 80,28.5)(0,0 ) performing inversion ( [ q1:inv ] ) of the loop momenta , we obtain the relations in fig .
[ f : m3i ] .
for example , the hqet ladder diagram with all indices @xmath135 is convergent ; its value at @xmath127 is related @xcite to a massive on - shell diagram ( fig . [ f : m3i2 ] ) by the second inversion relation .
this is one of the basis diagrams of fig .
[ f : m3b ] , and its value at @xmath127 is known ( fig . [
f : m3i2 ] ) . calculating this ladder diagram with ( sect .
[ s : h3 ] ) and solving for the most difficult hqet 3-loop basis integral in fig .
[ f : h3b ] , we obtain the @xmath1 expansion of this integral up to @xmath170 . this concludes the investigation of the basis integrals of fig . [ f : h3b ] ( sects . [
s : h3][s : in ] ) , and allows one to solve 3-loop propagator problems in hqet up to terms @xmath366 ( see @xcite ) .
( 92,50 ) ( 36,25)(0,0 ) ( 3.5,26)(0,0)[b]@xmath4 ( 18.5,26)(0,0)[b]@xmath5 ( 53.5,26)(0,0)[b]@xmath4 ( 68.5,26)(0,0)[b]@xmath5 ( -1.5,37.5)(0,0)[r]@xmath46 ( 23,37.5)(0,0)[l]@xmath47 ( 51,43)(0,0)[r]@xmath109 ( 49,39)(0,0)[r]@xmath121 ( 70.5,43)(0,0)[l]@xmath111 ( 72,39)(0,0)[l]@xmath122 ( 12,33.75)(0,0)[l]@xmath48 ( 62,33.75)(0,0)[l]@xmath48 ( 7.7,39)(0,0)[r]@xmath119 ( 15,39)(0,0)[l]@xmath120 ( 57.7,39)(0,0)[r]@xmath119 ( 65,39)(0,0)[l]@xmath120 ( 11,46)(0,0)[b]@xmath118 ( 61,46)(0,0)[b]@xmath123 ( 36,30)(0,0)@xmath373 ( 1,1)(0,0)[b]@xmath4 ( 11,1)(0,0)[b]@xmath46 ( 21,1)(0,0)[b]@xmath5 ( 51,1)(0,0)[b]@xmath4 ( 61,1)(0,0)[b]@xmath46 ( 71,1)(0,0)[b]@xmath5 ( -0.5,10)(0,0)[r]@xmath47 ( 22,10)(0,0)[l]@xmath48 ( 51.5,14)(0,0)[r]@xmath374 ( 47.5,9.5)(0,0)[r]@xmath124 ( 70,14)(0,0)[l]@xmath375 ( 73,9.5)(0,0)[l]@xmath125 ( 6.2,10)(0,0)[l]@xmath118 ( 15.8,10)(0,0)[r]@xmath119 ( 56.2,10)(0,0)[l]@xmath118 ( 65.8,10)(0,0)[r]@xmath119 ( 11,16)(0,0)[b]@xmath120 ( 61,16)(0,0)[b]@xmath126 ( 36,5)(0,0)@xmath373 ( 100,13 ) ( 1,0)(0,0)[b ] , @xmath127,title="fig : " ] ( 48,0)(0,0)[b ] , @xmath127,title="fig : " ] ( 27,4)(0,0)@xmath7 ( 65,4)(0,0)[l]@xmath376
in this section , we shall consider some mathematical methods useful for expanding massless and hqet diagrams in @xmath1 .
as you may have noticed , expansions of various feynman integrals in @xmath1 contain lots of @xmath2-functions .
they appear when expanding @xmath128-functions : @xmath377\ , .
\label{zeta : gamma}\ ] ] they also appear , together with their generalizations
multiple @xmath2 values , when expanding hypergeometric functions of unit argument with indices going to integers at @xmath0 ( sect .
[ s : exam ] , [ s : algo ] ) .
such hypergeometric functions appear in massless and hqet diagrams ( sect .
[ s : gn ] , [ s : jn ] , [ s : in ] ) . in this section
, we shall briefly discuss multiple @xmath2 values ; more details can be found in @xcite , where more general functions are also considered .
the riemann @xmath2-function is defined by @xmath378 let s define multiple @xmath2 values as @xmath379 and so on .
these series converge at @xmath380 .
the number of summations is called _ depth _ @xmath12 ; @xmath381 is called _ weight_. all identities we shall consider relate terms of the same weight ( where the weight of a product is the sum of the factors weights ) .
for example , there is 1 convergent sum at weight 2 : @xmath382 2 at weight 3 : @xmath383 4 at weight 4 : @xmath384 and 8 at weight 5 : @xmath385 suppose we want to multiply @xmath386 : @xmath387 here @xmath30 can be anywhere with respect to @xmath4 , @xmath5 .
there are 5 contributions : @xmath388 therefore , @xmath389 this process reminds shuffling cards ( fig .
[ f : stuffling ] ) .
the order of cards in the upper deck , as well as in the lower one , is kept fixed .
we sum over all possible shufflings . unlike real playing cards
, however , two cards may be exactly on top of each other . in this case
, they are stuffed together : a single card ( which is their sum ) appears in the resulting deck .
a mathematical jargon term for such shuffling with ( possible ) stuffing is _ stuffling _ , see @xcite .
( 163,122 ) ( 81.5,61)(0,0 ) for example , the stuffling relations for weights @xmath390 are @xmath391 now we are going to derive an integral representation of multiple @xmath2 values .
let s consider the integral @xmath392 it is easy to guess that @xmath393 but we need the sum ( [ zeta : zeta1 ] ) .
this is also easy : @xmath394 let s introduce short notations : @xmath395 all integrals will always have the integration region @xmath396 . then ( [ zeta : int ] ) becomes @xmath397 this can be generalized to multiple sums : @xmath398 and so on .
suppose we want to multiply @xmath399 : @xmath400 the ordering of primed and non - primed integration variables is not fixed .
there are 6 contributions : @xmath401 therefore , @xmath402 now we are multiplying integrals , not sums .
therefore , our cards are now infinitely thin , and can not be exactly on top of each other .
there are just 2 kinds of cards : @xmath403 and @xmath404 , and we sum over all possible shufflings of two decks ( fig .
[ f : shuffling ] ) .
( 163,122 ) ( 81.5,61)(0,0 ) for example , the shuffling relations for weight 5 are @xmath405 the integral representation allows us to derive another set of useful relations , even simpler than shuffling duality relations .
let s make the substitution @xmath406 .
then @xmath407 ; in order to preserve the ordering @xmath408 , we have to arrange all the @xmath227 factors in the opposite order . in other words , after writing down an integral representation for a multiple @xmath2 value , we may read it in the arabic fashion , from right to left , simultaneously replacing @xmath407 : @xmath409 duality relations are the only known relations which say that two distinct multiple @xmath2 values are just equal to each other .
let s summarize .
there are 3 distinct multiple @xmath2 values of weight 4 , @xmath410 due to duality ( [ zeta : duality ] ) .
the stuffling ( [ zeta : stuffle5 ] ) and shuffling ( [ zeta : shuffling2 ] ) relations yield @xmath411 \zeta_2 ^ 2 = 4 \zeta_{31 } + 2 \zeta_{22 } & \to & \displaystyle \zeta_{31 } = \frac{1}{4 } \zeta_4\ , , \end{array}\ ] ] where we have used @xmath412 this follows from @xmath413 therefore , all multiple @xmath2 values of weight 4 can be expressed via @xmath414 .
there are 4 distinct multiple @xmath2 values of weight 5 , @xmath415 due to duality ( [ zeta : duality ] ) .
the stuffling ( [ zeta : stuffle5 ] ) and shuffling ( [ zeta : shuffling3 ] ) relations yield @xmath416 solving this linear system , we obtain @xmath417 therefore , all multiple @xmath2 values of weight 5 can be expressed via @xmath418 and @xmath419 .
weight 5 is sufficient for calculating finite parts of 3-loop propagator diagrams .
we shall not discuss higher weights here .
there is an algorithm to expand in @xmath1 hypergeometric functions of unit argument whose indices tend to integers at @xmath0 .
the coefficients are linear combinations of multiple @xmath2 values .
we shall first consider an example : expanding the @xmath137 function ( [ gn : fe ] ) , which appears in @xmath154 ( sect .
[ s : gn ] ) , up to @xmath170 .
the function we want to expand is @xmath420 when arguments of two pochhammer symbols differ by an integer , their ratio can be simplified : @xmath421 * step 1*. we rewrite all pochhammer symbols @xmath422 via @xmath423 : @xmath424 then our function @xmath425 ( [ exam : f ] ) becomes @xmath426 in what follows , we shall consider @xmath427 , and return to @xmath425 at the end of calculation .
it is convenient to introduce the function @xmath428 : @xmath429 then @xmath430 * step 2*. we expand the rational function in front of @xmath431 s in @xmath1 : @xmath432 in this simple case , all @xmath30-dependent brackets are the same ( @xmath433 ) ; in more general cases , we should have to decompose each term into partial fractions with respect to @xmath30 . then @xmath434 because we may replace all @xmath435 in the @xmath170 term .
shifting the summation indices , we get @xmath436 * step 3*. we rewrite all @xmath437 via @xmath438 : @xmath439 we again expand in @xmath1 ( and , if necessary , decompose into partial fractions ) rational functions in front of @xmath431 s : @xmath440 * step 4*. we add and subtract a few first terms to have all summations start from @xmath320 : @xmath441 * step 5*. we expand @xmath438 in @xmath1 : @xmath442 where @xmath443 and so on .
these finite @xmath444 sums obey the same stuffling relations as infinite @xmath2 sums , for example , @xmath445 therefore , coefficients of expansion of ratios of products of @xmath431 functions can always be made linear in @xmath444 sums . in our simple case , @xmath446 therefore , @xmath447 recalling the duality relation ( [ zeta : duality ] ) @xmath448 , we obtain for our function @xmath425 ( [ exam : f ] ) @xmath449\\ & { } = 4 ( \zeta_2 - 1 ) + 6 ( -4\zeta_3 + 3\zeta_2 - 1 ) \varepsilon + \cdots\end{aligned}\ ] ] thus we have reproduced the first two terms of ( [ gn : fe ] ) .
now we shall formulate the algorithm in a general setting .
we want to expand @xmath450 in @xmath1 . * step 1*. we rewrite all pochhammer symbols @xmath422 via @xmath423 : @xmath451 introducing the function @xmath431 ( [ exam : p ] ) , we rewrite @xmath425 ( [ algo : f ] ) as @xmath452 where @xmath453 is a rational function of @xmath1 and @xmath30 .
* step 2*. we expand @xmath454 and decompose each @xmath455 into partial fractions with respect to @xmath30 .
there is a catch here : partial - fractioning of a convergent series can split it into a combination of logarithmically divergent ones . therefore , it is necessary to keep the upper limit of summation finite , and go to the limit only after combining such series together . because of shifts of the summation variable , cancellation of divergent series leaves us with a few terms around the upper limit .
they , however , tend to zero , and this means that we may formally manipulate logarithmically divergent series as if they were convergent .
shifting summation indices , we can write @xmath425 ( [ algo : f ] ) as a sum of terms of the form @xmath456 * step 3*. we rewrite all @xmath437 via @xmath438 : @xmath457 our @xmath425 becomes a sum of terms of the form @xmath458 where @xmath453 is a rational function of @xmath1 and @xmath30 . moving from terms with lower powers of @xmath1 to ones with higher powers ,
we expand @xmath453 s in @xmath1 : @xmath459 terms of higher orders in @xmath1 are decomposed in partial fractions again .
finally , @xmath425 becomes a sum of terms of the form @xmath460 * step 4*. adding and subtracting terms with @xmath30 from 1 to @xmath461 , we rewrite @xmath425 as a sum of terms of the form @xmath462 and rational functions of @xmath1 ( which can be trivially expanded in @xmath1 ) . *
step 5*. we expand each @xmath438 as ( [ exam : pe ] ) .
expansions of ratios of products of @xmath431 functions contain products of @xmath444 sums ; they are reduced to linear combinations of @xmath444 sums by stuffling relations ( e.g. , ( [ exam : stuffling ] ) ) . using @xmath463
we can express coefficients of @xmath1 expansion of @xmath425 as linear combinations of multiple @xmath2 values ( discussed in sect .
[ s : zeta ] ) and rational numbers .
_ q. e. d. _ a few historical comments . in 2000
, i needed to expand some @xmath137 functions of unit argument in @xmath1 ( in fact , they were ( [ jn : f ] ) and ( [ in : f ] ) ) .
i asked david broadhurst how to do this , and he replied : just expand pochhammer symbols in @xmath1 , the coefficients will be expressible via multiple @xmath2 values . following this advise , i implemented the algorithm described above in , and obtained the expansions ( [ jn : f ] ) and ( [ in : f ] ) .
presentation in sects .
[ s : exam ] and [ s : algo ] closely follows my notes from 2000 .
later this algorithm was described as algorithm a in @xcite ( this paper also contains more difficult algorithms ( b , c , d ) to expand some more complicated sums in @xmath1 ) .
all of these algorithms are implemented in the library @xcite values .
we follow @xcite here ; in @xcite the order of indices is reversed . ] .
it is based on the computer algebra library @xcite . in order to use with recent versions of , you should download a recent version from the url shown in @xcite .
unfortunately , no general algorithm is known to expand in @xmath1 hypergeometric functions of unit argument some of whose indices tend to half - integers at @xmath0 .
such functions appear in on - shell propagator calculations , see sects .
[ s : m2 ] , [ s : m2 m ] .
known expansions show that many constants more complicated than multiple @xmath2 values appear .
i am grateful to v.m .
braun for providing details of the calculation @xcite of @xmath265 ; to d.j .
broadhurst , k.g .
chetyrkin , a. czarnecki , a.i .
davydychev for collaboration on various multiloop projects ; to a.v .
kotikov for the discussion of @xcite ; to the organizers of calc-03 in dubna for inviting me to give these lectures and for organizing this excellent school .
|
i discuss methods of calculation of propagator diagrams ( massless , those of heavy quark effective theory , and massive on - shell diagrams ) up to 3 loops .
integration - by - parts recurrence relations are used to reduce them to linear combinations of basis integrals .
non - trivial basis integrals have to be calculated by some other method , e.g. , using gegenbauer polynomial technique .
many of them are expressed via hypergeometric functions ; in the massless and hqet cases , their indices tend to integers at @xmath0 .
i discuss the algorithm of their expansion in @xmath1 , in terms of multiple @xmath2 values .
these lectures were given at calc-03 school , dubna , 1420 june 2003 .
ttp03 - 19 lectures on multiloop calculations + andrey grozin + institut fr theoretische teilchenphysik , universitt karlsruhe + [email protected]
|
neutrino was proposed by w. pauli in 1930 to explain the continuous @xmath7 spectrum that had been observed .
however , its detection was not realized until 26 years later : in 1956 neutrino was first detected in the savannah river experiment led by f. reines and c. cowan .
today we know that there are three types or flavors of active neutrinos ( @xmath8 , @xmath9 and @xmath10 ) , and we also know neutrinos have masses ( with mass eigenstates @xmath11 , @xmath12 and @xmath13 )
. nevertheless , the neutrino flavor eigenstates are not their mass eigenstates . from the quantum mechanics , we know that neutrinos oscillate : the flavor components of the neutrinos will change as they propagate .
the neutrino flavor and mass eigenstates can be bridged as @xmath14 where @xmath15 and @xmath16 represent neutrino flavor and mass eigenstates , respectively .
u is a 3 @xmath4 3 unitary matrix called pontecorvo - maki - nakagawa - sakata ( pmns ) matrix that has the form of @xmath17 the pmns matrix contains three neutrino mixing angles ( @xmath18 , @xmath19 and @xmath20 ) and one phase ( @xmath21 ) , and the @xmath21 could lead to the cp violation . for the daya bay experiment , we focused on the measurement of the @xmath19 neutrino mixing angle using electron anti - neutrinos ( @xmath0 ) produced from reactors .
the @xmath19 can be extracted by measuring the @xmath0 survial probability after they propagate for certain distance , which has the form of @xmath22 @xmath23 where @xmath24 is the neutrino energy , @xmath25 is the neutrino propagation distance , and @xmath26 = @xmath27 - @xmath28 @xmath29 .
reactors mainly produce electron anti - neutrinos through the fissions of four isotopes ( @xmath30 , @xmath31 , @xmath32 and @xmath33 ) . on average ,
each fission can produce @xmath34 6 electron anti - neutrinos and release @xmath34 200 mev thermal energy .
hence for a 1 gigawatts ( gw ) thermal power reactor , it produces @xmath342 @xmath4 10@xmath35 electron anti - neutrinos per second with most neutrino energies below 8 mev . the detection of reactor anti - neutrinos is through the inverse beta decay ( ibd ) :
@xmath0 + p @xmath36 e@xmath37 + n that requires the @xmath0 s energy larger than 1.8 mev .
the ionization and annihilation of e@xmath37 in liquid scintillator forms a prompt signal , while the neutron will first get thermalized and then be captured by some nucleus and form a delayed signal . in the daya bay experiment , gd is used to capture ibd neutrons . when a neutron is captured by gd
, it will release several gamma rays with total energy @xmath34 8 mev .
the mean time difference between prompt and delayed signals is @xmath34 30 @xmath38s .
the coincidence of the prompt and the delayed signals forms a distinctive signature for ibd . in march , 2012 , daya bay discovered a non - zero @xmath1 with 5.2 @xmath39 significance using 55 days of data@xcite and provided the most precise measurement of this parameter at that time . with 217 days of data ,
daya bay updated the results in 2014 with a relative @xmath0 rate and spectrum analysis to measure both the @xmath1 and the oscillation frequency @xmath40 @xcite . in this article , we report the most recent results from daya bay using 621 days of data@xcite , which has 3.6 more times of statistics than our previous published results@xcite .
the daya bay neutrino experiment is located inside the daya bay nuclear power plant facility that is in the suburb of shenzhen , guangdong province , china .
the daya bay experiment detects the electron anti - neutrinos produced from six 2.9 gw thermal power commericial reactors , daya bay nuclear power plants ( two reactors ) , ling ao i nuclear power plants ( two reactors ) and ling ao ii nuclear power plants ( two reactors ) , using eight functional identical anti - neutrino detectors ( ads ) .
the ads are placed in three underground experimental halls , two near halls ( eh1 and eh2 ) and one far hall ( eh3 ) .
eh1 , which hosts two ads , is 363 m away from daya bay nuclear power plants and has an overburden of 250 meter water equivalent ( m.w.e . ) .
eh2 , with 265 m.w.e .
overburden , also has two ads and is about 500 m from ling ao i and ii nuclear power plants .
the other four ads are located in eh3 , which has an overburden of 860 m.w.e . and is 1615 and 1958 m from ling ao i and daya bay nuclear power plants , respectively .
the ad used in the daya bay experiment is a 3-zone cylindrical shaped detector .
the inner most part is a 3 @xmath4 3 m dimension inner acrylic vessel ( iav ) that holds 20 tons of gd doped liquid scintilltor ( gd - ls ) which is used as the detection target .
the iav is surrounded by 20 tons of none doped liquid scintillator ( ls ) that is used to capture the @xmath41 rays leaked from gd - ls region .
the ls is holded by a 4 @xmath4 4 m outer acrylic vessel ( oav ) .
the outmost region is a 5 @xmath4 5 m stainless steel vessel ( ssv ) which contains 40 tons of mineral oil ( mo ) to shield the gd - ls and ls regions from the radioactivities produced by pmt glasses and the stainless steel .
each ad contains 192 8-inch pmts that mounted on the side wall of ssv .
two optical reflectors are placed at the top and the bottom of ssv to increase the effective detector photo coverage to 12% .
160 photoelectrons are collected per mev on average which leads to the detector energy resolution of @xmath348%/@xmath42 .
three automated calibration units ( acus ) are mounted on the top of ssv lid which are used to calibrate the detector gd - ls center and edge as well as ls regions along three vertical axes .
each acu contains a led diffuser ball that is used to calibrate the pmt gain and timing , and three calibration sources : @xmath43ge , @xmath44co and @xmath45am@xmath46c for @xmath41 rays and neutron .
the calibration is performed weekly to all 8 ads simultaneously . besides weekly calibration using acu sources , during the summer of 2012
a special calibration campaign was performed on eh1 s two ads using @xmath47cs , @xmath48mn , @xmath49k , @xmath45am@xmath50be and @xmath51pu@xmath46c sources .
in addition , a manual calibration system ( mcs ) was installed on eh1-ad1 during this period , which allows for the full volume calibration of the detector . also , @xmath41 and @xmath52 peaks that from the radioactivities of the ls ( @xmath49k,@xmath53tl , @xmath54po , @xmath55po and @xmath56po ) and neutron captured on h , c and fe are included in the calibration .
overall the relative energy scale , the difference of the reconstructed energies between ads , is less than 0.2% for all sources .
the ads in each experimental hall are submerged in a purified water pool with at least 2.5 m of water coverage in all directions .
the water pool is optically separated to 2-zone : inner water pool ( iwp ) and outer water pool ( owp ) , by a layer of tyvek .
the water pool is used as a cherenkov detetor to veto cosmic muons as well as to shield the ads from the radioactivities that come from the surrounding rock .
each water pool is covered with 4-layer resistive plate chambers ( rpcs ) that can provide additional muon tagging information .
the detector energy nonlinearity , or relative scintillator response to different energies of different types of particles ( such as e@xmath37 , e@xmath57 and @xmath41 ) , is important to the interpretation of the observed prompt energy spectra .
there are two major sources that cause the detector nonlinearity response between the true energy and the reconstructed energies of a particle : the scintillator nonlinerity and the pmt read out electronics nonlinearity .
the scintillator nonlinearity is due to the scintillator quenching effect , which can be described by birk s law , and the cherenkov light emission .
the electronic nonlinearity is due to the interaction of the scintillation light time profile and the charge measurement scheme of the front - end electronics , which is modeled with monte carlo and measurement from single channel fadc in the daya bay experiment .
the detector nonlinearity is correlated among all ads , and is constrained to be @xmath34 1% in the daya bay experiment . for more information about daya bay experiment and its detector system , please check @xcite .
the daya bay experiment started to take data from december 2011 with 6-ad configuration , two ads at eh1 , one ad at eh2 and three ads at eh3 , and this period last for 217 days until july 2012 .
in the summer of 2012 , two new ads were installed : one at eh2 and the other at eh3 . from then on
, the daya bay experiment is running with full 8-ad configuration . in the following sections
, we will present the recent results from daya bay using 621 days of data : 217 days data with 6-ad configuration and 404 days data with 8-ad configuration .
to remove the irrelevant events and select the real ibd events that caused by the neutrinos , several cuts were applied to our ibd candidates .
first of all , we removed the events that caused by pmt flashers : the light spontaneously emitted from pmt .
then we required the prompt and delayed energies of the ibd candidates to fall into the ranges of 0.7 - 12 mev and 6 - 12 mev , respectively .
the prompt and delayed signals were also required to form a coincidence within a time window of 1 - 200 @xmath38s . to remove the events that caused by cosmic muons , muon veto cuts were applied .
events were removed if the delayed signal occurs within 600 @xmath38s after a water pool muon was detected , or the delayed signal happens within 1 ms after a reconstructed signal with energy @xmath58 20 mev in ad , or the delayed signal within 1 s after a reconstructed signal with energy @xmath58 2.5 gev in ad . finally , to further remove ambiguity ,
a multiplicity cut is applied to select only a single candidate pairs within each cut window .
the cut efficiency for the above mentioned cuts is 80.6% @xmath2 2.1% @xmath2 0.2% , where the 2.1% and 0.2% are corresponding to the correlated and uncorrelated uncertainties . during the 621 days of data taking
, we totally collected more than 1 million ibd events with @xmath34150 k events at far ads . the time dependent ibd rates are strongly correlated with the reactor @xmath0 flux expectation .
the daya bay experiment is a low background experiment .
there are five major backgrounds in the daya bay experiment that could mimic the ibd .
the largest background is accidental background , which accounts for 2.3% and 1.4% of the candidates in the far and near halls , respectively .
however , the accidental background can be statistically removed with high precision .
the second and third backgrouns are from @xmath50li/@xmath59he and fast neutrons that are generated from the cosmic muons account for 0.4% and 0.1% at both halls , respectively .
the fourth background is due to the natural radioactivities within the liquid scintillator , which accounts for 0.01% and 0.1% at near and far halls .
the last background is due to our @xmath45am@xmath46c calibration source , which is largely reduced during the 8-ad configuration period due to the removal of off - center @xmath45am@xmath46c sources at eh3 during the summer of 2012 .
the total background amount to 2% and 3% of the ibd candidates for the near and far halls .
in this section , we present the recent results from daya bay neutrino experiment that include the update of the measurement of @xmath1 and @xmath3 values from the oscillation analysis using 621 days of data@xcite , the light sterile neutrino searching results in the 3 + 1 neutrino framework@xcite and the absolute reactor @xmath0 flux and positron spectrum measurement@xcite using 217 days of data .
the oscillation analysis is performed by a relative measurement of the @xmath0 rate and spectrum between near and far ads using 621 days of data@xcite , which has 3.6 more times of the statistics than our previous published results@xcite .
the left plot of figure [ fig : osc_rate ] shows the ratio of the detected to expected @xmath0 events that assuming no oscillation versus the effective baseline for each ad .
the right plot , on top panel , of figure [ fig : osc_rate ] shows the comparision of the prompt spectra measured at far ads with the expected spectrum at far ads based on the measured spectrum at near ads with and without oscillation interpretations . while the bottom panel plot shows the ratio of the measured prompt spectrum at far ads to the weighted near ads spectrum without oscillation . from figure [ fig : osc_rate ] we can see that both the observed relative rate deficit and the relative spectrum distortion are highly consistent with the oscillation interpretation .
events that assuming no oscillation versus the effective baseline for each ad .
the open dots represent the ratio for each ad , and the curve shows the best fit .
two far ad dots are displaced @xmath2 50 m respectively on purpose to give a better view .
right : the top panel plot shows the measured prompt spectrum with background subtracted ( solid dots ) at far ads compared to the expected prompt spectra based on the near ads data with ( lower curve ) and without ( higher curve ) oscillation interpretation .
the bottom panel plot shows the ratio of the measured prompt spectrum at far ads to the weighted near ads spectrum without oscillation , and the curve shows the best fit . ] the left plot of figure [ fig : osc_fit ] shows the best fit values of sin@xmath60 and @xmath3 from our oscillation analysis , which yields @xmath1 = 0.084 @xmath2 0.005 and @xmath3 = ( 2.42 @xmath2 0.11 ) @xmath4 10@xmath5 ev@xmath6 .
the precision of the sin@xmath60 reaches 6% and is currently the most precise measurement of this parameter in the world .
the @xmath3 measurement , with a hight precision of @xmath34 5% , is consistent with the muon neutrino disappearance measurement of @xmath61 by minos@xcite and t2k@xcite experiments , converted to @xmath3
. more information of this daya bay oscillation analysis with 621 days of data can be found in reference @xcite .
daya bay also performed an independent relative measurement of @xmath0 rate between near and far ads of @xmath1 by detecting the ibd neutron captured on h ( nh rate analysis ) .
this analysis uses 217 days of data that with 6-ad configuration , and yields the @xmath1 value of 0.083 @xmath2 0.018 .
the @xmath1 from nh rate analysis is also a very precise measurement of this parameter and it is consistent with the result from the relative rate and spectrum measurement of ibd neutron captured on gd@xcite .
more information of the nh rate analysis with 217 days of data can be found in reference @xcite . and
@xmath3 , which gives @xmath1 = 0.084 @xmath2 0.005 and @xmath3 = 2.42 @xmath2 0.11 @xmath4 10@xmath5 ev@xmath6 .
the three bands surrounding the best fit value ( from inner to outer ) are corresponding to the allowed regions with 68.3% , 95.5% and 99.7% confidence level ( c.l . ) .
the adjoint panels show the dependence of @xmath62 on sin@xmath60 ( top ) and @xmath3 ( right ) .
the band on the right adjoint panel shows the 1 @xmath39 variation of @xmath3 .
the solid and dashed dots are the converted @xmath3 from @xmath61 , from minos@xcite and t2k@xcite experiments , assuming normal and inverted mass hierarchy , respectively .
right : the 90% c.l .
exclusion contours extracted from daya bay using the cls method ( solid ) and the feldman - cousin method ( long - dashed ) , the top short - dashed contour represents bugey-3 s 90% c.l .
exclusion contour using raster scan method@xcite .
, title="fig : " ] and @xmath3 , which gives @xmath1 = 0.084 @xmath2 0.005 and @xmath3 = 2.42 @xmath2 0.11 @xmath4 10@xmath5 ev@xmath6 .
the three bands surrounding the best fit value ( from inner to outer ) are corresponding to the allowed regions with 68.3% , 95.5% and 99.7% confidence level ( c.l . ) .
the adjoint panels show the dependence of @xmath62 on sin@xmath60 ( top ) and @xmath3 ( right ) .
the band on the right adjoint panel shows the 1 @xmath39 variation of @xmath3 .
the solid and dashed dots are the converted @xmath3 from @xmath61 , from minos@xcite and t2k@xcite experiments , assuming normal and inverted mass hierarchy , respectively .
right : the 90% c.l .
exclusion contours extracted from daya bay using the cls method ( solid ) and the feldman - cousin method ( long - dashed ) , the top short - dashed contour represents bugey-3 s 90% c.l .
exclusion contour using raster scan method@xcite .
, title="fig : " ] the oscillation analysis is performed in the 3 neutrino framework , what if there exists a fourth light sterile neutrino ( with mass of m@xmath63 ) , will this affect our oscillation results ? to answer these questions , daya bay also performed the light sterile neutrino search in the 3 ( active ) + 1 ( sterile ) neutrino framework using 217 days of data@xcite .
no significant signal had been observed for the existence of a light sterile neutrino .
however , the daya bay is able to set the most stringent limit on @xmath64 in the @xmath65 region from 10@xmath5 to 10@xmath66 ev@xmath6 .
the right plot of figure [ fig : osc_fit ] shows the 90% c.l .
exclusion contours extracted from daya bay using the cls method ( solid ) and the feldman - cousin method ( long - dashed ) , while the top short - dashed contour represents bugey-3 s 90% c.l .
exclusion contour using raster scan method@xcite .
the absolute reactor @xmath0 flux and spectrum measurements were also performed at daya bay . with 217 days of data , a total of 300k and 40k ibd candidates are detected at the near and far ads , respectively .
the measured @xmath0 flux is ( 1.55 @xmath2 0.04 ) @xmath4 10@xmath67 cm@xmath6/gw / day or ( 5.92 @xmath2 0.14 ) @xmath4 10@xmath69 cm@xmath6/fission ( see the top panel plot on the left side of figure [ fig : absolute ] ) , and it is consistent with previous short baseline reactor @xmath0 mesurements .
the ratio of the measured @xmath0 flux to the predcited flux from huber+mueller reactor @xmath0 models@xcite is 0.946 @xmath2 0.022 , which is consistent with the world average ratio of 0.942 @xmath2 0.009 ( see the bottom panel plot on the left side of figure [ fig : absolute ] ) .
the plot at the top panel on the right side of figure [ fig : absolute ] shows the comparison of the absolute reactor spectrum measured to the prediction with huber+mueller reactor @xmath0 model . for the whole energy range
, there is a 2.6 @xmath39 discrepancy between the measured and predicted spectra .
an excess structure has been observed in the prompt energy range from 4 to 6 mev , the discrepancy in this region yields a 4 @xmath39 ( see the plot at the bottom panel on the right side of figure [ fig : absolute ] ) . from the right plots of figure [ fig : absolute ] , we can see that the measured absolute reactor @xmath0 spectrum is not consistent with the prediction of all reactor @xmath0 flux models . however , this discrepancy has no impact on the sin@xmath60 and @xmath3 values , since they are extracted from the relative measurement of rate and spectrum between near and far ads . candidate events in the 6 ads with corrections for 3-flavor oscillations ( closed circles ) , and additionally for the variation of flux - weighted fission fractions at the different experimental halls ( open squares ) .
the average of the three near ads is shown as the straight line ( extended to the far ads ) with 1 systematic uncertainty band .
the rate predicted with the huber+mueller@xcite ( ill+vogel@xcite ) models and its uncertainty are shown at the right .
( bottom panel ) the measured reactor @xmath0 rate as a function of the distance from the reactor , normalized to the theoretical prediction with the huber+mueller model .
the rate is corrected for 3-flavor neutrino oscillations at each baseline .
the lower shaded region represents the global average and its 1 @xmath39 uncertainty .
the 2.7% model uncertainty is shown as a band around unity .
measurements at the same baseline are combined for clarity .
the daya bay measurement is shown at the flux - weighted baseline ( 573 m ) of the two near halls .
right : top panel : predicted and measured prompt - energy spectra .
the prediction is based on the huber+mueller model and normalized to the number of measured events .
the highest energy bin contains all events above 7 mev .
the error bars on the data points represent the statistical uncertainty .
bottom panel : ratio of the measured prompt - energy spectrum to the predicted spectrum ( huber+mueller model ) .
the curve shows the ratio of the prediction based on the ill+vogel model to that based on the huber+mueller model.,title="fig : " ] candidate events in the 6 ads with corrections for 3-flavor oscillations ( closed circles ) , and additionally for the variation of flux - weighted fission fractions at the different experimental halls ( open squares ) .
the average of the three near ads is shown as the straight line ( extended to the far ads ) with 1 systematic uncertainty band .
the rate predicted with the huber+mueller@xcite ( ill+vogel@xcite ) models and its uncertainty are shown at the right .
( bottom panel ) the measured reactor @xmath0 rate as a function of the distance from the reactor , normalized to the theoretical prediction with the huber+mueller model .
the rate is corrected for 3-flavor neutrino oscillations at each baseline .
the lower shaded region represents the global average and its 1 @xmath39 uncertainty .
the 2.7% model uncertainty is shown as a band around unity .
measurements at the same baseline are combined for clarity .
the daya bay measurement is shown at the flux - weighted baseline ( 573 m ) of the two near halls .
right : top panel : predicted and measured prompt - energy spectra .
the prediction is based on the huber+mueller model and normalized to the number of measured events .
the highest energy bin contains all events above 7 mev .
the error bars on the data points represent the statistical uncertainty .
bottom panel : ratio of the measured prompt - energy spectrum to the predicted spectrum ( huber+mueller model ) .
the curve shows the ratio of the prediction based on the ill+vogel model to that based on the huber+mueller model.,title="fig : " ]
the daya bay neutrino experiment has measured @xmath1 = 0.084 @xmath2 0.005 and @xmath40 = 2.42 @xmath2 0.11 @xmath4 10@xmath5 ev@xmath6 using 621 days of data .
this measurement is currently the most precise measurement of @xmath1 in the world .
the @xmath40 measurement also has a comparable precision and is consistent with the muon disappearance experiment results from minos and t2k experiments .
the light sterile neutrino searching analysis is able to set the most stringent limit on @xmath64 in the @xmath70 region from 10@xmath71 to 10@xmath66 ev@xmath6 .
the absolute reactor @xmath0 flux measurement is consistent with previous short - baseline reactor neutrino experiments .
while the absolute reactor @xmath0 spectrum is not consistent with predictions from all models .
daya bay will keep running until the end of 2017 , by that time , both @xmath1 and @xmath40 precisions are expected to reach 3% .
daya bay is supported in part by the ministry of science and technology of china , the u.s .
department of energy , the chinese academy of sciences , the national natural science foundation of china , the guangdong provincial gov- ernment , the shenzhen municipal government , the china general nuclear power group , key laboratory of particle and radiation imaging ( tsinghua university ) , the ministry of education , key laboratory of particle physics and particle irradiation ( shandong university ) , the ministry of education , shanghai laboratory for particle physics and cosmology , the research grants council of the hong kong special administrative region of china , the university development fund of the university of hong kong , the moe program for research of excellence at national taiwan university , national chiao - tung university , and nsc fund support from taiwan , the u.s .
national science founda- tion , the alfred p. sloan foundation , the ministry of education , youth , and sports of the czech republic , the joint institute of nuclear research in dubna , russia , the cnfc - rfbr joint research program , the national commission of scientific and technological research of chile , and the tsinghua university initiative scientific research program .
we acknowledge yellow river engineering consulting co. , ltd . , and china railway 15th bureau group co. , ltd . , for building the underground laboratory .
we are grateful for the ongoing cooperation from the china general nuclear power group and china light and power company .
|
the daya bay neutrino experiment has recently updated the oscillation analysis results with 621 days of data in 2015 , which has 3.6 times more statistics than the previous publication in 2014 .
the relative @xmath0 rate and spectrum measurement between the near and far detectors yielded the best fit values of @xmath1 = 0.084 @xmath2 0.005 and @xmath3 = ( 2.42 @xmath2 0.11 ) @xmath4 10@xmath5 ev@xmath6 .
this is currently the most precise measurement of @xmath1 in the world .
the measurement of @xmath3 also has a precision that is comparable to the measurements from minos and t2k experiments in 2014 .
daya bay also performed several other analyses such as the search for the light sterile neutrino in the 3 + 1 neutrino framework , and the measurements of the absolute reactor anti - neutrino flux and spectrum .
|
in recent years , there has been increasing interest of the possibility that evolution of species in an ecosystem may be a self organized critical phenomena . there are interactions among species in that ecosystem .
the most common such interactions are predation , competition for resources and mutualism . as a result of these interactions
the evolutionary adaptation of one species must affect its nearest neighbors .
these interactions can give rise to large evolutionary disturbances , termed _
most of these evolution models , like bs model [ 1 ] , considered only one fitness for each species .
bak and sneppen proposed a self organized model to explain the punctuated equilibrium of biological evolution .
they considered a 1- dimensional model with periodic boundary conditions , topologically a circle .
assign a fitness @xmath0 to each site @xmath1 , @xmath2 , where @xmath3 , is the number of species in the ecosystem . at each time step
look for the site with lowest fitness @xmath4 then replace its fitness together with the fitnesses of its nearest neighbors , @xmath5 by new ones which are uniformly distributed random variables .
@xmath6 after running the system for sufficiently long time most of the fitness are above certain threshold @xmath7 . also , the distribution of the distance between subsequent mutations and the avalanche sizes exhibit power laws .
several modification can be done to the bs model .
the first possible modification is to use extremal dynamics [ 2 , 3 ] which depends on the following idea : in real biological systems not only the lowest one who is updated but some of the low fitness species .
this number that changes is not fixed but random .
so , we will study this random version of bs model in section 2 .
also , in biology almost every optimization problem is multiobjective ( mob ) e.g. objective of foraging and of minimizing predation risk . in section 3
, we will apply the concept of mob to bs model .
here , we will study the first modification that can be done to the bs model . instead of finding exactly the site with lowest fitness , one may use the extremal dynamics . in this case a uniformly distributed random number is picked and all the sites with fitness less than this number has its fitness updated .
this dynamics has been used to explain the long term memory for the immune system [ 3 ] .
it has been also used to solve some optimization problems e.g. spin glass , graph coloring and graph partitioning [ 4 , 5 ] .
we run a system consisting of @xmath8 species for different sufficiently long time ( up to @xmath9 ) .
we find that most of the fitnesses are above a certain threshold value @xmath10 , as shown in figure 2 . in figure 1
we plotted the standard bak - sneppen model for reference .
in most evolution models e.g. bs model , only one fitness is considered i.e. single objective optimization .
almost every real life problem is multiobjective ( mob ) one [ 6 ] .
therefore it is important to generalize the standard single goal oligopoly studies to multiobjective ones .
methods for mob optimization are mostly intuitive .
the * first method * is lexicographic method . in this method objectives
are ordered according to their importance .
then the first objective is satisfied fully .
the second one is satisfied as much as possible given that the first objective has already been satisfied and so on .
a famous application is in university admittance where students with highest grades are allowed in any college they choose .
the second best group are allowed only the remaining places and so on .
this method is useful but in some cases it is not applicable .
the * second method * is the method of weights [ 7 ] .
assume that it is required to minimize the objectives @xmath11 , @xmath2 .
the problem of maximization is obtained via replacing @xmath11 by @xmath12 .
define @xmath13 where @xmath14 then the problem becomes to minimize @xmath15 .
this method is easy to implement but it has several weaknesses .
the first is that it may give a pareto dominated solution .
a solution @xmath16 is pareto dominated if there is another solution @xmath17 such that @xmath18 for all @xmath1 with at least one @xmath19 such that @xmath20 .
the second difficulty of this method is that it is difficult to apply for large @xmath3 .
the * third method * is to minimize only one objective while setting the other objectives as constraints e.g. minimize @xmath21 subject to @xmath22 , @xmath23 where @xmath24 are parameters to be updated .
the problem with this method is the choice of the thresholds @xmath24 . in the case of equality
i.e. @xmath25 this method is guaranteed to give a pareto optimal solution . the * fourth method * using fuzzy logic is to study each objective individually and find its maximum and minimum say @xmath26 , @xmath27 respectively . then determine a membership @xmath28 thus @xmath29 . then
apply @xmath30 . again
this method is guaranteed to give a pareto optimal solution .
this method is a bit difficult to apply for large number of objectives .
the bs model can be generalized to the multiobjective . assigning two fitnesses @xmath31 , @xmath32 , to each site instead of one .
the updating rule is if @xmath33 where @xmath34 , then update both @xmath31 , @xmath35 and @xmath36 , @xmath37 . in the updating rule
we have used the simple and widely used method , weighting method in mob .
multiobjective optimization is much more realistic than single objective ones . after running a system consisting of @xmath8 species for different sufficiently long time ( up to @xmath9 ) .
the distribution of the distance between subsequent mutations are shown in figure 3 .
we find that most of the fitnesses are above certain threshold value @xmath38 , as shown in figure 4 .
the size of avalanches are shown in figure 5 .
bak p. and sneppen k. ( 1993 ) , punctuated equilibrium and criticality in a simple model of evolution , phys .
71 , 4083 .
head d. ( 2000 ) , extremal driving as a mechanism for generating long term , j. phys .
a : math . gen .
33 , 387 . ahmed e. and hashish a.h . ( 2003 ) , on modeling of immune memory mechanisms , theor .
biosci.122 , 349 .
boettcher s. and percus a. ( 2001 ) , on extremal optimization , phys .
86 , 5211 . ahmed and el - alem m.(2002 ) , immune motivated optimization , int .
41 , 985 . zeleny m. ( 1982 ) ,
multiple criteria decision making , mcgraw hill , newyork .
zadeh l. ( 1963 ) , optimality and nonscalar performance criteria , ieee transactions .
automatic control , ac-8 , 59 . 1
. the results of bs model under our simulations in a system of size @xmath8 and @xmath39 , iteration .
( a ) the distribution of distances d(x ) between subsequent mutations x. ( b ) distribution of fitness in the critical state ( right curve ) d(f ) with the distribution of minimum fitness ( left curve ) . ( c ) distribution of avalanche sizes d(s ) in critical state .
( d ) mutation activity vs time measured as the total number of mutations ( @xmath40 and @xmath41 ) .
2 . the results of random bs model in a system of size @xmath8 and @xmath39 , iteration .
( a ) distribution of fitness in the critical state d(f ) .
( b ) distribution of minimum fitness in the critical state .
( c ) distribution of avalanche sizes d(s ) in critical state .
3 . the distribution of distances d(x ) between subsequent mutations in a system of size @xmath8 and @xmath39 , iteration with three different weights 0.3 , 0.5 , 0.9 . 4 . the distribution of fitness in the critical state ( right curve ) d(f ) with the distribution of minimum fitness ( left curve ) in a system of size @xmath8 and @xmath39 , iteration with three different weights 0.3 , 0.5 , 0.9 . 5 .
the distribution of avalanche sizes d(s ) in critical state in a system of size @xmath8 and @xmath39 , iteration with three different weights 0.3 , 0.5 , 0.9 .
|
self - organized criticality ( soc ) phenomena could have a significant effect on the dynamics of ecosystems .
the bak - sneppen ( bs ) model is a simple and robust model of biological evolution that exhibits punctuated equilibrium behavior . here
we will introduce random version of bs model .
also we generalize the single objective bs model to a multiobjective one .
* keyword * : self - organized criticality , evolution and extinction , bs model , multiobjective optimization .
|
correlation functions are some of the most widely used statistics within astrophysics ( see peebles 1980 for a extensive review ) .
they are often used to quantify the clustering of objects in the universe ( _ e.g. _ galaxies , quasars _ etc .
_ ) compared to a pure poission process .
more recently , they have also been used to measure fluctuations in the cosmic microwave background ( see szapudi et al . 2000 ) . on large scales ,
the higher order correlation functions ( 3-point and above ) can be used to test several fundamental assumptions about the universe ; for example , our hierarchical scenario for structure formation , the gaussianity of the initial conditions as well as testing various models for the biasing between the luminous and dark matter .
the reader is referred to szapudi ( 2000 ) , szapudi et al .
( 1999a , b ) and scoccimarro ( 2000 ; and references therein ) for an overview of the usefulness of @xmath0point correlation functions in constraining cosmological models . over the coming decade , several new , massive cosmological surveys will become available to the astronomical community . in this new era , the quality and quantity of data will warrant a more sophisticated analysis of the higher order correlation functions of galaxies ( and other objects ) over the largest range of scales possible .
our ability to perform such studies will be severely limited by the computational time needed to compute such functions and no - longer by the amount of data available . in this paper
, we address this computational `` bottle - neck '' by outlining a new algorithm that uses innovative computer science to accelerate the computation of @xmath0point correlation functions far beyond the naive @xmath1 scaling law ( where @xmath2 is the number of objects in the dataset and @xmath0 is the power of correlation function desired ) .
the algorithm presented here was developed as part of the `` computational astrostatistics '' collaboration ( see nichol et al . 2000 ) and is a member of a family of algorithms for a very general class of statistical computations , including nearest - neighbor methods , kernel density estimation , and clustering .
the work presented here was initially presented by gray & moore ( 2001 ) and will soon be discussed in a more substantial paper by connolly et al .
( 2001 ) . in this conference
proceeding , we provide a brief review of _ k_d - trees ( section [ kdtr ] ) , a discussion of the use of _ k_d - trees in range searches ( section [ range ] ) , an overview of the development of a fast @xmath0point correlation function code ( section [ npt ] ) as well as presenting the concept of controlled approximations in the calculation of the correlation function ( section [ apprx ] ) .
in section 6 , we provide preliminary results on the computation speed - up achieved with this algorithm and discuss future prospects for further advances in this field through the use of other tree structures .
our fast @xmath0point correlation function algorithm is built upon the _ k_d - tree data structure which was introduced by friedman et al .
k_d - tree is a way of organizing a set of datapoints in @xmath3-dimensional space in such a way that once built , whenever a query arrives requesting a list all points in a neighborhood , the query can be answered quickly without needing to scan every single point .
the root node of the _ k_d - tree owns all the data points .
each non - leaf - node has two children , defined by a splitting dimension @xmath4 and a splitting value @xmath5 .
the two children divide their parent s data points between them , with the left child owning those data points that are strictly less than the splitting value in the splitting dimension , and the right child owning the remainder of the parent s data points : @xmath6 < { { { { { n}}.{{\mbox{\scriptsize \sc splitvalue}}}}}}\mbox { and } { { { { { \mbox{\bf x}}}}_{i}}}\in { n}\\ { { { { { \mbox{\bf x}}}}_{i}}}\in { { { { { n}}.{{\mbox{\scriptsize \sc right } } } } } } & \leftrightarrow & { { { { { \mbox{\bf x}}}}_{i}}}[{{{{{n}}.{{\mbox{\scriptsize \sc splitdim } } } } } } ] \geq { { { { { n}}.{{\mbox{\scriptsize \sc splitvalue}}}}}}\mbox { and } { { { { { \mbox{\bf x}}}}_{i}}}\in { n}\end{aligned}\ ] ] as an example , some of the nodes of a _ k_d - tree are illustrated in figures 1 . _
k_d - trees are usually constructed top - down , beginning with the full set of points and then splitting in the center of the widest dimension .
this produces two child nodes , each with a distinct set of points .
this procedure is then repeated recursively on each of the two child nodes .
a node is declared to be a leaf , and is left unsplit , if the widest dimension of its bounding box is @xmath7 some threshold , @xmath8 .
a node is also left unsplit if it denotes fewer than some threshold number of points , @xmath9 .
a leaf node has no children , but instead contains a list of @xmath3-dimensional vectors : the actual datapoints contained in that leaf .
the values @xmath10 and @xmath11 would cause the largest _ k_d - tree structure because all leaf nodes would denote singleton or coincident points . in practice ,
we set @xmath8 to @xmath12 of the range of the data point components and @xmath9 to around 10 .
the tree size and construction thus cost considerably less than these bounds because in dense regions , tiny leaf nodes are able to summarize dozens of data points .
the operations needed in tree - building are computationally trivial and therefore , the overhead in constructing the tree is negligible . also , once a tree is built it can be re - used for many different analysis operations . since the introduction of _ k_d - trees , many variations of them have been proposed and used with great success in areas such as databases and computational geometry ( preparata & shamos 1985 ) .
r - trees ( guttman 1984 ) are designed for disk resident data sets and efficient incremental addition of data .
metric trees ( see uhlmann 1991 ) place hyperspheres around tree nodes , instead of axis - aligned splitting planes . in all cases ,
the algorithms we discuss in this paper could be applied equally effectively with these other structures .
for example , moore ( 2000 ) shows the use of metric trees for accelerating several clustering and pairwise comparision algorithms .
before proceeding to fast @xmath0point calculations , we will begin with a very standard _ k_d - tree search algorithm that could be used as a building block for fast 2-point computations . for simplicity of exposition we will assume the every node of the _ k_d - tree contains one extra piece of information : the bounding box of all the points it contains . call this box @xmath13 .
the implication of this is that every node must contain two new @xmath3 dimensional vectors to represent the lower and upper limits of each dimension of the bounding box .
the range search operation takes two inputs .
the first is a @xmath3-dimensional vector @xmath14 called the _ query point_. the second is a separation distance @xmath15 .
the operation returns the complete set of points in the _ k_d - tree that lie within distance @xmath15 of @xmath14 .
* @xmath16 + returns a set of points @xmath17 such that @xmath18 : = the closest distance from @xmath14 to @xmath13 . *
if @xmath19 then it is impossible that any point in @xmath20 can be within range of the query .
so simply return the empty set of points without doing any further work .
* else , if @xmath20 is a leaf node , we must iterate through all the datapoints in its leaf list . for each point , find if it is within distance @xmath15 of @xmath14 . if so , add it to the set of results .
* else , @xmath20 is not a leaf node .
then : * * let @xmath21 * * let @xmath22 * * return @xmath23 .
figure 2a shows the result of running this algorithm in two dimensions .
many large nodes are pruned from the search .
117 distance calculations were needed for performing this range search , compared with 499 that would have been needed by a naive method .
note that it is not essential that _ k_d - tree nodes have bounding boxes explicitly stored .
instead a hyper - rectangle can be passed to each recursive call of the above function and dynamically updated as the tree is searched .
range searching with a _ k_d - tree can be much faster than without if the range is small , containing only a small fraction of the total number of datapoints .
but what if the range is large ?
figure 2b shows an example in which _ k_d - trees provide little computational saving because almost all the points match the query and thus need to be visited . in general
this problem is unavoidable .
but in one special case it _ can _ be avoided
if we merely want to count the number of datapoints in a range instead of explicitly find them all .
we will add the following field to a _ k_d - tree node .
let @xmath24 be the number of points contained in node @xmath20 .
this is the first and simplest of a set of _ k_d - tree decorations we refer to as _ cached sufficient statistics _
( see moore & lee 1998 ) .
in general , we frequently stored the centroid of all points in a node and their covariance matrix .
once we have @xmath24 it is trivial to write an operation that counts the number of datapoints within some range without explicitly visiting them .
* @xmath25 + returns an integer : the number of points that are both inside the @xmath20 and also within distance @xmath15 of @xmath14 .
* let @xmath26 : = the closest distance from @xmath14 to @xmath13 . * if
@xmath19 then it is impossible that any point in @xmath20 can be within range of the query .
so simply return 0 .
* let @xmath27 : = the furthest distance from @xmath14 to @xmath13 . * if @xmath28 then every point in @xmath20 must be within range of the query .
so simply return @xmath24 .
* else , if @xmath20 is a leaf node , we must iterate through all the datapoints in its leaf list .
start a counter at zero .
for each point , find if it is within distance @xmath15 of @xmath14 .
if so , increment the counter by one .
return the count once the full list has been scanned .
* else , @xmath20 is not a leaf node .
then : * * let @xmath29 * * let @xmath30 * * return @xmath31 .
the same query that gave the poor range search performance in figure 2b gives good performance in figure 3 .
the difference is that a second type of pruning of the search is possible : if the hyperrectangle surrounding the n is either entirely outside _ or inside _ the range then we prune . ]
it is easy to see that the 2-point correlation function is simply a repeated set of range counts .
for example , given a minimum and maximum separation @xmath32 and @xmath15 we run the following algorithm : * @xmath33 + input @xmath34 is a dataset , represented as a matrix in which the @xmath3th row corresponds to the @xmath3th datapoint .
@xmath34 has @xmath2 rows and @xmath3 columns .
input @xmath20 is the root of a kdtree built from the data in @xmath34 .
output integer : the number of pairs of points @xmath35 such that @xmath36 .
* c : = 0 * for @xmath37 between @xmath38 and @xmath2 do : * * @xmath39 note that in practice we do not use two range counts at each iteration , but one slightly more complex rangecount operation _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ @xmath40 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ that directly counts the number of points whose distance from @xmath14 is between @xmath32 and @xmath15 .
the previous algorithm iterates over all datapoints , issuing a range count operation for each .
we can save further time by turning that outer iteration into an additional kd - tree search .
the new search will be a recursive procedure that takes two nodes , @xmath41 and @xmath42 , as arguments .
the goal will be to compute the number of pairs of points @xmath43 such that @xmath44 , @xmath45 , and @xmath46 .
* @xmath47 + returns an integer : the number of pairs of points @xmath43 such that @xmath44 , @xmath45 , and @xmath46 .
* let @xmath26 : = the closest distance between @xmath48 and @xmath49 . * if @xmath19 then it is impossible that any pair of points can match .
so simply return 0 .
* let @xmath27 : = the furthest distance between @xmath48 and @xmath49 . * if @xmath50 then it is again impossible that any pair of points can match .
so simply return 0 . *
if @xmath51 then all pairs of points must match .
use @xmath52 and @xmath53 to compute the number of resulting pairs @xmath54 , and return that value .
* else , if @xmath41 and @xmath42 are both leaf nodes , we must iterate through all pairs of datapoints in their leaf lists .
return the resulting ( slowly computed ) count .
* else at least one of the two nodes is a non - leaf .
pick the non - leaf with the largest number of points ( breaking ties arbitrarily ) , and call it @xmath55 . call the other node @xmath56 . then : * * let @xmath57 * * let @xmath58 * * return @xmath31 . computing a 2-point function on a dataset @xmath34
then simply consists of computing the value @xmath59 , where @xmath60 is a kd - tree built from @xmath34 , for a range of bins with minimum and maximum boundaries of @xmath32 and @xmath61 .
we note here that the 2-point correlation function , the quanity of interest is not simply @xmath62 , but @xmath63 ( the number of unique pairs of objects ) .
a further speed up can be obtained by simultaneously computing the @xmath64 over a series of bins .
we will discuss this in further detail in connolly et al .
( 2001 ) .
so far , we have discussed two operations exclusion and subsumption which remove the need to traverse the whole tree thus speeding up the computation of the correlation function .
another form of pruning is to eliminate node - node comparisons which have been performed already in the reverse order .
this can be done simply by ( virtually ) ranking the datapoints according to their position in a depth - first traversal of the tree , then recording for each node the minimum and maximum ranks of the points it owns , and pruning whenever @xmath41 s maximum rank is less than @xmath42 s minimum rank .
this is useful for all - pairs problems , but will later be seen to be _ essential _ for all - k - tuples problems .
this kind of pruning is not practical for single - tree search .
the advantages of dual - tree over single - tree are so far two fold .
first , dual - tree can be faster , and second it can exploit redundancy elimination .
but two more advantages remain .
first , we can extend the `` 2-tree for 2-point '' method up to `` n - trees for n - point '' .
second ( discussed in section [ se : approx ] ) , we can perform effective approximation with dual - trees ( or n - trees ) .
we now discuss the first of these advantages .
the @xmath0point computation is parameterized by two @xmath65 symmetric matrices : @xmath66 and @xmath67 .
we wish to compute @xmath68 where @xmath69 is zero unless the following conditions hold ( in which case it takes the value 1 ) : @xmath70 we will achieve this by calling a recursive function @xmath71 on an @xmath72-tuple of kdtree nodes @xmath73 .
this recursive function much return @xmath74 * @xmath75 + * let @xmath76 * for @xmath77 to @xmath72 do * * for @xmath78 to @xmath72 do * * * let @xmath26 : = the closest distance between @xmath79 and @xmath80 . * * * if @xmath81 then it is impossible that any @xmath72-tuple of points can match because the distance between the @xmath37th and @xmath82th points in any such @xmath72-tuple must be out of range .
so simply return 0 . *
* * let @xmath27 : = the furthest distance between @xmath79 and @xmath80 .
* * * if @xmath83 then similarly return 0 . * * * if @xmath84 then every @xmath72tuple has the property the the @xmath37th member and @xmath82th member match .
we are interested in whether this is true for all @xmath85 pairs and so the first time we are disappointed ( by discovering the above expression does not hold ) then we will update the @xmath86 flag .
thus the actual computation at this step is : + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if @xmath87 or @xmath88 then + @xmath89 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * if @xmath86 has remained true throughout the above double loop , we can be sure that every @xmath72tuple derived from the nodes in the recursive call must match , and so we can simply return @xmath90 * else , if all of @xmath91 are leaf nodes we must iterate through all @xmath72tuples of datapoints in their leaf lists .
return the resulting ( slowly computed ) count .
* else at least one of the nodes is a non leaf .
pick the non - leaf with the largest number of points ( breaking ties arbitrarily ) , and assume it has index @xmath92 . then : * * let @xmath93 * * let @xmath94 * * return @xmath31
. the full @xmath0point computation is achieved by calling @xmath71 with arguments consisting of an @xmath72tuple of copies of the root node .
we should note once again it is possible to save considerable amounts of computation by eliminating redundancy .
for example , in the 4-point statistic , the above implementation will recount each matching 4-tuple of points
@xmath95 in 24 different ways : once for each of the @xmath96 permutations of @xmath95 .
again , this excess cost can be avoided by ordering the datapoints via a depth - first tree indexing scheme and then pruning any @xmath72tuple of nodes violating that order .
but the reader should be aware of an extremely messy problem regarding how much to award to the count in the case that a subsume type of pruning can take place .
if all nodes own independent sets of points the answer is simple : the product of the node counts .
if all nodes are the same then the answer is again simple : @xmath97 , where @xmath72 is the number of points in the node .
somewhat more subtle combinatorics are needed in the case where some nodes in the @xmath72-tuple are identical and others are not . and fearsome computation is needed in the various cases in which some nodes are descendants of some other nodes .
in general , when the final answer comes back from @xmath71 , the majority of the quantity in the count will be the sum of components arising from large subsume prunes .
but the majority of the computational effort will have been spent on accounting for the vast number of small but unprunable combinations of nodes .
we can improve the running time of the algorithm by demanding that it also prunes it search in cases in which only a tiny count of @xmath72tuples is at stake .
this is achieved by adding a parameter , @xmath98 , to the @xmath71 algorithm , and adding the following lines at the start : * let @xmath99 * if @xmath100 then quit this recursive call .
this will clearly cause an inaccurate result , but fortunately it is not hard to maintain tight lower and upper bounds on what the true answer would have been if the approximation had not been made .
thus @xmath101 now returns a pair of counts @xmath102 where we can guarantee that the true count @xmath62 lies in the range @xmath103 .
suppose the true value of the @xmath0point function is @xmath62 but that we are prepared to accept a fractional error of @xmath104 : we will be happy with any value @xmath105 such that latexmath:[\[\label{awareapprox } possible to adapt the n - tree algorithm using a best - first iterative deepening search strategy to guarantee this result while exploiting permission to approximate effectively by building the count as much as possible from `` easy - win '' node pairs while doing approximation at hard deep node - pairs .
this is simply achieved by repeatedly calling the previous approximate algorithm with diminishing values of @xmath98 until a value is discovered that satisfies equation [ awareapprox ] .
is shown and agrees well with the observed data .
the naive @xmath107 law is also plotted for comparison [ fig4 ] ]
we plan to present a more detailed discussion of the techniques presented here in a forthcoming paper ( connolly et al .
2001 ) . that paper will also include a full analysis of the computational speed and overhead of our @xmath0point correlation function algorithm and compare those with existing software for computing the higher order correlation functions _
e.g. _ szapudi et al .
however , in figure [ fig4 ] , we present preliminary results on the scaling of computational timing needed for a 2-point correlation function as a function of the number of objects in the data set . for these tests , we computed all the data data pairs for random data sets and real , projected 2-dimensional galaxy data .
these data show that our 2-point correlation function algorithm scales as @xmath108 ( for projected 2-dimensional data ) compared to the naive all - pairs scaling of @xmath109 where here @xmath0 is the size of the dataset under consideration . to emphasis the speed up obtained by our algorithm ( figure [ fig4 ] ) , an all pairs count for a database of @xmath110 objects would take only 10 hours ( on our dec alpha workstation ) using our methodology compared to @xmath111 hours ( @xmath112 year ) using the naive @xmath107 method .
clearly , binning the data would also drastically increase the speed of analyses over the naive all
pairs @xmath109 scaling but at the price of lossing of resolution .
similar spectacular speed ups will be achieved for the 3 and 4point functions and we will report these results elsewhere ( connolly et al .
furthermore , controlled approximations can further accelerate the computations by several orders of magnitude .
such speed ups are vital to allow monte carlo estimates of the errors on these measurements . in summary ,
our algorithm now makes it possible to compute an exact , all pairs , measurement of the 2 , 3 and 4point correlation functions for data sets like the sloan digital sky survey ( sdss ) .
these algorithms will also help in the speed - up of cosmic microwave background analyses as outlined in szapudi et al .
( 2000 ) .
finally , we note here that we have only touched upon one aspect of how trees data structures ( and other computer science techniques ) can help in the analysis of large astrophysical data sets .
moreover , there are other tree structures beyond _ k_d - trees such as ball trees which could be used to optimize our correlation function codes for higher dimensionality data .
we will explore these issues in future papers .
|
we present here a new algorithm for the fast computation of @xmath0point correlation functions in large astronomical data sets .
the algorithm is based on _ k_d - trees which are decorated with cached sufficient statistics thus allowing for orders of magnitude speed
ups over the naive non - tree - based implementation of correlation functions .
we further discuss the use of controlled approximations within the computation which allows for further acceleration . in summary ,
our algorithm now makes it possible to compute exact , all pairs , measurements of the 2 , 3 and 4point correlation functions for cosmological data sets like the sloan digital sky survey ( sdss ; york et al .
2000 ) and the next generation of cosmic microwave background experiments ( see szapudi et al .
2000 ) .
|
most of the ternary intermetallic compounds with the chemical formula at@xmath0x@xmath0 ( @xmath7 alkaline earth metals , lanthanides or actinides , @xmath8 transition metal and @xmath9 , ga , si , ge , sn , p , as , ... ) crystallize in ternary variants of the tetragonal body - centred baal@xmath10 structure type , usually in the thcr@xmath0si@xmath0-structure one @xcite .
characteristic of this structure are layers of edge - connected tx@xmath10-tetrahedra , which alternate along the @xmath11-direction with planar square lattices of a - atoms .
prominent examples , concerning there physical properties , are cecu@xmath0si@xmath0 @xcite , uru@xmath0si@xmath0 @xcite and srfe@xmath0as@xmath0 @xcite , which all are model systems among the field of strongly correlated electron systems . within this field ,
the discovery of remarkable phases and transitions is often tightly coupled to the growth and characterization of novel materials . studying these emergent phenomena
necessitate the preparation of high - quality single crystals to investigate the physical properties as a function of crystal orientation and momentum . here
, we report on the details of the single crystal growth of pure ybrh@xmath0si@xmath0 and ybir@xmath0si@xmath0 as well as the alloy series yb(rh@xmath3ir@xmath4si@xmath0 , using a high - temperature indium - flux technique .
the growth of sizeable single crystals of these compounds is challenging , because of the high vapour pressure of yb , whose boiling point , @xmath12c , is well below the melting point of rh ( @xmath13c ) or ir ( @xmath14c ) .
thus , the growth from a stoichiometric melt is likely impossible .
furthermore , the solubility of si , rh and ir in the most appropriate low - melting solvents , in and sn , is small , making the growth of large crystals rather difficult .
ybrh@xmath0si@xmath0 also crystallizes in the thcr@xmath0si@xmath0 structure @xcite and was intensively studied in recent years , because it is situated on the magnetic side ( @xmath15 mk ) of , but very close to a quantum critical point ( qcp ) .
the proximity to the qcp leads to pronounced non - fermi - liquid behavior in transport and thermodynamic properties , such as the divergence of the electronic sommerfeld coefficient , @xmath16 , and a linear - in-@xmath2 resistivity @xcite .
the weak antiferromagnetic ( afm ) order can be continuously suppressed by applying a tiny magnetic field @xcite .
electrical transport and thermodynamic measurements have revealed multiple vanishing energy scales @xcite and , when the finite - temperature results are extrapolated to the @xmath17 limit , a discontinuity of the fermi volume across the qcp @xcite .
a specific feature of the ybrh@xmath0si@xmath0 single crystals are their suitability to cleave , yielding atomically flat and clean surfaces perpendicular to the crystallographic @xmath11-direction .
this made extensive angle - resolved photo - emission spectroscopy possible , which gives significant insight into the heavy - fermion behaviour of @xmath6-based correlated systems @xcite .
recent tunneling - spectroscopy experiments succeeded to directly image the transformation into the coherent ground state of the kondo lattice @xcite .
the isoelectronic compound ybir@xmath0si@xmath0 crystallizes in two different structure types , the thcr@xmath0si@xmath0 and the cabe@xmath0ge@xmath0 @xcite . in the latter
, an ir- and si - plane are interchanged with each other , therefore , the cabe@xmath0ge@xmath0 structure has a lower symmetry , with a missing mirror plane around the body centred yb - atom .
this structure is therefore named primitive ( or p - type ) in contrast to the body - centred ( i - type ) version . in fig.[fig1 ] , both structure types are presented and one clearly can see the interchanged ir-/si - atoms in the lower ir - si - plane .
low - temperature resistivity and specific - heat measurements of i - type ybir@xmath0si@xmath0 show strong similarities with ybrh@xmath0si@xmath0 .
however below 200mk , i - type ybir@xmath0si@xmath0 presents a magnetically non - ordered fermi - liquid ground state and is situated on the paramagnetic side of the qcp . a larger unit - cell volume , compared to ybrh@xmath0si@xmath0 , results in a higher kondo - energy scale , leading to a kondo - screened ground state .
therefore , the series yb(rh@xmath3ir@xmath4si@xmath0 is well suited to study the physical properties across this qcp .
first studies revealed a detaching of the afm qcp from the fermi - surface reconstruction and a new spin - liquid - type ground state for @xmath18 and @xmath19 @xcite . on the other hand
, p - type ybir@xmath0si@xmath0 orders antiferromagnetically at @xmath20k and do not present any non - fermi - liquid characteristics @xcite .
unit cell of the thcr@xmath0si@xmath0 and the cabe@xmath0ge@xmath0 structure types , denoted as i - type and p - type , respectively .
ybrh@xmath0si@xmath0 crystallizes with the i - type structure , whereas , both structures form in the case of ybir@xmath0si@xmath0.,scaledwidth=80.0% ]
the single crystal growth of ybrh@xmath0si@xmath0 and ybir@xmath0si@xmath0 was achieved using an in - flux technique @xcite .
yb ingots ( ames , 99.99% ) , rh / ir powder ( heraeus , 99.95% , si pieces ( alfa aesar , 99.9999% ) and in shots ( alfa aesar , 99.9999% ) were put together in a conical al@xmath0o@xmath21-crucible ( @xmath22ml ) .
the yb was stored and weighted in a glove box filled with purified argon gas .
great care was undertaken to keep the oxygen and water content below 0.1ppm .
afterwards , the al@xmath0o@xmath21-crucible was enclosed in a ta - tube , again under very pure ar - atmosphere , using arc welding .
subsequently , the ta - tube was put under ar in a vertical high - temperature resistance furnace , were temperatures of up to @xmath23c could be achieved .
the temperature was measured in - situ at the bottom of the crucible , independently of the furnace temperature . for cooling ,
a modified bridgman technique was utilized by slowly moving the ta - tube out of the furnace with velocities as low as 0.1mm / h .
after the growth , the in - flux was removed by etching the solidified melt in slightly diluted hydrochloric acid at moderate temperatures ( @xmath24c ) .
this dissolves also most of the binary and ternary foreign phases which has been formed due to the off - stoichiometric starting composition .
the ybt@xmath0si@xmath0 single crystals were found to be very stable in hcl , therefore , a complete etching for 1 - 2 days was possible .
finally , the single crystals were rinsed in an ultrasonic bath using acetone .
the growth of ybrh@xmath0si@xmath0 single crystals from a tin - flux was already reported in the first publication of its low - temperature properties in 2000 @xcite . in the meantime
, comprehensive studies were carried out on this model - type quantum - critical material , as discussed in the introduction .
most of these measurements were performed on single crystals , obtained from an improved flux - growth process , however , a description of the growth parameters was not yet reported . in this section ,
we discuss in detail the parameters of this high - temperature tin - flux technique .
many of the aforementioned experiments were possible only , because of the larger and purer single crystals obtained after the optimization .
two different research groups have also been working on the single crystal growth of ybrh@xmath0si@xmath0 .
et al . _ have grown mono - isotypic @xmath25ybrh@xmath0si@xmath0 single crystals from in - flux with a residual resistivity , @xmath26 cm , as well as very pure non - mono - isotypic crystals @xcite . on the latter , de - haas - van - alphen and
resistivity measurements under pressure were performed .
et al . _ reported on very large single crystals , which could be obtained using zn - flux @xcite . however , these crystals present only a residual resistivity , @xmath27 cm , likely resulting from zn - incorporation and no further physical measurements were published until now .
the optimization of the single crystal growth from a metallic flux focuses mainly on two parameters .
firstly , the initial stoichiometry and secondly , the temperature - time profile of the crystal growth .
these two parameter classes are not independent from each other , therefore , we first have optimized the initial stoichiometry and in a second step determined the best - possible temperature - time profile .
optimization of the initial stoichiometry of ybrh@xmath0si@xmath0 .
( a ) influence of the yb - content on the average crystal thickness .
( b ) influence of the rh - to - si ratio , @xmath28 , on the average crystal mass ( left axis ) and the resistivity ratio , @xmath29 ( right axis).,scaledwidth=80.0% ] in the first step , we have varied the initial atomic composition of yb , rh and si with respect to each other , but kept the mass ratio to the amount of in - flux constant , with @xmath30 . here
, @xmath31 is the sum of the mass of yb , rh and si , @xmath32 is the molar mass of indium , and @xmath33 is the molar mass of yb , rh and si in the given composition ( e.g. , for yb : rh : si=3:2:2 , @xmath34g / mol ) . during variation of the stoichiometry , we kept all other growth parameters constant . in fig .
[ ratio]a , we present the influence of the yb - content on the crystal thickness , @xmath35 , where we take an average value for the 10 largest crystals of each batch . the yb - content could be varied in a broad range from 1:2:2 to 5:2:2 ( yb : rh : si ) . in all cases stoichiometric ybrh@xmath0si@xmath0 single crystals
are formed , however , the different yb - content of the initial melt leads to single crystals with different thicknesses . from fig .
[ ratio]a it is evident that the thickest crystals ( @xmath36 mm ) could be obtained with a three - times higher yb - content , i.e. ( @xmath37 ) . for higher yb - contents , the thickness decreases and for @xmath38 only very tiny single crystals are formed . in a next step , we have optimized the ratio of rh to si , while keeping @xmath37 constant . here
we learned , that the applicable range is much narrower , compared to the yb one .
the @xmath28 ratio can only be varied from 0.9 to 1.35 . in this range
several crystal growths were carried out , with otherwise identical parameters .
we found that the mass of the crystals , @xmath39 , ( again we take an average over the 10 largest crystals ) is largest at @xmath40 corresponding to an initial ratio of yb : rh : si=3:2.2:1.8 ( fig .
[ ratio]b , left axis ) .
the thickness is nearly constant from @xmath41 to 1.22 ( @xmath42 mm ) but considerably smaller ( @xmath43 mm ) for @xmath44 and @xmath45 .
more important is the influence on the crystal quality , reflected in the resistivity ratio at 1.8k , @xmath29 , which increases from 5 for excess si to 17.5 for @xmath46 ( fig .
[ ratio]b , right axis ) .
however , ybrh@xmath0si@xmath0 is not anymore the main phase using this high rh - content and ybrhin@xmath47 is formed instead .
this is reflected in overall much smaller and thinner crystals for @xmath46 .
therefore , we decided to continue the optimization procedure of the temperature profile with the atomic ratio yb : rh : si=3:2.2:1.8 . optimized temperature - time profile for the single crystal growth of ybrh@xmath0si@xmath0 and ybir@xmath0si@xmath0 .
the temperature was measured in - situ at the bottom of the crucible.,scaledwidth=80.0% ] differential - thermal - analysis measurements were carried out to determine the liquidus temperature of ybrh@xmath0si@xmath0 in 96 mol% in .
we found an anomaly in the cooling curve at @xmath48c , which most likely is due to the formation of ybrh@xmath0si@xmath0 within the liquid indium . from several growths ,
not following a systematic optimization procedure , we observed that a reduction of the time at highest temperatures results in higher - quality single crystals .
therefore , we have optimized the temperature - time profile , which includes several steps ( solid line in fig . [ temp ] ) , by using a vertical two - zone furnace ( xerion company ) . first , we heated both zones to reach a temperature of @xmath49c throughout the crucible for one hour ( step ( i ) in fig .
[ temp ] ) .
subsequently , a temperature gradient was build up within the crucible by moving the sample out of the furnace ( step ( iia ) ) and by slowly decreasing the power of the lower heating zone ( step ( iib ) ) .
finally , we start the bridgman - type cooling by moving the ta - tube further out of the furnace , whereas the higher heating zone is kept at its initial temperature ( step ( iii ) ) .
the crystal growth is finished at around @xmath50c and the furnace is cooled down to room temperature ( step ( iv ) ) .
we assume that the following processes take place : step ( i ) provides an homogenization of the melt at temperatures well above @xmath51 .
the first seeds of ybrh@xmath0si@xmath0 can form at the bottom of the crucible when the temperature goes through @xmath51 ( step ( iib ) ) .
we use the second heating zone to build up a temperature gradient inside the crucible , which significantly reduces the total time at high temperatures .
the crystal grow during step ( iii ) , where the temperature - time curve has at the beginning a much smaller slope .
in contrast to a normal bridgman technique without a metallic flux , here , several seeds form at the bottom of the crucible , which result in many crystals of comparable size and quality .
the optimization of the temperature - time profile leads to single crystals with much higher resistivity ratios , reaching @xmath52 .
some of these crystals are pictured in fig .
these new generation of single crystals were deeply investigated at low temperatures and reveal a broad range of exciting physical phenomena .
the residual resistivity , extracted from measurements down to the mk range , reaches @xmath53 cm resulting in a resistivity ratio as high as @xmath54 @xcite . the extremely sharp anomaly in the specific heat at @xmath55 is a further proof of the high quality @xcite .
directly visible is the low defect density of these crystals in recent scanning tunneling measurements @xcite , where large areas of defect - free surfaces could be resolved .
some of the few defects could be assigned to rh occupation on si - sites , in agreement with the slight rh excess described above . in contrast , single crystals prepared with excess si show a lot more defects at the surface @xcite , which confirms that the resistivity ratio is a valid measure to access the quality of the crystals . ( a ) photograph of high - quality ybrh@xmath0si@xmath0 single crystals .
the edges are statistically aligned along 100 and 110 , which can be proved by laue back scattering images plotted in ( b ) . here
, the horizontal direction corresponds to the 100 direction.,scaledwidth=80.0% ] the growth of the related compound ybir@xmath0si@xmath0 is possible using a similar high - temperature in - flux method @xcite . here , we used an one - zone resistance furnace and a slightly different temperature - time profile during the crystal growth ( see dotted curve in fig .
[ temp ] ) .
the homogenization ( step ( i ) ) is done for 8 hours at @xmath56c .
the temperature gradient was build up by moving the sample out of the furnace , which leads to a temperature jump of @xmath57c ( ( step ( ii ) in fig .
[ temp ] ) .
subsequently , we start the bridgman - type cooling by slowly moving the sample tube further out of the furnace ( step ( iii ) ) . after 300 hours , the crystal growth is terminated at around @xmath58c . using a stoichiometric composition of the elements with 96mol% indium ,
gives quite large crystals with an average mass of 35 mg and an average thickness of 0.6 mm .
furthermore , the purest crystals obtained without optimization have already a residual resistivity of only @xmath59 cm giving a resistivity ratio , @xmath60 @xcite .
the reproducibility of the growth process is not as good as for ybrh@xmath0si@xmath0 .
nearly identical growth parameters lead sometimes to large but also to small crystals , making a quantitative optimization very difficult .
however , we varied also for ybir@xmath0si@xmath0 the initial composition and found that the applicable ranges are much smaller than for ybrh@xmath0si@xmath0 .
doubling the yb - content , @xmath61 , do not yield ybir@xmath0si@xmath0 crystals anymore , instead binary phases as e.g. , ybin@xmath0 , ybir@xmath21 and yb@xmath47si@xmath21 are formed .
a small excess of yb , @xmath62 , leads to the formation of p - type ybir@xmath0si@xmath0 . using an off - stoichiometric ratio of ir to si , @xmath63 , do no result in ybir@xmath0si@xmath0 crystals but to the formation of some unknown phases . from these results
it is obvious that the optimal crystal growth parameters are much different in ybrh@xmath0si@xmath0 and ybir@xmath0si@xmath0 .
however , the present output , concerning size and quality of the ybir@xmath0si@xmath0 crystals is already satisfactory , so no further optimization was undertaken .
the formation of the two structural variants , i - type and p - type , in ybir@xmath0si@xmath0 is not surprising , because these two different structures were observed in other rir@xmath0si@xmath0 compounds as well @xcite .
there , it was established , that the p - type configuration forms only at high temperatures , whereas the i - type is stabilized at lower temperatures . here , we are mainly interested in the i - type structure which is isostructural to ybrh@xmath0si@xmath0
. therefore , we have chosen the temperature - time profile such that a long annealing time below @xmath64c was achieved ( see fig .
[ temp ] ) .
it is presently unclear , why the excess of yb leads to the stabilization of the p - type phase .
however , we were able to transform the p - type single crystals into i - type by annealing the crystals at @xmath64c for 150 hours . for this purpose
the crystals extracted from the in - flux were enclosed in a ta - crucible under argon at ambient pressure .
we discuss the differences of i- and p - type in the x - ray diffraction patterns and the different physical ground states in the sections [ secxraypi ] and [ secpi ] , respectively .
as stated above , negative chemical pressure has to be applied to ybrh@xmath0si@xmath0 to reach the qcp in the absence of external magnetic field .
this can be achieved by various kind of substitutions .
ge - substitution on the si site was tried first @xcite , however , the maximal substitution level amounts to only about 2at% @xcite .
higher substitution can be achieved replacing the magnetic yb - atoms with non - magnetic la @xcite or lu @xcite . here
, we discuss a further substitution series , yb(rh@xmath3ir@xmath4si@xmath0 , where the rh atoms are replaced by the larger isoelectronic ir - atoms .
[ tabirdop ] for the crystal growth of yb(rh@xmath3ir@xmath4si@xmath0 we used the optimized growth parameters of ybrh@xmath0si@xmath0 , with the respective replacement of rh by ir .
the initial stoichiometry together with the properties of the resulting single crystals are given in table [ tabirdop ] .
for all growths the amount of in - flux was kept constant as described above , with @xmath65 .
the ir content was determined using energy dispersive x - ray ( edax ) spectra obtained in a scanning electron microscope ( philips xl30 ) with a si(li)-x - ray detector .
three crystals of each batch with a polished surface were analyzed at several positions .
the respective mean values result in the given @xmath66 values with an absolute error of less than @xmath67 .
these real ir - concentrations , which are used for @xmath68 throughout this article , are somewhat different than the nominal concentrations , @xmath69 , due to different solubilities of the elements in the in - flux .
the average thickness and mass of the single crystals are again determined by averaging over the 10 largest crystals .
the resistivity ratio , @xmath29 , reflects the amount of disorder in the single crystals .
unfortunately , the size of the crystals gets smaller with increasing ir - content , opposite to what was observed for the yb(rh@xmath70co@xmath71si@xmath0 series , where the crystal growth gives larger crystals , with increasing @xmath72 @xcite .
the value of the resistivity ratio decreases with increasing @xmath68 , which is expected due to the higher amount of disorder introduced by the statistical occupation of the ir atoms at the rh site .
an exception is the growth with @xmath73 , because @xmath29 is somewhat larger than for @xmath74 .
this is most probably caused by the higher rh / ir - content with respect to si ( @xmath75 ) , similar to what was observed in the pure ybrh@xmath0si@xmath0 , where we found a larger resistivity ratio for higher rh - content ( see fig .
[ ratio ] ) .
a series of 10 samples of different quality has been systematically analyzed by x - ray diffraction and electron microprobe analysis .
the lattice parameters were accurately determined in respect to the chemical composition and the homogeneity range of the phase ybrh@xmath0si@xmath0 was defined .
these results and the experimental details will be published elsewhere @xcite . in fig .
[ xtal]b , we present a x - ray laue backscattering image from a ybrh@xmath0si@xmath0 single crystal .
the reflection peaks are pin sharp rather than blurred , proving again the high quality of the synthesized crystal .
these pictures were used to orient the single crystals . in all cases
, the direction perpendicular to the surface corresponds to the crystallographic @xmath11-axis , proved by the fact that the 001 reflection is exactly in the middle of the aperture .
simulations demonstrate that the horizontal direction corresponds to the 100 direction , which however , not always corresponds to the edges of the crystal , visible in fig .
our experience from the orientation of more than 100 of these crystals for photo - emission spectroscopy , revealed that the crystal edges much more often corresponds to the 110 direction , although not exclusively .
comparison of three different x - ray powder diffraction patterns of ybir@xmath0si@xmath0 .
( a ) i - type after annealing .
( b ) i - type with small amount of p - type , leading to additional peaks at @xmath76 ( arrows ) .
( c ) p - type pattern ; the intensities of the peaks have changed considerably and all odd @xmath77 are observed .
however , small amount of i - type within the p - type phase can not be excluded.,scaledwidth=80.0% ] the distinction between the body - centred and the primitive structure type of ybir@xmath0si@xmath0 is possible using powder x - ray diffraction , which was performed on a stoe diffractometer in transmission mode using monochromated cu - k@xmath78 radiation ( @xmath79 ) . for the primitive structure additional peaks ( @xmath80 )
are observable . for the body - centred structure
these peaks with an odd sum of @xmath77 are absent . in fig .
[ fig5xrd ] we present three different diffractograms for a ) pure i - type ybir@xmath0si@xmath0 , b )
i - type with small content of p - type and c ) mainly p - type .
the diffraction pattern of the annealed single crystals only show very tiny odd @xmath77 peaks , confirming that in this case we have virtually pure i - type crystals .
when the crystal growth is terminated before ybir@xmath0si@xmath0 is completely transformed into the i - type structure , the diffraction pattern presents small peaks at the @xmath80 positions ( arrows in fig .
[ fig5xrd]b ) .
the growth with yb - excess leads to p - type ybir@xmath0si@xmath0 , the corresponding pattern is shown in fig .
[ fig5xrd]c . here ,
the intensities of the peaks have changed considerably and all odd @xmath77 peaks are observed .
however with powder diffraction , we can not finally prove , whether parts of the p - type crystal are already transformed into i - type .
this is because the lattice parameters of the two structures are nearly identical with @xmath81 , @xmath82 and @xmath83 , @xmath84 , which were refined by simple least squares fitting .
lattice parameters @xmath85 ( left axis ) and @xmath11 ( right axis ) for the yb(rh@xmath3ir@xmath4si@xmath0 series . with increasing ir content , @xmath85 linearly increases , whereas @xmath11 decreases.,scaledwidth=80.0% ] in fig .
[ fig6ac ] the lattice parameters @xmath85 ( left axis ) and @xmath11 ( right axis ) are plotted for the different ir - concentrations .
an important point is immediately visible , namely that @xmath85 and @xmath11 have opposite trends with increasing @xmath68 .
the lattice parameter @xmath85 , which describes the distance of the yb - ions in the plane ( see fig .
[ fig1 ] ) , increases linearly with @xmath68 and can be described with @xmath86( ) @xmath87 .
in contrast , @xmath11 decreases linearly , with @xmath88( ) @xmath89 .
the resulting volume of the unit cell increases with increasing @xmath68 , however , the overall volume change going from ybrh@xmath0si@xmath0 to ybir@xmath0si@xmath0 amounts to only @xmath90@xmath91 , which is less than 1% .
this is a rather small increase of the unit cell volume , because the variation of @xmath85 and @xmath11 cancels out each other . the corresponding hydrostatic pressure , if applied to ybir@xmath0si@xmath0 , amounts to @xmath92gpa , using a bulk modulus of 190gpa @xcite .
however , for increasing ir - concentration , the ratio of the lattice parameters , @xmath93 , strongly decreases , as @xmath11 decreases and @xmath85 increases with @xmath68 .
this results in a distortion of the unit cell , which gets less elongated with higher ir - content .
therefore , isoelectronic substitution is not equivalent to hydrostatic pressure , which would result in a constant @xmath93 ratio .
this is probably the reason why for ybir@xmath0si@xmath0 a huge pressure of 8gpa is needed to induce magnetic order @xcite , in contrast to the expected 1.6gpa mentioned above .
however , for small @xmath68 the relative change of @xmath93 is less than 0.1% , compared to pure ybrh@xmath0si@xmath0 , so that small ir - substitution is equivalent to hydrostatic pressure @xcite .
temperature dependence of the electrical - resistivity exponent , @xmath94 , for two single crystals with different residual resistivity . with higher sample quality ( lower @xmath95 ) pronounced deviations from the linear - in-@xmath2 resistivity
are evident below 10k.,scaledwidth=80.0% ] one of the hallmarks of non - fermi - liquid behaviour in ybrh@xmath0si@xmath0 is the linear - in-@xmath2 resistivity over more than an order of magnitude in temperature .
however , this was studied in a wider temperature range only for the very first generation of single crystals with @xmath96 cm @xcite . here , we present an analysis of the temperature exponent of the resistivity for one of the optimized single crystals with a residual resistivity as low as @xmath1 cm . in zero magnetic field ,
the resistivity shows a sharp drop at the afm phase transition ( @xmath97mk ) but a small critical magnetic field ( @xmath98mt , @xmath99 ) can be applied to tune the system towards quantum criticality , where the non - fermi - liquid behaviour is observed down to lowest temperatures @xcite .
generally , the temperature dependence of the resistivity can be described with @xmath100 .
therefore , the temperature exponent , @xmath101 , can be visualized by plotting @xmath102 , shown in fig .
[ fig7exp ] for two samples with different residual resistivity .
the data for the old sample presents a nearly constant @xmath103 below 7k , down to lowest measured temperatures .
in contrast , the exponent of the new sample exhibits a pronounced maximum at around 4k with @xmath104 , followed by a minimum around 0.3k with @xmath105 and a value of @xmath103 at lowest temperatures . for this analysis
we estimate @xmath95 by linear extrapolation to @xmath106 , however , the choice of @xmath95 influences @xmath107 only at lowest temperatures .
this is visible by the larger error bars of @xmath101 below 0.1k , which result from the uncertainty of @xmath95 .
beside this temperature - dependent analysis of the exponent , we additionally fit the data with a single @xmath101 in a larger temperature range . for the resistivity data below 0.1k
, we obtain an exponent of @xmath108 . a fit below 0.2k down to lowest temperatures gives @xmath109 , in agreement with the differential analysis .
we note that the evolution of the exponent for these two different samples presents qualitative similarities to quasi - classical calculations for a system in the vicinity of an antiferromagnetic qcp . there
, a. rosch studied the influence of strong anisotropic scattering due to spin fluctuations for different amount of disorder and he observed broad ranges of exponents between 1 and 2 together with oscillatory behaviour @xcite . it might be interesting to implement the specifics of ybrh@xmath0si@xmath0 to such a model to get further insight into the transport phenomena at this local qcp .
specific heat divided by temperature , @xmath110 , of ybir@xmath0si@xmath0 ( p - type ) on a linear temperature scale . at @xmath111k , a pronounced anomaly is visible which is suppressed in @xmath112 t .
the dc susceptibility , plotted in the inset , confirms the magnetic nature of this phase transition.,scaledwidth=80.0% ] ybir@xmath0si@xmath0 crystallizes in two different structure types , a bodycentred ( i - type ) and a primitive ( p - type ) variant ( see fig . [ fig1 ] and sec .
[ secyisgrowth ] ) .
interestingly , the magnetic ground states of these two derivatives of the baal@xmath10 structure are very different , although , the tetragonal lattice parameters are nearly identical , as discussed in sec .
[ secxraypi ] .
have shown that i - type ybir@xmath0si@xmath0 is a heavy - fermion metal situated on the non - magnetic side of a qcp with fermi - liquid behaviour below 0.4k and quite large single - ion anisotropy , reflected in a larger susceptibility perpendicular to the crystallographic @xmath11-direction @xcite .
on the other hand , the susceptibility of p - type ybir@xmath0si@xmath0 is rather isotropic @xcite and entropy as well as resistivity measurements suggest a much smaller kondo - temperature , @xmath113k , in contrast to @xmath114k for the i - type system @xcite .
furthermore , in p - type crystals a prominent anomaly was observed in the specific heat below 1k , which was suggested to be a magnetic phase transition . here , we further like to establish that p - type ybir@xmath0si@xmath0 indeed orders antiferromagnetically at @xmath115k by means of susceptibility measurements below 2k . for this purpose ,
p - type crystals were prepared with yb - excess as described in sec .
[ secyisgrowth ] and the dc susceptibility was measured down to @xmath116k , which was realized with the @xmath91he - option for the mpms ( quantum design ) designed by the iquantum corporation . in the inset of fig .
[ fig9ptype ] , we present this measurement with @xmath117 t perpendicular to @xmath11 .
a weak anomaly suggests that p - type ybir@xmath0si@xmath0 undergoes a magnetic phase transition at @xmath20k , although the transition is not very sharp .
specific heat measurements on these crystals confirm the earlier measurements in zero field @xcite and are shown in the main part of fig .
[ fig9ptype ] . applying a magnetic field of 0.1 t
does not affect the anomaly at @xmath55 , whereas a field of @xmath118 t suppresses the magnetic transition and shifts the entropy to higher temperatures .
the characteristics of this field dependence confirms the afm nature of the phase transition and classifies p - type ybir@xmath0si@xmath0 as a magnetically ordered system with localized yb@xmath119 moments .
however , the anomalies at @xmath55 are rather broad and we were not able to detect the magnetic transition in low - temperature resistivity measurements , probably due to significant amount of disorder , likely ir - si site exchanges .
in conclusion , we have optimized the single crystal growth of ybrh@xmath0si@xmath0 , using an indium - flux method with 96 mol% indium in closed ta - crucibles .
the optimal initial composition of the elements within the flux , yb : rh : si=3:2.2:1.8 , is far away from the stoichiometric ratio but was found to give large ( up to 100 mg ) single crystals .
prior to the optimization of the temperature - time profile , we determined the liquidus temperature ( @xmath120c ) of ybrh@xmath0si@xmath0 in 96 mol% in - flux by means of differential thermal analysis and use a two - zone resistance furnace to reduce the time of the crystal growth .
the resulting single crystals have a residual resistivity of only @xmath1 cm and present sharp anomalies in the transport and thermodynamic properties at the low - lying afm phase transition .
the analysis of the temperature - dependence of the electrical - resistivity exponent for this new generation of single crystals reveals a deviation from the linear - in-@xmath2 behaviour below 10k .
ybir@xmath0si@xmath0 was established to be a polymorphic compound which can crystallize in two different crystallographic modifications , with p - type crystals ( cabe@xmath0ge@xmath0 structure ) being the high - temperature and i - type ( thcr@xmath0si@xmath0 structure ) the low - temperature phase .
the latter can be stabilized by annealing selected crystals at @xmath64c for 150 hours .
we further found that the optimal growth parameters for ybrh@xmath0si@xmath0 and ybir@xmath0si@xmath0 from in - flux are much different .
for ybir@xmath0si@xmath0 , the applicable range of the initial composition is very narrow and small excess of yb leads to the formation of p - type crystals .
interestingly , both modifications have different physical ground states . in p - type ybir@xmath0si@xmath0
, the magnetic @xmath6-electrons undergo an afm phase transition at @xmath121k , observed in low - temperature specific - heat and susceptibility measurements , whereas i - type ybir@xmath0si@xmath0 presents a paramagnetic fermi - liquid ground state and is situated on the non - magnetic side of the quantum critical point .
the series yb(rh@xmath3ir@xmath4si@xmath0 , which crystallizes in the i - type structure , is therefore ideally suited to study the phenomena around this quantum phase transition .
the authors thank u. burkhardt and p. scheppan for energy dispersive x - ray analysis of the samples as well as n. caroca - canales and r. weise for technical assistance .
the presented low - temperature susceptibility measurements were performed with the help of c. klausnitzer .
we acknowledge valuable discussions with m. brando , m. deppe , j. ferstl , s. friedemann , p. gegenwart , a. rosch , and f. steglich . the dfg ( research unit 960 , `` quantum phase transitions '' ) is acknowledged for financial support .
s. danzenbcher , y. kucherenko , d.v .
vyalikh , m. holder , c. laubschat , a.n .
yaresko , c. krellner , z. hossain , c. geibel , x.j .
zhou , w.l .
yang , n. mannella , z. hussain , z.x .
shen , m. shi , l. patthey and s.l .
molodtsov , phys .
b 75 ( 2007 ) p.045109 . d.v .
vyalikh , s. danzenbcher , y. kucherenko , k. kummer , c. krellner , c. geibel , m.g .
holder , t.k .
kim , c. laubschat , m. shi , l. patthey , r. follath and s.l .
molodtsov , phys .
105 ( 2010 ) p.237601 .
|
we report on the single crystal growth of the heavy - fermion compounds ybrh@xmath0si@xmath0 and ybir@xmath0si@xmath0 using a high - temperature indium - flux technique .
the optimization of the initial composition and the temperature - time profile lead to large ( up to 100 mg ) and clean ( @xmath1 cm ) single crystals of ybrh@xmath0si@xmath0 .
low - temperature resistivity measurements revealed a sample dependent temperature exponent below 10k , which for the samples with highest quality deviates from a linear - in-@xmath2 behaviour .
furthermore , we grew single crystals of the alloy series yb(rh@xmath3ir@xmath4si@xmath0 with @xmath5 and report the structural details . for pure ybir@xmath0si@xmath0 , we establish the formation of two crystallographic modifications , where the magnetic @xmath6-electrons have different physical ground states .
ybrh@xmath0si@xmath0 , ybir@xmath0si@xmath0 , metal - flux technique , kondo - lattice systems , quantum critical point
|
since its introduction at the 2009 prague stringology conference @xcite , the problem of indexed binary jumbled pattern matching has been discussed in many top conferences and journals .
it asks us to preprocess a binary string such that later , given a number of 0s and a number of 1s , we can quickly report whether there exists a substring with those numbers of 0s and 1s and , optionally , return the position of one such substring or possibly even all of them .
the nave preprocessing algorithm takes quadratic time but researchers have reduced that bound to @xmath6 @xcite , @xmath7 @xcite , @xmath8 @xcite and finally @xmath9 with randomization or @xmath10 without @xcite .
researchers have also looked at indexing for approximate matching @xcite , indexed jumbled pattern matching over larger alphabets @xcite , indexing labelled trees and other structures @xcite , and how to index faster when the ( binary ) input string is compressible .
gagie et al .
@xcite gave an algorithm that runs in @xmath11 when the input is represented as a straight - line program with @xmath12 rules , and badkobeh et al .
@xcite gave one that runs in @xmath0 time when the input consists of @xmath2 runs , i.e. , maximal unary substrings ( we will denote later as @xmath2 the number of maximal substrings of 1s , for convenience ) . giaquinta and grabowski
@xcite gave two algorithms : one runs in @xmath13 time , where @xmath14 is a parameter , and produces an index that uses @xmath15 extra space and answers queries in @xmath16 time ; the other runs in @xmath17 time , where @xmath18 is the size of a machine word .
amir et al .
@xcite gave an algorithm that runs in @xmath19 time when the input is a run - length encoded binary string , or @xmath0 time when it is a plain binary string ; it builds an index that takes @xmath20 words and answers queries in @xmath21 time , however .
very recently , sugimoto et al .
@xcite considered the related problems of finding abelian squares , abelian periods and longest common abelian factors , also on run - length encoded strings .
we first review some preliminary notions in section [ sec : pre ] .
we present our main result in section [ sec : alg ] : a new and very simple indexing algorithm that runs in @xmath3 time , which matches giaquinta and grabowski s algorithm with the parameter @xmath22 and is thus tied as the fastest known when @xmath23 and the smallest straight - line program for the input has @xmath24 rules . for an input string of up to ten million bits , for example , if the average run - length is three or more then @xmath25 .
our algorithm takes only 17 lines of pseudocode , making it a promising starting point for investigating other possible algorithmic features . in section [ sec : witness ] , for example , we show how to extend our algorithm to store information that lets us report the position of a match ( if there is one )
. finally , in section [ sec : workspace ] , we show how we can alternatively adapt it to use only @xmath5 bits of space .
consider a string @xmath26 .
we denote by @xmath27 $ ] the substring of @xmath28 consisting of the @xmath29th through @xmath30th characters , for @xmath31 ; if @xmath32 , we can also write simply @xmath33 $ ] .
cicalese et al .
@xcite observed that , if we slide a window of length @xmath14 over @xmath28 , the number of 1s in the window can change by at most 1 at each step .
it follows that if @xmath34 $ ] contains @xmath35 copies of 1 and @xmath36 $ ] contains @xmath37 copies of 1 with @xmath38 then , for @xmath39 between @xmath35 and @xmath37 ( notice @xmath35 could be smaller than , larger than , or equal to @xmath37 ) , there is a substring of length @xmath14 in @xmath40 $ ] with exactly @xmath39 copies of 1 .
this immediately implies the following theorem : [ thm : table ] let @xmath35 and @xmath37 be the minimum and maximum numbers of 1s in any substring of length @xmath14 .
there is a substring of length @xmath14 with @xmath39 copies of 1 if and only if @xmath41 . by theorem [ thm : table ] ,
if we compute and store , for @xmath42 , the minimum and maximum numbers of 1s in a substring of @xmath28 of length @xmath14 then later , given a number of 0s and a number of 1s , we can report in constant time whether there exists a substring with that many 0s and 1s . for example , if @xmath43 then , as @xmath14 goes from 1 to @xmath44 , the minimum and maximum numbers of 1s are @xmath45 and @xmath46 , respectively . since the fifth numbers in these lists are 2 and 4 , we know , there are substrings of length 5 with exactly 2 , 3 and 4 copies of 1 , but none with 0 , 1 or 5 ( or more than 5 , obviously ) .
cicalese et al .
@xcite noted that , if we also store the positions of the substrings with the minimum and maximum numbers of 1s and a bitvector for @xmath28 that supports constant time rank queries , then via binary search in @xmath47 time we can find an example of a substring with any desired numbers of 0s and 1s , called a _ witness _ if such a substring exists .
( the query @xmath48 returns the number of 1s in @xmath49 $ ] ; see , e.g. , @xcite for more details of rank queries on bitvectors . ) for example , suppose we want to find a substring of length 5 with exactly 3 copies of 1 in our example string @xmath28 .
we have stored that there are substrings of length 5 with 2 and 4 copies of 1 starting at positions 1 and 4 , respectively , so we know there is a substring of length 5 with exactly 3 copies of 1 starting in @xmath50 $ ] .
we choose @xmath51 and check how many 1s there are in @xmath52 $ ] via two rank queries . in this case
, the answer is 3 , so we have found a witness in one step ; otherwise , we would know there is a witness starting in @xmath53 $ ] and we would recurse on that interval .
the same authors noted that in each step , the lists of minimum and maximum numbers can only stay the same or increment , so we can represent each list as a bitvector of length @xmath1 and support access to it using rank queries .
for example , the bitvector for the list of minimum numbers in our example is @xmath54 , so @xmath48 returns the @xmath29th number in the list .
since an @xmath1-bit bitvector takes @xmath5 bits of space , it follows that we can store our index in @xmath5 bits and still support constant - time queries , if we do not want a witness .
we note , however , that even though the input @xmath28 takes @xmath1 bits and the resulting index takes @xmath5 bits , all previous constructions have used @xmath55 _ words _ in the worst case .
a _ run _ in @xmath28 is a maximal unary substring and the run - length encoding @xmath56 is obtained by replacing each run by a copy of the character it contains and its length . although @xmath2 is usually used to denote the number of runs , for convenience , we use it to denote only the number of runs of 1s about half its normal value for binary strings and consider @xmath28 to begin and end with ( possibly empty ) runs of 0s .
for example , for our example string the run - length encoding is @xmath57 and @xmath58 ( instead of 9 ) .
we denote the lengths of the runs of zeroes and 1s as @xmath59 , \ldots , z [ \rho]$ ] and @xmath60 , \ldots , o [ \rho]$ ] , respectively .
since finding substrings with the minimum numbers of 1s is symmetric to finding substrings with the maximum numbers of 1s ( e.g. , by taking the complement of the string ) , we describe how , given a binary run - length encoded string @xmath61 $ ] , we can build a table @xmath62 $ ] such that @xmath63 = f ( k)$ ] , where @xmath64 denotes the maximum number of 1s in a substring of @xmath28 of length @xmath14 . the complete pseudo - code of our algorithm only 17 lines is shown as algorithm [ alg : index ] .
the starting point of our explanation and proof of correctness is the observation that , if the bit immediately to the left of a substring is a 1 , we can shift the substring one bit left without decreasing the number of 1s ; if the first bit of the substring is a 0 , then we can shift the substring one bit right ( shortening it on the right if necessary ) without decreasing the number of 1s .
it follows that , for @xmath42 , there is a substring of length at most @xmath14 containing @xmath64 copies of 1 and starting at the beginning of a run of 1s .
since we can remove any trailing 0s from such a substring also without changing the number of 1s , there is such a substring that also ends in a run of 1s .
therefore we have the following lemma : [ lem : starting ] for @xmath42 , there is a substring of length at most @xmath14 containing @xmath64 copies of 1 , starting at the beginning of a run of 1s and ending in a run of 1s .
applying lemma [ lem : starting ] immediately yields an @xmath65-time algorithm : set @xmath62 $ ] to all 0s ; for each position @xmath29 at the beginning of a run of 1s and each position @xmath30 in a run of 1s , set @xmath66 = \max ( t [ j - 1 + 1 ] , s [ i ] + \cdots + s [ j])$ ] ; finally , because @xmath67 is non - decreasing , make a pass over @xmath68 from @xmath69 $ ] to @xmath70 $ ] setting each @xmath71 = \max ( t [ i ] , t [ i - 1])$ ] . computing the number @xmath33 + \cdots + s [ j]$ ] of 1s in a substring @xmath72 $ ] starting at the beginning of a run of 1s and ending in a run of 1s is easy to do from the run - length encoding in amortized constant time . to speed this preliminary algorithm up to run in @xmath3 time , we first observe that , if @xmath73 is the length of a substring starting at the beginning of a run of 1s , ending in a run of 1s and containing @xmath74 copies of 1 , and @xmath75 is the length of a substring starting at the beginning of a run of 1s and ending at the end of a run of 1s , then @xmath76 .
( in fact this is true for any @xmath73 and @xmath77 , simply because @xmath78 for all @xmath35 . )
we then observe that , for some such @xmath79 , we have @xmath80 . to see why , consider any substring @xmath27 $ ] of length @xmath73 starting at the beginning of a run of 1s , ending within a run of 1s and containing @xmath74 copies of 1 :
let @xmath79 be the length of the substring starting at @xmath33 $ ] and ending at the end of the run of 1s containing @xmath81 $ ] , so @xmath80 .
[ lem : ending ] if @xmath73 is the length of a substring starting at the beginning of a run of 1s , ending in a run of 1s and containing @xmath74 copies of 1 , and @xmath75 is the length of a substring starting at the beginning of a run of 1s and ending at the end of a run of 1s , then @xmath76 .
furthermore , for some such @xmath79 , we have @xmath80 . with lemma [ lem : ending ]
, we can compute the number @xmath82 + \cdots + s[j]$ ] of 1s in each substring @xmath72 $ ] starting at the beginning of a run of 1s and ending in a run of 1s , in a total of @xmath3 time : again , set @xmath62 $ ] to all 0s ; for each position @xmath29 at the beginning of a run of 1s and each position @xmath30 at the end of a run of 1s , set @xmath66 = \max ( t [ j - 1 + 1 ] , s [ i ] + \cdots + s [ j])$ ] ; make a pass over @xmath68 from @xmath83 $ ] to @xmath84 $ ] setting each @xmath71 = \max ( t [ i ] , t [ i + 1 ] - 1)$ ] . computing the number @xmath33 + \cdots + s [ j]$ ] of 1s in a substring
@xmath27 $ ] starting at the beginning of a run of 1s and ending at the end of a run of 1s is again easy to do from the run - length encoding in amortized constant time . combining lemmas [ lem : starting ] and [ lem : ending ] , we have a complete algorithm for computing @xmath68 in @xmath3 time : set @xmath62 $ ] to all 0s ; for each position @xmath29 at the beginning of a run of 1s and each position @xmath30 at the end of a run of 1s , set @xmath66 = \max ( t [ j - 1 + 1 ] , s [ i ] + \cdots + s [ j])$ ] ; make a pass over @xmath68 from @xmath83 $ ] to @xmath84 $ ] setting each @xmath71 = \max ( t [ i ] , t [ i + 1 ] - 1)$ ] ( which sets @xmath85 $ ] correctly for every length @xmath73 of a substring starting at the beginning of a run of 1s , ending in a run of 1s and containing @xmath74 copies of 1 ) ; and make a pass over @xmath68 from @xmath69 $ ] to @xmath70 $ ] setting each @xmath71 = \max ( t [ i ] , t [ i - 1])$ ] ( which sets every entry in @xmath68 correctly ) .
once we have @xmath68 , we can convert it into a bitvector in @xmath5 time .
summarizing our results so far , we have the following theorem , which we adapt in later sections : [ thm : main ] given a binary string @xmath28 of length @xmath1 containing @xmath2 runs of 1s , we can build an @xmath5-bit index for constant - time jumbled pattern matching in @xmath3 time .
now we examine how our algorithm works on our example @xmath43 .
first we set all entries of @xmath68 to 0 , then we loop through the runs of 1s and , for each , loop through the runs of 1s not earlier , computing distance from the start of the first to the end of the second and the number of 1s between those positions . while doing this , we set @xmath84 = 1 $ ] , the number of 1s from the start to the end of the first run of 1s ; @xmath86 = 2 $ ] , the number of 1s from the start of the first run of 1s to the end of the second run of 1s ; @xmath87 = 5 $ ] , the number of 1s from the start of the first run of 1s to the end of the third run of 1s ; @xmath88 = 7 $ ] , the number of 1s from the start of the first run of 1s to the end of the fourth run of 1s ; @xmath89 = 4 $ ] , the number of 1s from the start of the second run of 1s to the end of the third run ; etc . when we have finished this stage , @xmath90 $ ] .
we then make a pass over @xmath68 from right to left , setting each @xmath71 = \max ( t [ i ] , t [ i + 1 ] - 1)$ ] . after this stage , @xmath91 $ ] .
finally , we make a pass over @xmath68 from left to right , setting each @xmath71 = \max ( t [ i ] , t [ i - 1])$ ] .
this fills in @xmath92 $ ] and leaves @xmath68 correctly computed as @xmath93 $ ] .
as described in section [ sec : pre ] , if together with computing the minimum and maximum number of 1s in a substring of length @xmath14 for @xmath42 , we also store the positions of substrings of length @xmath14 with those numbers of 1s , and a single bitvector for @xmath28 , then , together with confirming that @xmath28 contains a substring with a given number of 0s and 1s ( if it does ) , we can give the starting position of such a substring , still in constant time . in this section ,
we show how to modify our algorithm from section [ sec : alg ] to build also a table @xmath94 $ ] such that @xmath95 $ ] is the starting position of a substring of length @xmath14 containing @xmath64 copies of 1s .
computing and storing the starting position of a substring of length @xmath14 with the minimum number of 1s is symmetric .
first , notice that during the first stage of algorithm [ alg : index ] , whenever we set @xmath63 = f ( k)$ ] , we have found a substring of length @xmath14 containing @xmath64 copies of 1 , so we can set @xmath95 $ ] at the same time .
now consider the second stage of the algorithm , in which we make a right - to - left pass over @xmath68 setting @xmath71 = \max ( t [ i ] , t [ i + 1 ] - 1)$ ] for @xmath96 .
when we start this stage , for every positive entry in @xmath68 we have set the corresponding entry in @xmath97 .
therefore , by induction , whenever we set @xmath71 = t [ i + 1 ] - 1 $ ] , we have @xmath98 $ ] set to the starting position of a substring of length @xmath99 containing @xmath100 $ ] copies of 1 .
the substring of length @xmath29 starting at @xmath98 $ ] contains at least @xmath100 - 1 $ ] copies of 1 , so we can set @xmath101 = p [ i + 1]$ ] . in the last stage of the algorithm ,
in which we make a left - to - right pass over @xmath68 , we can almost use the same kind of argument and simply copy @xmath97 values when we copy @xmath68 values , except that we must ensure the starting positions we copy are far enough to the left of the end of the string ( i.e. , that the substrings have the correct lengths ) .
our modified algorithm is shown as algorithm [ alg : witness ] still only 25 lines and we now have the following theorem : [ thm : witness ] given a binary string @xmath28 of length @xmath1 containing @xmath2 runs of 1s , we can build an @xmath5-word index for constant - time jumbled pattern matching _ with witnessing _ in @xmath3 time . running our modified algorithm on our example @xmath43 , in the first stage we set @xmath90 $ ] and , simultaneously , @xmath102 $ ] , where 0 indicates an unset value in @xmath97 . in the second stage
, we set @xmath91 $ ] and @xmath103 $ ] .
finally , in the third stage , we fill in @xmath92 = t [ 11]$ ] , but we can not just set @xmath104 = p [ 11 ] = 2 $ ] because @xmath105 $ ] has length only 11 , so we set @xmath104 = 1 $ ]
. @xmath106 $ ]
it is frustrating that both @xmath28 and the index described in theorem [ thm : main ] take @xmath5 bits , but we use @xmath5 words to build the index . in this section ,
we show how to reduce this workspace to @xmath5 bits also , without increasing the time bound for construction by more than a constant factor .
suppose we divide @xmath68 into blocks of size @xmath107 and modify our algorithm such that , whenever we set a value @xmath71 $ ] , we ensure that each value @xmath108 $ ] in the same block with @xmath109 is at least @xmath71 - i + j$ ] and each value @xmath108 $ ] in the same block with @xmath110 is at least @xmath71 $ ] .
since we would eventually set each such @xmath108 $ ] to a value at least as great during the normal execution of the algorithm , this does not change its correctness , apart from perhaps slowing it down by an @xmath47 factor .
for any two consecutive values @xmath71 $ ] and @xmath100 $ ] in the same block now , however , we have @xmath71 \leq t [ i + 1 ] \leq t [ i ] + 1 $ ] . we can thus store each block by storing its first value and a binary string of length @xmath107 whose bits indicate where the values in the block increase .
therefore , we need a total of only @xmath5 bits to store all the blocks .
notice that , if we increase a value @xmath71 $ ] by more than @xmath107 , we reset the first value @xmath111 $ ] of the block to be @xmath71 - i + h$ ] , set the leading bits of the block to 1s to indicate that the values increase until reaching @xmath71 $ ] , and set the later bits of the block to 0s to indicate that the values remain equal to @xmath71 $ ] until the end of the block .
therefore , we can speed the algorithm up to run in @xmath3 time again , by using a universal table of size @xmath112 to decide how to update blocks when we set values in them .
[ thm : workspace ] given a binary string @xmath28 of length @xmath1 containing @xmath2 runs of 1s , we can build an @xmath5-bit index for constant - time jumbled pattern matching in @xmath3 time using @xmath5 bits of workspace .
in fact , it seems possible to make the algorithm run in @xmath3 time and @xmath5 bits of space even without a universal table , using ac0 operations on words that are available on standard architectures .
this workspace reduction makes little sense for a string as small as our example @xmath43 but , for the sake of argument , suppose we partition our array @xmath68 for it into three blocks of length 4 each .
we keep @xmath84 $ ] , @xmath89 $ ] and @xmath113 $ ] stored explicitly and represent the other entries of @xmath68 implicitly with three 3-bit binary strings @xmath114 , @xmath115 and @xmath116 .
initially we set @xmath84 = t [ 5 ] = t [ 9 ] = 0 $ ] and @xmath117 .
recall from section [ sec : alg ] that we first set @xmath84 = 1 $ ] , the number of 1s from the start to the end of the first run of 1s . at this point , we do not need to change @xmath114 .
we then set @xmath86 = 2 $ ] the number of 1s from the start of the first run of 1s to the end of the second run of 1s by setting @xmath118 : starting from @xmath84 = 1 $ ] , this encodes @xmath69 = t [ 1 ] + 0 = 1 $ ] , @xmath86 = t [ 1 ] + 0 + 1 = 2 $ ] and @xmath119 = t [ 1 ] + 0 + 1 + 0 = 2 $ ] .
next we set @xmath87 = 5 $ ] the number of 1s from the start of the first run of 1s to the end of the third run of 1s by setting @xmath89 = 3 $ ] and @xmath120 : starting from @xmath89 = 3 $ ] , this encodes @xmath121 = t [ 5 ] + 1 = 4 $ ] , @xmath87 = t [ 5 ] + 1 + 1 = 5 $ ] and @xmath122 = t [ 5 ] + 1 + 1 + 0 = 5 $ ] .
continuing like this , we set @xmath88 = 7 $ ] by setting @xmath113 = 5 $ ] and @xmath123 ; set @xmath89 = 4 $ ] and @xmath124 ; etc .
when we are finished this stage , @xmath84 = 1 $ ] , @xmath89 = 4 $ ] and @xmath113 = 6 $ ] , and @xmath125 , @xmath124 and @xmath126 , encoding @xmath93 $ ] . in this case
the final right - to - left and left - to - right passes have no effect , but there are cases ( e.g. , when we do not set any values in a certain block ) when they are still necessary .
amihood amir , alberto apostolico , tirza hirst , gad m landau , noa lewenstein , and liat rozenberg .
algorithms for jumbled indexing , jumbled border and jumbled square on run - length encoded strings . , 656:146159 , 2016 .
amihood amir , timothy m chan , moshe lewenstein , and noa lewenstein . on hardness of jumbled indexing . in _ proceedings of the international colloquium on automata , languages , and programming ( icalp ) _ ,
pages 114125 .
springer , 2014 .
david bremner , timothy m chan , erik d demaine , jeff erickson , ferran hurtado , john iacono , stefan langerman , mihai ptracu , and perouz taslakian .
necklaces , convolutions , and x+ y. , 69(2):294314 , 2014 .
pter burcsi , ferdinando cicalese , gabriele fici , and zsuzsanna liptk . on table arrangements , scrabble freaks , and jumbled pattern matching .
in _ proceedings of the international conference on fun with algorithms ( fun ) _ , pages 89101 .
springer , 2010 .
timothy m chan and moshe lewenstein .
clustered integer 3sum via additive combinatorics . in
_ proceedings of the forty - seventh annual acm on symposium on theory of computing ( stoc ) _ , pages 3140 .
acm , 2015 .
ferdinando cicalese , travis gagie , emanuele giaquinta , eduardo sany laber , zsuzsanna liptk , romeo rizzi , and alexandru i tomescu .
indexes for jumbled pattern matching in strings , trees and graphs . in _ proceedings of the international symposium on string processing and information retrieval ( spire ) _ , pages 5663 .
springer , 2013 .
ferdinando cicalese , eduardo laber , oren weimann , and raphael yuster .
near linear time construction of an approximate index for all maximum consecutive sub - sums of a sequence . in _ proceedings of the annual symposium on combinatorial
pattern matching ( cpm ) _ , pages 149158 .
springer , 2012 .
stephane durocher , robert fraser , travis gagie , debajyoti mondal , matthew skala , and sharma v thankachan .
indexed geometric jumbled pattern matching . in _ proceedings of the annual symposium on combinatorial pattern matching ( cpm ) _ ,
pages 110119 .
springer , 2014 .
tomasz kociumaka , jakub radoszewski , and wojciech rytter .
efficient indexes for jumbled pattern matching with constant - sized alphabet . in
_ proceedings of the european symposium on algorithms ( esa ) _ , pages 625636 .
springer , 2013 .
|
important papers have appeared recently on the problem of indexing binary strings for jumbled pattern matching , and further lowering the time bounds in terms of the input size would now be a breakthrough with broad implications .
we can still make progress on the problem , however , by considering other natural parameters .
badkobeh et al .
( ipl , 2013 ) and amir et al . ( tcs , 2016 ) gave algorithms that index a binary string in @xmath0 time , where @xmath1 is the length and @xmath2 is the number of runs , and giaquinta and grabowski ( ipl , 2013 ) gave one that runs in @xmath3 time .
in this paper we propose a new and very simple algorithm that also runs in @xmath4 time and can be extended either so that the index returns the position of a match ( if there is one ) , or so that the algorithm uses only @xmath5 bits of space .
|
the interaction of quantum systems with its immediate environment leads to decoherence or a decay of quantum phase coherences that destroys the delicate interference effects among its quantum states . as a consequence , macroscopic quantum superpositions
are strongly suppressed and classical behavior emerges from the quantum regime @xcite .
the physics at this borderline between the classical and quantum regime is not well understood yet and continues to attract research attention @xcite .
the quantum systems coupled to the environment lose their coherence exponentially fast @xcite and this is modelled by the decoherence factor of the form @xmath0 , where @xmath1 is the coherence time that generally depends on the system parameters including the strength of coupling to the environment . in many applications ,
it is necessary to sustain the quantum coherences for as long as possible and this might be possible by tuning @xmath1 .
one exciting area of application is to perform experimental quantum computation which requires quantum registers to maintain coherences and entanglements . in the last two decades , it has become possible to engineer suitable quantum reservoirs ( environments ) in experiments with desired control over the coupling to the system of interest @xcite .
this has allowed direct observation of decoherence dynamics in a mesoscopic system of harmonically trapped single atom coupled to a reservoir of random electric fields @xcite as also in an atom - optics based kicked rotor system @xcite . within the framework of exponential decoherence factors , these experiments tune the coherence time @xmath1 by controlling a system parameter or the coupling to the environment .
even so , exponential decay of coherences leads to classical regime too quickly .
however , recently it was shown that in principle it should be possible to introduce power law type decoherence factors to achieve relatively slow and controlled decoherence @xcite .
an unusual feature of the non - exponential decoherence factor is that it encompasses a regime in which the system might not complete its transition to the classical regime .
exploring the consequences of non - exponential decoherence might also lead to greater degree of control over decoherence .
non - stationary noise in a dissipative quantum two - level system also exhibits non - exponential decoherence @xcite . in this work ,
we present an atom - optics based experimental realization of the non - exponential decoherence using timing noise in the quantum kicked rotor system . kicked rotor ,
a particle periodically kicked by an external field , is a fundamental model of hamiltonian chaos @xcite .
classically it is a chaotic system and displays diffusive growth of mean energy whereas its quantized version suppresses the energy growth .
this is termed dynamical localization ( dl ) in analogy with the anderson localization in condensed matter physics @xcite .
dl has been experimentally realized using cold atoms in pulsed optical lattices @xcite . with its unambiguous and distinct signatures of energy growth in the classical regime and energy saturation in the quantum regime ,
atom - optics kicked rotor ( aokr ) system is a suitable test bed to study decoherence . in aokr , decoherence can be induced by spontaneous emission of the atoms @xcite , through noise in the amplitude of the periodic kicks in aokr @xcite , by the noise in the periodicity of the kicks @xcite and in the phase of the periodic kicks @xcite .
in contrast with these earlier studies , in this work , we induce decoherence in aokr by introducing a noisy scenario in which kicks do not act at certain time instants and the waiting time in between successive kicks taken from lvy distribution @xmath2 . as we show below , both through theory and experiments , this scenario leads to non - exponential decoherence rates and quantum sub - diffusive energy growth .
this scenario has not been realized in experiments before . aside from the questions related to the decoherence in quantum systems
, there is also considerable current interest in the transport and diffusion properties of chaotic quantum systems @xcite and disordered nonlinear hamiltonian lattices @xcite .
it has been shown that , in a periodically driven system relevant for cold atoms and optical lattices based experiments , long lasting exponential diffusion is possible in momentum basis @xcite .
though subdiffusion is generically not expected in a hamiltonian system , recently it was shown that under the effect of certain non - analytic potentials , it is possible to observe sub - diffusive energy growth in kicked rotor type of system @xcite .
thus , the results reported here also provide new insights related to energy diffusion in chaotic quantum systems .
we use the experimental system consisting of two level atoms in a pulsed laser field @xcite .
the hamiltonian for the dynamics of the center of mass of the atom , treated as a point particle , in a pulsed standing wave of near - resonant light of frequency @xmath3 can be written as @xmath4 if @xmath5 is the transition frequency of the two level atoms , the detuning is @xmath6 . in this
, @xmath7 is the wave number , @xmath8 is the period of pulsing , @xmath9 represents the square pulse of duration @xmath10 centred at @xmath11 and @xmath12 is the amplitude of kick with @xmath13 being the resonant rabi frequency . by appropriate rescaling
, we can obtain the dimensionless hamiltonian as @xmath14 where the scaled variables are , @xmath15 further , @xmath16 is the square pulse of unit amplitude and duration @xmath17 , and scaled kick strength is @xmath18 , recoil frequency is @xmath19 such that @xmath20 .
the hamiltonian in eq .
[ sham ] is for the atom - optics realization of kicked rotor with kicks imparted by square pulses in contrast to the standard kicked rotor with @xmath21-kicks @xcite .
this distinction leads to effective chaos parameter becoming @xmath22 , dependent on the momentum of the atom .
the qualitative classical dynamics of this system , _
i.e. _ , whether it is chaotic or regular or mixed , is determined by the scaled chaos parameter @xmath23 . for @xmath24 ,
the system is integrable and it is possible find an exact transformation to action - angle variables . if @xmath25 , the dynamics is typically mixed as the regular and chaotic regions coexist in phase space . when @xmath26 , the system becomes increasingly chaotic . for values of @xmath27 used in this work ,
the classical phase space is largely chaotic for the energy range that is experimentally accessible . for the quantum analogue of this system ,
the scaled planck s constant is @xmath28 such that the commutation relation is given by @xmath29 = i \hbar_s$ ] .
the behavior of kicked rotor in the chaotic regime for large @xmath23 is distinct in the classical and quantum regimes .
classically , the mean energy growth is diffusive and is linear in time while in the quantum domain energy growth is arrested due to dynamical localization @xcite .
as pointed out before , most of the earlier works that consider decoherence in aokr apply the kicking potential periodically and the noise affects either the amplitude or the phase between kicks @xcite .
in contrast , in this work , we study a noisy system in which kicks do not act at certain times .
as shown schematically in fig [ fig1](a ) , the waiting time between successive kicks is drawn from the lvy distribution @xmath30 which asymptotically decays as @xmath31 .
the number of kicks actually imparted in any finite time interval is a random variable and its mean is dependent on the value of exponent @xmath32 .
for @xmath33 , the mean time interval diverges .
properties of lvy distribution relevant for our purposes is reviewed in ref .
the experimental sequence consists of creating an ultracold cloud of @xmath34rb in an optical dipole trap and then subjecting the cloud to a periodically switched 1-d optical lattice @xcite . after the application of the pulses ,
the momentum distribution of the atoms is measured by absorption imaging of the cloud .
we load a standard magneto - optic trap ( mot ) with @xmath34rb atoms emerging out of the zeeman slower @xcite .
we load about 1 @xmath35 10@xmath36 atoms in 3 s in the mot .
the temperature of the cloud of atoms at this stage is @xmath37 250 @xmath38k .
we cool these atoms further down to @xmath37 30 @xmath38k by compressing the mot .
the primary mot laser beams operate on the @xmath39 transition , and the repumping transition operates on the @xmath40 transition .
the detuning of the laser beams is ramped from -3 @xmath41 down to -12 @xmath41 while the magnetic field gradient is changed from 20 g / cm to 22 g / cm ( where @xmath41 is 2@xmath42 @xmath35 6.1 mhz , the natural linewidth of the transition ) . at this stage
the cold sample of @xmath34rb atoms is transferred into a far - off - resonant crossed optical dipole - trap formed by intersecting two focused laser beams into the trap center .
these two dipole laser beams lie in the horizontal plane and are focused to get a gaussian beam - waist of 100 @xmath38 m in each beam .
the intensities are such that we achieve a trap depth of @xmath37 50 @xmath38k in the intersection region .
the primary laser source for the dipole beams is a 1064 nm , 20 w fiber - laser .
this laser beam is divided into equal parts and sent through two independent acousto - optic modulators ( aoms ) for intensity control and switching of the laser beams .
transfer of atoms from mot into the dipole trap is achieved in the following manner .
the dipole beams and the mot laser beams are switched on simultaneously during the start of the experimental sequence .
initial power in the dipole beams is set to 2 w in each beams . during the first 3 s of loading of the mot , the power in the dipole trap laser beams
is ramped up to a maximum of 5.2 w in both beams . after the compression sequence that lasts for 40 ms ,
the repumper laser operating on the @xmath40 transition is switched off briefly for a period of 1 ms .
this allows all the atoms to be pumped into the @xmath43 ground state .
this is done since , the loss rate due to two - body collision in this state is lower than what it would be if the atoms were in @xmath44 state @xcite .
the temperature and density of atoms at this stage in the dipole trap are @xmath37 30 @xmath38k and 10@xmath45 atoms / cc .
these atoms are further cooled down by forced evaporative cooling down to @xmath37 3@xmath38k in 3 s by lowering the intensity of the trapping beams .
the final power in the trapping beams is 100 mw at this stage .
the temperature and density of atoms is probed by standard absorption imaging after a certain time - of - flight ( tof ) imaging .
aokr system is simulated in the following manner : the periodic optical potential needed for the aokr system is generated by a 1-d standing wave created from a laser operating near 780 nm and detuned by 6 ghz from the @xmath40 transition .
the laser beam is generated from the same laser source that is used to trap atoms in the mot .
this laser beam is passed through an aom and sent to the experiment chamber through single mode optical fibers .
the intensity and the pulse - width and duty cycle of the light pulses are computer controlled by modulating the rf drive power driving the aoms .
the 1-d standing wave is created in the horizontal plane and it goes through the center of the intersection of the dipole trap .
the beam waist size of the lattice laser beam is 700 @xmath38 m
. vertical position of this 1-d lattice beam is lower by less than a beam waist size from the trap center .
the minimum pulse width that can be achieved in our experiment is @xmath37 40 ns .
we observe the localization of the momentum distribution in our system when we keep the period between the pulses of the standing wave constant as observed in ref .
for the results presented in this work , we modify the period between the pulses such that the distributions of the off - times between the pulses follows a lvi distribution .
the exponent @xmath32 of the power law @xmath46 describing the waiting time distribution between pulses is chosen to be 0.25 , 0.5 and 0.75 and 2.0 for this work .
this distribution is first generated numerically and fed into the pulse generator during the experimental sequence .
the on - time is set to be 220 ns and the intensity of the laser beams is set such that the kick strength is @xmath47 and @xmath48 .
the estimated uncertainty in the @xmath23 is 10 % .
after the application of the pulses with a waiting time of 4 ms , the density distribution of the cloud is measured . from the measured density distribution and the tof ,
the momentum distribution is calculated .
> m3 in > m3 in & + & +
in fig.[fig1 ] , we illustrate the central results of this work for the case of @xmath49 .
figure [ fig1](a ) displays the schematic of the position of kicks on the time axis for several values of @xmath32 .
note that the kicks are all of equal strength but do not act at certain times as dictated by the lvy distributed waiting times between successive kicks . for @xmath33 , the mean waiting time diverges in this regime . on the other hand , for @xmath50
relatively fewer kicks are skipped . the classical stroboscopic section ( fig .
[ fig1](b ) ) obtained by simulations for the aokr in eq .
[ sham ] with waiting time exponent @xmath51 shows that the phase space is predominantly chaotic for momenta up to @xmath52 beyond which resonances exist as a consequnce of the square pulse used for imparting kicks . corresponding classical energy displayed as solid line in fig .
[ fig1](c ) is consistent with the quasi - linear growth proportional to @xmath53 . on the other hand , the experimental result shown in fig .
[ fig1](c ) for @xmath49 displays sub - diffusive energy growth in good agreement with the theoretically obtained result @xmath54 , where @xmath55 and @xmath56 are constants .
when @xmath57 , the long time asymptote behaves as @xmath58 . for comparison purposes , we also show the experimental result for the standard case of periodic kicks which displays saturation of energy growth , i.e. , dynamical localization , beyond the break - time @xmath59 .
the experimentally measured momentum profile and corresponding quantum simulations are shown in fig . [ fig1](d ) . beyond the break time , the width of the momentum profiles increase much slower than for a classical evolution indicating slow spread over the momentum basis states .
the numerical simulations of the aokr system closely matching the experimental parameters ( discussed in section 2 ) were performed by repeated application of the floquet operator , @xmath60 ( written down explicitly in eq .
[ floq ] ) on an initial state @xmath61 taken to be of gaussian form .
the simulation results for @xmath51 ( solid lines in fig . [ fig1](c ) ) , averaged over an ensemble of initial states with random phases , also display sub - diffusive energy growth in agreement with the experimental data .
a similar agreement is also seen between the numerical simulations and experimental data for the case of periodic kicks .
uniformly distributed random waiting times results in complete loss of dynamical localization due to exponential decoherence to the corresponding classical limit . however , in contrast , for kicks with lvy distributed waiting times applied in this experiment sub - diffusive quantum mean energy growth results due to controlled decoherence . by tuning the value of @xmath32 , desired degree of subdiffusion can be realized . in a related theoretical work on
periodically kicked rotor with amplitude noise applied as per the lvy distributed waiting times @xcite , subdiffusive behaviour resulting from controlled decoherence was noted earlier . the starting point for theoretical analysis
is the modified delta - kicked rotor hamiltonian of the form @xmath62 where @xmath63 is a telegraph stochastic process that randomly switches between 0 and 1 . in this work ,
we take the waiting time between successive occurrences of 0 from lvy distribution whose asymptotic form is @xmath64 . for @xmath50 ,
it is a stationary stochastic process and if @xmath65 it is a non - stationary process with diverging mean waiting times .
if @xmath66 , then no kick acts at that instant and if @xmath67 then a kick of amplitude @xmath23 acts at that time instant .
the floquet operator corresponding to this system can be written as , @xmath68 where @xmath69 .
note that if @xmath67 for all times , then this reduces to the floquet operator for the standard kicked rotor and is given by , @xmath70 in refs .
@xcite , kicks act periodically and the amplitude noise is induced randomly at certain time instants dictated by the lvy waiting time distribution .
the strength of amplitude noise in refs .
@xcite falls within the range of intensity noise of lasers employed to create the standing waves .
experimental realization of the work in refs .
@xcite will require laser intensity fluctuations to be substantially reduced .
on the other hand , the experimental proposal in this work is to skip kicks entirely at certain time instants . in comparison to the former , this realization represents a much stronger perturbation to the standard atom - optics kicked rotor .
hence , from this perspective , it is surprising that notwithstanding the strong perturbation in our experimental realization , sub - diffusive energy growth and non - exponential decoherence can be realized . .
( a ) experimental data ( symbols ) and simulations of aokr ( lines ) , ( b ) experimental data ( symbols ) and theoretical fits to data ( lines ) from eqs .
[ varp1]-[varp2].,title="fig:",width=307 ] .
( a ) experimental data ( symbols ) and simulations of aokr ( lines ) , ( b ) experimental data ( symbols ) and theoretical fits to data ( lines ) from eqs .
[ varp1]-[varp2].,title="fig:",width=307 ] following the approach given in ref . @xcite and its extension to non - perturbative regime in ref .
@xcite , we relate the decoherence factor to the survival probability of quasi - energy eigenstates @xmath71 of the floquet operator in eq .
[ floq0 ] to obtain an expression for @xmath72 .
an important ingredient is the use of random phase approximation to estimate the contribution @xmath73 when the quasi - energy state responds to noise by transiting from state @xmath71 to @xmath74 . for our noise scenario ,
we obtain the decoherence factor for the quantum evolution from initial time @xmath75 until @xmath76 to be , @xmath77 @xmath78 where @xmath79 is the mean waiting time between kicks , the function @xmath80 is given by , @xmath81 and @xmath82 is the mittag - leffler function @xcite .
the decoherence factor is non - exponential though not necessarily a power - law for @xmath83 and is exponential for @xmath84 .
variation of mean energy @xmath85 with time is obtained by integrating the force - force correlations @xmath86 which is related to the decoherence factor @xcite . using the results in eqs .
[ decf1]-[decf2 ] , for @xmath87 , @xmath88 and @xmath89 , we obtain the mean energy growth as @xmath90 and for @xmath87 , @xmath50 and @xmath57 , we get @xmath91 where the constants @xmath92 and @xmath93 depend on the break - time @xmath59 for the corresponding standard kicked rotor system , mean waiting time @xmath79 , scaled planck s constant @xmath94 and @xmath32 .
these results are in good agreement with the experimental data presented in fig .
[ fig1 ] and in the figures dispayed further below . in order to compare the analytical results for @xmath85 with the experimental data , we use @xmath92 and @xmath93 as fitting parameter . at time ( a ) @xmath95 and ( b ) @xmath96 .
lines are momentum profiles obtained from simulations.,title="fig:",width=307 ] at time ( a ) @xmath95 and ( b ) @xmath96 .
lines are momentum profiles obtained from simulations.,title="fig:",width=307 ] a broader picture of the results illustrated in fig .
[ fig1 ] is shown in fig .
[ fig2](a ) that displays the experimentally measured mean energy growth for values of lvy exponent @xmath32=0.25 , 0.5 , 0.75 and 2.0 .
the quantum simulations performed for the aokr system using the experimental parameters displays a good agreement with measurements . in fig .
[ fig2](b ) , the experimental results ( same data as in fig .
[ fig2](a ) ) are compared with the analytical results in eq .
[ varp1]-[varp2 ] and we obtain a good agreement between the two .
each data point is a result of an average of at least five individual experiments for a given number of kicks . in all these cases ,
the quantum energy growth is sub - diffusive for @xmath33 and differs considerably from the normal diffusion exhibited by the corresponding classical system shown as black line in fig .
[ fig2](a , b ) .
in particular , the subdiffusion is highly pronounced in the regime of @xmath88 in which the decoherence is non - exponential and the mean waiting time between kicks diverges as well .
the classical stroboscopic plots ( not shown here ) have predominantly chaotic features and resonances appear for energies that are far outside the range of energies accessed by these experiments .
it is indeed surprising that the aokr system feels the lvy distributed waiting times in about 30 - 40 kicks and clearly distinguishes it from uniformly distributed waiting times through the different energy growth profiles .
the momentum profiles at times @xmath95 and @xmath97 are shown in fig .
[ fig3 ] for @xmath32 = 0.25 , 0.5 , 0.75 and 2.0 .
the momentum profiles obtained from aokr simulations ( lines ) have been superimposed on the experimentally determined data ( symbols ) .
we see a good agreement between the two . as @xmath32 increases ,
the width of the momentum distribution increases and is consistent with the results displayed in fig .
[ fig2](a , b ) .
the momentum profiles are well approximated by gaussian function , as would be expected in an experiment with spontaneous emission even as small as less than 1% @xcite .
the evolving momentum profiles have widths that are much smaller than the corresponding classical system .
the initial gaussian state encodes a specific realization of random phase information and time evolution under the action of lvy kicked rotor system leads to slow decay of these phase relations in comparison with the case for uniformly distributed random waiting times .
in this work , we have presented an experimental realization of a noisy atom - optics kicked rotor system in which the waiting times @xmath98 between subsequent kicks are chosen from lvy distribution whose asymptotic form is @xmath46 , where @xmath32 is the exponent . in all the earlier experiments on decoherence of quantum systems , the decoherence factor had an exponential decay in time . in the experimental results presented here , we have shown that it is possible to access the regime of non - exponential decoherence factors using atom - optics kicked rotor as a model system .
an atom - optics kicked rotor which is periodically kicked displays dynamical localization in the momentum basis and entirely suppresses the energy growth . in general ,
if we skip kicks randomly in time , we expect that the dynamical localization , a delicate destructive quantum interference effect , will be destroyed by decoherence effects .
this indeed happens if the waiting time between kicks is taken from a uniform distribution of a random variable . the experimental data in fig .
[ fig2 ] shows that in the presence of uniformly distributed timing noise , the quantum mean energy growth coincides with that of the corresponding classical system .
however , if the waiting times between kicks are lvy distributed , then the resulting random kick sequence does not lead to exponentially fast decoherence .
hence the quantum mean energy growth is partially inhibited . as pointed out earlier , the aokr system can clearly distinguish between lvy distribution and uniform distribution of waiting times in few tens of kicks , which makes this experiment feasible . in the regime where @xmath99 , the mean waiting time diverges and the decoherence factor is in general non - exponential . in this case , the quantum evolution of an arbitrary initial wavepacket is sub - diffusive and carries the signature of the lvy exponent .
in contrast , when @xmath50 , the quantum diffusion is well approximated to be linear in time . as figures [ fig2 ] and [ fig3 ] show , there is a good agreement between the experimental results , analytical results and the ab initio simulations of the quantum atom - optics kicked rotor . the sub - diffusive energy growth is also reflected in the slower increase of the width of momentum profiles
. it will be interesting to explore generic conditions under which non - exponential decoherence can be realised in experimental situations .
the authors would like to thank the department of science and technology , govt . of india through grants from emr/2014/000365 and nano mission .
sp would like to acknowledge university grants commission ( ugc ) and ms would like to acknowledge the council of scientific and industrial research ( csir ) , govt .
of india for research fellowship .
kienzler , d. , lo , h .- y . ,
keitch , b. , de clercq , l. , leupold , f. , lindenfelser , f. , marinelli , m. , negnevitsky , v. , and home , j. p. ( 2015 ) .
quantum harmonic oscillator state synthesis by reservoir engineering . _ science _ , * 347*(6217 ) , 5356 .
b asjad , m. and vitali , d. ( 2014 ) .
reservoir engineering of a mechanical resonator : generating a macroscopic superposition state and monitoring its decoherence .
_ journal of physics b : atomic , molecular and optical physics _ , * 47*(4 ) , 045502 .
myatt , c. j. , king , b. e. , turchette , q. a. , sackett , c. a. , kielpinski , d. , itano , w. m. , monroe , c. , and wineland , d. j. ( 2000 ) .
decoherence of quantum superpositions through coupling to engineered reservoirs .
_ nature _ , * 403 * , 269 273 .
moore , f. l. , robinson , j. c. , bharucha , c. , williams , p. e. , and raizen , m. g. ( 1994 ) .
observation of dynamical localization in atomic momentum transfer : a new testing ground for quantum chaos .
_ , * 73 * , 29742977 .
hensinger , w. k. , heckenberg , n. r. , milburn , g. j. , and rubinsztein - dunlop , h. ( 2003 ) .
experimental tests of quantum nonlinear dynamics in atom optics .
_ journal of optics b : quantum and semiclassical optics _ , *
5*(2 ) , r83 .
doherty , a. , vant , k. m. d. , h.ball , g. , christensen , n. , and leonhardt , r. ( 2000 ) .
momentum distributions for the quantum @xmath100kicked rotor with decoherence .
b : quantum semiclass . opt .
_ , * 2 * , 605 .
kumar , s. , sarkar , s. , verma , g. , vishwakarma , c. , noaman , m. , and rapol , u. ( 2015 ) .
bose - einstein condensation in an electro - pneumatically transformed quadrupole - ioffe magnetic trap .
_ new journal of physics _ , * 17*(2 ) , 023062 .
|
quantum systems upon interaction with the environment undergo decoherence or decay towards classical states .
the dynamics of the quantum coherence at the borderline between quantum and classical world is not fully understood yet . the quantum coherence is generally known to exponentially decay in time so that macroscopic quantum superpositions are not sustained for a long time . in many practical applications , it is necessary to maintain quantum coherences for as long as possible . in this work ,
we use the experimental test - bed of cold atoms in a pulsed one - dimensional optical lattice to realise an atom - optics kicked rotator .
we manipulate the waiting time between consequent kicks in this system by taking it as a random variable from a lvy distribution . in this noisy scenario , we show that coherence can be made to decay slower than an exponential form .
this manifests in the form of quantum subdiffusion that can be controlled through the lvy exponent .
the experimental results are in good agreement with the analytical expressions and numerical simulations for the mean energy growth of atom - optics kicked rotor .
|
optically thick pair plasma with baryon loading is assumed to power grbs in many models considered in the literature , see , e.g. , @xcite , @xcite , @xcite .
such plasma is self accelerated to large bulk lorentz factors due to initial energy dominance . due to relativistic kinematics in laboratory reference
frame it forms a shell with width approximately equal to its initial size .
when the plasma becomes transparent to compton scattering a flash of radiation is emitted . in the literature such emission is often associated with so - called precursors @xcite . in the fireshell model @xcite this emission
is called the proper grb ( p - grb ) . for typical baryonic loading parameters of observed grbs @xmath3 , where @xmath4 , @xmath5 is total baryonic rest mass energy
, @xmath6 is total energy released in the source of grb , the fraction of energy emitted at transparency can reach several percents of the total energy . considering
that typical duration of long grbs is of the order of hundreds of seconds , and p - grb typically lasts for less than few seconds @xcite luminosities of both events are comparable in magnitude . early treatments assuming the absence of baryons and thin shell approximation , and non instantaneous energy release @xcite produced estimations of p - grbs duration @xmath7 sec for the progenitor mass range @xmath8 , much larger than the naive estimation of the gravitational collapse time @xmath9 sec , where @xmath10 is newton s constant , @xmath11 is black hole mass , @xmath12 is the speed of light .
observed durations of p - grbs @xcite of most energetic grbs is of the order of several seconds .
main purpose of this paper is to outline a possibility of resolution of this tension considering different mechanisms of spreading of ultrarelativistically expanding shell .
the paper is organized as follows . in sec .
2 the hydrodynamic phase of grb is discussed , hydrodynamical spreading mechanism is recalled , and results of its application to grbs are shown . in sec .
3 the concept of thermal spreading is introduced , evaluation of spreading for relativistic maxwellian distribution function is performed , and the results of thermal spreading computation are applied to grbs .
discussion of the spreading effects on the duration of p - grb and conclusions follow in the last section .
energy , entropy and baryonic number conservation laws for each differential subshell of ultrarelativistically expanding shell imply @xcite @xmath13where @xmath14 and @xmath15 are respectively comoving energy density and pressure , @xmath16 and @xmath17 are respectively comoving entropy and comoving baryon number density , @xmath18 is lorentz factor , and @xmath19 is laboratory radial coordinate . assuming polytropic equation of state @xmath20 where @xmath21 is the thermal index , equations ( 1 )
can be solved for each differential subshell in ultrarelativistic and nonrelativistic asymptotic cases , respectively . recall that expanding plasma consists of two components : nonrelativistic baryons and ultrarelativistic electron - positron pairs and photons .
hence , in analogy with cosmology , see e.g. @xcite , these two limiting cases are characterized by two adiabatic indices , respectively : @xmath22 when comoving temperature is relativistic ( energy dominated phase ) , and @xmath23 when comoving temperature is nonrelativistic ( matter dominated phase ) . in the latter case the pressure is usually neglected , see e.g. @xcite . in what follows
we assume @xmath24 at the matter dominated phase .
therefore one has @xmath25 during the energy dominated phase and @xmath26 during the matter dominated phase .
these scaling laws are different from the ones in cosmology , see @xcite for details and comparison . the transition from energy to matter domination occurs at @xmath27 , where @xmath28 is initial size of the energy dominated region .
in @xcite we computed the optical depth and dynamics of p - grb for the case when @xmath29 is constant across the shell .
general expression for the optical depth of the shell is @xmath30 where @xmath31 is thomson s cross section and @xmath32 is the radius at which light ray emitted from the inner boundary of the shell at radius @xmath33 crosses its outer boundary . for large baryonic loading with @xmath34 where @xmath35 cm , @xmath36 erg , optical depth of the shell is @xmath37 where @xmath38 , @xmath39 is given by equation ( [ n0 ] ) with @xmath40 .
equation ( [ tauthin ] ) correspond to the photon thin asymptotics @xcite of equation ( [ tau ] ) defined by the condition @xmath41 .
the observed duration of p - grb is determined in this asymptotics by the process of radiative diffusion @xcite @xmath42 where @xmath43 and time of diffusion @xmath44 was found by solution of radiative transfer equation .
all these results are derived assuming constant lorentz factor across the shell .
hence the width of the shell remains constant during its expansion .
this fact has been used within the fireshell model and termed as the `` constant thickness approximation '' , provided that the baryon loading is not too heavy @xmath45 .
hydrodynamic analytical ( @xcite , @xcite ) and numerical ( @xcite , @xcite ) calculations show that lorentz factor gradient is developing during shell expansion that leads to spreading of the shell at sufficiently large radii at the matter dominated phase . at this phase of expansion each differential subshell is expanding with almost constant speed @xmath46 , so the spreading of the shell is determined by the radial dependence of the lorentz factor @xmath47 . in a variable shell
there can be regions with @xmath47 decreasing with radius and @xmath47 increasing with radius . at sufficiently large radii only the regions with increasing @xmath29 contribute to the spreading of the shell . from equations of motion of external and internal boundaries of this region
we obtain @xcite the thickness of the region as function of radial position of the region @xmath48 where @xmath49 and @xmath50 are lorentz factors at external and internal boundaries , @xmath51 is the width of the region at small @xmath33 , @xmath33 being the radial position of inner boundary . let us consider such a region in two limiting cases : * when relative lorentz factor difference is strong , + @xmath52 ; * when relative lorentz factor difference is weak , + @xmath53 . in the case
( a ) the second term in parenthesis in equation ( [ lr ] ) can be neglected , and we obtain that the spreading becomes efficient at @xmath54 , see @xcite and @xcite . in the case ( b ) we find the corresponding critical radius of hydrodynamical spreading @xmath55 . from equation ( [ lr ] )
one can see that in both cases for @xmath56 width of the shell is increasing linearly with radius @xmath57 . the discussion above corresponds to the case ( b ) since for weak lorentz factor difference @xmath58 .
in what follows we focus on the case ( a ) and derive corresponding relations assuming strong relative lorentz factor difference across the shell .
let us take an element of fluid with constant number of particles @xmath59 in the part of the shell with gradient of @xmath29 .
internal boundary of the element is moving with velocity @xmath60 , and external one is moving with velocity @xmath61 , where @xmath62 is the differential thickness at some fixed laboratory time @xmath63 and derivative @xmath64 is taken at the same laboratory time
. then at time @xmath65 the width of the element is @xmath66 , its radial position is @xmath67 , where @xmath68 ia initial radial position of the element , and corresponding laboratory density is @xmath69 where @xmath70 at large enough @xmath65 using @xmath71 we have in contrast with ( [ nrscalings ] ) @xmath72 in order to compute the integral ( [ tau ] ) we need to find the expression for baryonic number density along the light ray .
taking into account hydrodynamical spreading ( [ lr ] ) we obtain @xmath73 that is exact in ultrarelativistic limit .
notice the difference between equations ( [ nspralongtheray ] ) and ( [ nspr ] ) : in the former case @xmath74 is computed along the light ray , while in the latter case @xmath75 is computed along the radial coordinate at fixed laboratory time .
the expression ( [ nspralongtheray ] ) reduces to equation ( [ nrscalings ] ) when @xmath76 .
instead when the second term in the denominator of the expression ( [ nspralongtheray ] ) dominates , namely when @xmath77 , density radial dependence coincides with the one given by relations ( [ scalingsfaraway ] ) .
an estimate for @xmath74 can be given for strong relative lorentz factor difference @xmath78 in the shell @xmath79 where @xmath80 is the distance inside the shell along the light ray .
numerical results from @xcite , @xcite and analytical ones from @xcite support this estimate . integrating expression ( [ dgdr ] )
we obtain lorentz factor dependence on radial coordinate along the light ray @xmath81 since we are interested in the asymptotics when the hydrodynamical spreading is essential , we can assume in the integral ( [ tau ] ) @xmath82 and @xmath83 . under these conditions
the optical depth is @xmath84 this result coincides with equation ( [ tauthin ] ) up to a numerical factor .
however its physical meaning is different .
it represents photon thick asymptotics of equation ( [ tau ] ) , since @xmath83 @xcite .
transparency radius is defined by equating ( [ tauspreading ] ) to unity and is given by @xmath85 where @xmath86 is proton mass . at the radius of transparency the width of the shell ( [ lr ] ) spreads up to @xmath87 where @xmath88 . in this case
radiative diffusion is irrelevant and duration of p - grb is then given by the time of arrival of photons emitted from the shell all the way up to @xmath89 @xmath90 in contrast to equation ( [ tathin ] ) in the case of strong lorentz factor difference , namely with @xmath91 , this duration can be of the order of several seconds with agreement with observations .
we now determine the velocity spread of particles as a function of comoving temperature @xmath92 and bulk lorentz factor @xmath29 for relativistic maxwellian distribution . based on this result
we compute the value of thermal spreading of expanding shell .
the distribution of particles in the momentum space in the laboratory frame is a lorentz - boosted maxwellian @xmath93^{1/2}\biggr ) , \end{gathered}\ ] ] where we assumed that the relative motion of the frames is along their @xmath94-axes and @xmath95 is dimensionless comoving temperature .
velocity dispersion in the @xmath94-direction is @xmath96 where @xmath97 denotes the average value of @xmath98 , which is defined by the convolution with the distribution function ( [ maxwellboost ] ) @xmath99 the above written integrals can not be computed analytically , but their numerical approximations can be found .
numerical issues in the velocity dispersion calculations by ( [ dispersion ] ) arise from the fact that for high @xmath29 we need to subtract two numbers @xmath100 and @xmath101 which are very close to each other and to @xmath102 .
this leads to substantial reduction of accuracy . a different formula for dispersion@xmath103^{2 } )
\label{dispersiondef}\ ] ] proves to be more convenient .
the spread of particle velocities is then @xmath104 . for nonrelativistic comoving temperatures the correct asymptotics
is @xmath105 the case of ultrarelativistic comoving temperature ( @xmath106 ) is more interesting .
starting close to the maximal value @xmath107 , the velocity spread for @xmath108 reaches approximately @xmath109 which means that the dispersion is independent on the temperature . for @xmath110
the asymptotics ( [ asymptotics ] ) is restored just up to a multiplier close to unity @xmath111 our results suggest that ( [ urmediumasymptotics ] ) gives absolute upper limit for the velocity spread , and temperature dependence of ( asymptotics ) and ( [ urhighasymptotics ] ) reduce the spread even further . initial temperature of the plasma formed in the source of grb can be estimated from its initial size @xmath112 cm and total energy released @xmath113 assuming that the temperature is determined by @xmath114 pairs only @xmath115 we get for initial temperature @xmath116 at radiation dominated phase the comoving temperature of the plasma decreases as @xmath117 , at matter dominated phase as @xmath118 .
now we compute the thermal spreading at both phases .
for the first phase the reasonable approximation of lorentz factor is @xmath119 due to the nature of lorentz transformations in constantly accelerated frame the final spreading of the shell @xmath120 appears to be finite even if we extend this phase infinitely in time , and the main part of the spreading is connected with initial part of motion with relatively small @xmath29 .
while the temperature of protons in the source of grb is nonrelativistic @xmath121 , velocity spread is given by ( [ asymptotics ] ) which in energy dominated phase leads to the spreading @xmath122 in the matter dominated phase the additional spreading of the shell is @xmath123 when @xmath124 .
comparing to the hydrodynamical spreading for reasonable grb parameters the spreading coming from both ( [ nr1 ] ) and ( [ nr2 ] ) is negligible .
note that velocity dispersion in any case does not exceed the value given by equation ( [ urmediumasymptotics ] ) with @xmath125 @xmath126 which gives an absolute maximum of the thermal spreading on the energy dominated phase .
in this paper we considered two mechanisms of spreading of relativistically expanding plasma shell .
we also discuss their implications for the duration of electromagnetic signal from transparency of the plasma .
firstly , following the proposal of @xcite hydrodynamical spreading of relativistically expanding shell is estimated .
secondly , thermal spreading is considered .
assuming relativistic maxwellian distribution function we determined the velocity dispersion depending on temperature and the lorentz factor of the bulk motion .
we then applied these results to grbs within the framework of the fireshell model .
it is shown that thermal spreading provides negligible spreading for typical parameters of grbs . instead , hydrodynamical spreading results in the increase of the duration of p - grb . for nonspreading shells characterized by almost constant lorentz factor distribution within the shell the duration of p - grb
is determined by the time of diffusion , see equation ( [ tathin ] ) .
if strong lorentz factor difference is present within the shell , hydrodynamical spreading prevents occurrence of photon thin asymptotics and leads to duration of p - grb given by equation ( [ taspr ] ) .
our results imply that for high enough baryon loading and energy of the burst the duration of the p - grb is not determined by the initial size of the plasma @xmath28 , but by the value of the hydrodynamical spreading , reaching up to several seconds . ,
r. , aksenov , a. g. , bernardini , m. g. , bianco , c. l. , caito , l. , chardonnet , p. , dainotti , m. g. , de barros , g. , guida , r. , izzo , l. , patricelli , b. , lemos , l. j. r. , rotondo , m. , hernandez , j. a. r. , vereshchagin , g. , xue , s .- s . ,
may 2009 . the blackholic energy and the canonical gamma - ray burst iv : the `` long , '' `` genuine short '' and `` fake - disguised short '' grbs . in : novello , m. , perez , s. ( eds . ) , american institute of physics conference series .
1132 of american institute of physics conference series .
. 199266 .
|
optically thick energy dominated plasma created in the source of gamma - ray bursts ( grbs ) expands radially with acceleration and forms a shell with constant width measured in the laboratory frame . when strong lorentz factor gradients are present within the shell it is supposed to spread at sufficiently large radii .
there are two possible mechanisms of spreading : hydrodynamical and thermal ones .
we consider both mechanisms evaluating the amount of spreading that occurs during expansion up to the moment when the expanding shell becomes transparent for photons .
we compute the hydrodynamical spreading of an ultrarelativistically expanding shell . in the case of thermal spreading
we compute the velocity spread as a function of two parameters : comoving temperature and bulk lorentz factor of relativistic maxwellian distribution .
based on this result we determine the value of thermal spreading of relativistically expanding shell .
we found that thermal spreading is negligible for typical grb parameters .
instead hydrodynamical spreading appears to be significant , with the shell width reaching @xmath0 cm for total energy @xmath1 erg and baryonic loading @xmath2 . within the fireshell model such spreading will result in the duration of proper gamma - ray bursts up to several seconds .
gamma rays : bursts ; hydrodynamics ; radiative transfer ; relativity
|
the conductance of quantum dots is dominated by their discrete level spectra.@xcite particularly interesting features can be found at degeneracies of the discrete levels.@xcite at zero magnetic field spin - degenerate levels can be modeled with the anderson hamiltonian.@xcite at low temperatures such systems exhibit the intriguing phenomenon of the kondo effect , which has been extensively investigated .
@xcite on the contrary , much less is known about systems with orbital degeneracies .
the simplest system with such degeneracies is the two - level spinless quantum dot , which was studied theoretically in connection with phenomena such as phase lapses of the transmission phase,@xcite charge oscillations , @xcite and correlation induced conduction resonances.@xcite relations between these phenomena were discussed in refs . .
while refs . , investigated the properties of the system in the low - bias limit , the present paper deals with the effects of finite bias and temperature . for typical parameters applicable to experiments , we find that changing the bias might lead to qualitatively very different results .
finite bias was previously investigated in ref . , where the current and full counting statistics for the system were studied , but only to lowest order in lead - dot coupling .
the effect of finite bias and temperature in the co - tunneling regime has thus not previously been investigated .
a two - level quantum dot system , where the two levels were coupled to separate source and drain contacts , were studied in a recent paper.@xcite here the orbital quantum number has the same effect as spin , due to the separate contacts , and orbital kondo effect can be investigated .
the paper studied the level renormalization , kondo temperature , local density of states and conductance of the system .
on the experimental side , the degeneracy of orbital levels was recently studied for a gate - defined quantum dot in an insb nanowire,@xcite where a canyon of conductance suppression was found . here
large level - dependent g - factors enabled the study of degenerate orbital levels using the zeeman effect.@xcite similar systems of quantum dots embedded in nanowires have previously been experimentally realized.@xcite the model is not restricted to the two - level quantum dot but can also be applied to parallel quantum dots coupled to the same leads.@xcite previous theoretical investigations of the equal spin two - level quantum dot showed the existence of a complete conductance suppression around the electron - hole symmetry point in the limit of zero temperature.@xcite in this paper we investigate the current suppression at finite bias and temperature , as well as the current away from electron - hole symmetry , i.e. regimes not previously studied beyond sequential tunneling .
the aim of this paper is to determine whether the canyon of current suppression is generic or appears only for certain parameters . in the regime @xmath1 , where @xmath2 is the level broadening given by the coupling between dot levels and leads
, we find a complete current suppression close to degeneracy , both in the co - tunneling and the sequential tunneling regime , assuming that the two dot states couples to the leads with different parity . for weaker couplings to the leads , or for increased @xmath3 or @xmath4 , only partial current suppression is found .
results are compared with three limiting cases of bias .
i ) the low - bias limit , @xmath5 , is shown in fig .
[ electronholsym](a ) , where @xmath6 is the charging energy .
ii ) the high - bias limit @xmath7 , see fig .
[ electronholsym](b ) , where @xmath8 is the splitting between the two levels , so that both @xmath9 and @xmath10 are entirely inside the bias window .
iii ) the ultra - high - bias in which the bias is the largest parameter @xmath11 , as shown in fig . [ electronholsym](c ) . for the high- and ultra - high - bias limit
the current can be evaluated by analytical means , as first - order approximations in tunneling are valid .
we find that in the high - bias limit , the current suppression is completely independent of the couplings.@xcite in the ultra - high - bias limit , on the other hand , only partial suppression is found when the two states have different coupling strengths to the leads .
the remainder of the article is organized as follows .
the system hamiltonian and the modeling of transport is discussed in sec .
the canyon of current suppression is studied in sec .
iii using the second order von neumann method ( 2vn ) , where co - tunneling and coherence are included.@xcite the findings of this section are further explained in sections iv - vii . in sec .
iv an intuitive explanation for the conductance suppression is given by means of a schrieffer - wolff transformation.@xcite in sec .
v the current at electron - hole symmetry is analyzed as a function of bias , temperature , and couplings .
the current away from electron - hole symmetry is studied in sec .
vi . in sec .
vii focus is changed to the current peaks surrounding the canyon of current suppression . at increased bias
we discover an unexpected population inversion and investigate its importance for the peaks . here
the term population inversion means a higher population of the upper level .
it should be pointed out that this population inversion results from a finite bias and is different from what is studied in refs . .
in this section we discuss the hamiltonian of the spinless two - level quantum dot as well as the modeling of the current through the system . in the single - particle eigenbasis of the dot ,
the system hamiltonian is given by : @xmath12 where we have assumed that the couplings @xmath13 are independent of @xmath14 and @xmath15 with a constant density of states @xmath16 , for @xmath17 . in the simulations
a large bandwidth @xmath18 is used , assuming wide conduction bands of the leads .
the operators @xmath19 ( @xmath20 ) and @xmath21 ( @xmath22 ) are annihilation ( creation ) operators of electrons in the dot and leads , respectively . in eq .
( [ u ] ) the charging energy @xmath6 is due to coulomb interaction between the electrons when both dot states are filled .
the energy levels are parametrized by @xmath23 electron - hole symmetry , where the levels are symmetrically placed around the fermi energy , is given by @xmath24 and @xmath25 . at this point electrons and holes
contribute equally to the current through the quantum dot .
bias is applied symmetrically to the quantum dot so that the chemical potentials are @xmath26 and @xmath27 , for the left and right contact , respectively . in this paper
we mainly study couplings of the type @xmath28 where the asymmetry parameter @xmath29 is chosen to be real .
time - reversal symmetry assures that the coupling elements are real , implying that the two dot states couple to the leads with the same or opposite parity.@xcite this paper focuses on the case of opposite parity as it is essential for the canyon of current suppression at finite bias , see sec .
iii where also the effects of asymmetric couplings to left and right leads are briefly discussed .
furthermore , it was shown in ref .
that good agreement between theory and experiment could be achieved for real couplings . as the broadening of the levels is proportional to the square of the coupling elements we have @xmath30 .
the current through the dot is calculated using the second order von neumann ( 2vn ) method.@xcite this method is an extension of the generalized master equation .
starting from the von neumann equation an equation of motion is derived not only for the reduced density matrix but also for the elements of the total density matrix consisting of a single electron - hole excitation in the contact . in the latter system of equations two electron - hole excitation elements , corresponding to co - tunneling , enter . in the time evolution of the co - tunneling processes
there are contributions from three electron - hole excitation elements but these are ignored .
this results in a closed set of equations , which can be solved for the occupations of the dot levels , the coherences between these resulting from the coupling to the leads , and the current flowing through the quantum dot .
there are currently several other methods that account for tunneling of higher order than sequential processes .
the most widely used technique is the generalized master equation approach which can be derived in many different ways including the real - time diagrammatic technique@xcite and the bloch - redfield approach originally developed in refs . .
comparisons of different approaches have been performed in refs . .
the detailed relation between these methods and the 2vn - method will be the subject of future investigations , where we show that the 2vn - method is similar to other methods of @xmath31 order in the couplings @xmath32 , but contains certain diagrams up to infinite order .
for the anderson model with infinite @xmath6 it has been shown to give the same current as the resonant - tunneling approximation@xcite in the real - time diagrammatic approach.@xcite this method includes irreducible diagrams with an arbitrary number of correlated tunneling processes , but restricted to one electron - hole excitation at any given moment .
we proceed by investigating whether the canyon of current suppression observed in ref . can be found only in a certain parameter space or if it is generic for the studied two - level system .
we plot the normalized current @xmath33 , which equals the conductance @xmath34 in the low - bias limit . as a start , we restrict to couplings given by eq .
( [ couplings ] ) and study two different coupling strengths @xmath35 and @xmath36 . in both cases the relative coupling strength of the two levels is given by @xmath37 .
the reason for using @xmath38 is that we want to study the current suppression in the co - tunneling regime .
the complete suppression found close to degeneracy in the case of @xmath36 is also supported by the analytical arguments of sec .
furthermore the 2vn - method is non - perturbative and contains some classes of diagrams up to infinite order , which makes the analysis in the regime @xmath39 more reliable .
furthermore , both low bias @xmath40 and high bias @xmath41 are studied to investigate the effects of finite bias . in the left part of fig .
[ canyon ] , the results of the 2vn simulations for @xmath42 are shown . in order to study the role of the coulomb interaction , corresponding results for @xmath43
are shown in the right part of fig .
[ canyon ] .
here we applied the method of nonequilibrium green s functions ( negf)@xcite providing identical results to the numerically much more involved 2vn method ( for @xmath43 ) , for the simple system studied here .
we believe that this holds also for more complex @xmath44 @xmath45 systems , see also refs . .
in order to facilitate the comparison , we performed a cut of the noninteracting results at @xmath24 and shifted the upper / lower part by @xmath46 .
this mimics the presence of coulomb repulsion by one entirely filled or completely empty level , as expected in the respective energy range .
plotting the noninteracting results in this way we find indeed rather good agreement with the interacting results in the low bias regime , allowing us to explain the essential features in this region .
for high bias the current suppression is more pronounced in the interacting case .
the coulomb blockade regime present in the left figures is a result of the interaction - related correlations .
this regime can naturally not be studied using non - interacting methods .
the noninteracting results for weak bias , fig .
[ canyon](b , d ) , show a line of current suppression , which is bowed towards the weaker resonance .
this can be understood in terms of interference between the two levels .
the breit - wigner formula provides @xmath47 and the current can then be calculated using @xmath48t(e)de.\end{aligned}\ ] ] vanishing conductance in the low temperature limit corresponds to @xmath49 , i.e. complete conductance suppression can only occur for @xmath50 .
this explains the location of the strong current suppression found in fig .
[ canyon](b ) and ( d ) .
it is easily shown that if one level couples symmetrically to left and right , complete suppression is only found for couplings like eq .
( [ couplings ] ) , where the sign of @xmath51 and @xmath52 differs for @xmath53 .
this motivates the special attention given to these couplings . as the bias
is increased the current not only depends on @xmath54 . when the bias is larger than the width of the dip , @xmath55 at @xmath56 resulting from eq .
( [ breit - wigner ] ) , this results in only partial current suppression as can be seen in fig .
[ canyon](f , h ) .
[ canyon](a ) shows that the weak coupling and low - bias interacting result is well reproduced by the noninteracting result , fig .
[ canyon](b ) , in its region of applicability ( @xmath57 ) , apart from a slight shift of the suppression line . in the coulomb blockade regime ( @xmath58 ) almost no structure is visible due to the very weak co - tunneling . in sec .
v we see that this corresponds to the regime where @xmath4 dominates and only partial conductance suppression can be found .
[ canyon](c ) shows that the 2vn - method can sometimes give unphysical results corresponding to small negative currents .
this is expected as the 2vn method neglects some correlated transitions of more than two electrons / holes .
the contributions of these terms become important in the limit of large couplings and small temperature and bias .
apart from this , the 2vn method agrees well with the noninteracting results
. however , we see that the shift of the suppression line is larger when the couplings are strong . as explained in sec .
vi this can be attributed to the broadening of the levels . in the coulomb blockade regime a clear conductance suppression is found .
this is the regime where @xmath2 dominates , i.e. , the situation described by refs . .
in the high - bias limit , the interacting and noninteracting results differ significantly at the canyon of current suppression , as seen in fig .
[ canyon](e , g ) and fig .
[ canyon](f , h ) , respectively . in the sequential tunneling regime a strong current suppression is found in the interacting case .
this result has been explained by refs . and
is further discussed in sec .
this interaction - induced suppression is not included in the negf - method , resulting in only partial suppression , as can be seen from the gray colored canyon .
the current suppression in the co - tunneling region becomes weaker as we are approaching the region where @xmath3 dominates , which results in only partial current suppression , see sec .
v. this shows that the canyon of current suppression is found in all four cases albeit the current not always entirely drops to zero .
further investigations show that the canyon of suppression is not restricted to couplings like eq .
( [ couplings ] ) , but is present also for couplings which are asymmetric with respect to the left and right contact .
the importance of the relative coupling strength , @xmath29 , is investigated in sec .
v. for couplings where the two levels have the same parity , i.e. the two levels couple with the same sign to left and right lead , the situation is different , as shown in fig .
[ samepar ] . here
we still see the high - bias blockade in the sequential tunneling regime.@xcite for @xmath59 or @xmath60 interactions are of less importance and the blockade disappears .
one also sees that no canyon is visible in the coulomb blockade regime .
the reason is that the width of the dip in the transmission function becomes increasingly narrow as the singular coupling point is approached . at such couplings ,
one can at degeneracy completely decouple one level from the leads , resulting in a bistable system , see also refs . .
as a result , the canyon is present only for very low bias and temperature , i.e. a regime where the 2vn - method is not applicable .
[ samepar ] clearly shows that at finite bias the parity of the couplings is essential for the suppression , unlike the zero - bias limit .
two levels with the same parity will not be studied any further .
the remainder of the paper is devoted to the couplings described by eq .
( [ couplings ] ) . for such couplings ,
details of the canyon of current suppression is summarized in table i , which can be of use for futher experimental studies .
.a summary of the current suppression from fig .
[ canyon ] . for levels with couplings of different parity the current suppression is generic , but not always complete .
the high - bias limit of fig .
[ electronholsym](b ) implies to @xmath61 .
therefore the co - tunneling regime is not applicable ( n.a . ) in this limit . in the ultra - high - bias limit
the current is given by the sequential tunneling result .
the co - tunneling regime is thus not categorized in this limit .
the table also indicates in which sections the different regimes are investigated further .
[ cols="<,^,^ , > " , ]
in this section we study the current in the regime where @xmath62 and @xmath63 , so that the dot is singly occupied .
furthermore we will assume that @xmath64 .
this allows us to derive an effective low - energy hamiltonian by means of a schrieffer - wolff transformation .
low - energy implies that all tunneling events occur around @xmath65 .
the transformation requires that the charging energy of the dot is much larger than the broadening of the dot levels , so that the dot is singly occupied when the levels are placed as in fig .
[ electronholsym](a ) . by analyzing the effective hamiltonian in this regime
we will be able to identify the two - particle processes contributing to the current and give an intuitive explanation of the conductance suppression .
it is convenient to write the hamiltonian in terms of pseudo - spin operators@xcite defined by @xmath66 , @xmath67 , @xmath68 and @xmath69 .
the system hamiltonian then reads @xmath70 here @xmath71 , @xmath72 , @xmath73 , @xmath74 , and the problem has been mapped onto the one - lead anderson model with spin - dependent couplings .
a similar transformation can be found for arbitrary couplings .
we now perform a schrieffer - wolff transformation , i.e. a canonical transformation @xmath75 so that @xmath76 does not have a linear term in @xmath77 .
one finds that the following transformation @xmath78 , with @xmath79 and @xmath80 , will do the job . here
@xmath81 if @xmath82 and @xmath83 if @xmath84 .
writing the tunneling hamiltonian as @xmath85 , with @xmath86 and @xmath87 , the effective low - energy hamiltonian of the singly occupied subspace , when restricting to two - particle tunneling , is given by @xmath88+[s^{+},\hat{h}_{t}^{-}])}_{\hat{h}_{ts}}+o(\hat{h}_t^{3})\end{aligned}\ ] ] where the last term is the effective tunneling hamiltonian @xmath89 . performing the commutations and returning to our original basis
one finds : @xmath90 where @xmath91 indicates that the terms containing @xmath92 , corresponding to the renormalization of energy levels @xmath93 , are not included in the sum . in the limit of zero temperature and zero bias ,
the renormalization reads ( see appendix b ) @xmath94 it should be noted that the above expression does not hold if one of the levels is close to the chemical potential , as the schrieffer - wolff transform breaks down in this regime .
the effect of the renormalization will be further discussed in sec .
v , here we simply note that at electron - hole symmetry , where @xmath95 , the renormalization vanishes . each term in eq .
( [ lowenergy2 ] ) corresponds to a co - tunneling process , where the electron is transferred from one lead to the other , or returns to the original lead .
the first two sums correspond to elastic processes where the state of the dot is unchanged .
the next two sums correspond to inelastic processes where the state of the dot changes .
generally there is an effective overlap matrix element between the two dot states originating from the coupling to the leads . in refs .
it was shown that for couplings of the type eq .
( [ couplings ] ) , this overlap vanishes .
indeed , evaluating the @xmath96 inelastic terms in eq .
( [ lowenergy2 ] ) results in factors @xmath97 in the low - bias limit , see sec .
vii for a discussion concerning finite bias .
let us now analyze eq .
( [ lowenergy2 ] ) in detail : we see that only the inelastic processes affect the occupations of the dot , as @xmath98 and @xmath99 commutes with the other parts of @xmath89 . for @xmath100 the third and fourth sum in eq .
( [ lowenergy2 ] ) , corresponding to the inelastic processes , are symmetric in level 1 and 2 except for a sign change @xmath101 . in the low - bias limit
, this sign change does not affect the kinetics of the occupations and we conclude that both levels are half filled at degeneracy .
the current operator reads @xmath102 $ ] , resulting in @xmath103 ( c_{kr}^{\dagger}c_{k'l}+c_{kl}^{\dagger}c_{k'r } ) \nonumber \\ & + & \mathrm{h.c .
} , \nonumber\end{aligned}\ ] ] where @xmath104 has again been used .
the first row corresponds to elastic processes , while the second row constitutes the inelastic processes .
we adopt the notation @xmath105 , @xmath106 , @xmath107 , and @xmath108 , corresponding to the dot being empty , in state 1 , in state 2 , or doubly occupied .
the term @xmath109 represents the elastic processes @xmath110 .
the denominator includes the energy difference between the two states involved in the process , @xmath111 , and the numerator represents the fact that state @xmath107 has to be occupied for this process to occur . in the same way the term @xmath112 represents the process @xmath113 . assuming @xmath65 ,
corresponding to low bias and low temperature , we see that these terms cancel completely at the electron - hole symmetry point @xmath95 , owing to the equal population of the two levels .
we note that these two processes , involving tunneling into and out of state @xmath106 , use the couplings of this state . in the same way we see how the two processes @xmath114 and @xmath115 cancel .
the factor of @xmath116 enters here as it is state @xmath107 that is active in these processes .
we have seen that the elastic co - tunneling processes cancel completely at the electron - hole symmetry point , independent of the relative coupling strength of the two levels . away from electron - hole symmetry , we see that the processes no longer cancel
. however , at @xmath100 , we again find complete canceling for @xmath117 .
we now investigate the inelastic processes .
it can easily be seen that the terms involving @xmath118 and @xmath119 cancel at @xmath100 , due to the equal occupation of the two levels .
electron - hole symmetry is thus not required for the canceling of the inelastic terms , degeneracy of the two levels is sufficient .
to summarize we have shown that at @xmath95 the current suppression is complete , in the limit of low bias and low temperature at the electron - hole symmetric point . unlike previous explanations of this phenomena our derivation gives an intuitive explanation to the suppression in terms of canceling of co - tunneling processes .
it also explains why the suppression is not complete for finite bias and temperature or away from electron - hole symmetry .
this will be further studied in secs .
v and vi respectively .
the suppression is not an interference - effect between different processes , since the processes that cancel involve different initial and final states .
the processes instead cancel as a result of the equal population of the two levels . at degeneracy , but away from electron - hole symmetry , the elastic processes do not cancel unless @xmath117
. there is , however , partial canceling due to the different parity of the two levels .
we stress that the complete suppression at electron - hole symmetry is not an effect of this difference in parity .
the pseudo - spin operators are always orthogonal .
it is this orthogonality that gives equal occupations of the two levels and ensures the canceling of current carrying processes in the low - bias limit .
the canceling is due to the phase lapse between the elastic co - tunneling processes using @xmath105 and @xmath108 as the virtual intermediate state .
this shows the relationship between population switches , phase lapses , and vanishing conductance , which was previously investigated in ref . .
at electron - hole symmetry the vanishing conductance is thus an effect of the correlations induced by the coulomb interaction .
in the previous section we saw that the current suppression was most pronounced at the electron - hole symmetry point . in this section
we will investigate the suppression at this point in more detail .
the current is calculated as the flow of particles from the left contact into the dot . due to current conservation
this is , in the stationary case , of course equal to the current flowing from the dot into the right contact .
thus , we do not label the current with any lead - index . to quantify the extent of the current suppression we introduce the dimensionless ratio between the current through the two - level dot corresponding to eq .
( [ u ] ) and the current through a single level dot containing only level one @xmath120 this ratio measures how much the presence of level 2 suppresses the current . in the regime where sequential
tunneling is dominating the first order von neumann method ( 1vn ) can be used to approximately evaluate the current , see appendix a. using the couplings of eq .
( [ couplings ] ) , it is possible to derive eq .
( [ current2 ] ) . at electron - hole symmetry , @xmath95 , this simplifies to @xmath121\right .
\left . + w_{11}f_l(u/2)a^2-w_{22}[1-f_l(-u/2)]a^2 \right .
\nonumber \\ & & \left.+w_{22}f_l(u/2)-w_{dd}[1-f_l(u/2)](1+a^2)+2a\mathcal{r}\left\lbrace w_{12}\right\rbrace [ 1-f_l(-u/2)+f_l(u/2 ) ] \right\rbrace , \label{peter}\end{aligned}\ ] ] where @xmath122 , @xmath123 , @xmath124 and @xmath125 is the probability to find the dot empty , in state 1 , in state 2 and doubly occupied respectively , while @xmath126 is the coherence between level 1 and 2 . here
@xmath127 denotes the real part . in this paper
we refer to the coherence as the non - diagonal elements of the reduced density matrix .
the coherence is thus a basis - dependent quantity .
the first terms of eq .
( [ peter ] ) are identical to the corresponding pauli master equation and are easily interpreted in terms of transition rates .
the last term originating from the coherence of the two levels is more difficult to understand intuitively .
however , it is evident that a negative real part of the coherence results in a decreased current . at level degeneracy one can get rid of the coherence term by performing a basis change that diagonalizes the reduced density matrix .
here we focus at the case of electron - hole symmetry , @xmath95 , where level 1 and 2 have the same occupation and the coherence is real ( see appendix a ) .
the singly occupied part of the reduced density matrix can thus be written @xmath128 the diagonalization is given by @xmath129 where the diagonal elements of @xmath18 are the occupations in the new basis .
the couplings in the new eigenbasis are given by @xmath130 in the new basis one level couples strongly to the left , while the other couples strongly to the right lead .
this results in a strong current suppression and also explains the occupations in our new basis : the level which is strongly coupled to the lead with a high chemical potential has a high occupation .
a high bias thus results in a large coherence between the levels , cf .
the expression for the matrix @xmath18 . for @xmath117 ,
i.e. both levels have the same coupling strength , each level decouples completely from one lead . as a result
no current can flow through the quantum dot , see also ref . .
the effect of the coherence is thus to decrease the current .
as the coherence is zero in the new basis one can use the pauli master equation approach to derive the current . at electron - hole symmetry
we obtain @xmath131 . \label{easycurrent}\end{aligned}\ ] ] if only level 1 was present the current would be given by @xmath132,\end{aligned}\ ] ] and the resulting ratio is @xmath133 this result is valid when sequential tunneling is dominating . in fig .
[ qgamma ] the current suppression , @xmath134 , at electron - hole symmetry , @xmath95 , is shown as a function of the applied bias .
the numerator of @xmath134 is calculated using the 2vn method , while the denominator is calculated using negf which , however , agrees exactly with 2vn for the single level dot .
[ qgamma ] clearly shows that when @xmath135 , the ultra - high - bias regime , where sequential tunneling is dominating , is entered and the results of the 2vn simulations agree very well with the analytical results , eq .
( [ highbias ] ) . in the low - bias limit
the results are more complicated . for weak couplings
the contribution from co - tunneling processes is small and the first - order result eq .
( [ highbias ] ) is approached . as the bias
is increased the phase space for co - tunneling processes ( @xmath136 ) increases , so these can no longer be neglected .
this results in a stronger current suppression and a related decrease in @xmath134 for the weaker couplings . for strong couplings , i.e. @xmath137 , co - tunneling dominate and the low - energy hamiltonian of sec .
iv predicts complete suppression in the low - bias limit .
we stress that the suppression in the low - bias limit is not an artifact of neglecting higher order processes . as was shown in sec .
iv and refs . and
, the suppression is complete in the limit of low bias and low temperature , at the electron - hole symmetric point . in fig .
[ qkbt ] the simulated current suppression at electron - hole symmetry is shown for different temperatures as a function of bias .
as the temperature is not constant in figs .
5 and 6 we here express all energies in terms of @xmath138 .
it is clearly seen that the sequential tunneling result is reached in the high - temperature limit . in fig .
[ change_a ] we investigate the current suppression @xmath134 as a function of the difference @xmath29 in coupling strength between the two levels for different temperatures . for lower temperatures it can be seen that there is essentially complete suppression when both levels have equal coupling strength , i.e. @xmath139 . when @xmath140 and the difference in coupling strength is large , almost no suppression is observed .
this is expected as no current flows through the weakly coupled state , and due to electron - hole symmetry it does not matter if it is occupied or not . in this section
we investigated how a finite bias and temperature affects the suppression at electron - hole symmetry . in the limit of low / high bias and temperature we compared with previously established results / first order simulations , and found good agreement .
in intermediate regimes we found that the current suppression could show a non - monotonous behavior , see @xmath141 in fig . 4 . furthermore figs .
[ qgamma ] , [ qkbt ] and [ change_a ] show that @xmath134 increases with increasing temperature until the high temperature limit is reached . we have seen how the current suppression is given by analytical first order results in the regimes where sequential tunneling is dominating . in this regime
the suppression is coherence - induced and only partial unless both levels have the same coupling strength .
stronger couplings lead to an increased co - tunneling which results in a stronger suppression .
with a clear picture of the current suppression at electron - hole symmetry , we now change focus to the regime away from this symmetry point .
we will investigate the current at @xmath142 . at this point
the singly occupied levels are close to chemical potentials of the leads , and sequential tunneling contributes strongly to the transport .
thus , the 1vn - method gives reasonable results and we can compare first- and second - order von neumann to investigate the importance of the broadening in this regime .
[ shift](a ) shows the current at @xmath143 calculated with 2vn and 1vn methods . since level broadening
is not included in the first order method the peaks are too high and narrow . more interestingly , we see that the current minimum has been shifted away from degeneracy . comparing with the noninteracting results of fig .
[ canyon ] , where the dip is found at @xmath25 at @xmath144 , one draws the conclusion that the shift is related to the finite charging energy @xmath6 .
we will argue that this shift has three origins : \i ) a small part of the shift originates from a first order level renormalization and a bias induced anticrossing between the levels , as can be seen from the slightly different dip location of the two 1vn - curves in fig .
[ shift]a .
the first order simulations not including these effects also show a shift , however slightly smaller .
this suggests that there are additional mechanisms producing the shift .
the bias induced anticrossing is further discussed in sec .
vii in connection with the population inversion .
\ii ) the main shift within the first - order approach can be attributed to the asymmetry in couplings , which has to be compensated by different spectral weights in order to allow for full canceling of the individual channels .
as discussed in ref .
the spectral density of each state is divided among two peaks , situated at @xmath145 and @xmath146 .
increasing the occupation of one level results in a shift in the spectral function of the other level towards the upper peak at @xmath146 . at @xmath147
the upper peak is far from the fermi level of the leads and does not contribute to the transport through the dot . in order for the stronger / weaker coupled level ( i.e. , levels 1/2 for @xmath148 ) to have lower / higher spectral weight at the fermi energy , the weaker coupled level must have a higher occupation
this is naturally the case , if this level is lower in energy , thus at @xmath149 the current minimum requires a positive @xmath8 , c.f .
( [ levels ] ) .
this effect is reduced for increased bias . for @xmath150 and @xmath151 ,
complete conductance suppression can be found at degeneracy due to an increased occupation of the weakly coupled level , as discussed below .
\iii ) comparing the first- and second - order methods it is clear that the inclusion of broadening of the levels increases the shift of the conductance dip .
it is well known that in a broadened level the low energy part of the spectral density is mostly occupied.@xcite due to the coulomb interaction this results in a significantly higher occupation of the strongly coupled level at level degeneracy , see fig .
[ shift](b ) and refs . .
to compensate , the minimum is shifted so that the strongly coupled level has a higher energy than the weakly coupled . due to
the level broadening one can not expect the two levels to contribute equally to the current away from the degeneracy point .
as the levels are equally far from the bias window , the broader level will have a stronger contribution if the occupations of the two levels are equal .
thus the minimum can be found for values of @xmath8 such that the occupation of the weakly coupled level is slightly higher , in agreement with fig .
[ shift](b ) . in the regime
@xmath152 coulomb interaction is of minor importance but the dip is still found where the occupation of the weakly coupled level is higher , to compensate for the larger broadening of the strongly coupled level .
this implies the strong shift outside the coulomb - blockade region , as obtained by negf .
note that the above explanation holds not only in the sequential but also in the co - tunneling regime @xmath153 , but the argumentation has to be reversed for @xmath154 , where the current is mainly carried by holes through the doubly occupied states . at electron - hole symmetry
the singly and doubly occupied levels have the same distance from the bias window , which results in minimal conductance at level degeneracy .
the canceling of the different current paths through the dot is related to coherence between the levels . for the couplings considered here we concluded in sec .
v that a negative real part of the coherence led to a decreased current .
it is thus not surprising that fig .
[ shift](c ) shows that the current minimum coincides with a minimum in the real part of the coherence .
the relation between the charging energy and the shift of the dip can be better understood by studying the current at @xmath147 for different values of @xmath6 .
this is done in fig .
[ lowbias ] for @xmath155 . here
@xmath43 corresponds to electron - hole symmetry and the conductance suppression is not complete due to the finite bias . as the charging energy increases
, the current will mainly flow through the singly occupied dot states .
as predicted above , this results in a shift of the conductance dip towards lower energy of the weakly coupled state . due to the different occupations the current through the two levels cancel for larger @xmath6 , resulting in a strong conductance suppression .
thus a larger @xmath6 results in stronger conductance suppression at the dip , in agreement with fig .
[ lowbias ] .
we now focus on the case of higher bias . when both levels are in the bias window and @xmath156 the current can be calculated using the first - order results .
here it is important whether the charging energy is dominating over bias or not .
for @xmath157 , i.e. in the ultra - high - bias limit [ fig .
[ electronholsym](c ) ] , both the singly and doubly occupied levels are inside the bias window and the current can be calculated using eq .
( [ smallu ] ) , @xmath158}{2\left[(1+a^{2})^{2}+(e_1-e_2)^{2}/\gamma_1^{2}\right]}. \label{smallu2}\end{aligned}\ ] ] in agreement with ref . .
the dip is found at @xmath100 but the suppression is not complete .
( note that this result corresponds to electron - hole symmetry . ) for @xmath159 , i.e. the high - bias case [ fig . [ electronholsym](b ) ] , the doubly occupied state will be empty and the current is given by eq .
( [ gurvitz2 ] ) , @xmath160 again in agreement with ref . .
we note that the current is zero at @xmath100 .
this can be explained intuitively:@xcite at degeneracy we are free to work in any linear combination of our original single particle basis .
especially we can choose a basis such that one level , @xmath161 , is completely decoupled from the lead with the low chemical potential .
this level will always be occupied due to the high bias .
however , the other level can not be occupied as @xmath159 . as a result
neither level can carry any current through the dot , as seen for @xmath162 in fig .
[ gurvitz ] .
it is evident that if level 1 is more strongly coupled than level 2 , then @xmath163 implying that the weakly coupled level has a higher occupancy , as discussed above . at high bias , as in fig .
[ gurvitz ] , the conductance minimum can be found close to level degeneracy in agreement with eqs .
( [ smallu2 ] ) and ( [ gurvitz3 ] ) .
the symmetry with respect to mirroring in @xmath25 , for @xmath43 is broken by a finite @xmath6 . comparing figs .
[ lowbias ] and [ gurvitz ] shows that increasing the bias increases the width of the dip and broadens the peaks . in this section the current suppression away from electron - hole symmetry
was investigated . since the levels were close to the bias window ,
sequential tunneling was important , and different parity of the two levels was essential for the suppression . at finite bias , the suppression was not complete unless the levels had the same coupling strength .
however , a finite coulomb interaction shifted the dip away from degeneracy and resulted in a stronger suppression .
this shift , not previously discussed in the literature , was studied in detail .
three origins for the shift were found : ( i ) a renormalization of the energy levels , ( ii ) a shift required to give the weakly coupled level a higher occupation which reduces the spectral weight of the singly occupied state for the other level in the case of coulomb repulsion , and ( iii ) a shift related to the broadening of the levels .
furthermore , in the high - bias regime , fig .
[ electronholsym](b ) , there was complete current suppression at the degeneracy point .
in this section we change the focus from the dip to the surrounding current peaks . in the limit of zero bias and zero temperature and @xmath164 ,
it was shown in ref .
that , for the couplings of eq .
( [ couplings ] ) , the peaks correspond to a difference in occupation of @xmath165 . to verify
if the 2vn method can reproduce these results we must go to the regime @xmath166 , i.e. to the boundary of the region of validity of the 2vn - method . to reduce the possible occurrence of negative currents we use couplings with @xmath117 , where the current is inherently zero at degeneracy .
the result of the simulations is shown in fig .
[ peaksbias ] , where the black curve corresponds to low bias , @xmath167 .
closer examination reveals that the peaks are located at @xmath168 where @xmath169 in good agreement with the analytical prediction .
it is not surprising that the difference in occupation is smaller than predicted since there is also a finite probability that the dot is empty or doubly occupied .
a more surprising result is found when the occupations for higher values of @xmath3 are studied , see fig .
[ popinversion ] .
this is a regime of validity for the 2vn method due to the increased bias .
it should be pointed out that the regime described here , where the upper state is far below the chemical potentials , can not be studied in pure second - order perturbation in @xmath2 .
such methods do not include mechanisms for emptying the upper state.@xcite on each side of the level degeneracy there is a region of population inversion , i.e. the upper level has a higher occupation .
this population inversion has not previously been investigated for our system , however , a similar heating of single electron transistors due to inelastic co - tunneling processes was discussed in ref . .
the 2vn - simulation for the single - level dot with spin , where a magnetic field splits both spin directions , ( anderson model , not shown ) reveals no population inversion between the two spin levels .
this suggests that the coherence between the two orbital levels is essential for the inversion .
the coherence simply means that an electron can be in a superposition of the two states , which results from the overlap between the dot levels introduced by the couplings to the leads . for the level configuration of fig .
[ popinversion ] the dot can as a good approximation be considered to be singly occupied , which means that the schrieffer - wolff transformation of sec .
iv can be used to evaluate the overlap . as noted in sec .
iv , in connection with eq .
( [ lowenergy2 ] ) , the contribution to the overlap from state @xmath170 in the leads is proportional to @xmath171 , which vanishes in the low bias limit . as a result
the coherence and the overlap between the dot state vanishes . however , a finite bias results in an effective overlap between the two states . replacing the sums with integrals and using the notation of eq .
( [ levels ] ) one arrives at the following expression for the anticrossing @xmath172 @xmath173 in the additional term of the dot hamiltonian given by @xmath174 .
we have assumed @xmath175 .
as long as @xmath176 , which always is the case if eq .
( [ anticrossing ] ) is valid , @xmath172 is negative for @xmath177 . for small bias @xmath178 the anticrossing vanishes .
this is expected as inelastic processes where the state of the dot is changed is only possible as long as @xmath179 . to investigate the effect of this anticrossing we diagonalize the dot hamiltonian .
the resulting eigenvalues and eigenvectors are given by @xmath180 @xmath181 where the coefficients are given by @xmath182 respectively , with @xmath183 and @xmath184 being normalization constants .
clearly we always have @xmath185 .
the new eigenstates also get new effective couplings @xmath186 due to @xmath187 both coefficients @xmath188 and @xmath189 of the lower eigenstate @xmath190 have the same sign .
combined with the difference in parity of our two original states this results in a weak coupling of @xmath190 to the left lead , i.e. the lead with high chemical potential , and a strong coupling to the right lead . in the same way @xmath191 couples strongly to the left and weakly to the right , see fig .
[ popinversionfig2 ] .
the population inversion is real in the sense that it is present in the diagonal basis of the dot ( @xmath190 , @xmath191 ) , see fig .
[ popinversion ] .
it might at first seem surprising that the population inversion is so large . for @xmath192 in fig .
[ popinversion ] we have @xmath193 over the entire range of @xmath8 . the detuning @xmath194 at @xmath195 , i.e. it is almost the size of @xmath196 , and yet a strong population inversion is observed .
the cause for this population inversion is inelastic processes driving a current from left to right by filling the state @xmath191 and emptying @xmath190 . due to the asymmetric couplings of the ( @xmath190 , @xmath191)-basis such processes are very strong .
the populations of fig .
[ popinversion ] in the ( @xmath190 , @xmath191)-basis are calculated assuming @xmath172 given by eq .
( [ anticrossing ] ) .
deviations from single occupation of the quantum dot , due to the broadening of the dot levels and the finite bias , cause slightly different values of @xmath172
. however , at @xmath25 , @xmath190 and @xmath191 are the symmetric and antisymmetric combination of @xmath106 and @xmath107 respectively , independently of @xmath172 .
therefore , the population inversion is independent of @xmath172 at @xmath25 . for higher bias
this results in populations close to @xmath197 and @xmath198 . in the limit @xmath199 ,
it can not be considered as a population inversion since @xmath172 vanishes in this limit , resulting in @xmath200 . when @xmath201 , i.e. @xmath202 , @xmath190 is mainly composed of the original state @xmath106 , while @xmath191 is mainly composed of @xmath107 .
for @xmath203 the situation is the reverse .
thus , it is always the case that the upper state @xmath191 , that couples strongly to the left , is mainly composed of the upper state in the original basis .
this explains the observed population inversion in our original basis . a negative bias results in @xmath204 , which leads to a strong coupling of @xmath190 to the left .
however , for a negative bias the left lead has a low chemical potential .
thus the explanation holds also for negative bias .
the @xmath96 terms of eq .
( [ lowenergy2 ] ) are present also in first order perturbation in dot - lead couplings .
we have seen that the real part of the sum , ( or integral in the continuous limit ) , over @xmath96 gives rise to the population inversion .
we thus expect to see population inversion also in first order simulations if the real part is included .
however , first order perturbation theory can not be applied to the co - tunneling regime where eq .
( [ lowenergy2 ] ) is valid .
we instead investigate the ultra - high bias regime where all states are inside the bias window .
( [ vonneumann ] ) can be solved including the real part of the sums .
the result is shown in fig .
[ popinversionfig3 ] , where also the result of a 2vn - simulation is shown as comparison .
one observes that as the bias is increased , which means that the real parts of the sums becomes smaller , the occupations approach the infinite bias result of eq .
( [ smallu ] ) .
it is evident that the broadening results in a smaller population inversion .
we can therefore not completely rule out the possibility that the population inversion is an effect of neglecting higher order couplings . as the inversion is present also at very high bias , where the contribution from higher order terms is small , we believe , however , that it is a real effect . since it is the real part of of the sums of eq .
( [ vonneumann ] ) that is responsible for the renormalization one would not expect to see any population inversion when these are neglected .
indeed , already at @xmath205 one finds that the results of such simulations are indistinguishable from eq .
( [ smallu ] ) . these results
are therefore not shown in fig .
[ popinversionfig3 ] .
the population inversion is in fact related to the conductance peaks , by increasing the elastic co - tunneling .
let us assume that @xmath206 so that @xmath207 . then in eq .
( [ currentoperator ] ) the elastic processes with @xmath99 will dominate . if the levels are placed as in fig .
[ electronholsym](a ) , then @xmath208 and @xmath209 are both positive and there is constructive interference between these tunneling processes .
put simply , the population inversion activates elastic processes through the states close to the bias window , i.e. processes that contribute strongly to the current .
furthermore , the elastic processes have the same phase and interfere constructively . for large detuning , @xmath8 , the coherence decreases and as a result the occupation of the upper level will decrease . increased temperature , which also results in a lower coherence and a related decrease in population inversion , also leads to decreased peak height .
this scenario describes the formation of the conductance peaks at finite bias .
finally we would like to point out the experimental significance of the results presented in this section . by embedding the quantum dot in e.g. a microwave cavity
the dot can be connected to a photon mode . due to the population inversion
this couping might lead to stimulated emission of photons , i.e. a maser is formed .
previously it has been suggested that such a population inversion can be reached in double quantum dot systems , where the inversion results from an asymmetric couping of the dots to left and right lead.@xcite the results of this section suggest that an increased bias will modify the dot hamiltonian of two - level quantum dots so that such couplings are achieved , as long as the original dot levels have different parity .
the difference in parity guarantees a non - zero dipole element between the dot levels .
one would expect the dipole element to be larger in two - level quantum dots than in double quantum dots due to the spatial overlap of the two states .
this suggests an increased observability of population inversion in two - level quantum dots .
a thorough study of the transport through spinless two - level quantum dots has been presented .
regimes not previously studied in the presence of co - tunneling , such as the regimes of finite bias and temperature , as well as the transport away from electron - hole symmetry , was studied . to present a coherent picture of the transport through the system ,
comparison with previously derived results were made .
close to degeneracy a suppression of the current was found , while a slight detuning of the energy levels resulted in current peaks .
this resulted in a canyon of current suppression cutting through both the sequential and co - tunneling regime .
different mechanism for the suppression were found in the different regimes . in the coulomb blockade region
an intuitive explanation to the suppression was presented in the low - energy regime , in terms of canceling of different tunneling processes due to equal occupation of the two levels . in the high - bias limit with infinite
charging energy a complete suppression was found since the electron was trapped in a non - conducting state , preventing any transport .
these two mechanisms for the suppression are independent of the lead - dot couplings . when neither of the above mechanism were present , different parity of the two dot - levels had to be assumed to achieve current suppression . for such couplings the canyon
was found close to degeneracy in all four cases ( high and low bias , strong and weak couplings ) , independent of the level positions , albeit the current did not always entirely drop to zero . to quantify the current suppression
, simulations were performed using the second - order von neumann technique . in the regime where the lead - dot couplings dominate over temperature and applied bias , i.e. co - tunneling processes contribute strongly to the current , a strong suppression was observed both in the sequential and the co - tunneling regime , assuming different parity of the two levels . for weaker couplings , where co - tunneling could be neglected ,
only partial suppression was observed in agreement with analytical results .
away from the electron - hole symmetry point a shift of the conductance dip was observed .
the shift was shown to have three origins : ( i ) a renormalization of the energy levels , ( ii ) a shift required to give the weakly coupled level a higher occupation , and ( iii ) a shift related to the broadening of the levels .
finally the conductance peaks were investigated .
the peaks were found to be intimately related to an increased occupation of the upper level .
the increased population resulted from the bias - induced anticrossing between the two orbital levels .
sufficiently high bias resulted in an anticrossing large enough to cause population inversion .
we would like to thank henrik nilsson , hongqi xu and christian bergenfeldt for many discussions .
financial support from the swedish research counsil ( vr ) is gratefully acknowledged .
in this appendix we will derive expressions for the current assuming that the broadening can be neglected .
the results are thus of first order in @xmath2 . in the von neumann formalism
the markov approximation is made within the schrdinger picture , as compared to the redfield method where the markov limit is taken in the interaction picture .
this results in slightly different results when the levels are non - degenerate.@xcite eq .
( [ ht ] ) defines @xmath210 as the coupling of a single particle state @xmath211 to state @xmath14 in lead @xmath212 .
the couplings of the corresponding many particle states are given by @xmath213 where @xmath214 and @xmath215 are dot states .
we use the convention that the many particle state @xmath216 contains one more dot electron than @xmath29 and so on . assuming energy independent couplings , @xmath217 , the equation of motion for the reduced density matrix is in the first order von neumann method given by:@xcite @xmath218 where @xmath219 is the fermi distribution .
going to the continuum limit one replaces the sums over @xmath14 with integrals .
the real parts of the integrals constitute a renormalization of the energy levels .
however , the results of this appendix are applied either to electron - hole symmetry or to the high / ultra - high - bias case . in these situations
the renormalization vanishes , and for the couplings of eq .
( [ couplings ] ) the stationary 1vn equations read for the vacuum state @xmath220 , @xmath221 : @xmath222 \right\rbrace + w_{11}\left\lbrace 2-f_{l}(e_1)-f_{r}(e_1)\right\rbrace \\ & + & a^{2}w_{22}\left\lbrace 2-f_{l}(e_2)-f_{r}(e_2)\right\rbrace + a\mathcal{r}\left\lbrace w_{12}\right\rbrace \left\lbrace f_l(e_1)+f_l(e_2)-f_r(e_1)-f_r(e_2)\right\rbrace=0,\nonumber\end{aligned}\ ] ] for the single particle state @xmath223 , @xmath224 : @xmath225a^{2 } \right\rbrace\\ & + & a^{2}w_{dd}\left\lbrace 2-f_{l}(e_2+u)-f_{r}(e_2+u ) \right\rbrace -a\mathcal{r}\left\lbrace w_{12 } \right\rbrace \left\lbrace f_{l}(e_1)+f_{l}(e_2+u ) -f_{r}(e_1)-f_{r}(e_2+u)\right\rbrace=0,\nonumber\end{aligned}\ ] ] for the other single particle state @xmath226 , @xmath227 : @xmath228a^{2}+f_{l}(e_1+u)+f_{r}(e_1+u ) \right\rbrace \\ & + & w_{dd}\left\lbrace 2-f_{l}(e_1+u)-f_{r}(e_1+u ) \right\rbrace -a\mathcal{r}\left\lbrace w_{12 } \right\rbrace \left\lbrace f_{l}(e_2)+f_{l}(e_1+u ) -f_{r}(e_2)-f_{r}(e_1+u)\right\rbrace=0,\nonumber\end{aligned}\ ] ] and for the coherence @xmath223 , @xmath227 , ( @xmath226 , @xmath224 gives the same information since @xmath229 ) : @xmath230+\frac{e_1-e_2}{\gamma_1}w_{12}=0 . \nonumber\end{aligned}\ ] ] the equation for @xmath231 , @xmath232 ( doubly occupied state ) does not provide any new information .
one must instead use the normalization @xmath233 at @xmath95 , where @xmath234 , the equations of motion are symmetric in the two levels .
this results in an equal population of the two states .
furthermore the @xmath235-term of eq .
( [ last2 ] ) disappears , implying that the coherence is real as remarked in sec .
v. once the system of equations has been solved , the current can be calculated with@xcite @xmath236-w_{11}[1-f_{l}(e_1)]+w_{11}f_l(e_2+u)a^{2}-w_{22}[1-f_{l}(e_2)]a^{2}+w_{22}f_l(e_1+u)\right.\nonumber \\ \left .- w_{dd}[1-f_l(e_1+u)+[1-f_l(e_2+u)]a^{2 } ] + a\mathcal{r}\left\lbrace w_{12}\right\rbrace[2-f_{l}(e_1)-f_l(e_2)+f_{l}(e_1+u)+f_{l}(e_2+u)]\right\rbrace .~~ \label{current2}\end{aligned}\ ] ] we first consider the ultra - high - bias limit , i.e. @xmath237 and @xmath238 . inserting these occupations in eqs .
( [ first])-([last2 ] ) one obtains @xmath239 the imaginary part of @xmath126 only occurs in the last equation and can easily be expressed in terms of the real part of @xmath126 .
@xmath240 eq .
( [ w12 ] ) can then be rewritten as @xmath241\mathcal{r}\left\lbrace w_{12}\right\rbrace=0.\end{aligned}\ ] ] solving for the occupations one finds @xmath242 as noted in sec .
vii . using eq .
( [ current2 ] ) the current can be calculated as @xmath243.\end{aligned}\ ] ] solving eqs .
( [ a9])-([w12 ] ) results in @xmath158}{2\left[(1+a^{2})^{2}+(e_1-e_2)^{2}/\gamma_1^{2}\right]}. \label{smallu}\end{aligned}\ ] ] it is easy to show that the conductance minimum is found at @xmath100 where , for a general value of the bias , the current is given by eq .
( [ easycurrent ] ) .
we next turn to the high - bias limit @xmath244 , i.e. @xmath150 and @xmath245 . here
it trivially follows from the equation for @xmath246 that @xmath125 is zero .
furthermore eq .
( [ first ] ) is redundant .
the remaining equations for the occupations are given by @xmath247 from eq .
( [ w12b ] ) @xmath248 can be determined @xmath249 thus in the case @xmath244 the current is given by @xmath250 once again the conductance minimum is at degeneracy but for @xmath244 there is complete conductance suppression independently of the relative coupling strength @xmath29 , in agreement with refs . .
in this appendix we give the derivation of eq .
( [ renormalization ] ) .
the renormalization , given by the @xmath96-terms , reads where the factor of @xmath252 accounts for the hermitean conjugate in eq .
( [ lowenergy2 ] ) .
one sees that the sums are actually logarithmically divergent , which of course seems like a problem at first sight . however , a common renormalization of the two levels is just a shift of the zero point in the hamiltonian , that can be neglected .
the physically interesting quantity is the difference in renormalization of the two levels .
we first assume that level 2 is fixed , transferring the whole renormalization to level 1 . replacing the operators with their expectation values @xmath253 , in the low - bias limit , and using @xmath254 and a constant density of states ,
the renormalization is given by with @xmath256 denoting the principal value of the integral .
inserting @xmath257 for @xmath258 and @xmath259 for @xmath260 , corresponding to the low temperature limit , and adding the constant term @xmath261 we obtain eq .
( [ renormalization ] ) this expression can also be derived from first - order redfield which is similar to first order von neumann , except that the markov limit is taken in the interaction picture , while the schrdinger picture is used in the von neumann method .
redfield generates an equation of motion like eq .
( [ vonneumann ] ) , but with different energy denominators in the equations for the off - diagonal terms of the reduced density matrix .
|
motivated by the recent discovery of a canyon of conductance suppression in a two - level equal spin quantum dot system [ phys .
rev .
lett .
@xmath0 , 186804 ( 2010 ) ] the transport through this system is studied in detail . at low bias and
low temperature a strong current suppression is found around the electron - hole symmetry point independent of the couplings , in agreement with previous results . by means of a schrieffer - wolff transformation
we are able to give an intuitive explanation to this suppression in the low - energy regime . in the general situation ,
numerical simulations are carried out using quantum rate equations .
the simulations allow for the prediction of how the suppression is affected by the couplings , the charging energy , the position of the energy levels , the applied bias , and the temperature .
we find that away from electron - hole symmetry , the parity of the couplings is essential for the current suppression .
it is also shown how broadening , interference , and a finite interaction energy cause a shift of the current minimum away from degeneracy .
finally we see how an increased population of the upper level leads to current peaks on each side of the suppression line . at sufficiently high bias
we discover a coherence - induced population inversion .
|
ultracold atomic gases have attracted much interest since the successful realization of bose - einstein condensation in a bosonic @xmath0rb system @xcite . due to its high controllability in the interaction strength , particle number , and other parameters ,
many remarkable phenomena have been observed such as the mott transitions in the bosonic and fermionic systems @xcite .
recently , the spatial modulation of the interaction has also been realized in the @xmath1yb gas system @xcite , which stimulates further theoretical investigations on particle correlations in the ultracold atomic systems .
one of the interesting questions is how the spatial modulation in the interactions affects ground - state properties in the fermionic optical lattice .
although the optical lattice system with uniform interactions has been studied theoretically and experimentally , the system with alternating interactions , which may be one of the simplest systems , has not been discussed so far .
therefore , it is desired to clarify how the alternating interactions affect ground - state properties , in particular , the nature of the mott transition . motivated by background , we consider the infinite - dimensional hubbard model with alternating interactions . by means of the two - site dynamical mean - field theory ( dmft ) @xcite , we discuss how particle correlations affect the metal - insulator transition at half - filling . the nature of the mott phases is also discussed .
the paper is organized as follows . in sec . [ 2 ] , we introduce the model hamiltonian and briefly summarize our theoretical approach .
we demonstrate how particle correlations affect ground state properties in sec .
a brief summary is given in the last section .
we study the hubbard model with alternating interactions , as @xmath2\right ) , \label{hami}\end{aligned}\ ] ] where @xmath3 creates ( annihilates ) a fermion at the @xmath4-th site with spin @xmath5 .
@xmath6 is the hopping integral , and @xmath7 is the site - dependent onsite interaction in the sublattice @xmath8 . at half - filling ,
the systems with @xmath9 , @xmath10 , @xmath11 and @xmath12 are identified since the hamiltonian is invariant under the particle - hole transformations @xcite as @xmath13 and @xmath14 . to investigate the hubbard model eq .
( [ hami ] ) , we make use of dmft @xcite . which has successfully been applied to various strongly correlated electron systems . in the framework of dmft
, the lattice model is mapped to an effective impurity model , where local particle correlations are taken into account precisely .
the green function for the original lattice system is then obtained via self - consistent equations imposed on the impurity problem . in dmft ,
the green function in the lattice system @xcite is given as , @xmath15^{-1},\ ] ] where @xmath16 and @xmath17 are the chemical potential and the self - energy for the @xmath18-th sublattice , and @xmath19 is the dispersion relation for the bare band . in terms of the density of states (
dos ) @xmath20 , the local green function is expressed as , @xmath21 where @xmath22 in the following , we use the semicircular dos with the half bandwidth @xmath23 , @xmath24 , which corresponds to the infinite - coordination bethe lattice .
then the self - consistency condition for the sublattice @xmath18 is represented as , @xmath25 where @xmath26 is the non - interacting green function of the effective impurity model for the sublattice @xmath18 .
there are various numerical methods to solve the effective impurity problem . to discuss the mott transitions in the hubbard model at half - filling
, we use here the two - site dmft method @xcite , which provides us with a transparent view of the phase transitions @xcite . in this scheme
, we introduce a specific anderson impurity model for the sublattice @xmath18 , which is connected to only one host site , as @xmath27\right ) , \label{andhami}\end{aligned}\ ] ] where @xmath28 annihilates a fermion with spin @xmath29 in the effective bath ( localized band ) and @xmath30 .
note that the effective parameters in the impurity model such as the spectrum of host particles @xmath31 , the energy level of localized band @xmath32 and the hybridization @xmath33 , should be determined selfconsistently so that the obtained results properly reproduce the original lattice problem .
when the system is half - filling , @xmath34 .
this simple method provides sensible results as far as low - energy properties are concerned .
namely , for low - energy excitations around the fermi surface , the green function for @xmath35-fermions may be approximated by a single pole as @xmath36 where the residue @xmath37 corresponds to the quasi - particle weight for the sublattice @xmath18 .
therefore , the self - consistency condition is simplified as @xmath38 by estimating the quasi - particle weight for each anderson impurity model , we proceed to perform the dmft iterations . in the following
, we discuss the metal - insulator transition in the system with alternating interactions .
let us discuss the stability of a metallic state in the hubbard model with spatially - modulated interactions . by making use of the two - site dmft method
, we calculate the quasi - particle weight for each sublattice , which is defined as @xmath39 for the @xmath40 sublattice obtained in the systems with fixed ratios @xmath41 ( a ) , 2.0 ( b ) , -0.7 ( c ) and -2.0 ( d ) .
, width=415 ] the obtained results for fixed ratios @xmath42 and @xmath43 are shown in fig . [ fig1 ] . when @xmath44 , the system is reduced to the hubbard model with the homogeneous interactions . as the interactions are introduced , the quasi - particle weights are decreased , as shown in fig .
[ fig1 ] ( a ) . finally , the quasi - particle weights vanish , where the mott transition occurs at @xmath45 in the repulsive case and the pairing transition occurs at @xmath46 in the attractive case
. these are consistent with other numerical studies such as the numerical renormalization group @xmath47 @xcite and the exact diagonalization @xmath48 @xcite .
when the spatial dependence of interactions is introduced , the quasi - particle weights depend on the sublattices , as shown in figs .
[ fig1 ] ( b ) , ( c ) and ( d ) .
namely , the strong renormalization appears in the sublattice with a larger local interaction .
nevertheless , the quasi - particle weights vanish simultaneously in the strong coupling region .
this suggests that particle correlations induce a single phase transition . to clarify the nature of the single phase transition
, we focus on low energy properties around the fermi level . when the heavy metallic state in the vicinity of the critical point is considered , the quasi - particle weight in the effective anderson model eq .
( [ andhami ] ) can be expanded in @xmath33 @xcite , as @xmath49 in this case , the effective hybridization for the other sublattice is comparable since it is given as @xmath50 .
therefore , the fermions in both sublattices are simultaneously renormalized and different interactions are irrelevant around the critical points in this approximation .
the critical interactions for the transition are then given as @xmath51 the phase diagram is shown in fig .
[ pd ] . when the interaction is weak , the normal metallic state is realized . on the other hand , when the interactions in both sublattices are large , there exist four kinds of the mott states , which are denoted as the mott states ( i ) , ( ii ) , ( iii ) , and ( iv ) , as shown in fig .
we find that the mott state ( i ) [ ( iii ) ] in the repulsive ( attractive ) hubbard model is widely stabilized away from the symmetric limit @xmath52 .
in addition , the mott states ( ii ) and ( iv ) are also stabilized in the system where interactions are alternately arranged repulsive and attractive . to clarify the nature of these mott states
, we also calculate the double occupancy @xmath53 . for the @xmath40 sublattice obtained in the systems with fixed ratios @xmath41 ( a ) , 2.0 ( b ) , -0.7 ( c ) and -2.0 ( d).,width=415 ] in the non - interacting case @xmath54 , the quantity is quarter since the empty , single occupied and doubly occupied states are equally realized at each site . introducing the interaction , the quantity for each sublattice varies , as shown in fig .
when @xmath55 , the double occupancy for the sublattice @xmath56 rapidly decreases since the local coulomb interaction is large .
finally these quantities for both sublattices vanish simultaneously at the mott transition point . on the other hand ,
the double occupancy approaches half as the attractive interaction increases . in this case
, the pairing transition occurs at @xmath57 @xmath58 , where the single occupied states are forbidden and the empty or doubly occupied state is realized at each site .
it is also found that when the repulsive and attractive interactions alternate in the system @xmath59 , the mott and pairing transitions occur simultaneously in the corresponding sublattices , as shown in figs .
[ fig2 ] ( c ) and ( d ) . in this paper , we have clarified the nature of the mott transitions in the system with alternating interactions , by restricting our discussions to the paramagnetic case .
although this condition may be relevant for the optical lattice experiments , some ordered states are naively expected at very low temperatures .
therefore , it is necessary to clarify how the magnetically ordered state or the superfluid state is stabilized in the system only with repulsive or attractive interactions @xmath60 . on the other hand , the mott transition should be realized in the system with repulsive and attractive interactions @xmath61 since any ordered states are hard to be stabilized . in the case , the hole doping effect is one of the important problems since the local particle density varies in the optical lattice system with a confining potential .
furthermore , it is expected that particle correlations induce the commensurability in the doped system @xcite , which is now under consideration .
we have investigated the half - filled hubbard model with spatially - modulated interactions by means of the two - site dynamical mean - field theory .
it has been clarified that a single mott transition occurs when two kinds of the interactions are increased .
we have determined the ground state phase diagram and have discussed the nature of these mott states .
this work was partly supported by the grant - in - aid for scientific research 20740194 ( a.k . ) and the global coe program nanoscience and quantum physics " from the ministry of education , culture , sports , science and technology ( mext ) of japan .
u. schneider , l. hackermuller , s. will , th .
best , i. bloch , t. a. costi , r. w. helmes , d. rasch and a. rosch , metallic and insulating phases of repulsively interacting fermions in a 3d optical lattice , science * 322 * , 1520 ( 2008 ) .
|
we investigate the half - filled hubbard model with spatially alternating interactions by means of the two - site dynamical mean - field theory .
it is found that a single mott transition occurs when two kinds of interactions are increased .
this implies that the different interactions are essentially irrelevant at the critical point .
the nature of the mott states is also addressed .
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
|
a sequence @xmath0 formed as the concatenation of @xmath1 non - overlapping segments is given , where @xmath2 .
each segment is generated by some unknown stochastic process distribution .
the process distributions that generate every pair of consecutive segments are different . the index @xmath3 where one segment ends and another starts is called a _
change point_. the parameters @xmath4 specifying the change points @xmath5 are unknown and have to be estimated .
change point analysis is one of the core problems in classical mathematical statistics @xcite . in a typical formulation of the problem ,
the samples within each segment @xmath6 are assumed to be i.i.d . and the change refers to a change in the mean . in the literature on nonparametric methods for dependent data , the form of the change and/or the nature of dependence
are usually restricted , for example , strong mixing conditions are imposed @xcite . moreover , even for dependent time series , the finite - dimensional marginals are almost exclusively assumed different @xcite .
however , such strong assumptions do not necessarily hold in most of such real - world applications as bioinformatics , network traffic , market analysis , audio / video segmentation , fraud detection etc .
methods used in these applications are thus usually model - based or employ application - specific ad hoc algorithms .
more specifically , a theoretical framework to allow for the understanding of what is possible and under which assumptions is entirely lacking . in this paper
, we consider highly dependent time series , making as little assumptions as possible on how the data are generated .
each segment is generated by an ( unknown ) stationary ergodic process distribution .
the joint distribution over the samples can be otherwise arbitrary .
we make no such assumptions as independence , finite memory or mixing ; the samples can be arbitrarily dependent .
the marginal distributions of any given fixed size before and after the change points may be the same : the change refers to that in the time - series distribution .
we aim to construct an asymptotically consistent algorithm that simultaneously estimates all @xmath7 parameters @xmath8 consistently .
an estimate @xmath9 of a change point parameter @xmath10 is _ asymptotically consistent _ if it becomes arbitrarily close to @xmath10 in the limit as the length @xmath11 of the sequence approaches infinity .
the asymptotic regime means that the error is arbitrarily small if the sequence is sufficiently long , i.e. the problem is offline " and @xmath12 does not grow with time .
note that , in general , for stationary ergodic processes , rates of convergence are provably impossible to obtain ( see , for example , @xcite ) .
therefore , the asymptotic results of this work can not be strengthened . as follows from an impossibility result by @xcite , it is impossible to estimate the number of change points in the general setting that we consider .
thus , we assume that @xmath7 is known .
the case of @xmath13 has been addressed in @xcite where a simple consistent algorithm to estimate one change point is provided .
the general case of @xmath14 turns out to be much more complex . with the sequence containing multiple change points ,
the algorithm is required to simultaneously analyze multiple segments of the input sequence , with no a - priori lower bound on their lengths . in this case
the main challenge is to ensure that the algorithm is robust with respect to segments of arbitrarily small lengths .
usually in statistics this is done using methods based on the speed of convergence of sample averages to expectations . in the context of stationary ergodic processes ,
such tools are unavailable as no guarantees on the speed of convergence exist .
hence , the simultaneous analysis of segments of arbitrarily small lengths is conceptually much more difficult .
the problem is considerably simplified if additionally a lower bound on the minimum separation of the change points is provided . under this assumption
, an algorithm is proposed in @xcite that gives a list of possibly more than @xmath7 candidate estimates , whose first @xmath7 elements are asymptotically consistent , but it makes no attempt to estimate @xmath7 .
we use empirical estimates of the so - called distributional distance @xcite , which have proved useful in various statistical learning problems involving stationary ergodic time series @xcite . our method has a computational complexity that is at most quadratic in each argument .
we evaluate it on synthetic data generated by processes that , while being stationary ergodic , do not belong to any simpler " class , and can not be modeled as hidden markov processes with countable state spaces . moreover , the single - dimensional marginals before and after each change point are the same . to the best of our knowledge , none of the existing change point estimation algorithms work in this scenario .
the remainder of this paper is organized as follows . in section [ sec : pre ] we introduce preliminary notations and definitions . in section [ sec : protocol ]
we formalize the problem and describe the general framework considered . in section [ sec : results ] we present our algorithm , state the main consistency result , and informally describe how the algorithm works . in section [ sec : exp ] we provide some experimental results .
we prove the consistency of the proposed method in section [ sec : proofs ] .
let @xmath15 be a measurable space ( the domain ) ; in this work we let @xmath16 but extensions to more general spaces are straightforward . for a sequence @xmath17 we use the abbreviation @xmath18 .
consider the borel @xmath19-algebra @xmath20 on @xmath21 generated by the cylinders @xmath22 , where the sets @xmath23 are obtained via the partitioning of @xmath24 into cubes of dimension @xmath25 and volume @xmath26 ( starting at the origin ) .
let also @xmath27 .
process distributions are probability measures on the space @xmath28 . for @xmath29 and @xmath30
let @xmath31 denote the _ frequency _ with which @xmath12 falls in @xmath32 , i.e. @xmath33 a process @xmath34 is _ stationary _ if for any @xmath35 and @xmath36 , we have @xmath37 a stationary process @xmath34 is called _ stationary ergodic _
if for all @xmath38 with probability 1 we have @xmath39 by virtue of the ergodic theorem ( see , e.g. , @xcite ) , this definition can be shown to be equivalent to the standard definition for the stationary ergodic processes ( see , e.g. , @xcite ) . for a given @xmath40
we can define distributions on the space @xmath41 where the borel sigma - algebra @xmath42 is generated by @xmath43 the distributional distance between a pair of process distributions @xmath44 is defined as follows @xcite @xmath45 we let @xmath46 , but any summable sequence of positive weights may be used . in words , we partition the sets @xmath24 , @xmath47 into cubes of decreasing volume ( indexed by @xmath48 ) and take a weighted sum over the differences in probabilities of all the cubes in these partitions .
the differences in probabilities are weighted : smaller weights are given to larger @xmath25 and finer partitions .
we use _ empirical estimates _ of this distance defined as follows .
[ emd ] the empirical estimate of @xmath49 between @xmath50 and a process @xmath34 is given by @xmath51 and that between a pair of sequences @xmath52 .
is defined as @xmath53 where @xmath54 and @xmath55 are any sequences of integers that go to infinity with @xmath11 .
[ thm : constd ] let a pair of sequences @xmath56 and @xmath57 be generated by a distribution @xmath34 whose marginals @xmath58 are stationary and ergodic .
then @xmath59 the triangle inequality holds for the distributional distance @xmath60 and its empirical estimates @xmath61 , so that for all distributions @xmath62 and all sequences @xmath63 we have , @xmath64 the distributional distance @xmath60 and its empirical estimates @xmath61 are convex functions : for every @xmath65 for all distributions @xmath66 and all sequences @xmath67 with @xmath68 we have @xmath69 [ rem1 ] consider a pair of sequences @xmath70 with @xmath71 .
the computational complexity of @xmath72 is of order @xmath73 , where @xmath74 .
let @xmath75 correspond to the partition where each cell @xmath76 contains at most one element i.e. @xmath77 indeed , in all summands corresponding to @xmath78 are equal to 0 ; moreover , all summands corresponding to @xmath79 are equal .
thus , we already see that the number of required calculations is finite .
note that in practice @xmath75 is bounded by the length of the binary precision in approximating real numbers ( i.e. , the length of the mantissa ) . for a fixed @xmath80 for every sequence @xmath81 the frequencies @xmath82 for all @xmath83 may be calculated using suffix trees with @xmath84 worst case construction and search complexity ( see , e.g. , @xcite ) .
this brings the overall computational complexity of to @xmath85 .
furthermore , the practically meaningful choices of @xmath54 are of order @xmath86 . to see this , observe that for a fixed @xmath87 the frequencies @xmath88 of cells in @xmath89 corresponding to higher values of @xmath25 are not consistent estimates of their probabilities ( and thus only add to the error of the estimate ) .
indeed , for a pattern @xmath90 with @xmath91 of length @xmath25 the probability @xmath92 is asymptotically of the order @xmath93 where @xmath94 denotes the entropy rate of @xmath58 . by the above argument , one can set @xmath95 and @xmath96 in , bringing the overall complexity of calculating @xmath97 to @xmath98 .
we formalize the problem as follows . the sequence @xmath99 , generated by an unknown arbitrary process distribution , is formed as the concatenation of @xmath1 of sequences @xmath100 where @xmath2 , and @xmath7 is assumed known .
each of the sequences @xmath101 , is generated by an
_ unknown stationary ergodic _ process distribution .
formally , consider the matrix @xmath102 of random variables generated by some ( unknown ) stochastic process distribution @xmath34 such that , 1 .
the marginal distribution over every one of its rows is an unknown stationary ergodic process distribution ; 2 .
the marginal distributions over the consecutive rows are different , so that every two consecutive rows are generated by different process distributions .
note that the requirements are only on the marginal distributions over the rows ; the distribution @xmath34 is otherwise completely arbitrary .
the process distributions are unknown and may be dependent .
moreover , the means , variances , or , more generally , the finite - dimensional marginal distributions of any fixed size before and after the change points are not required to be different .
we consider the most general scenario where the _ process distributions are different_. the sequence @xmath12 is obtained by first fixing a length @xmath103 and then concatenating the segments @xmath104 that is , @xmath105 where for each @xmath106 , the _ segment _
@xmath107 is the sequence obtained as the first @xmath108 elements of the @xmath109 row of @xmath110 with @xmath111 . the parameters @xmath4
specify the _ change points _
@xmath5 which separate consecutive segments @xmath112 generated by different process distributions .
the change points are _ unknown _ and to be estimated .
let the minimum separation of the change point parameters @xmath113 be defined as @xmath114 since the consistency properties we are after are asymptotic in @xmath11 , we require that @xmath115 . note that this condition is standard in the change point literature , although it may be unnecessary when simpler formulations of the problem are considered , for example when the samples are i.i.d .
however , conditions of this kind are inevitable in the general setting that we consider , where the segments and the samples within each segment are allowed to be arbitrarily dependent : if the length of one of the sequences is constant or sub - linear in @xmath11 then asymptotic consistency is not possible in this setting .
however , @xmath116 is assumed unknown , and no ( lower ) bounds on it are available .
we also make no assumptions on the distance between the process distributions : they can be arbitrarily close .
our goal is to devise an algorithm that provides estimates @xmath117 for the parameters @xmath4 .
the algorithm must be _ asymptotically consistent _ so that @xmath118
in this section we present our method given by algorithm [ alg : kk ] , which as we show in theorem [ thm : kk ] , is asymptotically consistent under the general assumptions stated in section [ sec : protocol ] .
the proof of the consistency result is deferred to section [ sec : proofs ] . in this section
we give describe the algorithm , and intuitively explain why it works .
the following two operators namely , the score function denoted @xmath119 and the single - change point - estimator denoted @xmath120 are used in our method .
let @xmath121 be a sequence and consider a subsequence @xmath122 of @xmath12 with @xmath123 . 1 .
define the score function as the intra - subsequence distance of @xmath122 , i.e. @xmath124 2 .
define the single - change point estimator of @xmath122 as @xmath125 let us start by giving an overview of what algorithm [ alg : kk ] aims to do .
the algorithm attempts to simultaneously estimate all @xmath7 change points using the single - change point - estimator @xmath126 given by applied to appropriate segments of the sequence . for @xmath126 to produce asymptotically consistent estimates in this setting , each change point must be isolated within a segment of @xmath12 ,
whose length is a linear function of @xmath11 .
moreover , each segment containing a change point must be `` sufficiently far '' from the rest of the change points , w where `` sufficiently far '' means within a distance linear in @xmath11 .
this may be obtained by dividing @xmath12 into consecutive non - overlapping segments , each of length @xmath127 with @xmath128 for some @xmath129 $ ] where @xmath116 is given by . since , by definition , @xmath116 specifies the minimum separation of the change point parameters , the resulting partition has the property that every three consecutive segments of the partition contain _ at most one _ change point .
however , @xmath116 is not known to the algorithm . moreover , even if @xmath130 , not all segments in the partition contain a change point .
the algorithm uses the score function @xmath131 given by ( [ defn : delta ] ) to identify the segments that contain change points . as for @xmath116 , instead of trying to find it , the algorithm produces many partitions of @xmath12 ( using different guesses of @xmath116 ) , and produces a set of candidate change point estimates using each of them .
finally , a weighted combination of the candidate estimates is produced .
the weights are designed to converge to zero on iterations where our guess for a lower bound on @xmath116 is incorrect .
this last step of combining multiple estimates may be reminiscent of prediction with expert advice , @xcite , with the important difference that performance ( loss ) can not be measured directly in our setting .
* input : * @xmath132 , number @xmath7 of change points * initialize : * @xmath133 [ gamma : kk ] [ alg : estim ] * return : * @xmath134 algorithm [ alg : kk ] works as follows . given @xmath135 ,
it iterates over @xmath136 and at each iteration , produces a guess @xmath137 as a lower - bound on @xmath116 . for every fixed @xmath138 , a total of @xmath1 grids
are generated , each composed of evenly - spaced boundaries @xmath139 , that are @xmath140 apart for @xmath141 .
the grids have distinct starting positions @xmath142 for @xmath143 .
( as shown in the proof of theorem [ thm : kk ] , this ensures that for a fixed @xmath138 at least one of the grids for some @xmath144 has the property that the change points are not located at the boundaries . ) among the segments of the grid , @xmath7 segments of highest _ score _ , @xmath131 are selected ; @xmath131 is given by ) .
the single - change point estimator @xmath126 is used to seek a candidate change point parameter in each of the selected segments .
the weighted combination is given as the final estimate for every change point parameter @xmath4 .
two sets of weights are used , namely , an iteration weight @xmath145 and a score @xmath146 .
the former gives lower precedence to finer grids . to calculate the latter , at each iteration on @xmath138 and @xmath147 , for every fixed @xmath148 ,
a partition of the grid is considered , which is composed of the boundaries @xmath149 .
each partition , in turn , specifies a set of non - overlapping consecutive segments of length @xmath150 , for each of which a parameter @xmath151 is calculated as the @xmath152 highest intra - distance value @xmath153 of its segments ; the performance weight @xmath146 is obtained as @xmath154 .
( as shown in the proof , @xmath146 converges to zero on the iterations where either @xmath155 or there exists some change point on the boundary of one of the segments of the partition . ) [ thm : kk ] algorithm [ alg : kk ] is asymptotically consistent , provided that the correct number @xmath7 of change points is given : @xmath156 the proof is given in section [ sec : proofs ] ; an intuitive description follows .
first observe that the empirical estimate @xmath61 of the distributional distance is consistent .
thus , the empirical distributional distance between a given pair of sequences converges to the distributional distance between their generating processes . from this
we can show that the intra - subsequence distance @xmath131 corresponding to the segments in the grid that do not contain a change point converges to zero .
this is established in lemma [ prelem : nochpt : ii ] , provided in section [ sec : proofs ] . on the other hand ,
since the generated grid becomes finer as a function of @xmath138 , from some @xmath138 on , we have @xmath157 so that every three consecutive segments of the grid contain _ at most _ one change point . in this case , for every segment that contains a change point , the single - change point estimator @xmath126 produces an estimate that , for long enough segments , becomes arbitrarily close to the true change point .
this is shown in lemma [ lem2 ] , provided in section [ sec : proofs ] . moreover , for large enough @xmath11 , the performance scores associated with these segments are bounded below by some non - zero constant .
thus , the @xmath7 segments of highest @xmath131 , each contain a change point which can be estimated consistently using @xmath126 .
however , the estimates produced at a given iteration for which @xmath158 may be arbitrarily bad . moreover ,
even for @xmath159 , an appropriate grid to provide consistent estimates must be such that no change point is exactly at the start or at the end of a segment .
however , can not identify such grids directly .
we make the following observation .
the following observation is key to achieving this objective indirectly .
+ consider the partitioning of @xmath12 into @xmath7 consecutive segments where there exists at least one segment with more than one change point .
since there are exactly @xmath7 change points , there must exist at least one segment in this partitioning that does not contain any change points at all .
as follows from lemma [ prelem : nochpt : i ] , the segment that contains no change points has an intra - subsequence distance @xmath119 that converges to @xmath160 . on the iterations for which @xmath161 ,
at least one of the three partitions has the property that among every set of @xmath7 segments in the partition , there is _ at least _
one segment that contains no change points . in this case
, @xmath119 corresponding to the segment without a change point converges to @xmath160 .
the same argument holds for the case where @xmath162 , while at the same time a change point happens to be located exactly at the boundary of a segment in the grid .
observe that for a fixed @xmath138 , the algorithm forms a total of @xmath1 different grids , with the same segment size , but distinct starting points @xmath163 .
since there are @xmath7 change points , for all @xmath138 such that @xmath164 there exists at least one appropriate grid ( for some @xmath165 ) , that simultaneously contains all the change points within its segments . in this case
, @xmath166 converges to a non - zero constant .
the final estimate @xmath9 for each change point parameter @xmath10 is obtained as a weighted sum of the candidate estimates produced at each iteration .
two sets of weights are used in this step , namely @xmath146 and @xmath167 , whose roles can be described as follows . 1 .
@xmath146 is used to penalize for the ( arbitrary ) results produced on iterations on @xmath168 and @xmath144 where , either @xmath169 , or while we have @xmath164 there exists some @xmath10 for some @xmath170 such that @xmath171 .
as discussed , @xmath146 converges to zero only on these iterations , while it is bounded below by a non - zero constant on the rest .
@xmath167 is used to give precedence to estimates sought in longer segments .
since the grids are finer for larger @xmath138 , at some higher iterations the segments may not be long enough to produce consistent estimates .
thus , if @xmath11 is large enough , the final estimates @xmath172 converge to the true parameters , @xmath4 .
* computational complexity . *
the proposed method can be easily and efficiently implemented . for a fixed @xmath138 , a total of @xmath173 distance calculations
are done on segments of length @xmath174 , and a total of @xmath175 distance calculations are done to estimate each change point ; the procedure is repeated @xmath1 times . by remark [ rem1 ] , and summing over @xmath168 iterations , the overall complexity is of order @xmath176 .
the rest of the computations are of negligible order .
in this section we evaluate our method using synthetically generated data .
* generating the synthetic time - series . * in order to generate the data we use stationary ergodic process distributions that do not belong to any simpler " general class of time - series , and can not be approximated by finite state models , such as hidden markov processes with finite state - spaces .
moreover , the single - dimensional marginals of all distributions are the same throughout the generated sequence . to the best of our knowledge
, none of the existing algorithms are designed to work in this scenario , and as a result are bound to fail under this framework .
hence , we can not compare our method against other change point estimation algorithms .
we generate a segment @xmath177 as follows .
+ 1 . fix a parameter @xmath178 and two uniform distributions @xmath179 and @xmath180 .
2 . let @xmath181 be drawn randomly from @xmath182 $ ] .
3 . for each @xmath183 obtain @xmath184 ; draw @xmath185 from @xmath186 .
set @xmath187 if @xmath188 is irrational by a long double with a long mantissa . ]
this produces a real - valued stationary ergodic time - series .
similar families are commonly used as examples in this framework , ( e.g. @xcite ) . for the purpose of our experiment
, we fixed four parameters @xmath189 , @xmath190 and @xmath191 ( with long mantissae ) to correspond to @xmath192 different process distributions ; we used uniform distributions @xmath179 and @xmath180 over @xmath193 $ ] and @xmath194 $ ] respectively , ( deliberately chosen to overlap ) . to produce @xmath195 we randomly generated @xmath196 change point parameters @xmath4 at least @xmath197 apart .
every segment of length @xmath198 with @xmath111 was generated with @xmath199 , and using @xmath179 and @xmath180 . by this procedure ,
the single - dimensional marginals are the same throughout @xmath12 .
figure [ fig : synth ] demonstrates the average estimation error - rate of algorithm [ alg : kk ] as a function of the sequence length @xmath11 .
we calculate the error rate as @xmath200
in this section , we prove the main consistency result .
the proof depends upon some technical lemmas stated below .
[ prelem : nochpt : i ] let @xmath132 be generated by a stationary ergodic process @xmath34 . for all @xmath178
the following statements hold with @xmath34-probability 1 : 1 .
[ prelem : nochpt : i1 ] @xmath201 for every @xmath202 .
[ prelem : nochpt : i1d ] @xmath203 .
[ prelem : nochpt : ii ] @xmath204 to prove part we proceed as follows .
assume by way of contradiction that the statement is not true .
therefore , there exists and some @xmath205 , @xmath202 and sequences @xmath206 and @xmath207 with @xmath208 , such that with probability @xmath209 we have @xmath210 using the definition of @xmath211 it is easy to see that the following inequalities hold @xmath212 for every @xmath213 and all @xmath214 .
fix @xmath215 . for each @xmath216
we can find a finite subset @xmath217 of @xmath218 such that @xmath219 for every @xmath220 , there exists some @xmath221 such that for all @xmath222 with probability one we have @xmath223 define @xmath224 and let @xmath225 ; observe that @xmath226 .
let @xmath227 consider the sequence @xmath228 . 1 . for every @xmath216 we have @xmath229 2 . on the other hand , by all @xmath230 we have @xmath231 increase @xmath232 if necessary to have @xmath233 for all @xmath234 and @xmath235
for all @xmath230 we obtain @xmath236 where follows from ; follows from and ; and follows from , , , summing over the probabilities , and observing that @xmath237 for all @xmath238 . observe that holds for any @xmath215 , and it particular it holds for @xmath239 .
therefore , we have @xmath240 contradicting .
part follows .
fix @xmath241 , @xmath178 and @xmath242 .
we can find some @xmath202 such that @xmath243 by part of lemma [ prelem : nochpt : i ] , there exists some @xmath232 such that for all @xmath230 we have @xmath244 + from and , for all @xmath230 we have @xmath245 and part of the lemma follows .
fix @xmath241 , @xmath178 . without loss of generality assume that @xmath246 .
observe that for every @xmath247 we have @xmath248 .
therefore , by there exists some @xmath232 such that for all @xmath249 we have @xmath250 it remains to use the definition of @xmath131 ( [ defn : delta ] ) and the triangle inequality to observe that @xmath251 for all @xmath230 , and follows .
[ prelem : chpt : dist ] assume that a sequence @xmath121 has a change point @xmath252 for some @xmath253 so that the segments @xmath254 , @xmath255 are generated by two different process distributions @xmath34
, @xmath256 respectively .
if @xmath34 , @xmath256 are both stationary ergodic then with probability one , for every @xmath257 we have 1 .
[ prelem : chpt : dist : i ] @xmath258 2 .
[ prelem : chpt : dist : iii ] @xmath259 fix @xmath241 , @xmath253 , @xmath257 .
there exists some @xmath202 such that @xmath260 to prove part we proceed as follows . by the definition of @xmath211 given by , for all
@xmath261 and all @xmath262 we have @xmath263 therefore , for all @xmath261 and all @xmath262 we obtain @xmath264 where the first inequality follows from the fact that @xmath265 , the second inequality follows from the definition of @xmath211 given by @xmath266 and the third inequality follows from .
observe that @xmath267 for all @xmath268 .
therefore , by part of lemma [ prelem : nochpt : i ] , there exists some @xmath269 such that for all @xmath270 we have @xmath271 similarly , @xmath272 for all @xmath273 .
therefore , by part of lemma [ prelem : nochpt : i ] , there exists some @xmath274 such that for all @xmath275 we have @xmath276 note that @xmath277 for all @xmath261 .
therefore , we have @xmath278 for all @xmath279 , @xmath268 and @xmath273 we have @xmath280 let @xmath281 . by , , , and , for all @xmath230 we have @xmath282 finally , by and for all @xmath230 we obtain @xmath283 and part of lemma
[ prelem : chpt : dist ] follows .
the proof of the second part is analogous .
[ lem2 ] consider a sequence @xmath284 with @xmath7 change points .
let @xmath285 , be a sequence of indices with @xmath286 for some @xmath178 , such that @xmath287 for some @xmath242 . 1 .
[ lem2:i ] with probability one we have @xmath288 where @xmath289 denotes the minimum distance between the distinct distributions that generate @xmath12 .
[ lem2:ii ] assume that we additionally have @xmath290 \subseteq [ \theta_{k-1},\theta_{k+1}]\ ] ] where @xmath291 denote the elements of @xmath292 that appear immediately to the left and to the right of @xmath5 respectively . with probability
one we obtain @xmath293 ) .
fix some @xmath170 .
define @xmath294 . following the definition of @xmath295 given by we have @xmath296 to prove part of lemma [ lem2 ] ,
we show that for large enough @xmath11 , with probability @xmath297 we have @xmath298 let @xmath299 .
to prove for the case where @xmath300 we proceed as follows .
by assumption of the lemma , we have @xmath301 hence , it is easy to see that @xmath302 fix @xmath241 .
observe that as follows from the definition of @xmath303 and @xmath304 , and our assumption that @xmath300 , the segment @xmath305 is fully generated by @xmath306 . by ,
the condition of part of lemma [ prelem : nochpt : i ] hold for @xmath305 .
therefore , there exists some @xmath307 such that for all @xmath249 we have @xmath308 similarly , from and we have @xmath309 by and , the conditions of part of lemma [ prelem : chpt : dist ] hold for @xmath310 .
therefore , there exists some @xmath311 such that for all @xmath312 we have @xmath313 by we have @xmath314 moreover , we obtain @xmath315 where the inequality follows from ( [ lem2:fract ] ) and the definition of @xmath289 as the minimum distance between the distributions . let @xmath316 . for all @xmath230
we obtain @xmath317 where and follow from applying the triangle inequality on @xmath61 , follows from and , and follows from .
since holds for every @xmath215 , this proves ( [ objective1 ] ) in the case where @xmath300 . the proof for the case where @xmath318 is analogous .
since holds for every @xmath170 , part of lemma [ lem2 ] follows .
+ ( [ lem2:ii ] ) .
fix some @xmath170 .
following the definition of @xmath126 given by ( [ defn : phi ] ) we have @xmath319 to prove part of the lemma , it suffices to show that for every @xmath320 with probability @xmath297 , for large enough @xmath11 , we have @xmath321 for all @xmath322 . to prove ( [ objective ] ) for @xmath323 we proceed as follows .
fix some @xmath324 and @xmath241 .
first note that for all @xmath325 we have @xmath326 note that by the sequence @xmath327 is a subsequence of @xmath328 .
consider the segment @xmath327 .
observe that by the conditions of part of lemma [ prelem : nochpt : i ] are satisfied by all @xmath329 .
therefore , there exists some @xmath307 such that for all @xmath249 we have @xmath330 similarly , consider @xmath331 .
observe that by definition of @xmath304 we have @xmath332 ; moreover , by the segment is a subsequent of @xmath333 .
therefore , by part of lemma [ prelem : nochpt : i ] , there exists some @xmath311 such that for all @xmath312 we have @xmath334 by , there is a single change point @xmath335 within @xmath336 .
therefore , every @xmath337 has a linear distance from @xmath338 , i.e. @xmath339 for all @xmath337 . on the other hand , @xmath340 .
therefore by part of lemma [ prelem : chpt : dist ] there exists some @xmath341 such that @xmath342 let @xmath343 . by ( [ lem2ii : firstdiff1 ] ) , ( [ lem2ii : firstdiff2 ] ) and the subsequent application of the triangle inequality on @xmath61 for all @xmath230 we obtain @xmath344 by applying the triangle inequality on @xmath61 , for all @xmath230 we obtain @xmath345 where follows from , and follows from .
we also have @xmath346 where the inequality follows from and the definition of @xmath289 as the minimum distance between the distributions that generate the data .
finally , from ( [ lem2ii : firsthalf ] ) , ( [ lem2ii:2ndhalf ] ) and ( [ lem2ii : dr ] ) for all @xmath230 we obtain , @xmath347 since ( [ lem2ii : final ] ) holds for every @xmath215 , this proves ( [ objective ] ) for @xmath323 . the proof for the case where @xmath348 is analogous .
since ( [ objective ] ) holds for every @xmath170 , part follows . on each iteration
@xmath349 the algorithm produces a set of estimated change points .
we show that on some iterations these estimates are consistent , and that estimates produced on the rest of the iterations are negligible .
we partition the set of iterations into three sets as described below .
first recall that for every @xmath168 and @xmath144 the algorithm generates a grid of boundaries @xmath350 such that for all @xmath168 and @xmath144 we have @xmath351 therefore , the segments @xmath352 have lengths that are linear functions of @xmath11 .
more specifically , for @xmath136 and @xmath144 define @xmath353 ( note that @xmath354 can also be zero . ) for all @xmath355 we have @xmath356 such that @xmath357 this first subset of the set of iterations @xmath136 corresponds to the higher iterations where @xmath137 is too small . in this case
the resulting grids are too fine , and the segments may not be long enough for the estimates to be consistent .
these iterations are penalized by small weights @xmath167 , so that the corresponding candidate estimates become negligible .
* the second subset corresponds to the iterations where * a. * @xmath358 $ ] _ and _ * b. * the segments are long enough for the candidate change point parameter estimates to be consistent .
let @xmath359 where @xmath116 defined by specifies the minimum separation of the change points . for all @xmath360 we have @xmath361 .
therefore , at every iteration on @xmath360 and @xmath144 , for every change point @xmath362 we have @xmath363 \subseteq [ \theta_{k-1},\theta_{k+1}]\ ] ] where @xmath364 and @xmath365 are defined in lemma [ lem2 ] .
we further partition the set of iterations on @xmath144 into two subsets as follows .
for every fixed @xmath366 we identify a subset @xmath367 of the iterations on @xmath368 at which the change point parameters @xmath4 are estimated consistently and the performance scores @xmath146 are bounded below by a nonzero constant .
moreover , we show that if the set @xmath369 is nonempty , the performance scores @xmath146 for all @xmath366 and @xmath370 are arbitrarily small . 1 . to define @xmath367 we proceed as follows . for every @xmath4 we can uniquely define @xmath371 and @xmath372 so that @xmath373 .
therefore , for any @xmath374 with @xmath375 , we have @xmath376 . observe that we can only have @xmath7 distinct residues @xmath377 . therefore , any subset of @xmath378 with @xmath1 elements , contains at least one element @xmath379 such that @xmath380 .
it follows that for every @xmath381 there exists at least one @xmath144 such that @xmath382 . for every @xmath383 ,
define @xmath384 let @xmath385 and define @xmath386 .
note that @xmath387 . by , and hence part of lemma [ lem2 ] , for every @xmath383 there exists some @xmath388 such that for all @xmath389 we have @xmath390 where @xmath289 denotes the minimum distance between the distinct distributions that generate the data . recall that , as specified by algorithm [ alg : kk ] we have @xmath391 .
hence by ( [ thm : constj1 ] ) for all @xmath230 we have @xmath392 by lemma [ lem2 ] there exists some @xmath393 such that for all @xmath394 we have @xmath395 2 .
define @xmath369 for @xmath383 .
it may be possible for the set @xmath396 to be nonempty on some iterations on @xmath366 . without loss of generality ,
define @xmath397 for all @xmath366 with @xmath398 .
observe that by definition , for all @xmath366 such that @xmath399 , we have @xmath400 where @xmath354 is given by .
this means that on each of these iterations , there exists some @xmath335 for some @xmath170 such that @xmath401 for some @xmath402 . since @xmath130 for all @xmath383
, we have @xmath403 and @xmath404 .
therefore , by part of lemma [ prelem : nochpt : i ] , there exists some @xmath405 such that for all @xmath406 we have , @xmath407 thus , for every @xmath383 and all @xmath406 we have @xmath408 * step 3 . *
consider the set of iterations , @xmath409 . recall that it is desired for a grid to be such that every three consecutive segments contain at most one change point .
this property is not satisfied for @xmath410 , since by definition on these iterations we have @xmath411 .
we show that for all these iterations , the performance score @xmath412 becomes arbitrarily small . for all @xmath413 and @xmath143 , define the set of intervals @xmath414 and consider its partitioning into @xmath415 . observe that , by construction for every fixed @xmath416 , every pair of indices @xmath417 specifies a segment @xmath418 of length @xmath419 and the elements of @xmath420 index non - overlapping segments of @xmath12 . since for all @xmath409 we have @xmath411 , at every iteration on @xmath421 and @xmath422 , there exists some @xmath423 such that the segment @xmath418 contains more than one change point . since there are exactly @xmath7 change points , in at least one of the partitions @xmath424 for some @xmath148 we have that within any set of @xmath7 segments indexed by a subset of @xmath7 elements of @xmath424 , there exists at least one segment that contains no change points .
therefore , by ( [ thm : lem1cond ] ) , ( [ thm : lineardist ] ) and hence lemma [ prelem : nochpt : ii ] , for every @xmath421 there exists some @xmath425 such that for all @xmath426 we have @xmath427 let @xmath428 and @xmath429 .
let @xmath430 . by , and that @xmath431 , for all @xmath230 we have @xmath432 recall that by definition we have @xmath433 which , as follows from is nonzero .
therefore we have @xmath434 by and for all @xmath230 we have @xmath435 note that @xmath436 and that @xmath437 . therefore , by and for all @xmath230 we obtain @xmath438 similarly , from and we obtain @xmath439 let @xmath440 . by , ,
and we have @xmath441 since the choice of @xmath442 is arbitrary , the statement of the theorem follows .
we have presented an asymptotically consistent method to locate the changes in highly dependent time - series data .
the considered framework is very general and as such is suitable for real - world applications .
note that , in the considered setting , rates of convergence ( even of frequencies to respective probabilities ) are provably impossible to obtain .
therefore , unlike in the traditional settings for change - point analysis , the algorithms developed for this framework are forced not to rely on any rates of convergence .
we see this as an advantage of the framework as it means that the algorithms are applicable to a much wider range of situations . at the same time
, it may be interesting to derive the rates of convergence of the proposed algorithm under stronger assumptions ( e.g. , i.i.d . data , or some mixing conditions ) .
we conjecture that our method is optimal ( up to some constant factors ) in such settings ( although it is clearly suboptimal under parametric assumptions ) ; however , this is left as future work .
|
given a heterogeneous time - series sample , it is required to find the points in time ( called change points ) where the probability distribution generating the data has changed .
the data is assumed to have been generated by arbitrary , unknown , stationary ergodic distributions .
no modeling , independence or mixing are made .
a novel , computationally efficient , nonparametric method is proposed , and is shown to be asymptotically consistent in this general framework ; the theoretical results are complemented with experimental evaluations .
|
one of the main concerns in the study of ultra - relativistic collisions is the investigation of the quark - gluon plasma ( qgp ) properties . in this regard the thermodynamical aspects of the plasma is specially interesting due not only to the possibility of studying the deconfinement process but also because it can give important information to other fields , like hydrodynamical models of the qgp , cosmological models of the early universe and models of massive objects in astrophysics .
the non extensive statistics has been applied to a large number of problems since the seminal paper by c. tsallis in 1988 @xcite .
an updated list of applications and studies on the subject can be found in @xcite . in high energy physics ( hep )
nonextensivity was introduced by bediaga , curado and miranda ( bcm ) in 2000 @xcite . in that work
the authors used the well - known transverse momentum ( @xmath2 ) distribution from hagedorn s theory @xcite and formally substituted the exponential function by the q - exponential function that appears in the tsallis statistics . with the new distribution obtained by bcm
it is possible to describe the whole @xmath2-distribution measured in hep experiments .
a large number of works reporting the use of the non extensive formalism have been published since the bcm work , all of them showing a good agreement with experiments ( see for instance @xcite ) .
more recently a non extensive generalization of the hagedorn s theory @xcite was developed in @xcite showing that not only a limiting ( or critical ) temperature , @xmath3 , exists but also that there is an entropic index , @xmath4 , which characterizes the hadronic systems at least for the confined regime .
also a new formula for the hadronic mass spectrum in terms of @xmath3 and @xmath4 was derived from that theory and in ref .
@xcite it is shown that this formula can describe quite well the spectrum of known hadronic states with masses from the pion mass up to @xmath5 .
the non extensive self - consistent theory @xcite imposes much more restrictive tests to the applicability of the tsallis statistics in hep and a number of analysis of experimental data @xcite have shown that the theoretical predictions are in agreement with the experimental findings .
the non extensive thermodynamics of hadronic matter was already explored for null chemical potential ( @xmath6 ) systems in ref .
@xcite and compared with lattice - qcd data , showing a reasonable agreement . in the present work we extend the thermodynamics to finite chemical potential systems , which is of importance in the study of nucleus - nucleus collisions and of astrophysical objects
. an important class of compact objects are protoneutron stars .
the understanding of their evolution in time from the moment they are born as remnants of supernova explosions until they completely cool down to stable neutron stars , has been a matter of intense investigation .
all sorts of phenomenological equations of state ( eos ) , relativistic and non - relativistic ones , have been used to describe protoneutron star matter .
these eos are normally parameter dependent and are adjusted so as to reproduce nuclear matter bulk properties , as the binding energy at the correct saturation density and incompressibility as well as ground state properties of some nuclei @xcite .
the present work provides the necessary formalism for the investigation of how the non extensive statistics affects stellar matter .
the paper is organized as follows : in section [ sec : pf ] we introduce the partition function and show that it is in agreement with the @xmath2-distribution used in previous works @xcite ; in section [ sec : thermodynamics ] we derive the thermodynamical functions of interest and compare with previous results in the literature ; in section [ sec : thermo_had_sys ] we establish the phase transition line between confined and deconfined regimes in the @xmath7 diagram and discuss the properties of the eos of hadronic matter ; and finally we present our conclusions in section [ sec : conclusions ] .
we next outline the main formulas necessary for the development and application of the non - extensive formalism to hadronic matter .
consider the exponential function defined as @xmath8^{1/(q-1 ) } \qquad\;\;\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , , \ ; x\geq 0\ , , \\ & e_q^{(-)}(x)=\frac{1 } { e_q^{(+)}(|x|)}=[1+(1-q)x]^{1/(1-q)}\qquad\ , , \ ; x<0\ , .
\end{cases } \label{eq : eq}\ ] ] we define the q - logarithm as @xmath9 which would correspond to the inverse function of the q - exponential if @xmath10 were defined by @xmath11 for @xmath12 ( @xmath13 for @xmath14 ) . in the following we make use of these two definitions of the q - exponential , but do not consider in general their definition regimes in this way .
it follows straightforwardly that @xmath15^{2-q } \,,\\ & \frac{d}{dx}e_q^{(-)}(x)=[e_q^{(-)}(x)]^{q } \,,\\ & \frac{d}{dx}\log^{(+)}_q(x)=x^{q-2 } \,,\\ & \frac{d}{dx}\log^{(-)}_q(x)=x^{-q}\ , .
\end{cases}\label{usefull}\ ] ] the relations above are used many times in the following and specially for the derivation of the identities : @xmath16^{q}\,,\ ] ] for @xmath17 , and @xmath18^{2-q}\,,\ ] ] for @xmath19 . in these expressions and in the following
we take @xmath20 for bosons and fermions respectively .
we then define the grand - canonical partition function for a non extensive ideal quantum gas as @xmath21 where @xmath22 , the particle energy is @xmath23 , with @xmath24 being the hadron mass and @xmath6 the chemical potential , and @xmath25 is the step function .
the partition function for bosons is defined only for the case where @xmath26 , therefore the term with @xmath27 in the integrand is applied only for fermions , and it only contributes if @xmath1 . in the limit
@xmath28 the q - exponential reduces to the exponential function and the q - logarithm reduces to the logarithm function . in this limit
( [ partitionfunc ] ) reduces to the well - known fermi - dirac and bose - einstein partition functions for fermions and bosons , respectively , and in fact it is the partition function for the ideal quantum gas in tsallis statistics , as discussed below .
there is an intrinsic gain in the knowledge of the non extensive partition function , since it is closer to the methods of statistical mechanics , however in order to go deeper into more fundamental aspects one needs knowledge about nonperturbative qcd that are not available at present .
to show that the partition function defined in eq . ( [ partitionfunc ] ) corresponds to the one for a quantum gas in non extensive statistics we derive the occupation number , average number of particles , energy density and the entropy , and show that the results are identical to the ones obtained in refs @xcite in the sector @xmath29 .
however , some differences are found and discussed for @xmath30 .
the average number of particles can be obtained through the relations @xmath31 using relations ( [ usefull ] ) it is easy to show that @xmath32\ , , \label{occnumb}\ ] ] where @xmath33 the @xmath34-independent term @xmath35 demands some explanation .
the integrand in eq .
( [ partitionfunc ] ) is a discontinuous function in @xmath36 , i.e. in @xmath37 . as a consequence
, one has to be careful when acting with the derivative with respect to @xmath6 .
let us denote by @xmath38 the integrand in eq .
( [ partitionfunc ] ) , then one has @xmath39
\label{eq : ninte}\\ & = & - \frac{\mu } { \beta \sqrt{\mu^2-m^2 } } \left [ f^{(+)}(x=0^+ ) - f^{(-)}(x=0 ^ -)\right ] + \beta^{-1 } \int_0^\infty dp \ , \sum_{r=\pm 1 } \frac{\partial}{\partial\mu}f^{(r)}(p,\mu ) \ , \ , .
\nonumber\end{aligned}\ ] ] the first term in the last equality produces the contribution @xmath40 in eq .
( [ occnumb ] ) , which is non vanishing only when @xmath1 .
note also that @xmath41 is vanishing in the boltzmann - gibbs limit @xmath42 , as in this case the integrand in eq .
( [ partitionfunc ] ) is a continuous function in @xmath34 . from eq .
( [ occnumb ] ) one gets the average occupation number @xmath43 where the signals @xmath44 and @xmath45 correspond to @xmath46 and @xmath30 , respectively . in the case
@xmath17 , this result is identical to the one obtained by cmp @xcite and by cleymans and worku ( cw ) @xcite .
in addition , observe that for @xmath47 and @xmath34 sufficiently high one can write the occupation number as @xmath48 ^{-q } \,,\ ] ] which is exactly the equation used by bcm in ref .
@xcite to describe the hep @xmath2-distributions .
thus we conclude that the occupation number derived from the partition function defined in eq .
( [ partitionfunc ] ) can correctly describe the results obtained in ultra relativistic collisions , corresponding to the regime of high temperature and low baryonic chemical potential .
regarding the case @xmath19 , which is relevant for the regime of high baryonic chemical potential , there are some discrepancies with previous references that will be explained in section [ subsec : comparison ] .
the entropy can be obtained through the relation @xmath49 resulting @xmath50^{\tilde{q } } \log_q^{(-r)}\left(\bar{n}_q^{(r)}(x)\right ) + \xi [ 1 + \xi \bar{n}_q^{(r)}(x ) ] ^{\tilde{q } } \log_q^{(-r)}\left(1 + \xi \bar{n}_q^{(r)}(x)\right ) \bigg]\ , , \label{eq : s}\end{aligned}\ ] ] where we have defined @xmath51^{1/\bar{q}}$ ] .
this result is identical to the cmp entropy defined in ref .
@xcite , either for @xmath52 or @xmath27 , but written in terms of q - log functions and for bosons / fermions , as it can be easily checked . finally ,
for the sake of completeness , we show also the result for the average energy . it can be computed from the relation @xmath53 and the result
is @xmath54\ , , \label{occe}\ ] ] with @xmath55 .
we have verified the thermodynamic consistency of these expressions by checking that @xmath56 . if the @xmath34-independent terms @xmath35 and @xmath57 were dropped off , then the thermodynamic consistency would not be preserved for @xmath1 .
it is important to notice that the @xmath34-independent contributions @xmath58 and @xmath59 , which appear for fermions when @xmath1 in @xmath60 and @xmath61 respectively , have been disregarded in previous works , see e.g. @xcite .
these terms are consequence of the discontinuity in the integrand of eq .
( [ partitionfunc ] ) , which appears because of the definition of the q - exponential and q - logarithm functions we are using in eqs .
( [ eq : eq ] ) and ( [ eq : logq ] ) .
note that we are considering the same definitions for these functions as the ones used in many previous references , in particular @xcite .
the existence of this discontinuity was already pointed out and discussed in @xcite ( see fig . 2 and eqs .
( 25 ) and ( 26 ) in that reference ) .
however , in this and in other works it has not been taken into account the effect coming from the fact that the location of the discontinuity is a function of the chemical potential @xmath6 , as in the momentum variable the discontinuity appears at @xmath62 .
this means that the discontinuity will lead to some contribution proportional to it when deriving with respect to the chemical potential , which should be added to the contribution from the first derivative of the integrand , as it is written in eq .
( [ eq : ninte ] ) . using these arguments
, we have corrected some inconsistencies of previous references , and have obtained fully thermodynamical consistent expressions for fermions when @xmath1 .
it would be possible to get a continuous integrand in eq .
( [ partitionfunc ] ) by defining the q - log function as @xmath63 for all values of its argument .
however , the motivation to consider different expressions for @xmath64 and @xmath65 , as well as for @xmath66 and @xmath67 , was already stressed in the literature , and it is based on the fact that they fulfill the desirable logarithm - like @xmath68 and exponential - like @xmath69 relations .
see e.g. @xcite for a discussion . one way to avoid the explicit appearance of the p - independent terms in eqs .
( [ occnumb ] ) and ( [ occe ] ) is by modifying the definition of the occupation number for @xmath19 , in the form @xmath70 this occupation number contributes only in the interval @xmath71 in the integration in momentum , see eq .
( [ eq : ninte ] ) , and @xmath72 is any function which fulfills @xmath73 then the results of eqs .
( [ occnumb ] ) and ( [ occe ] ) are reproduced as it can be easily checked .
the point is that @xmath72 is not uniquely determined .
there is not a good criterium to choose one particular expression for @xmath72 , so we prefer to leave the expressions for @xmath74 and @xmath75 as defined in eqs .
( [ occnumb ] ) and ( [ occe ] ) .
the partition function can be obtained also from the total number of particles by integrating in @xmath6 , as follows from eq .
( [ eq : avn ] ) .
the entropy is then obtained by deriving the partition function with respect to temperature , as in eq .
( [ eq : sthermo ] ) . when making the computation one can realize that if one considered @xmath76 instead of @xmath77 in the occupation number for @xmath19 , see eq .
( [ eq : nq ] ) , as it is explicitly written in @xcite , the expression of the entropy which is obtained is not the one postulated by these authors , but a much more complicated expression which writes @xmath78^q + \frac{[e_q^{(-)}(x)]^{(1 - 2q)}}{2q-1 } { } _ { 2}f_{1}[q,2q-1,2q , \xi \cdot [ e_q^{(-)}(x)]^{-1 } ] \right ] \
, , \label{eq : sm}\ ] ] where @xmath79 is the gauss hypergeometric function .
so , we assume that the presence of @xmath76 instead of @xmath77 in the expression for @xmath60 in @xcite was a typo , which was wrongly assumed also by @xcite .
only when using @xmath77 in the occupation number , eq .
( [ eq : nq ] ) , one gets exactly the tsallis expression of the entropy written in @xcite ( apart from the term @xmath58 which appears for the reasons discussed above ) .
the entropy of eq .
( [ eq : sm ] ) has also the peculiar issue that it leads to negative values in the zero temperature limit , while the entropy of @xcite does not , see fig .
[ fig : entropy ] .
on the other hand , it is worth mentioning that the entropy of @xcite ( eqs .
( 21 ) and ( 24 ) in that reference ) is not the same as the one in @xcite ( eq . ( 4 ) in that reference ) for @xmath19 , as it can be easily seen by considering that in the former case they define the q - log function the same for all values of its argument .
this means that these authors are not consistent with each other . a computation of the occupation number derived from the entropy of @xcite for @xmath19 leads to the result @xmath80^{-1 } - \xi}\bigg)^{q}\,,\ ] ] which is different from the expression of @xcite and from the one the authors of @xcite claim . with these results
we conclude that the partition function in eq .
( [ partitionfunc ] ) does represent the relevant function for an ideal quantum gas in tsallis statistics .
observe that for @xmath47 and @xmath34 sufficiently high the partition function defined here is similar to those used in refs .
@xcite to extend hagedorn s theory to non extensive statistics .
therefore definition given here is in accordance with the non extensive self - consistent thermodynamics . in the following
we explore some of the features of the thermodynamical systems described by the partition function written in eq .
( [ partitionfunc ] ) .
there is an intrinsic gain in the knowledge of the non extensive partition function , since it is closer to the methods of statistical mechanics .
however , in order to go deeper into more fundamental aspects one needs knowledge about nonperturbative qcd that are not available at present .
before studying the properties of a thermodynamically equilibrated hadronic system in the non extensive thermodynamics we have to find the region where this system can exist .
in fact , due to the transition from confined to deconfined regimes , the hadronic matter can be found only below the phase transition line .
there are different proposals for the conditions determining the transition line .
cleymans and redlich @xcite pointed out that the transition line can be determined by the condition that @xmath81 .
this result was obtained through a systematic analysis of particle yields from hep experiments . in @xcite
the transition line was obtained in terms of the total baryon density with the help of the hadron resonance gas model . in @xcite
, the freeze - out condition was determined from an interpolation between a resonance gas used at low densities and repulsive nucleonic matter at low temperatures . in @xcite ,
the author proposed that the chemical freeze - out of hadrons in heavy - ion collisions is characterized by the entropy density and its value was taken from lattice qcd ( lqcd ) calculations at zero chemical potential . in an interesting and more recent analysis @xcite ,
higher order multiplicity moments obtained with the hadron resonance gas were used in the calculation of the standard deviation , the variance ( or susceptibility ) , the skewness and the kurtosis , quantities which are related to the cumulants @xcite , that can be experimentally determined and are also of interest in lqcd calculations .
the chemical freeze - out curve was then described in terms of the susceptibility of the system . in the present work we use two different methods to find the transition line : the condition that @xmath81 , as in refs .
@xcite , and the entropy density condition proposed in ref .
these conditions were compared with a systematic analysis of particle yields from hep experiments , where the boltzmann s statistics was used , but since yields are calculated by the integration over all energy states , and considering that the differences between the distributions obtained through boltzmann statistics and tsallis statistics are relevant only in the high energy tail , we can assume with some confidence that the same relation holds in the non extensive case .
of course this assumption must be checked by a similar analysis of experimental data , now using the non extensive formulas derived here , a task that is beyond the scope of the present work .
however we present below some evidences that this hypothesis is correct . in the following
we assume that the entropic index , @xmath4 , is a fixed property of the hadronic matter with its value determined in the analysis of @xmath2-distributions and in the study of the hadronic mass spectrum in ref .
@xcite , so we set @xmath82 , although in some cases we analyze the behavior of some quantities for different values of @xmath76 . in the following we refer to boltzmann - gibbs statistics also by @xmath83 .
the system of interest here is a gas composed by different hadronic species in thermodynamical and chemical equilibrium .
the partition function is then given by @xmath84 where @xmath85 refers to the chemical potential for the _ i - th _ hadron .
the lowest - lying hadrons considered in our calculations are taken from the particle data group @xcite , and some of them are presented in tables [ tablemesons ] and [ tablebaryons ] .
the total numbers of hadronic states considered are 808 for mesons and 1168 for baryons
( @xmath86 anti - baryons ) , including degeneracies , which correspond to a maximum value of the mass about @xmath87 and @xmath88 respectively .
the computation will be performed by restricting the ensemble summation in eq .
( [ eq : logz ] ) to zero strangeness ( see @xcite for details )
. we will focus first on a study of the phase transition by assuming @xmath47 for all mesons and considering that all baryons have the same chemical potential value @xmath89 .
after that , the effect of a nonzero value of the chemical potential for pions will be explored as well .
we have not compared our results with lqcd data because these data are available only up to baryon chemical potentials much lower than the ones we need to cover the whole @xmath90 plane @xcite . in @xcite ,
for instance @xmath91 reaches 2 , what means that @xmath6 is of the order of @xmath92 mev corresponding to @xmath93 of the order of @xmath94 mev .
in addition , one can observe in fig .
[ txmu ] that for small chemical potential the effective temperature remains practically unchanged , so the theoretical results would not differ sensitively from those obtained with the usual boltzmann - gibbs statistics after the temperature scalability observed in this work is taken into account . for the case of null chemical potential a comparison with lqcd
was already done in ref .
@xcite . the phase transition line obtained with the energy per particle condition , as described above ,
is reported in fig .
[ txmu ] ( left ) , and it is compared to the boltzmann s result , corresponding to @xmath95 . for the sake of clarity we refer to the temperature obtained with bg statistics as @xmath96 and to the temperature obtained with tsallis statistics as @xmath97 .
the relation between @xmath97 and @xmath96 was already investigated in ref .
@xcite , and it was found that for a fixed chemical potential there is a linear correspondence between both quantities . the relation between both quantities when varying @xmath93 is shown in fig .
[ txmu ] ( right ) . for the sake of comparison
the transition line obtained through the entropy condition was calculated , and its results are shown in fig .
[ txmu ] . we observe that both conditions lead to transition lines that are in agreement with the available experimental data .
the main differences are in the high chemical potential region , where the entropy condition leads to lower temperature , approaching zero around @xmath98 , while the energy per particle condition gives higher temperatures in the same region .
.list of the lowest - lying mesons used in eq .
( [ eq : logz ] ) .
we include some of their properties : mass , strangeness ( @xmath99 ) and degeneracy ( @xmath100 ) .
the corresponding anti - mesons with @xmath101 are not shown in the table but they are considered as well in the computation .
the dots indicate that heavier mesons are included in the computation , although they are not explicitly shown in this table due to lack of space . [
cols="^,^,^,^,^,^,^,^,^",options="header " , ] in fig .
[ pxt ] ( left ) we plot the pressure as a function of the temperature for the chemical potentials fixed at the value @xmath102 for the baryonic chemical potential and @xmath47 for all mesons with @xmath83 and @xmath103 .
we observe that the pressure increases faster in the non extensive case , @xmath103 , than in the extensive one , @xmath83 . in fig .
[ pxt ] ( right ) we show the results for the pressure as a function of the energy density for @xmath104 . also here
we observe that the pressure increases faster in the case of @xmath103 in comparison with the case @xmath83 .
we show in fig .
[ p3d ] the pressure as a function of the temperature and chemical potential in the regime @xmath105 and @xmath106 , in boltzmann - gibbs ( left ) and in tsallis with @xmath103 ( right ) .
the more rapid growth of the pressure in tsallis statistics in comparison with boltzmann - gibbs is clear for all the values of @xmath93 displayed , but this effect becomes much stronger for @xmath107 , see also fig .
[ pxt ] ( left ) . from figs .
[ pxt ] and [ p3d ] one can conclude that the eos for hadronic matter obtained with the tsallis statistics is harder than the one obtained from the bg case .
up to now we have studied the equation of state including the spectrum of hadrons in table [ tablemesons ] and [ tablebaryons ] .
it would be interesting to analyze also the case in which only protons , neutrons and possibly pions contribute to the equation of state , all of them with nonzero chemical potential .
we have studied the finite pion chemical potential in two different cases : a ) @xmath108 , b ) @xmath109 ; and @xmath110 , @xmath111 in both cases .
these values are relevant for the study of protoneutron stars @xcite .
the phase transition lines in boltzmann - gibbs and in tsallis statistics in the regime of high baryonic chemical potential are plotted in fig .
[ txmumu ] .
note that the effect of finite pion chemical potential is to increase the transition line to higher values of temperature , and the curves with @xmath112 and @xmath113 coincide , as we have not introduced electrons in the computation .
we show in fig .
[ fig : pmubmu ] the result for the equation of state considering zero and finite pion chemical potential .
the effect of pions is to increase the values for the pressure with respect to the case with only protons and neutrons .
when considering a nonzero value for the pion chemical potential this leads to a noticeable effect on the eos , as it becomes harder either in boltzmann - gibbs or in tsallis statistics , see fig .
[ fig : pmubmu ] ( right ) .
in this work we developed the non extensive thermodynamics for an ideal quantum gas for both bosons and fermions from the partition function defined here for the first time .
then we showed that the partition function and the thermodynamics derived from it is equivalent to the thermodynamics derived from the entropy proposed by conroy , miller and plastino @xcite , and also by cleymans and worku @xcite when @xmath0 . for @xmath1 and for fermions some inconsistencies of previous references
were addressed , and our result is fully thermodynamically consistent . in the limit of high energies ,
the partition function is in accordance with that proposed in ref .
@xcite , and thus the partition function defined here is self - consistent in the sense proposed by hagedorn .
some thermodynamical functions are derived from the partition function for hadronic systems with different values of chemical potentials .
particularly we analyze how pressure and energy densities vary when the entropic index or the chemical potentials vary , and we obtain the chemical freeze - out line by using two different hypotheses . a discussion about a discontinuity observed in the first derivatives of the partition function is done for the first time .
the results presented in this work can be applied to stellar matter , where high pressures are necessary to compensate for the gravitational force so that protoneutron star stability and correct macroscopic properties are attained .
this study is performed in other work @xcite , and the results compared with the ones existing in the literature @xcite .
this work has been supported by plan nacional de altas energas ( fpa2011 - 25948 ) , junta de andaluca grant fqm-225 , generalitat de catalunya grant 2014-sgr-1450 , spanish mineco s consolider - ingenio 2010 programme cpan ( csd2007 - 00042 ) , centro de excelencia severo ochoa programme grant sev-2012 - 0234 , by cnpq ( brazil ) and fapesc ( brazil ) under project 2716/2012,tr 2012000344 , and fapesp ( brazil ) under grant 2013/24468 - 1 .
a.d . acknowledges the support from cnpq under grant 305639/2010 - 2 .
would like to thank the instituto de fsica of the universidade de so paulo for their hospitality and support during the completion of parts of this work .
the research of e.m . is supported by the juan de la cierva program of the spanish mineco grant jci-2010 - 06699 .
99 c. tsallis , j. stat .
52 ( 1988 ) 479 .
i. bediaga , e.m.f .
curado e j.m
. de miranda , physica a 286 ( 2000 ) 156 .
r. hagedorn , nuovo cimento suppl . 3 ( 1965 ) 147 .
b. abelev
_ et al . _
( alice collaboration ) , phys .
b * 717 * , 162 ( 2012 ) .
b. abelev _
( alice collaboration ) , eur .
j. c * 72 * ( 2012 ) 2183 .
k. aamodt _
( alice collaboration ) , eur .
j. c * 71 * ( 2011 ) 1655 .
v. khachatryan _
( cms collaboration ) , jhep 05 ( 2011 ) 064 .
( atlas collaboration ) , nucl .
b 850 ( 2011 ) 387 - 444 .
r. hagedorn , lect .
notes phys .
* 221 * , 53 ( 1985 ) .
n. o. agasian , phys .
b * 519 * , 71 ( 2001 ) .
w. broniowski , w. florkowski and l. y. .glozman , phys .
d * 70 * , 117503 ( 2004 ) .
a. tawfik , phys .
d * 71 * , 054502 ( 2005 ) .
e. megias , e. ruiz arriola and l. l. salcedo , phys .
* 109 * , 151601 ( 2012 ) .
e. megias , e. ruiz arriola and l. l. salcedo , nucl .
suppl . * 234 * , 313 ( 2013 ) .
e. megias , e. ruiz arriola and l. l. salcedo , phys .
d * 89 * , 076006 ( 2014 ) .
a. deppman , physica a 391 ( 2012 ) 6380 .
see also a.
deppman , physica a 400 ( 2014 ) 207208 .
l. marques , e. andrade - ii e a. deppman , phys .
d 87 ( 2013 ) 114022 . i. sena e a. deppman , eur .
j. a 49 ( 2013 ) 17 .
i. sena e a. deppman , aip conf .
1520 , ( 2013 ) 172
. j. cleymans e d. worku , j. phys .
( 2012 ) 025006 .
a. deppman , j. phys .
g : nucl . part .
phys . * 41 * , 055108 ( 2014 ) .
n. k. glendenning , compact stars , springer - verlag , new - york , ( 2000 ) .
p. haensel , a. y. potekhin , d. g. yakovlev : _ neutron stars , equation of state and structure _ , springer , new york ( 2006 ) .
b. d. serot and j. d. walecka , adv .
phys . 16 , 1 ( 1986 ) ; j. boguta and a. r. bodmer , nucl . phys . a 292 , 413 ( 1977 ) .
|
the non extensive thermodynamics of an ideal gas composed by bosons and/or fermions is derived from its partition function for systems with finite chemical potentials .
it is shown that the thermodynamical quantities derived in the present work are in agreement with those obtained in previous works when @xmath0 .
however some inconsistencies of previous references are corrected when @xmath1 . a discontinuity in the first derivatives of the partition function and its effects
are discussed in detail .
we show that at similar conditions , the non extensive statistics provide a harder equation of state than that provided by the boltzmann - gibbs statistics .
|
the pion _ electromagnetic _ form factor is defined by @xmath15 where @xmath16 is the electromagnetic current , @xmath17 .
the form factor is normalized as @xmath18 . as function of the complex variable @xmath19 ,
the form factor @xmath20 has a cut in the complex @xmath21-plane starting at the two - pion threshold @xmath22 which corresponds to two - pion intermediate states .
there are also cuts related to @xmath23 intermediate states and multimeson states ( @xmath24 , etc ) .
the form factor in the timelike region @xmath25 is @xmath26 where @xmath27 is the phase . for the theoretical description of the form factor in different regions of momentum transfers
different theoretical approaches are used . at large spacelike momentum transfers ,
@xmath28 , perturbative qcd ( pqcd ) gives rigorous predictions for the asymptotic behaviour of the form factor @xcite @xmath29 where @xmath30 is the qcd coupling parameter and @xmath31 mev is the pion decay constant defined by the relation @xmath32 as the spacelike momentum transfer becomes smaller , the form factor turns out to be the result of the interplay of perturbative and nonperturbative qcd effects , with a strong evidence that nonperturbative qcd effects dominate in the region @xmath33 gev@xmath11 @xcite . the picture based on the concept of constituent quarks which effectively account for nonperturbative dynamics
has proven to be efficient for the description of the form factor in this region ( see for instance @xcite ) . at large timelike momentum transfers , @xmath34 gev@xmath11 , @xmath20 can be obtained from the analytic continuation of the pqcd formula ( [ pqcd ] ) . at
small timelike momentum transfers the situation is more complicated since there dynamical details of the confinement mechanism are crucial .
quarks and gluons are no longer the degrees of freedom of qcd leading to a simple description of the form factor . at timelike momentum transfers we are
essentially in the region of hadronic singularities and typically one relies on methods based on hadronic degrees of freedom . in the region of interest to us here , @xmath35 gev@xmath11 , the lightest pseudoscalar mesons are most important .
there are many approaches to understand the behaviour of the pion form factor at timelike momentum transfers from 0 to 1.5 gev@xmath11 .
a time honoured approach is based on the vector meson dominance ( vmd ) model @xcite . in the simplest version of vmd
one assumes just the @xmath4-meson dominance , which leads to @xmath36 this simple formula works with a good accuracy both for small spacelike momentum transfers and timelike momentum transfers below the @xmath37 threshold : @xmath38 gev@xmath39 . for @xmath21 near the @xmath37 threshold one should take into account effects of the virtual pions . in this region momenta of the intermediate pions are small and a consistent description of the form factor is provided by chiral perturbation theory ( chpt ) @xcite , the effective theory for qcd at low energies . for higher @xmath21 , in the region of @xmath4 and @xmath14 resonances , a similar rigorous treatment of the form factor is still lacking , and one has to rely on model considerations .
contributions of the two - pion intermediate states may be consistently described by dispersion representations .
the application of dispersion relations has led to the famous gounaris - sakurai ( gs ) formula @xcite which takes into account @xmath4-meson finite width corrections due to the virtual pions @xmath40 the function @xmath41 corresponds to the two - pion loop diagram .
the corresponding feynman integral is linearly divergent , but its imaginary part is defined in a unique way . the real part is then reconstructed by a doubly - subtracted dispersion representation . the gounaris - sakurai prescription to fix the subtraction constants reads @xmath42 the phase of the form factor @xmath43 for the gs prescription agrees well with the experimental data in the region @xmath44 gev@xmath11 . but ( [ piff_b ] ) gives too small a value ( by @xmath45 ) for @xmath46 at @xmath21 around @xmath47 .
on the other hand , one can consider a simple vmd ansatz taking only the @xmath4-meson contribution into account .
this should be a good approximation in the region @xmath48 gev@xmath11 , except for the narrow interval near @xmath49 where the @xmath7 mixing effects are important @xcite .
the simple vmd ansatz is then very similar to ( [ piff_b ] ) , but with the numerator replaced by the @xmath50 transition matrix element : @xmath51 here @xmath52 and @xmath53 are defined according to @xmath54 where @xmath55 is the @xmath4-meson polarization and @xmath56 is the 4-momentum vector .
now @xmath46 from ( [ piff_c ] ) describes well the data for @xmath57 . but extrapolating the eq .
( [ piff_c ] ) to @xmath58 gives @xmath59 in gross violation of the normalization condition @xmath18 .
thus , neither ( [ piff_b ] ) nor ( [ piff_c ] ) can describe the form factor for all @xmath60 gev@xmath11 : namely , ( [ piff_b ] ) leads to a too small value of @xmath61 at @xmath62 , whereas the form factor given by ( [ piff_c ] ) is far above unity at @xmath58 .
there were many attempts to modify the vector meson dominance or to use related approaches in order to bring the results on the pion form factor in agreement with the data ( see @xcite and papers quoted therein ) .
the pion form factor in the region @xmath60 gev@xmath11 is one of the main sources for obtaining masses and coupling constants of vector mesons .
however , with different assumptions on the form of the vector - resonance contribution to the pion form factor one obtains different values of masses and couplings .
therefore a consistent description of the pion form factor in this region in terms of the low - lying mesons ( @xmath63 ) is crucial for extracting reliable values of these parameters .
interesting results have been obtained by the authors of @xcite who noticed that an effective momentum - dependent @xmath2 coupling appears in the framework of the effective lagrangian approach .
this momentum - dependent @xmath2 coupling considerably improves the description of the pion form factor at timelike momentum transfers in the region @xmath64 gev@xmath11 .
our analysis is similar in spirit to the approach of @xcite .
the goal of our paper is twofold : first , to analyse the pion form factor in a dispersion approach and to establish in this way a general form of the resonance contribution to the form factor , paying special attention to the analysis of the existing ambiguities .
second , to apply the results obtained to the extraction of the @xmath4 and @xmath14 masses and couplings from the pion form factor data .
we make use of a dispersion approach to the pion form factor in a model with @xmath0 , @xmath1 , @xmath65 , and gauge - invariant @xmath66 , @xmath67 and @xmath7 couplings .
our approach allows an exact resummation of pion and kaon loops .
we pay a special attention to the analysis of ambiguities related to subtractions in linearly divergent meson loop diagrams and show that they may be reabsorbed in the physical masses and the effective couplings in the expression for the pion form factor . after taking into account the @xmath7 mixing effects the pion form factor in the range @xmath68 gev@xmath11
is well described both in magnitude and phase by a formula which is similar to the vmd expressions ( [ piff_b ] ) and ( [ piff_c ] ) but avoids their pitfalls .
our model makes use of conventional methods of dispersion theory .
first we make an ansatz for the effective couplings of the pseudoscalar mesons , vector mesons and the photon .
these couplings are used in essence only to calculate the absorptive parts of the amplitudes .
the complete amplitudes are then obtained by dispersion relations and a dyson resummation .
we want to make clear from the outset that our effective couplings discussed below are not to be compared directly to the effective lagrangian of chpt @xcite and resonance theory in the framework of chpt @xcite .
we shall see , however , that our model , used as explained above , respects all the known results from chpt for the pion form factor .
thus our model can be seen as an alternative to the one of @xcite where chpt results are extended to @xmath20 in the range @xmath69 gev@xmath11 using again a resummation scheme . in our model pions
interact with the @xmath4-mesons and generate in this way the finite @xmath4 meson width .
are essentially saturated by the meson resonance exchange . ]
the @xmath13-meson is coupled to the conserved vector current of charged pions as follows : @xmath70 where @xmath71 is the conserved vector field describing the @xmath4 meson .
we denote in this section @xmath72 . matching to the one - loop chpt
@xcite leads to the relation @xmath73 the photon is coupled to the charged pion through the usual minimal coupling , @xmath74 we also add a direct gauge - invariant @xmath75 coupling of the form @xmath76 where @xmath77 this model is similar to the model of @xcite .
no @xmath78-parity violating @xmath65 or direct @xmath79 couplings are included at this stage .
c as explained above , we calculate the electromagnetic form factor in our model by the sum of the diagrams of fig .
[ figure-1:piff ] . summing all the pion loop insertions
, we obtain @xmath80 the quantites @xmath41 and @xmath81 correspond to one - loop @xmath2 and @xmath82 self energy diagrams generated by the pion loop .
the imaginary parts of these diagrams can be calculated by setting the intermediate pions on mass shell .
the full functions @xmath41 and @xmath81 are constructed from their imaginary parts by means of the spectral representation with a suitable number of subtractions and by adding the corresponding subtraction constants . for the @xmath37 intermediate states the imaginary parts of the functions @xmath41 and @xmath81 satisfy the relations @xmath83 where @xmath84 for the realistic description we have to take into account also contributions of @xmath85 and @xmath86 intermediate states . the coupling constant @xmath87 can not be measured directly .
we use the relation @xmath88 which is valid in the su(3 ) limit . repeating the procedure described above , summing the pion and kaon loops we find with ( [ gsu3 ] ) @xmath89 and
hence @xmath90 it follows from ( [ rel1 ] ) that the difference @xmath91 is a polynomial in @xmath21 determined by the subtraction conditions .
hence the numerator of the pion form factor ( [ fff ] ) is also a real polynomial .
therefore , the phase of the form factor is completely determined by the denominator .
the latter is the usual propagator of the @xmath4 meson with the finite width corrections taken into account .
let us now consider subtraction constants .
the function @xmath81 describes the coupling of the pion to the conserved electromagnetic current .
therefore we must set @xmath92 such that the charge of the pion remains unrenormalized by higher order corrections .
the function @xmath41 determines the behaviour of the @xmath37 elastic @xmath93 partial wave amplitude in which the @xmath4-meson pole is known to be present in the zero - width limit .
therefore , we require meson mass . notice that various definitions of the mass and the width of a resonance are used in the literature .
for a different definition see e.g. @xcite ] @xmath94 let us point out that this is essentially our definition of the @xmath4 meson mass .
one should take into account that various definitions of the mass and the width of a resonance of nonzero width without a loss of generality the second subtraction constant may be fixed by setting @xmath95 any other condition would just lead to rescaling of the parameters in the formula for the form factor .
thus , in a model we are considering when pions interact via a single @xmath4-meson exchange the general expression for the form factor incorporating subtraction ambiguities in the @xmath37 and @xmath96 loop diagrams contains three constants @xmath47 , @xmath97 , and @xmath53 @xmath98 here @xmath99 @xmath100 is defined by ( [ is ] ) , and @xmath101 where v.p . means the principle value .
let us point out that the numerator of the form factor in ( [ rhodominance ] ) is not a constant , but a linear function of @xmath21 .
this @xmath21-dependence appears as the direct consequence of the current conservation .
( [ rhodominance ] ) can be written in the form of the modified gs formula meson mass , whereas eq .
( 5.23 ) contains both bare and physical masses . ] @xmath102 with the effective @xmath21-dependent @xmath103 coupling constant @xmath104 two remarks are in order here : first , one should be careful with the interpretation of this result : as is clear from ( [ bgammarho ] ) , there is no direct transition of the @xmath4 meson to the _ real _ photon as a consequence of the gauge invariant @xmath66 coupling @xcite . on the other hand ,
the _ effective _ coupling @xmath105 is clearly nonzero at @xmath58 .
therefore the pion form factor _ looks _ as the direct @xmath66 coupling also for the real photon .
this is just the usual vector meson dominance .
the latter thus emerges as the direct consequence of our assumption that the vector meson couples to the same pion current as the photon . for further discussions of the relationship between vmd and gauge invariance we refer to @xcite .
if we use the chpt relation ( [ chpt ] ) , which agrees perfectly with the measured value of @xmath106 , then ( [ frhoeff ] ) leads to an interesting relation @xmath107 note that the phase of @xmath20 in ( [ piff_our ] ) is still given by eq .
( [ phase ] ) and is completely determined by the function @xmath41 .
second , it should be noted that the @xmath4-meson contribution to the pion form factor given by ( [ piff_our ] ) does not fall down sufficienty fast in the limit @xmath108 .
thus , for a realistic description of the pion form factor in a broader range of timelike momentum transfers , coupling of pions to higher vector resonances @xmath109 etc should be added . however ,
if a vector - meson coupling to the photon is described by the gauge - invariant expression ( [ rhogamma ] ) , contributions of higher resonances to @xmath61 are negligible in the region @xmath110 gev@xmath11 ( see recent analyses @xcite ) .
we are interested only in this region and therefore do not take higher resonances @xmath109 etc into consideration .
in section [ section:2 ] we discussed the @xmath4 contribution to the pion form factor ( the bare @xmath4 plus the effects of the @xmath4-meson width due to the light - meson loops ) .
this analysis is sufficient for describing the pion form factor of the charged vector current using the cvc relation . for the electromagnetic pion form factor
it is necessary to take into account the @xmath7 mixing effects .
the @xmath14 is coupled to the pions and the photon similarly to the @xmath13-meson [ cf .
( [ rhopipi ] ) and ( [ rhogamma ] ) ] @xmath112 @xmath113 being a conserved vector field describing the @xmath14-meson and @xmath114 .
it has proven useful to classify various contributions to hadronic amplitudes according to their formal order in the @xmath115 expansion @xcite , where @xmath116=3 is the number of colours in qcd . in the language of the @xmath115 expansion the analysis of the previous section corresponds to taking into account the leading order @xmath115 process ( resonance in a zero - width approximation ) and the subleading @xmath117 effects of the meson loops . and of order @xmath118 of the momentum expansion . ] performing a resummation of these meson loops gave our dispersion description of the form factor .
a corresponding treatment of the @xmath7 mixing effects then requires taking into account the leading and subleading @xmath115 effects as well . to leading order in @xmath119
, meson loops do not contribute and therefore the only effect is the direct @xmath7 transition described in terms of the direct coupling ( fig .
2 ) . at subleading @xmath115 order several meson loop diagrams shown in fig . 2 emerge .
we make use of spectral representations for loop diagrams , i.e. we calculate directly the imaginary parts and then reconstruct the full function by means of the spectral integral with a relevant number of subtractions .
subtraction constants are then either fixed by physical constraints or determined by the experimental data .
let us point out an important feature related to our dispersion calculation : the direct @xmath7 coupling ( a leading @xmath115 process ) and the _ real part _ of the @xmath7 mixing loop diagrams at @xmath120 ( a subleading @xmath115 process ) contribute to the form factor precisely the same way , such that only their sum has a physical meaning .
we therefore account for the net effect of these two contributions by a single subtraction constant and do not consider the direct @xmath7 coupling separately .
we have analysed in section [ section:2 ] the @xmath4-meson self - energy function @xmath121 which determines the propagator of the interacting @xmath4 meson .
let us now discuss a similar self - energy function of the @xmath14-meson @xmath122 and the off - diagonal @xmath7 function @xmath123 which describes the @xmath7 mixing .
the function @xmath122 determines the @xmath14-propagator @xmath124 in the absence of the @xmath7 mixing effects . the main contribution to @xmath125
is given by the three - pion intermediate states .
this @xmath125 should then be inserted into a dispersion integral to get @xmath122 . however , because of the small width of the @xmath14-resonance , it is sufficient for our analysis to consider as a simple ansatz a constant @xmath122 @xmath126 possible processes which contribute to the @xmath7 mixing amplitude @xmath127 are shown in fig . [
fig : mixing1 ] .
the coupling contants which determine the relative strength of the diagrams in fig .
[ fig : mixing1 ] are shown in table [ table : pdg ] . [ cols="^,^,^,^,^,^,^,^",options="header " , ] +
we are grateful to v. anisovich , a. donnachie , m. jamin and e. de rafael for useful discussions and to j. urheim for correspondence on the cleo data .
the work was supported by the alexander von humboldt - stiftung and the german bundesministerium fr bildung und forschung under project 05 ht 9 hva3 .
30 a. v. efremov , a. v. radyushkin , jetp lett , * 25 * , 210 ( 1977 ) ; + s. brodsky , g. p. lepage , phys . lett . * 87b * , 359 ( 1979 ) . v. v. anisovich , d. i. melikhov ,
v. a. nikonov , phys . rev .
* d52 * , 5295 ( 1995 ) , phys . rev . * d55 * , 2918 ( 1997 ) .
f. cardarelli , e. pace , g. salme , s. simula , phys .
* b357 * , 267 ( 1995 ) ; d. melikhov , phys .
rev * d53 * , 2460 ( 1996 ) . j. j. sakurai , ann . phys . * 11 * , 1 ( 1960 ) ; m. gell - mann , f. zakhariasen , phys . rev . * 124 * , 953 ( 1961 ) .
j. gasser and h. leutwyler , nucl .
b250 * , 465 ( 1985 ) .
g. j. gounaris and j. j. sakurai , phys .
* 21 * , 244 ( 1968 ) .
j. lefrancois , in _ proceedings of the 1971 internatinal symposium on electron and photon interactions at high energies _ , cornell university , ithaca , n.y .
, august 23 - 27 , 1971 .
f.klingl , n. kaiser , and w. weise , z. phys .
bf a356 , 193 ( 1996 ) .
m. benayoun , h. b. oconnell , and a. g. williams , phys . rev . *
d59 * , 074020 ( 2001 ) .
a. pich and j. portoles , phys
. rev . * d63
* , 093005 ( 2001 ) ; + j. a. oller , e. oset , j. a. palomar , phys . rev . * d63 * , 114006 ( 2001 ) ; + j. j. sanz - cillero and a. pich , hep - ph/0208199 . g. ecker et al , nucl .
b321 * , 311 ( 1989 ) .
n. m. kroll , t. d. lee , b. zumino , phys . rev . * 157 * , 1376 ( 1967 ) .
a. v. anisovich , v. v. anisovich , a. v.sarantsev , z. phys . *
a359 * , 173 ( 1997 ) .
d. melikhov , phys . lett .
* b516 * , 61 ( 2001 ) .
particle data group ( k. hagiwara et al . ) , phys . rev . *
d66 * , 010001 ( 2002 ) .
m. gourdin , in _ hadronic interactions of electrons and photons _ , j. cumming and h. osborn , eds .
, academic press , london , p. 395
a. donnachie and h. mirzaie , z. phys .
* c33 * , 407 ( 1987 ) ; g. kulzinger , h. g. dosch , h. j. pirner , epj * c7 * , 73 ( 1999 ) .
r. barate et al , z. phys .
* c76 * , 15 ( 1997 ) .
s. d. protopopescu et al , phys . rev . * d7 * , 1279 ( 1973 ) ; + p.
estabrooks and a. d. martin , nucl . phys . *
b79 * , 301 ( 1974 ) .
s. amendolia et.al .
, nucl . phys . *
b277 * , 168 ( 1986 ) .
r. r. akhmetshin et al .
, hep - ex/9904027 . s. anderson et al , phys . rev .
* d61 * , 112002 ( 2000 ) .
|
we consider a model with @xmath0 , @xmath1 , and gauge - invariant @xmath2 couplings and obtain the pion form factor @xmath3 at timelike momentum transfers by resummation of pion and kaon loops .
we use a dispersion representation for the loop diagrams and analyse ambiguities related to subtraction constants .
the resulting representation for @xmath3 is shown to have the form of the conventional vector meson dominance formula with one important distinction - the effective @xmath4-meson decay constant @xmath5 turns out to depend on the momentum transfer squared @xmath6 . for the electromagnetic pion form factor
we include in addition the @xmath7 mixing effects .
we apply the representations obtained to the analysis of the data on the pion form factors from @xmath8 annihilation and @xmath9 decay in the region @xmath10 gev@xmath11 and extract the @xmath12 , @xmath13 and @xmath14 masses and coupling constants .
|
little is known about the influence of individual element abundances on the integrated flux of a stellar population . as present
, it is impossible to derive an accurate age within 10% from the integrated light of a cluster - like single - age and single - abundance stellar population , and the primary reason is the complication due to abundance ratio effects @xcite .
the secondary reason is that the input ingredients and choices made in even the most sophisticated stellar evolution calculations induce scatter in the results @xcite .
the tertiary reason is the set of difficulties associated with stellar flux knowledge , such as colors , line strengths , or spectra , that are needed at each stellar evolutionary phase to represent the component stars .
additional uncertainties exist , such as dust extinction , stellar rotation and activity , the blue straggler frequency , and the effects of close binaries .
paper i in this series , @xcite , explored the effects of 12 chemical mixtures on the @xmath0 , @xmath1 , and lifetimes of stars along stellar evolutionary isochrones and upon the opacities needed to calculate the stellar models .
the mixtures explored were solar , @xmath2-element enhanced , and ten cases where only one element at a time was enhanced , for elements c , n , o , ne , mg , si , s , ca , ti , and fe .
the mixtures were re - scaled so that the mass fraction of heavy elements , @xmath3 , was constant . and variable @xmath4 and @xmath5 but constant [ fe / h ] with 0.2 dex more carbon , 0.3 dex more nitrogen and 0.3 dex more oxygen ( see @xcite ) . ]
the conclusions from that paper could be summarized by splitting the elements that were investigated into three categories .
`` displacers '' include o , c , n , and ne . these are elements that are abundant but supply less opacity per unit mass than heavier elements . at fixed @xmath3 , boosting a displacer element will therefore decrease the opacity .
this leads to shorter stellar lifetimes and hotter stars .
`` boosters '' include mg and si .
these elements are good opacity sources , so that boosting a booster will increase opacity and make cooler stars that live somewhat longer .
`` oddball '' elements defy these trends , and include ca and ti , both of which make cooler stars , that , nevertheless , have shorter lifetimes . another oddball is s , which decreases low - temperature opacity but increases high - temperature opacity so that its ultimate effect on temperatures and lifetimes is small .
an element worth special mention is fe , because its high - temperature opacity has considerable structure .
one of its opacity peaks corresponds to the temperatures characteristic of the outer edge of the convective core that develops in stars slightly more massive than the sun .
increasing the fe abundance therefore strongly couples to the onset of convective core overshooting effects , and leads to a strong luminosity effect in the region of the main sequence turnoff .
the temperature , luminosity , and lifetime effects are illustrated in paper i , but further consequences will be illuminated in this paper . besides isochrones ,
the other major ingredient in any population synthesis effort is some representation of stellar flux for each star in the isochrone .
the effects of individual element abundances on the spectra of stars have been studied for years , and enormous progress has been made in finding the structure of stellar atmospheres and calculating the emergent flux .
the two most - cited programs for constructing model atmospheres of stars are atlas @xcite and marcs @xcite and we use results from these code in this work . for spectral synthesis of the emergent flux we use synthe @xcite ,
fantom @xcite , and ssg @xcite . for ordinary population synthesis , empirical spectra or colors
can be used @xcite , but for investigating element - by - element effects , using synthetic spectra is clearly the way forward .
investigations into the effects of individual elements on stellar spectra , but with the intent of applying the results to galaxy spectra , include @xcite and @xcite , who gauged the effects of ten individual elements on the lick system of 25 pseudo - equivalent width indices @xcite .
@xcite investigated 24 elements in spectra that ran from 3500 to 9000 with velocity smoothing appropriate for dynamically hot systems such as elliptical galaxies .
the present paper combines the new isochrones described in paper i with greatly extended and updated synthetic spectra in order to obtain _ ab initio _ population synthesis models as a function of _ individual _ element abundances - elements over a range of metallicity . ] .
all parts of the models , including high temperature and low temperature opacities , stellar evolutionary models , and synthetic spectra , include the altered abundances in the same way . at present
the grid includes only solar @xmath3 , but calculations are underway to extend to many different abundances .
this paper describes the new spectra and their color index results in @xmath6 2 , lick index results in @xmath6 3 , concluding with a discussion and summary section .
the new isochrones were described in paper i @xcite , and we refer the reader to that paper for details .
twelve chemical mixtures were explored self - consistently with both high - temperature and low - temperature opacities adjusted properly , and evolution is complete only to the end of the red giant branch .
it should be emphasized that the models are thus _ incomplete and should not be used blindly when comparing to real stellar populations _ until the helium - burning phases are properly incorporated .
this is planned for the ongoing improvement of these models . in careful , differential ways ,
the present models give us important pointers to the behavior of stellar populations with variable element mixtures , but the absolute values of colors and indices are not to be trusted .
the mixtures explored were solar , @xmath2-element enhanced , and ten cases where only one element at a time was enhanced ( c , n , o , ne , mg , si , s , ca , ti , and fe ) .
the mixtures were re - scaled so that the mass fraction of heavy elements , @xmath3 , was held constant .
new for the present work is a collection of synthetic spectra from three different sources .
spectra for the coolest stars , at 3000 and 3190 k for log @xmath7 = 0 , were calculated using the fantom synthesis code and @xcite model atmospheres . at 3500 and 3750
k for log @xmath7 = 0 and 5 spectra were calculated using the fantom synthesis code and atlas model atmospheres exactly as described in @xcite , from 3000 to 10000 in 0.02- steps .
spectra for the medium - temperature stars were calculated using the ssg synthesis code and marcs model atmospheres from 3000 to 10000 in 0.01-steps .
there were 9 temperatures in this range , from 4000 k to 6000 k , in 250 k steps .
gravities of log @xmath7 = 4.5 and 2.5 were used for all nine temperatures , but log @xmath7 = 0.5 was included for the three coolest .
spectra for hot stars , at 7000 k , 8000 k , 10000 k , and 20000 k , were calculated using the synthe synthesis code starting from atlas model atmospheres in high resolution with logarithmic wavelength spacings .
gravities of @xmath8 = 4.5 and 2.5 were used except the 20000 k models , where 4.5 and 3.0 were used .
the lines lists were not completely homogeneous between the three regimes , although they share much in common .
the hot stars were computed with r. kurucz ( kurucz.harvard.edu ) line lists , the medium - temperature stars included custom modifications by r. bell , m. tripicco , and m. houdashelt , and the cool stars benefitted from the tio re - scalings described in @xcite .
these spectra were then rebinned to a common wavelength range of 3000 to 10000 and a common linear wavelength binning of 0.01 per flux point for future high - resolution studies , and then rebinned again to a common linear wavelength binning of 0.5 per flux point for more convenient use for working with colors and spectral indices . for each star in the grid of spectra described above , multiple realizations of the same spectrum
were calculated with single - element abundance adjustments , so that the effects of c , n , o , ne , mg , si , s , ca , ti , and fe could be individually gauged . for neon enhancement , we use the scaled - solar spectra because of the small effect on the optical stellar spectrum from neon . any isochrone - level temperature , luminosity , or stellar lifetime effects that we depicted in paper i from neon are , however , included .
thus , at solar metallicity , 35 gravity and temperature combinations @xmath9 10 element adjustments ( counting the non - adjusted solar abundance spectrum ) makes a total of 350 spectra .
another convenience is the definition of @xmath10 , which stands for `` generic heavy element . ''
usually , one refers to relative abundance changes relative to iron , such as [ o / fe ] or [ mg / fe ] .
however , this stops making sense in the case of wanting to change fe with respect to other elements ; i.e. , using [ fe / fe ] leads to nonsense .
so we use the `` generic heavy element '' r to stand for `` all heavy elements that remain scaled with the solar mixture , where all other element tweaks must be specified . '' with that notation , [ fe / r ] is perfectly acceptable , and the specification of constant [ r / h ] means that all elements except the one ( or , more generically , ones ) under consideration are held in solar lockstep .
an interpolation routine was constructed so that spectra of arbitrary temperature , gravity , and abundance mixture could be produced on demand .
linear interpolation in the log of the flux seemed to give the most predictive results .
interpolation was done for all input variables : temperature , log @xmath7 , and the whole set of abundance parameters .
tests in which one star in the grid was compared with its interpolated version calculated from the flanking temperature grid points indicate that , within the dense 4000 - 6000 k part of the grid , interpolations yield better than 1% accuracy at any given flux point , since the spectra change more with temperature than with the other parameters . given the nonlinear temperature response of many lines , e.g. , @xcite , this is encouragingly good .
doubtless , in the future , we will demand a denser temperature grid , but 1% accuracy is better than we need for the present study . toward the more fundamental question of how well the synthetic spectra
match real spectra , there is little we can add to discussions in @xcite .
@xcite did discover some too - strong lines due to chromium ( that is , mistakes in the line list ) via stellar comparisons , but these lines do not affect the present conclusions at all .
more worrisome is silicon because the features that show si effects are mostly sih lines in the blue , but these line lists may be immature ( r. peterson , private communication ) . all conclusions regarding si
should be regarded with great suspicion until this question is satisfactorily resolved . in figures 1 and 2
, we compare the high resolution spectra of the sun and arcturus with the grid - interpolated synthetic spectra near the regions of the h@xmath11 and the mg @xmath12 lines , respectively . for the solar parameter we employ t@xmath13 = 5777 k , @xmath14 , and [ fe / h ] = 0 , and for arcturus t@xmath13 = 4290 k , @xmath15 , and [ fe / h ] = @xmath160.7 with [ @xmath2/fe ] = + 0.4 dex ( e.g. , @xcite ) .
the top panels are the solar spectra and the bottom panels are arcturus .
the thicker ( black ) lines are the observations and the thinner ( red ) ones are the synthetic .
the degraded low - resolution spectra , gaussian - smoothed to 8 fwhm or about 200 km s@xmath17 somewhat better than the best lick - system resolution , are overlaid on top of the high - resolution spectra .
we investigate the individual element effects over a rather broad wavelength region for the lick indices and broadband colors in this paper . in this context
, it is useful to find that the notable mismatches between the synthetic and the observed spectra in the high - resolution comparisons are smeared out and become subtle at low [email protected] ] .
table [ tab : color1 ] presents some comparisons between synthetic colors and empirical color behavior for the stars in the temperature range 4000 to 6000 k. columns 2 - 5 refer to _ synthetic _ spectra , and columns 6 - 8 refer to the _ empirical _ calibration of @xcite .
the synthetic spectra were converted to @xmath19 colors as in @xcite using bessell s ( 1990 ) filter responses zeroed to vega s colors .
there is at least a 10 mmag uncertainty just from that procedure , and probably at least 20 mmag for @xmath20 .
the `` dwarf '' is @xmath21 , and the `` giant '' is @xmath22 .
columns 2 and 6 present the colors for the star listed , synthetic and empirical , respectively .
the most serious mismatches are the @xmath20 color for the 6000 k dwarf and for the 4000 k and 5000 k giants .
these facts suggest some room for improvement in model atmospheres and line lists ( especially at shorter wavelengths ) . the remaining columns present color changes induced by either an abundance increase of 0.3 dex or a temperature increase of 250 k ( color changes in milli - magnitudes ) .
column 4 is a complete scaled - solar enhancement of 0.3 dex in every element and with new atmospheres calculated , while column 3 is a 0.3-dex enhancement of only the 24 elements that are explicitly traced in the spectral library , a superset of those followed in the isochrone library .
comparison of columns 3 and 4 indicates that , for the heart " of the spectral library between 4000 and 6000 k , the sum of the available one - by - one element tweaks approximately equals the scaled - solar analog operation .
the list of 24 elements in linear combination thus appears to approximate the full - blown calculation of a spectrum in most cases .
this allows the approximate calculation of a new spectrum at arbitrary composition in an eyeblink rather than many minutes for a whole new synthetic spectrum calculation . in general ,
the synthetic colors are within a few hundredths of a magnitude of the empirical colors and track the empirical color responses to abundance and temperature in an approximate way .
the worst outlier is the @xmath20 color for the 6000 k giant , in which the color gets redder rather than bluer with increasing temperature ( see also @xcite ) .
strong conclusions are not possible with this comparison , but the agreement is fairly encouraging
. tables [ tab : color2 ] and [ tab : color3 ] show the spectral effects of element - by - element enhancements of 0.3 dex ( except carbon which is increased by 0.15 dex , see footnote # 4 ) on some selected color indices for the stars in the temperature range 4000 to 6000 k. in table [ tab : color2 ] the elements were rescaled to constant heavy - element fraction @xmath3 , while in table [ tab : color3 ] @xmath3 was allowed to increase .
one technical subtlety is that , since neon is not tracked in the spectral library , the only effects that appear in table [ tab : color2 ] for ne are due to the re - scaling , and there are no effects at all in table [ tab : color3 ] .
table 3 can be visually explained by taking a look at the [email protected] ] .
figure 3 displays one example . here
the flux ratio between 0.3 dex mg - enhanced spectra and that of the solar - scaled are shown for a 4000 k giant star .
the @xmath23 becomes bluer ( e.g. , more @xmath24-band flux over @xmath25-band flux ) and @xmath26 , @xmath27 become redder ( e.g. , comparably more @xmath10- and @xmath28-band fluxes over @xmath25-band flux ) .
the @xmath29 also becomes similarly bluer just like @xmath23 because the @xmath24-band filter response curve goes all the way to @xmath305400 .
now the effects of mg that we read from table 3 are clearly understood from this illustration .
it is confirmed that mg is the most important @xmath2-element that influence the @xmath20 and @xmath31 colors as already described in @xcite .
comparing corresponding entries in the two tables , especially at the 6000 k giant , one sees that oxygen have little effect on the spectrum by themselves ; it is the effect of the decrease of the rest of the elemental abundances that causes the bulk of the spectral change in the constant-@xmath3 case .
unsurprisingly , it is also clear from the table that cooler stars are more susceptible to element - by - element effects than warmer stars , at least at optical wavelengths .
table [ tab : color4 ] presents color results for the isochrone - summed models assuming a @xcite initial mass function and using the @xcite machinery for some disparate ages of populations .
the bolometric correction ( bc ) row includes both isochrone effects and direct stellar spectral effects with the caveat that our synthetic spectra are limited in wavelength coverage , so the fluxes at given wavelengths were normalized to the older low - resolution @xcite flux library at the same @xmath3 .
thus , only effects that affect wavelengths between 3000 and 10000 are considered , and there may be additional , small effects present that are not accounted for . however , in absolute value , all of the bc shifts are quite small , the largest being 6% for the case of all @xmath2-elements enhanced .
it is unlikely that element ratio effects will be of major concern for the mass estimation of clusters and galaxies from their luminosities ( see also @xcite ) .
the other immediate conclusion from table [ tab : color4 ] is that older populations show larger spectral effects due to element - by - element abundance changes .
this is a straightforward consequence of the earlier conclusion that cooler stars are more sensitive ; the light of older populations is dominated by stars that are cooler than those present in younger ones 1 gyr . for younger ages , ir fluxes from agb and tp - agb stars complicate this status ( see also @xcite ) . ] .
oxygen and neon tend to make the colors bluer , while species that contribute more to the lines in the spectrum generally make them redder .
exceptions can be easily explained if one knows where the lines contribute the most .
for instance , magnesium has most of its absorption near 5100 , i.e. , in the @xmath25 band , so adding mg makes @xmath31 bluer while it makes @xmath32 and @xmath33 redder .
the former , the bluer @xmath31 is also partly due to the increase in the flux in the region around 4000 from the mg - enhanced spectra as shown in figure 3 and described in @xcite .
the latter , the redder @xmath32 and @xmath33 also reflect the cooler red giant branch as illustrated in figure 8 of paper i. figure 4 displays this mg - enhanced integrated case at 5 gyr ( thick line ) over the single star case of a 4000 k giant ( thin line ) .
the populations respond rather like the stars do , as can be seen from figure 4 .
based on this , one would suspect that isochrone - caused effects are relatively minor though non - negligible , and this conclusion will be confirmed and amplified in the following section .
the effects of element by element enhancement on the lick indices @xcite are described in this [email protected] ] .
the synthetic indices are not very accurate in absolute predictions ( c.f . * ? ? ?
* ; * ? ? ?
* ) so we employ a differential approach in which the fitting functions of @xcite and @xcite are used as the zero point , and delta - index information as a function of element ratio is incorporated via measuring the synthetic spectral library .
this procedure is similar to that of previous investigations ( e.g. , @xcite ) but more sophisticated since an entire grid of delta - index information was used .
that is , 350 spectra at solar @xmath3 , plus similar grids for 4 other @xmath3 values for this work , as opposed to 2 or 3 synthetic stars at solar abundance only for previous works . to maintain an exact correspondence with paper i , all the elements are 0.3 dex enhanced , except carbon , which is enhanced 0.2 dex0.04 instead of near 0 in terms of [ fe / h ] . ] . in the cases of carbon- , nitrogen- , and oxygen - enhancement , we investigate both at fixed total metallicity and at fixed [ fe / h ] ( and [ r / h ] ) .
the cases at fixed [ fe / h ] are denoted with plus sign in the figures ( e.g. , c+ , n+ , o+ ) .
also in the figures , solar - scaled solar metallicity predictions are connected by a solid line from 1 gyr to 12 gyr and the element - enhanced cases are marked at 1 , 2 , 5 , and 12 gyr .
we have selected fe5406 as a reference index in most plots . among eight lick iron indices (
fe4383 , fe4531 , fe5015 , fe5270 , fe5335 , fe5406 , fe5709 , fe5782 ) we predict that fe5406 is insensitive to every element except iron ( see figures 10 @xmath16 11 and table 5 and online spectra described in footnote # 9 ) , making it a convenient independent variable . as has already been mentioned in the literature @xcite , for most of cases the isochrone effects are relatively minor compared to the stellar spectral effects ( see also @xcite )
. for some cases , however , isochrone effects are non - negligible and quite important ( cf .
h@xmath11 in figure 11 ; see also figure 17 in @xcite )
. it may be worthwhile to mention for clarity that `` isochrone effects '' indicate the temperature , luminosity , and stellar lifetime effects from element ratio changes , and , in the case of the iron - enhanced mixture , the altered [ fe / h ] value that goes into the empirical fitting functions in our experiment .
when calculating observables , we would still use a scaled - solar - ratio spectral library .
these isochrones were presented in paper i. in the present paper ii we add the detailed spectral effects due to element by element enhancement at the stellar atmosphere / stellar flux level .
the `` direct stellar spectral effects '' are the ones that come purely from the emergent spectra with isochrones held fixed .
[ carbon and nitrogen : cn@xmath34 : lick indices cn@xmath35 and cn@xmath36 have identical central bandpasses , but cn@xmath36 has a narrower blue continuum which makes the cn@xmath36 somewhat less prone to abundance variations other than those due to elements c and n. we compare cn@xmath36 with g4300 as carbon , nitrogen , oxygen , silicon and iron are enhanced in figure [ fig05 ] .
the upper panels show that cn@xmath36 is both carbon- and nitrogen - sensitive .
nitrogen s effect is more prominent ( partly because carbon is enhanced only by 0.2 dex compared to the 0.3 dex nitrogen enhancement ) .
the bottom right panel also suggests that silicon and iron affect cn@xmath36 in a non - negligible way when stellar populations become older than 5 gyr .
we can tell from table 5 that the silicon effect is mostly from the stellar spectra , while the iron effect is largely from the isochrones .
the oxygen - enhanced case is displayed in the bottom left panel .
it shows that simply adding o alters cn@xmath36 and g4300 by a small amount ( increases at young ages and decreases at old ages ; the latter via consumption of more c into the co molecule via molecular equilibrium balance ) , but if @xmath3 is held constant both indices decrease much more due to the displacement of c and n to lower abundance because of the o - enhancement and fixed sum .
g4300 is n - insensitive . also , we note that cn@xmath36 has little sensitivity to mg- , s- , ca- , or ti - enhancement . [
carbon : g4300 vs. c@xmath364668 ] : in figure [ fig06 ] , g4300 and c@xmath364668 are plotted as a function of fe5406 .
the left panels show c- and fe - enhanced models , while the right panels show the effects of si and ti enhancement .
g4300 and c@xmath364668 are known to be good carbon indicators among lick indices along with cn@xmath35 and cn@xmath36 .
the left panels show that this is indeed the case . in the bottom left panel
, however , it can be seen that c@xmath364668 is also highly iron sensitive , as indicated by its former name , fe4668 @xcite4668 ( also cn@xmath36 and fe5015 ) is mostly an isochrone effect . however , it is , in fact , not because of the temperature and/or luminosity changes , but because of the [ fe / h ] changes that go into the fitting function that we use for the index calculation . as one can find from figure 1 of paper
i , [ fe / h ] = 0.268 for the fe - enhanced isochrones compared to @xmath160.225 for the @xmath2-enhanced ones at constant solar metallicity , @xmath3 . clearly , future fitting - function work will need to be more meticulously defined in terms of abundance parameters , and not locked to [ fe / h ] necessarily . ] .
contrary to c@xmath364668 , g4300 shows a negligible iron sensitivity .
furthermore , the right panels show that c@xmath364668 is influenced by si and ti , whereas g4300 shows only slight ti sensitivity .
both g4300 and c@xmath364668 show little sensitivity by n- , mg- , s- , or ca - enhancement .
the prospects for disentangling c , n , and o are now fairly good , because it appears that there is a lot of sensitivity among various indices .
however , at the moment it appears that , with the indices available there will be considerable degeneracy among the three quantities . in the future , adding nh , ch , and co features may help considerably ( e.g. , @xcite ) . [
calcium : ca4227 vs. ca4455 ] : figure [ fig07 ] compares ca4227 and ca4455 with fe5406 .
carbon- , nitrogen- , and oxygen - enhanced cases at fixed [ fe / h ] are displayed at the left panels , while right panels depict calcium- and iron - enhanced cases at fixed @xmath3 .
it is clear from upper right panel of figure 7 that ca4227 is significantly boosted with calcium enhancement , the effect increasing with age . to a lesser degree
, ca4227 is also affected by c , n , o , and fe .
carbon- and nitrogen - enhancement make ca4227 weaker by 0.3 , whereas the oxygen - enhancement does the opposite mostly due to their effects at the blue - continuum ( see also @xcite ) .
contrary to ca4227 , ca4455 is hardly influenced by any of those elements . according to this study ,
ca4455 is found to be the most element - enhancement - free lick index .
this is consistent with the previous findings by @xcite and @xcite although our presentation is based on a large spectral grid weighted by isochrones rather than 3 stars .
one concern , however , is that two recent data sets of milky way globular clusters , by @xcite and @xcite , both significantly disagree with theoretical model predictions and also each other ( figure 3 of @xcite ) .
according to tables 1 @xmath16 3 of @xcite , ca4455 has the strongest dependence on the bandpass placement ( wavelength shift error ) among lick indices .
this is because the blue - continuum and the index bandpass of ca4455 share the strong ca absorption line feature near 4455 and consequently cancel out its effect , but only if the wavelength match is perfect .
this phenomenon at the stellar level with 10 km / sec velocity dispersion is displayed in figure 8 .
both ca4227 and ca4455 are rather insensitive to mg- , si- , s- , or ti - enhancement . [ titanium : fe4531 and fe5015 ] : fe4531 and fe5015 are displayed in figure 9 with fe5406 .
tb95 , @xcite , kmt05 , lw05 , and @xcite indicate that these two indices are titanium - sensitive , and we confirm that their titanium sensitivity is indeed strong and almost comparable to the iron sensitivity .
fe5406 , on the other hand , demonstrates little titanium sensitivity .
[ balmer lines : h@xmath11 , h@xmath37 , h@xmath38 , h@xmath39 , h@xmath40 : balmer lines are widely used as an age indicator because of their nonlinear temperature sensitivity in stars , tracing better than many indices the temperature of the main - sequence turnoff .
however , @xcite and earlier work @xcite found that they are also abundance sensitive to some degree . in figures
[ fig10][fig18 ] , we look into their element by element sensitivity in detail .
effects of the individual 10 chemical elements enhancement on the h@xmath11 and fe5406 are shown in figure 10 .
the effects of carbon , nitrogen , and oxygen enhancement are of great importance but they are relatively difficult to understand here because our experimental setup preserves the total metallicity . in other words , it is not straightforward to determine whether we are seeing the effects of c- , n- , o - enhancement or whether we are seeing other elements ( such as mg and fe ) counter - effects due to their depression .
hence , in figure 11 , we display c- , n- , o - enhancement cases both at the fixed total metallicity and at the fixed [ fe / h ] ( the plus signs ) .
it is seen from left panels of figure 11 that unlike nitrogen , oxygen at young ages and carbon at old ages influence h@xmath11 .
the top and bottom right panels of figure 11 illustrates that it is mostly the isochrone - level effects of carbon and oxygen enhancement that affect h@xmath11 .
figure 10 further shows that h@xmath11 is similarly altered by the mg enhancement and in this case it is mostly due to the synthetic spectra ( see also table 5 and figure 12 ) . in the lower right panel of figure 10 ( also in the bottom panels of figure 11 ) ,
a grid with a range of metallicity ( @xmath41 } = -2.0 $ ] , @xmath161.0 , @xmath160.5 , 0.0 , and 0.5 ) is displayed at 1 , 2 , 5 , and 12 gyr .
these additional calculations are of _ solar - scaled _ chemical mixtures @xcite .
we have also depicted @xmath2-elements enhanced cases both at the fixed @xmath3 and at the fixed [ fe / h ] ( alpha+)-elements enhanced case at the fixed [ fe / h ] ( alpha+ ) is @xmath42 = + 0.4 $ ] dex at @xmath41 } = 0 $ ] using the @xcite .
the [ fe / h ] of the @xmath2-elements enhanced case at the fixed @xmath3 ( alpha ) is @xmath160.225 using the paper i. ] .
it is intriguing to note that the effect due to the fe - enhancement by 0.3 dex is seen over the @xmath41 } = 0.5 $ ] grid line .
this is because these grids only reflect the _ mere isochrone effects _ of iron variation with _ solar - scaled _ chemical mixtures , while the 0.3 dex fe - enhancement case shows both isochrone and spectral effects combined .
we find that the h@xmath11 becomes weaker in the @xmath2-elements enhanced case at fixed [ fe / h ] mostly because of the effects of mg ( both isochrone and spectral as one could see from table 5 ) , while it stays nearly unchanged in the @xmath2-elements enhanced case at the fixed @xmath3 mostly because of the reflection of depression of fe . at 12 gyr , a comparison of integrated spectra uncovers that adding mg brings down the blue continuum , making h@xmath11 weaker , but adding fe brings down the central bandpass as well as the red continuum , the net effects of which tend to cancel out effects from fe .
alpha - enhancement at fixed @xmath3 mostly reflects the decrease of fe due to dilution rather than overt , direct spectral effects from @xmath2-elements .
they are displayed in figures 12 to 14 .
the solid lines are the index bandpass and the dashed lines are the blue- and the red - continuum edges .
figures 15 and 16 , on the contrary , display that both h@xmath43 and h@xmath44 become mildly ( @xmath30 0.7 ) stronger with mg - enhancement . furthermore , top panels of figures 15 and 16 illustrate that the broader @xmath45 indices ( h@xmath37 , h@xmath39 ) show sensitivity to fe - enhancement and become significantly weaker with increasing age , by up to 3 at 12 gyr for the case of the h@xmath39 .
this is mostly due to spectral effects , as can be seen from table 5 .
they are illustrated in figures 17 ( h@xmath43 ) to 18 ( h@xmath44 ) . from this experiment
, it seems that for the high order balmer lines , the @xmath46 indices ( h@xmath38 , h@xmath47 ) are less prone to abundance changes because of their narrower index definition and therefore may possibly serve as more robust age indicators .
furthermore , it is found that h@xmath47 among balmer lines has the least sensitivity to c , n , and o ( see also @xcite , @xcite ) . in figure 19 ,
we show some selected diagrams for illustrating element enhancement . in the upper left panel , cn@xmath36 and g4300 are compared and the effects of carbon and nitrogen enhancement are depicted .
the plot shows that cn@xmath36 is sensitive to both carbon and nitrogen .
cn@xmath36 index values go up by 0.04 mag at 12 gyr with 0.3 dex nitrogen enhancement here .
g4300 , on the contrary , is primarily a carbon sensitive index .
ca4227 and g4300 are contrasted in the lower left panel of figure 19 and the effects of sulfur , calcium , and titanium are illustrated .
it is seen that ca4227 is significantly affected by calcium enhancement whereas g4300 is not much altered by anything but carbon . in the upper right panel , mg @xmath12 and fe5406
are compared and the effects of magnesium , and silicon are shown .
it is clear that mg @xmath12 is notably affected by the magnesium enhancement .
h@xmath11 and fe5406 are contrasted in the lower right panel and , as we have seen in figures 10 and 13 , h@xmath11 is not fe - enhancement sensitive , while fe5406 is significantly affected by fe .
figure 20 is basically same as figure 19 , but here the isochrone effects are decoupled from those of synthetic spectra .
the lines show the combination of isochrones and stellar spectral effects , while the points depict the isochrone effects alone at 1 , 2 , 5 , and 12 gyr .
it is seen again from figure [ fig20 ] that the synthetic spectra make the spectral lines stronger and/or weaker , while isochrones play a comparatively minor role .
the displacement of points in the bottom right panel is mostly due to the different [ fe / h ] values that go into the index value calculations rather than the isochrone changes ( see figure 1 of paper i ) .
we have commenced this project to make stellar population models which incorporate flexible chemistry , so that almost any interesting chemical mixture can be interpolated .
paper i dealt with the effects on the stellar evolution models , examining the temperature and luminosity changes due to the altered opacites when ten chemical elements are individually tweaked to the end of the red giant branch . in this paper
we combine those isochrone effects with the stellar spectral effects in order to investigate their mixed effects on the integrated spectrophotometric indices as well as in their integrated spectra themselves .
we again emphasize here that the models in this study are _ incomplete in terms of inclusion of all stellar evolutionary phases and should not be used blindly when comparing to real stellar populations _ until the helium - burning phases are properly incorporated . a version with a full range of metallicity and with horizontal - branch and asymptotic giant branch stars included is planned .
comparison of our models with observations of virgo cluster galaxies will be presented as well ( serven et al . in preparation ) . within our spectral coverage ( 3000 to 10000 )
, we have investigated the broadband color behaviors in the @xmath19 filter set .
as one would expect , older populations show larger spectral effects due to element - by - element abundance changes .
this mostly reflects temperature effects .
we have also confirmed that mg is the most important @xmath2-element that shapes the @xmath20 and @xmath31 colors as already depicted in @xcite . from our investigation of lick indices using the _ integrated _ spectra , we find that ( 1 ) cn@xmath36 is a useful nitrogen indicator once we have good carbon abundances from g4300 and c@xmath364668 , but good silicon ( and also titanium for the c@xmath364668 ) abundances are also needed , ( 2 ) ca4227 is a robust calcium indicator with some good constraints of c , n , and o , ( 3 ) fe4531 and fe5015 are very useful titanium indicators where an independent iron abundances is provided , ( 4 ) mg @xmath12 and fe5406 are good magnesium and iron indicator , respectively , and ( 5 ) the variation of individual elements affects the balmer lines .
we defer the investigation of nad and tio@xmath35 and tio@xmath36 indices until we have the na - enhanced isochrones and can model tio molecular effects with confidence .
below , we illuminate some points that we have described here with the help of full spectra .
fe4531 and fe5015 : from figure 9 we see that both fe4531 and fe5015 prove to be good titanium indicators . they are almost equally sensitive to titanium and iron .
figures 2123 show the sauron @xcite spectral range , 4810 to 5350 , that includes h@xmath11 , fe5015 , mg@xmath35 , mg @xmath12 , fe5270 and a part of fe5335 . here
the integrated spectra of solar - scaled , mg- , fe- , and ti - enhanced cases are shown at 2 gyr with 300 km / sec velocity dispersion normalized at 4750 .
h@xmath11 is not influenced much by the enhancement of these elements at this age , as figure 10 also shows .
mg @xmath12 is mostly sensitive to mg , as figure 21 shows , while fe5015 , fe5270 , and fe5335 are notably sensitive to fe - enhancement , although fe5015 is equally sensitive to titanium , as figures 22 and 23 display .
moreover , we have found that the fe - enhanced and ti - enhanced spectra show their centroids at different wavelengths in the case of fe4531 , opposite sides from the scaled - solar spectra , because of different locations of iron and titanium lines in their index bandpass .
figure [ fig24 ] displays this [email protected] and http://astro.wsu.edu/hclee/fe45@[email protected] ] .
we have preliminary hints from high - s / n , high - resolution spectra of virgo cluster galaxies taken at the kitt peak 4-m telescope ( serven et al . in preparation ) that galaxies with the higher velocity dispersion tend to have their centroids at the longer wavelengths , indicating higher ti - enhancements .
if confirmed , this would indicate a ti-@xmath48 relation , similar to the well - known mg-@xmath48 relation ( see also @xcite ) .
h@xmath11 : the bottom right panel of figure 10 shows that , in an @xmath2-enhanced mixture at the fixed [ fe / h ] ( alpha+ ) , h@xmath11 becomes weaker .
this sounds a bit different from the conclusion of lw05 where h@xmath11 is @xmath2-element insensitive .
as we can see from table 5 , the effects that we see from the bottom right panel of figure 10 are mostly isochrone effects , especially at young ages .
the h@xmath11 index becomes weaker because of the decrease of temperature of stellar isochrones from the @xmath2-enhanced case at fixed [ fe / h ] ( see also @xcite ; @xcite ) .
lw05 used only solar - scaled isochrones at solar and super - solar metallicities , although these authors employed modified stellar spectra because of the @xmath2-enhancement .
moreover , it is worth emphasizing that @xmath2-enhanced isochrones at fixed @xmath3 are now no longer significantly hotter than the solar - scaled ones .
this has been revealed by paper i , in its figure 11 , after incorporating the latest development of low - temperature opacities by @xcite .
if this behavior is indeed true , it will certainly affect the conclusions of earlier works ( e.g. , @xcite ) that are based upon the @xcite isochrones , which were later shown by @xcite to present problems .
h@xmath43 and h@xmath44 : fe - enhancement makes these indices significantly weaker , especially for the broader indices , h@xmath49 and h@xmath50 ( see also @xcite ) .
figures 17 and 18 show that this is because of fe - lines that inhabit the continuum region .
it is suggested therefore that the narrower indices ( h@xmath51 and h@xmath52 ) may serve as better age indicators in case the stellar systems are of wildly different chemical mixtures .
our goals are to narrow down the uncertainties of mean age estimations to 10% when derived from a single integrated light spectrum .
we believe that we can reach these goals as we are able to take into account the effects from individual chemical elements .
the ultimate goal then obviously would be the understanding of the chemical enrichment and star formation histories of galaxies , stepping away from the study of mean ages and mean metallicities to begin to tackle the more realistic problem of multi - age , multi - metallicity stellar populations as they exist in galaxies .
we believe that several efforts such as speckmap @xcite , moped @xcite , and starlight @xcite will be even more useful once they use sophisticated stellar population models with flexible chemistry as inputs .
we thank jim rose , jon fulbright , and andreas koch for useful discussions on the mg - enhancement of the metal - rich cool giants .
we are also grateful to the referee , ricardo schiavon , for a thorough reading which greatly helped us to improve the paper .
support for this work was provided by the nsf through grant ast-0307487 , the new standard stellar population models ( nsspm ) project .
lee also acknowledges jp blakeslee s support via nasa grant .
pc acknowledges the support by the european community under a marie curie international incoming fellowship .
bacon , r. et al .
2001 , , 326 , 23 bell , r. a. , paltoglou , g. , & tripicco , m. j. 1994 , , 268 , 771 bessell , m. s. 1990 , , 102 , 1181 carretta , e. , gratton , r. g. , bragaglia , a. , bonifacio , p. , & pasquini , l. 2004 , , 416 , 925 cassisi , s. , salaris , m. , castelli , f. & pietrinferni .
a. 2004 , , 616 , 498 cayrel , r. , perrin , m .-
n . , barbuy , b. , & buser , r. 1991 , , 247 , 108 charlot , s. , worthey , g. , & bressan .
a. 1996 , , 457 , 625 cid fernandes , r. et al . 2008 , ( arxiv:0802.0849 ) coelho , p. , barbuy , b. , melndez , j. , schiavon , r. p. , & castilho , b. v. 2005 , , 443 , 735 coelho , p. , bruzual , g. , charlot , s. , weiss , a. , barbuy , b. , & ferguson , j.w .
2007 , , 382 , 498 cohen , j. g. , blakeslee , j. p. , & ryzhov , a. 1998 , , 496 , 808 dotter , a. , chaboyer , b. , jevremovi , d. , baron , e. , ferguson , j. w. , sarajedini , a. , & anderson , j. 2007a , , 134 , 376 dotter , a. , chaboyer , b. , ferguson , j. w. , lee , h .- c .
, worthey , g. , jevremovi , d. , & baron , e. 2007b , , 666 , 403 ( paper i ) ferguson , j. w. , alexander , d. r. , allard , f. , barman , t. , bodnarik , j. g. , hauschildt , p. h. , heffner - wong , a. , & tamanai , a. 2005 , , 623 , 585 fulbright , j. p. , mcwilliam , a. , & rich , r. m. 2007 , , 661 , 1152 graves , g. j. , & schiavon , r. p. 2008 , ( arxiv : 0803.1483 ) gray , d. f. , & brown , k. 2001 , , 113 , 723 griffin , r. e. m. , & lynas - gray , a. e. 1999 , , 117 , 2998 gustafsson , b. , bell , r. a. , eriksson , k. , & nordlund , 1975 , , 42 , 407 koch , a. , & mcwilliam , a. 2008 , , 135 , 1551 korn , a. j. , maraston , c. , & thomas , d. 2005 , , 438 , 685 ( kmt05 ) kurucz , r. l. 1970 , sao special report # 309 lee , h .- c . , & worthey , g. 2005 , , 160 , 176 ( lw05 ) lee , h .- c .
, worthey , g. , trager , s. c. & faber , s. m. 2007a , , 664 , 215 lee , h .- c .
, worthey , g. , trager , s. c. , dotter , a. , chaboyer , b. , ferguson , j. w. , jevremovi , d. , baron , e. , coelho , p. , & briley , m. m. 2007b , in stellar populations as building blocks of galaxies , iau symposium 241 , ed .
r. f. peletier , a. vazdekis , in press , ( astro - ph/0701666 ) mrmol - queralt , e. , et al .
2008 , , in press ( arxiv:0806.0581 ) martell , s. l. , smith , g. h. & briley , m. m. 2008 , , in press ( arxiv:0806.0012 ) martins , l. p. , & coelho , p. 2007
, , 381 , 1329 milone , a. , barbuy , b. , & schiavon , r. p. 2000 , , 120 , 131 ocvirk , p. , pichon , c. , lancon , a. , & thiebaut , e. 2006 , , 365 , 74 panter , b. , jimenez , r. , heavens , a. f. , & charlot , s. 2007 , , 378 , 1550 peterson , r. c. , dalle ore , c. m. , & kurucz , r. l. 1993 , , 404 , 333 plez , b. 1992 , , 94 , 527 prochaska , l. c. , rose , j. a. , & schiavon , r. p. 2005
, , 130 , 2666 prochaska , l. c. , rose , j. a. , caldwell , n. , castilho , b. v. , concannon , k. , harding , p. , morrison , h. , & schiavon , r. p. 2007
, , 134 , 321 proctor , r. n. , & sansom , a. e. 2002 , , 333 , 517 puzia , t. h. , et al .
2002 , , 395 , 45 salasnich , b. , girardi , l. , weiss , a. , & chiosi , c. 2000 , , 361 , 1023 salpeter , e. e. 1955 , , 121 , 161 schiavon , r. p. 2007
, , 171 , 146 schiavon , r. p. , faber , s. m. , castilho , b. v. , & rose , j. a. 2002 , , 580 , 850 serven , j.l . , worthey , g. , & briley , m. m. 2005 , , 627 , 754 thomas , d. , & maraston , c. 2003 , , 401 , 429 thomas , d. , maraston , c. , & bender , r. 2003 , , 339 , 897 thomas , d. , maraston , c. , & korn , a. j. 2004 , , 351 , l19 trager , s. c. 1997 , ph.d .
thesis , univ .
california , santa cruz trager , s. c. , faber , s. m. , worthey , g. , gonzlez , j. j. 2000a , , 119 , 1645 trager , s. c. , faber , s. m. , worthey , g. , gonzlez , j. j. 2000b , , 120 , 188 tripicco .
m. j. , & bell , r. a. 1995 , , 110 , 3035 ( tb95 ) vazdekis , a. 1999 , , 513 , 224 weiss , a. , salaris , m. , ferguson , j. w. , & alexander , d. r. 2006 , ( astro - ph/0605666 ) worthey , g. 1994 , , 95 , 107 worthey , g. 1998 , , 110 , 888 worthey , g. , faber , s. m. , gonzlez , j. j. , & burstein , d. 1994 , , 94 , 687 worthey , g. , & lee , h .- c .
2008 , , submitted ( astro - ph/0604590 ) worthey , g. , & ottaviani , d. l. 1997 , , 111 , 377 yong , d. , grundahl , f. , johnson , j. a. & asplund , m. 2008 , , in press ( arxiv:0806.0187 ) @xmath20 & 1.26 & 21 & 81 & -106 & @xmath53 & 58 & -85 + @xmath31 & 1.43 & 43 & 35 & -137 & @xmath54 & 22 & -81 + @xmath32 & 0.89 & 16 & 24 & -94 & @xmath55 & 3 & -74 + @xmath33 & 1.60 & 13 & 24 & -176 & @xmath56 & -5 & -176 + @xmath20 & 0.61 & 123 & 111 & -179 & @xmath57 & 160 & -169 + @xmath31 & 0.91 & 55 & 45 & -90 & @xmath58 & 62 & -101 + @xmath32 & 0.52 & 21 & 19 & -62 & @xmath59 & 11 & -55 + @xmath33 & 0.98 & 15 & 14 & -103 & @xmath58 & 13 & -97 + @xmath20 & 0.02 & 113 & 95 & -33 & @xmath60 & 114 & -75 + @xmath31 & 0.58 & 39 & 37 & -50 & @xmath61 & 42 & -70 + @xmath32 & 0.33 & 15 & 11 & -28 & @xmath62 & 4 & -33 + @xmath33 & 0.66 & 16 & 5 & -55 & @xmath63 & 1 & -60 + @xmath20 & 1.34 & 115 & 137 & -182 & @xmath64 & 52 & -194 + @xmath31 & 1.45 & 39 & 57 & -166 & @xmath65 & 22 & -103 + @xmath32 & 0.83 & 5 & 21 & -120 & @xmath66 & 5 & -82 + @xmath33 & 1.51 & -4 & 20 & -198 & @xmath67 & 4 & -182 + @xmath20 & 0.57 & 159 & 147 & -159 & @xmath68 & 161 & -194 + @xmath31 & 0.93 & 52 & 60 & -107 & @xmath69 & 63 & -116 + @xmath32 & 0.48 & 18 & 12 & -46 & @xmath70 & 10 & -52 + @xmath33 & 0.92 & 9 & -1 & -80 & @xmath71 & 11 & -87 + @xmath20 & 0.16 & 79 & 83 & 19 & @xmath72 & 109 & -81 + @xmath31 & 0.55 & 48 & 54 & -64 & @xmath73 & 41 & -85 + @xmath32 & 0.31 & 12 & 9 & -35 & @xmath74 & 4 & -36 + @xmath33 & 0.61 & 12 & 4 & -69 & @xmath75 & 1 & -70 + @xmath20 & -13 & 8 & 19 & -2 & -66 & -36 & 0 & -40 & 15 & 136 & -103 + @xmath31 & 17 & -2 & -50 & -4 & -19 & 35 & 0 & -7 & 15 & 22 & -33 + @xmath32 & -2 & -1 & -7 & -1 & 31 & -1 & 0 & -14 & 4 & 12 & 8 + @xmath33 & -5 & -1 & -5 & -1 & 31 & -1 & 0 & -14 & 4 & 14 & 10 + @xmath20 & 13 & 18 & -61 & -13 & -26 & -39 & -2 & -4 & 11 & 61 & -120 + @xmath31 & 17 & 1 & -40 & -6 & -32 & 5 & -1 & 6 & 7 & 18 & -58 + @xmath32 & -3 & -5 & -8 & -2 & 8 & -2 & 0 & 0 & 3 & 10 & -3 + @xmath33 & -12 & -11 & 0 & -1 & 10 & -1 & 0 & -1 & 3 & 15 & 7 + @xmath20 & -4 & 10 & -48 & -13 & 5 & -19 & -2 & 1 & 8 & 30 & -59 + @xmath31 & 8 & 1 & -20 & -4 & -14 & -7 & 0 & 6 & 5 & 11 & -34 + @xmath32 & 0 & -1 & -7 & -1 & 0 & -1 & 0 & 0 & 2 & 7 & -9 + @xmath33 & 0 & -3 & -8 & -1 & 0 & 0 & 0 & 0 & 2 & 10 & -10 + @xmath20 & 7 & 11 & -50 & -13 & -26 & -50 & -2 & -20 & 22 & 112 & -129 + @xmath31 & 17 & -1 & -47 & -4 & -30 & 37 & 0 & -3 & 12 & 19 & -37 + @xmath32 & -10 & -5 & 2 & 0 & 21 & 0 & 0 & -8 & 3 & 13 & 15 + @xmath33 & -21 & -9 & 14 & 1 & 23 & 0 & 0 & -9 & 2 & 17 & 29 + @xmath20 & 10 & 23 & -74 & -17 & 24 & -39 & -3 & 6 & 11 & 19 & -82 + @xmath31 & 18 & 5 & -34 & -6 & -22 & -6 & -1 & 8 & 6 & 8 & -52 + @xmath32 & 1 & -10 & -8 & -1 & 1 & 0 & 0 & 0 & 2 & 9 & -6 + @xmath33 & -7 & -23 & 1 & 0 & 6 & 2 & 0 & 0 & 3 & 16 & 10 + @xmath20 & -5 & 6 & -34 & -9 & 12 & -9 & -2 & 6 & 8 & 9 & -25 + @xmath31 & 1 & 0 & -23 & -5 & -5 & -3 & -1 & 9 & 5 & 9 & -23 + @xmath32 & 0 & 0 & -5 & -1 & -1 & 0 & 0 & 0 & 2 & 5 & -7 + @xmath33 & 0 & -1 & -6 & -1 & 0 & 0 & 0 & -1 & 2 & 7 & -9 + @xmath20 & -11 & 10 & 30 & & -65 & -35 & 0 & -40 & 15 & 139 & -97 + @xmath31 & 22 & 0 & -31 & & -17 & 38 & 0 & -7 & 15 & 26 & -5 + @xmath32 & 0 & 0 & 0 & & 31 & 0 & 0 & -13 & 4 & 14 & 19 + @xmath33 & -3 & 0 & 1 & & 32 & -1 & 0 & -14 & 4 & 16 & 19 + @xmath20 & 28 & 26 & -10 & & -20 & -33 & 0 & -3 & 11 & 74 & -57 + @xmath31 & 24 & 4 & -14 & & -30 & 8 & 0 & 6 & 7 & 24 & -23 + @xmath32 & -1 & -3 & 2 & & 9 & -1 & 0 & 0 & 3 & 12 & 11 + @xmath33 & -10 & -10 & 7 & & 11 & 0 & 0 & -1 & 3 & 16 & 18 + @xmath20 & 8 & 17 & 1 & & 11 & -13 & 0 & 2 & 8 & 41 & 8 + @xmath31 & 13 & 3 & 0 & & -12 & -5 & 0 & 7 & 5 & 15 & -8 + @xmath32 & 2 & 0 & 0 & & 0 & 0 & 0 & 0 & 2 & 8 & 0 + @xmath33 & 1 & -2 & 0 & & 0 & 0 & 0 & 0 & 2 & 11 & 0 + @xmath20 & 21 & 18 & 1 & & -20 & -44 & 0 & -20 & 22 & 125 & -63 + @xmath31 & 22 & 0 & -27 & & -28 & 39 & 0 & -3 & 12 & 23 & -10 + @xmath32 & -9 & -4 & 5 & & 21 & 0 & 0 & -8 & 3 & 13 & 19 + @xmath33 & -22 & -9 & 11 & & 23 & 0 & 0 & -9 & 2 & 17 & 26 + @xmath20 & 29 & 33 & -9 & & 31 & -32 & 0 & 7 & 11 & 34 & 5 + @xmath31 & 24 & 8 & -9 & & -19 & -3 & 0 & 8 & 6 & 13 & -20 + @xmath32 & 4 & -9 & 0 & & 2 & 0 & 0 & 0 & 2 & 11 & 5 + @xmath33 & -6 & -22 & 6 & & 7 & 2 & 0 & 0 & 3 & 17 & 17 + @xmath20 & 3 & 11 & 1 & & 16 & -5 & 0 & 7 & 8 & 17 & 25 + @xmath31 & 7 & 2 & 0 & & -3 & 0 & 0 & 10 & 5 & 14 & 9 + @xmath32 & 1 & 0 & 0 & & 0 & 0 & 0 & 0 & 2 & 6 & 0 + @xmath33 & 1 & 0 & 0 & & 0 & 0 & 0 & 0 & 2 & 8 & 0 + @xmath76 & & -0.3 & 0.0 & 5.5 & 1.9 & 2.9 & 2.6 & 1.6 & -1.2 & 0.8 & -3.8 & 8.0 + @xmath77 & 0.050 & -3.7 & 2.3 & -14.9 & -4.0 & 1.6 & -5.5 & -1.9 & 2.6 & 2.2 & 14.8 & -20.7 + @xmath78 & 0.361 & -1.0 & -0.2 & -34.6 & 1.3 & 5.7 & 9.1 & 6.2 & 6.0 & 2.4 & -1.1 & -11.6 + @xmath79 & 0.223 & -1.1 & -0.2 & -17.1 & 2.4 & 7.6 & 7.4 & 4.8 & 1.0 & 1.9 & -1.7 & -2.3 + @xmath80 & 0.458 & -4.0 & 0.4 & -26.3 & 5.8 & 15.2 & 15.2 & 10.7 & 2.0 & 7.2 & -7.6 & 4.5 + @xmath76 & & -9.9 & -4.5 & 11.1 & 0.7 & 12.2 & 4.5 & 0.6 & -17.5 & 15.5 & 3.9 & 46.6 + @xmath20 & 0.171 & 0.5 & 12.1 & -30.2 & 4.9 & 7.7 & -6.6 & 4.8 & -2.8 & 16.7 & 12.5 & -18.4 + @xmath31 & 0.664 & 16.1 & 6.3 & -20.6 & 12.9 & -10.0 & 10.4 & 8.2 & -2.4 & 15.2 & -2.5 & -10.4 + @xmath32 & 0.428 & 0.5 & -1.6 & -1.9 & 6.6 & 8.8 & 7.1 & 4.2 & -14.2 & 14.2 & 4.6 & 22.0 + @xmath33 & 0.877 & -15.5 & -4.9 & 10.0 & 8.5 & 23.3 & 13.9 & 6.2 & -43.3 & 47.2 & 9.8 & 97.5 + @xmath76 & & -21.1 & -11.1 & 10.5 & -2.8 & 12.0 & 0.0 & -3.0 & -2.7 & 20.2 & -1.6 & 48.8 + @xmath77 & 0.369 & -10.7 & 5.7 & -75.5 & -13.1 & 14.9 & -19.8 & -8.1 & -2.7 & 5.1 & 41.1 & -71.3 + @xmath78 & 0.861 & 9.7 & -4.8 & -52.3 & -4.3 & -15.6 & 5.8 & -5.7 & 1.5 & -3.0 & 15.1 & -54.8 + @xmath79 & 0.537 & -9.1 & -10.8 & -15.6 & -3.3 & 11.2 & 1.4 & -3.7 & -3.9 & 6.2 & 8.7 & 4.2 + @xmath80 & 1.087 & -40.9 & -22.5 & -3.5 & -8.1 & 18.4 & -0.4 & -8.4 & -5.7 & 40.6 & 5.0 & 71.9 + @xmath76 & & -22.5 & -10.4 & 11.8 & -3.5 & 21.4 & 2.8 & -1.5 & -1.2 & 30.4 & 7.2 & 55.3 + @xmath20 & 0.599 & 2.5 & 14.7 & -101.8&-15.7 & 31.1 & -14.7 & -7.1 & -0.8 & 9.2 & 68.8 & -99.0 + @xmath31 & 0.988 & 21.1 & 0.8 & -58.1 & -4.9 & -8.3 & 15.7 & -3.4 & 4.2 & -0.4 & 26.3 & -65.8 + @xmath32 & 0.604 & -6.1 & -8.4 & -15.0 & -3.6 & 22.4 & 5.3 & -1.8 & -2.3 & 11.7 & 17.1 & 5.8 + @xmath33 & 1.200 & -37.2 & -17.3 & 1.5 & -8.4 & 36.3 & 7.9 & -3.9 & -1.6 & 57.2 & 22.2 & 78.6 + cn@xmath35 & -0.1729 & -0.0013 & 0.0052 & -0.0003 & 0.0015 & -0.0091 & 0.0096 & 0.0029 & 0.0027 & 0.0014 & 0.0004 & -0.0001 & 0.0037 & -0.0030 & 0.0317 + cn@xmath36 & -0.1099 & -0.0008 & 0.0031 & 0.0001 & 0.0009 & -0.0052 & 0.0061 & 0.0020 & 0.0017 & 0.0010 & 0.0005 & 0.0000 & 0.0017 & -0.0014 & 0.0210 + ca4227 & 0.3702 & -0.0054 & 0.0184 & -0.0032 & 0.0029 & -0.0492 & 0.0247 & 0.0100 & 0.0098 & 0.0061 & 0.0018 & -0.0003 & 0.0391 & -0.0349 & 0.1102 + g4300 & -0.1319 & -0.0663 & 0.2024 & -0.0232 & 0.0580 & -0.4061 & 0.3611 & 0.1018 & 0.0976 & 0.0473 & 0.0105 & -0.0048 & 0.2476 & -0.2056 & 1.1606 + fe4383 & 0.0769 & -0.0418 & 0.0922 & -0.0086 & 0.0337 & -0.2781 & 0.1782 & 0.0566 & 0.0289 & 0.0066 & 0.0178 & 0.0081 & 0.2371 & -0.2220 & 0.7064 + ca4455 & 0.4773 & -0.0219 & 0.0285 & -0.0098 & 0.0062 & -0.1185 & 0.0428 & 0.0098 & 0.0069 & 0.0040 & 0.0023 & -0.0006 & 0.1553 & -0.1151 & 0.1653 + fe4531 & 1.4918 & -0.0271 & 0.0517 & -0.0082 & 0.0124 & -0.1626 & 0.0814 & 0.0261 & 0.0209 & 0.0124 & 0.0067 & 0.0005 & 0.1565 & -0.1401 & 0.3023 + c@xmath36 4668 & 0.3137 & -0.0962 & 0.0786 & -0.0405 & 0.0117 & -0.4479 & 0.1322 & 0.0245 & 0.0041 & 0.0059 & 0.0118 & -0.0001 & 0.7050 & -0.4688 & 0.5675 + h@xmath11 & 6.0151 & -0.0298 & -0.1859 & -0.0043 & -0.0386 & 0.1792 & -0.2727 & -0.1091 & -0.1003 & -0.0425 & -0.0073 & 0.0115 & 0.0807 & -0.0603 & -1.0362 + fe5015 & 2.4643 & -0.0762 & 0.0633 & -0.0264 & 0.0157 & -0.3389 & 0.1203 & 0.0219 & 0.0105 & 0.0090 & 0.0116 & 0.0009 & 0.4497 & -0.3421 & 0.4301 + mg@xmath35 & 0.0013 & 0.0006 & 0.0004 & 0.0006 & -0.0001 & 0.0003 & 0.0000 & 0.0011 & 0.0006 & 0.0004 & 0.0006 & 0.0002 & -0.0014 & -0.0001 & 0.0059 + mg@xmath36 & 0.0578 & -0.0001 & 0.0017 & 0.0006 & 0.0001 & -0.0034 & 0.0018 & 0.0022 & 0.0013 & 0.0009 & 0.0010 & 0.0005 & 0.0023 & -0.0034 & 0.0140 + mg @xmath12 & 0.9845 & -0.0053 & 0.0357 & 0.0078 & 0.0074 & -0.0440 & 0.0625 & 0.0421 & 0.0296 & 0.0197 & 0.0147 & 0.0091 & -0.0036 & -0.0129 & 0.2977 + fe5270 & 1.0894 & -0.0378 & 0.0317 & -0.0028 & 0.0107 & -0.1627 & 0.0665 & 0.0233 & 0.0133 & 0.0136 & 0.0156 & 0.0075 & 0.1864 & -0.1699 & 0.2598 + fe5335 & 0.8774 & -0.0442 & 0.0295 & -0.0055 & 0.0099 & -0.1727 & 0.0648 & 0.0208 & 0.0106 & 0.0116 & 0.0159 & 0.0080 & 0.2445 & -0.1872 & 0.2509 + fe5406 & 0.3673 & -0.0203 & 0.0169 & -0.0005 & 0.0047 & -0.0884 & 0.0364 & 0.0153 & 0.0077 & 0.0080 & 0.0111 & 0.0045 & 0.1183 & -0.0949 & 0.1593 + fe5709 & 0.4454 & -0.0258 & -0.0001 & -0.0005 & 0.0045 & -0.0581 & 0.0299 & 0.0061 & 0.0010 & 0.0044 & 0.0101 & 0.0061 & 0.0859 & -0.0698 & 0.0844 + fe5782 & 0.2933 & -0.0147 & -0.0003 & 0.0017 & 0.0015 & -0.0373 & 0.0145 & 0.0052 & 0.0007 & 0.0033 & 0.0076 & 0.0037 & 0.0465 & -0.0478 & 0.0562 + nad & 1.1216 & 0.0455 & 0.0516 & 0.0090 & -0.0037 & -0.0638 & -0.0174 & 0.0305 & 0.0231 & 0.0132 & 0.0051 & -0.0049 & 0.0294 & -0.0684 & 0.1801 + tio@xmath35 & 0.0114 & 0.0034 & 0.0020 & -0.0008 & -0.0011 & -0.0010 & -0.0035 & -0.0001 & -0.0001 & -0.0009 & -0.0015 & -0.0013 & 0.0004 & 0.0000 & -0.0005 + tio@xmath36 & -0.0075 & 0.0079 & 0.0046 & -0.0020 & -0.0026 & -0.0026 & -0.0080 & -0.0001 & -0.0001 & -0.0021 & -0.0032 & -0.0031 & 0.0006 & -0.0001 & -0.0004 + h@xmath50 & 8.7930 & 0.0021 & -0.3210 & -0.0001 & -0.0908 & 0.3505 & -0.5622 & -0.1823 & -0.1727 & -0.0756 & -0.0182 & 0.0092 & 0.1019 & -0.0687 & -2.0267 + h@xmath49 & 9.0232 & -0.0102 & -0.4023 & -0.0039 & -0.1157 & 0.4442 & -0.6934 & -0.2312 & -0.2128 & -0.0889 & -0.0256 & 0.0143 & 0.1671 & -0.1038 & -2.6081 + h@xmath52 & 5.8139 & 0.0277 & -0.2032 & 0.0104 & -0.0573 & 0.2863 & -0.3554 & -0.1075 & -0.1054 & -0.0484 & -0.0079 & 0.0064 & -0.0646 & 0.0497 & -1.1933 + h@xmath51 & 6.1194 & 0.0129 & -0.2280 & 0.0024 & -0.0605 & 0.2801 & -0.3875 & -0.1305 & -0.1233 & -0.0575 & -0.0135 & 0.0062 & -0.0013 & 0.0067 & -1.3902 + cn@xmath35 & -0.1729 & -0.0009 & 0.0061 & 0.0016 & 0.0036 & -0.0121 & 0.0082 & 0.0037 & 0.0030 & 0.0012 & -0.0013 & 0.0019 & 0.0080 & -0.0051 & 0.0322 + cn@xmath36 & -0.1099 & -0.0005 & 0.0043 & 0.0022 & 0.0036 & -0.0108 & 0.0050 & 0.0031 & 0.0023 & 0.0008 & -0.0014 & 0.0021 & 0.0054 & -0.0053 & 0.0225 + ca4227 & 0.3702 & -0.0349 & 0.0020 & -0.0190 & -0.0060 & -0.0895 & 0.0319 & 0.0048 & 0.0035 & 0.0034 & 0.1021 & 0.0050 & 0.0892 & 0.0048 & 0.3323 + g4300 & -0.1319 & 0.0953 & 0.4853 & -0.0681 & 0.0570 & -0.7360 & 0.3229 & 0.0262 & 0.0345 & 0.0270 & 0.0934 & 0.1973 & 0.2422 & -0.4542 & 1.3367 + fe4383 & 0.0769 & -0.0556 & 0.1236 & -0.0304 & 0.0319 & -0.4444 & 0.1498 & 0.0248 & -0.0315 & -0.0029 & -0.0478 & 0.1234 & 0.4107 & -0.4621 & 0.5671 + ca4455 & 0.4773 & -0.0417 & 0.0275 & -0.0196 & 0.0059 & -0.1871 & 0.0428 & -0.0278 & -0.0350 & -0.0003 & 0.0162 & 0.0251 & 0.0500 & -0.2267 & 0.1421 + fe4531 & 1.4918 & -0.0932 & 0.0524 & -0.0416 & 0.0130 & -0.4076 & 0.0831 & -0.0291 & -0.0344 & -0.0025 & 0.0058 & 0.2521 & 0.1914 & -0.2945 & 0.5801 + c@xmath36 4668 & 0.3137 & 0.0475 & 0.2777 & -0.0673 & 0.0115 & -0.6511 & 0.0977 & 0.0472 & -0.1423 & 0.0234 & 0.0520 & 0.0896 & 0.9129 & -0.6611 & 0.5911 + h@xmath11 & 6.0151 & -0.0295 & -0.1969 & 0.0020 & -0.0367 & 0.2201 & -0.2554 & -0.1288 & -0.0934 & -0.0403 & -0.0094 & 0.0493 & -0.0273 & 0.0027 & -0.9932 + fe5015 & 2.4643 & -0.1859 & 0.0965 & -0.0993 & 0.0160 & -0.8476 & 0.1272 & -0.0770 & -0.1042 & -0.0227 & 0.0789 & 0.2735 & 0.9007 & -0.8013 & 0.8161 + mg@xmath35 & 0.0013 & 0.0021 & 0.0032 & -0.0001 & -0.0002 & -0.0044 & -0.0008 & 0.0048 & -0.0011 & 0.0002 & 0.0001 & 0.0008 & 0.0001 & -0.0036 & 0.0106 + mg@xmath36 & 0.0578 & -0.0008 & 0.0031 & -0.0006 & 0.0001 & -0.0112 & 0.0014 & 0.0163 & -0.0017 & 0.0004 & 0.0010 & 0.0021 & 0.0034 & -0.0017 & 0.0394 + mg @xmath12 & 0.9845 & -0.0484 & 0.0239 & -0.0079 & 0.0078 & -0.1458 & 0.0643 & 0.4731 & -0.0047 & 0.0120 & 0.0248 & 0.0143 & -0.0420 & 0.2288 & 1.0441 + fe5270 & 1.0894 & -0.1036 & 0.0349 & -0.0354 & 0.0133 & -0.4167 & 0.0637 & -0.0172 & 0.1169 & 0.0096 & 0.1063 & 0.0534 & 0.3233 & -0.2704 & 0.6047 + fe5335 & 0.8774 & -0.1122 & 0.0176 & -0.0350 & 0.0094 & -0.3570 & 0.0868 & -0.0184 & 0.1227 & -0.0009 & 0.0030 & 0.0324 & 0.3424 & -0.3239 & 0.4283 + fe5406 & 0.3673 & -0.0620 & 0.0008 & -0.0134 & 0.0050 & -0.1852 & 0.0360 & 0.0022 & 0.0242 & 0.0021 & 0.0072 & -0.0149 & 0.3128 & -0.2230 & 0.1452 + fe5709 & 0.4454 & -0.0470 & 0.0006 & -0.0130 & 0.0033 & -0.1317 & 0.0296 & 0.0218 & -0.0033 & 0.0153 & 0.0092 & 0.0142 & 0.1678 & -0.1249 & 0.1635 + fe5782 & 0.2933 & -0.0271 & 0.0009 & -0.0058 & 0.0010 & -0.0829 & 0.0138 & 0.0198 & -0.0051 & 0.0002 & 0.0046 & 0.0005 & 0.0177 & -0.0967 & 0.0828 + nad & 1.1216 & 0.0303 & 0.0632 & -0.0003 & -0.0004 & -0.1616 & -0.0210 & 0.0017 & -0.0057 & 0.0101 & -0.0263 & -0.0161 & -0.0019 & -0.2582 & 0.0620 + tio@xmath35 & 0.0114 & 0.0028 & 0.0018 & -0.0008 & -0.0008 & -0.0023 & -0.0030 & -0.0003 & 0.0004 & -0.0010 & -0.0016 & 0.0008 & 0.0006 & 0.0004 & 0.0040 + tio@xmath36 & -0.0075 & 0.0071 & 0.0038 & -0.0020 & -0.0025 & -0.0033 & -0.0072 & -0.0003 & -0.0004 & -0.0024 & -0.0043 & -0.0001 & 0.0014 & -0.0001 & 0.0029 + h@xmath50 & 8.7930 & 0.0011 & -0.3916 & 0.0191 & -0.1020 & 0.5938 & -0.5279 & -0.1339 & -0.1226 & -0.0616 & 0.0439 & -0.0549 & -0.2367 & 0.2959 & -1.8286 + h@xmath49 & 9.0232 & -0.1151 & -0.6314 & 0.0461 & -0.1130 & 0.8066 & -0.6397 & -0.1076 & -0.1489 & -0.0668 & -0.0552 & -0.1770 & -0.0191 & 0.2940 & -2.6081 + h@xmath52 & 5.8139 & 0.0552 & -0.2155 & 0.0300 & -0.0562 & 0.4313 & -0.3398 & -0.0890 & -0.0728 & -0.0396 & 0.0489 & -0.0276 & -0.2790 & 0.2848 & -1.0708 + h@xmath51 & 6.1194 & -0.0558 & -0.3441 & 0.0210 & -0.0584 & 0.4152 & -0.3578 & -0.0815 & -0.0958 & -0.0493 & -0.0315 & -0.0107 & -0.0773 & 0.2051 & -1.2780 + cn@xmath35 & -0.0855 & 0.0024 & 0.0040 & 0.0017 & 0.0027 & -0.0060 & 0.0082 & -0.0005 & 0.0019 & 0.0035 & 0.0001 & 0.0038 & 0.0082 & -0.0017 & 0.0234 + cn@xmath36 & -0.0429 & 0.0019 & 0.0026 & 0.0018 & 0.0022 & -0.0041 & 0.0067 & -0.0012 & 0.0012 & 0.0032 & 0.0001 & 0.0036 & 0.0065 & -0.0009 & 0.0180 + ca4227 & 0.7521 & 0.0227 & 0.0315 & 0.0048 & 0.0113 & -0.0269 & 0.0223 & 0.0168 & 0.0210 & 0.0125 & -0.0144 & 0.0168 & 0.0759 & -0.0141 & 0.1168 + g4300 & 2.6706 & 0.0532 & 0.1604 & 0.0379 & 0.0784 & -0.2405 & 0.2681 & 0.0432 & 0.0821 & 0.0915 & -0.0094 & 0.0894 & 0.3417 & -0.1276 & 0.7597 + fe4383 & 2.3834 & 0.0201 & 0.1097 & 0.0502 & 0.0732 & -0.3131 & 0.2280 & -0.0104 & 0.0423 & 0.1130 & -0.0306 & 0.1131 & 0.5479 & -0.2635 & 0.7914 + ca4455 & 1.0187 & 0.0101 & 0.0357 & 0.0020 & 0.0172 & -0.0957 & 0.0347 & 0.0032 & 0.0113 & 0.0153 & -0.0184 & 0.0251 & 0.2121 & -0.0953 & 0.1391 + fe4531 & 2.4730 & 0.0260 & 0.0566 & 0.0145 & 0.0292 & -0.1097 & 0.0627 & 0.0054 & 0.0261 & 0.0353 & -0.0307 & 0.0434 & 0.2686 & -0.1070 & 0.2052 + c@xmath36 4668 & 2.4464 & -0.0308 & 0.0922 & -0.0211 & 0.0615 & -0.5312 & 0.1239 & -0.0415 & -0.0199 & 0.0283 & -0.1289 & 0.1142 & 1.2324 & -0.5513 & 0.5766 + h@xmath11 & 3.4442 & -0.1493 & -0.1619 & -0.0499 & -0.0689 & 0.0140 & -0.1484 & -0.0572 & -0.1026 & -0.0778 & 0.0084 & -0.0804 & -0.0126 & -0.1345 & -0.5049 + fe5015 & 3.9434 & -0.0178 & 0.0455 & 0.0004 & 0.0400 & -0.2780 & 0.0648 & -0.0140 & -0.0205 & 0.0082 & -0.0997 & 0.0669 & 0.5568 & -0.2650 & 0.3656 + mg@xmath35 & 0.0439 & 0.0030 & 0.0032 & 0.0016 & 0.0015 & -0.0010 & 0.0029 & 0.0005 & 0.0023 & 0.0026 & -0.0012 & 0.0027 & 0.0106 & -0.0015 & 0.0077 + mg@xmath36 & 0.1326 & 0.0039 & 0.0050 & 0.0018 & 0.0021 & -0.0049 & 0.0048 & 0.0014 & 0.0025 & 0.0029 & -0.0048 & 0.0043 & 0.0232 & -0.0045 & 0.0212 + mg @xmath12 & 2.0402 & 0.0364 & 0.0486 & 0.0154 & 0.0221 & -0.0648 & 0.0711 & 0.0351 & 0.0275 & 0.0328 & -0.0648 & 0.0538 & 0.2308 & -0.0252 & 0.3892 + fe5270 & 2.0614 & -0.0059 & 0.0285 & 0.0194 & 0.0261 & -0.1232 & 0.0658 & -0.0104 & 0.0078 & 0.0389 & -0.0319 & 0.0447 & 0.2649 & -0.1341 & 0.2021 + fe5335 & 1.8585 & -0.0153 & 0.0322 & 0.0140 & 0.0261 & -0.1486 & 0.0678 & -0.0079 & 0.0089 & 0.0359 & -0.0388 & 0.0442 & 0.3550 & -0.1747 & 0.1918 + fe5406 & 1.0693 & 0.0062 & 0.0258 & 0.0140 & 0.0196 & -0.0807 & 0.0499 & -0.0136 & 0.0091 & 0.0309 & -0.0168 & 0.0333 & 0.1989 & -0.0935 & 0.1176 + fe5709 & 0.7623 & -0.0270 & -0.0089 & 0.0082 & 0.0066 & -0.0558 & 0.0446 & -0.0249 & -0.0058 & 0.0220 & 0.0168 & 0.0140 & 0.0845 & -0.0766 & 0.0474 + fe5782 & 0.6065 & -0.0069 & 0.0004 & 0.0095 & 0.0079 & -0.0326 & 0.0265 & -0.0167 & -0.0005 & 0.0182 & -0.0059 & 0.0162 & 0.0971 & -0.0482 & 0.0300 + nad & 2.1600 & 0.1099 & 0.1152 & 0.0102 & 0.0364 & -0.1702 & -0.0175 & 0.0072 & 0.0417 & 0.0317 & -0.0799 & 0.0453 & 0.5593 & -0.2156 & 0.0953 + tio@xmath35 & 0.0261 & 0.0075 & 0.0050 & -0.0014 & 0.0007 & -0.0003 & -0.0061 & 0.0029 & 0.0014 & -0.0030 & -0.0085 & 0.0010 & 0.0184 & 0.0031 & 0.0000 + tio@xmath36 & 0.0293 & 0.0177 & 0.0117 & -0.0031 & 0.0017 & -0.0013 & -0.0132 & 0.0058 & 0.0037 & -0.0058 & -0.0166 & 0.0021 & 0.0409 & 0.0055 & -0.0009 + h@xmath50 & 3.4142 & -0.1725 & -0.3488 & -0.1243 & -0.1800 & 0.2600 & -0.5152 & -0.1256 & -0.2074 & -0.2125 & 0.0263 & -0.1989 & -0.2548 & -0.0748 & -1.6128 + h@xmath49 & 1.7203 & -0.2679 & -0.4682 & -0.1819 & -0.2445 & 0.3604 & -0.6567 & -0.1653 & -0.2646 & -0.2901 & 0.0721 & -0.2717 & -0.2918 & -0.1146 & -2.3564 + h@xmath52 & 2.8724 & -0.0923 & -0.1963 & -0.0585 & -0.0947 & 0.1466 & -0.2873 & -0.0819 & -0.1200 & -0.1046 & 0.0111 & -0.0961 & -0.1627 & -0.0260 & -0.8223 + h@xmath51 & 2.4631 & -0.1487 & -0.2424 & -0.0789 & -0.1168 & 0.1471 & -0.3364 & -0.1001 & -0.1562 & -0.1424 & 0.0165 & -0.1327 & -0.1709 & -0.0683 & -1.0344 + cn@xmath35 & -0.0855 & 0.0080 & 0.0114 & 0.0132 & 0.0154 & -0.0193 & 0.0021 & -0.0019 & 0.0046 & 0.0030 & -0.0033 & 0.0058 & 0.0085 & -0.0164 & 0.0155 + cn@xmath36 & -0.0429 & 0.0075 & 0.0104 & 0.0140 & 0.0160 & -0.0191 & -0.0001 & -0.0034 & 0.0058 & 0.0026 & -0.0038 & 0.0060 & 0.0061 & -0.0164 & 0.0117 + ca4227 & 0.7521 & -0.1012 & -0.0610 & -0.0716 & -0.0503 & -0.0975 & 0.0686 & 0.0048 & 0.0004 & 0.0056 & 0.2690 & 0.0222 & 0.1701 & 0.1554 & 0.6581 + g4300 & 2.6706 & 0.5505 & 0.7466 & 0.0031 & 0.0839 & -0.7120 & 0.1122 & -0.1588 & -0.0841 & 0.0722 & 0.0821 & 0.3310 & 0.1024 & -0.6526 & 0.4429 + fe4383 & 2.3834 & 0.0775 & 0.2567 & 0.0018 & 0.0693 & -0.6734 & 0.1963 & -0.2398 & -0.1479 & 0.0921 & -0.1526 & 0.2185 & 1.2160 & -1.0831 & 0.1032 + ca4455 & 1.0187 & -0.0234 & 0.0310 & -0.0143 & 0.0158 & -0.2026 & 0.0384 & -0.0364 & -0.0265 & 0.0087 & 0.0091 & 0.0575 & 0.0398 & -0.2344 & 0.1708 + fe4531 & 2.4730 & -0.0488 & 0.0539 & -0.0195 & 0.0325 & -0.3997 & 0.0635 & -0.1001 & -0.0858 & 0.0188 & -0.0422 & 0.3638 & 0.3113 & -0.3423 & 0.3601 + c@xmath36 4668 & 2.4464 & 0.7551 & 1.0555 & -0.1144 & 0.0595 & -1.2788 & -0.0475 & -0.1114 & -0.3407 & 0.0118 & -0.1040 & 0.3611 & 1.3397 & -1.4541 & 0.4946 + h@xmath11 & 3.4442 & -0.1495 & -0.1757 & -0.0424 & -0.0680 & 0.0788 & -0.1291 & -0.1227 & -0.1021 & -0.0747 & 0.0093 & -0.0370 & -0.0788 & -0.0796 & -0.5181 + fe5015 & 3.9434 & -0.2229 & -0.0170 & -0.0696 & 0.0419 & -0.7226 & 0.1514 & -0.2716 & -0.1163 & -0.0233 & -0.0512 & 0.5726 & 0.9488 & -0.5685 & 0.9104 + mg@xmath35 & 0.0439 & 0.0150 & 0.0179 & 0.0001 & 0.0014 & -0.0162 & -0.0021 & 0.0168 & -0.0010 & 0.0021 & -0.0034 & 0.0014 & 0.0011 & -0.0077 & 0.0150 + mg@xmath36 & 0.1326 & 0.0062 & 0.0119 & -0.0005 & 0.0022 & -0.0241 & 0.0026 & 0.0369 & -0.0034 & 0.0019 & -0.0059 & 0.0062 & 0.0136 & 0.0036 & 0.0664 + mg @xmath12 & 2.0402 & -0.0569 & 0.0150 & -0.0121 & 0.0243 & -0.2529 & 0.1000 & 0.9302 & -0.0697 & 0.0194 & -0.0693 & 0.0569 & 0.0326 & 0.5222 & 1.7033 + fe5270 & 2.0614 & -0.0455 & 0.0662 & -0.0014 & 0.0451 & -0.4508 & 0.0351 & -0.1186 & 0.0032 & 0.0294 & 0.0706 & 0.0922 & 0.5833 & -0.4600 & 0.2179 + fe5335 & 1.8585 & -0.0857 & 0.0333 & -0.0231 & 0.0248 & -0.4247 & 0.0640 & -0.1098 & 0.0047 & 0.0201 & -0.0508 & 0.0682 & 0.7263 & -0.5722 & 0.0612 + fe5406 & 1.0693 & -0.0319 & 0.0293 & -0.0036 & 0.0228 & -0.2486 & 0.0414 & -0.0677 & -0.0178 & 0.0216 & -0.0279 & 0.0117 & 0.5092 & -0.3938 & -0.0389 + fe5709 & 0.7623 & -0.0553 & -0.0115 & -0.0133 & -0.0021 & -0.1575 & 0.0396 & -0.0153 & -0.0096 & 0.0220 & 0.0142 & 0.0348 & 0.1979 & -0.1642 & 0.0991 + fe5782 & 0.6065 & -0.0290 & 0.0024 & -0.0054 & 0.0049 & -0.1244 & 0.0212 & -0.0028 & -0.0034 & 0.0128 & -0.0089 & 0.0057 & 0.0666 & -0.1526 & 0.0400 + nad & 2.1600 & 0.0791 & 0.1744 & -0.0118 & 0.0582 & -0.5319 & -0.0528 & -0.0788 & -0.0302 & 0.0130 & -0.1102 & 0.0081 & 0.4549 & -0.8234 & -0.2008 + tio@xmath35 & 0.0261 & 0.0033 & 0.0024 & -0.0011 & 0.0019 & 0.0003 & -0.0007 & 0.0028 & 0.0009 & -0.0033 & -0.0087 & 0.0087 & 0.0167 & 0.0103 & 0.0193 + tio@xmath36 & 0.0293 & 0.0109 & 0.0073 & -0.0034 & 0.0026 & -0.0015 & -0.0052 & 0.0053 & 0.0019 & -0.0064 & -0.0178 & 0.0122 & 0.0403 & 0.0118 & 0.0197 + h@xmath50 & 3.4142 & -0.2562 & -0.5249 & -0.1412 & -0.2482 & 0.6961 & -0.4302 & 0.1315 & -0.0090 & -0.1889 & 0.1551 & -0.2975 & -1.0744 & 0.8685 & -0.7611 + h@xmath49 & 1.7203 & -0.6871 & -1.0336 & -0.1228 & -0.2541 & 1.0152 & -0.5040 & 0.2810 & -0.0097 & -0.2575 & 0.0399 & -0.5986 & -0.6753 & 0.8979 & -1.6287 + h@xmath52 & 2.8724 & -0.0784 & -0.2105 & -0.0426 & -0.0950 & 0.2732 & -0.2815 & -0.0012 & 0.0166 & -0.0969 & 0.1023 & -0.1642 & -0.4546 & 0.3483 & -0.4492 + h@xmath51 & 2.4631 & -0.4012 & -0.5422 & -0.0592 & -0.1180 & 0.3864 & -0.2484 & 0.0823 & -0.0491 & -0.1323 & -0.0062 & -0.1572 & -0.2800 & 0.4076 & -0.5306 + cn@xmath35 & -0.0160 & -0.0080 & 0.0015 & -0.0040 & 0.0006 & -0.0170 & 0.0057 & -0.0010 & 0.0000 & -0.0020 & -0.0010 & -0.0010 & 0.0250 & -0.0170 & 0.0123 + cn@xmath36 & 0.0160 & -0.0090 & 0.0004 & -0.0040 & 0.0014 & -0.0160 & 0.0053 & -0.0010 & 0.0000 & -0.0020 & -0.0010 & -0.0010 & 0.0250 & -0.0160 & 0.0096 + ca4227 & 1.1160 & -0.0030 & 0.0619 & -0.0170 & 0.0095 & -0.0720 & 0.0261 & 0.0500 & 0.0340 & -0.0060 & -0.0060 & -0.0040 & 0.0940 & -0.0700 & 0.1238 + g4300 & 4.6200 & -0.1490 & 0.1445 & -0.0610 & 0.0251 & -0.3510 & 0.2432 & 0.0590 & 0.0470 & -0.0380 & -0.0120 & -0.0130 & 0.3920 & -0.3100 & 0.4973 + fe4383 & 4.5220 & -0.3140 & 0.0489 & -0.1290 & 0.0059 & -0.6340 & 0.1811 & 0.0170 & 0.0090 & -0.0710 & -0.0210 & -0.0420 & 0.8690 & -0.7030 & 0.4978 + ca4455 & 1.4560 & -0.0370 & 0.0378 & -0.0290 & 0.0060 & -0.1600 & 0.0192 & 0.0180 & 0.0110 & -0.0130 & -0.0090 & -0.0090 & 0.2240 & -0.1840 & 0.0959 + fe4531 & 3.1680 & -0.0680 & 0.0393 & -0.0450 & 0.0030 & -0.2240 & 0.0203 & 0.0250 & 0.0160 & -0.0220 & -0.0140 & -0.0220 & 0.2740 & -0.2570 & 0.1394 + c@xmath36 4668 & 4.1090 & -0.3150 & 0.0401 & -0.1610 & 0.0200 & -0.8720 & 0.0818 & -0.0310 & -0.0510 & -0.0910 & -0.0400 & -0.0270 & 1.4740 & -1.0680 & 0.2875 + h@xmath11 & 2.2820 & -0.1130 & -0.1915 & 0.0040 & -0.0420 & 0.0820 & -0.0787 & -0.0940 & -0.0880 & -0.0090 & 0.0160 & -0.0060 & -0.0930 & 0.0170 & -0.2373 + fe5015 & 4.9640 & -0.1970 & -0.0111 & -0.0820 & -0.0064 & -0.4030 & 0.0322 & -0.0070 & -0.0290 & -0.0500 & -0.0200 & -0.0250 & 0.5470 & -0.4970 & 0.1363 + mg@xmath35 & 0.0800 & -0.0030 & 0.0025 & -0.0030 & 0.0001 & -0.0110 & 0.0002 & 0.0030 & 0.0020 & -0.0020 & -0.0020 & -0.0020 & 0.0120 & -0.0120 & 0.0095 + mg@xmath36 & 0.2050 & -0.0080 & 0.0049 & -0.0050 & 0.0007 & -0.0200 & 0.0037 & 0.0050 & 0.0030 & -0.0020 & -0.0020 & -0.0020 & 0.0250 & -0.0220 & 0.0199 + mg @xmath12 & 3.2300 & -0.1530 & 0.0614 & -0.0600 & 0.0025 & -0.2520 & 0.1154 & 0.0720 & 0.0380 & -0.0330 & -0.0070 & -0.0160 & 0.3280 & -0.2640 & 0.3600 + fe5270 & 2.7650 & -0.1510 & -0.0339 & -0.0550 & -0.0152 & -0.2320 & 0.0187 & -0.0080 & -0.0140 & -0.0310 & -0.0100 & -0.0310 & 0.2640 & -0.2780 & 0.1220 + fe5335 & 2.5330 & -0.1630 & -0.0323 & -0.0640 & -0.0164 & -0.2680 & 0.0216 & -0.0020 & -0.0130 & -0.0360 & -0.0120 & -0.0330 & 0.3600 & -0.3270 & 0.1322 + fe5406 & 1.5470 & -0.0980 & -0.0167 & -0.0440 & -0.0085 & -0.1850 & 0.0105 & -0.0040 & -0.0080 & -0.0250 & -0.0130 & -0.0240 & 0.2290 & -0.2210 & 0.0842 + fe5709 & 0.9590 & -0.1070 & -0.0565 & -0.0280 & -0.0137 & -0.0920 & 0.0209 & -0.0280 & -0.0250 & -0.0200 & -0.0040 & -0.0160 & 0.1130 & -0.1180 & 0.0379 + fe5782 & 0.7840 & -0.0690 & -0.0345 & -0.0230 & -0.0087 & -0.0790 & 0.0008 & -0.0160 & -0.0140 & -0.0160 & -0.0060 & -0.0150 & 0.0860 & -0.1020 & 0.0194 + nad & 2.8340 & 0.1030 & 0.1704 & -0.0390 & 0.0247 & -0.3660 & -0.0616 & 0.0520 & 0.0520 & 0.0030 & -0.0230 & -0.0210 & 0.5570 & -0.4300 & 0.0894 + tio@xmath35 & 0.0240 & 0.0160 & 0.0131 & 0.0020 & 0.0043 & -0.0010 & -0.0036 & 0.0050 & 0.0050 & 0.0020 & -0.0010 & 0.0020 & 0.0060 & -0.0010 & -0.0038 + tio@xmath36 & 0.0250 & 0.0360 & 0.0288 & 0.0040 & 0.0089 & -0.0040 & -0.0088 & 0.0110 & 0.0100 & 0.0040 & -0.0020 & 0.0050 & 0.0130 & -0.0030 & -0.0086 + h@xmath50 & -0.3970 & 0.1820 & -0.3275 & 0.1110 & -0.0640 & 0.6440 & -0.3895 & -0.1350 & -0.1160 & 0.0570 & 0.0340 & 0.0190 & -0.8510 & 0.5680 & -0.9208 + h@xmath49 & -3.6950 & 0.1560 & -0.4981 & 0.1270 & -0.0870 & 0.7590 & -0.4980 & -0.2270 & -0.1880 & 0.0620 & 0.0450 & 0.0310 & -0.8820 & 0.6230 & -1.1746 + h@xmath52 & 0.9500 & 0.0750 & -0.1692 & 0.0430 & -0.0354 & 0.2540 & -0.2056 & -0.0810 & -0.0670 & 0.0230 & 0.0090 & 0.0040 & -0.2930 & 0.1900 & -0.4549 + h@xmath51 & -0.2360 & 0.0820 & -0.2690 & 0.0620 & -0.0464 & 0.3570 & -0.2836 & -0.1350 & -0.1110 & 0.0290 & 0.0190 & 0.0140 & -0.4020 & 0.2630 & -0.6679 + cn@xmath35 & -0.0160 & 0.0040 & 0.0187 & 0.0180 & 0.0260 & -0.0390 & -0.0063 & -0.0060 & 0.0070 & -0.0030 & -0.0060 & 0.0010 & 0.0190 & -0.0400 & -0.0009 + cn@xmath36 & 0.0160 & 0.0030 & 0.0187 & 0.0200 & 0.0287 & -0.0420 & -0.0084 & -0.0090 & 0.0120 & -0.0030 & -0.0070 & 0.0020 & 0.0180 & -0.0410 & -0.0021 + ca4227 & 1.1160 & -0.2170 & -0.1256 & -0.1620 & -0.1190 & -0.1410 & 0.1285 & 0.0330 & -0.0010 & -0.0160 & 0.4660 & 0.0000 & 0.1940 & 0.2300 & 0.9642 + g4300 & 4.6200 & 0.5190 & 0.8674 & -0.0710 & 0.0368 & -0.8280 & -0.0776 & -0.2190 & -0.1660 & -0.0490 & 0.1060 & 0.2570 & -0.0520 & -0.9350 & -0.0886 + fe4383 & 4.5220 & -0.2140 & 0.2685 & -0.1920 & -0.0011 & -1.1120 & 0.1418 & -0.4310 & -0.3010 & -0.0970 & -0.2120 & 0.0520 & 2.1490 & -1.9830 & -0.7360 + ca4455 & 1.4560 & -0.0800 & 0.0303 & -0.0490 & 0.0037 & -0.2930 & 0.0269 & -0.0320 & -0.0260 & -0.0210 & 0.0310 & 0.0320 & 0.0020 & -0.3440 & 0.1431 + fe4531 & 3.1680 & -0.1420 & 0.0349 & -0.0760 & 0.0082 & -0.5120 & 0.0198 & -0.1300 & -0.1460 & -0.0380 & -0.0470 & 0.3320 & 0.3180 & -0.5500 & 0.2122 + c@xmath36 4668 & 4.1090 & 0.8430 & 1.5366 & -0.2990 & 0.0157 & -1.9260 & -0.1637 & -0.1800 & -0.5090 & -0.1330 & -0.0200 & 0.3330 & 1.5040 & -2.2890 & -0.0120 + h@xmath11 & 2.2820 & -0.1150 & -0.2097 & 0.0130 & -0.0409 & 0.1530 & -0.0547 & -0.2150 & -0.0920 & -0.0050 & 0.0190 & 0.0410 & -0.1150 & 0.0510 & -0.3246 + fe5015 & 4.9640 & -0.4440 & -0.1407 & -0.1400 & -0.0014 & -0.7460 & 0.1583 & -0.4480 & -0.1170 & -0.0780 & 0.0230 & 0.6160 & 0.8830 & -0.7110 & 0.6491 + mg@xmath35 & 0.0800 & 0.0150 & 0.0260 & -0.0050 & -0.0002 & -0.0320 & -0.0095 & 0.0330 & -0.0030 & -0.0020 & -0.0050 & -0.0060 & -0.0090 & -0.0180 & 0.0248 + mg@xmath36 & 0.2050 & -0.0020 & 0.0167 & -0.0080 & 0.0008 & -0.0460 & -0.0019 & 0.0600 & -0.0050 & -0.0040 & -0.0040 & -0.0010 & 0.0040 & -0.0090 & 0.0802 + mg @xmath12 & 3.2300 & -0.2680 & 0.0058 & -0.0890 & 0.0059 & -0.4470 & 0.1651 & 1.2620 & -0.0990 & -0.0480 & -0.0180 & -0.0110 & -0.0050 & 0.4790 & 1.9536 + fe5270 & 2.7650 & -0.1710 & 0.0252 & -0.0700 & 0.0108 & -0.5900 & -0.0448 & -0.1810 & -0.0840 & -0.0440 & 0.1070 & 0.0090 & 0.7220 & -0.7290 & -0.0166 + fe5335 & 2.5330 & -0.2270 & -0.0200 & -0.1030 & -0.0178 & -0.5810 & -0.0030 & -0.1700 & -0.0880 & -0.0530 & -0.0290 & -0.0130 & 0.9090 & -0.8720 & -0.1698 + fe5406 & 1.5470 & -0.1330 & -0.0004 & -0.0650 & -0.0039 & -0.3920 & -0.0078 & -0.0970 & -0.0610 & -0.0360 & -0.0310 & -0.0480 & 0.6130 & -0.6190 & -0.1545 + fe5709 & 0.9590 & -0.1340 & -0.0566 & -0.0510 & -0.0244 & -0.2040 & 0.0108 & -0.0230 & -0.0300 & -0.0250 & -0.0080 & 0.0130 & 0.2280 & -0.2180 & 0.0879 + fe5782 & 0.7840 & -0.0930 & -0.0307 & -0.0410 & -0.0125 & -0.1900 & -0.0106 & -0.0060 & -0.0170 & -0.0220 & -0.0100 & -0.0300 & 0.0540 & -0.2300 & 0.0204 + nad & 2.8340 & 0.0680 & 0.2474 & -0.0670 & 0.0521 & -0.8180 & -0.1204 & -0.0810 & -0.0410 & -0.0210 & -0.0580 & -0.0770 & 0.3990 & -1.1940 & -0.3016 + tio@xmath35 & 0.0240 & 0.0100 & 0.0081 & 0.0020 & 0.0058 & 0.0010 & 0.0039 & 0.0050 & 0.0030 & 0.0010 & -0.0010 & 0.0130 & 0.0040 & 0.0100 & 0.0189 + tio@xmath36 & 0.0250 & 0.0270 & 0.0213 & 0.0030 & 0.0098 & -0.0030 & 0.0001 & 0.0100 & 0.0080 & 0.0030 & -0.0040 & 0.0200 & 0.0140 & 0.0070 & 0.0175 + h@xmath50 & -0.3970 & 0.0210 & -0.6357 & 0.0430 & -0.2047 & 1.2520 & -0.2448 & 0.3900 & 0.2490 & 0.0890 & 0.2480 & -0.1100 & -2.4010 & 2.0720 & 0.5809 + h@xmath49 & -3.6950 & -0.4320 & -1.2520 & 0.1800 & -0.1069 & 1.5670 & -0.1989 & 0.5510 & 0.2050 & 0.0970 & -0.0050 & -0.4030 & -1.6820 & 2.0990 & 0.1432 + h@xmath52 & 0.9500 & 0.0750 & -0.1937 & 0.0510 & -0.0403 & 0.3450 & -0.2139 & 0.0560 & 0.2200 & 0.0300 & 0.1370 & -0.0910 & -0.7540 & 0.7100 & 0.1785 + h@xmath51 & -0.2360 & -0.2810 & -0.6841 & 0.0770 & -0.0494 & 0.6430 & -0.1163 & 0.1720 & 0.0500 & 0.0380 & -0.0150 & -0.0170 & -0.6380 & 0.9330 & 0.1314 + cn@xmath35 & 0.0230 & -0.0120 & -0.0055 & -0.0030 & 0.0007 & -0.0210 & 0.0055 & -0.0010 & -0.0020 & -0.0020 & 0.0000 & 0.0010 & 0.0340 & -0.0200 & 0.0070 + cn@xmath36 & 0.0550 & -0.0140 & -0.0069 & -0.0030 & 0.0016 & -0.0210 & 0.0079 & 0.0010 & -0.0020 & -0.0020 & 0.0010 & 0.0020 & 0.0370 & -0.0210 & 0.0080 + ca4227 & 1.4710 & 0.0270 & 0.0843 & -0.0110 & 0.0059 & -0.0880 & 0.0026 & 0.0920 & 0.0620 & -0.0050 & -0.0050 & -0.0030 & 0.0960 & -0.0630 & 0.1600 + g4300 & 5.7320 & -0.1830 & -0.0670 & -0.0360 & 0.0166 & -0.3040 & 0.1150 & 0.0810 & 0.0280 & -0.0230 & 0.0200 & 0.0290 & 0.2600 & -0.2100 & 0.2130 + fe4383 & 6.0160 & -0.4690 & -0.1835 & -0.1080 & 0.0360 & -0.7020 & 0.1713 & 0.1560 & 0.0310 & -0.0520 & 0.0560 & 0.0660 & 0.8990 & -0.6830 & 0.3990 + ca4455 & 1.7870 & -0.0250 & 0.0382 & -0.0190 & 0.0066 & -0.1700 & 0.0073 & 0.0480 & 0.0260 & -0.0100 & -0.0010 & 0.0020 & 0.2290 & -0.1720 & 0.0980 + fe4531 & 3.6960 & -0.0770 & 0.0233 & -0.0310 & 0.0125 & -0.2440 & 0.0177 & 0.0840 & 0.0390 & -0.0150 & 0.0070 & 0.0100 & 0.3030 & -0.2470 & 0.1440 + c@xmath36 4668 & 4.9720 & -0.3210 & -0.0312 & -0.1210 & 0.0114 & -0.9250 & 0.1003 & -0.0170 & -0.0670 & -0.0750 & -0.0050 & 0.0200 & 1.6420 & -1.0790 & 0.2410 + h@xmath11 & 1.7130 & -0.2470 & -0.2583 & -0.0070 & 0.0029 & 0.0740 & 0.0319 & -0.0800 & -0.0930 & 0.0010 & 0.0450 & 0.0430 & -0.0280 & -0.0090 & -0.1820 + fe5015 & 5.4890 & -0.2120 & -0.0932 & -0.0440 & 0.0142 & -0.3710 & 0.0766 & 0.0390 & -0.0160 & -0.0260 & 0.0230 & 0.0400 & 0.5760 & -0.4240 & 0.1250 + mg@xmath35 & 0.1160 & -0.0040 & 0.0027 & -0.0020 & 0.0012 & -0.0140 & 0.0023 & 0.0090 & 0.0040 & -0.0010 & 0.0000 & 0.0000 & 0.0160 & -0.0130 & 0.0150 + mg@xmath36 & 0.2700 & -0.0090 & 0.0021 & -0.0030 & 0.0019 & -0.0230 & 0.0055 & 0.0120 & 0.0060 & -0.0010 & 0.0010 & 0.0020 & 0.0280 & -0.0210 & 0.0240 + mg @xmath12 & 4.1370 & -0.1980 & -0.0534 & -0.0420 & 0.0196 & -0.2520 & 0.1093 & 0.1370 & 0.0600 & -0.0120 & 0.0340 & 0.0460 & 0.3190 & -0.2200 & 0.3210 + fe5270 & 3.2950 & -0.2250 & -0.1150 & -0.0400 & 0.0185 & -0.2420 & 0.0640 & 0.0720 & 0.0100 & -0.0150 & 0.0360 & 0.0400 & 0.3040 & -0.2530 & 0.1280 + fe5335 & 3.0500 & -0.2380 & -0.1058 & -0.0460 & 0.0179 & -0.2780 & 0.0770 & 0.0820 & 0.0120 & -0.0200 & 0.0350 & 0.0380 & 0.3990 & -0.3010 & 0.1490 + fe5406 & 1.9140 & -0.1380 & -0.0613 & -0.0310 & 0.0130 & -0.1940 & 0.0503 & 0.0530 & 0.0080 & -0.0150 & 0.0190 & 0.0230 & 0.2620 & -0.2050 & 0.1040 + fe5709 & 1.0960 & -0.1680 & -0.1259 & -0.0210 & 0.0090 & -0.0900 & 0.0651 & 0.0090 & -0.0240 & -0.0110 & 0.0250 & 0.0270 & 0.1290 & -0.1000 & 0.0350 + fe5782 & 0.9090 & -0.1060 & -0.0679 & -0.0160 & 0.0067 & -0.0780 & 0.0369 & 0.0140 & -0.0110 & -0.0090 & 0.0130 & 0.0150 & 0.1060 & -0.0890 & 0.0290 + nad & 3.5260 & 0.1960 & 0.3170 & -0.0360 & 0.0104 & -0.4510 & -0.1153 & 0.1110 & 0.1090 & -0.0040 & -0.0380 & -0.0370 & 0.6460 & -0.5090 & 0.1800 + tio@xmath35 & 0.0140 & 0.0290 & 0.0253 & 0.0020 & -0.0013 & -0.0010 & -0.0111 & -0.0020 & 0.0030 & 0.0000 & -0.0060 & -0.0060 & 0.0030 & -0.0020 & -0.0020 + tio@xmath36 & 0.0020 & 0.0660 & 0.0572 & 0.0050 & -0.0018 & 9.4270 & -0.0238 & -0.0020 & 0.0090 & 0.0010 & -0.0110 & -0.0100 & 0.0090 & -0.0030 & -0.0050 + h@xmath50 & -2.8110 & 0.1500 & -0.1331 & 0.0750 & -0.0324 & 0.7440 & -0.1547 & -0.2430 & -0.1470 & 0.0470 & 0.0020 & -0.0220 & -0.7950 & 0.5340 & -0.6520 + h@xmath49 & -6.5830 & 0.0410 & -0.2453 & 0.0640 & -0.0327 & 0.7250 & -0.1017 & -0.2980 & -0.1940 & 0.0430 & 0.0170 & -0.0040 & -0.6860 & 0.4890 & -0.6550 + h@xmath52 & -0.1300 & 0.0610 & -0.0451 & 0.0280 & -0.0136 & 0.2780 & -0.0637 & -0.0930 & -0.0590 & 0.0180 & -0.0030 & -0.0090 & -0.2650 & 0.1780 & -0.2530 + h@xmath51 & -1.8520 & 0.0330 & -0.1164 & 0.0310 & -0.0195 & 0.3440 & -0.0767 & -0.1770 & -0.1170 & 0.0180 & 0.0010 & -0.0100 & -0.2910 & 0.1910 & -0.3920 + cn@xmath35 & 0.0230 & 0.0100 & 0.0236 & 0.0310 & 0.0383 & -0.0540 & -0.0143 & -0.0110 & 0.0130 & -0.0040 & -0.0070 & 0.0030 & 0.0240 & -0.0540 & -0.0140 + cn@xmath36 & 0.0550 & 0.0080 & 0.0240 & 0.0330 & 0.0422 & -0.0620 & -0.0143 & -0.0140 & 0.0220 & -0.0030 & -0.0070 & 0.0050 & 0.0240 & -0.0590 & -0.0080 + ca4227 & 1.4710 & -0.2730 & -0.1853 & -0.2220 & -0.1875 & -0.1190 & 0.1724 & 0.0770 & 0.0110 & -0.0150 & 0.6420 & -0.0040 & 0.1600 & 0.4320 & 1.3030 + g4300 & 5.7320 & 0.4910 & 0.6277 & -0.0310 & 0.0331 & -0.8630 & -0.3989 & -0.1610 & -0.1260 & -0.0290 & 0.1830 & 0.3650 & -0.2060 & -0.7810 & -0.3130 + fe4383 & 6.0160 & -0.3750 & 0.0179 & -0.1750 & 0.0263 & -1.2200 & 0.1250 & -0.4790 & -0.3520 & -0.0770 & -0.2130 & 0.1250 & 2.6260 & -2.3680 & -1.3250 + ca4455 & 1.7870 & -0.0740 & 0.0280 & -0.0430 & 0.0035 & -0.3180 & 0.0181 & -0.0160 & -0.0220 & -0.0190 & 0.0470 & 0.0530 & -0.0230 & -0.3530 & 0.1470 + fe4531 & 3.6960 & -0.1520 & 0.0185 & -0.0610 & 0.0181 & -0.5310 & 0.0162 & -0.1260 & -0.1760 & -0.0310 & -0.0500 & 0.4480 & 0.3590 & -0.5860 & 0.1700 + c@xmath36 4668 & 4.9720 & 1.0990 & 1.7348 & -0.2790 & 0.0052 & -2.1610 & -0.2334 & -0.2170 & -0.6250 & -0.1310 & 0.0210 & 0.4600 & 1.6750 & -2.5670 & -0.2460 + h@xmath11 & 1.7130 & -0.2480 & -0.2812 & 0.0040 & 0.0048 & 0.1630 & 0.0612 & -0.2890 & -0.0950 & 0.0050 & 0.0500 & 0.0970 & -0.0250 & -0.0080 & -0.3730 + fe5015 & 5.4890 & -0.4750 & -0.2428 & -0.0970 & 0.0215 & -0.6820 & 0.2140 & -0.6200 & -0.1050 & -0.0530 & 0.0710 & 0.8380 & 0.9100 & -0.6330 & 0.5730 + mg@xmath35 & 0.1160 & 0.0180 & 0.0307 & -0.0050 & 0.0009 & -0.0410 & -0.0103 & 0.0570 & -0.0030 & -0.0020 & -0.0050 & -0.0040 & -0.0110 & -0.0160 & 0.0450 + mg@xmath36 & 0.2700 & -0.0010 & 0.0173 & -0.0070 & 0.0020 & -0.0550 & -0.0021 & 0.0840 & -0.0040 & -0.0030 & -0.0030 & 0.0020 & 0.0010 & -0.0030 & 0.0990 + mg @xmath12 & 4.1370 & -0.3100 & -0.1128 & -0.0660 & 0.0244 & -0.4250 & 0.1648 & 1.4420 & -0.0780 & -0.0250 & 0.0190 & 0.0550 & -0.0480 & 0.6850 & 2.0170 + fe5270 & 3.2950 & -0.2320 & -0.0428 & -0.0510 & 0.0469 & -0.6220 & -0.0147 & -0.1610 & -0.0730 & -0.0290 & 0.1730 & 0.0650 & 0.8510 & -0.7940 & -0.1220 + fe5335 & 3.0500 & -0.3020 & -0.0871 & -0.0890 & 0.0165 & -0.6300 & 0.0424 & -0.1510 & -0.0810 & -0.0380 & 0.0060 & 0.0550 & 1.0550 & -0.9860 & -0.2880 + fe5406 & 1.9140 & -0.1720 & -0.0390 & -0.0530 & 0.0182 & -0.4310 & 0.0254 & -0.0760 & -0.0550 & -0.0270 & -0.0080 & 0.0010 & 0.7100 & -0.6880 & -0.2110 + fe5709 & 1.0960 & -0.1930 & -0.1245 & -0.0450 & -0.0018 & -0.2050 & 0.0531 & 0.0120 & -0.0310 & -0.0170 & 0.0190 & 0.0640 & 0.2460 & -0.2030 & 0.0870 + fe5782 & 0.9090 & -0.1300 & -0.0630 & -0.0350 & 0.0026 & -0.2020 & 0.0225 & 0.0200 & -0.0150 & -0.0160 & 0.0090 & -0.0050 & 0.0690 & -0.2420 & 0.0100 + nad & 3.5260 & 0.1580 & 0.4102 & -0.0720 & 0.0387 & -0.9830 & -0.1881 & -0.0610 & 0.0060 & -0.0320 & -0.0860 & -0.1160 & 0.4590 & -1.4500 & -0.3080 + tio@xmath35 & 0.0140 & 0.0210 & 0.0191 & 0.0020 & -0.0001 & 0.0010 & -0.0021 & -0.0020 & 0.0020 & -0.0010 & -0.0070 & 0.0100 & 0.0000 & 0.0120 & 0.0280 + tio@xmath36 & 0.0020 & 0.0540 & 0.0479 & 0.0040 & -0.0010 & -0.0020 & -0.0129 & -0.0020 & 0.0060 & 0.0010 & -0.0130 & 0.0130 & 0.0090 & 0.0100 & 0.0320 + h@xmath50 & -2.8110 & -0.0710 & -0.5336 & -0.0450 & -0.2458 & 1.5700 & 0.0769 & 0.5310 & 0.3480 & 0.0870 & 0.3360 & -0.1930 & -3.0410 & 2.7180 & 1.5670 + h@xmath49 & -6.5830 & -0.5090 & -0.9420 & 0.1160 & -0.0585 & 1.7290 & 0.3864 & 0.6840 & 0.1980 & 0.0780 & -0.0620 & -0.5500 & -2.0090 & 2.2860 & 0.9150 + h@xmath52 & -0.1300 & 0.0660 & -0.0658 & 0.0330 & -0.0243 & 0.3670 & -0.0854 & 0.0570 & 0.4080 & 0.0250 & 0.1770 & -0.1240 & -0.9350 & 0.9120 & 0.6730 + h@xmath51 & -1.8520 & -0.3450 & -0.5269 & 0.0450 & -0.0228 & 0.7050 & 0.1915 & 0.1800 & 0.0320 & 0.0250 & -0.0530 & -0.0540 & -0.6990 & 0.9790 & 0.5160 +
|
the first paper in this series explored the effects of altering the chemical mixture of the stellar population on an element by element basis on stellar evolutionary tracks and isochrones to the end of the red giant branch .
this paper extends the discussion by incorporating the fully consistent synthetic stellar spectra with those isochrone models in predicting integrated colors , lick indices , and synthetic spectra .
older populations display element ratio effects in their spectra at higher amplitude than younger populations . in addition , spectral effects in the photospheres of stars tend to dominate over effects from isochrone temperatures and lifetimes , but , further , the isochrone - based effects that are present tend to fall along the age - metallicity degeneracy vector , while the direct stellar spectral effects usually show considerable orthogonality .
|
over the past few years our preliminary knowledge on the frequency of extrasolar planets in the low mass part of the hr - diagram has increased significantly . an ever increasing number of m dwarfs
is being monitored by various doppler searches with high radial velocity ( rv ) precision .
this has led to several discoveries of m dwarf planets that cover an enormous and surprising range in mass , almost comparable to the mass range of the planets in our solar system .
jovian planets were found around gj 876 ( delfosse et al .
1998 , marcy et al . 1998 , 2001 ) ,
gj 849 ( butler et al .
2006 ) and gj 317 ( johnson et al .
2007 ) , and neptune - mass planets around gj 436 ( butler et al . 2004 ) , gj 581 ( bonfils et al .
2005b ) and gj 674 ( bonfils et al .
2007 ) . the low primary masses of m
dwarfs combined with state - of - the - art rv precision even allowed the detection of additional planets with minimum masses below @xmath3 ( so - called `` super - earths '' ) in the gj 876 ( rivera et al .
2005 ) and gj 581 ( udry et al . 2007 ) systems .
in this paper we report the detection of a new low mass planet in a @xmath2 d orbit around the m dwarf gj 176 .
this is the fourth planet in the class of neptune - mass companions orbiting m dwarfs .
gj 176 is already the third ( previously unknown ) planet host in our het m dwarf planet search ( endl et al .
2003 , 2006 ) sample .
gj 176 ( hd 285968 , hip 21932 , lhs 196 ) is a @xmath4 m2.5ve star at a distance of @xmath5 parsec , according to the hipparcos parallax of @xmath6 mas ( perryman et al .
the star has a @xmath7 of @xmath8 and an absolute @xmath9 magnitude of @xmath10 .
the 2mass magnitudes for gj 176 are @xmath11 , @xmath12 and @xmath13 mag ( cutri et al . 2003 ) . using the v - band and k - band mass - luminosity relationships of delfosse et al .
( 2000 ) , we estimate a mass of @xmath14 and @xmath15 , respectively . we adopt the mean value of @xmath16 as the mass of the star
. a photometric study by weis ( 1994 ) did nt find significant variability for gj 176 .
the v - band scatter measured by weis ( 1994 ) is 0.006 mags , equal to the measurement uncertainties .
the rosat all - sky - survey catalog of nearby stars ( hnsch et al . 1999 ) reports a moderate coronal x - ray emission level of @xmath17 .
gj 176 is thus a moderately active star , possibly exhibiting starspots and flares , a quite typical behavior for m dwarfs .
rauscher & marcy ( 2006 ) measured the equivalent widths of the ca ii h and k lines with @xmath18 and @xmath19 , respectively . based on the v - band mass - metallicity relationship of bonfils et al . ( 2005a ) we estimate an [ fe / h]@xmath20 for gj 176 , so roughly solar metallicity .
we observed gj 176 as part of our on - going doppler search for planets around m dwarfs ( endl et al .
2003 , 2006 ) using the het ( ramsey et al .
1998 ) and its hrs spectrograph ( tull 1998 ) .
we started observations of gj 176 in 2003 october 15 and collected a total of 28 spectra over 5 years .
all observations were performed using our standard planet search setup and data reduction pipeline decribed in detail in cochran et al .
we use the common i@xmath21 cell technique to obtain high precision differential rv measurements ( e.g. butler et al .
1995 , endl et al .
2000 ) .
[ rvs ] shows the time series of our het / hrs rv measurements with the small secular acceleration of the rv of @xmath22 , as computed from the hipparcos parallax and proper motion information , already subtracted .
the data have an overall rms - scatter of @xmath23 and average internal errors of @xmath24 . the total scatter is more than twice the measurement uncertainties , indicative of intrinsic rv variability of this target .
the het rv data are listed in table [ data ] .
we searched the rv data of gj 176 for any significant periodicities using the classic lomb - scargle periodogram ( lomb 1976 ; scargle 1982 ) .
[ lomb ] displays the resulting power spectrum .
a strong peak is visible at a period of @xmath2 d. we estimate the false - alarm - probability ( fap ) of this signal with the bootstrap randomization scheme ( e.g. krster et al .
after 100,000 bootstrap re - shuffling runs we find that the fap of the @xmath2 d signal is only 0.0004 . as the next step we use gaussfit ( jeffereys et al .
1988 ) to find a keplerian orbital solution to our rv data .
a circular orbit fit yields a @xmath25 of @xmath26 ( @xmath27 ) and a residual rv scatter of @xmath28 .
a slightly better fit is obtained with an eccentric orbit with @xmath29 : @xmath25 of @xmath30 ( @xmath31 ) and residual rms of @xmath32 .
the large uncertainity in @xmath33 and the fact that a circular orbit yields a lower @xmath34 urges us to remain cautious concerning the reality of the non - circular orbit .
however , future observations will allow us to determine whether the orbit is indeed eccentric .
a moderately eccentric orbit could be an indication for additional planets around this star and thus warrants intensive follow - up monitoring .
[ phase ] shows the rv measurements and both orbital solutions ( circular and eccentric ) phased to the best - fit period .
the circular orbital solution yields a minimum mass for the companion of @xmath35 , while an eccentric orbit would lower the minimum mass slightly to @xmath36 .
the orbital parameters are summarized in table [ tab : planet2 ] .
the case of gj 674 has demonstrated clearly that star spots can introduce low level signals in high precision rv data of m dwarfs ( bonfils et al .
can the gj 176 signal also be caused by rotational modulation due to star spots and not a planet ?
this scenario is very unlikely because a rotation period of @xmath37 d would mean that gj 176 would rotate more than 3 times faster than gj 674 , and this should lead to a higher activity level than gj 674 .
this is not the case , as gj 176 has a slightly lower coronal x - ray emission than gj 674 ( hnsch et al .
1999 ) . as mention before , weis ( 1994 ) did nt detect photometric variability in gj 176 and we also did not find any significant variability or periodicity in the hipparcos photometry for this star ( the highest peak at @xmath38 d has a fap of @xmath39 . ) moreover , the vast majority of the data of bonfils et al .
( 2007 ) was collected over a relatively short period of time ( @xmath40 d ) , where rotational modulation can mimic a keplerian signal of a planet because the active regions on the star remain constant for this period of time . however , over a larger amount of time , active regions will reconfigure and emerge at different stellar longitudes causing a phase as well as amplitude change in the rv signal .
the gj 176 signal remains stable in phase and amplitude over 5 years , which makes the planet hypothesis much more plausible as the origin of the rv signal .
remarkably , gj 176 is already the third m dwarf with a neptune - mass companion that was included in our het sample of @xmath41 m dwarfs .
gj 436 and gj 581 were part of our m dwarf sample before the planets around them were announced .
the frequency of short - periodic neptune - mass planets around m dwarfs in our het sample is hence @xmath42 , which is higher than the frequency of hot jupiters around fgk - type stars of @xmath43 ( marcy et al .
neptunes and super - earths could be relatively common around low mass stars .
of course , with m dwarfs we are still limited by small number statistics , as compared to the few thousands of fgk - type stars already observed by various doppler surveys around the world .
we are currently increasing the size of the het m dwarf sample , and combined with the results of other programs observing m dwarfs , we should be able to derive important constraints for planet formation models .
the present information about neptune - mass planets on short - period orbits suggests that they may be the tip of the terrestrial planet distribution , for several reasons .
first , a number of hot neptunes have 2 or 3 sibling gas giant planets orbiting at much greater distances : @xmath44 1 cnc , @xmath45 ara , and gj 876 all have such systems .
this planetary system architecture is the same as our own solar system , with inner terrestrial planets and outer gas giant planets , and suggests that the hot neptunes in these systems formed inside their gas giants , making them likely to be rocky planets .
theoretical models of the collisional accumulation of terrestrial planets predict that rocky planets as massive as about @xmath46 would result from a protoplanetary disk with a surface density of gas and solids high enough to form a gas giant planet by core accretion in a few million years ( wetherill 1996 ; inaba et al .
formation of the neptunes interior to their gas giants seems to be a much more likely scenario than formation as ice giants on orbits outside the orbits of the gas giants , followed by migration somehow past the gas giants to their current short - period orbits . while gj 876 is the only m dwarf of these three stars
, one might expect a common explanation for its hot neptune as a rocky world , as well as those of gj 436 and gj 581 .
this interpretation bodes well for the eventual detection of earth - mass planets on habitable orbits around low mass stars . for
m dwarfs with masses of @xmath47 we already have the sensitivity to detect @xmath48 planets in the habitable zone ( endl & krster 2007 , in prep . ) .
while m dwarfs appear to have hot neptunes at a higher rate than g dwarfs , they appear to have short - period gas giant planets at a somewhat smaller rate than g dwarfs ( endl et al .
2006 ; johnson et al .
2007 ) . similarly
, microlensing surveys have detected planetary companions at asteroidal distances to m dwarf stars at frequencies that suggest that planets at such locations are more likely to be neptune - mass than jupiter - mass .
these results may simply be a result of m dwarfs have lower mass protoplanetary disks than g dwarfs , with the consequent result that the planets that form tend to have a mass distribution that is shifted downward .
however , this apparent preference for cold super - earths over jupiters can also be explained at present by both the core accretion and disk instability formation mechanism for gas giant planets . in the case of core accretion
, the lengthened orbital periods around m dwarfs could lead to most cores growing too slowly to accrete significant gaseous envelopes , making failed cores the usual outcome , rather than gas giants ( laughlin et al .
disk instability explains the appearance of both gas giants and super - earths at asteroidal distances around m dwarfs by appealing to rapid formation of gaseous protoplanets by disk instability , followed by conversion to ice giants ( cold super - earths ) by photoevaporation at asteroidal distances in region of high mass star formation ( boss 2006 ) .
because most m dwarfs form in regions of high mass formation , they should be orbited at asteroidal distance primarily by super - earths .
m dwarfs that form in region of low mass star will have gas giant planets at those distances , if disk instability is operative ( boss 2006 ) .
it remains for these theoretical speculations to be tested by completing the extrasolar planetary census .
this material is based on work supported by the national aeronautics and space administration under grants nng04g141 g , nng05g107 g issued through the terrestrial planet finder foundation science program and grant nnx07al70 g issued through the origins of solar systems program . the hobby - eberly telescope ( het ) is a joint project of the university of texas at austin , the pennsylvania state university , stanford university , ludwig - maximilians - universitt mnchen , and georg - august - universitt gttingen .
the het is named in honor of its principal benefactors , william p. hobby and robert e. eberly .
we would like to thank the mcdonald observatory tac for generous allocation of observing time .
cochran , w. d. , endl , m. , mcarthur , b. , paulson , d. b. , smith , v. v. , macqueen , p. j. , tull , r. g. , good , j. , booth , j. , shetrone , m. , roman , b. , odewahn , s. , deglman , f. , graver , m. , soukup , m. , & villarreal , m. l. 2004 , , 611 , l133 d ( the data are plotted twice for a second cycle ) .
the semi - amplitude @xmath51 of @xmath52 corresponds to a minimum mass of @xmath53 for the companion .
all parameters of the orbital solution are summarized in table [ tab : planet2 ] .
the residual scatter around the fit is @xmath54 ( circular ) and @xmath55 ( eccentric orbit ) .
[ phase ] ] rrrrrr 52927.82553 & -19.88 & 4.88 & 53669.99475 & 7.67 & 4.86 + 52935.80776 & -3.78 & 3.83 & 53682.74654 & -0.38 & 5.20 + 52939.79789 & -4.69 & 4.44 & 53687.75180 & 5.35 & 4.84 + 52941.98273 & 2.78 & 4.49 & 53718.67107 & -8.00 & 4.76 + 53254.93830 & -1.95 & 3.99 & 53719.66762 & -0.47 & 4.98 + 53297.80620 & -17.60 & 4.88 & 53721.64443 & 16.26 & 4.54 + 53302.79940 & 4.63 & 4.84 & 53730.82258 & 18.79 & 5.29 + 53310.78943 & 4.09 & 5.37 & 54049.74686 & 12.15 & 4.51 + 53313.97346 & 2.26 & 4.69 & 54136.72087 & -8.74 & 4.74 + 53315.77643 & 2.73 & 6.56 & 54328.97056 & -3.20 & 4.27 + 53330.71871 & 4.76 & 4.30 & 54330.96817 & -16.47 & 3.95 + 53330.72892 & 10.60 & 4.89 & 54338.95808 &
-2.98 & 4.16 + 53350.66330 & 2.13 & 5.19 & 54342.93427 & -0.15 & 3.62 + 53603.95259 & -19.79 & 5.67 & 54347.92539 & 1.41 & 3.70 + lr@@xmath56lr@@xmath56l p & 10.2369 & 0.0039 days & 10.2366 & 0.0038 days + t & 2454550.6672 & 0.39 bjd & 2455037.7979 & 1.1 bjd + k & 11.62 & 1.61 ms@xmath57 & 11.72 & 1.62 ms@xmath57 + e & 0.0 & 0.0 ( fixed ) & 0.232 & 0.127 + @xmath58 & 0.0 & 0.0 degrees ( fixed ) & 210.4 & 32.5 degrees + @xmath59 & 0.0771 & 0.0122 @xmath60 & 0.0757 & 0.0096 @xmath60 + @xmath59 & 24.5 & 3.9 @xmath61 & 24.1 & 3.1 @xmath61 + a & 0.0727 & 0.0007 au & 0.0727 & 0.0007 au + rms & & +
|
we report the detection of a planetary companion with a minimum mass of @xmath0 to the nearby ( @xmath1 pc ) m2.5v star gj 176 .
the star was observed as part of our m dwarf planet search at the hobby - eberly telescope ( het ) .
the detection is based on 5 years of high - precision differential radial velocity ( rv ) measurements using the high - resolution - spectrograph ( hrs ) .
the orbital period of the planet is @xmath2 d. gj 176 thus joins the small ( but increasing ) sample of m dwarfs hosting short - periodic planets with minimum masses in the neptune - mass range .
low mass planets could be relatively common around m dwarfs and the current detections might represent the tip of a rocky planet population .
|
the radio source 0108 + 388 ( ra=@xmath0 , dec=@xmath1 ( j2000 ) , @xmath2=0.699 ) is an example of a class of extragalactic radio source referred to as gigahertz - peaked spectrum ( or gps ) sources .
gps sources , which are discussed at length in a thorough review by odea ( 1998 ) , are characterized by large radio luminosities , compact structure , and spectra that turn over at gigahertz frequencies with steep spectra on either side of the peak ( odea , baum , & stanghellini 1991 ; odea 1998 ) .
the steep slope at high frequencies is generally accepted to be due to synchrotron losses , but there are two main competing theories for the cause of the low frequency turnover .
either the lower frequency radio emission is absorbed by ionized gas in the line of sight in the host galaxies , with conditions comparable to that of emission - line regions in active galaxies ( van breugel , heckman , & miley 1984 , bicknell , dopita , & odea 1997 ; kameno et al .
2000 ) , or the nuclei of these galaxies have exceptionally large magnetic fields , causing synchrotron self - absorption at ghz frequencies ( hodges , mutel , & phillips 1984 ; mutel , hodges , & phillips 1985 ; odea et al .
1991 ; readhead et al.1996 ; de vries , barthel , & odea 1997 ; snellen et al .
absorption by induced compton scattering in combination with free - free absorption has also been proposed by kuncic , bicknell , & dopita ( 1998 ) . as gps sources
have been studied in more detail since the discovery of a large population of such sources in the 1980 s ( gopal - krishna , patnaik , & steppe 1983 , spoelstra , patnaik , & gopal - krishna 1985 ) , it has become apparent that they constitute a diverse group , with morphological and spectral variances .
the object we discuss here , 0108 + 388 , is a canonical case of a morphological group referred to by wilkinson et al .
( 1994 ) as compact symmetric objects ( or cso s ) , which have small ( @xmath3 1 kpc ) symmetric radio emission structure .
multi - epoch vlbi observations of a number of cso s have revealed expansion rates that imply ages of the order of 1000 years ( owsianik , conway , & polatidis 1998 ; owsianik & conway 1998 ; taylor et al.2000 ) , supporting the earlier hypothesis that cso s are precursors to the young classical double radio galaxies ( phillips & mutel 1982 ; carvalho 1985 ; readhead et al .
vlbi observations of 0108 + 388 by owsianik et al .
( 1998 ) and by taylor et al .
( 2000 ) infer apparent separation speeds of the outer components of 0.197 @xmath4 0.026 @xmath5 c and 0.24 @xmath4 0.04 @xmath5 c , respectively , from which these authors derive kinematic ages of just 367 @xmath4 48 yr and 310 @xmath4 70 yr , respectively .
this is the youngest age yet measured for a radio galaxy .
however , with vla and wsrt observations of 0108 + 388 baum et al .
( 1990 ) detected extended emission 20 arcsec ( 78 @xmath5 kpc ) from the core , casting an apparent contradiction to the young classical double radio galaxy model .
baum et al .
propose that either the activity in 0108 + 388 is recurrent or has recently been smothered by infalling gas .
high resolution imaging of 0108 + 388 was first undertaken at 5 ghz as part of the pr vlbi survey ( pearson & readhead 1988 ) and found to have a simple double structure .
conway et al . ( 1994 ) and taylor , readhead , & pearson ( 1996 ) made more sensitive and higher resolution images at higher frequencies that revealed a steeply inverted core component connected to both outer components by a faint chain of components .
overall , 0108 + 388 has an ` s ' symmetry that is relatively common among csos ( readhead et al .
1996 ) .
the spectrum of the total radio emission from the compact structure in 0108 + 388 was analyzed by baum et al .
they measured a spectral index below the turnover that was steep and roughly constant , in conflict with the exponential turnover expected from free - free absorption .
however , their observed spectrum is the total from several components in the source .
an exponential turnover in the free - free absorption case will apply only if the absorbing medium uniformly covers the entire radio source .
we report here on a set of vlba observations designed to study the absorption at each point in the map .
we have observed 0108 + 388 with the vlba at five frequencies : two above the turnover , two below the turnover , and one at the turnover frequency itself .
the maps made with these observations reveal the morphology of the radio emission from 0108 + 388 and how it is obscured by absorption .
we find that the spectra of 0108 + 388 at all positions in our maps are consistent with free - free absorption .
additionally , our maps reveal that the absorption is non - uniform and is suggestive of an edge - on disk centered on the core .
other evidence for a large amount of thermal gas along the line of sight to the core of 0108 + 388 has been reported by carilli et al .
( 1998 ) , who measured significant absorption by hi , and by cawthorne et al .
( 1993 ) , who find no measurable polarization of the radio emission .
the existence of a circumnuclear gas disk , with likely instabilities , could also be the source of intermittent fueling of the central engine and explain the recent onset of activity along with the past activity .
we observed 0108 + 388 with the very long baseline array ( vlba ) at 8.417 , 4.983 , 2.267 , and 1.663 ghz on 1996 june 24 and at 15.359 and 8.415 ghz on 1997 february 03 . in the 1996 observations ,
the source signal was recorded at each station for a total of 100 to 112.5 minutes at each frequency , with an integration time of 2 seconds . in the 1997 observations a total of 330 minutes on source was obtained at both 8.415 and 15.359 ghz .
the on - source scans at all frequencies and both epochs were spread across a wide range in hour angle in order to obtain good @xmath6-@xmath7 coverage .
fringe fitting was performed with aips using the small and bright source j0136 + 478 as a fringe calibrator . in the fringe searches we generally used solution intervals of 3 minutes , moderately large delay and rate windows ( 100 ns and 40 mhz ) , and a minimum signal - to - noise ratio of 7 .
good solutions were generally found at small and stable delays and rates .
we applied _ a priori _ amplitude calibration , also with aips , using the system temperatures and gains measured by the nrao staff .
we assume that the amplitude calibration is accurate to within 5% .
the detected flux densities on the shortest baselines were all greater than 92% of previously reported flux densities from low resolution observations ( baum et al .
1990 ; odea et al . 1990 ; umrao database ) .
editing , modelfitting , and mapping were done using difmap ( shepherd , pearson , & taylor 1994 , 1995 ; shepherd 1997 ) . starting with the 1997 15-ghz data , we used an initial model of seven elliptical gaussian components , based on the appearance of the clean map made by taylor et al .
the model parameters were then allowed to vary to yield a best rms - fit in the @xmath6-@xmath7 plane to the self - calibrated data .
we found best - fit models for each data set independently except that we checked for general consistency in the positions of components at all frequencies .
the final model for each data set fit the self - calibrated data with an rms difference of around @xmath8 .
we fed the best - fit models as input models for hybrid clean ( readhead & wilkinson 1978 ) , which we used to obtain our final image maps .
we initially used small clean boxes that barely encompassed the apparent emission in the dirty map .
after 100 clean components were selected , with a gain of just 0.05 , we increased the sizes of the clean boxes while always staying marginally larger than the area of the apparent flux in the residual map .
we excluded apparent spurious features from the clean boxes the first few cycles of self - cal and clean .
cleaning was generally stopped a few clean components before the first negative component .
the maps were of sufficiently high quality that all the detected flux on the shortest baselines was collected in the cleaning by this point .
we did not apply amplitude self - calibration until the model agreed with the data with an rms difference less than @xmath9 the final clean component models fit the data to better than @xmath10 the noise in the maps range from 0.2 mjy / beam in the 8.4-ghz maps to 0.7 mjy / beam in the low frequency maps , yielding signal - to - noise ratios from 300 to over 1000 .
we also made spectral index maps between neighboring frequency bands using the aips task comb .
we first edited the data at each frequency , removing the larger @xmath6-@xmath7 spacings from the higher frequency data and the smaller @xmath6-@xmath7 spacings from the lower frequency data , and remapped in difmap .
the maps at the two frequencies were convolved with identical clean beams . in this way , we produced spectral index maps between 1.7 and 2.3 ghz , between 2.3 and 5.0 ghz , between 5.0 and 8.4 ghz from the 1996 june observations , and between 8.4 and 15.4 ghz from the 1997 february observations . since moderate percentage errors in the observed maps
propagate to large errors in the spectral index , we masked the spectral index maps where the signal in either clean map was less than 10@xmath11 .
since the observations were not phase - referenced to a nearby calibrator , no absolute position information is available for each map .
each map , therefore , can be arbitrarily shifted in ra and dec , and so the correct alignment of the maps is not known _
a priori_. this is a critical issue in making spectral index maps .
if the maps are misaligned , extremely large and extremely small spectral indices will be inferred , particularly at the edges of the emission regions . at the higher frequencies ( from 5.0 to 15.4 ghz ) the relative positions of the component peaks were consistent such that they could be used to determine the alignments of the maps .
the resultant spectral index maps were quite reasonable in that they did not show surprisingly large spectral indices at the edges or sides of the features . at the lower frequencies , however , the separations of the centroids of the components varied . for these spectral index maps ,
we chose alignments that yielded the minimum magnitudes of the extreme spectral indices . additionally , since the flux - density distribution in the maps appears relatively symmetric , especially in the lower resolution maps , we also checked the alignment of the maps by examining the spectral index map and watching for dipolar distributions of inferred spectral indices .
we successfully found relative alignments that minimized the spectral index extremes and that yielded roughly symmetrical spectral indices .
the clean maps for each observation are shown in figure 1 . as noted by previous investigators of this source ( pearson & readhead 1988 ; conway et al .
1994 ; taylor et al .
1996 ; owsianik & conway 1998 ; owsianik et al .
1998 ) the intensity distribution of the source at all frequencies consistently involves two major components . following the notation of taylor et al .
( 1996 ) , we refer to these components as c1 ( in the southwest ) and c7 ( in the northeast ) . at the higher frequencies ( 15.4 and 8.4 ghz ) we obtained models with a total of seven components , while at the three lower frequencies we could reasonably fit only four components , due to blending of the smaller components .
the spectral index maps are shown in figure 2 .
the uncertainties in the inferred spectral indices , calculated by propagating both uncertainties in amplitude calibration and the map noise , are listed in table 1 .
the spectral index map between the two highest frequencies shows that the two major components have steeply declining spectra , with spectral indices @xmath12 , where @xmath13 . a small region between the two major components , centered on component c3 , in the notation of taylor et al .
( 1996 ) , has an inverted spectrum with @xmath14 getting as large as + 1.8 , suggesting that the core of this active galaxy is located here , as previously noted by taylor et al .
( 1996 ) . between 8.4 and 5.0 ghz ,
the spectra of the regions containing the major components are generally flatter than between 15.4 and 8.4 ghz , although the spectra get steeper towards the outer edges . at 5.0 ghz
the core region has small flux densities , so most of this central region is masked in the spectral index map . between 5.0 and 2.3 ghz
, the spectral index map shows that the spectra of all but the easternmost and westernmost edges are inverted ( @xmath15 ) . at the very edges of the eastern and western extremes the spectral indices are still seen to get as low as @xmath160.7 .
the region with inverted spectra appears to run across the map in the north - south direction , but with the steepest parts located in two areas north and south of the core at the edge of the mapped region .
the core is not detected at these frequencies due to its low luminosities and steeply inverted spectrum . between 2.3 and 1.7 ghz
, the spectral index map looks very similar to the 2.3 - 5.0 ghz spectral index map , except that all the spectral indices across the entire map are larger and the region of inverted spectra encompasses almost the entire source .
the two areas of the steepest inverted spectra are more evident and appear even more symmetrical about the core . as clearly shown in figure 2 the turnover of the composite spectrum of 0108 + 388 at 5.0 ghz results from the majority of the source , including the two brightest components of the source , turning over at this frequency .
this is in conflict with the expectations for synchrotron self - absorption , in which the small regions with the largest magnetic fields and electron densities would turn over at higher frequencies .
the nature of the turnover in 0108 + 388 , however , is consistent with the absorption being due to free - free absorption by a foreground screen of ionized gas .
additionally , the peak spectral index between 2.3 and 1.7 ghz , even at the moderately low resolution of the map in figure 2d , is 3.2 @xmath4 0.5 .
this is larger than the maximum spectral index allowed by classical self - absorption models , although it can be fit to synchrotron self - absorption with special physical conditions ( see , e.g. , de kool & begelman 1989 ) .
the spectral index maps also suggest the shape and structure of the absorbing gas , which appears to be extended north - south with two density peaks on either side of the center .
this is suggestive of an edge - on disk centered on the core of the galaxy .
however , the spectral index only measures the ratio of the flux at two frequencies and does not directly indicate the densities of the absorbing gas . to directly map the absorbing gas we made opacity maps , which we discuss in the following section .
the creation of opacity maps involves comparison of maps at all the observed frequencies ( as described below ) . for this purpose , all data sets were re - mapped with the same resolution .
two methods were used .
first , we simply convolved all clean maps with the same beam as the clean beam at 1.7 ghz .
this approach is flawed , in general , because the different frequency observations ( with the same array ) probe structures of different sizes , and so even with the same clean beam different features are being compared . as a second approach we edited all data sets to have the same minimum and maximum @xmath6-@xmath7 distances ;
all data sets were flagged to have the same maximum @xmath6-@xmath7 distance as in the 1.7-ghz data set and the same minimum @xmath6-@xmath7 distance as in the 15.4-ghz data .
these truncated data sets had aspect ratios of only 9.0 and so this method could result in the maps missing much of the structure , both on large and on small scales .
for example , the detected flux densities on the remaining shortest baselines at 2.7 and 1.3 ghz were 87% and 78% , respectively , of those in the full data sets .
interestingly , there were no significant differences in the final maps made with these two methods and the maps looked very much like the 2.3 and 1.7 ghz clean maps in figure 1 .
furthermore the signal to noise ratios in the final maps made from the edited data sets were , in general , higher than in the clean maps from the untruncated data sets , especially at the mid - frequencies ; this resulted primarily because the editing removed the data at the extremes in the @xmath6-@xmath7 plane , which corresponds to where the density of data is lowest .
the high fidelity of these maps resulted from the simplicity of the structure of this source at these low frequencies .
the clean maps made from the truncated data sets are shown in figure 3 . to infer optical depths
we need flux - density maps in the absence of the absorption .
we obtain such maps by extrapolating the spectra at the higher frequencies .
since the spectra of the major components between 15.4 and 8.4 ghz are steeply declining with increasing frequency , the absorption at these frequencies is small and so the flux densities at these frequencies provide a close approximation of the unabsorbed emission spectrum . if @xmath17 is the spectral index between 15.4 and 8.4 ghz at some position @xmath18 in the map , and @xmath19 is the flux density at 8.4 ghz at this position , then the approximate unabsorbed flux density at 2.3 ghz is @xmath20 ^{\alpha_{15 - 8}(x , y)}.\eqno(1)\ ] ] we then obtain an approximation of the optical depth at 2.3 ghz at each point in the map by @xmath21 however , since the optical depths at 8.4 and 15.4 ghz are not really zero , the extrapolation of the spectrum produces a small error .
we improve the inferred optical depths by iterating the procedure ; we use the inferred optical depths at 2.3 ghz to infer the optical depths at 8.4 and 15.4 ghz , assuming free - free absorption .
we then correct the 8.4 and 15.4 ghz maps for the non - zero optical depths , yielding maps of the unabsorbed flux - densities at these frequencies .
we then used these modified maps to recreate the 15.4 - 8.4 ghz spectral index map and the unabsorbed 1.7 and 2.3 ghz flux density maps , which we used , in turn , to create improved opacity maps .
the opacity maps were made from the maps shown in figure 3 by using a number of steps of the aips task comb which reproduced the math discussed in the previous paragraph . the low - frequency flux densities ( at 5.0 , 2.3 , and 1.7 ghz )
were extrapolated from the 8.4-ghz flux densities of the same epoch ( 1996 june ) while the spectral indices between 15.4 ghz and 8.4 ghz ( @xmath17 ) were calculated using the 8.4-ghz data of the same epoch ( 1997 february ) as the 15.4-ghz data .
the resultant opacity maps at 2.3 and 1.7 ghz are shown in figure 4 .
the uncertainties in the optical depths are primarily due to the 5% uncertainty in amplitude calibration , which yields a constant uncertainty across each map .
the calculated uncertainties due to amplitude calibration errors are listed in table 2 .
the opacity maps at 5.0 , 2.3 , and 1.7 ghz are nearly identical , except that the optical depths are larger at lower frequencies , as expected .
the optical depths at 5.0 ghz are small , relative to the uncertainty , and so the 5.0-ghz opacity map is dominated by noise and so is not displayed in figure 4 .
the opacity maps are nearly identical in appearance .
these maps also indeed show that the maximum absorption occurs at two points north and south of the center and that the optical depth decreases in the direction perpendicular to the line connecting the optical depth peaks . at each point in the map
, we have a measure of the optical depth at three different frequencies , and so we can check whether free - free absorption fits the observed spectra everywhere .
the optical depth of free - free absorption , in a simple single - component model , depends on frequency according to @xmath22 where @xmath23 ghz and is given in ghz , @xmath24 is in units of @xmath25 , and @xmath26 is in pc ( mezger & henderson 1967 ) . assuming this model ,
we can remove the frequency dependence in our opacity maps , and check for consistency between the data at different frequencies , by multiplying the optical depths in the maps of figure 4 by @xmath27 ( using the aips task comb ) .
if each position in these maps has a single free - free absorption component , then the resultant images will display maps of the value @xmath28 .
these maps are shown in figure 5 ( again , because of the large uncertainty relative to calculated values we do not display the map at 5.0 ghz ) .
the uncertainties in these values , calculated by propagating through just the amplitude calibration errors , are listed in table 2 ( the total uncertainties actually include errors due to map noise , which are significant only near the edges of the source ) .
the maps in figure 5 are essentially identical . in figure 6
we display the difference between the two maps in figure 5 divided by the total uncertainty , due to both amplitude calibration errors and map noise .
the contours displayed in figure 6 represent difference values of @xmath160.5 , 0.5 , 1 , and 2 times the calculated uncertainty .
the differences are typically @xmath29 across the map , with a mean of 1.1@xmath11 , and get as large as 2.1@xmath11 only near the southwest edge .
the rms value in figure 6 , and hence the rms deviation between the maps in figure 5 , is 1.26@xmath11 .
as is evident in figures 4 and 5 , the optical depth is not uniform over the map and most likely varies on scales smaller than the resolution of these maps .
therefore , one should not expect the single - component free - free absorption model to perfectly fit the spectra at each point in our moderately low resolution maps .
therefore , a reduced @xmath30 value for the agreement between this simple model and our data across our entire map of just 1.26 is striking and strongly supports the premise that free - free absorption is at work . only a mild level of complexity , such as a variable electron density across our synthesized beam , would be needed to produce a fit with an rms deviation of less than 1@xmath11 . as an additional test
, one can fit a free - free absorption model to individual spectra . by stacking the maps in figure 3 , with the same alignments that we used to make the opacity maps ,
we can obtain individual spectra at each location .
however , as we discussed in the previous section , one should expect that the optical depth varies on scales below the low resolution of the maps in figure 3 .
one should not , therefore , expect the precise fitting of a single - component free - free absorption model to the observed spectra to produce a perfect fit . in figure 7
, we display the spectra at the locations of the peaks in emission at 5.0 ghz . to correct for any flux variations in the source between our two observation epochs the flux densities at 15.4 ghz in our 1997 observation
were multiplied by a scale factor of 0.97 , which is the ratio of the 8.4-ghz flux densities in 1996 to that in 1997 .
the displayed error bars correspond to the total uncertainty in the measured flux densities , due to both the 5% amplitude calibration uncertainty and the rms in the map flux densities .
overlaid on the observed spectra in figure 7 are best - fit curves for single - component models of free - free absorption in the foreground of optically thin synchrotron emission sources .
the curves in figure 4 are given by @xmath31 .
since our goal is to determine the cause of the absorption at low frequencies , we kept @xmath32 relatively fixed to ensure that the curve passes through the measured 15.4-ghz flux density .
variation in @xmath33 , though , causes slight changes in the model 15.4-ghz flux density , so we needed to allow @xmath32 to vary a little , as @xmath33 varied , to ensure that the flux - density at 15.4 ghz matched the observed value .
we , therefore , had two free parameters .
the values of these parameters in the best fit curves are listed in table 3 .
although the curves appear by visual inspection to fit the data well , the reduced @xmath30 of the fits , listed in table 3 , are quite large . almost the entire contribution to these large @xmath30 values are in the disagreements at 2.3 and 1.7 ghz , where the slopes of the curves are very steep and small adjustments in the model parameters yield large differences in the curve values . however , a slightly more complex model , such as an electron density that varies by a factor of two on scales below the map resolution , yields nearly perfect fits to the data .
we , therefore , find that the observed spectra at these two positions are consistent with a reasonable free - free absorption model .
the maps in figure 5 , which show the morphology of the quantity @xmath28 projected onto the sky plane , show that the absorbing medium is elongated north - south and is concentrated in two spots symmetrical about the core . in figure 8
we display the 1.7-ghz map from figure 5 in contours overlaid on a grey - scale plot of the 15.4-ghz clean map from figure 3 .
even with the large clean beam of these maps , the denser regions of the absorbing medium are clearly shown to be located to the sides of the radio emission structure .
such a morphology in optical depth is often symptomatic of a disk centered on the core .
if this is indeed a disk , it is roughly perpendicular to the jet axis .
one must keep in mind that this medium is probed in our maps only through its absorption of the radio emission from the jet , and so the medium can , and most likely does , extend well beyond the regions indicated in our maps .
additionally , the absorption is caused only by ionized gas and any such disk is most likely to have a large neutral component that is far more extensive .
in fact , the observations of carilli et al .
( 1998 ) indicate that a large amount of neutral gas does in fact exist in this galaxy . in interpreting the morphology of the opacity maps
, one must keep in mind that these maps were made with a resolution of 4.4 mas in the north - south direction and 3 mas in the east - west direction .
for example , the peaks in optical depth in figure 4 appear to be located where there is no detectable emission at 15.4 ghz in the high resolution map of figure 1a .
however , since in the process with which we made the opacity maps the input maps were masked where the signal was below 10@xmath11 , this does nt seem possible .
the reason that we appear to detect absorption far from the radio emission structure seen in figure 1a is because of the low resolution of the opacity maps ; the 15.4-ghz map of figure 3a was used in making the opacity maps , not figure 1a .
if we were able to make opacity maps at the resolution of figure 1a , we would find the optical depth peaks to occur somewhere at the edges of the 15.4-ghz emission . in the opacity maps of figure 4 ,
the precision of the position information is given by the width of the clean beam , and so a true location of the optical depth peaks 1 mas away from the apparent location is reasonable .
the morphological information in our opacity maps of which we can be sure is that indicated in figure 8 , i.e. that the peaks in the optical depth are offset from the peaks in the radio emission . in order for the optical depth peaks to appear at the edges of our opacity maps ,
the true optical depth structure must have a similar gradient .
the peaks in optical depth are located about 10 @xmath4 4.4 mas apart . therefore , if this is indicative of a disk , then the disk appears to have an inner radius of about 5 @xmath4 2.2 mas . with a redshift of 0.699 and a hubble constant of @xmath34 km s@xmath35 mpc@xmath35
, this corresponds to an inner radius of @xmath36 50 @xmath4 20 @xmath5 pc . if we assume an axially symmetric disk of gas with a uniform structure of temperature @xmath37 , free electron density @xmath38 , and radius @xmath39 , with a hole in the middle of radius @xmath40 .
the optical depth at each point is @xmath41 where @xmath26 is the path length through the ionized part of the disk to the radio emission in pc .
the values in figure 5 , then , correspond to @xmath42 which is roughly the emission measure ( @xmath43 ) times @xmath44 times a constant .
for convenience of discussion we refer to the quantity @xmath45 $ ] as @xmath46 .
when we average the maps in figure 5 together we find peak values of @xmath46 of 10.6 to the north and 10.1 to the south .
reasonable values of @xmath47 and @xmath24 for the inner ionized region of a gas disk of order 100 pc in radius in a radio galaxy nucleus are @xmath48 k and 500 @xmath25 , respectively , which yield a path length of @xmath49 pc . we therefore find that entirely reasonable physical parameters fit the observed free - free optical depths . comparing our results to the hi absorption results of carilli et al .
( 1998 ) , we have @xmath50 where @xmath51 is the covering factor of the neutral gas and @xmath52 is the average spin temperature of the hydrogen atoms along the line of sight .
these results are reasonably consistent , as can be demonstrated by the application of saha equation with some simplifying assumptions . for argument s sake
, we assume @xmath53 , @xmath51 = 1 , and @xmath54 ( i.e. that the inner 1% of the disk is ionized ) . then , @xmath55 if we further approximate the gas as containing only hydrogen , then the saha equation can be written as @xmath56 equating these two expressions yields an average gas temperature of @xmath57 k. the spectrum of the core region is already inverted at the observed frequencies above 5 ghz .
the cause for the inverted spectrum of the core , therefore , seems unrelated to the ionized disk of gas . since the core is the region most likely to have large magnetic fields and large densities of relativistic energies
, we suspect that the inverted spectrum of the core is due to synchrotron self - absorption . with the same 15.4-ghz data reported here ,
taylor et al . ( 2000 ) infer an upper limit to the magnetic field in the core , assuming synchrotron self - absorption , of @xmath58 gauss .
the suggestion of a dense circumnuclear disk of gas in 0108 + 388 may provide further insight into some of the unique characteristics of this cso .
we speculate that instabilities in such a disk could result in periodic infall of gas , which will produce renewed apparent activity .
this would simultaneously explain the extended emission 20 arcsec ( 78 @xmath5 kpc ) to the east detected by baum et al .
( 1990 ) and the young age of the compact structure inferred by owsianik et al .
( 1998 ) and taylor et al .
( 2000 ) .
as the results presented here demonstrate , when a radio source s flux density is produced by a number of components , absorption on large scales is needed in order for the composite spectrum to have a sharp turnover .
synchrotron self - absorption depends on the internal conditions of the source , and these conditions will vary from one part of the source to another .
such a large - scale absorption is therefore unlikely to be due to this process .
on the other hand , we have also found that the core of this particular gps source is probably synchrotron self - absorbed .
its spectrum is already turned over at the frequency where the free - free absorption becomes significant .
the implication for other , and possibly all , gps sources is that free - free absorption is the likely cause for the high - frequency turnovers of the composite spectrum , but synchrotron self - absorption may still be important on smaller scales ; the turnover of a weak core component will not contribute significantly to the integrated flux densities at low frequencies .
by extrapolating our high frequency maps , we obtained opacity maps of 0108 + 388 at two frequencies below the spectral turnover which strongly suggest that free - free absorption , possibly by a disk of ionized gas , is indeed the cause of the spectral turnover , and perhaps in other csos with a ghz - peaked spectrum .
the spectra all across the map fit free - free absorption well .
reasonable values for temperature ( 10@xmath59 k ) and electron density ( 500 @xmath25 ) in a disk of order 100 pc in size fit the observed optical depths .
the presence of a dense gas disk in 0108 + 388 , which is also consistent with the large column densities implied by the hi absorption measurements of carilli et al .
( 1998 ) , may be related to the recent onset of activity in this object , as indicated by the age estimates by owsianik et al .
( 1998 ) and taylor et al .
the accumulated evidence thus strongly supports the conclusions that csos in general are intrinsically young objects that evolve into classical kpc - scale radio galaxies , and that the radio galaxy 0108 + 388 , in particular , is the youngest source known .
this research was aided , in part , by funding from the dudley observatory .
this research has also made use of data from the university of michigan radio astronomy observatory which is supported by funds from the university of michigan .
we are also grateful for the friendly assistance of vivek dhawan and the staff at the national radio astronomy observatory .
baum , s. a. , odea , c. p. , murphy , d. w. , & de bruyn , a. g. 1990 , a&a , 232 , 19 bicknell , g. v. , dopita , m. a. , & odea , c. p. 1997 , apj , 485 , 112 carvalho , j. c. 1985 , mnras , 215 , 463 cawthorne , t. v. , wardle , j. f. c. , roberts , d. h. , gabuzda , d. c. , & brown , l. f. 1993 , apj , 416 , 496 carilli , c. l. , menten , k. m. , reid , m. j. , rupen , m. p. , & yun , m. s. 1998 , apj , 494 , 175 conway , j. e. , myers , s. t. , pearson , t. j. , readhead , a. c. s. , unwin , s. c. , & xu , w. 1994 , apj , 425 , 568 de kool , m. & begelman , m. c. 1989 , nature , 338 , 484 de vries , w. h. barthel , p. d. , & odea , c. p. 1997 , a&a , 321 , 105 gopal - krishna , patnaik , a. r. , & steppe , h. 1983 , a&a , 123 , 107 hodges , m. w. , mutel , r. l. , & phillips , r. b. 1984 , aj , 89 , 1327 kameno , s. , horiuchi , s. , shen , z .- q . , inoue , m. , kobayashi , h. , hirabayashi , h. , & murata , y. 2000 , pasj , 52 , 209 kuncic , z. , bicknell , g. v. , & dopita , m. a. 1998 , apj , 495 , l35 mezger , p. g. & henderson , a. p. , 1967 ,
apj , 143 , 471 mutel , r. l. , hodges , m. w. , & phillips , r. b. 1985 , apj , 290 , 86 odea , c. p. 1998 , pasp , 110 , 493 odea , c. p. , baum , s. a. , & stanghellini , c. 1991 , apj , 380 , 66 owsianik , i. & conway , j. e. 1998 , a&a , 337 , 69 owsianik , i. , conway , j. e. , & polatidis , a. g. 1998 , a&a , 336 , l37 pearson , t. j. & readhead , a. c. s. 1988 , apj , 328 , 114 phillips , r. b. & mutel , r. l. 1982 , a&a , 106 , 21 readhead , a. c. s. & wilkinson , p. n. 1978 , apj , 223 , 25 readhead , a. c. s. , taylor , g. b. , pearson , t. j. , wilkinson , p. n. 1996 , apj , 460 , 634 shepherd , m. c. , pearson , t. j. , & taylor , g. b. 1995 , baas , 26 , 987 shepherd , m. c. , pearson , t. j. , & taylor , g. b. 1995 , baas , 27 , 903 shepherd , m. c. 1997 , in a.s.p .
125 , astronomical data analysis software and systems vi , ed .
g. hunt & h. e. payne , ( san francisco : asp ) 77 snellen , i. , tschager , w. , schilizzi , r. , rttgering , h. , & miley , g. 2000 in astrophysical phenomena revealed by space vlbi , ed .
h. hirabayashi , p. g. edwards , & d. w. murphy ( tokyo : isas ) , 79 spoelstra , t. a. t. , patnaik , a. r. , & gopal - krishna 1985 , a&a , 152 , 38 taylor , g. b. , readhead , a. c. s. , & pearson , t. j. 1996 , apj , 463 , 95 taylor , g. b. , marr , j. m. , readhead a. c. s. , & pearson , t. j. 2000 , apj , 541 , 112 van breugel , w. , heckman , t. , & miley , g. 1984 , apj , 276 , 79 wilkinson , p. n. , polatidis , a. g. , readhead , a. c. s. , xu , w. , & pearson , t. j. 1994 , apj , 432 , l87 .uncertainties in spectral index maps due to uncertainties in the observed flux densities .
uncertainties in observed flux densities result both from possible errors in the amplitude calibration and from map noise .
the uncertainties in @xmath14 due to amplitude calibration errors , listed in column 2 , are constant across each map . the uncertainties in @xmath14 due to map noise varies and depends on the ratio of the observed flux density to the map rms .
since the spectral index maps were truncated where the map signals were @xmath60 , the maximum uncertainty due to map noise , which occurs at the edges of the source , are calculated and are listed in column 3 . [ cols="^,^,^",options="header " , ]
|
we have observed the canonical gigahertz - peaked spectrum source 0108 + 388 with the vlba at a range of frequencies above and below the spectral peak .
the activity that dominates the radio emission from 0108 + 388 , which is also classified as a compact symmetric object , is thought to be less than 1000 years old .
we present strong evidence that the spectral turnover in 0108 + 388 results from free - free absorption by non - uniform gas , possibly in the form of a disk in the central tens of parsecs .
|
recently there has been considerable interest in dimensionally reduced yang - mills theories as a means to obtain non - perturbative information on superstring theory and m - theory .
the possible relevance of these systems to quantum gravity appeared in @xcite through the light cone quantization of the 11-dimensional supermembrane .
it was argued in @xcite that 10-dimensional @xmath0 super yang - mills theory reduced to one dimension
i.e. matrix quantum mechanics correctly quantizes the supermembrane in the large @xmath2 limit .
the very same system , first studied ( without reference to applications to quantum gravity ) in @xcite , has recently been interpreted as a non - perturbative attempt at m - theory @xcite .
unfortunately , on the technical side , very little is known about this model : it is suspected @xcite , @xcite that a novel , intricate large @xmath2 limit is required , but only few concrete results are available .
this motivates the study of a simpler system : the complete reduction of yang - mills theory to @xmath3 dimensions .
in addition , the reduction of the ten - dimensional susy gauge theory path integral to a matrix integral has been at the heart of an alternative proposal to directly define non - perturbative iib string theory @xcite .
more generally we may study the complete reduction of d - dimensional @xmath0 yang - mills theory .
then the path integral of the field theory simplifies to an integral over the group s lie algebra , with a flat measure : a _ yang - mills integral_. denoting the gauge potential by @xmath4 and their superpartners by @xmath5 we obtain @xmath6 . \label{susyint}\ ] ] with the euclidean `` action '' @xmath7 [ x_\mu , x_\nu ] - \tr \psi_{\alpha } [ \gamma_{\alpha \beta}^{\mu } x_{\mu},\psi_{\beta } ] .
\label{action}\ ] ] the _ a priori _ allowed dimensions for the reduced supersymmetric gauge theory are @xmath8 corresponding to @xmath9 real supersymmetries . for the bosonic , i.e. the non - supersymmetric case @xmath10 , we omit the grassmann variables @xmath5 , and may study all dimensions @xmath11 .
there are numerous further reasons for being interested in the integrals eq.([susyint ] ) : @xmath12 the susy integrals are crucial for the computation of the witten index of the above mentioned quantum mechanical gauge theories , as they contribute to the so - called bulk part of the index @xcite ( cf also @xcite for an earlier calculation ) .
@xmath12 in the maximally supersymmetric case , the system describes to leading order the statistical distribution of a system of @xmath2 d - instantons ( or `` @xmath13-branes '' ) .
@xmath12 the @xmath14 integral appears in very recent work developing a multi - instanton calculus for the @xmath15 , @xmath16 @xmath17 conformal gauge theory @xcite , and again in the large @xmath2 limit of sp@xmath18 and so(@xmath2 ) @xmath15 susy gauge theory @xcite .
@xmath12 finally we can regard the integrals eq.([susyint ] ) as a version of the eguchi - kawai reduced gauge theory . the original work @xcite focused on a lattice formulation and employed unitary matrices , while the above integrals use the hermitean gauge connections @xmath4
this is similar to @xcite ; however , we apply neither gauge fixing nor quenching prescriptions to the above integrals .
the interesting question is whether the models eq.([susyint ] ) encode universal information on the full gauge field theory as @xmath19 .
the integrals of eq.([susyint ] ) appear to be singular due to the `` valleys '' of the action , i.e. the directions in the configuration space of the @xmath4 where all @xmath20 matrices commute .
recent work has however proven this intuition wrong : yang - mills integrals _ do _ exist in many interesting cases .
indeed , the rigorous results of @xcite ( see also @xcite ) show that for the gauge group @xmath1 the susy integrals converge in dimensions @xmath21 .
the calculations are easily repeated for the bosonic case @xcite , and the convergence condition @xmath22 is found .
unfortunately , no rigorous methods exist to date for higher rank gauge groups @xmath23 . in @xcite , @xcite,@xcite we developed methods to numerically test convergence of singular multidimensional integrals .
the idea is to perform a metropolis random walk weighted by the integrand , and to merely measure the autocorrelation function of subsequent configurations . in this approach ,
a unit autocorrelation function signals the presence of a nonintegrable singularity . as an illustration we plot in fig.1 the autocorrelation function of the @xmath1 bosonic integrals .
( 200,180 ) ( -300,-120 ) ( 40,-120 ) ( 360,-120 ) fig .
1 _ autocorrelation functions versus monte carlo time for the @xmath1 bosonic integral with , from the left , @xmath24 . + _ we are clearly able to reproduce the convergence condition @xmath22 : for @xmath25 the configurations decorrelate well and the whole integration space is properly sampled .
( one observes increasingly improved decorrelation for @xmath26 , not shown in fig.1 . )
in contrast for @xmath27 the system gets quickly trapped in a singular configuration : the markov chain gets lost in a valley , and the integral is divergent .
@xmath16 shows marginal divergence , which agrees with the exact analytical results . applying the same method to higher rank bosonic models , and to the supersymmetric models , we are able to map out the convergence conditions for the yang - mills integrals .
they read @xmath28 in particular the @xmath27 susy integral is divergent ( see @xcite,@xcite,@xcite for a more detailed discussion of this point )
. it would be nice to have a rigorous mathematical proof of the conditions ( [ intcond ] ) .
some understanding may be gained by considering one - loop perturbative estimates of the integrals eq.([susyint ] ) .
one has for the supersymmetric case @xcite @xmath29~\times \nonumber\\ & \times~\sum_{g:\;{\rm maximal \ ; tree}}~ \prod_{(ij):\;{\rm link \ ; of } \ ; g}~{1 \over ( x^i - x^j)^{3 ( d-2 ) } } + \ldots \label{susyfluct}\end{aligned}\ ] ] where the @xmath30 are the diagonal components of the matrices @xmath4 , while it is found in @xcite that the bosonic integrals are approximated by @xmath31~ \prod_{i < j}~{1 \over ( x^i - x^j)^{2 ( d-2)}}. \label{fluct}\ ] ] powercounting for large separations @xmath32 yields precisely eq.([intcond ] )
. however , it should be stressed that this does not prove the convergence conditions , since one has to worry about configurations where some separations are small , so that the one - loop approximation becomes invalid . in the supersymmetric case the value for the integrals
is believed to be known : @xmath33 the @xmath14 expression was conjectured by green and gutperle @xcite based on a calculation of the d - instanton effective action of the superstring .
a derivation of the terms to the right of the curly bracket in eq.([index ] ) , based on cohomological deformation techniques , was given in @xcite .
this calculation has still an important loophole ( see comments in @xcite ) .
however , the formula eq.([index ] ) was numerically checked by monte carlo techniques in @xcite,@xcite up to @xmath34 . for the bosonic case no exact value of the partition function
is known except for @xmath1 @xcite , where the result reads @xmath35 it would be exciting to find the generalization of this result to higher rank gauge groups this is after all the `` zero - mode '' contribution to the yang - mills partition function on a d - dimensional torus .
let us shift attention from the partition functions to the correlation functions of the models .
the simplest correlators are @xmath0 invariant one - matrix correlators : the moments @xmath36 of one matrix , say the @xmath20-th : @xmath37 .
these are directly related to the distribution of eigenvalues of the matrix : if the eigenvalues of @xmath37 are @xmath38 , the eigenvalue density is defined for all @xmath2 as @xmath39 the non - zero moments of @xmath40 are then given by @xmath41 in an ordinary wigner type matrix model all moments exist .
yang - mills integrals are more intricate .
analytical calculations have only been performed for @xmath1 , and we found the following surprising results . in the @xmath16
susy integral all moments are infinite , even though the integral itself exists , as argued above . in the @xmath42 susy
integral the first two moments are finite and one finds @xmath43 while all higher moments diverge . for the @xmath44 susy integral we have exactly twelve finite moments which are
@xmath45 it would be interesting to find a geometrical or combinatorial interpretation for these numbers . which densities give rise to this convergence behavior ?
for @xmath1 we can go farther and find the exact densities : @xmath46 \nonumber\\ \rho_{d=10}^{\rm{susy}}(\lambda)&= & \frac{1287}{64\cdot 2^{3/4}\ , \sqrt{\pi } } \ , \lambda^2\ , \bigl [ 546\ , u(\frac{17}{4 } , \frac{1}{2 } , 8\lambda^4 ) \nonumber \\ & & \qquad \qquad \qquad -147\ , \frac{17\cdot 19}{8}\ , u(\frac{21}{4 } , \frac{1}{2 } , 8\lambda^4 ) \nonumber\\ & & \qquad \qquad \qquad + 45\ , \frac{17\cdot 19\cdot 21\cdot 23}{256}\ , u(\frac{25}{4 } , \frac{1}{2 } , 8\lambda^4 ) \nonumber \\ & & \qquad \qquad \qquad -\frac{17\cdot 19\cdot 21\cdot 23\cdot 25\cdot 27}{2048}\ , u(\frac{29}{4 } , \frac{1}{2 } , 8\lambda^4)\ , \bigl ] \nonumber\end{aligned}\ ] ] where @xmath47 is the kummer-@xmath47 function defined as @xmath48 now we see that the above finiteness properties of the moments result from a rather curious powerlike behavior of the densities at large values of @xmath49 .
we have for @xmath50 @xmath51 this power - like behavior is very different from wigner type systems where the fall - off at infinity is at least exponential . for the @xmath20-dimensional bosonic models the density
can be worked out as well albeit less explicitly than in eqs.([exactrho ] ) , and one finds the asymptotic behavior @xmath52 . moving on to higher values of @xmath2 , we are unable to analytically calculate the eigenvalue densities with presently known techniques .
we can , however , find numerically exact densities using monte carlo methods . in fig.2
we illustrate this by plotting the @xmath53 @xmath16 susy half - densities ( we only plot the @xmath54 part since the densities are symmetric functions ) . in the @xmath1 case the exact expression of eq.([exactrho ] ) and the monte carlo data can not be separated on the scale of the figure .
now we would like to know how the @xmath1 result eq.([asymp ] ) generalizes to other values of @xmath2 .
it is impossible to extract the asymptotics from histograms such as fig.2 , since the tails comprise only a small number of samples . instead
, we can go back to the markov - chain technique of the last section and measure the finiteness of the moments eq.([corr ] ) for various values of @xmath2 .
we find that for the susy integrals the qualitative behavior of the @xmath1 case persists : in the @xmath21 integrals only the first , respectively , @xmath55 moments are finite .
we thus conclude , in view of eq.([corr ] ) , that the asymptotic behavior _ eq.([asymp ] ) is valid for all @xmath2_. this is very different from wigner type random matrix models , where as @xmath2 increases , the density condenses onto a compact interval . at the sharply defined ( at @xmath56 )
edge of the interval wigner distributions show universal behavior .
we have argued that in susy yang - mills matrix models no such edge exists , indicating that _ the large @xmath2 physics of these models is indeed very different_. ( 200,180 ) ( -320,220 ) ( 20,220 ) ( 360,220 ) fig .
2 _ eigenvalue ( half-)densities for susy @xmath16 , from the left , @xmath53 + _ let us furthermore compare supersymmetric and non - supersymmetric yang - mills integrals .
how does the asymptotic behavior of the density eq.([asymp ] ) change in the absence of susy ?
we have already mentioned above that for @xmath1 this behavior is powerlike as well .
actually we can guess a general formula by looking once more at the effective one - loop estimate of eq.([fluct ] ) . for one - matrix correlators
the most dangerous configuration stems from pulling away one coordinate @xmath57 from a bulk configuration of all other @xmath58 coordinates .
powercounting leads to the guess @xmath59 the same procedure applied to the susy estimate eq.([susyfluct ] ) reproduces eq.([asymp ] ) .
we then verified the validity of eq.([bosasymp ] ) by the same monte carlo random walk procedure as above , measuring the finiteness of moments .
we thus notice a marked difference to the susy situation : in the bosonic case _ all moments exist _ as @xmath19 for all @xmath60 .
in particular we expect the eigenvalue distribution to condense onto a compact support , much like for wigner type models .
a further natural set of correlation functions of yang - mills matrix integrals are wilson loops . due to the eguchi - kawai mechanism @xcite,@xcite , one _ naively _ expects them to correspond at @xmath56 to wilson loops in the unreduced gauge field theory . in the proposal of @xcite for a non - perturbative definition of the iib superstring
, they have been interpreted as string creation operators @xcite . despite the dimensional reduction of the field theory to zero dimensions
we are still able to define an infinite set of independent wilson loops dependent on an arbitrary contour @xmath61 in @xmath20-dimensional euclidean space : @xmath62 due to the non - commutative nature of the connections @xmath63 and the path - ordering @xmath64 , this is a non - trivial functional of the contour @xmath61 despite the fact that the @xmath63 are spacially constant . in the special case of a rectangular contour with lengths @xmath65 and @xmath66 in the @xmath67 plane
this simplifies to @xmath68 we would like to understand how the loops @xmath69 behave as a function of @xmath2 and as a functional of the shape of the contours , in particular whether planar loops satisfy an area - law .
we would also like to see whether there are any telltale differences between the supersymmmetric and the bosonic loops . in the previous sections we have shown how a number of exact results may be derived for the simplest gauge group @xmath1 .
we have not been able to analytically calculate a wilson loop for an yang - mills integral even for @xmath1 . on the other hand it is possible to obtain high precision numerical results for low values of @xmath2 . in fig.3a
we plot the wilson loop for a square ( @xmath70 ) in the case of susy @xmath16 as a function of @xmath65 for various values of @xmath2 .
( 200,180 ) ( -200,-40 ) ( 200,-40 ) fig . 3 @xmath16 susy wilson loops _ ( a ) square of side length @xmath65 with @xmath71 ( top to bottom ) ( b ) square , regular 16-gon and 64-gon enclosed in circle of radius @xmath72 with @xmath73 . + _
the behavior for small area is easy to understand . indeed , for an arbitrary planar loop enclosing a small area of size @xmath74 , it is straightforward to show , using stokes theorem , that @xmath75 ^ 2 \rangle + { { \cal o}}(x^6)$ ] .
this immediately gives @xmath76 and agrees to high precision with the numerical data .
a very curious feature of fig.3a is that the loops tend to a constant for large area .
the existence of this constant can be demonstrated analytically .
it should be considered a finite @xmath2 artifact for the following reasons : ( a ) the constant decreases with @xmath2 , as seen in fig.3a .
( b ) the constant depends in various ways on the shape of the contour .
we checked that by distorting the rectangle to a slightly irregular quadrangle the constant drops to zero for all @xmath2 as the size increases .
in fig3.b we show various regular polygons approximating a circle of radius @xmath72 : as we increase the number of edges the constant goes to zero for large @xmath72 .
it is clear from the mentioned features of the wilson loops that there is an area law neither for very small nor very large areas .
an intermediate region in which an area law holds might still be present .
we checked , by going to rather larger @xmath2 , that this is not the case for the bosonic models .
there has been an interesting suggestion @xcite that such an intermediate region may exist for the @xmath16 supersymmetric model .
we showed how numerical markov chain methods can be used to verify _ non - perturbative _ convergence conditions for yang - mills integrals with and without supersymmetry .
the same methods may be applied to establish the convergence properties of correlation functions . applying the technique to invariant correlators of a single matrix , we are able to accurately predict the asymptotic behavior of the eigenvalue density of yang - mills matrix models .
the results demonstrate an unusual powerlaw behavior which , in the supersymmetric cases , persists for large @xmath2 .
this indicates that the large @xmath2 limit of these `` new '' matrix models might indeed be very different from the one of the `` old '' wigner type models .
we also demonstrated that monte carlo methods are capable to rather accurately compute various quantities relevant to these models such as partition functions , correlation functions , spectral distributions and wilson loops .
as opposed to yang - mills quantum mechanics @xcite,@xcite,@xcite we are confronted to a system which allows some non - perturbative analysis , at least for finite @xmath2 .
yang - mills integrals are thus an ideal laboratory for exploring new large @xmath2 techniques .
powerful analytical methods will have to be developed if we are to verify or , maybe more importantly , if we are to bring to good use the ideas presented in @xcite,@xcite,@xcite .
this work was supported in part by the eu under contract fmrx - ct96 - 0012 .
77 b. de wit , j. hoppe and h. nicolai , _ on the quantum mechanics of supermembranes _
b305 ( 1988 ) 545 . m. claudson and m. halpern , _ supersymmetric ground state wave functions _ , nucl .
b250 ( 1985 ) 689 ; m. baake , m. reinecke and v. rittenberg , _ fierz identities for real clifford algebras and the number of supercharges _ , j. math .
phys . 26 ( 1985 ) 1070 ; r. flume , _ on quantum mechanics with extended supersymmetry and non - abelian gauge constraints _ , annals phys . 164 ( 1985 ) 189 .
t. banks , w. fischler , s. shenker and l. susskind , _ m - theory as a matrix model : a conjecture _
d55 ( 1997 ) 5112 , hep - th/9610043 .
n. ishibashi , h. kawai , y. kitazawa and a. tsuchiya , _ a large n reduced model as superstring _
, nucl . phys .
b498 ( 1997 ) 467 , hep - th/9612115 .
yi , _ witten index and threshold bound states of d - branes _ , nucl . phys . b505 ( 1997 ) 307 , hep - th/9704098 ; s. sethi and m. stern , _
d - brane bound state redux _ , commun
( 1998 ) 675 , hep - th/9705046 .
smilga , _ witten index calculation in supersymmetric gauge theory _ , yad .
42 ( 1985 ) 728 , nucl .
b266 ( 1986 ) 45 ; a.v .
smilga , _ calculation of the witten index in extended supersymmetric yang - mills theory _ , ( in russian ) yad .
43 ( 1986 ) 215 . n. dorey , t.j
. hollowood , v.v .
khoze , m.p .
mattis , and s. vandoren , _ multi - instantons and maldacena s conjecture _ , jhep 9906 ( 1999 ) 023 , hep - th/9810243 ; _ multi - instanton calculus and the ads / cft correspondence in _
@xmath15 _ superconformal field theory _ ,
b552 ( 1999 ) 88 , hep - th/9901128 . t.j .
hollowood , v.v .
khoze and m.p .
mattis , _ instantons in @xmath15 sp@xmath18 and so@xmath18 theories and the ads / cft correspondence _
, hep - th/9910118 .
t. eguchi and h. kawai , _ reduction of dynamical degrees of freedom in the large n gauge theory _ ,
48 ( 1982 ) 1063 .
gross and y. kitazawa , _ a quenched momentum prescription for large n theories _
b206 ( 1982 ) 440 .
w. krauth , h. nicolai and m. staudacher , _ monte carlo approach to m - theory _ , phys .
b431 ( 1998 ) 31 , hep - th/9803117 . w. krauth and m. staudacher , _ finite yang - mills integrals _ , phys .
b435 ( 1998 ) 350 , hep - th/9804199 . w. krauth and m. staudacher , _ eigenvalue distributions in yang - mills integrals _ , phys .
b453 ( 1999 ) 253 , hep - th/9902113 .
h. aoki , s. iso , h. kawai , y. kitazawa and t. tada , _ space - time structures from _ iib _ matrix model _ , prog .
99 ( 1998 ) 713 , hep - th/9802085 .
t. hotta , j. nishimura , a. tsuchiya , _ dynamical aspects of large @xmath2 reduced models _
b545 ( 1999 ) 543 , hep - th/9811220 .
m. b. green and m. gutperle , _ d - particle bound states and the d - instanton measure _ , jhep 9801 ( 1998 ) 005 , hep - th/9711107 .
g. moore , n. nekrasov and s. shatashvili , _ d - particle bound states and generalized instantons _ , hep - th/9803265 . m. fukuma , h. kawai , y. kitazawa and a. tsuchiya , _ string field theory from iib matrix model _ ,
b510 ( 1998 ) 158 , hep - th 9705128 .
a. ambjorn , k. anagnostopoulos , w. bietenholz , t. hotta and j. nishimura , _ monte carlo studies of susy large n reduced models _
, poster presented by j. nishimura at strings 99 .
|
@xmath0 yang - mills integrals form a new class of matrix models which , in their maximally supersymmetric version , are relevant to recent non - perturbative definitions of 10-dimensional iib superstring theory and 11-dimensional m - theory .
we demonstrate how monte carlo methods may be used to establish important properties of these models .
in particular we consider the partition functions as well as the matrix eigenvalue distributions .
for the latter we derive a number of new exact results for @xmath1 .
we also report preliminary computations of wilson loops .
aei-1999 - 38 + +
|
in simple terms , our experimental procedure is to drop a cloud of cold cesium atoms from a magnetooptic trap ( mot ) into a high - finesse optical cavity , while continuously monitoring the cavity s complex susceptibility with a weak probe laser @xcite . by limiting the number of atoms in the initial cloud
, we can easily reach an operating regime in which atoms transit the cavity only one at a time . using broadband heterodyne detection and a high - speed digitizer
, we continuously record both the amplitude and phase of the transmitted probe beam during @xmath12 ms time windows .
each window typically contains from zero to five atom transit signals .
the elementary theoretical description of such a measurement employs steady - state solutions of the nonselective master equation for a _ stationary _ two - level atom coupled to a single electromagnetic mode via the jaynes - cummings interaction hamiltonian .
atomic center - of - mass motion and optical pumping among zeeman states can only be included in this treatment by allowing for a time - dependent atom - cavity coupling strength .
although this type of approach can not make predictions about dynamical variations in the coupling strength , it does provides a quantitative basis for interpreting some time - independent features of our data .
our use of this `` adiabatic '' model may be justified to a certain extent by the separation of timescales that we achieve in optical cavity qed with laser - cooled atoms . in optimal cases ,
the atom - cavity coupling strength should vary by as little as a factor of @xmath13 over the system damping time @xmath14 ns .
if we let @xmath15 denote the density operator for the joint state of the atom and cavity , the nonselective master equation ( in the electric dipole and rotating - wave approximations ) reads : @xmath16 + \gamma_\perp\left(2{\hat\sigma}\rho { \hat\sigma}^\dagger - { \hat\sigma}^\dagger{\hat\sigma}\rho - \rho{\hat\sigma}^\dagger { \hat\sigma}\right)\nonumber\\ & & + \left(\kappa_a+\kappa_b+\kappa_c\right)\left(2{\hat a}\rho{\hat a}^\dagger - { \hat a}^\dagger { \hat a}\rho - { \hat a}^\dagger{\hat a}\rho\right),\label{eq : me}\\ { \hat h}_0&=&\hbar\delta{\hat a}^\dagger{\hat a } + \hbar\theta{\hat\sigma}^\dagger{\hat\sigma } + \sqrt{2\kappa_a}{\cal e}\left({\hat a}+ { \hat a^\dagger}\right ) + h_{\rm int},\\ h_{\rm int}&=&\hbar g_0 e^{-\left(y^2+z^2\right)/w^2 } \cos\left(k_lx\right)\left[{\hat a}{\hat\sigma}^\dagger+{\hat a}^\dagger { \hat\sigma}\right].\end{aligned}\ ] ] here @xmath17 is the atomic dipole decay rate , @xmath18 is the cavity field decay rate through the input mirror ( through which the probe laser is injected ) , @xmath19 is the cavity field decay rate through the output mirror , @xmath20 is the cavity field decay rate due to intracavity scattering / absorption losses , @xmath21 is the atom - probe detuning , @xmath22 is the cavity - probe detuning , and the coupling strength @xmath23 is equal to half the maximum single - photon rabi frequency . we treat the atomic center - of - mass coordinates @xmath24 as c - number parameters , with the @xmath25-axis conciding with the cavity axis and @xmath26 parallel to gravity .
the gaussian waist of our cavity mode is @xmath27 @xmath2 m .
note that we have written the master equation in a frame rotating at the drive frequency , so @xmath28 is a constant term proportional to the complex amplitude of the driving field . to find the steady - state density operator as a function of driving strength and various detunings , we simply set @xmath29 and solve for @xmath30 using linear algebra .
the expected amplitude and phase of the heterodyne photocurrent may then be computed as @xcite @xmath31{\rm tr}\left[\rho_{ss}{\hat a}\right ] , \label{eq : hetphot}\ ] ] where @xmath32 represents the overall photodetection efficiency ( including propagation losses between the cavity and photodetectors , heterodyne efficiency , and detector quantum efficiency ) , @xmath33 and @xmath34 represent the photon flux and phase of the ( optical ) local oscillator , and @xmath35 is the frequency of the optical local oscillator relative to the rotating frame ( cavity driving field ) . in the experiment , we mix @xmath36 with an rf local oscillator at the frequency @xmath35 ( which ranges between 40 and 190 mhz ) and separately record the in - phase and quadrature components of the slowly - varying envelope ( with an analog bandwidth of 300 khz ) . note that the presence of an intracavity atom can influence the heterodyne photocurrent only via the interaction hamiltonian @xmath37 , where @xmath38 in a two - level approximation for the atomic internal dynamics , and for a classical treatment of the atomic center - of - mass degrees of freedom , _ all _ steady - state properties of the atom - cavity system are strictly determined by the value of @xmath39 once the parameters ( @xmath18,@xmath19,@xmath17,@xmath40,@xmath41,@xmath28 )
have been specified .
this includes the quantity @xmath42 , which represents one possible measure of the `` interaction energy '' between atom and cavity .
given that @xmath39 appears to be a more fundamental measure of the interaction strength , however , we have adopted the convention @xmath43 .
the atom - cavity interaction can be treated semiclassically using the optical bistability state equation ( obse ) @xcite .
the obse is traditionally written in terms of the scaled field variables @xmath25 and @xmath44 ( not to be confused with the atomic coordinates ) , with the correspondence @xmath45 where the saturation photon number @xmath46 is given by @xmath47 . for a given driving strength @xmath28
, the expected intracavity field amplitude can be found by inverting the equation @xmath48 where we again work in a rotating frame at the drive frequency , @xmath49 represents the scaled atom - probe detuning , @xmath50 represents the scaled cavity - probe detuning , and the `` cooperativity '' @xmath51 is defined by @xmath52 note that the dependence of @xmath53 ( and thereby @xmath51 and @xmath46 ) on the atomic position and internal state is implicit . a semiclassical prediction for the heterodyne photocurrent
is obtained by substituting @xmath54 for @xmath55 $ ] in equation ( [ eq : hetphot ] ) .
figure [ fig : apparat ] provides a general overview of the apparatus , indicating the schematic arrangement of various components to be described below . the diode laser setup for forming the cs mot
is not shown .
we use a fabry - perot high - finesse microcavity ( `` physics cavity '' ) consisting of two spherical mirrors with 1 m radius of curvature @xcite .
the cavity was constructed with a mean length @xmath56 @xmath2 m , which we inferred from the cavity s measured free spectral range of @xmath57 hz .
the measured @xmath58 and specified radii of curvature geometrically determine the cavity s electromagnetic mode volume for tem@xmath59 modes near 852 nm @xcite .
together with the dipole decay rate @xmath60 mhz for the cs 6p@xmath61 level @xcite , this determines our optimal coupling constant @xmath62 to be @xmath63 mhz for @xmath64 transistions ( specifically the @xmath65 ) and @xmath66 mhz for @xmath67 transitions ( @xmath68 ) within the d2 zeeman manifold @xcite . in order to allow cold atoms to fall into such a short cavity
, we found it necessary to have the mirror manufacturer reduce the substrate diameters from the standard value of 7.75 mm down to 3 mm @xcite .
this reduced the `` sagittal depth '' of the curved mirror substrates and allowed us to maintain a gap of @xmath69 @xmath2 m around the edge of the cavity .
machining of the mirror substrates was performed _
after _ they had been superpolished and coated , but this process did not seem to degrade the mirror reflectivities significantly . the nominal combined transmission and loss per mirror , before machining , was @xmath70 .
direct measurements of the cavity finesse yield @xmath71 at an optical wavelength of 852.36 nm , consistent with a combined mirror transmission and loss @xmath72 .
this value of @xmath73 is inferred from the measured value of @xmath58 and the measured cavity hwhm @xmath74 mhz .
the cavity used throughout the work described in this paper suffers from a rather pronounced birefringence , which for tem@xmath59 modes near 852 nm induced a splitting of @xmath75 mhz between linearly - polarized eigenmodes .
it is not entirely clear whether this birefrigence is a result of the substrate machining for diameter - reduction , a property of the coatings , or something associated with the mirror - mounting procedure used for this particular cavity .
we note that recent efforts by other members of our group @xcite have produced a cavity of approximately the same finesse with greatly - reduced birefringence ( by a factor @xmath76 ) , using mirrors from a different coating run and with great care taken to minimize cavity misalignments and stress on the mirror substrates .
our two mirror substrates are mounted in vee - grooves atop independent aluminum blocks , with a piezoelectric actuator between the blocks for active servo - control of the mirror separation ( cavity length ) .
the cavity mount sits on a stack of alternating ofhc copper blocks and viton o - rings for passive vibration isolation , all within an ion - pumped vacuum chamber whose background pressure was typically @xmath77 torr ( inferred from the ion pump current ) . in order to bring the mot as close as possible to the central axis of the physics cavity
, we had to use a rather open ( and non - magnetic ) mount design , leading to some compromises in the way of mechanical stability .
sitting on the vibration - isolation stack and under vacuum , we found that the native noise spectrum of the cavity length extended out to about 4 khz , with one prominent resonance at 50 hz ( which we attribute to a transmission resonance of the isolation stack ) . due to a set of pzt - actuator resonances above 10 khz , we have ultimately been limited to a unity - gain bandwidth @xmath78 khz for the cavity stabilization servo . although the principle aim for this experiment is to stabilize the cavity length at some precise offset @xmath79 mhz from the cs d2 resonance at 852.359 nm , the strong atom - cavity coupling places severe restrictions on the optical power that can be used for the purpose of generating an error signal . on resonance ,
the saturation intracavity photon number for our cavity is as small as @xmath80 which sets a fiducial cavity throughput of @xmath81 pw . with such low optical power
it would be extremely difficult to obtain a high - quality error signal for locking the physics cavity .
other experiments performed in our group have circumvented this problem by using a chopped locking scheme , in which a strong `` lock beam '' alternates with a weak `` probe beam '' at 50% duty cycle and @xmath82 khz frequency @xcite .
however , such a strategy inherently limits the servo unity - gain bandwidths to @xmath69 hz at best , and would not be appropriate for future experiments with atoms trapped inside the cavity for long periods of time . in this work
we have developed an alternative locking scheme for the physics cavity , which employs an auxilliary diode laser at 836 nm to monitor the cavity length on a different longitudinal mode than that which couples strongly to intracavity cs atoms .
we use a commercial grating - stabilized diode produced by new focus ( santa clara , california ) . with a 16 nm detuning
, we can send @xmath83 nw through the cavity and incur an ac stark shift of only @xmath84 khz for the atomic resonance at @xmath85 . using an eg&g avalanche - photodiode / transimpedance amplifier module ( model c30998 ) for ac detection of transmitted 836 nm light , we obtain an fm error signal ( modulation frequency 3.8 mhz ) with signal - to - noise ratio ( snr ) @xmath86 at 30 khz bandwidth .
the 836 nm diode laser is stabilized to an auxilliary `` transfer cavity , '' which consists of a pair of 25 cm radius - of - curvature mirrors at @xmath87 cm separation .
one of the mirrors is mounted on a piezoelectric actuator to allow cancellation of dc drift and low - frequency noise .
the transfer cavity has a linewidth @xmath88 khz at both 836 nm and 852 nm , with an overall mode spacing @xmath89 mhz .
the transfer cavity is also used for pre - stabilization of the ti : sapphire laser , and some of the ti : sapphire light is used in a cs modulation - transfer spectrometer @xcite to provide an absolute reference for the transfer cavity length . from run to run we used one or two acoustooptic modulators to offset the ti : sapphire / transfer - cavity lock point by + 140 , + 87 , or + 43 mhz relative to the cs 6s@xmath90 6p@xmath91 transition . by comparison of the ti : sapphire and diode laser error signals in their respective locks to the transfer cavity , we infer that the relative rms jitter between them is @xmath92 khz . for both laser locks we use the pound - drever - hall technique @xcite of detecting an fm signal in reflection from the tranfser cavity .
the stability of the transfer cavity resonances with respect to atomic cs lines in a vapor cell was such that we did not see any relative jitter beyond the measurement noise in our modulation - transfer spectrometer ( snr @xmath93 in 30 khz bandwidth ) .
the ti : sapphire stabilization employs two feedback loops , one with @xmath94 khz bandwidth to the tweeter inside the ring laser and another with @xmath69 khz bandwidth to a vco - driven , double - passed acoustooptic modulator ( aom ) just outside the laser cavity @xcite .
the mean frequency of the aom is 76 mhz , and the error signal going to the vco has a lower cutoff of @xmath94 khz to prevent dc drifts .
we note that use of the aom is crucial for achieving high stability of the ti : sapphire frequency .
the diode laser servo utilizes both feedback to the grating pzt and direct modulation of the injection current , achieving an overall unity - gain bandwidth @xmath95 mhz .
having locked both the diode laser and ti : sapphire to modes of the transfer cavity , which itself is locked to cs , we use a travelling - wave electrooptic modulator to generate an rf sideband of the diode laser at @xmath96 mhz . either the upper or lower sideband is used to derive an fm error signal for locking the physics cavity by dithering @xmath97 at 3.8 mhz , thus allowing us to achieve arbitrary placement of the physics cavity mode near 852 nm via the tunability of @xmath97 .
our basic requirement for the quality of the physics - cavity servo was that relative jitter of the cavity resonance and the probe laser frequency should not contribute a significant amount of noise in the heterodyne photocurrent .
hence the relevant comparison to make is between the noise in both quadratures of a demodulated beatnote and the photocurrent fluctuations produced by the local oscillator alone . at 200 khz bandwidth and with a probe beam strength such that @xmath98 in the empty cavity
, the standard deviations of the phase and amplitude quadratures of the transmitted probe beam were measured to be 1.01 and 1.39 ( respectively ) relative to those of the local oscillator alone .
note that we have estimated the standard deviation of the quadrature - amplitude signals produced by our optical local oscillator to be only a factor of 1.05 above the theoretical shot - noise limit ( see below ) .
we therefore believe that our overall excess noise factor @xmath99 . by taking some simultaneous recordings of the heterodyne photocurrent and the physics - cavity error signal , we were able to verify directly that the atomic transits do not affect the physics cavity servo .
to provide a source of cold cs atoms , we used a standard magnetooptic trap loaded directly from a thermal beam @xcite .
our choice of thermal - beam loading , as opposed to loading from a background vapor , was driven by an attempt to prevent accidental coating of the physics cavity mirrors with cs . in more than two years of service
, we did not detect any significant ( @xmath100 ) change in the cavity finesse .
pre - cooling of the cs beam was not necessary for this experiment , as we required only a very low rate of delivering single cold atoms into the cavity mode volume .
our mot employs a six - beam configuration , and we orient the anti - helmholtz coils for the trap so that their symmetry axis is parallel to that of the optical cavity .
this leads to a mot laser beam geometry with one beam axis running parallel to and just above the cavity , plus two beam axes in the plane of the mirror surfaces ( figure [ fig : motconfig ] ) .
the light for the mot was provided by a pair of grating - stabilized diode lasers ( sdl 5421-g2 ) , one tuned to the cs 6s@xmath906p@xmath91 cycling transition for trapping and the other to 6s@xmath1016p@xmath102 for repumping .
each trapping beam had @xmath95 cm diameter and anywhere from 40 @xmath2w to 4 mw of optical power , depending on how many atoms we were trying to send into the physics cavity .
we typically used a cs reservoir temperature of 60 - 80 c for the thermal beam , which effused through a 200 @xmath2 pinhole and travelled an overall distance of @xmath84 cm to the trapping region ( with a cold mechanical collimator in the way to reduce loading of the ion pump ) . with an anti - helmholtz field of
around 25 g / cm , we could load up to @xmath103 atoms into a millimeter - sized cloud , whose mean temperature we estimate to be @xmath69 @xmath2k based on fluorescent imaging of free expansion .
this temperature estimate is also supported by the spread in arrival times of individual atoms falling into the cavity .
when running the experiment we would load the mot for about 0.5 s , then drop it by quickly turning off the trapping beams with an aom ( using an rf switch with @xmath104 db attenuation ) .
after the trapping beams were thus extinguished , we would ramp down the anti - helmholtz field according to an rc - filtered step with @xmath105 ms time constant .
the repumping beam was left on all the time , so that falling atoms would be shelved in the @xmath106 ground hyperfine level before entering the cavity .
no specific preparation was performed with respect to the atomic zeeman states .
dropping @xmath103 atoms we would generally see 30 - 50 atoms falling through the central part of the cavity mode volume , so for single - atom transit data we had to reduce the trapping beam power by a factor of @xmath107 to reach 0 - 2 atoms per drop .
the overall repitition rate for the trap - drop cycle was typically 0.6 hz .
we used a balanced - heterodyne setup in order to achieve high - efficiency , zero - background photodetection of @xmath95 pw levels of 852 nm light transmitted through the physics cavity . the frequency difference between cavity probe light and the optical local oscillator for
heterodyne detection was between 40 - 190 mhz , depending upon our choice for the atom - probe detuning .
the probe light was generated from the ti : sapphire output by cascading a + 200 mhz aom and a tunable travelling - wave electrooptic modulator , which was driven between -245 mhz and -440 mhz to produce the desired atom - probe detuning .
this indirect method was required to prevent contamination of the heterodyne photocurrent by electronic noise at the heterodyne frequency .
light leaving the physics cavity first hit a color - separation mirror which reflected @xmath108 of the 852 nm light but transmited @xmath109 of the 836 nm light , allowing us to recover an error signal for locking the physics cavity ( see above ) without compromising the overall detection efficiency for the probe field .
residual 836 nm light going to the heterodyne setup amounted to only @xmath14 nw and had negligible effect on the photocurrent of interest .
the local oscillator ( lo ) for the optical heterodyne was spatially cleaned by a @xmath110 fabry - perot cavity ( linewidth @xmath95 mhz ) , which also served to strip off spectral noise at 76 mhz associated with the aom servo for stabilization of the ti : sapphire frequency .
the cleaning cavity was locked using the pound - drever - hall method @xcite with fm sidebands at 24 mhz , which likewise had to be kept weak in order not to saturate the ac gain of the heterodyne photodetectors .
we used a total of @xmath111 mw in the lo , which generated a shot - noise level @xmath112 db above the electronic noise of the photodetectors in the frequency range of interest .
the difference photocurrent from the balanced heterodyne detectors was amplified up to -50 dbm or higher , then divided by a @xmath113 rf splitter .
an independent signal generator was used to produce an rf local oscillator at the heterodyne frequency , and it was halved using a @xmath114 rf splitter .
the two identical copies of the photocurrent were mixed with the in - phase and quadrature copies of the rf lo to produce an orthogonal pair of quadrature amplitude ( qa ) signals at baseband .
the qa signals were further amplified , and passed through 300 khz analog filters with a roll - off of 12 db / octave .
we used a 12-bit adc to sample both qa s simultaneously at a rate of 10 mhz per channel , which is sufficiently high to avoid signal aliasing completely . following each drop of the trap , we continuously recorded both qa s for a data acquisition window of 50 ms and streamed the data to a hard drive for offline processing following the experimental run .
ideally , we would like the data acquisition procedure just described to yield directly the amplitude and phase quadrature amplitudes of light transmitted through the cavity . if we write the transmitted optical field as @xmath115 , where @xmath116 is a slowly - varying complex amplitude , the amplitude @xmath117 and phase @xmath118 quadrature - amplitudes are defined by @xmath119 . with respect to the quantum - mechanical theory of the master equation ( [ eq : me ] ) , @xmath120 .
note that we define @xmath121 to have zero phase when the cavity is empty , so that @xmath118 should have zero mean when there are no intracavity atoms .
given the way that we generate the probe beam , however , we have no way of generating a phase - locked rf local oscillator to recover @xmath117 and @xmath118 directly .
the phase of the heterodyne photocurrent differs from the phase that the light has just after it leaves the cavity because of fluctuations in the relative optical path length travelled by the signal beam and optical local oscilator in reaching the photodetectors .
so the two signals produced by mixing the photocurrent with the shifted and unshifted copies of our rf local oscillator correspond to an orthognal , but _ rotated _ pair of quadrature amplitudes @xmath122 , @xmath123 : @xmath124 luckily , the characteristic timescales for fluctuations in the phase @xmath125 are quite long ( @xmath126 ms , corresponding to acoustic disturbances ) compared to the 250 @xmath2s duration of an individual atom - transit signal . in processing the recorded data to produce the plots discussed below , we have therefore used an `` adaptive '' definition of the amplitude and phase quadrature amplitudes . within a window of
2 ms preceding the signal of interest , we estimate the instantaneous value of @xmath125 by determining the rotation of @xmath122,@xmath123 that produces one quadrature @xmath127 with zero mean and one quadrature @xmath128 with positive mean .
then @xmath127 is operationally defined to be the phase quadrature photocurrent , and @xmath128 is the amplitude quadrature photocurrent .
figure ( [ fig : hmtlong ] ) shows an example of a 15 ms segment of our quadrature amplitude data .
note that some excess low - frequency noise can still be seen in the phase quadrature , which constrains the lower end of our measurement bandwidth to @xmath95 khz .
six prominent atom - transit signals , characterized by a sharp drop in the amplitude quadrature and a simultaneous increase in the phase quadrature , can be seen between @xmath129 sec and @xmath130 sec . in our subsequent discussions of the data ,
we shall focus on individual signal `` events '' of this type .
in order to make quantitative comparisons between data and theory , we need to calibrate the strength of the driving field used in each experimental run . as we will ultimately choose to ignore dc optical phase offsets , the relevant quantity for us will be the number of photons @xmath131 that builds up when the cavity is empty .
this corresponds to @xmath132 in the master equation case and @xmath133 for the obse .
our strategy for determining the intracavity photon number during experimental runs has been to work backwards from the heterodyne photocurrent observed with no atoms in the cavity , using the expression @xcite @xmath134 where @xmath135 is the mean value of the demodulated amplitude quadrature signal ( i.e. the heterodyne photocurrent is of the form @xmath136 ) , @xmath137 is the mean - squared power of fluctuations in the same signal due to ( optical ) shot noise , @xmath138 is the measurement interval , and @xmath139 . note that the above expression is valid for a coherent signal beam , which we assume to be the state of the light transmitted through the cavity when no atoms are present . for an accurate calibration , we thus need to know the output mirror transmission @xmath19 , the total cavity loss ( which can be determined from measurements of the cavity finesse ) , and the overall photodetection efficiency @xmath32 . in determining @xmath131 from the data , we typically chose @xmath138 on the order of 1 ms .
the following subsections provide further detail for each critical aspect of the calibration .
ideally , the quantity @xmath137 appearing in equation ( [ eq : hetsn ] ) should coincide with the photocurrent noise power @xmath140 observed when the signal beam is blocked .
this would allow us to determine the intracavity photon number without having to calibrate the exact gains of the photocurrent amplifiers , etc .
but the procedure is invalid if the optical local oscillator carries excessive technical noise , so we have empirically checked the scaling of our nominal @xmath137 with the dc optical power of the lo .
a fit of the data to the functional form @xmath141 yields @xmath142 mw@xmath143 . given our typical operating lo power of 1 mw per detector , and considering the relative magnitudes of other uncertainties , we approximate @xmath144 .
note that the lo does pass through an intensity stabilizer with @xmath145 hz bandwidth just before entering the cleaning cavity .
we directly determined the total cavity field decay rate @xmath146 by making a calibrated measurement of the hwhm of a cavity tem@xmath59 resonance . in order to minimize systematic errors , we did this by using the heterodyne detectors to monitor the transmitted optical power of a fixed - frequency ti : sapph probe beam while scanning the cavity length .
the cavity length was always under servocontrol during the measurement , as we generated the scan by stepping the frequency of the rf going to the travelling wave modulator for the locking diode laser .
the total cavity field decay rate is then given by the resonance hwhm measured in terms of the modulator rf scan , times a correction factor of the ratio of the diode laser and ti : sapph wavelengths .
we find @xmath147 mhz .
as the mirrors used to construct the cavity should be identical , we assign half the total losses to each mirror .
unfortunately , we did not manage to characterize the ratio of intracavity losses @xmath20 to transmission losses @xmath148 before we accidentaly damaged the cavity .
given that the mirror coatings we have previously received from the same manufacturer have displayed very low scattering / absorption loss ( @xmath149 ) in the wavelength range of interest @xcite , we have assumed @xmath150 mhz .
three principal factors determine the value of @xmath151 : the spatial overlap @xmath152 between the signal beam and the optical local oscillator , the photodetector quantum efficiency @xmath153 , and the ( power ) efficiency @xmath154 with which we transfer light from the cavity ouput to the photodetectors .
using an optical powermeter we measured @xmath155 , as well as @xmath156 for our two photodetectors .
we measured @xmath157 by producing a dc fringe between the signal beam and the optical local oscillator , with both beams adjusted to a power level of 1 @xmath2w .
the power adjustments were made upstream of the physics cavity and lo cleaning cavity , ensuring that no optical misalignments were introduced in the process .
we obtained an independent measurement of @xmath158 from the heterodyne signal - to - noise ratio obtained with a measured signal power of @xmath159 nw .
note that the reasonable agreement between this latter value and the directly - measured @xmath160 provides further indication that our optical local oscillator bears minimal excess noise , and that our assumptions @xmath161
, @xmath162 are valid .
although it could have been avoided , we did suffer one additional loss of detection efficiency due to gain and phase imbalance in the subtraction of heterodyne photocurrents .
several weeks after we took the data sets shown below , we realized that one photodetector output had an rf signal 4 db higher than the other ( this figure includes both the deviation of our heterodyne beamsplitter from being 50/50 , the difference in photodiode quantum efficiencies , and the difference in transimpedance gains ) , and a phase offset of 1 radian . if we write the two photocurrents ( for a coherent signal beam ) as @xmath163 and @xmath164 ( where @xmath165 and @xmath166 are uncorrelated gaussian white noises with zero mean and unit variance ) , the difference photocurrent is @xmath167 where @xmath168 is again a gaussian white noise process with zero mean and unit variance .
the complex imbalance @xmath169 should thus reduce the effective photodetection efficiency ( for the purpose of evaluating the heterodyne signal - to - noise ratio ( [ eq : hetsn ] ) ) by @xmath170 using the measured @xmath171 and @xmath172 rad , @xmath173 , and we quote an overall value of @xmath174 .
in order to facilitate the interpretation of our data , we generated some rudimentary monte carlo simulations of the heterodyne signals that we should see as atoms fall through the cavity .
our code simulates three - dimensional , classical center - of - mass motion of individual atoms under the influence of fluctuating forces due to their strong interactions with the cavity field .
gravity is also included in the kinematics , but plays only a minor role over the @xmath95 ms duration of the simulations .
the basic scheme of the simulations is to precompute the values of the mean cavity - induced force , the cavity - field - induced diffusion coefficient , and the steady - state ( complex ) amplitude for the intracavity field as a function of atomic position .
hence , _ every _ other degree of freedom in the simulation is adiabatically eliminated and slaved to the atomic motion , which is assumed to be the slowest and the `` stiffest '' process in the dynamics . the mean force and intracavity field
are derived from steady - state solutions of the master equation ( [ eq : me ] ) , and the cavity diffusion is computed using the quantum regression theorem @xcite . in each timestep
, the code first performs an interpolation on the precomputed tables of values to determine the appropriate change in atomic momentum ( which includes a stochastic increment consistent with the local value of the diffusion constant ) , records an appropriate value for the cavity output field , and then updates the atomic position and momentum .
the simplest possible integration scheme is used ( explicit ito - euler ) , yielding order @xmath175 convergence in the timestep .
we chose a timestep of 7.5 ns in order to keep the run - times for the simulations reasonable under matlab on our pentium ii workstations , and this should have been sufficiently small to keep the integration error below other sources of inaccuracy .
the principle shortcoming of this scheme is that it assumes the atomic velocity will remain small enough that variations in the coupling strength @xmath53 will be negligible over timescales on the order of @xmath176 and @xmath177 .
if this condition is violated , then the steady - state values of the mean cavity force and cavity output are no longer appropriate .
another major approximation was made in treating only a two - level atom optical pumping and/or the associated opto - mechanical effects ( _ e.g. _ , spontaneous forces ) could certainly play an important role in determining the shapes of our observed transit signals .
a third , less severe approximation is that we have treated the stochastic process associated with recoil from spontaneous - emissions as being statistically independent from the one associated with dipole - force fluctuations .
the possibility that interesting effects could arise due to quantization of the atomic motion is of course ignored as well .
given the relative simplicity of our simulation scheme , it would be inappropriate to draw any strong conclusions about our experimental data soley on the basis of apparent similarities with the numerical results .
however , the simulations do provide an appealing _ model _ for the atomic center - of - mass dynamics and associated heterodyne signals , with predictions that seem to be fully consistent with what we see in the actual experiment .
this model suggests some interesting interpretations for qualitative features of the data .
for example , the marked asymmetry ( in time ) of most of our experimental transit signals may be associated with an initial phase of atomic `` channeling '' in the dipole potential provided by the intracavity optical field , followed by a sudden escape from local confinement due to cavity - field - induced momentum diffusion . such a dynamical process , if it could be confirmed in some way , would bear a strong resemblence to the levy walk behavior predicted in @xcite and observed in @xcite .
our simulations also suggest that isolated dips and steps in the observed transit signals may be associated with sudden changes in atomic localization relative to the cavity standing wave . before presenting some results of the atomic - motion simulations ,
let us first describe the numerical method used to compute the cavity diffusion coefficients . following doherty _
et al _ @xcite , we computed the diffusion coefficient associated with dipole - force fluctuations according to @xmath178 where @xmath179 is the force operator @xmath180 and we define @xmath181 .
\label{eq : gofr}\ ] ] note that the vector nature of the force operator comes only from the gradient @xmath182 . unlike the computation of mean forces and the expected intracavity field amplitude , the evaluation of ( [ eq : diffco ] ) requires an actual time - integration of the cavity - qed master equation ( [ eq : me ] ) .
note that @xmath183 - \langle{\bf f}\rangle^2 , \label{eq : heisen}\ ] ] where @xmath184 is the liouville superoperator defined by @xmath185 with reference to the cavity - qed master equation ( [ eq : me ] ) .
knowing the steady - state atom - cavity density matrix @xmath30 , one can evaluate the correlation function in ( [ eq : heisen ] ) by integrating the master equation for a time @xmath154 with @xmath186 as an initial condition , multiplying the result by @xmath30 , and finally taking the trace .
the numerical integration was performed by a c++/mpi code on an sgi / cray origin-2000 cluster .
we used a truncated basis of 25 fock states for the cavity mode , using an explicit euler integration with a 1 ps timestep and 5 @xmath2s total integration time .
figure [ fig : cavdiff1 ] shows the results of such a calculation for the diffusion coefficient as a function of atomic position along the standing wave , for an atom located on the cavity axis .
recall that the cavity mode function varies as @xmath187 along the cavity axis , and as a gaussian in the transverse dimensions .
starting from the top , the four curves represent @xmath188 , @xmath159 , @xmath189 , and @xmath190 mhz , all with a probe strength such that an average of 2 photons would build up in the cavity if it were empty and with @xmath191 .
figure [ fig : cavdiff2 ] shows similar results , all for @xmath192 mhz , but with an average intracavity photon number ( for the empty cavity ) of @xmath193 for the top curve , 4 for the middle , and 2 for the lowest . as the variation of @xmath53 along the standing wave is much more rapid than in the transverse ( gaussian ) directions , we approximated @xmath194 in expression ( [ eq : diffco ] ) .
the computation of @xmath179 for the mean force field was of course three - dimensional .
note that expression ( [ eq : diffco ] ) gives the diffusion coefficient associated with dipole force fluctuations only it takes no account of recoils from spontaneous emission .
we therefore incorporate a second diffusion process in the simulations , whose coefficient is computed from the expectation value of the atomic excitation at each point in space : @xmath195 here @xmath196 is the atomic spontaneous emission rate ( einstein a coefficient ) , and the number @xmath197 comes from averaging over the angular distribution pattern for dipole radiation @xcite .
figure [ fig : simex1 ] shows an example of the atomic position along the standing - wave @xmath25 and the atomic velocity along the standing - wave @xmath198 from one typical simulation . also shown
are the mean value of the intracavity phase quadrature amplitude @xmath199 , and a simulated heterodyne signal with gaussian noise added at the level appropriate to our experimental parameters ( see equation ( [ eq : hetsn ] ) ) .
note that the quantity ( @xmath200 plus noise ) is proportional to the photocurrent @xmath127 that would be recorded in our experiment . in the bottom row of the figure , @xmath200 and @xmath200 plus shot - noise
are shown at an analog bandwidth 300 khz in order to illustrate the effects of finite detection bandwidth on the qualitative features of the data @xcite .
the results clearly suggests that the gross features of the simulated signals are most strongly affected by atomic motion along the standing wave , in that the overall gaussian profile associated with motion in the @xmath26-direction becomes distorted by the `` envelope '' of the oscillatory variations due to motion along @xmath25 ( recall @xmath201 $ ] ) .
looking at the results for @xmath202 and @xmath203 , we see that the atom in this particular simulation was mechanically confined within one well of the cavity standing wave until @xmath204 @xmath2s .
that is , @xmath25 did not vary by more than @xmath205 and @xmath198 displays the oscillations that one would expect to see for an atom trapped within a potential well . at time @xmath204 @xmath2s , however , we see that momentum diffusion finally pushes @xmath198 above some threshold value such that the atom is able to escape from local confinement and `` fly above '' the periodic dipole potential associated with the cavity standing wave . generally speaking
, one expects this type of escape to occur as the atom passes through the cavity axis in the @xmath206-direction and the standing - wave potential wells begin to decrease in depth .
the expectation value @xmath200 clearly reflects the qualitative features of the atomic trajectory @xmath202 , although we also see that the details are lost in the measured signal due to heterodyne shot - noise .
what survives in the simulated heterodyne signal is an overall asymmetry in the envelope of the transit - signal , with a sudden `` step '' downwards at @xmath207 @xmath2s ( as suggested by the dotted vertical line ) .
figure [ fig : sims ] shows four simulated signals with @xmath208 mhz and an average intracavity photon number of @xmath209 , and two simulated signals with @xmath192 mhz and @xmath210 ( @xmath191 in all cases ) . in each subplot , the upper trace shows the intracavity amplitude quadrature @xmath211 versus time , and the lower trace ( with zero mean value ) shows the intracavity phase quadrature @xmath212 versus time .
again , gaussian noise has been added to the simulated signals in order to reproduce the overall signal - to - noise ratio predicted by equation ( [ eq : hetsn ] ) . in all simulations ,
the atom was started at a position 7 gaussian waists above the cavity with a vertical velocity of -47 cm / s .
the initial transverse position , transverse velocity , position along the cavity standing - wave axis , and velocity along the standing - wave were randomly assigned within parameter ranges that were narrow enough to ensure that most simulated atom - drops produced a sizeable signal .
the simulated signals in figure [ fig : sims ] may be compared directly to the real data shown in figure [ fig : hmthbw ] . like the simulation shown in figure [ fig : simex1 ] , the signals in figure [ fig : sims ] display a generic asymmetry and often contain an isolated dip or step .
the features of this type that are marked by an arrow near the horizontal axis are again associated with sudden changes in the atomic localization relative to the cavity standing wave ( as determined by examining @xmath202 from the simulations ) .
although we have the benefit of knowing both the `` actual '' trajectory of an atom and the corresponding heterodyne signal in our simulations , _ a priori _ knowledge of the atomic trajectory is of course unavailable in our experiment . hence , the results of our numerical simulations can only provide general guidelines for how we might try to infer something about individual atomic trajectories from our heterodyne transit signals .
first of all , it appears that signficant asymmetries in the observed signals could provide evidence that we do in fact observe mechanical aspects of the atom - cavity coupling within individual atomic transits .
relative to the approximations involved in our simulations however , one should bear in mind that such asymmetries could also arise from optical pumping processes that may occur during at atomic transit through the cavity .
second , the simulations suggest that the transit signal envelopes are most strongly influenced by atomic motion along the cavity standing wave , as opposed to details of the motion along @xmath44 or @xmath26 .
third , isolated dips or steps in the signal could be indicative of sudden changes in the atomic localization with respect to the standing wave .
again however , we should be careful to note that such features might also be caused by intracavity optical pumping .
moving on to the experimental data , let us first discuss some transit signals displayed in the same fashion as the simulations of figure [ fig : sims ] .
the quality of this data illustrates our experimental ability to perform continuous , nearly quantum - noise limited measurement of the atom - cavity interaction energy during individual scattering events . in the second subsection we shall display transit data on the complex plane , and
present a quantitative comparison to theoretical predictions of the atom - cavity system s complex optical susceptibility .
we shall see that in optimal cases , sufficient information may be obtained from an _ individual _ atom - transit signal to distinguish unambiguously between quantum and semi - classical models of cavity qed .
note that all of the data shown here were taken with cavity - probe detuning @xmath213 , while the atom - probe detuning @xmath214 and probe power @xmath131 are varied .
the basic unit of our experimental data is a continuous stream of quadrature - amplitude values versus time , recorded for @xmath12 ms windows following each dropping of the mot .
one @xmath215 ms segment of such data is shown in figure [ fig : hmtlong ] .
this particular data segment was taken with detunings @xmath216 mhz and with a probe power such that an average of @xmath217 photons would build up in the cavity if it were empty ( @xmath218 ) .
figure [ fig : hmthbw ] shows six of the largest single - atom transit signals from our entire data set , which covers detunings @xmath40 from @xmath219 to @xmath220 mhz and probe powers @xmath131 from @xmath217 to @xmath221 photons .
the particular values of @xmath40 and @xmath131 for each signal in figure [ fig : hmthbw ] are displayed by the y - axis .
the data are shown at our full analog bandwidth of 300 khz and sampling rate of 10 mhz ( 12 bit resolution ) .
note that we have displaced the amplitude quadrature signals @xmath128 by @xmath222 in order to prevent them from overlapping with the phase quadrature signals @xmath127 .
we have found that transit signals of maximal contrast tend not to have very much internal structure , although there is a fairly prominant dip at the point indicated by an arrow in subplot `` dat4 , '' a set of three dips indicated by the arrow in subplot `` dat6 , '' and oscillatory structure in the signal of subplot `` dat5 . ''
the signal of subplot `` dat4 '' is really not so different in its overall structure than the simulations shown in figure [ fig : simex1 ] and in the `` traj10 '' subplot of figure [ fig : sims ] .
one can see that the shapes of the overall signal envelopes do vary substantially and are generally asymmetric in time .
the distribution of the atom - transit signal between amplitude and phase quadratures clearly depends upon the probe detuning , as will be discussed in greater detail below . for the data event recorded with @xmath188
, one sees that there is only a reduction in the power of the transmitted probe without any shift in its phase . at all other detunings ,
the atom - induced reduction of amplitude quadrature is partly offset by an increase in the phase quadrature , indicating a significant phase shift of the transmitted probe beam . in subplot `` dat6 '' taken with @xmath192 mhz
, the signal is seen to primarly reside in the phase quadrature . for signals such as those shown in figure [ fig :
hmthbw ] , we may estimate a full - signal to rms - noise ratio of approximately 2.5 for the phase quadrature and 4 for the amplitude quadrature . combining the two signals
, we have an overall ratio of 4.5 .
the signal bandwidth is 300 khz , implying a relative sensitivity of @xmath223 hz@xmath224 . assuming that the largest signals correspond to atoms reaching the maximal coupling strength of @xmath225 mhz , this sets our broadband sensitivity @xmath226 to time - varations in @xmath227 to be @xmath7 khz/@xmath11 . given our overall photodetection efficiency @xmath174 and our independent assesment of the excess noise factor @xmath99 ( corresponding to technical noise on the optical local oscillator power and on the physics cavity lock ) , we find that this value lies only a factor of @xmath228 above the fundamental quantum noise floor .
we maintain this high sensitivity over 2.5 decades of signal frequency , from our full photodetection bandwidth of 300 khz down to 1 khz ( at which point we are limited by residual technical noise in our measurements of the phase quadrature ) .
although a complete theoretical description of our experiment at the level of a stochastic master equation @xcite could in principle be formulated to include three - dimensional quantized atomic center - of - mass motion , cavity birefringence , and the full manifold of internal atomic zeeman states , the resulting model would almost certainly be too complex to be useful for quantitative interpretations of our existing data .
nevertheless , we wish here to stress that only a theory that explicitly treats the _ quantum conditioning _
@xcite of the atom - cavity evolution on the stochastic component of the heterodyne photocurrent could provide a full account of the statistics and autocorrelation properties of our atom - transit data .
in this crucial sense , we would claim that our recent work may be distinguished from all other efforts in contemporary experimental physics .
of course , much remains to be done before we can offer definitive proof , so our continuing efforts will largely be directed towards the unambiguous demonstration of quantum conditional dynamics in cavity qed .
in addition to displaying our atom - transit signals in the format of photocurrent versus time , we can also construct parametric plots of @xmath128 versus @xmath127 to examine the correlation induced between these two quantities by the atom - cavity interaction .
this is equivalent to compiling a histogram of the complex amplitude of the optical field transmitted through the cavity during an atomic transit , and hence to a continuous monitoring of the atom - cavity system s complex optical susceptibility .
the underlying probability distributions for such histograms are determined by the ( evolving ) husimi q - function for the intracavity field @xcite .
many of our measurements have been conducted in a regime of strong cavity - driving fields ( @xmath229 ) , with the consequence that significant effects of saturation and optical nonlinearity can be seen in our experimental data . as has been extensively discussed in the cavity - qed literature , the nonlinear optical response of
the atom - cavity system represents an important experimental signature that may be used to distinguish between quantum and semiclassical theoretical models for the atom - cavity interaction .
prior to this work , four groups @xcite have reported nonlinear measurements in cavity qed . in each of these previous experiments ,
the information gained per atomic transit about the details of the system s nonlinear response was far less than in the data we shall present here .
hence , we wish again to stress that the techniques developed in our most recent experimental work have opened new possibilities for exploring quantum dynamics at the level of single quantum realizations ( trajectories ) , as opposed to the level of ensemble averages .
in figures [ fig : hmtfig3 ] and [ fig : hmtfig2 ] , we display single - transit data on the complex plane .
the experimental data are displayed as gray dots , with each dot representing the values of @xmath128 and @xmath127 at some particular time during an atomic transit . in order to produce a set of dots from a transit signal such as those shown in figure [ fig : hmthbw ] , the quadrature amplitude signals were first passed through an anti - aliasing filter and then resampled every 10 @xmath2s to produce a discrete set of points @xmath230 . for every @xmath231 , a gray dot is placed at radius @xmath232 and polar angle @xmath233 .
these transit phasors represent a simple way of looking at the correlation between the amplitude and phase of light transmitted through the cavity , with time removed from the picture .
we believe that this type of plot is the best format for comparisons to theory , because atomic internal ( optical pumping among zeeman states ) and external ( motion through the cavity eigenmode ) dynamics should be factored out . in the simplest approximation
, these two types of processes merely induce fluctuations in the atom - cavity coupling @xmath53 , which should only move the locus of our gray dots in and out along a curve parametrized by @xmath53 .
the overall shape of this curve should be dictated by the interaction hamiltonian for the atom and cavity mode , and is therefore quite easy to compute . in figure
[ fig : hmtfig3 ] , we show a set of transit phasors taken at various probe detunings . the data overlay theoretical curves predicted by the quantum and semiclassical theories .
the quantum - mechanical predictions are computed by finding the steady - state solution of the master equation ( [ eq : me ] ) for values of @xmath53 in the range @xmath234 $ ] and appropriate values of @xmath40 and @xmath131 .
the solid curves represent an interpolation through the discrete set of computed values for @xmath212 and @xmath200 as a function of @xmath53 .
the semiclassical prediction is computed in the analogous manner , using the optical bistability state equation ( [ eq : obse ] ) rather than the quantum master equation .
note that both curves must agree for @xmath235 , and this common point of origin is marked in each subplot by a triangle .
the @xmath236 endpoint of the quantum curve is marked by a circle , and the @xmath236 endpoint of the semiclassical curve is marked by an @xmath237 . in the subplots of figure [ fig : hmtfig3 ] ,
the grey dots represent an overlay of data from two individual atom - transit signals .
the two particular signals displayed in each subplot were selected on the basis of having maximal `` contrast '' among all the signals from a given data set , under the assumption that the atoms causing maximal signals should have come the closest to actually achieving @xmath236 on their way through the cavity .
quantum theory is seen to predict the observations quite well for all the detunings shown in figure [ fig : hmtfig3 ] .
the quantum and semiclassical predictions shown in figure [ fig : hmtfig3 ] do not differ significantly except for the case of @xmath188 .
generally speaking , the two theories are known to agree in their predictions for weak - field response of the atom - cavity system but to differ in their predictions regarding _ saturation _ of the optical response .
we here present experimental evidence _ from single atom transits _ for the quantum character of saturation in the atom - cavity system s response to near - resonant driving fields . in figure
[ fig : hmtfig2 ] we show a set of transit phasors for fixed detuning @xmath208 but variable probe strength @xmath238 , @xmath239 , @xmath240 , and @xmath221 . a clear discrepancy between the quantum and semiclassical theory may be seen in this sequence of plots , with the data ( gray dots ) showing significant preference for the quantum - mechanical predictions .
each subplot in figure [ fig : hmtfig2 ] represents an overlay of data from three individual atom - transit signals , in order to fill in the overall `` shape '' of the experimental transit phasors .
it should be clear , however , that just a single data trace would suffice to determine that the experiment matches much more closely with the quantum theory than the semiclassical theory .
we plan to elaborate this result in separate publication @xcite .
in summary , we have described the details of our recent experimental work to perform continuous measurement of the interaction energy @xmath241 between one atom and an optical cavity during individual transit events .
we displayed heterodyne transit signals in two complementary formats , one of which highlights the large bandwidth and nearly quantum - noise limited signal - to - noise ratio achieved in tracking the time - evolution of the amplitude and phase quadratures ( @xmath242 ) of light transmitted through the cavity .
the second data format ( transit phasors ) displays the correlation induced between @xmath128 and @xmath127 by the atom - cavity interaction , and our experimental results show that we are able to distinguish unambiguously between quantum and semiclassical models of cavity qed in only a few ( or even just one ) atomic transits . we have furthermore presented the results of rudimentary numerical simulations of atomic motion under the influence of mechanical forces and momentum diffusion associated with the strong atom - cavity coupling , and we examined the interpretations suggested by these simulations for certain qualitative features of our experimental data ( _ i.e. _ steps and asymmetries ) .
the primary conclusion we wish to draw in comparing the simulation results with our experimental data is that we have experimentally reached a regime of measurement sensitivity and bandwidth in which details of the atomic center - of - mass trajectories really should be visible . even though we have no incontrovertible means of proving that steps and asymmetries seen in our experimental data should be associated with dynamical processes like channeling and diffusive escape , we are now motivated in our continuing work to develop some means of _ actively influencing _ the atomic motion _ while _ it is still inside the cavity .
this would allow us to produce deliberate displacements of an atom along the cavity standing wave , and to examine the induced variations of the heterodyne signal in order to verify our inferred displacement sensitivity of @xmath10 m/@xmath11 .
we anticipate that the combined abilities of monitoring and influencing atomic position relative to the cavity standing wave will enable the investigation of schemes for _ real - time feedback control _ of quantized atomic center - of - mass motion . finally , let us note that the standard quantum limit for overall observation time of the position of a single atom with sensitivity @xmath10 m/@xmath11 and 300 khz bandwidth should be @xmath243 @xmath2s @xcite .
this implies that quantitative experimental investigations of conditional quantum dynamics ( as described in @xcite ) should indeed become possible in our cavity qed system once we are able to reliably prepare individual atoms in well - defined initial states of motion .
the ideal situation in this regard would be to trap and localize atoms _ within _ the cavity , releasing them at a node or antinode of the standing wave , on the cavity axis , and with an initial position uncertainty that is small compared to @xmath205 .
current efforts in our group focus on trying to achieve this level of control via optical dipole - force traps and/or far - detuned optical lattices inside the cavity .
we wish to acknowledge q. a. turchette s vital participation in the early stages of this work , and to thank a. c. doherty , c. j. hood , and t. w. lynn for valuable discussions .
much of the numerical work described in this paper was performed on resources located at the advanced computing laboratory of the los alamos national laboratory .
this research was supported by the national science foundation under grant no .
phy97 - 22674 , by the onr , and by darpa through the quic initiative ( administered by aro ) .
jun ye is supported by an r. a. millikan fellowship from the california institute of technology .
h. j. kimble : in _ cavity quantum electrodynamics_. san deigo , academic press : 1994 d. m. meekhof , c. monroe , b. e. king , w. m. itano , d. j. wineland : phys .
lett . * 76 * , 1796 ( 1996 ) d. leibfried , d. m. meekhof , b. e. king , c. monroe , w. m. itano , d. j. wineland : phys . rev . lett .
* 77 * , 4281 ( 1996 ) w. e. moerner : science * 265 * , 46 ( 1994 ) b. c. stipe , m. a. rezaei , w. ho : science * 279 * , 1907 ( 1998 ) m. r. andrews , d. m. kurn , h. j. miesner , d. s. durfee , c. g. townsend , s. inouye , w. ketterle : phys . rev .
* 79 * , 553 ( 1997 ) m. kozuma , k. nakagawa , w. jhe , m. ohtsu : phys .
a * 57 * , r24 ( 1998 ) s. guibal , c. triche , l. guidoni , p. verkerk , g. grynberg : opt
* 131 * , 61 ( 1996 ) c. m. caves , g. j. milburn : phys .
a * 36 * , 5543 ( 1987 ) h. m. wiseman : quantum semiclass
* 8 * , 205 ( 1996 ) a. m. herkommer , h. j. carmichael , w. p. schleich : quantum semiclass
* 8 * , 189 ( 1996 ) a. c. doherty , a. s. parkins , s. m. tan , d. f. walls : phys .
a ( in press ) h. j. carmichael , _ an open systems approach to quantum optics_. berlin , springer : 1993 h. m. wiseman , g. j. milburn : phys .
a * 47 * , 642 ( 1993 ) h. m. wiseman , g. j. milburn : phys .
a * 47 * , 1652 ( 1993 ) p. zoller , c. w. gardiner : in _ quantum fluctuations , session lxiii of lecole dete de physique des houches_. amsterdam , north - holland : 1997 h. m. wiseman : phys .
lett . * 75 * , 4587 ( 1995 ) r. alicki , d. makowiec , w. miklaszewski : phys .
77 * , 838 ( 1996 ) r. schack , c. m. caves : phys .
e * 53 * , 3257 ( 1996 ) h. mabuchi : quantum semiclass
* 8 * , 1103 ( 1996 ) h. m. wiseman : phys . rev .
a * 49 * , 2133 ( 1994 ) h. mabuchi , p. zoller : phys .
lett . * 76 * , 3108 ( 1996 ) j. a. dunningham , h. m. wiseman , d. f. walls : phys .
a * 55 * , 1398 ( 1997 ) k. s. wong , m. j. collett , d. f. walls : opt .
* 137 * , 269 ( 1997 ) a. h. zewail : j. phys
100 * , 12701 ( 1996 ) s. d. gensemer , p. l. gould : phys
* 80 * , 936 ( 1998 ) m. a. m. marte , p. zoller : appl .
b * 54 * , 477 ( 1992 ) p. storey , t. sleator , m. collett , d. walls : phys . rev .
a * 49 * , 2322 ( 1994 ) c. c. gerry : phys .
a * 53 * , 1179 ( 1996 ) h. mabuchi , q. a. turchette , m. s. chapman , h. j. kimble : opt
* 21 * , 1393 ( 1996 ) c. j. hood , m. s. chapman , t. w. lynn , h. j. kimble : phys .
lett . * 80 * , 4157 ( 1998 ) m. j. collett , r. loudon , c. w. gardiner : j. mod . opt .
* 34 * , 881 ( 1987 ) l. a. lugiato : prog
. opt . * 21 * , 71 ( 1984 ) superpolished bk7 substrates with multi - layer dielectric coatings fabricated by research electrooptics in boulder , co. a. e. siegman : _ lasers_. mill valley , university science books : 1986 .
r. j. rafac , c. e. tanner , a. e. livingston , k. w. kukla , h. g. berry , c. a. kurtz : phys .
a * 50 * , r1976 ( 1994 ) o. schmidt , k .-
knaak , r. wynands , d. meschede : appl .
b * 59 * , 167 ( 1994 ) q. a. turchette , r. j. thompson h. j. kimble : appl . phys .
b * 60 * , s1 ( 1995 ) j. h. shirley : opt .
* 7 * 537 ( 1982 ) r. w. p. drever , j. l. hall , f. v. kowalski , j. hough , g. m. ford , a. j. munley , h. ward : appl .
b * 31 * , 97 ( 1983 ) w. vassen , c. zimmermann , r. kallenbach , t. w. hnsch : opt . commun . *
75 * , 455 ( 1990 ) t. l. boyd , h. j. kimble : opt
* 16 * , 808 ( 1991 ) a. cable , m. prentiss , n. p. bigelow : opt . lett .
* 15 * , 507 ( 1990 ) e. s. polzik , j. carri , h. j. kimble : appl
b * 55 * , 279 ( 1992 ) g. rempe , r. j. thompson , h. j. kimble : opt .
* 17 * , 363 ( 1992 ) c. w. gardiner : _ quantum noise_. berlin , springer : ( 1991 ) s. marksteiner , k. ellinger , p. zoller : phys . rev .
a * 53 * , 3409 ( 1996 ) a. c. doherty : private communication ( 1996 ) h. katori , s. schlipf , h. walther : phys .
lett . * 79 * , 2221 ( 1997 ) a. c. doherty , a. s. parkins , s. m. tan , d. f. walls : phys . rev .
a * 56 * , 833 ( 1997 ) q. a. turchette , c. j. hood , w. f. lange , h. mabuchi , h. j. kimble : phys .
lett . * 75 * 4710 ( 1995 ) m. brune , f. schmidt - kaler , a. maali , j. dreyer , e. hagley , j. m. raimond , s. haroche : phys .
lett . * 76 * , 1800 ( 1996 ) r. j. thompson , q. a. turchette , o. carnal , h. j. kimble : phys .
a * 57 * , 3084 ( 1998 ) h. mabuchi , j. ye , and h. j. kimble : in preparation h. mabuchi : phys .
a ( in press )
|
we report the use of broadband heterodyne spectroscopy to perform continuous measurement of the interaction energy @xmath0 between one atom and a high - finesse optical cavity , during individual transit events of @xmath1 @xmath2s duration .
we achieve a fractional sensitivity @xmath3 to variations in @xmath4 within a measurement bandwidth that covers 2.5 decades of frequency ( 1300 khz ) .
our basic procedure is to drop cold cesium atoms into the cavity from a magnetooptic trap while monitoring the cavity s complex optical susceptibility with a weak probe laser . the instantaneous value of the atom - cavity interaction energy , which in turn determines the coupled system s optical susceptibility , depends on both the atomic position and ( zeeman ) internal state .
measurements over a wide range of atom - cavity detunings reveal the transition from resonant to dispersive coupling , via the transfer of atom - induced signals from the amplitude to the phase of light transmitted through the cavity . by suppressing all sources of excess technical noise , we approach a measurement regime in which the broadband photocurrent may be interpreted as a classical record of _ conditional _ quantum evolution in the sense of recently - developed quantum trajectory theories .
= 10000 optical cavity quantum electrodynamics ( qed ) in the strong coupling regime @xcite provides a unique experimental paradigm for _ real - time _ observation of quantum dynamical processes at the _ single - atom _ level . while spectacular advances have certainly been made in the preparation and tomography of quantum states of motion for a single trapped ion @xcite , all such experiments have involved the accumulation of ensemble - averaged data over many successive realizations of the process being studied .
recent studies of single - molecule dynamics have likewise demonstrated the `` immediate '' detection of photochemical @xcite or conformational @xcite events , but such experiments presently lack the potential that cavity qed provides for observing quantum processes on a timescale that makes coherent control / intervention a tangible possibility .
we wish here to look beyond the mere detection of quantum jumps , and to focus on the development of a broadband , _ single - shot _ measurement technique that achieves signal - to - noise ratio @xmath5 over a bandwidth that includes all characterstic frequencies of a quantum dynamical process .
real - time observation of quantum dynamics in _ many - atom _ systems has recently become an important theme in atomic physics , with notable demonstrations involving vibrational excitations of a trapped bose - einstein condensate @xcite and the decay of coherent oscillations of an ensemble of atoms in an optical lattice @xcite .
in contrast to programs like these , for which the scientific emphasis lies on noninvasive observation of a system s intrinsic dynamical processes , experiments in single - atom cavity qed hold great potential for enabling precise investigations of how measurement backaction _ alters _ the dynamical behavior of a continuously - observed open quantum system @xcite .
a sophisticated theoretical basis for understanding such issues is presently maturing in the form of quantum trajectory theories @xcite , but significant technical challenges remain to be solved before definitive experiments can be performed in the lab . our purpose in the present work is to report substantial progress towards surmounting such obstacles in the context of cavity qed , and hence towards achieving the essential experimental capabilities required to perform quantitative tests of measurement - based stochastic master equations .
we ultimately hope to be able to implement some recently - proposed `` applications '' of the continuous observation of dissipative quantum dynamics , in fields such as quantum measurement @xcite , quantum chaos @xcite , and quantum feedback control @xcite .
this article focuses on a detailed description of our recent experiments that record the complete time - evolution of interaction energy between one atom and a high - finesse optical cavity , during individual transit events of @xmath1 @xmath2s duration . with characteristic atom - cavity interaction energies @xmath6 mhz , we achieve measurement sensitivities @xmath7
khz@xmath8 over a bandwidth that covers the dominant rates of variation in @xmath0 ( 1300 khz ) .
unlike typical pump - probe measurements of scattering dynamics in real ( _ e.g. _ diatomic ) molecular systems @xcite , our experiments on the jaynes - cummings `` molecule '' yield a continuous time - domain record of the atom - cavity coupling during each individual `` scattering '' event ( transit ) .
the data clearly illustrate variations caused by atomic motion through the spatial structure of the cavity eigenmode and/or optical pumping among the atomic internal ( zeeman ) states . in certain parameter regimes of the detuning and probe power ,
distinctive indications of the quantum - mechanical nature of the atom - cavity coupling can be seen in the photocurrent recorded from just a _ single _ atomic transit . for large ( @xmath9 mhz ) atom - cavity detunings
we obtain phase - contrast signals induced by individual atomic transits , corresponding to a regime of strong but dispersive coupling .
phase - quadrature measurements of atomic motion have been widely discussed in the quantum optics literature @xcite , but the present work provides the first experimental demonstration at the single - atom level .
because of the standing - wave spatial structure of the cavity eigenmode , and the corresponding rapid varation of the atom - cavity coupling strength over sub - wavelength distances , our data should in principle display a _ sensitivity _ of @xmath10 m/@xmath11 to atomic displacements along the cavity axis . unfortunately we can not realize this figure as a _ precision _ for monitoring the atomic position , as we do not presently have any means of separating signal variations due to motion through the standing wave from `` background '' contributions due to transverse motion or optical pumping . in our concluding section , we shall briefly discuss our motivations for further work to disambiguate the nature of rapid variations in our data .
|
g349.7 + 0.2 is one of the three brightest radio supernova remnants ( snrs ) in our galaxy @xcite , yet until recently very little was known about it .
it is a small diameter remnant ( @xmath10 ) classified as shell - type with a brighter eastern snr limb @xcite .
early hi observations indicated that g349.7 + 0.2 is located at a distance of @xmath11 kpc @xcite .
oh(1720 mhz ) maser emission has been detected towards g349.7 + 0.2 at a radial velocity of around + 16 kms@xmath9 , implying that the snr is interacting with dense molecular gas @xcite , and establishing a kinematic distance of 22.4 kpc .
the molecular complex associated with the remnant was mapped in the co 10 line @xcite , while near - infrared and millimeter - line observations have identified the shocked portion of the cloud located towards the center of the remnant @xcite .
the shocked gas has temperature around 45 k and density around @xmath12 @xmath3 , which agrees well with conditions required for oh(1720 mhz ) maser production @xcite . in the x - ray band g349.7 + 0.2
was first detected in the _ asca _ galactic plane survey @xcite .
follow - up pointed observations with _
asca _ showed that the remnant is one of the most luminous snrs in the x - ray band @xcite , with a thermal spectrum dominated by strong k@xmath13 lines from si , s and fe .
the global spectrum was well fitted by both single equilibrium and nonequilibrium ionization models , with a plasma temperature @xmath14 kev .
the _ asca _
data indicated an age of @xmath15 yr , an unabsorbed x - ray luminosity of @xmath16@xmath8 ergss@xmath9 , and an energy release of @xmath17 ergs in the supernova explosion .
the inferred abundances were consistent with solar values , and the large derived x - ray emitting mass ( @xmath18@xmath19 ) showed the gas to be dominated by swept - up interstellar material .
these results are consistent with the snr location near a molecular cloud . however , a detailed morphological assessment and spatially resolved spectral analysis was not possible due to the poor spatial resolution of _
asca_. in this paper
we present high - resolution x - ray observations of g349.7 + 0.2 with the _ chandra x - ray observatory_. we discuss the snr morphology and plasma properties , compare x - ray data with available atomic and molecular data , and compare the properties of g349.7 + 0.2 with those of similar snrs .
g349.7 + 0.2 was observed with the advanced ccd imaging spectrometer ( acis ) on board the _ chandra x - ray observatory _ on 13 may 2002 ( obsid 2785 ) . the snr was positioned on the acis - s3 chip , which has the best sensitivity and energy resolution in the soft x - ray bands .
data were taken in full - frame timed - exposure mode with the standard integration time of 3.2 s. we used the contributed cxc software corr_tgain @xcite to correct for time - dependence of the acis gain @xcite . after applying the corrections ,
the overall lightcurve was examined for possible contamination from a time - variable background .
after these data reduction processes , the effective exposure was 56 ks from which 48606 photons were observed for the whole remnant .
data were further reduced using standard threads in the _ chandra _ interactive analysis of observations ( ciao ) software package v.3.0.2 , and calibration database ( caldb ) v.2.26 which includes the corrected fef files for the s3 ccd used for making the acis response matrices .
we used weighted response matrices for spectral fitting to account for cti effects across the ccd .
the low - energy ( e @xmath20 1 kev ) quantum efficiency of the acis has been degraded by molecular contamination on the optical blocking filter , and correction for this is integrated in the v.3.0.2 .
ciao tools for making auxiliary response files .
in figure [ fig - chandra ] we show the exposure - corrected 1.08.0 kev band acis image of g349.7 + 0.2 .
the image has a pixel size of @xmath21 and was smoothed with a gaussian filter with a fwhm of 1 .
the emission is dominated by the eastern snr shell , with the peak emission located to the southeast .
the bright region appears complex with several knots of emission .
the western snr shell is weak and diffuse , with no noticeable structure .
as shown later , this x - ray morphology is very similar to that in the radio band .
the images in different energy bands imply that the emission from the eastern - most part of the snr shell may be softer with respect to the rest of the snr .
five unresolved sources have been detected in the field of g349.7 + 0.2using the celldetect procedure in ciao .
their properties are listed in table [ tab - psou ] , and three sources are marked with circles in figure [ fig - chandra ] .
source ps4 ( cxou j171801.0372617 ) is the only source located , in projection , within the snr shell .
all the sources have about the same number of counts , except for source ps1 .
the positions of sources were compared with the us naval observatory guide star catalogue and the 2mass all - sky point source catalogue . for sources within 2 of the telescope axis , which is true for almost all of the sources ( table [ tab - psou ] ) the radius ( 90% confidence ) of the _ chandra _ position uncertainty is 06 ( see _
chandra _ proposer s observatory guide ) .
only source 1 has an optical / nir counterpart ( usno 052527848790 ) .
the rest of the sources do not have counterparts within 5 of their position down to 22 mag in @xmath22 band .
their spectra are hard , so they could be pulsars or background active galactic nuclei .
source 4 is of particular interest since it is located within the snr and we give it more consideration in the next section . the spatially - averaged ( global ) spectrum from the entire remnant is shown in figure [ fig - sp - whole ] .
a background spectrum was extracted from the outer regions of the s3 chip . as observed previously by _
asca _ , the snr spectrum is dominated by k@xmath13 line features from si , s and fe . in addition , the _ chandra _ spectrum shows weaker lines from mg , ar and ca . for spectral
fitting we used models for an optically thin thermal plasma that has reached collisional ionization equilibrium ( vraymond ; * ? ? ? * ) and for a plasma that is still being ionized ( vpshock ; * ? ? ?
vpshock is a plane - parallel time - dependent ionizing plasma model for which the ionization timescale , @xmath23 , describes the progress of the plasma towards equilibrium and can assume a range of values between the specified lower ( usually set to zero ) and upper limits . since the version v.1.1 of the vpshock model we used for
the spectral fitting does not include the ar line , we added a narrow gaussian component at @xmath24 kev in our fits .
single - component vraymond or vpshock models with solar abundances , acceptable for analysis of the overall _ asca _ spectrum from g349.7 + 0.2 , were not able to give a satisfactory fit to the overall _ chandra _ spectrum .
a good fit was obtained using a vpshock model with enhanced si abundance , and is significantly improved by adding an equilibrium thermal component with solar abundances ; the best - fit parameters are listed in table [ tab - whole ] .
this result implies that the plasma in g349.7 + 0.2 is comprised of multiple components , with a softer component having plasma temperature around 0.8 kev , and a harder component having plasma temperature around 1.4 kev .
we show the contribution of both components to the global spectrum in figure [ fig - sp - whole ] .
as expected , the harder component ( dashed line ) dominates the spectrum above 3 kev , while the softer component ( dotted line ) dominates below that .
we find no strong nonthermal emission in g349.7 + 0.2and estimate the power law component contribution to the observed flux from the whole snr region to be less than 2.6% for photon index values between 1.5 and 3 at the 3@xmath25 confidence level . the si abundance in g349.7 + 0.2 is clearly enhanced ; the @xmath26 value of the vpshock+vray model drops by 46 when one additional free parameter ( the si abundance ) is added .
additional elemental abundances were thawed one by one based on the inspection of the residuals , but only the additional si in the vpshock ( hard ) component improved the fit significantly .
this enhanced si abundance implies that we have detected for the first time a sn ejecta component in this snr .
this also implies that the soft spectral component , which has solar abundances , is most likely related to the forward shock , i.e. , shocked interstellar material .
the ionization timescales obtained for the two components are consistent with this scenario .
the forward shock is encountering interstellar material that is denser than that of the snr interior , and will therefore reach ionization equilibrium faster than shocked ejecta that is expanding in the rarefied snr interior .
the hydrogen column density obtained from the global _ chandra _ spectrum is consistent with that obtained from the _ asca _ data @xcite .
such a large column density confirms a considerable absorption of the soft emission from the snr and is consistent with its relative distance .
the x - ray luminosity derived from the global snr spectrum is @xmath27@xmath8 ergss@xmath9 .
we note that the value obtained for the plasma temperature from the global _ asca _ spectrum represents an intermediate value between the two temperatures detected with _
chandra_. the spectra from the whole snr
can therefore be fitted with two plasma components .
could these two components be separate parts of a single evolving shock ? to test this
we also fitted the spectral regions with a single planar , initially unequilibrated shock ( from * ? ? ? * ) . in this model the temperatures of the electrons and protons equilibrate downstream through coulomb collisions , providing a natural range in temperatures and ionization timescales behind the shock .
however , even in this model the best - fit ionization timescales of the si and s lines differ significantly such that no single component model can accurately fit both line complexes at once .
the high spatial resolution of the _ chandra _ data allow us to perform a spatially resolved spectral analysis of g349.7 + 0.2 to search for spectral variations across the snr .
we selected 6 extraction regions , shown in figure [ fig - chandra ] , positioned along the snr brightness gradient and each containing around 5000 counts .
spectra were re - grouped to include at least 25 counts per bin . as for the global snr spectrum ,
the spectrum used for background subtraction was extracted from the outer regions of the s3 chip . unlike the overall snr spectrum , spectra from the individual snr regions can be well fitted with single - component thermal models whose parameters vary from region to region ; the fit parameters are listed in table [ tab - fit ] and the spectra are shown in figure [ fig - spectra ] .
all the regions , except the eastern - most region ( region 1 ) , favor the vpshock model over the vraymond model ; this implies that most of the plasma in g349.7 + 0.2 has not yet reached ionization equilibrium and is still evolving .
furthermore , all the regions except the eastern - most region prefer the presence of enhanced si .
similarly , only the western - most region ( region 6 ) prefers , in addition to si , an enhanced abundance of s. for completeness , in table [ tab - fit ] we list the values of si and s abundances for all the regions .
plasma temperatures , ionization timescales and column densities measured from the individual regions and the global snr spectrum are plotted in figure [ fig - temp ] .
the temperatures derived from individual regions have values intermediate to those of the soft and hard spectral components derived from the global spectrum .
this suggests that the individual regions do not have sufficient statistics to distinguish between the two plasma components detected in the global snr spectrum , and represent an average of the two components summed along the line of sight .
the two eastern regions , regions 1 and 2 , appear to have a lower plasma temperature than the other regions , suggesting that the softer component dominates in the eastern snr region .
similarly , the fact that these two regions do not show the presence of fe - k line emission like the other regions suggests that plasma temperature in these two regions is lower than elsewhere in the remnant .
figure [ fig - temp ] also implies that there may be a small variation in ionization timescale and column density across the snr ( e.g. , regions 1 , 2 and 3 versus regions 4 and 6 ) . to investigate the spatial distribution of si and s line emission across the remnant on a finer scale than that of the spectral extraction regions , we produced continuum - subtracted ( cs ) and equivalent - width ( ew ) images ( see e.g. , * ? ? ?
* ) , shown in figure [ fig - ew ] .
narrow - band images used for deriving the cs and ew images were filtered in energy according to table [ tab - ew ] , and binned by 8 to provide a sufficient number of counts in the faint snr regions .
we found that the cs images for si and s are quite similar to the whole band image .
the ew images , which represent line strength compared to the continuum , do not show any spatial variation for the si and s lines , which appears inconsistent with the results obtained from spectral modeling , at least for s. however , the ew of the emission lines does not depend just on the total number of emitting atoms that are present in the plasma , but also on the temperature and ionization state of the plasma .
region 6 , where an enhanced abundance of s is derived from the spectral fit , is the same region where the ionization timescale is low , which also makes the line - to - continuum ratio lower than expected for an equilibrium plasma .
the net result is that the s ew is relatively uniform across the remnant .
this is confirmed by measuring ews directly from spectra of the two representative snr regions .
the derived ews for the si(s ) line in regions 2 and 6 are @xmath28(@xmath29 ) kev and @xmath30(@xmath31 ) kev , respectively . to derive mean electron density @xmath32 for the two plasma components , we approximate the x - ray emitting volume @xmath33 as a sphere 6.4 pc in radius , and denote @xmath34 and @xmath35 as filling factors for the soft and hard emission components , respectively .
the emission measure for each component ( @xmath36 ) is @xmath37 . assuming that the soft and hard thermal components are roughly in pressure equilibrium ( @xmath38 )
, we can derive their relative filling factors : @xmath39 .
here we assume that the plasma temperature we measure represents the temperature of all species ( electrons and ions ) , although the x - ray fits only provide the electron temperature . substituting in the fitted values from table [ tab - whole ] , we find that @xmath40 . however , in the case that the whole snr volume is not emitting x - rays , we set @xmath41 to account for the unknown volume filling factor .
the electron densities are then @xmath42@xmath8@xmath43 @xmath3 and @xmath44@xmath8@xmath43 @xmath3 . the x - ray emitting mass can be found from @xmath45 , which gives @xmath46@xmath19 mass in the soft component and @xmath47@xmath19 mass in the hard component .
@xcite showed that the global characteristics of a snr expanding into a density gradient will be similar to those of the sedov ( adiabatic ) snr , because there is a mutual compensation of the emission deficit from the low density regions by the enhanced emission from the high density regions .
the shock temperature is thus related to the plasma temperature as in the @xcite model : @xmath48 k , for @xmath49 kev ( using the temperature of the soft spectral component which is most probably the forward shock component , as suggested previously ) .
the corresponding shock velocity can be calculated from @xmath50 kms@xmath9 , where @xmath51 is the mean atomic weight .
the snr age is proportional to the shock radius and velocity , @xmath52 .
we adopt the snr radius of 1 , for which its dimensional size is @xmath53@xmath8pc , and derive an snr age of @xmath54@xmath8 yr .
this is in agreement with the upper limits derived from the ionization timescales of the two components of @xmath55 yr .
the explosion energy is approximately @xmath56@xmath8@xmath57 ergs , using the value of @xmath58 derived above .
this is a slightly lower energy than the canonical value of @xmath59 ergs , although the uncertainty on the density @xmath58 could modify this estimate somewhat .
the x - ray emission from the point source cxou j171801.0372617 found inside the snr shell is hard ( table [ tab - psou ] ) , and there is a possibility that this source is a neutron star associated with the snr . indeed , in our _ asca _ study we suggested that g349.7 + 0.2 could be the remnant of a core - collapsed supernova @xcite . because the total number of counts from the point source is too low for spectral analysis , we estimated the unabsorbed flux using the count rate ( @xmath60@xmath61 counts s@xmath9 ) , the effective area and response matrix files for acis - s .
neutron stars are found with a range of photon index values , with classical young pulsars having values between 1.11.7 ( e.g. * ? ? ?
* ) and more exotic objects like compact central objects ( ccos ) and anomalous x - ray pulsars ( axps ) having a softer photon index of @xmath62 and a blackbody component with temperature between 0.20.4 kev @xcite .
thus , assuming the same column density as for the snr ( 7@xmath5 @xmath6 ) and a power law photon index between 1.1 and 4.0 we obtained an unabsorbed integrated flux in the 0.510.0 kev range of @xmath63(0.4 - 4.1)@xmath64 ergs@xmath6s@xmath9 , which at a distance of 22 kpc gives an x - ray luminosity of @xmath65@xmath66 ergss@xmath9 .
this luminosity is lower than the values found from the other young pulsars , and is closer to the luminosities found from ccos , such as the one in snr g347.30.5 @xcite .
there is no obvious compact source at the location of our point source in the radio image , but some difuse emission is present at the 15 mjybeam@xmath9 level . using the number of sources as a function of flux density from the _ chandra _ multi - wavelength plane survey @xcite , we derive a chance probability of @xmath67 for finding an unrelated source with the x - ray brightness of @xmath68@xmath69 ergs@xmath6s@xmath9 ( assuming a power law photon index of 1.7 ) within the boundary of the snr .
thus , the x - ray luminosity , and lack of an optical and radio counterpart at the present sensitivity of the observations is consistent with a possibility that point source in g349.7 + 0.2 is an associated cco .
we also examined 21-cm hi emission towards the g349.7 + 0.2 region from the southern galactic plane survey ( sgps ) data with angular resolution of @xmath70 ( e.g. , * ? ? ? * ) .
the 21 cm spectrum toward the remnant shows very strong absorption at all negative velocities ( @xmath71 ) and at positive velocities up to + 8 kms@xmath9 .
this places the snr slightly beyond the solar circle on the far side . adopting the molecular cloud velocity of + 16 kms@xmath9 and using a flat rotation curve beyond the solar circle ,
the kinematic distance is 23.0 kpc , as found by @xcite .
the galactic center co survey of @xcite covers this region with beamwidth ( 1/8 ) sampling .
integrating the nearest co spectrum over the range @xmath72 to @xmath73 kms@xmath9 and using the co - to - mass conversion factor @xmath74 @xmath6k@xmath9 kms@xmath9 @xcite we obtain : @xmath75 the 21 cm velocity - integrated intensity varies very smoothly in longitude and latitude through the position of the remnant , and to obtain the atomic column density we used the average 21 cm spectrum over a small rectangle ( @xmath76 ) surrounding the remnant , but excluding a smaller region ( @xmath77 ) of absorption directly toward the remnant
. a lower limit on the atomic column density can be obtain by assuming the 21 cm emission is optically thin .
this yields : @xmath78 adopting a standard hi spin temperature of 140 k yields an atomic column density of @xmath79 therefore : @xmath80 which agrees well with the value derived from x - ray observations .
the x - ray morphology of g349.7 + 0.2 revealed with the _ chandra _ observations is strikingly similar to that of the radio image @xcite , shown as contours in figure [ fig - radio ] .
previous studies in the radio band have suggested that the structure of g349.7 + 0.2 is consistent with two overlapping rings , with the fainter , larger ring corresponding to the side where the snr shock is encountering a lower - density medium @xcite .
this interpretation is consistent with theoretical modeling of snr expansion in a medium with a density gradient ( e.g. , * ? ? ?
since the snr shock moves slowly in the dense cloud and faster in the lower density medium , two half shells with different radii are expected to be seen , with the shell radii depending on the initial location of the explosion and the steepness of the density gradient .
in addition , @xcite performed 2-d hydrodynamical modeling of the evolution and x - ray emission of snrs expanding in a large - scale density gradient , as seen by an observer under different viewing angles . compared to their model ,
the morphology of g349.7 + 0.2 is broadly consistent with a large scale density gradient seen by an observer at an angle of @xmath81 .
g349.7 + 0.2 shows some morphological resemblance to another galactic snr associated with oh(1720 mhz ) masers , ic 443 .
the two remnants are of similar age and spatial size , and radio and x - ray morphology of ic 443 is also consistent with expansion into a large scale density gradient @xcite produced by an hi cloud on the east and south - east side of the snr ( e.g. , * ? ? ?
* ) . in ic 443 ,
an interaction with a co cloud is occurring in the central region , where shocked molecular gas and oh(1720 mhz ) maser emission is found ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in figure
[ fig - mol]a we show the distribution of the molecular cloud associated with g349.7 + 0.2 , as traced by the co 10 emission @xcite , superimposed on the _ chandra _ image .
the molecular cloud interaction with g349.7 + 0.2 also appears to be limited to the central part of the remnant .
the location of oh(1720 mhz ) masers @xcite , also shown in figure [ fig - mol]a , point to the on - going interaction between the co cloud and the snr .
oh masers originate from shocked molecular clumps with density around @xmath12 @xmath3 and temperature of 50125 k , which have encountered the shock propagating transverse to the line of sight @xcite . shocked h@xmath82
2.12 @xmath83 emission has also been detected at the locations of the oh masers @xcite .
the velocity - integrated h@xmath82 10 s(1 ) emission is shown in figure [ fig - mol]b , superimposed on the _ chandra _ image .
the h@xmath82 line emission originates from the warm ( @xmath84 k ) gas just behind the part of the shock front expanding into the molecular cloud .
the h@xmath82 emission is also situated in the center of the snr where x - ray emission is very faint .
it is possible that due to high density in this region the snr shock became radiative and we see the transmitted shock through the molecular cloud , that radiates mostly in ir band ; mid - ir and far - ir emission has also been detected from this part of the snr @xcite .
thus , the density gradient responsible for the morphology of g349.7 + 0.2 is probably produced by hi clouds , as seen in ic 443 .
velocity - integrated ( + 14 to + 20 kms@xmath9 ) hiemission is shown in figure [ fig - co+hi ] overlaid with contours of co integrated over the same velocity range and the 18 cm radio image of the snr .
the overall large scale morphology of the atomic gas is consistent with that of the molecular gas .
the hi gas follows the distribution of the large co shell , with hi emission peaks not always coinciding with the co peaks , as is commonly found between atomic and molecular gas .
thus , hi material is present in the region towards g349.7 + 0.2 .
the density , size , and luminosity of g349.7 + 0.2 are similar to those proposed by @xcite for galactic snrs expanding into an intercloud medium , and to those of compact snrs in other galaxies @xcite .
these remnants are expanding into molecular clouds , but because of the clumpy structure of molecular clouds and the small filling factor of dense clumps , the remnant is mostly expanding into an intercloud medium with densities 525 @xmath3 @xcite . indeed , the molecular cloud associated with g349.7 + 0.2 has a density of @xmath68@xmath85 @xmath3 , whereas densities determined from x - ray emitting plasma of @xmath86 @xmath3 are more characteristics of an intercloud hi medium .
this high density of ambient medium results in a rapid snr evolution and a large amount of swept - up material producing very high luminosity ( @xmath87 ergss@xmath9 ) in a short period just before the snr becomes radiative and starts to cool rapidly .
such remnants can be recognized by a high radio surface brightness ( e.g. , ic 443 and w44 ; * ? ? ? * ) or a strong ir flux @xcite .
the x - ray observations of g349.7 + 0.2 imply that this snr is still dynamically young enough to produce a significant x - ray flux , in addition to strong radio flux .
the radiative part of the shell should be accompanied by a hi shell
. however , high - resolution observations towards g349.7 + 0.2 are needed to investigate the association between the hi emission and the snr shock .
the solar abundances and large x - ray emitting mass derived from _ asca _ observations imply that the x - ray emitting gas in g349.7 + 0.2 is dominated by swept - up interstellar material @xcite .
the implication is that , similar to the case of the snr n132d in the large magellanic cloud ( e.g. * ? ? ?
* ) , the sn ejecta component , expected from such a young snr , is largely hidden by the glow of massive amounts of swept - up material . in the case of n132d
, optical studies showed clear evidence of fast - moving ejecta synthesized in the collapse of a massive star . in x - rays , there was initially no detection of ejecta , but high resolution spectral studies indicated enhanced levels of oxygen @xcite , and the presence of an additional higher temperature spectral component which could be related to the ejecta @xcite . with the _ chandra _ observations
, we clearly detect enhanced abundances of si and possibly of s in g349.7 + 0.2 , which provides the first evidence for the presence of ejecta in this remnant .
similar evidence for the presence of ejecta in two middle - aged large magellanic cloud snrs , 054870.4 and 0534 - 69.9 , with substantial amount of swept - up material ( @xmath88@xmath19 ) has been presented by @xcite , who derive enhanced fe abundance based on the _ chandra _ spectra of these remnants .
the detection of the enhanced abundances only in the harder spectral component suggests that this component is associated with the sn ejecta , which means that the softer spectral component is related to the forward shock and the swept - up material . if the ambient medium into which the snr is expanding has much lower density than that of the ejecta , the forward shock will be heated to a significantly higher temperature than the ejecta ; this is the case in the young snr cas a ( e.g. , * ? ? ?
* ) . however , g349.7 + 0.2 is expanding into a dense medium and it is possible that the forward shock is cooler than the ejecta .
the assumed density gradient in the ambient medium could yield enhanced s in just the western - most region , as hinted at in the data ; the amount of swept - up material is lowest there because of the low ambient density , which enables the ejecta component to have a larger filling factor relative to the eastern regions where the ambient density is higher .
indeed , the fits to all the other regions ( 1 through 5 ) could accommodate a contribution from the plasma component of region 6 with an emission measure an order of magnitude lower than the components listed in table 3 .
the lower limit on the fractional mass of ejecta can be inferred from @xmath89 , where @xmath90 is the measured abundance , and @xmath91 is the standard solar mass fraction ; for si and s the mass fractions are 6.96@xmath61 and 4.79@xmath61 , respectively @xcite .
using the measured abundances from region 6 , we derive @xmath92@xmath61 and @xmath93@xmath61 .
the observed emission measure in region 6 implies a gas density of @xmath94 @xmath3 , which for an emission volume of @xmath1@xmath95 @xmath96 ( @xmath97 ) gives a total mass in the region of @xmath98@xmath19 .
the mass of the observed si and s ejecta is thus @xmath99@xmath19 and @xmath100@xmath19 , respectively .
this is a significant fraction of the amount produced by 1315@xmath19stars ( si : 0.0470.07@xmath19 , s : 0.0260.023@xmath19 ; * ? ? ?
oh(1720 mhz ) maser emission such as that associated with g349.7 + 0.2 and ic 443 , has been found in about 10% of the galactic snrs ( e.g. * ? ? ?
* ; * ? ? ?
seven of these maser - emitting ( me ; * ? ? ?
* ) snrs have been studied in detail in x - rays , and all of them are found to be of the mixed - morphology ( mm ; * ? ? ?
* ) class of snrs .
mm remnants , also known as thermal composites , are identified by shell emission in the radio band , and centrally brightened thermal emission in the x - ray band with little or no shell brightening .
the evolutionary characteristics which lead to mixed - morphology x - ray properties in snrs are not well understood , and different models have been suggested to explain individual cases .
one such model invokes evaporation of clouds which are left relatively intact after the passage of the snr blast wave and are evaporating in the snr interior @xcite .
another model explains the emission for the mixed morphology snr w44 @xcite as due to the effect of thermal conduction in the remnant interior .
another model for mixed morphology snrs invokes shell - like snr evolution at the edge of a molecular cloud , in which the ambient density gradient does not lie in the projection plane @xcite . a possible correlation between me and mm snrs has been suggested , because both types of snrs appear to be associated with molecular clouds @xcite .
the fact that some mm snrs do not have oh masers is not unexpected because very specific physical conditions are required for the maser excitation .
however , it has been suggested that all me snrs could be of the mm class @xcite .
g349.7 + 0.2 is then an example of a me snr which does not exhibit mm morphology .
its morphology , which is not centrally peaked , can be explained by the projection effect model used to explain mm snrs as well @xcite .
as shown in previous sections , we are likely seeing the snr expanding into a density gradient of an angle of @xmath81 . if we were to observe the expansion along the direction of the density gradient , we would see centrally peaked x - ray emission rather than two overlapping shells .
another example is the only me snr in the large magellanic cloud , n49 @xcite .
this remnant also has an irregular shell morphology in both radio and x - rays , with a bright region to the south - east ( e.g. , * ? ? ?
* ) , which can similarly be explained as resulting from expansion in a density gradient produced by the presence of a molecular cloud , as mapped by @xcite .
thus , if me and mm snrs are the same class of the snrs and there is a single mechanism to explain their appearance and spectral characteristics , the observations of g349.7 + 0.2 would support the projection model .
identifying more cases of snrs expanding into density gradients seen under different angles might provide further support for the projection model and contribute towards understanding the nature of mm snrs .
we present high - resolution _ chandra _ observations of a small - diameter snr , g349.7 + 0.2 .
its x - ray morphology is almost identical to that at radio wavelengths , and is consistent with expansion into a density gradient viewed at an angle of @xmath81 .
we investigated the atomic and molecular material in the vicinity of the snr , and suggest that the snr belongs to the class of snrs expanding into an intercloud medium discussed by chevalier ( 1999 ) .
the interaction with the molecular cloud is probably limited to the central portion of the snr , where the snr shock became radiative and the shocked material radiates mostly in the ir band @xcite .
g349.7 + 0.2 belongs to a class of maser - emitting snrs which have been suggested to be related to the mixed - morphology class of snrs .
g349.7 + 0.2 does not exhibit mixed - morphology , but its brightness distribution can be explained with the projection model for mixed - morphology snrs . the spectrum from the whole snr can be fitted with two thermal components , produced by distinct swept - up and ejecta plasma contribution . the softer component has a plasma temperature @xmath0 kev and is in ionization equilibrium , while the harder component has plasma temperature @xmath1 kev , ionization timescale @xmath101 @xmath3s and enhanced abundances of si .
the slightly enhanced s in just the western - most part of the snr could , if correct , be explained by a higher filling factor of the ejecta component relative to the swept - up component because of the lower density of the ambient medium in that region .
the column density to the snr of @xmath102@xmath5 @xmath6 derived from x - ray data agrees well with the value derived from hi and co data .
we find a point source with a small number of counts ( @xmath103 ) close , in projection , to the snr center .
using the column density value found towards the snr and the observed range of photon index values from neutron stars , we estimate the luminosity of the point source in the 0.510.0 kev band to be @xmath104@xmath8 ergss@xmath9 .
this luminosity and the lack of an optical and radio counterpart suggests that the x - ray point source could be the compact object formed by the explosion that also produced in g349.7 + 0.2 .
we thank cara rakowski for use of her shock model , john raymond , oleh petruk and sangwook park for helpful discussions , estela reynoso for the co image and bryan gaensler for the hi data .
this work was supported in part from _ chandra _ grants go2 - 3080a ( jsl ) and go2 - 3080b ( jph ) , and nasa contract nas 8 - 39073 ( pos ) . 1 & cxou j171816.2372556 & 17 18 16.19 & @xmath10537 25 55.84 & @xmath106 & 23 & @xmath107 & yes , @xmath22(mag)=10 + 2 & cxou j171807.3372707 & 17 18 07.32 & @xmath10537 27 07.66 & @xmath108 & 15 & @xmath109 & no + 3 & cxou j171805.7372437 & 17 18 05.67 & @xmath10537 24 37.41 & @xmath110 & 12 & @xmath111 & no + 4 & cxou j171801.0372617 & 17 18 01.00 & @xmath10537 26 16.62 & @xmath112 & 08 & @xmath113 & no + 5 & cxou j171746.1372506 & 17 17 46.10 & @xmath10537 25 05.90 & @xmath114 & 37 & @xmath115 & no + ll @xmath116 ( @xmath117 @xmath6 ) & 7.1@xmath118 + + @xmath119 ( kev ) & 0.76@xmath120 + @xmath121 ( @xmath122 @xmath3 ) & 9.9@xmath123 + + @xmath119 ( kev ) & 1.44@xmath124 + @xmath125 ( @xmath126 @xmath3s ) & 4.5@xmath127 + si & 2.1@xmath128 + @xmath121 ( @xmath122 @xmath3 ) & 2.9@xmath129 + + @xmath130 ( kev ) & 3.1 ( frozen ) + @xmath131 ( kev ) & @xmath132 ( frozen ) + @xmath133 ( phot @xmath6s@xmath134 ) & 2.9@xmath135@xmath136 + @xmath137/dof & 1.22/339 + + @xmath138 ( ergs@xmath6s@xmath9 ) & 6.5@xmath139 + @xmath140 ( ergs@xmath6s@xmath9 ) & 1.7@xmath139 + @xmath116 ( @xmath117 @xmath6 ) & 6.7@xmath141 & 6.7@xmath141 & 6.5@xmath142 & 7.8@xmath143 & 7.2@xmath144 & 7.8@xmath145 + @xmath119 ( kev ) & 0.92@xmath146 & 0.94@xmath147 & 1.28@xmath148 & 1.28@xmath149 & 1.13@xmath150 & 1.18@xmath151 + si & 1.4@xmath128 & 1.6@xmath141 & 1.7@xmath142 & 1.6@xmath152 & 1.6@xmath152 & 2.4@xmath153 + s & 1.2@xmath154 & 0.9@xmath155 & 0.8@xmath118 & 1.2@xmath156 & 1.2@xmath154 & 1.5@xmath128 + @xmath125 ( @xmath3s ) & cie & 1.2@xmath157 & 1.1@xmath158 & 3.6@xmath159 & 8.0@xmath160 & 3.8@xmath161 + @xmath121 ( @xmath122 @xmath3 ) & 1.2@xmath156 & 1.2@xmath154 & 0.6@xmath162 & 0.8@xmath118 & 0.9@xmath162 & 0.9@xmath118 + @xmath137/dof & 1.12/134 & 1.02/130 & 1.19/137 & 1.27/143 & 0.88/143 & 0.97/145 + and residuals from the best - fit models . fits parameters are listed in table [ tab - fit].,title="fig:",height=264 ] and residuals from the best - fit models .
fits parameters are listed in table [ tab - fit].,title="fig:",height=264 ] + and residuals from the best - fit models .
fits parameters are listed in table [ tab - fit].,title="fig:",height=264 ] and residuals from the best - fit models .
fits parameters are listed in table [ tab - fit].,title="fig:",height=264 ] + and residuals from the best - fit models .
fits parameters are listed in table [ tab - fit].,title="fig:",height=264 ] and residuals from the best - fit models .
fits parameters are listed in table [ tab - fit].,title="fig:",height=264 ] + for the cs images and in kev for the ew images .
the images are overlaid with contours from the 18 kev band image binned in the same manner .
contour levels are : 70 , 180 , 300 , 500 , 700 and 900 counts arcmin@xmath9.,title="fig:",height=188 ] for the cs images and in kev for the ew images .
the images are overlaid with contours from the 18 kev band image binned in the same manner .
contour levels are : 70 , 180 , 300 , 500 , 700 and 900 counts arcmin@xmath9.,title="fig:",height=188 ] for the cs images and in kev for the ew images .
the images are overlaid with contours from the 18 kev band image binned in the same manner .
contour levels are : 70 , 180 , 300 , 500 , 700 and 900 counts arcmin@xmath9.,title="fig:",height=188 ] for the cs images and in kev for the ew images .
the images are overlaid with contours from the 18 kev band image binned in the same manner .
contour levels are : 70 , 180 , 300 , 500 , 700 and 900 counts arcmin@xmath9.,title="fig:",height=188 ]
|
we present high - resolution x - ray observations of the small - diameter supernova remnant ( snr ) g349.7 + 0.2 with the _ chandra x - ray observatory_. the overall snr spectrum can be described by two spectral components .
the soft component is in ionization equilibrium and has a temperature of @xmath0 kev ; the hard spectral component has a temperature of @xmath1 kev , an ionization timescale of @xmath2 @xmath3s and enhanced abundances of si .
the spatially resolved spectral modeling shows that s may also be enhanced , at least in some regions .
the enhanced abundances clearly point to the presence of an ejecta component in this remnant . using the available hi and co data towards g349.7
+ 0.2 we derive a column density of @xmath4@xmath5 @xmath6 along the line of sight to the snr , which is consistent with our x - ray data .
the x - ray morphology of g349.7 + 0.2 is strikingly similar to that at radio wavelengths an irregular shell with a brighter eastern side which is consistent with expansion in a medium with a large scale density gradient .
the remnant is known to be interacting with a molecular cloud ( from the presence of oh(1720 mhz ) masers ) , but this interaction is probably limited to the central portion of the snr , as seen in snr ic 443 .
we found that hiclouds are present in the snr region , which supports the notion that g349.7 + 0.2 belongs to a class of remnants evolving in the intercloud medium ( such are ic 443 and w44 ; * ? ? ?
* ) , which is also responsible for the remnant s morphology .
g349.7 + 0.2 does not have the mixed - morphology found for other maser - emitting snrs studied to date in x - rays , but its morphology can be explained by a projection model for mixed - morphology snrs .
we have identified a point source close to the center of the snr with a luminosity of @xmath7@xmath8 ergss@xmath9 , which is consistent with that of the compact central objects found in a few other galactic snrs .
|
over the last decades various interesting effects related to long - range spiral - like forms of magnetic order have been revealed in multiple fields of condensed matter physics .
recent examples include multiferroic compounds such as tbmno@xmath0 @xcite , magnetic surfaces @xcite that are interesting for applications in spintronics , e.g. the construction of a spin field effect transistor@xcite , magnetic insulators such as ba@xmath6cuge@xmath6o@xmath7 @xcite or itinerant magnets such as mnsi@xcite .
further , helimagnetism is also extensively discussed theoretically , here examples encompass the proposal of a chiral universality class @xcite , non - centrosymmetric superconductors @xcite and new helical goldstone modes , so - called helimagnons@xcite .
the latter have recently also been observed experimentally@xcite .
+ spiral magnetic order that in addition exhibits chirality , i.e. it breaks the spatial inversion symmetry , may exist in a right- and left - handed version that are interrelated via the inversion operation .
as long as the underlying chemical structure is centrosymmetric both versions are energetically degenerate and right- and left - handed domains should be observed in equal fractions@xcite .
however , in case the chemical structure is non - centrosymmetric itself , the magnetic structure is expected to show a single handedness .
since 65 of the known 230 crystallographic space groups are non - centrosymmetric ( sohncke groups)@xcite one would expect a great number of chiral magnetic compounds with a single chirality .
yet , only a few such compounds have been observed so far , with the most prominent example being mnsi .
mnsi exhibits many interesting physical properties such as an extended non - fermi liquid phase @xcite probably associated with a new kind of metallic state @xcite , a magnetic skyrmion lattice @xcite and the observation of helimagnons@xcite , that are all closely related to its magnetic chirality@xcite .
more recently monochirality was also observed in the non - centrosymmetric compound ba@xmath0nbfe@xmath0si@xmath6o@xmath8 that exhibits the interesting coexistence of two forms of chiral magnetism , within triangles in the basal plane and and along the @xmath9-axis as a magnetic helix@xcite .
a further example is uptge that , however , displays a cycloid magnetic structure without true chirality , but with a single turning sense of the cycloid , that was also explained in terms of its non - centrosymmetric crystal structure@xcite .
+ in this work we demonstrate that ndfe@xmath0(@xmath1bo@xmath0)@xmath2 is a new compound that displays a chiral magnetic structure with a single chirality . below a magnetic commensurate ( c ) to incommensurate ( ic ) phase transition at @xmath1013.5
k the collinear antiferromagnetic structure of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 transforms into an antiferromagnetic helix that exists with a single chiral domain .
furthermore , the observation of third - order harmonics below @xmath11 suggest that the emergence of the helical magnetic structure is accompanied by the formation of a magnetic soliton lattice below @xmath11
. + ndfe@xmath0(@xmath1bo@xmath0)@xmath2 belongs to the family of borates rm@xmath0(bo@xmath0)@xmath2 ( r = y , la - lu , m = al , ga , cr , fe , sc ) . these borates are interesting in their own right and have previously been studied mainly due to their special optical properties .
rare - earth ions , in general , and nd@xmath12 , in particular , have excellent characteristics to generate infrared laser action and to serve in nonlinear optics@xcite .
the subfamily of ferroborates ( m = fe ) is equally interesting with respect to their magnetic properties due to competing magnetic sublattices . here
gdfe@xmath0(bo@xmath0)@xmath2 and ndfe@xmath0(bo@xmath0)@xmath2 are especially fascinating as both materials show a large magneto - electric effect @xcite .
+ the ferroborates rfe@xmath0(bo@xmath0)@xmath2 crystallize in the trigonal space group _ r32 _ ( group no .
155 ) , that is they belong to the structural type of the mineral huntite camg@xmath0(bo@xmath0)@xmath2 @xcite .
note that this structure is missing a center of inversion . in our previous work @xcite
we performed an unpolarized neutron diffraction study with both powder and single crystal samples in order to investigate the magnetic structure of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 .
we demonstrated that ndfe@xmath0(@xmath1bo@xmath0)@xmath2 exhibits long - range antiferromagnetic order with the magnetic propagation vector @xmath13 $ ] below @xmath14 30 k. however , our combined magnetic representational and rietveld analysis@xcite yields different magnetic structures that explain the data equally well ( s. fig . [ fig : old_mag_struc ] ) .
all models have in common that the magnetic moments of all three fe sublattices and the nd sublattice are parallel to the hexagonal basal plane and are coupled antiferromagetically in adjacent planes .
this is also in agreement with easy - plane type magnetic anisotropy that was observed in several studies of the magnetic susceptibility @xcite . for a more detailed description
we refer to refs . .
in addition the study revealed that below approximately t = 19 k the magnetic structure becomes incommensurate with the magnetic propagation vector @xmath15 $ ] .
however , since the study was performed with thermal neutrons only , the experimental resolution did not allow for an exact study of the propagation vector as a function of temperature or the nature of the ic magnetic structure .
+ the purpose of this study is mainly to identify the correct magnetic structure in both the c and ic magnetic phases . for this task
we performed a detailed neutron diffraction study with both unpolarized neutrons and full spherical neutron polarimetry applied .
further high resolution diffraction was carried out to study the temperature dependence of the c - ic phase transition with better resolution .
the unpolarized neutron single crystal diffraction data were additionally used to verify the low - temperature chemical structure of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 , mostly as previous results indicated that the overall chemical symmetry might only be _ r3 _ @xcite .
+ finally all our magnetic single neutron diffraction results were evaluated by a comparative analysis with both the standard fullprof package@xcite and jana2006@xcite .
the latter was only recently extended for the purpose of analysis of magnetic neutron scattering data and emphasizes the use of magnetic symmetries .
in this study two different samples of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 were studied .
a large single crystal of approximate dimensions 8 x 8 x 8 mm@xmath16 that has been already used in our previous work@xcite ( sample1 ) was reinvestigated .
further , a new smaller sample of the size 3 x 5 x 4mm@xmath16 was prepared for additional experiments ( sample2 ) .
the process that was used for the preparation of the studied crystals is already described in detail in our previous publication@xcite . +
in order to reduce strong extinction effects , diffraction measurements were carried out on the smaller sample2 of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 , in a four - circle setup on the thermal neutron diffractometer trics @xcite at the continuous swiss spallation neutron source sinq@xcite and on the heidi hot neutron diffractometer@xcite , situated at the frm ii reactor at munich .
the corresponding neutron@xcite wavelengths were 1.18 and 0.55 , respectively .
the data evaluations were performed by means of current versions of fullprof @xcite and jana2006 @xcite . for the refinements the low - temperature lattice parameters a = 9.594 and c = 7.603 of ref . were employed .
absorption corrections were neglected .
nuclear and magnetic extinction may be quite different , cf . e.g. @xcite .
we will show that in case of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 such a difference is almost not significant for hot neutrons , but turned out to be important for thermal neutrons .
in contrast to fullprof , jana2006 permits different types of isotropic extinction corrections .
however , the corresponding differences were not found to be essential in case of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 and thus type ii extinction@xcite was used . in order to limit the number of parameters , only isotropic temperature factors were refined .
+ the recent extension of the jana2006 refinement program allows commensurate and incommensurate magnetic structures to be described by four - dimensional magnetic superspace groups analogous to occupationally modulated chemical structures@xcite . a more detailed discussion of these aspects of the refinement is given in the appendix together with a comparison to the corresponding fullprof results .
+ further , sample1 was investigated in a high resolution diffraction experiment carried out on the triple - axis spectrometer tasp@xcite , situated at the end position of a cold super mirror guide of sinq .
the spectrometer was operated in its elastic mode with fixed incident and final wave vector k@xmath17 = 1.2 @xmath18 .
additionally 20 soller collimators were installed in the incident beam , in front of the analyzer and the detector .
the second order contamination was removed from the beam by means of a beryllium filter that was inserted between the sample and the analyzer .
the use of a triple - axis spectrometer for diffraction experiments is justified by the excellent signal - to - noise ratio that is achieved by the use of an additional analyzer crystal .
this experimental setup only allows access to bragg reflections in a single scattering plane . for this measurement
the single crystal was oriented with the reciprocal axis b@xmath19(k ) and c@xmath19(l ) within the scattering plane .
+ in order to perform full polarization analysis on ndfe@xmath0(@xmath1bo@xmath0)@xmath2 , the spherical neutron polarimetry ( snp ) option mupad @xcite available at sinq was mounted on tasp .
the neutron beam was polarized and analyzed via two polarizing supermirror benders that were installed after the monochromator and in front of the analyzer , respectively .
a final wave vector k@xmath17 = 1.97 @xmath18 was chosen to maximize both the intensity and the polarization of the neutron beam .
no additional filter for second order suppression was used because the benders already act as such .
the orientation of the crystal was identical to the unpolarized measurements .
+ only the combination of unpolarized and polarized neutron diffraction measurements allowed us to find the model for magnetic structure that gives the best agreement with all data collected during the course of this investigation .
however , for the sake of clarity we will describe the unpolarized and the snp measurements in separate sections , referring to the corresponding other sections when necessary .
the low temperature chemical structure was verified by measuring extended sets of nuclear neutron intensities on heidi at 22 k ( @xmath20 = 0.55 , 287 bragg peaks ; data set n1 ) and on trics at 6 k ( @xmath20 = 1.18 , 151 bragg peaks ; data set n2 ) . for the fits of both data sets with the space group _ r32 _ we refined 11 parameters with two isotropic temperature factors ( one for the heavy atoms nd and fe and one for the light atoms ) and one isotropic extinction parameter .
the resulting structural parameters are summarized in table [ tab : structure ] together with the agreement factors .
note , that for the data set ( n1 ) we performed fits with both fullprof and jana2006 .
this is discussed in more detail in the appendix .
+ our findings confirm that also at low temperatures the chemical structure of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 is well described within the space group _
r32_. the positional parameters derived from the two data sets are in good agreement .
they are close to the room - temperature values published in our previous work@xcite .
in contrast to the trics measurement , but in agreement with previous powder investigations @xcite , the heidi single crystal refinement tended for the isotropic temperature factor of nd to negative values . compared to the hot neutron results , the larger temperature factor values obtained from the thermal neutron data are presumably due to the neglected larger absorption in the latter case .
refinement of occupation factors did not indicate essential deviations from stoichiometry . by means of careful cold neutron scattering experiments on the triple - axis spectrometer tasp carried out on the bigger sample1 , we investigated the temperature dependence of the magnetic c - ic transition .
elastic @xmath21-scans along the reciprocal l - direction around the magnetic bragg reflection ( 0,0,@xmath22 ) as a function of temperature were used to determine the transition temperature as illustrated in fig .
[ fig : k_tempdep](a ) .
the propagation vector is commensurate down to the temperature t@xmath5
@xmath4 13.5 k below which the magnetic bragg reflection splits into two incommensurate satellite peaks .
here we determined t@xmath5 as the temperature where the maximum intensity of the commensurate magnetic reflection is reached and then starts to decrease as the reflection splits up into the two incommensurate satellite peaks . below t@xmath5 the scans show that @xmath23 changes continuously at the c - ic phase transition in ndfe@xmath0(@xmath1bo@xmath0)@xmath2 and can be well described via @xmath24 where @xmath25 this is demonstrated by the black solid line in fig .
[ fig : k_tempdep](a ) .
the splitting of the magnetic bragg reflections is only observed along the @xmath26-direction down to t = 1.6 k as shown by the intensity map around the ( 0,0,@xmath22 ) provided in the inset of fig .
[ fig : k_tempdep](a ) .
the splitting at t = 1.6 k is @xmath27 = 0.00667 .
+ in addition to the principal magnetic satellites also reflections at higher order harmonics were observed as shown in fig .
[ fig : nd_second_order_sats ] .
their shift with respect to @xmath28 was found to be 3@xmath27 = 0.02 .
the integrated intensity of the third order harmonics is largest near the transition temperature t@xmath5 and decreases fast as a function of decreasing temperature .
we note that the integrated intensity of the principle satellites remains approximately constant down the lowest temperature as would be expected for a magnetic soliton lattice @xcite .
+ investigations performed on trics with thermal neutrons on the smaller sample2 of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 clearly show broadening of magnetic peaks at temperatures below 12.5 k , confirming the existence c - ic magnetic phase transition in ndfe@xmath0(@xmath1bo@xmath0)@xmath2 .
this is illustrated in fig .
[ fig : k_tempdep](b ) .
however , a splitting of the ( -1,0,0.5 ) peak as observed before on the sample1@xcite , could not be reproduced under similar conditions or on the heidi neutron diffractometer .
this indicates a certain sample dependence of the k - vector magnitude in the incommensurate phase . at the lowest temperature @xmath23
may be approximated as 1.502 for sample2 .
+ corresponding to the results in the preceding section , we assumed commensurate and incommensurate magnetic ordering for temperatures above and below t@xmath5 , respectively , for the fits of the data discussed in the following .
moreover , we used the structural parameters from ( n1 ) ( heidi results ) .
+ the spherical neutron polarimetry experiments that will be discussed in section [ snp ] showed that only the model ( m1 ) for the magnetic structure is able to explain the data correctly .
therefore , we will focus on the fits performed for this model here . as shown in refs . for ( m1 )
the magnetic structure can be conveniently expressed as spirals of the form @xmath29 , \end{split } \label{eq : nd_fe_spiral}\ ] ] that propagate along the hexagonal @xmath9-axis for all three magnetic fe moments and the magnetic nd moment ( @xmath30 ) and also for both the ( c ) and ( ic ) magnetic phases . here
@xmath31 and @xmath32 are unit vectors of the hexagonal lattice and @xmath33 describes the polar takeoff angle from the hexagonal @xmath34 axis . + in the incommensurate case we superimposed @xmath35 and @xmath36 satellites in each magnetic peak , as here @xmath37 are inequivalent .
further , we note , that for such spirals the magnetic neutron intensity is modulated with the factor @xmath38 , where @xmath39 represents the angle of the magnetic scattering vector to the spiral axis@xcite .
the direction of the three parallel magnetic fe moments , that is described via the polar angle @xmath40 in our model , can not be determined from the present unpolarized neutron diffraction data . however , recent bulk magnetic measurements of tristan et al .
@xcite on a single crystal of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 at low temperatures show easy magnetization along the @xmath34-axis and we therefore fixed the value of the polar angle to @xmath40 = 0 . in our previous work@xcite we found a small angle between magnetic moments of the fe and nd ions .
hence , we performed fits in two configurations , respectively , where we either kept the nd magnetic moment parallel to the fe moments ( model ( m1a ) ) or allowed for a refinable angle @xmath41 between the fe and nd magnetic moments ( model m1(b ) ) . starting with a single magnetic domain and using the nuclear isotropic extinction parameters , thus 2 or 3 parameters
were refined ( ( m1a ) and ( m1b ) , respectively ) . in principle
up to six magnetic orientation domains would be possible according to the threefold- and twofold rotation axes of the paramagnetic space group _ r32 _ in the latter case : @xmath42 ; @xmath43 ; @xmath44 ; @xmath45 ; @xmath46 ; @xmath47 in direct space notation . for the incommensurate case only the three threefold rotations around the crystallographic @xmath9-axis are present@xcite . however , the corresponding fullprof calculation did not yield improved fits .
this can be understood by considering that the incommensurate magnetic spiral described by eq
. breaks the translation symmetry of the underlying chemical lattice along the c direction .
the broken translation symmetry implies that any rotation of the spiral around the @xmath9-axis can be compensated by a translation and the spin structure is conserved .
the additional fits by means of jana2006 gave identical results .
in particular the superspace description for the magnetic structure in jana2006 gives a natural explanation concerning the single populated magnetic orientation domain(s .
appendix and ref . ) .
+ in the commensurate case with @xmath13 $ ] , @xmath48 is equivalent to @xmath49 and this implies that the magnetic moments in adjacent planes are arranged antiparallel . here
the refinement according to model ( m1a ) yield three in good approximation equally distributed magnetic domains according to the threefold rotation axes .
+ at various temperatures sets of 50 magnetic and 5 nuclear bragg peaks were measured on heidi .
corresponding characteristic refinement results obtained by both fullprof and jana2006 are given in the appendix and are discussed subsequently . + for the commensurate magnetic phase of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 we may conclude @xmath40 = @xmath41 = 0 according to model ( m1a ) .
we note our previously published powder neutron diffraction data@xcite may be equally well fitted with @xmath40 = @xmath41 = 0 ( cf .
ref . additionally ) .
model ( m1a ) is furthermore supported by the spherical neutron polarimetry results that will be discussed in section [ snp ] .
+ with respect to incommensurate magnetic ordering in ndfe@xmath0(@xmath1bo@xmath0)@xmath2 , the thermal neutron data at 6 k indicate almost significance for the introduction of a non - zero angle @xmath41(m1b ) . on the other hand ,
the difference between the two different heidi refinements is considerably smaller , due to less important extinction corrections at shorter neutron wavelength .
thus the simpler magnetic structure with @xmath40 = @xmath41 = 0 ( m1a ) holds most probably also in the incommensurate phase of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 ( s. appendix for details ) .
+ the resulting temperature dependencies of the ordered magnetic moments of the fe@xmath12 and nd@xmath12 ions in ndfe@xmath0(@xmath1bo@xmath0)@xmath2 are shown in fig .
[ fig : magmom ] . the deviations between the magnetic moments derived from the 5.5 k trics data and those from heidi at 5 k are presumably due to the larger extinction corrections in case of thermal neutrons .
the magnetic moments of both fe and nd seem to vary smoothly as a function of temperature , also at the c - ic magnetic phase transition at approximately t@xmath5 @xmath4 13.5 k.
snp measurements were performed above and below the c - ic transition , respectively . in the ic phase of ndfe@xmath0(@xmath1bo@xmath0)@xmath2
an excellent @xmath21-resolution is necessary to observe the small splitting ( @xmath27 = 0.00667 ( r.l.u . )
@xmath50 0.0055@xmath18 ) of the magnetic bragg reflections .
for the small final wave vectors k@xmath17 that are required for such a high resolution setup , the polarizing benders used at tasp perform non - ideal with respect to transmission and polarizing efficiency . the best tradeoff between resolution and transmission / polarization
is reached when tasp is operated with k@xmath17 = 1.97 @xmath18 .
however , the relatively moderate @xmath21-resolution associated with k@xmath17 = 1.97 @xmath18 is not sufficient to perform separate snp measurements on the magnetic satellites @xmath37 and consequently a superposition of intensities from both satellites will be observed .
nevertheless important information could be extracted by performing polarized @xmath21-scans over the magnetic satellites .
+ we first discuss the results obtained by conventional snp measurements in the commensurate phase of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 in section [ snp_c ] , the findings for the incommensurate phase are described in section section [ snp_ic ] . the final polarization vector after the scattering process at the sample is described by the blume - maleyev - equations@xcite that may be given in the following convenient form @xmath51 where @xmath52 is the polarization tensor , which describes the rotation of the initial polarization vector @xmath53 in the scattering process and @xmath54 is the polarization created in the scattering process at the sample .
for purely magnetic reflections , like the ones observed in ndfe@xmath0(@xmath1bo@xmath0)@xmath2 , @xmath55 and @xmath54 are reduced to @xmath56 here @xmath57 is the magnetic interaction vector defined as @xmath58 , where @xmath59 is the fourier transform of the magnetization density @xmath60 of the investigated sample and @xmath61 is a unit vector parallel to the scattering vector @xmath62 .
the set of polarization axes is defined to have @xmath63 parallel to @xmath62 , @xmath26 perpendicular to the scattering plane and @xmath64 completing the right - handed set .
we note , that the term @xmath65 is only non - zero for magnetic structures that display chirality and is therefore often denoted as the chiral term . finally , the measured quantity is the polarization matrix , namely the components of the final polarization vector after the scattering process for all three directions of the incident beam polarization , @xmath66 where @xmath67 and @xmath68 ( @xmath69 ) denote the directions of the incident and final polarization vectors , respectively .
+ for the commensurate magnetic phase of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 we measured the polarization matrix at six magnetic bragg reflections and at t = 20 , 25 and 30 k. the measured polarization matrices proved to be independent of temperature ( within the error bars ) for t @xmath70 20 k. thus , only the data for t = 20 k will be discussed in the following .
the corresponding matrices are provided in table [ tab : polartensors ] . from the measured polarization
matrices several constraints on the magnetic structure can be derived : 1 . on
all measured bragg peaks the elements @xmath71 and @xmath72 are equal to zero .
this implies that the magnetic structure is not chiral at all or that it is a chiral structure with equally populated chiral domains . 2 . for the magnetic reflection ( 0,0,-1.5 )
the scattering vector @xmath21 is directed parallel to the crystallographic c direction .
hence , the magnetic interaction vector only contains components in the basal plane . @xmath73
= @xmath74 . here
the reduction from @xmath75 is due the polarizing benders which have a non - ideal polarization efficiency of approximately 0.966 .
thus , @xmath76 is fully polarized whereas @xmath77 and @xmath78 are fully depolarized .
this would foremost lead to the assumption @xmath79 but as the elements @xmath80 and @xmath81 are also equal to zero it also suggests the presence of spin domains in the basal plane .
3 . on the magnetic reflection ( 0,4,0.5 )
the scattering vector @xmath21 is approximately parallel to the reciprocal b@xmath82 axis .
as the z axis which is perpendicular to the scattering plane lies always within the basal hexagonal plane in the chosen scattering geometry the y - axis is approximately parallel to the crystallographic @xmath9-axis ( @xmath83 ) .
the polarization tensor shows @xmath84 which indicates that @xmath85 0 and hence the magnetic interaction vector is directed along z. therefore the magnetic moments are confined in the basal plane .
we note that these constraints are satisfied by both possible magnetic structures ( m1a / b ) and ( m2 ) . in order to calculate the expected polarization matrices for both models
, we used lattice constants and structural parameters from table [ tab : structure ] .
the magnitude of the magnetic moments for the fe and nd ions were set to the values as obtained from the unpolarized diffraction data .
since snp is generally insensitive to absolute moment sizes in case of pure magnetic reflections they were fixed in subsequents fits .
both models fail to explain the observed polarization matrices when no orientation domains were considered . introducing the three or six orientation domains with statistical population in the calculation for models ( m1 ) and ( m2 ) , respectively ( cf .
appendix ) , resulted in polarization matrices given in table [ tab : polartensors ] .
further , we considered the non - ideal polarization efficiency of the used polarizing benders in the calculation . +
as demonstrated in table [ tab : polartensors ] the magnetic model ( m2 ) was not able to explain the measured polarization tensors on the two magnetic reflections ( 0 , 1 , -2.5 ) and ( 0 , 1 , 0.5 ) as indicated by the bold entries in table [ tab : polartensors](@xmath86 = 9.5 ) .
the model introduces small chiral contributions ( @xmath71 and @xmath72 elements of the tensors ) due to its slightly canted spins ( cf fig .
[ fig : old_mag_struc](b ) ) , that are not observed in the experiment . performing a fit from this starting values did not result in a better agreement between model and data , as the fit diverged .
hence , model ( m2 ) can be excluded .
+ for the magnetic model ( m1 ) we fixed @xmath40 = 0 and left @xmath41 free for the fits , thus corresponding to ( m1b ) .
the fit converges to a solution with @xmath41 = 0 ( @xmath86 = 3.8 ) which corresponds to ( m1a ) .
attempts to additionally determine the orientation of the fe magnetic moments via the angle @xmath40 gave no conclusive results .
the angle @xmath40 describes the absolute orientation of the magnetic moments in the crystallographic ab - plane with respect to chemical structure .
thus , the indeterminacy of @xmath40 is most probably related to the presence of the three orientation domains in the hexagonal basal plane .
+ in summary our snp data for the commensurate phase is best explained by the model ( m1a ) when @xmath40 = @xmath41 = 0 which is also in agreement with the unpolarized neutron single crystal diffraction results described in section [ unpol_magstruc ] .
the corresponding magnetic structure is illustrated in fig .
[ fig : old_mag_struc](a ) . + fig .
[ fig : chiral](a ) shows polarized constant - energy - scans along the l - direction performed around the reciprocal space position ( 0 , 0 , -1.5 ) at t = 1.5 k in the ic - phase .
four different polarization channels were measured , namely @xmath87 , @xmath88 , @xmath89 and @xmath90 .
the scans for the channels @xmath88 and @xmath89 appear to be slightly shifted towards the positions of the magnetic satellite reflections ( 0,0,-3)+@xmath91 ( @xmath88 ) and ( 0,0,0)-@xmath91 ( @xmath89 ) , respectively , in agreement with the observations made in the high resolution measurement in section [ highres ] .
+ the effect is more significant in the sum of the four terms given by @xmath92 that is equivalent of measuring the difference of the two polarized neutron cross - sections with the initial polarization vector @xmath93 directed parallel and antiparallel to the @xmath63 direction ( thus parallel or antiparallel to @xmath21 ) .
the resulting scan is shown in fig .
[ fig : chiral](b ) . + by means of eqns . and
we further obtain the result @xmath94 demonstrating that @xmath95 is proportional to the chiral term .
thus , by carrying out the polarized scans we were able to separate the resolution limited peak at ( 0,0,-1.5 ) in two satellites ( 0,0,0)-@xmath91 ( i@xmath96 ) and ( 0,0,-3)+@xmath91 ( i@xmath97 ) . since two distinct extrema with opposite signs for the two satellites can be distinguished the collinear antiferromagnetic structure of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 observed in the commensurate phase appears to transform to an antiferromagnetic long - period spiral below t@xmath5 .
in addition the non - zero chiral term also suggests that at least unequally populated chiral domains are present in ndfe@xmath0(@xmath1bo@xmath0)@xmath2 .
+ in order to investigate this in more detail we performed calculations based on the magnetic model ( m1a ) , however together with the incommensurate propagation vector @xmath91 that leads to a small rotation of the magnetic moments between neighboring layers .
we calculated the integrated intensities for the four measured polarization channels @xmath87 , @xmath88 , @xmath89 and @xmath90 for both satellites ( 0,0,0)-@xmath91 and ( 0,0,-3)+@xmath91 , where we used @xmath27 = 0.0667 in @xmath91 = [ 0 , 0 , @xmath22 + @xmath27 ] as deduced from our high resolution diffraction data .
the calculated integrated intensities are given in table [ tab : chiral_scan_ints](a ) for the case of only a single chirality domain being populated .
in addition we assumed only one orientation domain , similar as for the unpolarized results .
for comparison we give the measured integrated intensities from the scans for the four polarization channels in table [ tab : chiral_scan_ints](b ) . due to the limited resolution of the setup the measured values for the individual polarization channels
are integrated over both peaks and can not be compared directly to the calculations .
this is especially true for the difference of the intensities .
however , two reasonable assumptions can be made to allow for a comparison : 1 .
the measured integrated intensity in the channel ` @xmath99 ' is only due to the peak at ( 0,0,0)-@xmath91 whereas the channel ` @xmath100 ' is only due to the peak at ( 0,0,-3)+@xmath91 . this assumption is justified as the measured points of each respective polarization channel ( red squares and blue circles in fig .
[ fig : chiral](a ) ) are shifted towards the direction of the corresponding peak .
the channels ` @xmath101 ' and ` @xmath102 ' are equally contributed by both satellite reflections as they are centered on top of each other and between the other two channels ( black circles and green triangles in fig .
[ fig : chiral](a ) ) . based on these assumptions we are able to calculate the ratio @xmath103 = 28(4 ) for the intensities measured on the magnetic bragg reflection ( 0,0,-3)+@xmath91 . note that the division by 2 is due to assumption ( ii ) .
the same ratio for the calculated intensities for this peak amounts to @xmath104 = 24 . for the satellite ( 0,0,0)-@xmath91 we obtain the measured ratio @xmath105 = 28(4 ) and the calculated ratio @xmath106 = 24
this indicates that our model of a long - period antiferromagnetic helix propagating along the hexagonal @xmath9-axis with the magnetic moments parallel to the hexagonal basal plane and single chirality domain is in good agreement with our data .
+ in addition we verified our assumptions by performing a convolution of the calculated integrated intensities for each of the four polarization channels ( cf .
table [ tab : chiral_scan_ints ] ) with the four - dimensional resolution function of the spectrometer in a simulation .
the results of the simulations are the solid lines in fig .
[ fig : chiral](a ) . to match the intensity a single scale factor four all four polarization channels was introduced in the calculation .
similar as for the previous calculations the simulation was performed with only one of the two chiral domains being populated , and a single orientation domain .
the solid line in fig . [ fig : chiral](b ) was obtained by summating the individual curves for each polarization channel with respect to eq . .
the peculiar shape of difference curve is due to the slightly asymmetric shape of the magnetic satellite peaks .
after the asymmetric peak shape was taken account for within the simulation good agreement between simulation and experimental data was achieved .
the present single - crystal neutron diffraction investigations did not detect significant deviations from space group _ r32 _ concerning the chemical structure of multiferroic ndfe@xmath0(@xmath1bo@xmath0)@xmath2 at low temperatures .
with respect to magnetic ordering this study shows that only the magnetic model ( m1a ) is in agreement with our polarized neutron data on the noncentrosymmetric ndfe@xmath0(@xmath1bo@xmath0)@xmath2 .
+ however , in contrast to the previous neutron powder diffraction study of ndfe@xmath0(@xmath1bo@xmath0)@xmath2@xcite , we may conclude from our combined investigation with unpolarized and polarized neutrons that in addition , ferromagnetic alignment of the magnetic fe and nd sublattices holds .
this implies low magnetic symmetry such as _ r12 _ ( monoclinic _ c2 _ ) in the commensurate phase .
our conclusion is further supported by measurements of the magnetic susceptibility by tristan et al.@xcite that yield an easy magnetization along the @xmath34-axis .
in addition , we should emphasize that not only fe@xmath12 , but also the nd@xmath12 ions show antiferromagnetic long - range order below the nel temperature in case of neodymium ferroborate . + concerning the incommensurate phase the polarized @xmath21-scans over the positions of the magnetic peaks
could be well explained via the magnetic model ( m1a ) that was found for the commensurate phase simply by introducing the incommensurate propagation vector @xmath91 .
this suggests that the magnetic structure transforms into a long - period antiferromagnetic helix that propagates along the hexagonal @xmath9-axis with the magnetic moments perpendicular to it .
the incommensurate magnetic propagation vector @xmath91 = [ 0 , 0 , @xmath22 + @xmath27 ] is therefore associated with a rotation of the magnetic moments about 180@xmath107 + @xmath108 around the @xmath9-axis between adjacent hexagonal planes that are interrelated via trigonal translations .
the measured value of the splitting @xmath27 = 0.00667 corresponds to @xmath108 @xmath4 0.8@xmath107 and the full period of the helix amounts to approximately 1140 .
the mere observation of a chiral contribution by means of polarization analysis signifies unequally populated chirality domains .
our data further suggest that only one of the two chirality domains is populated .
the antiferromagnetic helix in the ic phase of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 therefore exists with an unique handedness .
a single chirality domain is in principle not expected , since left- and right - handed spirals are energetically degenerate , however in the case of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 this might be related to the fact that the chemical structure is non - centrosymmetric .
this is similar to the magnetic spirals in mnsi or uptge@xcite , and the more recent example ba@xmath0nbfe@xmath0si@xmath6o@xmath8@xcite that all three possess no inversion symmetry .
in particular the jana2006 analysis has shown that in addition the incommensurate magnetic structure exists in a single magnetic orientation domain with full superspace group symmetry _
r32(00@xmath108)t0 _ and in full agreement with the polarized results .
+ the determined saturation value of the magnetic fe moment of approximately 4.2 @xmath109 ( see fig . [
fig : magmom ] ) is less than 5 @xmath109 which would be expected for a free fe@xmath110 ion , in contrast to our previous powder results@xcite .
moreover , the ferromagnetic alignment of the fe and nd magnetic moments implies considerably smaller magnitudes for nd than derived from the powder diffraction data .
the latter are not affected by extinction .
the difference may be to a certain extent due to the extinction effects , as the even smaller saturation value of fe in case of thermal neutrons is caused by the considerably larger extinction , compared to hot neutrons . on the other hand , for tbfe@xmath0(@xmath1bo@xmath0)@xmath2 at 2 k , ritter at al . determined by means of powder neutron diffraction an antiparallel alignment of the fe and tb magnetic moments with ordered magnitudes @xmath111
= 4.39(4 ) @xmath109 and @xmath112 = 8.53(5 ) @xmath109@xcite .
despite the bondvalence result @xmath113 for fe derived by these authors , the ordered fe magnetic moment is also reduced .
hence , another possible reason could be frustration effects , yielding partially disordered magnetic moments .
+ furthermore , the observation of third order harmonics of the magnetic satellites at the positions ( 0 , 0 , 3/2 @xmath114 3@xmath27 ) in the incommensurate phase additionally suggest the formation of a magnetic soliton lattice in ndfe@xmath0(@xmath1bo@xmath0)@xmath2@xcite .
a soliton is the appearance of localized or topological defects in periodic structures due to the presence of non - linear forces .
such non - linear forces can be due to an external magnetic field that interacts with the magnetic moments or due to magnetic anisotropy as shown theoretically by izyumov and laptev@xcite .
indeed ndfe@xmath0(@xmath1bo@xmath0)@xmath2 exhibits magnetic anisotropy in the hexagonal basal plane as demonstrated by the results of tristan et al.@xcite .
the calculation of izyumov and laptev is based on a magnetic helix that forms due to the the presence of the dzyaloshinsky - moriya interaction ( dmi ) . until now no explicit statement about the existence of the dmi in ndfe@xmath0(@xmath1bo@xmath0)@xmath2 has been made .
however , as ndfe@xmath0(@xmath1bo@xmath0)@xmath2 is non - centrosymmetric the presence of the dmi is allowed from symmetry .
therefore , the formation of the observed magnetic helix is possibly driven by the dmi . +
in figs .
[ fig : nd_second_order_sats ] ( b ) to ( d ) we see that the intensities of the second order satellites are highest for temperatures t @xmath115 13.5 k and the distortions of the incommensurate periodic structures seem to be largest near to the c - ic phase transition .
we therefore assume that at t@xmath5 the interaction that favors a magnetic order that is incommensurate with respect to the underlying crystal lattice becomes non - negligible and leads to non - linear forces onto the magnetic subsystem slightly below t@xmath5 as it still wants to remain in its commensurate magnetic order . within a small temperature regime below
t@xmath5 the magnetic structure consequently is not yet completely incommensurate but can be rather viewed as a distorted commensurate magnetic structure with domain walls .
alternating periods of commensurate parts and domains walls then lead to the observed third order harmonics .
the observation of a magnetic soliton lattice without the application of external forces like magnetic fields or mechanical stress are rather unlikely and to the best of our knowledge the only other compound for which a magnetic soliton lattice was reported without the application of an external magnetic field is cub@xmath6o@xmath2 @xcite .
the observed temperature dependence of the propagation vector is continuous and described by eq . .
this behavior is close to @xmath116 reported in ref . .
in addition , similar to cub@xmath6o@xmath2 the commensurate phase is realized when the temperature is increased , which is in contradiction to the prediction of the theory @xcite . for cub@xmath6o@xmath2
it was proposed that the difference to the theory can be explained by assuming that the change of the propagation vector is not due to a temperature dependent magnetic anisotropy as in ref .
but rather due to the magnitude of the dmi that decreases as a function of increasing temperature @xcite .
we assume that is is similarly true for ndfe@xmath0(@xmath1bo@xmath0)@xmath2 . in summary our experimental results
are well described by the assumption of a magnetic soliton lattice .
+ the origin of the less pronounced incommensurability in case of the small crystal studied ( sample2 ) in the present work , compared to the larger one ( sample1 ) that was also used in ref . , is not yet clear and should be clarified by future systematic investigations of possible sample dependencies .
our neutron diffraction results show that the long - range magnetic order of multiferroic ndfe@xmath0(@xmath1bo@xmath0)@xmath2 observed below @xmath14 30 k consists of antiferromagnetic stacking along the @xmath9-axis , where the magnetic moments of all three fe@xmath12 sublattices and the nd@xmath12 sublattice are aligned ferromagnetically and parallel to the hexagonal basal plane , corresponding to model ( m1a ) . below t@xmath117 13.5
k the magnetic structure turns into an incommensurate antiferromagnetic helix propagating along the @xmath9-axis with a period of approximately 1140 . +
our polarized neutron diffraction data further suggests that the helix is monochiral , i.e. only one of the two possible chiral domains is fully populated .
the single magnetic chirality in neodymium ferroborate can be explained in terms of its non - centrosymmetric chemical structure , similar as for mnsi@xcite , and the more recent example ba@xmath0nbfe@xmath0si@xmath6o@xmath8@xcite . to the best of our knowledge
the former two materials are the only examples apart from ndfe@xmath0(@xmath1bo@xmath0)@xmath2 that show this peculiar property .
+ in the case of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 the commensurate - incommensurate magnetic phase transition is possibly accompanied by the formation of a magnetic soliton lattice , as indicated by the observation of third order harmonics of the magnetic bragg peaks .
this further suggests that the magnetic helix in ndfe@xmath0(@xmath1bo@xmath0)@xmath2 may be driven by the dzyaloshinskii - moriya interaction , which would be allowed by symmetry .
+ in conclusion , we identified the new monochiral compound ndfe@xmath0(@xmath1bo@xmath0)@xmath2 that provides us with a new model system to investigate the interesting properties of magnetic chirality in condensed matter . [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ] @xmath118 @xmath41 was fixed at the given value different from @xmath41 = 0 .
partially this work has been performed at the swiss spallation neutron source sinq@xcite ( instruments trics@xcite , tasp@xcite , mupad@xcite ) and another part at the single crystal neutron diffractometer heidi@xcite for hot neutrons , situated at the forschungsneutronenquelle heinz maier - leibnitz ( frm ii ) .
further we are thankful to severian gvasaliya for providing experimental support for the measurements performed on tasp .
mj is grateful to ben taylor and dominik bauer for useful discussions .
here we want to discuss the results of our magnetic refinements for ndfe@xmath0(@xmath1bo@xmath0)@xmath2 by means of fullprof in more detail .
we have chosen several characteristic data sets that highlight the results of the fits provided in table [ tab : fullprof_ref ] .
+ first we will discuss the data sets measured on heidi with hot neutrons ( @xmath20 = 0.55 ) . for
the commensurate magnetic phase of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 we performed fits with @xmath41 fixed to zero and free , respectively .
the corresponding sets in table [ tab : fullprof_ref ] are ( f1a ) and ( f1b ) at t = 22.5 k , respectively .
the refinements were carried out with the nuclear scale factor and extinction parameter 2.9(1 ) that was determined by the fit ( n1 ) ( cf .
table [ tab : structure ] ) .
the determined angle @xmath41 for ( f1b ) is of similar magnitude as the value 76(3)@xmath119 of the previous powder neutron diffraction results@xcite .
however , according to hamilton s signficance test@xcite , one should consider the ratio 1.01 of the weighted r - factors . then r@xmath120 @xmath41.01 from table 1 in ref
. indicates only a marginal significance level of 0.25 that ( f1b ) is the correct result .
further refinements at lower temperatures support this as will be demonstrated in the following .
further , variation of also the extinction parameter yields the value 3.0(8 ) and @xmath86 = 1.07 . therefore in case of
ndfe@xmath0(@xmath1bo@xmath0)@xmath2 the nuclear and magnetic extinction parameters agree within error limits .
thus we use the former also for the refinements of the subsequent small heidi data sets with 5 nuclear and 50 magnetic peaks .
the former were used to obtain the scale factor .
+ at t = 15 k we performed three different sets of fits where the angle @xmath41 was consequently fixed to 0 , 15 and 30 @xmath119 ( cf . sets ( f2a , f2a,f2a ) in table [ tab : fullprof_ref ] ) .
there is apparently a very flat minimum centered at @xmath41 = 0 which proves the assumption @xmath41 = 0 to be correct .
+ the same is valid for the incommensurate magnetic phase of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 .
fits of the data at 5 k show better agreement factor when @xmath41 is fixed to zero ( cf . sets ( f3a+b ) in table [ tab : fullprof_ref ] ) . + finally , the results obtained by fits of the trics data measured with thermal neutrons ( @xmath20 = 1.18 ) at t = 5.5 k gives the following results .
the nuclear fit ( n2 ) yielded the scale factor 40(1 ) and a considerably larger extinction parameter 18(1 ) , compared to the hot neutron heidi data .
best fits of the magnetic neutron intensities had been obtained with temperature parameter b = 0 of the metal atoms .
the positional x - parameter of fe had been taken from ( n2 ) , table [ tab : structure ] .
good fits were only obtained by refining also the extinction parameter , resulting in the considerably smaller magnetic value 2.3(2 ) , compared to the almost an order larger nuclear value .
the results of the refinement of the magnetic intensities is given in table [ tab : fullprof_ref ] sets ( f4a ) and ( f4b ) , corresponding again to fixed and free polar angle @xmath41 , respectively .
for ( f4b ) the extinction parameter was fixed to the value from ( f4a ) .
the latter does not change essentially , if it is varied too .
although refinement ( f4a ) seems with an additional parameter somewhat better , we think that it is hardly significant with respect to the comparable hot neutron 5 k results , which in principle are considerably less affected by extinction .
in the following we will discuss our additional refinements that were performed with jana2006@xcite . recently
the option to refine magnetic structures has been implemented .
it should be emphasized that magnetic structures have to be defined in a different way in jana2006 compared to fullprof . in jana2006
commensurate and incommensurate magnetic structures are described with superspace groups@xcite similar to occupationally modulated chemical structures .
the modulation function of the magnetic axial vector configuration is represented by means of a fourier expansion according to the observed k - vectors , similar to ref . .
moreover , the irreducible representations of the chemical structure with their little group @xmath121 and the associated magnetic basis functions as well as time inversion are taken into account .
the atomic magnetic moments are described in a polar coordinate system .
+ the incommensurate and commensurate magnetic structures of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 correspond to a complex one - dimensional and to a two - dimensional representation , respectively . keeping the chemical structure according to space group _ r32 _ ,
the former is described in by means of a four - dimensional superspace approach with cosine and sine waves associated with the k - vectors @xmath91 = [ 0,0,@xmath1141.502 ] . for ferromagnetic coupling of the magnetic fe and nd moments in the easy ( a , b)-plane of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 , the magnetic superspace group _ r32(00@xmath108)t0 _ was used . with the nd moments deviating from @xmath40 = 0 in the ( a
, b)-plane , magnetic superspace group _ r3(00@xmath108)t _ holds . here
the symmetry implies one or two magnetic domains , respectively . in the latter case
they were found to be statistically populated .
+ in the magnetic commensurate case with @xmath28 = [ 0,0,3/2 ] , being equivalent to the two - fold superstructure , only a cosine wave with magnetic superspace group _ r12(@xmath122,@xmath122,@xmath108)00 _ ( @xmath122=@xmath123=0;@xmath108=3/2 ) is appropriate for ferromagnetic alignment of the magnetic nd and fe moments in the ( a , b)-plane . for magnetic monoclinic superspace group _ r1(@xmath122,@xmath123,@xmath108)t _ ( @xmath122=@xmath123=0;@xmath108=3/2 )
the magnetic nd and fe moments are not oriented parallel in the ( a , b)-plane .
the two cases imply three or six magnetic domains , respectively .
they were found to be statistically populated .
+ as an example we discussed the results of combined refinement of 5 nuclear ( to obtain the scale factor ) and 50 magnetic peaks at 15 k ( heidi measurements ) that are given in table [ tab : jana_ref ] .
here we used the extinction parameter rhiiso = 0.109(8 ) and the atom positions from table [ tab : structure ] .
( j2a ) and ( j2b ) correspond to fixed and free polar angle @xmath41 .
the resulting r - factors of the refinement for this temperature show obviously no significance for an additional parameter @xmath41 and again indicate that the ferromagnetic alignment of the fe and nd magnetic moments is correct .
+ for the data at 5 k a non - zero angle @xmath41 shows a small improvement of the refinement ( j3b ) , however , in view of the additional refinement parameter @xmath41 , one may question whether the small improvement is significant . based on hamilton s significance test@xcite ,
the @xmath124-ratio 8.2/7.9 = 1.04 should be considered . in cases ( j3a ) and ( j3b ) we have 2 and three parameters , respectively and 55 reflections ,
i.e @xmath125 @xmath4 1.04 from table 1 in ref . indicates the correctness of @xmath41=0 ( j3a ) at a significance level of approximately 0.05 .
+ on the other hand ( j4b ) appears to be better than ( j4a ) .
presumably this is related to the considerably larger extinction effects associated with thermal neutrons compared to hot neutrons .
apart from this temperature , the jana2006 data evaluation yields values for the magnetic moments that agree within the error bars with the values shown in fig .
[ fig : magmom ] which is based on the fullprof evaluation .
it should be noted that the superspace group description yields a natural explanation for the number of magnetic domains@xcite .
the latter were found to be statistically occupied .
|
we have performed an extensive study of single - crystals of the magneto - electric ndfe@xmath0(@xmath1bo@xmath0)@xmath2 by means of a combination of single - crystal neutron diffraction and spherical neutron polarimetry .
our investigation did not detect significant deviations at low temperatures from space group _ r32 _ concerning the chemical structure .
with respect to magnetic ordering our combined results demonstrate that in the commensurate magnetic phase below t@xmath3
@xmath4 30 k all three magnetic fe moments and the magnetic nd moment are aligned ferromagnetically in the basal hexagonal plane but align antiferromagnetically between adjacent planes .
the phase transition to the low - temperature incommensurate magnetic structure observed at t@xmath5 @xmath4 13.5 k appears to be continuous . by means of polarized neutron studies
it could be shown that in the incommensurate magnetic phase the magnetic structure of ndfe@xmath0(@xmath1bo@xmath0)@xmath2 is transformed into a long - period antiferromagnetic helix with single chirality .
close to the commensurate - incommensurate phase transition third - order harmonics were observed which in addition indicate the formation of magnetic solitons .
|
during a few recent years , artificial media with negative real parts of material parameters ( called veselago media , backward - wave media , double negative media ) have attracted much attention in view of the first experimental realizations of such media and because of potential applications in sub - wavelength imaging .
some other applications have been proposed , including improvement of antenna performance . in paper
@xcite , radiation from a small electric dipole inside a spherical shell of such material has been considered , with some conclusions regarding the antenna bandwidth .
a two - dimensional model of an antenna coated by a dispersive material shell has been considered in @xcite .
these and many other issues involve considerations of the reactive energy stored in materials with negative parameters , and the goal of this paper is to develop a method for calculation of the stored energy density in complex materials and to discuss difficulties in the interpretation of the properties of exotic materials as well as the validity of the used models of materials with negative parameters .
a general approach that we introduce in this paper allows to determine the time - averaged energy density of time - harmonic electromagnetic fields in dispersive and lossy materials with various dispersion laws .
also , the possibility to create artificial materials with dispersion - free and negative parameters using _ active _ inclusions @xcite is revisited .
it is well known that the field energy density in materials can be uniquely defined in terms of the effective material parameters only in case of small ( negligible ) losses ( e.g. @xcite ) .
this is because in the general case when absorption can not be neglected , the terms _
e_dt + _ h_btdescribe both the rate of changing the stored energy and the absorption rate .
only if the absorption is negligible , we can write _ e_dt + _
h_bt = w_et+w_mt , where @xmath0 and @xmath1 are the energy densities of the electric and magnetic fields , respectively . for artificial materials based on metal or dielectric inclusions of various shapes absorption can be neglected when the operational frequency is far from the resonant frequencies of the inclusions and from the lattice resonances , if the material is periodical . for electromagnetic fields
whose spectrum is concentrated near a certain frequency @xmath2 , the time - averaged energy density in a material with scalar frequency - dispersive parameters @xmath3 and @xmath4 reads ( e.g. @xcite ) w=12.d(())|_=_0|e|^2 + 12.d(())|_=_0|h|^2 .
wdif in the vicinity of the operating frequency @xmath2 the frequency dispersion can be neglected and @xmath5 and @xmath6 can be assumed to be independent from the frequency , simplifies to w=12|e|^2 + 12|h|^2 .
wnthe validity of this formula is restricted to positive values of @xmath5 and @xmath6 because no passive media in thermodynamic equilibrium can store negative reactive energy , as this is forbidden by the thermodynamics ( the second principle ) @xcite .
actually , this means that frequency dispersion can not be neglected when estimating the stored energy in the frequency regions where the material parameters are negative .
if the material has considerable losses near the frequency of interest , it is not possible to define the stored energy density in a general way ( more precisely , it is not possible to express that in terms of the material permittivity and permeability functions ) @xcite .
knowledge about the material microstructure is necessary to find the energy density , and this problem is far from trivial . in the literature , the energy density expressions have been derived for an absorbing classical dielectric ( lorentz dispersion ) with a single resonant frequency @xcite and for the case where also the permeability obeys the same dispersion law as the permittivity @xcite .
the known artificial materials with negative real parts of the material parameters have different and more general dispersion laws , which means that we need to develop a more general approach suitable for calculations of the stored energy density in general dispersive and lossy materials .
such method will be presented in this paper . in the next section
we start from a brief review of the general properties of the constitutive parameters of passive low - loss materials , when the energy density can be defined in terms of the effective material parameters by formula .
in this section we consider passive linear materials in thermodynamic equilibrium in the frequency regions where absorption can be neglected and the field energy density can be found in terms of the effective permittivity and permeability functions . for simplicity of writing , we restrict the analysis to isotropic media . landau and lifshitz in @xcite give a proof that for all linear _ passive _ materials in the frequency regions with weak absorption the value of @xmath7 in is not only always positive , but it is always larger than the energy density of the same fields @xmath8 and @xmath9 in vacuum .
indeed , the following inequalities can be derived from the causality requirement assuming negligible losses @xcite : > 0 ( eq .
( 84.1 ) in @xcite ) and > 2(_0- ) 842 ( eq . ( 84.2 ) in @xcite ) . summing these two inequalities
one finds that @xcite > _ 0 .
limit the same is true for the permeability as well : > _ 0 .
brokenthis has a clear physical meaning : to create fields in a material , work must be done to polarize the medium , which means that in the absence of losses more energy will be stored in material than in vacuum .
this result is very general and applies also to passive low - loss metamaterials with negative parameters .
inequality can be cast in equivalent form > 2_0- ( ) .
2edepending on the value of @xmath5 , either or is stronger . as is seen from , when the permittivity is negative and large in the absolute value , the permittivity must be very dispersive .
considering plane electromagnetic waves in transparent isotropic dispersive materials , sivukhin @xcite gave one more limitation on the relative material parameters : + _ r_r d(_r())>0 .
this relation holds if _ both _ @xmath10 and @xmath11 are either positive or negative .
if in a certain model the material parameters are assumed to be completely lossless , the above inequalities can become equalities .
for example , the lossless plasma permittivity function ( ) = _ 0(1-_p^2 ^ 2)is just on the allowed limit , because in this case = _ 0(1 + _ p^2 ^ 2)=2_0-()it is easy to check that the lossless lorentz permittivity model ( ) = _
0(1+_p^2)satisfies all the above inequalities at all frequencies .
modeling of artificial magnetic materials requires more care because the very notion of permeability loses its meaning at high frequencies .
thus , the model permeability expressions obtained from quasi - static considerations do not necessarily satisfy the basic physical requirements at high frequencies .
an important example is the effective permeability of a mixture of chiral or omega particles @xcite or split - ring resonators @xcite , or of arrays of swiss rolls " @xcite : = _ 0(1+a^2 ) .
mu_srr this function has a physically sound behavior at low frequencies [ @xmath12 and near the resonance , but in the limit @xmath13 it does not tend to @xmath14 however , in the limit of extremely high frequencies materials can not be polarized at all because of inertia of electrons , so the parameters must tend to @xmath15 and @xmath14 . as a result ,
this expression becomes non - physical [ due to instantaneous response of the material , condition is not satisfied ] at frequencies larger than @xmath16 . as explained in @xcite , the integrals in the kramers - krnig relations should be truncated at a high enough frequency where the permeability becomes nearly real and constant ( formula ( 82.17 ) in @xcite ) .
note , however , that inequality > 0 for the permeability function is still satisfied at all frequencies . in this section
we will introduce a circuit model for complex passive metamaterials needed to study the energy density in complex media and use it first to show that although field solutions of the maxwell equations for materials with non - dispersive negative material parameters exist , they are unstable .
this fact appears to be quite obvious , but for some reason it appears to be neglected in the literature .
indeed , let us consider a material model with constant and negative @xmath5 and @xmath6 , as e.g. in @xcite . within the frame of this model ,
the energy density is given by and it is negative @xcite . if due to a fluctuation the field at a certain moment of time increases , the stored energy will _ decrease_. in this situation the field will increase even more , so that the total energy of the system be minimized .
as a first simple example , let us consider a parallel - plate capacitor filled by a _
hypothetical _ material with a negative permittivity @xmath17 which does not depend on the frequency .
let us assume that the capacitor is charged and the voltage between its terminals is @xmath18 .
at @xmath19 it is connected to a resistor ( resistance @xmath20 ) .
the voltage satisfies the differential equation u(t)+cr du(t)=0,whose solution is u(t)=u(0)e^-t.under our assumption , capacitance @xmath21 is negative , which means that the capacitor does not discharge but , on the contrary , charges itself , collecting free charges from the conductor and resistor .
this is an expected result , because in this process the total field energy in the system _
decreases_. the same conclusion follows if the transient regimes in these systems are analyzed .
figure [ cap ] shows a capacitor that is connected to a time harmonic voltage source @xmath22 at time @xmath19 ( there is no charge in the capacitor at @xmath23 ) .
the current in the circuit reads i(t)=v_0c(t)+rc(t)-e^-t .if the capacitance is negative , the current exponentially grows .
this result shows that the time - harmonic solutions in media with negative _ non - dispersive _ parameters are _ unstable _ , because a small increase in the source amplitude due to noise will exponentially grow .
this is the reason for the know fact @xcite that finite - difference time - domain schemes for media with negative and non - dispersive material parameters are unstable . in reality
, these instabilities do not exist because all passive materials with negative parameters are frequency dispersive .
to illustrate this , let us consider a more realistic case where the negative permittivity corresponds to a lossless plasma : ( ) = _ 0_r=_0 ( 1-_p^2).at the frequencies @xmath24 the permittivity is real and negative .
the real parameter @xmath25 is called _ plasma frequency . _ at microwaves , dense arrays of ideally conducting wires can be described ( although only for specific excitations @xcite ) by this model ( e.g. @xcite ) . if we fill a capacitor with such material , its impedance becomes z=1=1 , z_wireswhere @xmath26 is the capacitance of the same capacitor filled by vacuum .
obviously , this corresponds to a parallel connection of a capacitor with capacitance @xmath26 and an inductor with @xmath27 , as illustrated in figure [ cap_dis ] .
connection of a capacitor filled by a lossless plasma to a source is equivalent to connection of a usual parallel resonant circuit to the same source .
although the filling material has a negative permittivity at @xmath24 , both the capacitance and inductance in the equivalent circuit are positive .
naturally , the solution for the current in this circuit ( that can be readily found using the laplace transform , for example ) does not contain any growing exponents . similarly to the case of capacitors filled with _
hypothetical _ non - dispersive materials with negative permittivities , current through an inductor filled with a _
hypothetical _ non - dispersive material with a negative permeability exponentially grows in time , if such inductor is connected to a voltage source via a resistor . for a more realistic consideration ,
let us assume that the magnetic material that fills the inductor s coil can be modeled by the permeability function , where the magnitude factor @xmath28 does not depend on the frequency . in a certain frequency region ( higher than the resonant frequency @xmath2 )
the permeability is negative .
let us consider an inductor filled by this material and connected via a resistor to a time - harmonic voltage source ( figure [ l ] , left ) .
the impedance of this inductor is z()=jl_0_r ( ) = jl_0 + j^3 l_0 a , where @xmath29 is the inductance of the same coil without any magnetic material inside .
this is the same as the impedance of an inductor magnetically coupled to a resonant circuit ( figure [ l ] , right ) @xcite : z()=jl_0 + j^3 m^2/l.thus , the circuit shown on the right of figure [ l ] which contains only usual constant and positive inductances and a capacitance is equivalent to the circuit containing a magnetic core with permittivity ( figure [ l ] , left ) if we choose @xmath30 . as in the case of materials with negative permittivity , it is obvious that no instabilities occur in this simple passive circuit .
next we will determine the stored energy density in particular realizations of veselago media at microwaves , taking into account dispersion and dissipation .
for media without magnetoelectric interactions that can be adequately characterized by two materials parameters : the permittivity and permeability , it is possible to consider energies stored in the electric and magnetic fields separately .
indeed , the properties of linear media do not depend on what particular external field we apply . having the full freedom to choose the external sources
, we can always realize a situation where in a certain ( small ) volume only electric or magnetic field is non - zero . because we deal with _ effective materials , _ the period of the microstructure or the average distance between inclusions is considerably smaller than the wavelength , otherwise one can not introduce effective permittivity and permeability .
thus , we can take a representative sample of the material that contains many inclusions but whose size is still much smaller than the wavelength , and probe its properties in ( nearly ) uniform electric and magnetic fields .
negative effective permittivity is most often realized by dense arrays of parallel thin metal wires . for plane
electromagnetic waves whose wave vector is orthogonal to the wires , the effective permittivity for electric fields directed along the wires can be modeled by the plasma permittivity function = _ 0 .
eps_plathere exist several models for the equivalent plasma frequency @xmath25 , and here we will use the quasi - static model @xcite that is not limited to the case of small wire radius and allows to estimate the loss factor .
for example , if the skin effect in the wires can be neglected ( uniform current distribution over the wire cross section ) , the effective parameters read @xcite _ p^2=2 , omega_p= 2 .
gamma here @xmath31 is the array period , @xmath32 is the wire radius , @xmath33 is the conductivity of the wire material , and @xmath15 and @xmath14 are the parameters of the matrix .
the matrix is assumed to be a lossless magnetodielectric , so @xmath15 and @xmath14 are real numbers . to determine the stored field energy density in this material , we position a small ( in terms of the wavelength or the decay factor in the effective medium ) piece of this material in a parallel - plate capacitor .
generalizing formula using , we have for the admittance y = jc_0 + 1 , where @xmath26 is now the capacitance of the capacitor filled with the matrix material ( permittivity @xmath15 ) , and l=1/(_p^2c_0 ) , r=/(_p^2c_0 ) .
paramobviously , the equivalent circuit is a parallel connection of a capacitor and an inductor with a loss resistor , figure [ para ] .
this circuit has the same input impedance as the actual capacitor filled by a material sample .
however , before using this circuit in order to calculate the stored reactive energy in the medium , we must ensure that the circuit structure indeed corresponds to the microstructure of the material under study .
it is well known that from the input impedance of a circuit it is impossible to uniquely determine the circuit structure . in other words
, different circuits can have the same input impedance at all frequencies ( e.g. , @xcite ) . in the context of this study
this means that our circuit model correctly describes the input impedance of the material - loaded capacitor , but it may fail to properly model the material microstructure . in our particular case
the material is realized as an array of wires along the electric field direction ( that is , running from one plate of the capacitor to the other ) .
apparently , this array of wires possesses some inductance and resistance connected in series , so we see that our model indeed corresponds to the microstructure of the medium , and we can use it . in the time
- harmonic regime the time - averaged stored reactive energy is w=12(c_0|v_c|^2+l|i_l|^2 ) , where @xmath34 is the voltage amplitude on the capacitor and @xmath35 is the amplitude of the current through the inductor ( see the equivalent circuit in figure [ para ] ) .
this can be written as w=12c_0|v_c|^2.wwfor a parallel - plate capacitor ( the plate area @xmath36 , the distance between the plates @xmath37 ) , we have @xmath38 and @xmath39 .
the total energy is the energy density @xmath0 multiplied by the capacitor volume @xmath40 : w = w_esd=12_0 sd|e|^2d^2 .wwthus , the energy density reads w_e=_0 2 ( 1+_p^2)|e|^2 , w_wirewhere we have substituted the values of the circuit parameters from .
parameters @xmath25 and @xmath41 are given by and , respectively .
if the losses can be neglected ( @xmath42 ) , the same result follows from , since = _ 0 ( 1+_p^2 ) .
let us next consider a wire medium where the matrix is a lossy dielectric , with the permittivity @xmath43 , where @xmath44 is the conductivity of the matrix material .
following the approach of @xcite , we find that the effective permittivity is = _ 0-j_d , where the plasma parameters @xmath25 and @xmath41 remain the same as for wires in a lossless matrix ( the physical reason for this is that the negative permittivity appears due to the inductance of the wire array , and that inductance does not depend on the dielectric loss ) . in the equivalent circuit shown in figure [ para ]
, the matrix loss will be reflected by an additional loss resistance @xmath45 connected in parallel with the capacitor @xmath26 .
this means that the expression for the stored energy density does not change if the matrix has some non - zero conductivity : dielectric losses in the matrix have no effect on the stored energy density function , while losses in the wires have a strong effect .
let us assume next that the matrix has no conductivity , but there are some magnetic losses : the matrix parameters are @xmath15 and @xmath46 , where @xmath15 and @xmath14 are real .
the effective permittivity of the wire medium in this matrix can be found using formulas of @xcite with substitution @xmath47 where @xmath48 is the inductance per unit length of the wire array in the matrix with the permeability @xmath14 .
the result is the same as where the plasma frequency does not depend on @xmath49 and is given by , but the loss factor @xmath41 is different : = 2 + .
in addition to the loss factor due to resistive wires , there is a factor measuring magnetic losses in the background medium .
the structure of the equivalent circuit in figure [ para ] does not change , but the resistance r=_p^2c_0 is now _ frequency - dependent_. however , the stored energy density can be still calculated using formula , because the additional impedance is _
real_. negative permittivity can be alternatively realized using artificial dielectrics with resonant inclusions .
frequency dispersion in such materials is described by the lorentz formula = _
0(1+_p^2 ) , lorwhich is widely used as a model of natural materials in solid state physics .
we will consider microwave materials designed as a collection of short metal needles oriented along the electric field direction and distributed in a lossless matrix . if the length of the needles is much smaller than the wavelength and the distance between the needles is much larger than the needle length but much smaller than the wavelength , formula is a good estimate for the effective permittivity .
parameters @xmath25 , @xmath2 and @xmath41 can be estimated in terms of the particle dimensions and the inclusion concentration using the antenna model of an individual inclusion @xcite and an appropriate mixing rule ( e.g. , @xcite ) .
making use of the same approach as above , we consider a capacitor filled with a material having this dispersion law .
its admittance is y = jc_0=jc_0 + jc_0_p^2 .apparently , this is the admittance of a parallel connection of a capacitor @xmath26 and a series resonant circuit with the elements c = c_0_p^2/_0 ^ 2 , l=1/(_p^2c_0 ) , r=/(_p^2c_0 ) . pl it differs from the equivalent circuit for wire media ( figure [ para ] ) by the additional capacitance @xmath21 in series with @xmath48 and @xmath20 .
this equivalent circuit is a valid model for the microstructure of this material because currents along needles are modeled by inductance @xmath48 and charges at the ends of the needles by capacitance @xmath21 .
the loss is due to non - ideally conducting material of the needles ( the matrix material is assumed to be lossless ) , so it is appropriately modeled by resistor @xmath20 in series with the inductance .
the stored reactive energy is the sum of the energies stored in all reactive elements : w = w_esd=12(c_0|v_c_0|^2+l|i_l|^2+c|v_c|^2 ) , where @xmath50 and @xmath34 are the voltages at the respective elements . solving for @xmath35 and @xmath34 and substituting the equivalent circuit parameters we find w_e=_02|e|^2.this result coincides with that obtained earlier in @xcite , where the motion equation for the electric polarization was directly solved . for the case of negligible losses ( @xmath51 ) ,
the same result follows from .
the present method extends to the case of many resonant frequencies ( multi - phase mixtures of inclusions of several different sizes ) by simply adding more parallel @xmath52 branches to the equivalent circuit .
dense arrays of split rings and other similar structures can be modeled in the quasi - static regime by the following effective permeability ( e.g. @xcite ) : = _ 0(1+a^2 ) , lorentzwhere the magnitude factor @xmath28 and the loss factor @xmath41 do not depend on the frequency . similarly to the approach introduced above for artificial dielectrics ,
we position a small ( in terms of the wavelength or the decay length in the effective medium ) sample in the magnetic field of a solenoid with inductance @xmath29 .
the inductance becomes z()=jl_0_r ( ) = jl_0 + j^3 l_0 a. zl an equivalent circuit with the same impedance is shown in figure [ ll ] .
indeed , the input impedance seen by the source is z = jl_0+j^3 m^2/l.this is the same as if = l_0a,=_0 ^ 2,=. lparathis is a correct equivalent representation from the microscopic point of view , because the material which we model is a collection of capacitively loaded loops magnetically coupled with the incident magnetic field .
the total stored reactive energy is the sum of the energies stored in all reactive elements : w=12(l_0|i|^2 + l|i_l|^2 + c|v_c|^2 ) .expressing @xmath35 and @xmath34 in terms of @xmath53 , we get , similarly to the derivations in @xcite , w=12|i|^2 .
rearranging terms and substituting the equivalent parameters , this can be written as w=12l_0|i|^2.considering the stored energy in one unit - length section of the solenoid , we have w = w_ms=12_0 n^2s |h|^2n^2 , where @xmath36 is the solenoid cross section area and @xmath54 is the number of turns per unit length .
we have substituted the solenoid inductance per unit length ( a tightly wound long solenoid ) @xmath55 and used the relation @xmath56 between the current @xmath53 and the magnetic field inside the solenoid @xmath57 .
finally , the stored field energy density is found to be w_m=_02 |h|^2.correct_srr it is important to note that in this particular case formula leads to an incorrect expression even if the losses are negligible ( @xmath51 ) . for @xmath58 the stored energy density obtained from is less than the energy stored in vacuum , and it becomes even negative at still higher frequencies .
this is a manifestation of the failure of the quasistatic permeability model discussed above in section [ limitations ] .
formula should be used even in the case of small losses .
in paper @xcite it was shown that in principle it is possible to realize a material with _
non - dispersive _ negative material parameters using artificial molecules " that contain electronic circuits ( impedance inverters ) . in these active materials
, the basic physical limitation does not apply . in the light of the previous consideration , it is of interest to study
if in this case the stored energy density can be negative and clarify the physical meaning of this effect .
let us consider an artificial molecule in form of a short dipole antenna loaded by a bulk impedance ( figure [ opam ] , left ) .
to realize a wide - band non - dispersive negative permittivity , the load must be a _ negative _ capacitance @xcite : z_load=-1- nl_eff^2 , zlwhere @xmath59 is the input capacitance of the dipole antenna , @xmath60 is the number of particles per unit volume , @xmath61 is the effective antenna length , and @xmath5 is the effective negative permittivity that we want to realize [ the low - density limit of the clausius - mossotti formula was used in @xcite to derive ] . on the right of figure [ opam ] , the equivalent circuit of the whole molecule is shown .
the negative capacitance is realized by an impedance inverter , which is connected to an equivalent circuit of a short dipole antenna ( @xmath59 is its input capacitance ) .
@xmath62 is the loss resistance of the antenna ( the radiation resistance is compensated by the interaction field , as we suppose that the array of particles is regular or enough dense @xcite ) .
the total impedance connected to the voltage source ( external electric field ) reads z = r_loss - nl_eff^2,that is , effectively we have a capacitor filled with a non - dispersive negative - permittivity material excited by a voltage source . as we know from section
[ unstable ] , this system is _ unstable _ , which is an expected result .
this is in agreement with experiments @xcite .
a general approach that allows to determine the stored energy density in complex composite microwave materials has been been presented .
the method is based on an equivalent circuit representation of small material samples excited by electric and magnetic fields .
introduction of equivalent circuit parameters for specific microstructures of media is physically equivalent to an appropriate averaging procedure , needed to determine the properties of the effective medium .
particular examples of wire media ( negative - epsilon material ) and arrays of split - rings ( negative - mu material ) have been considered , as well as the usual lorentzian dielectrics with losses .
the last case has been considered in the literature using a different approach , and the present result agrees with the known formula .
the above derivations show how the energy density can be found for any passive and lossy composite , if its microstructure is known .
the energy density is determined in terms of the energy stored in the reactive elements of the equivalent circuits .
naturally , in all cases the stored energy is positive , as it should be in all passive materials .
this conclusion appears to be very natural if one remembers that passive metamaterials exhibiting negative material parameters are anyway made from usual materials like metals or dielectrics . on the microscopic level ,
the stored energy is the electromagnetic field energy in the matrix material ( normally a dielectric ) and in the inclusions ( normally a metal of another dielectric ) .
this energy is a strictly non - negative definite function .
the energy stored in a sample of the effective medium is the average of the corresponding microscopic quantity , and there is no reason to expect that for some specific shapes of metal inclusions the effective material will store negative energy .
for example , the use of passive metamaterials in the design of antennas basically means adding some extra metal or dielectric elements like metal wires or split - ring resonators to a simpler antenna . on the fundamental level , this means only changing the antenna shape .
this work has been partially funded by the academy of finland and tekes through the center - of - excellence program .
helpful discussions with prof .
nefedov , dr . s.i .
maslovski , and prof .
simovski are very much appreciated .
ziolkowski and a.d .
kipple , application of double negative materials to increase the power radiated by electrically small antennas , _ ieee trans .
antennas propag .
, _ vol .
51 , no .
10 , pp . 2626 - 2640 , 2003 .
tretyakov , s.i .
maslovski , a.a .
sochava , c.r .
simovski , the influence of complex material coverings on the bandwidth of antennas , submitted to _
ieee trans .
antennas propagation .
_ preprint available at http://arxiv.org/pdf/physics/0401144 .
kadomtsev , a.b .
mikhailovski , a.v .
timofeyev , negative energy waves in dispersive media , _ zhurnal teoretich . and experim .
fiziki , _ vol .
6(12 ) , pp . 2266 - 2268 , 1964 ( in russian .
english translation in _ sov .
zetf _ ) .
tretyakov , f. mariotte , c.r .
simovski , t.g .
kharina , and j .-
heliot , analytical antenna model for chiral scatterers : comparison with numerical and experimental data , _ ieee transactions on antennas and propagation , _ vol .
44 , no . 7 , pp . 1006 - 1014 , 1996 . d.r .
smith , s. schultz , p. marko , c.m .
soukoulis , determinaiton of effective permittivity and permeability of metamaterials from reflection and transmission coefficients , _ phys . rev .
b , _ vol . 65 , p. 195104
, 2002 .
belov , r. marqus , s.i .
maslovski , i.s .
nefedov , m. silveirinha , c.r .
simovski , and s.a .
tretyakov , strong spatial dispersion in wire media in the very large wavelength limit , _ physical review b , _ vol .
67 , pp . 113103(1 - 4 ) , 2003
. p.a .
belov , s.a .
tretyakov , a.j .
viitanen , dispersion and reflection properties of artificial media formed by regular lattices of ideally conducting wires , _ j. electromagnetic waves and applications , _ vol . 16 , no . 8 , pp . 1153 - 1170 , 2002 .
|
general relations for the stored reactive field energy density in passive linear artificial microwave materials are established .
these relations account for dispersion and absorption effects in these materials , and they are valid also in the regions where the real parts of the material parameters are negative .
these relations always give physically sound positive values for the energy density in passive metamaterials .
the energy density and field solutions in active metamaterials with non - dispersive negative parameters are also considered .
basic physical limitations on the frequency dispersion of material parameters of artificial passive materials with negative real parts of the effective parameters are discussed .
it is shown that field solutions in hypothetical materials with negative and non - dispersive parameters are unstable . artificial materials , negative material parameters , energy density , equivalent circuit , instability .
|
a _ relative group presentation _ is a presentation of the form @xmath0 where @xmath8 is a group , @xmath9 a set disjoint from @xmath8 and @xmath10 is a set of cyclically reduced words in the free product @xmath11 where @xmath12 denotes the free group on @xmath9 .
if @xmath13 denotes the group defined by @xmath3 then @xmath13 is the quotient group @xmath14 where @xmath15 denotes the normal closure in @xmath11 of @xmath10 .
a relative presentation is defined in [ 2 ] to be _ aspherical _ if every spherical picture over it contains a _
dipole_. if @xmath3 is aspherical then statements about @xmath13 can be deduced and the reader is referred to [ 2 ] for a discussion of these ; in particular torsion in @xmath13 can easily be described .
we will consider the case when both @xmath9 and @xmath10 consists of a single element .
thus @xmath16 where @xmath17 where @xmath18 , @xmath19 and @xmath20 implies @xmath21 for @xmath22 , subscripts @xmath23 . if @xmath24 or if @xmath25 then , modulo some open cases , a complete classification of when @xmath3 is aspherical has been obtained in [ 1 ] , [ 2 ] , [ 7 ] and [ 11 ] .
the case @xmath26 for @xmath27 was studied in [ 12 ] and @xmath28 ( @xmath29 ) was studied in [ 5 ] .
the authors of [ 10 ] used results from [ 7 ] in which @xmath30 to prove asphericity for certain log groups . in this paper
we continue the study of asphericity and consider @xmath31 .
observe that @xmath32 if and only if @xmath33 so replacing @xmath34 by @xmath35 it follows that we can work modulo @xmath36 and @xmath37 .
a standard approach is to make the transformation @xmath38 and then consider the subgroup @xmath5 of @xmath8 generated by the resulting coefficients
. in our case @xmath39 becomes @xmath40 and so @xmath41 .
one then usually proceeds according to either when @xmath5 is non - cyclic or when @xmath5 is cyclic .
( note that @xmath42 is cyclic if and only if @xmath43 is cyclic . )
the latter case seems to be the more complicated indeed the open cases referred to in the above paragraph almost all involve @xmath5 being cyclic .
our results reflect this difference in difficulty .
when @xmath5 is non - cyclic we obtain a complete answer except for the following case ( modulo @xmath36 , @xmath37 ) in which @xmath44 : 1 .
@xmath45 ; @xmath46 ; @xmath47 ; @xmath48 . *
* theorem 1.1**__let @xmath3 be the relative presentation _
_ @xmath49 _ where @xmath18 ( @xmath50 ) , @xmath51 , @xmath52 and @xmath53 .
let @xmath54 and assume that @xmath5 is non - cyclic and the exceptional case ( * e * ) does not hold
. then @xmath3 is aspherical if and only if ( modulo @xmath36 , @xmath37 ) none of the following conditions holds : _ 1 .
@xmath55 _ and @xmath56 ; _ 2 .
@xmath57 , @xmath58 _ and @xmath59 ; _ 3 .
@xmath60 .
now let @xmath5 be cyclic . before stating the theorem we make a list of exceptions ( modulo @xmath36 , @xmath37 ) .
1 . @xmath61 ; @xmath62 ; @xmath63 .
2 . @xmath64 ; @xmath47 ; @xmath65 .
3 . @xmath64 ; @xmath47 ; @xmath66 .
4 . @xmath67 ; @xmath68 ; @xmath66 .
observe that ( * e1 * ) implies @xmath69 ; ( * e2 * ) and ( * e3 * ) imply @xmath70 ; and ( * e4 * ) implies @xmath71 . * * theorem 1.2**__let @xmath3 be the relative presentation _ _ @xmath72 _ where @xmath18 ( @xmath50 ) , @xmath51 , @xmath52 and @xmath53 .
let @xmath73 be a cyclic group .
suppose that none of the exceptional conditions ( * e1*)(*e4 * ) holds
. then @xmath3 is aspherical if and only if either @xmath5 is infinite or @xmath5 is finite and ( modulo @xmath36 , @xmath37 ) none of the following conditions holds : _ [ cols= " > , < , > , < " , ] let @xmath74 be a reduced connected spherical picture over @xmath3 .
then the vertices ( discs ) of @xmath74 are given by figure 2.1(i ) and ( ii ) where @xmath75 denotes @xmath34 for @xmath76 ; and the star graph @xmath77 is given by figure 2.1(iii ) .
we make the following assumptions . 1 .
@xmath74 has a minimum number of vertices .
if @xmath78 then , subject to ( * a1 * ) , @xmath74 has a maximum number of regions of degree 2 with label @xmath79 .
observe that ( * a1 * ) implies that @xmath80 , @xmath81 where @xmath82 and @xmath83 in @xmath8 can not occur as sublabels of a region . for otherwise a sequence of bridge moves [ 4 ]
can be applied to produce a dipole which can then be deleted to obtain a picture with fewer vertices .
moreover if @xmath84 then ( * a2 * ) implies that @xmath79 can not be a proper sublabel and @xmath85 where @xmath86 , @xmath83 in @xmath8 can not be a sublabel of a region in @xmath74 . for otherwise bridge moves can be applied to increase the number of regions labelled @xmath79 while leaving the number of vertices unchanged . to prove asphericity we adopt the approach of [ 6 ]
let each corner in every region of @xmath74 be given an angle .
the _ curvature _ of a vertex is defined to be @xmath87 less the sum of the angles at that vertex .
the curvature @xmath88 of a @xmath89-gonal region @xmath90 of @xmath74 is the sum of all the angles of the corners of this region less @xmath91 .
our method of associating angles is to give each corner at a vertex of degree @xmath92 an angle @xmath93 . this way the vertices have zero curvature and we need consider only the regions .
thus if @xmath90 is a @xmath89-gonal region of @xmath74 ( a @xmath89-gon ) , denoted by @xmath94 , and the degree of the vertices of @xmath90 are @xmath95 ( @xmath22 ) then @xmath96 in fact since each @xmath97 ( @xmath22 ) we obtain @xmath98 so if @xmath99 then @xmath100 .
it follows from the fundamental curvature formula that @xmath101 is where the sum is taken over all the regions @xmath90 of @xmath74 .
our strategy to show asphericity will be to show that the positive curvature that exists in @xmath74 can be sufficiently compensated by the negative curvature .
to this end , as a first step , we located the regions @xmath90 of @xmath74 satisfying @xmath102 , that is , of positive curvature .
for each such @xmath90 we distribute all of @xmath88 to regions @xmath103 near @xmath90 . for such regions @xmath103 of @xmath74
define @xmath104 to equal @xmath105 plus all the positive curvature @xmath103 receives in the distribution procedure mentioned above with the understanding that if @xmath103 receives no positive curvature then @xmath106 .
observe then that the total curvature of @xmath74 is at most @xmath107 where the sum is taken over all regions @xmath103 of @xmath108 that are not positively curved regions . therefore to prove @xmath3 is aspherical it suffices to show that @xmath109 for each @xmath103 .
using the star graph @xmath77 of figure 2.1(iii ) we can list the possible labels of regions of small degree ( up to cyclic permutation and inversion ) .
@xmath110 @xmath111 allowing each element in @xmath112 to be either trivial or non - trivial yields 512 possibilities .
this number can be reduced without any loss as follows . 1 .
work modulo _
t - equivalence _ , that is , @xmath113 , @xmath114 .
2 . delete any combination that implies @xmath115 or @xmath116 .
3 . delete any combination that yields a contradiction ( for example @xmath117 , @xmath118 , @xmath119 .
4 . delete any combination that yields @xmath120 and @xmath118 or @xmath121 and @xmath122 ( see lemma 3.1(i ) ) .
when @xmath123 is cyclic it can be assumed that @xmath5 is finite ( see lemma 3.4(i ) ) .
it can be readily verified that there remain 23 cases partitioned according to the existence in @xmath74 of regions of degree 2 and are listed below .
@xmath124 : there are no regions of degree two
. + + ( * a0 * ) @xmath125 , @xmath126 , @xmath127 , @xmath128 , @xmath129 .
+ ( * a1 * ) @xmath130 , @xmath126 , @xmath127 , @xmath128 , @xmath129 . + ( * a2 * ) @xmath130 , @xmath131 , @xmath127 , @xmath128 , @xmath129 . + ( * a3 * ) @xmath125 , @xmath126 , @xmath127 , @xmath118 , @xmath132 , @xmath129 .
+ ( * a4 * ) @xmath125 , @xmath126 , @xmath127 , @xmath133 , @xmath134 , @xmath129 . + ( * a5 * ) @xmath130 , @xmath126 , @xmath127 , @xmath118 , @xmath132 , @xmath129 . + ( * a6 * ) @xmath130 , @xmath126 , @xmath127 , @xmath133 , @xmath134 , @xmath129 .
+ ( * a7 * ) @xmath130 , @xmath126 , @xmath127 , @xmath128 , @xmath135 , @xmath136 .
+ ( * a8 * ) @xmath130 , @xmath131 , @xmath127 , @xmath133 , @xmath134 , @xmath129 .
+ ( * a9 * ) @xmath130 , @xmath131 , @xmath127 , @xmath133 , @xmath134 , @xmath135 , @xmath137 .
+ ( * a10 * ) @xmath125 , @xmath126 , @xmath127 , @xmath133 , @xmath134 , @xmath135 , @xmath136 .
+ + * case b * : regions of degree two are possible .
+ + ( * b1 * ) @xmath78 , @xmath126 , @xmath127 , @xmath128 , @xmath129 . + ( * b2 * ) @xmath78 , @xmath138 , @xmath127 , @xmath128 , @xmath129 .
+ ( * b3 * ) @xmath78 , @xmath131 , @xmath127 , @xmath128 , @xmath129 .
+ ( * b4 * ) @xmath78 , @xmath126 , @xmath127 , @xmath139 , @xmath129 .
+ ( * b5 * ) @xmath78 , @xmath126 , @xmath127 , @xmath128 , @xmath135 , @xmath137 . + ( * b6 * ) @xmath78 , @xmath131 , @xmath127 , @xmath128 , @xmath135 , @xmath136 . + ( * b7 * ) @xmath125 , @xmath126 , @xmath140 , @xmath128 , @xmath129 .
+ ( * b8 * ) @xmath78 , @xmath126 , @xmath140 , @xmath128 , @xmath129 .
+ ( * b9 * ) @xmath130 , @xmath126 , @xmath140 , @xmath128 , @xmath129 .
+ ( * b10 * ) @xmath78 , @xmath138 , @xmath140 , @xmath128 , @xmath129 .
+ ( * b11 * ) @xmath78 , @xmath131 , @xmath140 , @xmath128 , @xmath129 .
+ ( * b12 * ) @xmath130 , @xmath131 , @xmath140 , @xmath128 , @xmath129 .
we first exhibit spherical diagrams corresponding to some of the conditions of theorems 1.1 and 1.2 . * note * : when drawing figures the discs ( vertices ) will often be represented by points ; the edge arrows shown in figure 2.1 will be omitted ; and regions with label @xmath141 will be labelled simply by @xmath142 . * lemma 3.1 * 1 . _
if any of the following conditions holds then @xmath3 fails to be aspherical : _ 1 . _ @xmath120 and @xmath118 ; _ 2 .
@xmath120 _ and @xmath143 ; _ 3 .
@xmath120 and @xmath144 .
if @xmath145 and @xmath146 is any of the following then @xmath3 fails to be aspherical .
@xmath147 _ _ ; _ _ 2 .
@xmath148 _ _ ; _ _ 3 .
@xmath149 ( @xmath150 ) _ ; _ 4 .
@xmath151 ( @xmath152 ) _ ; _ 5 .
@xmath153 ( @xmath150 ) _ ; _ + _ where _ @xmath154 .
if @xmath155 and @xmath146 is any of the following then @xmath3 fails to be aspherical .
@xmath147 _ _ ; _ _ 2 .
@xmath156 _ _ ; _ _ 3 .
@xmath157 ( @xmath152 ) _ ; _ 4 .
@xmath158 ( @xmath150 ) _ ; _ 5 .
@xmath149 ( @xmath150 ) _ , _ + _ where _ @xmath159 .
_ proof _ ( a)(i ) if @xmath160 there is a spherical picture over @xmath3 consisting of two @xmath161-gons labelled @xmath162 together with @xmath163 3-gons labelled @xmath164 .
the sphere for @xmath165 is shown in figure 3.1(i ) .
( ii ) if @xmath160 then there is a spherical picture over @xmath3 consisting of two regions with labels @xmath166 and between them two layers .
each layer has @xmath161 4-gons and @xmath161 3-gons with labels @xmath167 , @xmath168 and @xmath169 , @xmath170 .
the sphere for @xmath171 is shown in figure 3.1(ii ) .
( iii ) if @xmath172 then there is a spherical picture over @xmath3 consisting of a region labelled @xmath173 , a region labelled @xmath174 and a layer of @xmath163 regions between them .
the layer consists of @xmath161 2-gons labelled @xmath175 and @xmath161 4-gons labelled @xmath176 .
the sphere for @xmath171 is given by figure 3.1(iii ) .
( b)(i ) if @xmath160 then there is a spherical picture over @xmath3 consisting of two regions with label @xmath173 and between them a layer consisting of @xmath161 4-gons with label @xmath177 , @xmath161 2-gons with label @xmath178 and @xmath163 4-gons with label @xmath179 .
the case @xmath165 is shown in figure 3.1(iv ) .
( ii ) if @xmath180 then there is a spherical picture over @xmath3 consisting of two regions with label @xmath181 and between them a layer consisting of @xmath161 2-gons with label @xmath178 , @xmath161 2-gons with label @xmath182 and @xmath163 4-gons with label @xmath179 .
the case @xmath183 is shown in figure 3.1(v ) .
\(iii ) if @xmath184 a spherical picture is given by figure 3.2(i ) in which @xmath185 , @xmath186 , @xmath187 and @xmath188 .
if @xmath189 a spherical picture is given by figure 3.2(ii ) in which @xmath190 . if @xmath191 there is a spherical picture consisting of 30 a regions , 60 b regions , 20 c regions and 12 @xmath192 regions , half of which is shown in figure 3.2(iii ) .
\(iv ) if @xmath193 a spherical picture is given by figure 3.3(i ) in which @xmath194 , @xmath186 , @xmath195 and @xmath190 .
if @xmath196 there is a spherical picture consisting of 20 @xmath197 regions , 60 @xmath198 regions , 30 @xmath199 regions and 12 @xmath192 regions , half of which is shown in figure 3.3(ii ) .
( v ) spherical pictures for @xmath184 and 4 are given in figure 3.4(i ) and ( ii ) in which @xmath194 and @xmath200 ( respectively ) , @xmath186 , @xmath195 and @xmath188 .
if @xmath191 there is a spherical picture consisting of 12 @xmath201 regions , 60 @xmath198 regions , 30 @xmath199 regions and 20 @xmath108 regions half of which is shown in figure 3.4(iii ) .
( c)(i ) if @xmath202 then there is a spherical picture over @xmath3 consisting of two 4@xmath161-gons each with label @xmath203 together with @xmath161 2-gons with label @xmath204 , @xmath161 2-gons with label @xmath205 and @xmath163 2-gons with label @xmath206 . the sphere for @xmath183
is shown in figure 3.5(i ) .
( ii)if @xmath160 then there is a sphere consisting of two @xmath161-gons with label @xmath207 and between them a layer consisting of @xmath161 8-gons with label @xmath208 , @xmath163 2-gons with label @xmath209 and @xmath161 2-gons with label @xmath195 . the sphere for @xmath210 is shown in figure 3.5(ii ) .
\(iii ) a spherical picture for @xmath193 is given by figure 3.5(iii ) in which @xmath209 , @xmath211 , @xmath195 and @xmath190 .
if @xmath196 there is a sphere consisting of 60 @xmath197 regions , 20 @xmath198 regions , 30 @xmath199 regions and 12 @xmath192 regions half of which is given by figure 3.5(iv ) .
( iv ) spherical pictures for @xmath184 and 4 are given by figure 3.6(i ) and ( ii ) in which @xmath209 , @xmath208 , @xmath187 and @xmath188 and @xmath212 ( respectively ) .
if @xmath191 there is a spherical picture consisting of 60 @xmath197 regions , 30 @xmath198 regions , 20 @xmath199 regions and 12 @xmath192 regions half of which is given by figure 3.6(iii ) .
( v ) spherical pictures for @xmath184 and 4 are given by figure 3.7(i ) and ( ii ) in which @xmath209 ; @xmath213 and @xmath214 ( respectively ) , @xmath195 and @xmath188 .
if @xmath191 there is a spherical picture consisting of 60 @xmath197 regions , 12 @xmath215 regions , 30 @xmath199 regions and 20 @xmath108 regions half of which is given by figure 3.7(iii ) .
@xmath216 it follows from theorem 1(2 ) in [ 1 ] that if @xmath217 in @xmath13 then @xmath3 fails to be aspherical .
we apply this fact in the proof of the next lemma . * * lemma 3.2**__if any of the following conditions hold then @xmath3 fails to be aspherical . _ _ 1 .
@xmath218 , @xmath219 _ and _ @xmath220 .
@xmath120 _ and _ @xmath221 .
3 . @xmath120 _ and _ @xmath222 .
4 . @xmath223 .
5 . @xmath224 .
@xmath225 _ and _ @xmath226 . 7 .
@xmath227 _ and _ @xmath228
@xmath229 _ and _ @xmath230 .
_ proof _ 1 .
it is enough to show that the group @xmath231 has order @xmath232 .
now @xmath233 and @xmath234 = 1 \rangle$ ] .
let @xmath235 denote the covering 2-complex associated with @xmath236 [ 3 ] .
then @xmath235 has edges @xmath237 ( @xmath238 ) and 2-cells @xmath239 , @xmath240 , @xmath241 where @xmath242 and @xmath238 and the @xmath92 subscripts are mod @xmath243 . collapsing the maximal subtree
whose edges are @xmath244 ( @xmath238 ) , @xmath245 ( @xmath246 ) and using the lifts of @xmath182 shows that @xmath247 . using the lifts of the second relator
it is easily shown that @xmath248 where @xmath249 and , finally , using the lift of the third relator @xmath250 one can show that @xmath251 where @xmath252 .
we omit the details .
it is enough to show that @xmath253 has order @xmath254 .
now @xmath255 and @xmath256 .
let @xmath257 denote the covering complex associated with @xmath236 .
then @xmath257 has edges @xmath258 ( @xmath238 ) and 2-cells @xmath259 ( @xmath238 ) . collapsing the maximal tree
whose edges are @xmath260 implies @xmath261 .
the lifts of @xmath262 yield the relators @xmath263 for @xmath246 and @xmath264 .
the lifts of @xmath265 yield the relators @xmath266 .
it follows that @xmath267 where @xmath268 .
3 . here @xmath269 and @xmath120 implies @xmath217 .
( v ) a spherical picture for ( iv ) , ( v ) is shown in figure 3.8(i ) , ( ii ) ( respectively ) .
( viii ) for these cases we use gap [ 9 ] . for ( vi ) , @xmath270 together with the conditions yields @xmath271 ; for ( vii ) @xmath272 and @xmath273 implies @xmath274 ( respectively ) ; and for ( viii ) @xmath275 and @xmath230 implies @xmath276 . @xmath216 * * lemma 3.3**__if any of the conditions ( i)(iii ) of theorem 1.1 or ( i)(x ) of theorem 1.2 holds then @xmath3 fails to be aspherical . _ _ _ proof _ consider theorem 1.1 .
if ( i ) holds then @xmath3 fails to be aspherical by lemma 3.1(a)(i ) .
if ( ii ) holds then @xmath3 fails to be aspherical by lemma 3.2(i ) .
this leaves condition ( iii ) . if @xmath277 and @xmath278 then ( iii ) does not hold ; and if @xmath279 then @xmath5 is cyclic .
let @xmath140 .
since @xmath280 is @xmath281-equivalent to @xmath282 it can be assumed without any loss that @xmath283 .
the resulting ten cases are dealt with by lemma 3.1(c ) .
let @xmath284 . since @xmath285 is @xmath281-equivalent to @xmath286 it can again be assumed without any loss that @xmath283 .
the resulting ten cases are dealt with by lemma 3.1(b ) .
now consider theorem 1.2 .
if ( i ) holds then @xmath3 is aspherical by lemma 3.1(a)(i ) ; if ( ii ) holds then by lemma 3.1(a)(ii ) ; if ( iii ) holds then by lemma 3.2(ii ) ; if ( iv ) holds then by lemma 3.2(i ) ; if ( v ) holds then by lemma 3.2(iv ) ; if ( vi ) holds then by lemma 3.2(v ) ; if ( vii ) holds then by lemmas 3.1(a)(iii ) and 3.2(iii ) ; if ( viii ) holds then by lemma 3.2(vi ) ; if ( ix ) holds then by lemma 3.2(vii ) ; and if ( x ) holds then by lemma 3.2(viii ) .
@xmath216 a _ weight function _
@xmath287 on the star graph @xmath77 of figure 2.1(iii ) is a real - valued function on the set of edges of @xmath77 .
denote the edge labelled @xmath288 by @xmath289 ( respectively ) . the function @xmath287 is _ weakly aspherical _ if the following two conditions are satisfied : 1 . @xmath290 ; 2 .
each admissible cycle in @xmath77 has weight at least 2 .
if there is a weakly aspherical weight function on @xmath77 then @xmath3 is aspherical [ 2 ] .
* * lemma 3.4**__if any of the following conditions holds then @xmath3 is aspherical . _ _ 1 .
@xmath291 _ _ ; _ _ 2 .
@xmath292 _ and _ @xmath293 _ _ ; _ _ 3 . _
@xmath121 , @xmath120 and @xmath294 . _ _ proof _ the following functions @xmath287 are weakly aspherical .
1 . @xmath295 , @xmath296 .
2 . @xmath297 , @xmath298 , @xmath299 .
3 . @xmath300 , @xmath301 .
@xmath216 the following lemmas will be useful in later sections . * * lemma 3.5**__let @xmath302 where @xmath103 is a region of the spherical picture @xmath74 over @xmath3 .
_ _ 1 . _ if @xmath103 receives at most @xmath303 across each edge and @xmath304 then @xmath109
if @xmath103 receives at most @xmath303 across at most two - thirds of its edges , nothing across the remaining edges and @xmath305 then @xmath109 .
if @xmath103 receives at most @xmath306 across each edge and @xmath307 then @xmath109 .
if @xmath103 receives at most @xmath306 across at most two - thirds of its edges , nothing across the remaining edges and @xmath304 then @xmath109 .
_ 5 . _ if @xmath103 receives at most @xmath306 across at most half of its edges , nothing across the remaining edges and @xmath305 then @xmath109
if @xmath103 receives at most @xmath308 across at most half of its edges , nothing across the remaining edges and @xmath309 then @xmath109 .
if @xmath103 receives at most @xmath308 across at most three - fifths of its edges , nothing across the remaining edges and @xmath307 then @xmath310 .
_ _ proof _ recall that @xmath311 .
( i ) @xmath312 for @xmath304 .
( ii ) @xmath313 for @xmath304 . if @xmath191 then @xmath103 receives at most @xmath303 across at most three edges and so @xmath314 .
( iii ) @xmath315 for @xmath307 .
( iv ) @xmath316 for @xmath304 .
( v ) @xmath317 for @xmath304 .
if @xmath191 then @xmath103 receives at most @xmath306 across at most two edges and @xmath318 .
( vi ) @xmath319 for @xmath307 . if @xmath320 then @xmath103 receives at most @xmath308 across at most three edges and @xmath321 .
( vii ) @xmath322 for @xmath323 .
if @xmath324 then @xmath103 receives at most @xmath308 across five edges and @xmath325 ; if @xmath326 then @xmath103 receives across at most four edges and @xmath327 .
@xmath216 * * remark**we will use the above lemmas as follows .
suppose that @xmath103 receives positive curvature across its edge @xmath328 .
if we know that it then never receives curvatures across @xmath329 or across @xmath330 then we can apply the `` half '' results ; or if we know that it receives positive curvature across at most one of @xmath331 then we can apply the `` two - thirds '' results .
let @xmath103 be a region of @xmath74 and let @xmath332 be an edge of @xmath103 . if @xmath103 receives no curvature across @xmath332 then @xmath332 is called a _ gap _ ; if at most @xmath303 then @xmath332 is called a _ two - thirds gap _ ; and if at most @xmath306 then @xmath332 is called a _
half gap_. * * lemma 3.6 ( the four gaps lemma)**__if @xmath103 has a total of at least four gaps ( in particular , four edges across which @xmath103 does not receive any curvature ) and the most curvature that crosses any edge is @xmath308 then @xmath333__. _ proof _ it follows that @xmath334 . @xmath216
checking the star graph shows that we will have the following list for the labels of regions of degree @xmath89 where @xmath335 : if @xmath336 then @xmath337 . if @xmath338 then @xmath339 . if @xmath340 then @xmath341 . if @xmath342 then @xmath343 . if @xmath344 then @xmath345 . where , for example @xmath346 yields the eight labels @xmath347 , @xmath348 , @xmath349 , @xmath350 .
we will use the above list throughout the following sections often without explicit reference .
in this section we prove theorems 1.1 and 1.2 for case a , that is , we make the following assumptions : @xmath351 this implies that @xmath352 for each region @xmath90 of the spherical diagram @xmath74 .
if @xmath338 then we will fix the names of the fifteen neighbouring regions @xmath353 ( @xmath354 ) of @xmath90 as shown in figure 4.1(i ) and we use this notation throughout the section .
we treat each of the cases ( * a0*)(*a10 * ) in turn . *
( a0 ) * @xmath355 , @xmath356 , @xmath357 , @xmath358 , @xmath359.in this case @xmath360 for all regions @xmath90 .
since the degree of each vertex equals @xmath361 it follows that @xmath362 and so @xmath3 is aspherical . * ( a1 ) * @xmath363 , @xmath364 , @xmath357 , @xmath358 , @xmath359.if @xmath338 then @xmath365 , @xmath366 and @xmath90 is given by figure 4.1(ii ) .
observe that if @xmath367 for @xmath368 then @xmath369 and so @xmath370 .
( see the list of section 3 . )
but @xmath371 implies @xmath372 , a contradiction .
observe further that at most one of @xmath373 , @xmath374 , @xmath375 equals @xmath376 otherwise there is a contradiction to @xmath130 or @xmath126 .
this leaves the following cases : ( i ) @xmath377 ; @xmath378 ; ( ii ) @xmath379 ; ( iii ) @xmath380 ; ( iv ) @xmath381
. consider case ( i ) . here
@xmath382 for @xmath368 of figure 4.1(ii ) , so add @xmath383 to each of @xmath384 across the @xmath385 edge as shown .
observe from figure 4.1(ii ) that @xmath386 , say , does not receive any curvature from @xmath387 or @xmath388 .
therefore if @xmath103 receives positive curvature , since @xmath389 , it follows that @xmath333 by lemma 3.5(ii ) .
now consider case ( ii ) , @xmath390 . if @xmath382 for at least two of the @xmath353 where @xmath368 , say @xmath386 and @xmath391 , then add @xmath392 to each of @xmath393 and @xmath394 across the @xmath385 edge as in figure 4.1(iii ) . otherwise by symmetry
it can be assumed without any loss of generality that @xmath395 .
then add @xmath392 firstly to @xmath394 and then on to @xmath396 across the @xmath385 and @xmath397 edges as shown in figure 4.1(iv ) .
if @xmath398 then add the remaining @xmath392 to @xmath399 across the @xmath385 edge ; otherwise @xmath400 and similarly add the @xmath306 firstly to @xmath399 then on to @xmath401 . now observe that in figure 4.1(iii ) @xmath386 does not receive positive curvature from @xmath387 or @xmath388 and in figure 4.1(iv ) that @xmath387 does not receive positive curvature from @xmath386 or @xmath402 .
it follows that if @xmath103 receives positive curvature then @xmath103 receives curvature across at most half of its edges and so if @xmath389 , it follows that @xmath333 by lemma 3.5(v ) .
this leaves the case when @xmath403 and @xmath404 .
therefore @xmath405 . in each case @xmath406 together with @xmath407 implies @xmath131 , a contradiction .
now consider ( iii ) , @xmath380 .
as before , if @xmath408 and @xmath409 then add @xmath392 to each of @xmath393 and @xmath394 across the @xmath385 edge as in figure 4.1(iii ) . if @xmath410 then add @xmath392 firstly to @xmath393 and then on to @xmath396 across the @xmath385 and @xmath411 edges as shown in figure 4.2(i ) .
the remaining @xmath392 is distributed either to @xmath412 when @xmath413 or to @xmath414 via @xmath412 when @xmath415 .
now observe that in figure 4.1(iii ) @xmath386 does not receive positive curvature from @xmath387 or @xmath388 ; and in figure 4.2(i ) that @xmath387 does not receive positive curvature from @xmath391 or @xmath416 .
it follows that if the region @xmath103 receives positive curvature then it does so across at most half of its edges and therefore if @xmath389 then @xmath333 by lemma 3.5(v ) .
this leaves the case when @xmath417 and @xmath418 .
therefore @xmath419 and @xmath5 is cyclic .
if this occurs then add @xmath392 to @xmath420 as shown in figure 4.2(ii ) .
if @xmath421 then @xmath422 which , together with @xmath423 , contradicts the ( * a1 * ) assumptions . also note that in figure 4.2(ii ) , @xmath424 does not receive positive curvature from @xmath425 or from @xmath426 and so if @xmath103 receives positive curvature it does so across at most half its edges and @xmath109 . finally consider case ( iv ) , @xmath381 .
first assume that @xmath427 .
again if @xmath408 and @xmath409 then add @xmath392 to each of @xmath393 and @xmath394 across the @xmath385 edge as in figure 4.1(iii ) .
if @xmath395 then add @xmath392 firstly to @xmath393 and then on to @xmath396 across the @xmath385 and @xmath206 edges as shown in figure 4.2(iii ) and add the remaining @xmath428 to @xmath412 or @xmath414 as in the above .
now observe that in figure 4.1(iii ) @xmath386 does not receive positive curvature from @xmath387 or @xmath388 and in figure 4.2(iii ) that @xmath387 does not receive positive curvature from @xmath391 or @xmath416 .
thus if @xmath103 receives positive curvature then it does so across at most half of its edges so @xmath389 implies @xmath429 by lemma 3.5(v ) .
this leaves the case when @xmath417 and @xmath430 .
therefore @xmath431 and there is a contradiction to @xmath427 .
now assume that @xmath432 .
the distribution in this case is different in that we add @xmath433 to @xmath434 for @xmath435 as in figure 4.1(ii ) across the @xmath385 edge .
if @xmath436 , say , then distribute the @xmath303 further on to @xmath437 across the @xmath205 edge as shown in figure 4.2(iv ) .
therefore if @xmath103 receives positive curvature and if @xmath438 , it follows that @xmath333 by lemma 3.5(i ) .
this leaves the case when @xmath439 and @xmath440 . if @xmath403 then @xmath441 and @xmath442 therefore @xmath443 and there is a sphere by lemma 3.1(b)(iii ) .
so let @xmath444 . if @xmath103 receives curvature across at most three edges then @xmath445 .
checking ( the list ) now shows that @xmath429 except when @xmath446 and in this case we obtain a sphere by lemma 3.1(b)(iii )
. in conclusion @xmath3 fails to be aspherical in this case if and only if @xmath447 and @xmath448 .
note that if these conditions hold then @xmath5 is non - cyclic for otherwise we would obtain @xmath449 , a contradiction . *
( a2 ) * @xmath363 , @xmath450 , @xmath357 , @xmath358 , @xmath359.if @xmath338 then @xmath90 is given by figures 4.1(ii ) and 4.3(i ) . if @xmath340 and @xmath451 then @xmath452 otherwise there is a contradiction to one of the assumptions .
( throughout this case unless otherwise stated this means one of the ( * a2 * ) assumptions . )
the cases to be considered are ( where for example case ( ii ) means @xmath374 is the only member of @xmath453 to equal 1 ) : ( i ) @xmath378 , @xmath454 , @xmath455 , @xmath456 ; ( ii ) @xmath380 ; ( iii ) @xmath381 ; ( iv ) @xmath457 ; ( v ) @xmath458 ; ( vi ) @xmath459 ; ( vii ) @xmath460 ; ( viii ) @xmath461 ; ( ix ) @xmath457 , @xmath459 ; ( x ) @xmath457 , @xmath460 ; ( xi ) @xmath458 , @xmath459 ; ( xii ) @xmath458 , @xmath460 .
note that any other combination of these conditions gives a contradiction to one of the assumptions . moreover , ( ii ) is t - equivalent to ( viii ) ; ( iv ) is t - equivalent to ( vi ) ; ( v ) is t - equivalent to ( vii ) ; and ( x ) is t - equivalent to ( xi ) .
so it remains to consider ( i ) , ( ii ) , ( iii ) , ( iv ) , ( v ) , ( ix ) , ( x ) and ( xii ) .
now let @xmath462 and so @xmath463 . in cases ( i ) , ( ii ) , ( iv ) and ( v ) add @xmath433 to @xmath384 for @xmath368 as shown in figures 4.1(ii ) and 4.3(i ) . if @xmath382 then no further distribution takes place .
suppose without any loss of generality that @xmath436 .
this can not happen in case ( i ) ; in case ( ii ) @xmath386 is given by figure 4.3(ii ) and so add the @xmath303 from @xmath88 to @xmath437 across the @xmath385 and @xmath464 edges noting that @xmath465 otherwise @xmath466 which contradicts one of the assumptions ; in case ( iv ) @xmath386 is given by figure 4.3(iii ) and so add the @xmath303 from @xmath88 to @xmath396 across the @xmath385 and @xmath411 edges noting that @xmath467 otherwise @xmath468
which contradicts one of the assumptions or yields case ( ix ) or ( x ) ; and in case ( v ) @xmath90 is given by figure 4.3(iv ) and so add the @xmath303 from @xmath88 to @xmath437 across the @xmath385 and @xmath469 edges noting @xmath465 otherwise @xmath470 which contradicts one of the assumptions or yields case ( xi ) or ( xii ) . therefore if the region @xmath103 receives positive curvature then it receives @xmath303 across each edge and so if @xmath438 then @xmath333 by lemma 3.5(i ) .
this leaves the case when @xmath444 . after checking for vertex labels that contain the sublabels @xmath471 , @xmath472 , @xmath473 and @xmath474 corresponding to the edges crossed in figures 4.1(ii ) and 4.3(i)(iv ) we obtain @xmath475
this completes cases ( i ) , ( ii ) , ( iv ) and ( v ) .
consider case ( iii ) , @xmath381 .
if @xmath432 then we obtain a sphere by lemma 3.1(b)(iii ) .
note also that @xmath5 is non - cyclic in this case otherwise we obtain @xmath476 , a contradiction .
suppose then that @xmath427 .
first let @xmath366 .
add @xmath383 to @xmath384 for @xmath368 as in figure 4.1(ii ) .
if @xmath382 then no further distribution takes place .
suppose that @xmath436 . then add @xmath303 from @xmath88 to @xmath477 across the @xmath385 and @xmath206 edges as shown in figure 4.4(i ) noting that @xmath478 , otherwise @xmath479 which contradicts @xmath427 or one of the assumptions ; and if @xmath480 or @xmath400 in figure 4.1(ii ) then similarly add @xmath303 to @xmath481 or @xmath482 .
secondly , let @xmath483 .
add @xmath383 to @xmath384 for @xmath368 as in figure 4.3(i ) .
if @xmath382 then no further distribution takes place .
suppose without any loss of generality that @xmath436 .
then add @xmath303 from @xmath88 to @xmath396 across the @xmath397 and @xmath484 edges as shown in figure 4.4(ii ) , noting @xmath467 , otherwise @xmath485 which contradicts @xmath427 or one of the assumptions .
if @xmath480 or @xmath400 then similarly add @xmath303 to @xmath482 or @xmath414 .
if @xmath103 receives positive curvature and @xmath438,it follows by lemma 3.5(i ) that @xmath429 .
it remains to check @xmath444 . after checking for vertex labels that contain the sublabels @xmath471,@xmath472,@xmath486 and @xmath487 corresponding to the edges crossed in figures 4.1(ii ) , 4.3(i ) and 4.4(i)(ii )
we obtain @xmath475 or @xmath488 and this contradicts one of the assumptions .
therefore @xmath429 .
consider case ( ix ) , @xmath457 and @xmath459 . if @xmath382 for at least two of @xmath353 where @xmath368 , say @xmath386 and @xmath391,then add @xmath428 to each of @xmath393 and @xmath394 across the @xmath385 and @xmath397 edges as shown in figures 4.1(iii ) and 4.4(iii ) . by symmetry
it can be assumed that @xmath395 .
the two possibilities are given in figures 4.4(iv ) and 4.5(i ) and in both cases add @xmath392 to @xmath396 across the @xmath385 , @xmath411 or @xmath397 , @xmath464 edges as shown .
if @xmath398 then add the remaining @xmath392 to @xmath399 ; or if @xmath400 then apply the above to @xmath386 and @xmath489 to distribute the remaining @xmath392 similarly to @xmath477 .
now observe that if @xmath386 receives positive curvature from @xmath90 then it does not receive positive curvature from @xmath387 ; and if @xmath387 receives positive curvature from @xmath90 ( as in figures 4.4(iv ) and 4.5(i ) ) then it does not receive positive curvature from @xmath391 .
it follows that if the region @xmath103 receives positive curvature then it does so across at most two - thirds of its edges and therefore if @xmath103 receives positive curvature and if @xmath438 then @xmath333 by lemma 3.5(iv ) .
note that @xmath467 in figures 4.4(iv ) and 4.5(i ) otherwise @xmath490 which contradicts one of the assumptions .
so there remains the case @xmath444 and @xmath491 .
checking shows that in each case @xmath492 .
consider ( x ) , @xmath457 and @xmath460 .
first consider @xmath483 .
if at least two of the @xmath353 where @xmath368 have degree greater than four , say @xmath386 and @xmath391 , then add @xmath392 to @xmath393 and @xmath394 as shown in figure 4.4(iii ) .
so assume otherwise and without any loss of generality let @xmath395 as shown in figure 4.5(ii ) where @xmath467 otherwise @xmath493 which contradicts @xmath494 .
so add @xmath392 to @xmath396 as shown in figure 4.5(ii ) .
if @xmath398 in figure 4.5(ii ) add the remaining @xmath392 to @xmath399 otherwise use the same argument as above for @xmath386 and @xmath489 and add @xmath392 to @xmath477 .
now consider @xmath495 .
if at least two of the @xmath496 where @xmath497 have degree @xmath498 , say , @xmath499 and @xmath500 , then add @xmath392 to @xmath501 and @xmath502 as in figure 4.1(iii ) .
suppose exactly two of the @xmath496 have degree @xmath503 , say , @xmath499 and @xmath500 .
add @xmath392 to @xmath504 . if @xmath505 then add the remaining @xmath392 to @xmath506 as in figure 4.4(iv ) . if @xmath507 then add @xmath392 to @xmath508 as in figure 4.5(iii ) .
if now @xmath509 then @xmath510 and so add the @xmath392 to @xmath511 as in figure 4.6(i ) .
observe that @xmath512 forces @xmath513 otherwise there is a contradiction to @xmath494 .
finally suppose that @xmath514 for @xmath515 .
then @xmath392 is added to either @xmath481 , @xmath516 or @xmath517 exactly as above ; similarly @xmath518 is added to @xmath414 , @xmath425 or @xmath519 as shown in figure 4.6(i ) .
now observe that in figures 4.4(iii ) and 4.1(iii ) @xmath499 does not receive positive curvature from @xmath481 ; in figures 4.5(ii ) and 4.4(iv ) @xmath387 does not receive positive curvature from @xmath391 ; in figure 4.5(iii ) @xmath516 does not receive positive curvature from @xmath520 ; and in figure 4.6(i ) @xmath402 does not receive positive curvature from @xmath387 .
observe that if @xmath103 receives positive curvature then @xmath389 .
it follows from lemma 3.5(iv ) that if @xmath438 then @xmath521 so let @xmath444 .
if @xmath103 receives across at most two edges then @xmath429 so it remains to check if @xmath103 receives curvature from more than two edges . from the
above we see that positive curvature is transferred across @xmath522-edges .
the only two labels that contain more than two such sublabels and do not yield a contradiction are @xmath523 and @xmath524 as shown in figures 4.6(ii)(iii ) .
let @xmath525 as in figure 4.6(ii ) .
here @xmath90 receives nothing from @xmath526 or @xmath527 . if @xmath528 then @xmath90 receives nothing from @xmath529 and so @xmath429 . if @xmath530 then @xmath531 and @xmath90 receives nothing from @xmath532 via @xmath533 as in figure 4.4(iv ) and again @xmath429 .
let @xmath534 as in figure 4.6(iii ) .
here @xmath90 receives nothing from @xmath533 or @xmath527 . if @xmath535 then @xmath90 receives nothing from @xmath526 and so @xmath429 . if @xmath536 then @xmath528 and @xmath90 receives nothing from @xmath529 via @xmath532 and again @xmath429 .
finally consider case ( xii ) , @xmath537 .
then @xmath538 and so @xmath539 . if now @xmath432 then we obtain @xmath540 and @xmath541 is cyclic .
assume first that @xmath5 is non - cyclic so , in particular , @xmath542 .
add @xmath383 to @xmath384 for @xmath368 as in figures 4.1(ii ) and 4.3(i ) across the @xmath385 and @xmath397 edges . if , say , @xmath408 then no further distribution takes place . if @xmath436 then the @xmath433 is added to @xmath396 if @xmath366 across the @xmath385 and @xmath206 edges , or to @xmath477 if @xmath483 across the @xmath397 and @xmath175 edges as shown in figures 4.7(i)(ii ) .
observe that @xmath467 and @xmath478 .
if @xmath103 receives positive curvature and @xmath438 , it follows by lemma 3.5(i ) that @xmath429 .
it remains to check @xmath444 . after checking for vertex labels that contain the sublabels @xmath471,@xmath472,@xmath543 and
@xmath544 corresponding to the edges crossed in figures 4.1(ii ) , 4.3(i ) and 4.7(i)(ii ) it follows either that @xmath103 receives at most @xmath545 and so @xmath429 or @xmath546 which in each case yields a contradiction to @xmath5 non - cyclic .
now assume that @xmath5 is cyclic . if at least two of the @xmath496 where @xmath497 have degree greater than four , say @xmath499 and @xmath500 , then add @xmath392 to each of @xmath501 and @xmath502 as shown in figures 4.1(iii ) and 4.4(iii ) . by symmetry
assume then that @xmath547 .
the two possibilities are in figures 4.7(iii ) and 4.5(ii ) and in each case add @xmath428 to @xmath506 as shown . if @xmath413 then add the remaining @xmath392 to @xmath504 ; or if @xmath415 then similarly distribute the remaining @xmath392 via @xmath412 to @xmath482 or @xmath414 .
now observe that if @xmath499 receives positive curvature from @xmath90 it does not receive curvature from @xmath481 ; and if @xmath481 receives positive curvature from @xmath90 it does not receive positive curvature from @xmath499 in figure 4.7(iii ) or from @xmath500 in figure 4.5(ii ) .
it follows that if @xmath548 then @xmath549 by lemma 3.5(iv ) .
now @xmath550 in figures 4.7(iii ) and 4.5(ii ) so there remains the case @xmath444 and @xmath551 .
but checking shows that in all cases @xmath552
. in conclusion @xmath3 fails to be aspherical in this case if and only if @xmath553 ( and @xmath5 is non - cyclic ) . *
( a3 ) * @xmath355 , @xmath364 , @xmath357 , @xmath554 , @xmath555 , @xmath359.if @xmath338 then @xmath556 and @xmath90 is given by figure 4.7(iv ) .
add @xmath557 to @xmath393 across the @xmath558 edge .
if @xmath489 receives no curvature across at least four edges then there are at least four _ gaps _ as defined in section 3 and @xmath559 by lemma 3.6 .
this leaves the case when @xmath560 . if @xmath561 then there is a sphere by lemma 3.1(a)(i ) ; and if @xmath562 then @xmath5 is cyclic
. therefore @xmath3 is aspherical if and only if @xmath293 . *
( a4 ) * @xmath355 , @xmath364 , @xmath357 , @xmath563 , @xmath564 , @xmath359.if @xmath338 then @xmath565 and @xmath90 is given by figure 4.8(i ) .
first assume that @xmath5 is non - cyclic .
add @xmath557 to @xmath393 across the @xmath566 edge and note that @xmath567 forces @xmath568 , otherwise @xmath569 which contradicts @xmath126 or @xmath5 non - cyclic .
observe also that @xmath386 does not receive curvature from @xmath387 or @xmath388 and so if @xmath570 , it follows that @xmath571 by lemma 3.5(vi ) ; if @xmath572 then @xmath573 contains at most one occurrence of @xmath574 and so @xmath575 ; and if @xmath576 then @xmath573 contains at most two occurrence of @xmath574 and so @xmath577 .
now let @xmath5 be cyclic . if @xmath408 in figure 4.8(i ) then add @xmath557 to @xmath393 .
assume otherwise and @xmath436 .
then @xmath578 which yields a contradiction to the ( * a4 * ) assumptions except when @xmath579 .
if @xmath580 then there is a sphere by lemma 3.1(a)(ii ) so assume otherwise .
thus if @xmath436 then @xmath581 . in this case as shown in figure 4.8(ii ) add @xmath557 firstly to @xmath393 then on to @xmath477 . observe that @xmath478 otherwise @xmath582 which either contradicts @xmath125 , @xmath126 or @xmath129 , or implies the exceptional case ( * e1 * ) when @xmath583 . now observe that if @xmath386 receives positive curvature from @xmath90 then it does not receive positive curvature from @xmath387 or @xmath388 in figure 4.8(i ) ; and if @xmath388 receives positive curvature from @xmath90 then it does not receive positive curvature from @xmath489 or @xmath519 in figure 4.8(ii ) .
therefore if @xmath103 receives positive curvature then it does so across at most half of its edges which implies that if @xmath584 , then @xmath333 by lemma 3.5(vi ) .
if @xmath444 then checking shows that @xmath406 contains at most one occurrence of @xmath574 or @xmath585 .
it follows that @xmath333 .
if @xmath586 then checking shows that @xmath406 contains at most two occurrences of @xmath574 or @xmath585 and so @xmath333 . in conclusion in this case @xmath3 fails to be aspherical if and only if @xmath587 ( and so @xmath5 is cyclic ) . *
( a5 ) * @xmath363 , @xmath364 , @xmath357 , @xmath554 , @xmath555 , @xmath359.if @xmath588 then there is a sphere by lemma 3.1(a)(i ) and if @xmath293 then @xmath3 is aspherical by lemma 3.4(ii ) . *
( a6 ) * @xmath363 , @xmath364 , @xmath357 , @xmath563 , @xmath564 , @xmath359.if @xmath338 then @xmath90 is given by figure 4.1(ii ) and 4.8(iii ) .
first assume that either @xmath5 is non - cyclic or @xmath5 is cyclic but @xmath589 .
if @xmath366 then add @xmath433 to @xmath384 for @xmath368 across the @xmath385 edge .
note that if @xmath367 then @xmath590 which contradicts @xmath591 or @xmath589 and so @xmath592 . if @xmath565 then add @xmath557 to @xmath399 across the @xmath469 edge and note that if @xmath400 then @xmath593 which contradicts @xmath591 or @xmath589 and so @xmath594 .
observe now that @xmath386 does not receive curvature from @xmath387 or @xmath388 in figure 4.1(ii ) ; and @xmath489 does not receive curvature from @xmath595 or @xmath388 in figure 4.8(iii ) .
therefore if @xmath103 receives positive curvature then it does so across at most half of its edges which implies that if @xmath584 , then @xmath333 by lemma 3.5(vi ) .
it remains to consider @xmath596 .
if @xmath444 then checking shows that @xmath406 contains at most one occurrence of @xmath597 or @xmath598 .
it follows that @xmath333 .
if @xmath586 then checking shows that @xmath406 contains at most two occurrences of @xmath597 or @xmath598 and so @xmath599 and @xmath3 is aspherical .
now let @xmath600 .
this implies @xmath601 and @xmath449 .
it follows that if @xmath602 then @xmath603 . if @xmath604 then add @xmath392 to @xmath501 and to @xmath504 as shown in figure 4.8(iv ) .
let @xmath495 .
if at least two of @xmath605 , say @xmath606 , then add @xmath428 to @xmath501 and to @xmath502 as shown in figure 4.1(iii ) ; if say @xmath547 and @xmath607 then add @xmath428 to @xmath504 and @xmath428 to @xmath506 via @xmath500 as shown in figure 4.1(iv ) ; and if @xmath514 for @xmath497 then add @xmath383 to each of @xmath608 for @xmath609 similarly .
therefore if @xmath103 receives positive curvature then @xmath610 .
observe that @xmath499 does not receive positive curvature from @xmath481 in figure 4.1(iii ) ; @xmath611 does not receive positive curvature from @xmath612 respectively in figure 4.8(iv ) ; and @xmath481 does not receive positive curvature from @xmath517 in figure 4.1(iv ) .
it follows that if @xmath103 receives positive curvature then it does so across at most two - thirds of its edges .
therefore @xmath548 implies @xmath109 by lemma 3.5(iv ) .
since there are at most two occurrences of @xmath613 , @xmath614 , @xmath615 , @xmath616 in each of @xmath617 , @xmath618 , @xmath619 , @xmath620 it follows that if @xmath444 then again @xmath109 .
finally let @xmath621 .
this implies @xmath622 and @xmath449 .
it follows that if @xmath623 then @xmath624 .
let @xmath625 .
if at least one of @xmath496 where @xmath497 has degree @xmath626 , say @xmath627 , then add @xmath557 to @xmath501 as shown in figure 4.9(i ) ; or if @xmath628 for @xmath629 then add @xmath383 to each @xmath630 as shown in figure 4.9(ii ) in which @xmath631 .
let @xmath632 .
if @xmath633 then add @xmath557 to @xmath502 as in figure 4.8(v ) ; if @xmath634 and @xmath635 then add @xmath557 to @xmath636 via @xmath500 as in figure 4.9(iii ) ; and if @xmath634 and @xmath637 then add @xmath557 to @xmath504 as in figure 4.9(iv ) . finally if @xmath444 , @xmath638 and @xmath103 receives positive curvature from two neighbouring regions then add @xmath308 from @xmath104 to @xmath639 as shown in figure 4.10(i ) .
now observe that @xmath499 does not receive positive curvature from @xmath414 in figure 4.9(i ) ; @xmath499 , for example , does not receive from @xmath414 or @xmath481 in figure 4.9(ii ) ; @xmath500 does not receive from @xmath482 in figure 4.8(v ) ; @xmath482 does not receive from @xmath412 in figure 4.9(iii ) ; @xmath412 does not receive from @xmath482 or @xmath414 in figure 4.9(iv ) ; and @xmath640 does not receive from @xmath641 or @xmath642 in figure 4.10(i ) .
it follows that if @xmath103 receives positive curvature across two consecutive edges then @xmath643 and @xmath103 is given by figure 4.10(ii ) in which @xmath103 does not receive from @xmath640 or @xmath641 . since two sublabels of the form @xmath644 must be separated by a sublabel of length at least 2 it follows that if @xmath103 receives positive curvature then it does so across at most @xmath645 of its edges . therefore if @xmath646 then @xmath109 by lemma 3.5(vii ) ; and if @xmath406 does not involve @xmath647 then lemma 3.5(vi ) applies and @xmath648 for @xmath649 . if @xmath444 then either @xmath109 or @xmath103 is given by figure 4.10(i ) and @xmath650 .
if @xmath406 does not involve @xmath644 it remains to consider @xmath586 .
but checking shows that each candidate for @xmath406 contains at most two occurrences of @xmath615 , @xmath651 or @xmath652 so @xmath109 . finally if @xmath653 and @xmath586 then @xmath654 and @xmath109 ; or if @xmath655 then @xmath656 and @xmath657 . in conclusion
@xmath3 is aspherical for this case . *
( a7 ) * @xmath363 , @xmath364 , @xmath357 , @xmath358 , @xmath658 , @xmath659.if @xmath338 then @xmath660 and @xmath90 is given by figures 4.1(ii ) and 4.10(iii ) .
note that if @xmath661 , @xmath662 and @xmath514 then @xmath663 , @xmath664 ( respectively ) otherwise there is a contradiction to @xmath591 .
first assume that @xmath665 .
if @xmath495 then add @xmath433 to @xmath630 for @xmath497 across the @xmath385 edge as shown in figure 4.1(ii ) ; and if @xmath666 then add @xmath557 to @xmath504 as in figure 4.10(iii ) .
observe then that @xmath499 does not receive positive curvature from @xmath481 or @xmath414 in figure 4.1(ii ) ; and @xmath412 does not receive from @xmath482 or @xmath414 in figure 4.10(iii ) .
it follows that @xmath109 for @xmath649 by lemma 3.5(vi ) . if @xmath444 then checking shows that either @xmath406 contains at most one occurrence of @xmath597 or @xmath667 and so @xmath333 or @xmath668 . if @xmath586 then checking shows that @xmath406 contains at most two occurrences of @xmath597 or @xmath667 and so @xmath333 . thus if @xmath669 then @xmath103 is given by figure 4.10(iv ) and add @xmath670 to @xmath639 as shown .
now @xmath640 does not receive positive curvature from @xmath641 or @xmath642 in figure 4.10(iv ) and the above statements still hold for @xmath496 of figure 4.1(ii ) and @xmath412 of figure 4.10(iii ) .
therefore if @xmath103 receives positive curvature it does so across at most half of its edges and so @xmath109 for @xmath649 .
let @xmath671 . if @xmath406 does not involve @xmath672 then checking shows that @xmath109 so assume @xmath673 .
but if now @xmath674 then @xmath406 together with @xmath675 contradicts @xmath676 ; and if @xmath586 then the only labels with more than two occurrences of @xmath677 , @xmath615 or @xmath678 that do not yield a contradiction are @xmath679 , @xmath680 and @xmath681 in which case @xmath682 .
if @xmath683 then , since @xmath684 and @xmath685 , there is a sphere by lemma 3.1(a)(ii ) .
this leaves the case @xmath284 which implies @xmath449 and @xmath601 . if @xmath666 then again add @xmath557 for @xmath504 as in figure 4.10(iii ) .
let @xmath495 .
if at least two @xmath686 , say @xmath499 and @xmath500 , then add @xmath392 to @xmath501 and to @xmath502 as in figure 4.1(iii ) .
if , say , @xmath547 then add @xmath392 to @xmath506 via @xmath499 as in figure 4.2(iii ) and @xmath392 to @xmath504 if @xmath687 or to @xmath688 via @xmath412 if @xmath415 also . observe that @xmath412 does not receive positive curvature from @xmath482 or @xmath414 in figure 4.10(iii ) ; @xmath499 does not receive any from @xmath414 in figure 4.1(iii ) ; and @xmath481 does not receive any from @xmath500 in figure 4.2(iii ) .
it follows from lemma 3.5(v ) that if @xmath649 then @xmath109 . if @xmath586 then checking the allowable labels shows that @xmath406 involves at most two occurrences of @xmath615 , @xmath689 or @xmath690 and so @xmath109 . if @xmath444 then @xmath109 except when @xmath103 is given by figure 4.11(i ) in which case add @xmath691 to @xmath639 across the @xmath385 edge as shown ; and note that @xmath640 does not receive positive curvature from @xmath641 ( and so none of the above is affected ) .
if @xmath692 then we are done , otherwise @xmath693 .
let @xmath694 .
then add @xmath695 to @xmath641 via @xmath640 as shown in figure 4.11(ii ) noting that @xmath641 does not receive positive curvature from the region @xmath90 .
it follows that if @xmath669 then @xmath696 and @xmath697 . but
this implies that @xmath698 and @xmath109 . finally if @xmath640 is given by figure 4.11(iii )
then add @xmath699 to @xmath642 as shown and repeat the above argument .
this procedure must terminate at a region @xmath700 , say , where either @xmath701 or @xmath702 as in figure 4.11(ii ) , in which case @xmath703 is added to @xmath704 where @xmath705 and so @xmath701 . in conclusion
@xmath3 is aspherical if and only if @xmath706 .
( note that if @xmath380 then @xmath5 is cyclic . ) * ( a8 ) * @xmath363 , @xmath450 , @xmath357 , @xmath563 , @xmath564 , @xmath359.if @xmath338 then @xmath707 and @xmath90 is given by figures 4.1(ii ) , 4.3(i ) and 4.8(iii ) . if @xmath366 then add @xmath433 to @xmath384 for @xmath368 across the @xmath385 edge as shown in figure 4.1(ii ) .
note that if @xmath367 then @xmath590 which contradicts @xmath131 or @xmath708 and so @xmath592 . if @xmath565 then add @xmath557 to @xmath399 across the @xmath469 edge as shown in figure 4.8(iii ) and note that if @xmath400 then @xmath593 which again contradicts @xmath131 or @xmath708 and so @xmath594 . if @xmath483 then there are three subcases .
first assume that @xmath592 for @xmath368 .
then add @xmath709 to each @xmath384 across the @xmath397 edge as shown in figures 4.3(i ) .
now suppose that at least two of the @xmath353 have degree four , without loss of generality @xmath386 and @xmath391 .
then @xmath710 ( otherwise we will have a contradiction to @xmath708 ) which in turn forces @xmath711 and so @xmath467 otherwise @xmath712 which contradicts @xmath131 or @xmath708 . in this case add @xmath392 to @xmath396 and @xmath399 across the @xmath397 and @xmath385 edges as shown in figure 4.11(iv ) and if @xmath415 then in the same way add @xmath392 to @xmath688 .
this leaves the case where exactly one of @xmath353 has degree four and without loss assume @xmath436 .
then add @xmath392 to each of @xmath394 and @xmath399 as shown in figure 4.11(v ) .
observe that in figure 4.1(ii ) @xmath386 does not receive any positive curvature from @xmath387 or @xmath388 ; in figure 4.3(i ) @xmath386 does not receive positive curvature from @xmath387 ; in figure 4.8(iii ) @xmath489 does not receive any positive curvature from @xmath388 ; in figure 4.11(iv ) @xmath387 does not receive any positive curvature from @xmath391 or @xmath416 and @xmath489 does not receive any positive curvature from @xmath388 ; and in figure 4.11(v ) @xmath391 , @xmath489 ( respectively ) does not receive any positive curvature from @xmath595 , @xmath388 ( respectively ) .
let @xmath103 receive positive curvature and first suppose that @xmath406 does not involve @xmath713 .
it follows that @xmath103 receives positive curvature across at most half of its edges and so , since @xmath389 , @xmath333 by lemma 3.5(v ) .
now assume that @xmath714 and put @xmath302 . since positive curvature crosses @xmath471 , @xmath472 and @xmath473 edges only it follows that if @xmath406 involves at least two occurrences of @xmath713 then there are at least four gaps and @xmath333 by lemma 3.6 , so assume otherwise . moreover , @xmath714 forces @xmath103 to have at least two gaps and so @xmath715 which implies that if @xmath716 then @xmath333 . if @xmath444 then @xmath717 which contradicts @xmath718 or @xmath708 . if @xmath586 then checking shows that @xmath103 receives curvature across at most two edges and so @xmath333 . in conclusion in this case
@xmath3 is aspherical . *
( a9 ) * @xmath363 , @xmath450 , @xmath357 , @xmath563 , @xmath564 , @xmath658 , @xmath719.since @xmath5 is cyclic @xmath3 is aspherical by lemma 3.2(ii ) . * ( a10 ) * @xmath355 , @xmath364 , @xmath357 , @xmath563 , @xmath564 , @xmath658 , @xmath719.since @xmath5 is cyclic , @xmath120 and so @xmath3 is aspherical by lemma 3.2(ii ) .
it follows from ( * a0*)(*a10 * ) above that either @xmath3 is aspherical or modulo @xmath281-equivalence one of the conditions from theorem 1.1(i ) , ( iii ) or theorem 1.2(i ) , ( ii ) , ( iii ) is satisfied and so theorems 1.1 and 1.2 are proved for case a.
in this section we prove theorems 1.1 and 1.2 for case b , that is , we make the following assumption : at least one of @xmath204 , @xmath182 , @xmath672 equals 1 in @xmath5 .
if @xmath336 then we will fix the names of the four neighbouring regions @xmath353 ( @xmath720 ) of @xmath90 as shown in figure 5.1(i ) and we use this notation throughout the section .
* * remark**recall that if @xmath117 then @xmath79 can not be a proper sublabel .
this fact will be used often without explicit reference . * ( b1 ) * @xmath721 , @xmath364 , @xmath357 , @xmath358 , @xmath722.if @xmath336 then @xmath90 is given by figure 5.1(ii ) .
observe that if @xmath367 for @xmath723 then @xmath724 .
but @xmath725 implies @xmath372 , a contradiction .
observe further that at most one of @xmath726 equals 1 otherwise there is a contradiction to @xmath727 .
this leaves the following cases : 1 . in this case
@xmath382 for @xmath353 ( @xmath242 ) of figure 5.1(ii ) so add @xmath729 to each of @xmath384 ( @xmath730 ) .
observe from figure 5.1(ii ) that @xmath353 does not receive positive curvature from @xmath731 for @xmath732 .
it follows that if @xmath103 receives positive curvature then it does so across at most half of its edges and so @xmath733 implies that + @xmath429 by lemma 3.5(vi ) . checking ( the list of section 3 ) shows that if @xmath444 then @xmath103 receives positive curvature across at most one edge and so @xmath429 .
also if @xmath586 then checking shows @xmath103 receives positive curvature across at most two edges and so @xmath429 .
2 . suppose that @xmath734 , @xmath378 .
if @xmath408 and @xmath467 then add @xmath735 to @xmath393 and @xmath396 as in figure 5.1(ii ) .
if say @xmath736 as in figure 5.1(iii ) then @xmath737 which forces @xmath738 .
first assume that @xmath739 .
then @xmath740 and so add @xmath308 to @xmath394 via @xmath481 . if @xmath436 then add @xmath308 to @xmath636 via @xmath499 in a similar way .
observe that @xmath499 does not receive positive curvature from @xmath391 or @xmath595 in figure 5.1(ii ) ; and @xmath500 does not receive positive curvature from @xmath386 or @xmath489 in figure 5.1(iii ) .
it follows that if @xmath103 receives positive curvature then it does so across at most half of its edges and so @xmath733 implies that @xmath429 by lemma 3.5(vi ) .
it remains to study @xmath741 .
checking shows that if @xmath444 then either the label contradicts @xmath742 or @xmath103 receives positive curvature across at most one edge and so @xmath429 . also if @xmath586 then checking shows that @xmath103 receives positive curvature across at most two edges and so @xmath429 . now assume that @xmath461 , in which case @xmath743 , @xmath600 and @xmath744 .
the distribution of curvature is exactly the same except when @xmath634 in figure 5.1(iii ) . in this case
add @xmath745 to @xmath501 and @xmath746 to @xmath504 via @xmath500 as shown in figure 5.1(iv ) . together with the observations above ( which still hold ) we also have that @xmath499 does not receive positive curvature from @xmath500 , @xmath482 or @xmath425 and that @xmath412 does not receive positive curvature from @xmath517 or @xmath516 in figure 5.1(iv ) .
an argument similar to those given in the proof of lemma 3.5 now shows that if @xmath646 then @xmath109 ; and that if @xmath406 does not involve @xmath747 then @xmath109 for @xmath649 by lemma 3.5(vi ) .
the conditions on @xmath748 , @xmath749 and @xmath92 imply that if @xmath750 then @xmath751 so if @xmath406 does not involve @xmath747 it remains to consider @xmath586 . but
checking shows that @xmath406 will then either involve at most two non - adjacent occurrences of @xmath615 , @xmath752 or + @xmath753 and so @xmath109 or @xmath754 in which case @xmath755 . finally if @xmath756 and @xmath757 then @xmath758 and @xmath759 .
3 . now suppose that @xmath760 , @xmath377 .
first assume that @xmath761 .
add @xmath729 to @xmath393 and @xmath396 as in figure 5.1(ii ) .
if say @xmath736 then @xmath762 .
first let @xmath763 as in figure 5.2(i ) .
this forces @xmath764 and so @xmath765 forces @xmath766 .
but if @xmath621 then @xmath743 and there is a sphere by lemma 3.2(iv ) , so it can be assumed that @xmath767
. so add @xmath308 to @xmath636 via @xmath481 as shown .
suppose now that @xmath768 as in figure 5.2(ii ) .
this forces @xmath769 and so @xmath770 , otherwise there is a contradiction to @xmath761 .
so add @xmath308 to @xmath636 via @xmath481 as shown .
similarly add @xmath771 to @xmath502 if @xmath772 .
observe that @xmath386 does not receive positive curvature from @xmath391 or @xmath595 in figure 5.1(ii ) ; @xmath387 does not receive positive curvature from @xmath391 or @xmath595 in figure 5.1(ii ) ; and @xmath595 does not receive positive curvature from @xmath386 or @xmath416 in figures 5.2(i ) , ( ii ) .
it follows that if @xmath103 receives positive curvature then it does so across at most half of its edges and so @xmath733 implies that @xmath429 by lemma 3.5(vi ) .
it remains to study @xmath741 .
if @xmath773 then checking shows that if @xmath444 then either the label contradicts one of the ( * b1 * ) assumptions or @xmath103 receives positive curvature across at most one edge and so @xmath429 .
checking shows that if @xmath586 then @xmath103 receives positive curvature across at most two edges or @xmath754 contradicting @xmath774 , and so @xmath429 .
let @xmath718 .
if @xmath5 is cyclic then @xmath775 implies @xmath621 and we obtain a sphere as before , so assume otherwise . if @xmath718 and @xmath776 then we obtain a sphere by lemma 3.1(b)(iv ) .
so let @xmath718 , @xmath777 .
distribute curvature from @xmath90 as above and as shown in figure 5.2 .
checking shows that if @xmath444 then either the label contradicts @xmath718 or @xmath777 or @xmath103 receives positive curvature across at most one edge and so @xmath429 .
checking shows that if @xmath586 then @xmath103 receives positive curvature across at most two edges and so @xmath429 except when @xmath778 . this case is shown in figure 5.3(i ) where @xmath779 and so add @xmath780 to @xmath781 for @xmath782 across the edge @xmath783 . if @xmath784 then @xmath785 .
suppose that @xmath786 as in figure 5.3(ii ) .
then @xmath787 and @xmath788 otherwise there is a contradiction to @xmath5 non - cyclic .
so add @xmath303 to @xmath789 across the edge @xmath566 .
if @xmath790 as in figure 5.3(iii ) then @xmath791 and @xmath788 otherwise there is a contradiction to @xmath777 .
so add @xmath303 to @xmath789 across the edge @xmath558 .
observe that if @xmath103 receives positive curvature then it receives @xmath308 across the edges @xmath206 , @xmath411 or @xmath385 ; and receives @xmath303 across the edges @xmath783 , @xmath566 or @xmath792 .
thus there is a gap preceding @xmath793 , @xmath748 , @xmath794 and a gap after @xmath793 , @xmath795 , @xmath796 and there is a two - thirds gap across the edges @xmath783 , @xmath566 and @xmath792 .
also there is always a gap between two @xmath92 s ( other than when the subword is @xmath797 ) .
now since @xmath204 can not be a proper sublabel it follows that if there are at least two occurrences of @xmath793 then we obtain four gaps .
suppose now that there is at most one occurrence of @xmath793 .
if there is exactly one occurrence of @xmath749 and either no occurrences of @xmath748 or no occurrences of @xmath92 then @xmath5 is cyclic , a contradiction ; and if there are no occurrences of @xmath749 and exactly one occurrence of @xmath92 or of @xmath748 then again @xmath5 is cyclic , a contradiction .
so assume otherwise .
it follows that @xmath406 contains at least four gaps or @xmath798 .
but since @xmath799 each of these labels contradicts @xmath5 non - cyclic , @xmath777 or @xmath718 except when @xmath800 . in this case
if @xmath801 then it can be assumed without loss of generality that @xmath103 is given by figure 5.3(iv ) and @xmath103 receives @xmath802 from @xmath803 .
this implies that @xmath804 and @xmath805 .
so add @xmath303 from @xmath105 to @xmath789 . since this @xmath303 is across an @xmath783 edge and since @xmath806 it follows from the above that @xmath807 .
if @xmath808 in figure 5.3(iv ) then a similar argument applies to @xmath809 . finally let @xmath810 .
in particular , @xmath5 is non - cyclic for otherwise @xmath811 , a contradiction . if @xmath588 then we obtain a sphere by lemma 3.1(b)(i ) and if @xmath293 then @xmath3 is aspherical by lemma 3.4(ii ) . in conclusion
@xmath3 is aspherical in this case except when @xmath5 is non - cyclic , @xmath812 and either @xmath718 , @xmath813 or @xmath810 , @xmath120 ; or when @xmath5 is cyclic , @xmath621 , @xmath743 and @xmath744 . *
( b2 ) * @xmath721 , @xmath814.if @xmath5 is cyclic then @xmath538 .
therefore @xmath815 and there is a sphere by lemma 3.2(i ) .
so suppose from now on that @xmath5 is non - cyclic .
if @xmath336 then @xmath90 is given by figures 5.1(ii ) and 5.4(i ) .
moreover , if @xmath367 for @xmath723 then @xmath816 .
but @xmath817 implies that @xmath476 which is a contradiction .
this leaves the following cases : 1 . in this case @xmath408 and @xmath467 in figures 5.1(ii ) and 5.4(i ) so add @xmath735 to each of @xmath393 and @xmath396 .
observe that @xmath386 and @xmath387 do not receive positive curvature from @xmath391 or @xmath595 in figures 5.1(ii ) and 5.4(i ) .
it follows that if @xmath103 receives positive curvature then it does so across at most half of its edges and so @xmath584 implies that @xmath429 by lemma 3.5(vi ) .
it remains to study @xmath741 .
checking shows that if @xmath444 then either the label contradicts @xmath820 or @xmath103 receives positive curvature across at most one edge and so @xmath429 . also if @xmath586 then @xmath103 receives positive curvature across at most two edges and so @xmath429 .
2 . in this case the labels @xmath821 can occur .
if @xmath822 then we obtain spheres by lemma 3.1(b)(ii ) and if @xmath823 then @xmath3 is aspherical by lemma 3.4(iii ) .
3 . in this case the labels @xmath371 can occur and @xmath5 is non - abelian . if @xmath408 and @xmath467 as in figures 5.1(ii ) and 5.4(i ) then add @xmath735 to @xmath393 and @xmath396 .
if say @xmath736 as in figure 5.4(ii ) then @xmath824 and so @xmath409 , otherwise there is a contradiction to @xmath825 or @xmath826 .
so add @xmath308 to @xmath394 via @xmath481 as shown . if @xmath772 then similarly add @xmath729 to @xmath636 via @xmath499 .
we see from figures 5.1(ii ) and 5.4(i)(ii ) that if @xmath103 receives positive curvature then it does so across the edges @xmath827 or @xmath828 .
it follows that the only word of length 3 that contains no gaps is @xmath829 .
suppose that @xmath830 where @xmath831 and @xmath832 have length 2 and @xmath833 has length at least 0 .
then there is a gap preceding @xmath796 in @xmath834 and after @xmath794 in @xmath835 .
moreover , if @xmath831 does not contain a gap then @xmath836 and if @xmath832 does not contain a gap then @xmath837 .
it follows that if @xmath838 then @xmath406 contain at least 4 gaps and @xmath429 . if @xmath839 and @xmath840 then @xmath841 and there are 4 gaps ; if @xmath839 and @xmath842 then @xmath843 and again there are 4 gaps . if now @xmath844 and @xmath697 then either @xmath845 , a contradiction , or there are 4 gaps .
now suppose that @xmath406 does not contain the subword @xmath829 .
then @xmath103 receives positive curvature across at most half of its edges and so @xmath584 implies @xmath429 by lemma 3.5(vi ) .
so let @xmath846 and @xmath847 .
if @xmath848 then checking shows that there is a contradiction to @xmath5 non - abelian ; and if @xmath586 then checking shows that @xmath103 receives positive curvature across at most two edges and so @xmath429 .
4 . in this case
the labels @xmath849 can occur and @xmath5 is non - abelian .
if @xmath408 and @xmath467 as in figures 5.1(ii ) and 5.4(i ) then add @xmath735 to @xmath393 and @xmath396 .
if say @xmath736 then @xmath850 which forces @xmath851 and so @xmath409 , otherwise there is a contradiction to @xmath826 or @xmath852 .
so add @xmath308 to @xmath401 via @xmath481 as shown in figure 5.4(iii ) .
if @xmath772 then similarly add @xmath853 to @xmath502 via @xmath499 .
we see from figures 5.1(ii ) , 5.4(i ) and 5.4(iii ) that if @xmath103 receives positive curvature then it does so across the edges @xmath385 , @xmath464 or @xmath397 .
it follows that the only word of length 3 that contains no gaps is @xmath854 .
suppose that @xmath855 where @xmath831 and @xmath832 have length 2 and @xmath833 has length at least 0 .
then there is a gap preceding @xmath748 in @xmath856 and after @xmath795 in @xmath857 .
moreover , if @xmath831 does not contain a gap then @xmath858 and if @xmath832 does not contain a gap then @xmath859 .
it follows that if @xmath838 then @xmath406 contain at least 4 gaps and @xmath429 .
if @xmath839 and @xmath860 then @xmath861 and there are 4 gaps ; if @xmath839 and @xmath862 then @xmath863 and again there are 4 gaps .
if @xmath864 and @xmath846 then there are 4 gaps .
now suppose that @xmath406 does not contain the subword @xmath854 .
then @xmath103 receives positive curvature across at most half of its edges and so @xmath584 implies @xmath429 by lemma 3.5(vi ) .
so let @xmath846 and @xmath865 .
if @xmath848 then checking shows that there is a contradiction to @xmath5 non - abelian ; and if @xmath586 then checking shows that @xmath103 receives positive curvature across at most two edges and so @xmath429 .
5 . in this case the labels
@xmath371 and @xmath849 can occur . if @xmath294 then @xmath3 is aspherical by lemma 3.4(iii ) ; and if @xmath866 then there is a sphere by lemma 3.2(i ) . *
( b3 ) * @xmath721 , @xmath868.if @xmath336 then @xmath90 is given by figure 5.1(ii ) . if @xmath338 then @xmath90 is given by figure 4.3(i ) .
moreover , if @xmath367 and @xmath661 or @xmath869 then@xmath816 .
but each of @xmath870,@xmath725 and @xmath819 implies a contradiction to one of the ( * b3 * ) assumptions .
thus we have the following cases : 1 . in this case @xmath408 and @xmath467 in figure 5.1(ii ) , so add @xmath735 to each of @xmath393 and @xmath396 . in figure 4.3(i ) @xmath408 , @xmath409 and @xmath398 so add @xmath872 to each of @xmath393 , @xmath394 and @xmath399 . observe that @xmath386 and @xmath387 do not receive positive curvature from @xmath391 or @xmath595 in figure 5.1(ii ) .
also @xmath386 , @xmath391 and @xmath489 do not receive positive curvature from @xmath873 for @xmath874 in figure 4.3(i ) .
it follows that if @xmath103 receives positive curvature then it does so across at most half of its edges and so @xmath584 implies that @xmath429 by lemma 3.5(vi ) .
it remains to study @xmath741 .
checking shows that if @xmath444 then either the label contradicts + @xmath133 or @xmath103 receives positive curvature across at most one edge and so @xmath429 . also if @xmath586 then @xmath103 receives positive curvature across at most two edges and so @xmath429 .
2 . in this case the labels @xmath821 can occur .
if @xmath5 is cyclic then @xmath875 , @xmath876 and there is a sphere by lemma 3.2(v ) , so assume that @xmath5 is non - cyclic . if @xmath877 then we obtain spheres by lemma 3.1(b)(i ) , ( v ) , so assume that @xmath878 . in figure 5.1(ii ) if @xmath408 and @xmath467 then add @xmath735 to @xmath393 and @xmath396 as shown .
if say @xmath408 and @xmath736 as in figure 5.5(i ) then add @xmath308 to @xmath393 .
this implies that @xmath762 as shown .
this forces @xmath879 and so @xmath409 , otherwise there is a contradiction to @xmath878 so add @xmath308 to @xmath394 via @xmath481 as shown .
if @xmath880 then add @xmath308 to @xmath434 for @xmath732 as shown in figure 5.5(ii ) .
the one exception to the above is when @xmath881 and @xmath882 .
then @xmath770 and in this situation add the @xmath308 from @xmath88 to @xmath401 via @xmath499 as shown in figure 5.5(iii ) .
the same applies to @xmath387 . in figure 4.3(i ) if @xmath408 , @xmath409 and @xmath398 then add @xmath872 to each of @xmath393 , @xmath396 and @xmath399 . if say @xmath436 , @xmath409 and @xmath398 then @xmath883 and @xmath467 otherwise there is a contradiction to @xmath878 so add @xmath303 to @xmath396 , @xmath394 and @xmath399 as shown in figure 5.6(i ) .
now suppose that @xmath436 and @xmath480 .
this implies that @xmath884 as shown in figure 5.6(ii ) .
so add @xmath303 to @xmath396 , @xmath401 and @xmath399 .
if @xmath885 then similarly add @xmath303 to @xmath886 for @xmath874 .
we now see that if @xmath103 receives positive curvature then it receives at most @xmath308 across @xmath597 and @xmath887 ; and it receives at most @xmath303 across @xmath888 and @xmath889 .
thus there is always a gap immediately preceding @xmath749 and @xmath890 ; and there is a gap immediately after @xmath891 and @xmath92 .
this implies that if there are at least four occurrences of @xmath793 or @xmath892 then @xmath406 contains at least four gaps and so @xmath333 by lemma 3.6 .
suppose that there are at most three occurrences of @xmath793 or @xmath892 in @xmath406 .
observe that in addition to the four gaps mentioned above the following sublabels yield gaps : @xmath893 and @xmath713 each yields a gap ; @xmath894 yields two gaps ( see figure 5.5(iii ) ) ; and @xmath888 and @xmath889 + each yields the equivalent of a two - thirds gap . if @xmath895 where @xmath896 then @xmath406 obtains at least four gaps since @xmath878 . if @xmath897 then @xmath5 is cyclic so it can be assumed that @xmath406 involves either two or three occurrences of @xmath793 or @xmath892 .
it follows that if there are three occurrences then @xmath109 ; or if exactly two occurrences then either @xmath109 or@xmath898 .
but each of these labels forces @xmath5 cyclic or @xmath899 or a ( * b3 * ) contradiction , therefore @xmath333 by lemma 3.6 .
3 . in this case
the label @xmath167 can occur .
first assume that @xmath5 is non - cyclic . in figure 5.1(ii )
@xmath408 and @xmath467 otherwise there is a contradiction to @xmath78 or @xmath131 , so add @xmath729 to @xmath393 and @xmath396 . in figure 4.3(i ) if @xmath408 , @xmath409 and @xmath398 add @xmath872 to each of @xmath393 , @xmath396 and @xmath399 .
if say @xmath436 only then add @xmath306 to @xmath394 and @xmath399 .
now suppose that @xmath395 .
this implies that their label is @xmath167 which forces @xmath900 as shown in figure 5.6(iii ) .
so add @xmath306 to @xmath396 via @xmath499 and to @xmath399 . finally if @xmath885 then in a similar way add @xmath303 to @xmath886 for @xmath874 .
observe that @xmath386 does not receive positive curvature from @xmath391 and @xmath387 does not receive positive curvature from @xmath595 in figure 5.1(ii ) . in figure 4.3(i ) @xmath386 does not receive positive curvature from @xmath387 . in figure 5.6(iii
) @xmath387 does not receive positive curvature from @xmath391 .
since @xmath103 receives @xmath308 across the @xmath385 edge and @xmath306 across the @xmath397 , @xmath484 edges it follows that @xmath103 receives an average of @xmath306 across each of its edges , so @xmath901 implies that @xmath429 .
it remains to study @xmath902 .
checking shows that if @xmath444 then either the label contradicts @xmath131 or @xmath5 non - cyclic or @xmath103 receives positive curvature across at most one edge and so @xmath429 , except when @xmath903 as in figure 5.7(i ) . in this case
@xmath103 receives @xmath308 from @xmath803 and @xmath306 from @xmath904 . if @xmath542 then this implies that @xmath905 and so add @xmath306 to @xmath906 noting that this is a similar edge to the one crossed in figure 5.6(iii ) so there is no change to the above argument and @xmath429 in this case .
suppose now that @xmath810 and @xmath907 as in figure 5.7(ii ) .
if + @xmath908 then add @xmath306 to @xmath789 across the @xmath909 edge .
if @xmath910 then @xmath911 which implies that @xmath912 and so if @xmath913 then add @xmath306 to @xmath809 as in figure 5.7(iii ) .
if @xmath914 then @xmath915 and this forces @xmath916 and so @xmath917 otherwise there is a contradiction to @xmath742 .
so add @xmath306 to @xmath918 again as shown in figure 5.7(iii ) .
observe that @xmath533 in figure 5.7(ii ) can now receive @xmath306 from @xmath105 , however it receives no positive curvature from @xmath532 or any other region across the @xmath909 edge .
moreover , it is clear from figure 5.7(iii ) that @xmath527 receives only the @xmath306 from @xmath533 across its @xmath397 edge ; and @xmath919 receives only the @xmath306 from @xmath527 across its @xmath484 edge . finally observe that figures 5.7(ii)(iii ) do not alter the fact that @xmath386 does not receive positive curvature from @xmath391 and @xmath387 does not receive positive curvature from @xmath595 in figure 5.1(ii ) .
therefore the average positive curvature that @xmath103 receives across each edge is still @xmath306 and so if @xmath901 then @xmath429 . it remains to check @xmath920 for the sublabels @xmath921 and @xmath922 , @xmath923 .
checking shows that if @xmath586 then the most curvature that @xmath103 can receive is either @xmath924 or @xmath925 or @xmath926 and so @xmath310 . if @xmath655 then the most curvature received is @xmath927 or @xmath928 or @xmath929 or @xmath930 and @xmath648 except for @xmath931 ; but this implies @xmath932 , a contradiction .
now let @xmath5 be cyclic
. then @xmath933 and @xmath876 .
again add @xmath771 to each of @xmath501 , @xmath506 as in figure 5.1(ii ) . in figure 4.3(i ) if say @xmath627 then add @xmath557 to @xmath501 and so it can be assumed that @xmath934 for @xmath935 in which case @xmath936 .
if say @xmath772 then add @xmath557 to @xmath688 via @xmath499 as shown in figure 5.7(iv ) . it can be assumed then that @xmath628 for @xmath497 in which case add @xmath433 to each @xmath105 via @xmath496 where @xmath497 as shown in figure 5.8(i ) . if say @xmath937 and @xmath938 then repeat the above , that is , add the @xmath303 from @xmath88 across another @xmath397 edge and continue in this way until @xmath303 is eventually added to a region @xmath700 where either @xmath939 ( and so the process terminates ) or @xmath702 in which case the @xmath303 from @xmath88 is added to @xmath704 as shown in figure 5.8(ii ) , where @xmath184 .
if @xmath940 then the process terminates ( and note that @xmath941 ) ; otherwise @xmath942 and the @xmath303 from @xmath88 is added to @xmath943 where @xmath944 is the region shown in figure 5.8(ii ) with @xmath184 .
observe that @xmath945 so @xmath946 and the process terminates .
this completes the distribution of curvature that occurs .
it follows that if @xmath103 receives positive curvature across an edge @xmath328 say then @xmath103 does not receive any curvature across the adjacent edges @xmath329 , @xmath330 except when @xmath103 is given by @xmath947 in figure 5.8(ii ) .
therefore if @xmath406 does not involve @xmath747 then lemma 3.5(vi ) applies and @xmath109 for @xmath649 ; and if @xmath586 then checking for @xmath615 , @xmath752 and @xmath948 shows that @xmath103 receives positive curvature across at most two edges and @xmath109 . finally if @xmath949 then we see from figure 5.8(ii )
that the maximum amount @xmath103 receives is on average @xmath950 across @xmath951 of its edges and so if @xmath646 then @xmath109 by lemma 3.5(vii ) .
checking shows that if @xmath952 then @xmath953 and so if @xmath954 then @xmath103 receives curvature across at most @xmath955 edges ( respectively ) and @xmath109 . *
( b5 ) * @xmath721 , @xmath364 , @xmath357 , @xmath358 , @xmath658 , @xmath659.if @xmath336 then @xmath90 is given by figure 5.1(ii ) in which case add @xmath729 to each of @xmath501 and @xmath506 as shown ; and if @xmath338 then @xmath90 is given by figure 4.10(iii ) in which case add @xmath557 to @xmath504 as shown .
assume that @xmath5 is non - cyclic .
if say @xmath772 in figure 5.1(ii ) then @xmath960 which contradicts @xmath591 or @xmath5 non - cyclic ; and if @xmath415 in figure 4.10(iii ) then @xmath961 which contradicts @xmath591 or @xmath5 non - cyclic .
observe that in figure 5.1(ii ) @xmath499 say does not receive positive curvature from @xmath500 or @xmath482 ; and in figure 4.10(iii ) @xmath412 does not receive any from @xmath482 or @xmath414 .
therefore @xmath310 for @xmath649 by lemma 3.5(vi ) and so it remains to consider @xmath962 .
but checking for subwords @xmath963 shows that if @xmath964 then @xmath103 receives positive curvature across @xmath965 edges respectively except when @xmath966 .
but @xmath5 would then be cyclic , so @xmath109 .
now assume that @xmath5 is cyclic . if @xmath380 then the conditions are @xmath281-equivalent to those of lemma 3.1(a)(ii ) and there is a sphere , so assume otherwise .
then @xmath882 in figure 5.1(ii ) and @xmath413 in figure 4.10(iii ) .
the argument above now applies except when @xmath967 in which case add @xmath771 to @xmath502 via @xmath499 as shown in figure 5.8(iii ) .
we see from figures 4.10(iii ) , 5.1(ii ) and 5.8(iii ) that if @xmath103 receives positive curvature then it does so across at most half of its edges and so @xmath109 for @xmath649 by lemma 3.5(vi ) , so let @xmath968 .
then either @xmath969 and @xmath970 ; or @xmath586 , @xmath103 receives across at most two edges ( checking for @xmath971 ) and again @xmath109 . * ( b6 ) * @xmath721 , @xmath450 , @xmath357 , @xmath358 , @xmath658 , @xmath659.if @xmath336 then @xmath90 is given by figure 5.1(ii ) .
if @xmath338 then @xmath90 is given by figures 4.3(i ) and 4.10(iii ) . in figure 5.1(ii ) if @xmath408 and @xmath467 then add @xmath735 to @xmath393 and @xmath396 .
if @xmath408 and @xmath736 then @xmath973 and so @xmath974 otherwise there is a contradiction to @xmath131 . therefore @xmath387 is given by figure 5.1(iii ) forcing @xmath738 and so @xmath409 , otherwise there is a contradiction to @xmath78 , so add @xmath308 to @xmath393 and to @xmath394 via @xmath481 . if @xmath436 then add @xmath729 to @xmath636 via @xmath499
observe in figure 4.10(iii ) that if @xmath975 then @xmath976 which contradicts @xmath78 so add @xmath557 to @xmath399 as shown . in figure 4.3(i ) if @xmath367 where @xmath497 then @xmath977 which contradicts @xmath78 so add @xmath978 to @xmath384 as shown .
we see from figures 4.3(i ) , 4.10(iii ) and 5.1(ii)(iii ) that if @xmath103 receives positive curvature then it does so across the edges @xmath385 , @xmath397 or @xmath979 .
it follows that the only word of length 3 that contains no gaps is @xmath829 .
suppose that @xmath830 where @xmath831 and @xmath832 have length 2 and @xmath833 has length at least 0 .
then there is a gap preceding @xmath796 in @xmath834 and after @xmath794 in @xmath835 .
moreover , if @xmath831 does not contain a gap then @xmath836 and if @xmath832 does not contain a gap then @xmath837 .
it follows that if @xmath838 then @xmath406 contain at least 4 gaps and @xmath429 .
if @xmath839 and @xmath840 then @xmath841 and there are 4 gaps ; if @xmath839 and @xmath842 then @xmath843 and again there are 4 gaps .
if @xmath980 and @xmath846 then either we obtain a contradiction to @xmath78 or @xmath981 and there are 4 gaps . now suppose that @xmath406 does not contain the subword @xmath829 .
then @xmath103 receives positive curvature across at most half of its edges and so @xmath584 implies @xmath429 by lemma 3.5(vi ) .
so let @xmath846 and @xmath847 . if @xmath848 then checking shows that there is a contradiction to a ( * b6 * ) assumption or @xmath982 and there are four gaps ; and if @xmath586 then checking for @xmath983 shows that either there is a contradiction to a ( * b6 * ) assumption or @xmath103 receives positive curvature across at most two edges and so @xmath429
. therefore @xmath3 is aspherical . *
( b7 ) * @xmath355 , @xmath364 , @xmath984 , @xmath358 , @xmath359.if @xmath336 then @xmath90 is given by figure 5.9(i ) . here
@xmath985 and @xmath986 .
first assume that @xmath5 is non - cyclic .
this implies that @xmath408 , @xmath467 , otherwise there is a contradiction to @xmath5 non - cyclic , @xmath125 , or @xmath126 .
so add @xmath771 to @xmath393 and @xmath396 .
observe that each of @xmath386 and @xmath387 does not receive positive curvature from @xmath731 for @xmath732 .
it follows that if @xmath103 receives positive curvature then it does so across at most half of its edges and so @xmath584 implies that @xmath429 by lemma 3.5(vi ) .
checking shows that if @xmath444 then @xmath406 contradicts @xmath125 , @xmath126 or @xmath5 non - cyclic . also if @xmath586 then @xmath103 receives positive curvature across at most two edges and so @xmath429 .
now assume that @xmath5 is cyclic . if @xmath987 there is a sphere by lemma 3.1(a)(iii ) and lemma 3.2(iii ) so assume from now on that @xmath988 .
if @xmath989 and @xmath990 then @xmath3 is aspherical by the above argument .
if @xmath601 and @xmath991 then @xmath992 , a contradiction , so assume that @xmath601 and @xmath990
. then @xmath109 as above except when @xmath993 and @xmath103 is given by figure 5.9(ii ) . in this case
add @xmath994 to @xmath639 as shown noting that @xmath640 does not receive any curvature from @xmath90 .
therefore if a region receives positive curvature across an edge it is across a @xmath995 or @xmath783 or @xmath996 edge ; and if across consecutive edges then the sublabel is @xmath997 .
it follows that if @xmath406 does not involve @xmath998 and @xmath649 then @xmath109 .
observe that each occurrence of @xmath999 contributes two gaps .
it follows that if @xmath406 contains at least two occurrences of @xmath998 or if @xmath406 contains exactly one occurrence and @xmath646 then there are four gaps and @xmath109 .
but if @xmath1000 and @xmath655 then checking possible @xmath406 shows that again there are four gaps and @xmath109 . if @xmath1001 then either there is a contradiction to one of our assumptions or @xmath692 except when @xmath1002 . if @xmath1003 or @xmath1004 then @xmath1005 and there is a sphere by lemma 3.2(vii ) .
this leaves @xmath1006 and these are given by figure 5.9(iii ) , ( iv ) .
add @xmath1007 to @xmath1008 as shown .
repeat the argument for @xmath641 noting that as before @xmath641 does not receive positive curvature from the corresponding region @xmath90 .
this procedure will terminate at a region @xmath700 , say , such that @xmath1009 for @xmath1010 where @xmath1011 and @xmath1012 .
repeat this for each copy of the region @xmath640 to conclude that @xmath3 is aspherical . if @xmath991 and @xmath989 then the argument is the same by symmetry .
* ( b8 ) * @xmath721 , @xmath364 , @xmath984 , @xmath358 , @xmath359 .
+ if @xmath336 then @xmath90 is given by figures 5.1(ii ) and 5.9(i ) . in figure 5.9(i ) @xmath985 and @xmath986 .
this implies that @xmath408,@xmath467 , otherwise there is a contradiction to @xmath591 so add @xmath771 to @xmath393 and @xmath396 . in figure 5.1(ii ) @xmath1015 .
this similarly implies that @xmath408 and @xmath467 , so add @xmath771 to @xmath393 and @xmath396 . observe that in figure 5.9(i ) @xmath386 does not receive positive curvature from @xmath595 ; and @xmath387 does not receive positive curvature from @xmath595 . observe also that if @xmath103 receives positive curvature then it does so across the edges @xmath1016 , @xmath783 or @xmath385 .
thus there is always a gap immediately preceding @xmath891 and @xmath794 ; and there is a gap after @xmath749 and @xmath796 .
this implies that if there are at least four occurrences of @xmath793 then @xmath406 contains at least four gaps and so @xmath429 by lemma 3.6 .
we will proceed according to the number of occurrences of @xmath793 in @xmath406 .
if there are no occurrences of @xmath793 then either @xmath1017 where @xmath1018 and there are four gaps , or @xmath1019 where @xmath1020 @xmath1021 .
but since there is always at least one gap between any two occurrences of @xmath892 it follows that again there are four gaps or @xmath1022 , a contradiction .
so @xmath333 in this case .
assume first that @xmath5 is non - cyclic .
if there is exactly one occurrence of @xmath793 in @xmath406 then @xmath5 is cyclic so suppose that there are either two or three occurrences of @xmath793
. then either the label contains at least four gaps or it contradicts one of the ( * b8 * ) assumptions or one of the following cases @xmath1023 ( @xmath1024 ) occurs : consider ( 1)(4 ) .
these yield the relator @xmath1034 and it follows that @xmath1035 for @xmath1036 .
moreover if @xmath120 then there is a sphere by lemma 3.1(c)(ii ) so it can be assumed that @xmath293 . in case ( 1 ) @xmath640 is given by figure 5.10(i ) where , given that @xmath1037 , it can be assumed that @xmath1038 and at least one of @xmath1039 , @xmath1040 equals 2 .
add @xmath308 from @xmath639 to @xmath1041 as shown in figure 5.10(i ) ; and if @xmath1042 add a further @xmath308 of @xmath639 to @xmath1043 as shown . in cases ( 2)(4 )
@xmath1044 where @xmath1045 and @xmath308 is added from @xmath105 to @xmath1041 as shown in figure 5.10(ii ) .
observe that @xmath1046 in figure 5.10(i ) , ( ii ) for otherwise @xmath204 would be a proper sublabel , and so @xmath1047 .
if @xmath1048 then the sublabel @xmath411 yields a gap so let @xmath1049 .
then either @xmath792 yields a gap or @xmath1050 and @xmath308 is added to @xmath1041 from @xmath1051 . continuing this way , since @xmath293
, eventually we get a sublabel @xmath411 or @xmath792 which contributes a gap .
consider @xmath1052 .
if it contains an odd number of occurrences of @xmath749 then @xmath1053 implies that @xmath1054 and @xmath5 is cyclic .
this leaves the case when there are exactly two occurrences of @xmath749 and @xmath1055 for @xmath1056 .
if @xmath1057 then there are four gaps and @xmath1058 ; and if @xmath1059 , @xmath1060 this implies @xmath120 , a contradiction . consider ( 5)(8 ) .
these yield the relator @xmath1061 and @xmath5 is abelian . observe that @xmath1062 yields ( * e * ) , so assume otherwise . in each case add @xmath1063 to @xmath1064 as shown in figure 5.10(iii)(vi ) .
observe that @xmath1065 receives no curvature from @xmath499 or @xmath481 ; that @xmath1066 implies @xmath1067 otherwise there is a contradiction to @xmath591 ; and there is still a gap between each pair of occurrences of @xmath92 .
if @xmath1068 contains an odd number of occurrences of @xmath749 then @xmath5 is cyclic so it can be assumed that @xmath1068 yields the relator @xmath1069 . if @xmath1070 and @xmath1071 then there are four gaps and if @xmath1072 then @xmath1073 , so this leaves the case @xmath1074 , @xmath1075 .
again there are four gaps except when @xmath1076 and this is shown in figure 5.10(vii ) : add @xmath1077 to @xmath1078 and observe that @xmath1079 does not receive positive curvature from @xmath90 .
consider @xmath1080 .
if there are at least four occurrences of @xmath749 then @xmath1081 ; and if there is an odd number of occurrences then @xmath5 is cyclic .
suppose firstly that there are no occurrences of @xmath749 in @xmath1082 .
since @xmath1083 , if there is one occurrence of @xmath748 then @xmath1084 ( @xmath305 ) and there are four gaps ; and since each @xmath677 yields a gap and each @xmath1085 ( @xmath1086 ) yields a gap it follows that if there are at least two occurrences of @xmath748 then again @xmath1087 .
suppose finally that there are two occurrences of @xmath749 and so @xmath1088 where @xmath1089 and @xmath1090 . if @xmath1071 then there are four gaps ; and if @xmath1091 then @xmath1092 , otherwise there is a contradiction to @xmath1093 , and again there are four gaps , so @xmath1087
. finally consider case ( 9 ) . in this case
@xmath1094 is given by figure 5.10(viii ) .
suppose that @xmath1095 .
then it can be assumed that @xmath1096 for @xmath1097 and @xmath779 so add @xmath780 to @xmath1098 for @xmath1099 . in this case
if @xmath776 then we obtain a sphere by lemma 3.1(c)(iii ) . now
if @xmath1100 then as shown in figure 5.10(viii ) @xmath1101 and @xmath205 will contribute two - thirds of a gap .
if there are now at least two occurrences of @xmath749 then either @xmath1102 or @xmath5 is cyclic , a contradiction , or there are four gaps ; if there is exactly one occurrence of @xmath749 then this contradicts @xmath5 non - cyclic ; and if there are no occurrences of @xmath749 then @xmath1103 where @xmath1104,@xmath1105 .
since @xmath1106 it follows that there are at least four gaps and @xmath1107 . now let @xmath5 be cyclic .
if @xmath601 or @xmath743 then there is a sphere by @xmath281-equivalence and lemma 3.2(vii ) , ( viii ) ; and @xmath1108 is ( * e4 * ) , so assume otherwise .
in particular , @xmath1093 .
we follow the same argument as above and so if @xmath406 contains no occurrences or at least four occurrences of @xmath749 then , as before , @xmath310 ; and if @xmath406 contains an odd number of occurrences of @xmath749 then @xmath1109 for some @xmath1110 which implies there are at least four gaps and @xmath109 .
suppose then that @xmath406 involves @xmath749 exactly twice .
subcases ( 1)(4 ) imply @xmath811 and ( 9 ) implies @xmath743 , a contradiction .
this leaves subcases ( 5)(8 ) .
add @xmath994 to @xmath1064 as in figure 5.10(iii)-(vi ) .
since there is still a gap between each pair of occurrences of @xmath92 it follows from the above paragraph and the previous argument that @xmath1111 except when @xmath1112 . again
add @xmath1113 to @xmath1078 as shown in figure 5.10(vii ) .
if @xmath1080 involves at least three occurrences of @xmath749 then , since @xmath1114 does not receive positive curvature from @xmath90 in figure 5.10(vii ) , there are at least four gaps and @xmath1081 . otherwise checking the possible labels for @xmath1115 shows that there are four gaps or a contradiction to @xmath1093 or @xmath1116 . *
( b9 ) * @xmath363 , @xmath364 , @xmath984 , @xmath358 , @xmath359.if @xmath336 then @xmath90 is given by figure 5.9(i ) .
if @xmath338 then @xmath90 is given by figure 4.1(ii ) . in figure 5.9(i ) @xmath985 and @xmath986 which implies that @xmath408 and @xmath467 otherwise there is a contradiction to @xmath126 or @xmath130 , so add @xmath771 to @xmath393 and @xmath396 . in figure 4.1(ii ) @xmath1120 which implies that @xmath382 where @xmath368 otherwise there is a contradiction to @xmath126 so add @xmath433 to @xmath384 . observe that if @xmath103 receives positive curvature then it receives @xmath308 across the edges @xmath1016 , @xmath783 or receives @xmath303 across the edge @xmath385 .
thus there is always a gap immediately preceding @xmath891 and @xmath794 ; and there is a gap after @xmath749 and @xmath796 .
also there is a two thirds of a gap across the edge @xmath385 .
this implies that if there are at least four occurrences of @xmath793 then @xmath406 contains at least four gaps and so @xmath429 by lemma 3.6 .
we will proceed according to the number of occurrences of @xmath793 in @xmath406 .
if there are no occurrences of @xmath793 then either @xmath1017 where @xmath1121 and there are four gaps , or @xmath1019 where @xmath1020 @xmath1021 .
but since there is always at least one gap between any two occurrences of @xmath892 it follows that again there are four gaps or @xmath1022 , a contradiction .
so @xmath333 in this case .
on the other hand if there are between one and three occurrences of @xmath1122 inclusive and @xmath669 then either there is a contradiction to one of the ( * b9 * ) first suppose that @xmath1124 ( in particular , @xmath5 is non - cyclic ) as in figure 5.11(i ) . if @xmath1126 then it can be assumed that @xmath1096 for @xmath50 and @xmath1127 , so add @xmath1128 to @xmath1129 for @xmath1130
. therefore @xmath1131 contributes a two - thirds gap .
if @xmath776 then we obtain a sphere by lemma 3.1(c)(iv ) .
assume now that @xmath1100 and without any loss of generality assume that @xmath1132 of figure 5.11(i ) .
if @xmath1133 has at least two occurrences of @xmath749 then either there is a contradiction to @xmath5 non - cyclic or one of the ( * b9 * ) assumptions or there are four gaps and @xmath1134 ; if there is exactly one occurrence of @xmath749 then this contradicts @xmath5 non - cyclic ; and if there are no occurrences of @xmath749 then @xmath1135 where @xmath1104 , @xmath1105 .
since @xmath1106 it follows that there are at least four gaps and @xmath1107 . now let @xmath743 .
it follows from the above argument and checking possible vertex labels that if @xmath640 receives positive curvature then @xmath692 except when @xmath1136 and @xmath670 .
add @xmath1137 to @xmath1008 as shown in figure 5.11(ii ) ( where we note that @xmath641 receives no curvature from @xmath90 ) .
checking @xmath1138 with the understanding that there is a gap preceding the @xmath205 and each @xmath205 contributes a two - thirds gap it follows that either there is a contradiction or @xmath1139 except when @xmath1140 . if @xmath641 is given by figure 5.11(iii ) then @xmath1141 , so let @xmath641 be given by figure 5.11(iv ) .
observe that the neighbour @xmath499 has label @xmath1142 and no longer has degree 3 and that if either @xmath1143 or @xmath1144 then @xmath1145 and so figure 5.11(iv ) shows the only possibility for @xmath1146 .
now repeat the argument for @xmath641 and @xmath642 .
the conditions on the neighbouring regions @xmath499 , @xmath481 and @xmath500 ensure that this procedure will terminate at a region @xmath700 , say , such that @xmath1009 for @xmath1010 where @xmath1011 and @xmath1147 . repeat this for each copy of the region @xmath640 to conclude @xmath3 is aspherical . *
( b10 ) * @xmath721 , @xmath1149.if @xmath1150 then we obtain a sphere by lemma 3.1(c)(i ) , so assume otherwise ( in particular , @xmath5 is non - cyclic ) . observe that if a label involves @xmath749 and is not
@xmath79 then it yields the relation @xmath1151 and so it follows that if @xmath102 then @xmath90 is given by figure 5.1(ii ) in which @xmath1152 .
add @xmath771 to @xmath501 and @xmath506 as shown .
this completes the distribution of curvature .
consider @xmath499 of figure 5.1(ii ) .
since positive curvature only crosses the edge @xmath615 it follows that if @xmath1153 involves at least four occurrences of @xmath92 then @xmath1154 .
this leaves the case when @xmath1153 involves exactly two occurrences of @xmath92 .
but @xmath182 can not be a sublabel since there would then be a region with label @xmath869 , and so @xmath1155 has four gaps and @xmath1154 . in conclusion
@xmath3 is aspherical in this case if and only if @xmath1156 . *
( b11 ) * @xmath721 , @xmath1157.if @xmath5 is cyclic then there is a sphere by lemma 3.2(vi ) , so assume otherwise .
if @xmath336 then @xmath90 is given by figures 5.1(ii ) and 5.9(i ) ; and if @xmath338 then @xmath90 is given by figure 4.3(i ) . in figure 5.9(i ) @xmath985 and @xmath986 .
this implies that @xmath408 and @xmath467 otherwise there is a contradiction to @xmath131 so add @xmath771 to @xmath393 and @xmath396 . in figure 5.1(ii ) @xmath1015 .
this implies that @xmath408 and @xmath467 otherwise there is a contradiction to @xmath131 so add @xmath771 to @xmath393 and @xmath396 . in figure 4.3(i ) @xmath1158 .
this implies that @xmath408 , @xmath409 and @xmath398 otherwise there is a contradiction to @xmath131 so add @xmath433 to @xmath393 , @xmath394 and @xmath399 .
observe that if @xmath103 receives positive curvature then it does so across the edges @xmath1016 , @xmath783 , @xmath385 , @xmath397 and there is a two - thirds gap in @xmath397 .
if there is only one occurrence of @xmath749 or @xmath92 then this contradicts @xmath5 non - cyclic . also if @xmath1159 then we obtain a sphere by lemma 3.1(c)(ii),(v ) .
so assume that @xmath1160 .
if there are between two and five occurrences of @xmath749 inclusive then the resulting relator contradicts @xmath5 non - cyclic or @xmath1161 except for @xmath1162 and @xmath1163 . but
each of these yields four gaps . since at least six occurrences of @xmath749 yields four gaps
, this leaves the case when @xmath749 does not occur and this implies that there is at least one gap between any two occurrences of @xmath92 .
it follows that there can only be exactly three occurrences of @xmath92 .
checking shows there are four gaps except when the following subcases occur : consider ( 1 ) , this case is given by figure 5.12(i ) . if @xmath1167 then @xmath333 .
assume otherwise and add @xmath779 to @xmath803 as shown .
observe that @xmath1168 and @xmath1169 .
if @xmath1170 then @xmath526 can not receive positive curvature across the edges @xmath206 or @xmath411 .
suppose that @xmath1171 .
then @xmath526 can not receive positive curvature across the edge @xmath783 otherwise @xmath1172 , a contradiction .
so there will be a gap between @xmath749 and the next occurrence of @xmath92 and the previous argument does not change .
it follows that @xmath1173 .
consider ( 2 ) , this case is given by figure 5.12(ii ) .
if @xmath1167 or @xmath1174 then @xmath333 . assume otherwise and add @xmath779 to @xmath904 as shown .
observe that @xmath1175 and the same argument as for ( 1 ) can be applied to deduce that @xmath1176 .
consider ( 3 ) , this case is given by figure 5.12(iii ) .
if @xmath1167 or @xmath1177 or @xmath1178 then @xmath333 . assume otherwise and add @xmath780 to @xmath781 for @xmath1179 as shown .
consider @xmath1180 where @xmath1181 . if @xmath1182 then @xmath1183 , a contradiction .
if @xmath1184 then exactly as above @xmath1185 either contains four gaps or there is a contradiction to either @xmath5 non - cyclic , @xmath1186 or one of the ( * b11 * ) assumptions and so @xmath1187 . finally let @xmath1171 . if @xmath1188 then again as above @xmath1187
so assume otherwise .
but if @xmath1189 then we are back in subcase ( 1 ) and if @xmath1190 then we are back in subcase ( 2 ) , the only difference being that @xmath103 in figures 5.12(i ) , ( ii ) this time receives @xmath1191 from @xmath481 rather than @xmath1192 . * ( b12 ) * @xmath1194 , @xmath1157.if @xmath5 is cyclic then @xmath987 and so there is a sphere by lemma 3.1(a)(iii ) or lemma 3.2(iii ) , so assume otherwise . if @xmath336 then @xmath90 is given by figure 5.9(i ) ; and if @xmath338 then @xmath90 is given by figures 4.1(ii ) and 4.3(i ) . in figure 5.9(i ) @xmath985 and @xmath986 .
this implies that @xmath408 @xmath467 , otherwise there is a contradiction to @xmath5 non - cyclic or @xmath130 or @xmath131 so add @xmath771 to @xmath393 and @xmath396 . in figure 4.1(ii ) @xmath1120 .
this implies that @xmath408 , @xmath409 and @xmath398 otherwise there is a contradiction to @xmath5 non - cyclic or @xmath130 or @xmath131 so add @xmath433 to @xmath393 , @xmath394 and @xmath399 . in figure 4.3(i ) @xmath1158 .
this implies that @xmath408 , @xmath409 and @xmath398 otherwise there is a contradiction to @xmath5 non - cyclic or @xmath130 or @xmath131 so add @xmath433 to @xmath393 , @xmath394 and @xmath399 .
observe that if @xmath103 receives positive curvature then it does so across the edges @xmath1016 , @xmath783 , @xmath385 , @xmath397 and there is a two - thirds gap in @xmath385 and @xmath397 .
if there is only one occurrence of @xmath749 or @xmath92 in @xmath406 then this contradicts @xmath5 non - cyclic so assume otherwise . also if @xmath1195 then we obtain a sphere as by lemma 3.1(c)(iv ) .
so assume that @xmath1196 .
checking now shows that there is at least one gap between any two occurrences of @xmath749 or between any two occurrences of @xmath92 ; and there is at least a two - thirds gap between any two occurrences of @xmath749 and @xmath92 .
it follows that if there are at least four occurrences of @xmath749 or of @xmath92 then @xmath310 .
suppose there are three occurrences of @xmath749 .
if there are three occurrences of @xmath92 then there are at least @xmath1197 gaps ; and if there are two occurrences of @xmath92 then there are four gaps or @xmath406 yields the relator @xmath1198 which forces @xmath5 to be cyclic .
suppose there are two occurrences of @xmath749 .
if there are three occurrences of @xmath92 then there are four gaps or @xmath406 yields the relator @xmath1199 which forces @xmath5 to be cyclic ; and if there are two occurrences of @xmath92 then @xmath406 forces @xmath1148 or @xmath5 cyclic or is one of @xmath1200 , @xmath1201 and there are four gaps .
this leaves the case when @xmath749 does not occur . checking @xmath406 shows that there is a contradiction to @xmath131 or there are at least four gaps except when @xmath1164 as shown in figure 5.12(iv ) . in this case
add @xmath1202 to @xmath803 as shown .
since @xmath303 is distributed across @xmath397 it follows that @xmath1173 as above .
it follows from ( * b1*)(*b12 * ) that either @xmath3 is aspherical or modulo @xmath281-equivalence one of the conditions of theorem 1.1 ( i)(iii ) or theorem 1.2 ( i ) , ( ii ) , ( iv)(x ) is satisfied and so theorems 1.1 and 1.2 are proved for case b. 99 y g baik , w a bogley and s j pride , on the asphericity of length four relative group presentations , int .
j. algebra comput .
* 7 * ( 1997 ) , 277312 .
w a bogley and s j pride , aspherical relative presentations , proc .
edinburgh math .
* 35 * ( 1992 ) , 139 .
d e cohen , combinatorial group theory : a topological approach , lms student texts * 14 * ( cup ) ( 1989 ) .
d j collins and j huebschmann , spherical diagrams and identities among relations , math . ann .
* 261 * ( 1982 ) , 155183 .
p j davidson , on the asphericity of a family of relative group presentations , int . j. algebra comput . * 19 * ( 2009 ) , 159185 .
m edjvet , equations over groups and a theorem of higman , neumann and neumann , proc .
london math .
* 62 * ( 1991 ) , 563589 .
m edjvet , on the asphericity of one - relator relative presentations , proc .
edinburgh sec .
* 124 * ( 1994 ) , 713728 .
m edjvet and j howie , the solution of length four equations over groups , trans .
* 326 * ( 1991 ) , 345369 .
the gap group : gap groups , algorithms and programming ; version 4.6.4 , 2013 .
n d gilbert and j howie , log groups and cyclically presented groups , j. algebra * 174 * ( 1995 ) , 118131 .
j howie and v metaftsis , on the asphericity of length five relative group presentations , proc .
london math .
* 82 * ( 2001 ) , 173194 .
v metaftsis , on the asphericity of relative group presentations of arbitrary length , int .
j. algebra comput .
* 13 * ( 2003 ) , 323339 .
|
we consider the relative group presentation @xmath0 where @xmath1 and @xmath2 .
we show modulo a small number of exceptional cases exactly when @xmath3 is aspherical .
if @xmath4 then the exceptional cases occur when @xmath5 is isomorphic to one of @xmath6 or @xmath7.mathematics subject classification : 20f05 , 57m05.keywords : relative group presentations ; asphericity .
* asphericity of a length four relative group presentation * abd ghafur bin ahmad , muna a al - mulla and martin edjvet
|
ladders ( or ladders - like ) systems are of great interest in condensed matter physics . on one hand ,
spin ladders materials with intriguing magnetic properties have been found or synthesized @xcite . on the other hand , due to the advances of cold atom technology , optical ladders
might be provided as one of the simplest accessible objects that exhibit fascinating properties of quantum matters @xcite .
apparently , real compounds are not so flexible as the cold atoms in optical lattices in exploring the underlying mechanism of theoretical models .
progresses in these two fields could interplay and lead to a promising boom of new findings . in this work
, we investigate a two - leg spin ladders in the strong anisotropic limit .
we construct the ground state phase diagram and investigate the properties of rich phases , including the full polarized ( fp ) , singlet dimer ( sd ) , luttinger liquid ( ll ) , triplon solid ( ts ) , and non - trivial antiferromagnet ( non - trivial af ) ones .
we set up the low - energy effective models for describing different phases within the frame of bond operator theory .
numerical methods including exact diagonalization ( ed ) and quantum monte carlo ( qmc ) are employed to verify our new findings .
we also discuss how our interesting results could be related to experiment .
the model we are concerned with is the dimer @xmath0 spin-1/2 model on two - leg ladders @xmath1
\nonumber \\ & & + \sum_{j}k\mathbf{s}_{1,j}\cdot \mathbf{s}_{2,j}. \label{h}\end{aligned}\ ] ] where @xmath2 denotes the @xmath3 spin operator of @xmath4 ( @xmath5 ) component at site @xmath6 ( @xmath7 ) of @xmath8-th ( @xmath9 ) leg . @xmath10 are the interactions strength along rail in transverse and longitude directions respectively , @xmath11 is the heisenberg interaction strength along rung and will be set as the energy unit , @xmath12 denotes the zeeman field . for convenience ,
we consider the case of @xmath13 and @xmath14 , and a bipartite lattice without frustration , i.e. @xmath15 is even when periodic boundary condition is imposed .
we will focus on the strong anisotropic cases , e.g. @xmath16 and @xmath17 , which are not readily to be found in real compounds .
recent progress in the technology of cold atoms are encouraging @xcite .
the above model can be translated to an identical hardcore bosonic @xmath18-@xmath19 model by a matsubara - matsuda mapping @xcite . although some experiment on hardcore bosons have been realized @xcite , but the interaction terms in this model need more sophisticated methods to realize @xcite .
one may consider a ladder - shaped optical lattice formed by standing wave lasers @xcite .
the superexchange mechanism could be mimicked by controlled tunneling of cold bosonic or fermionic atoms with two internal states , such as @xmath20rb in @xmath21 or @xmath22 , where @xmath23 is a vacuum . so that we can define a spin on a site in a bilinear form @xmath24 , where @xmath25 , is a pauli matrix and @xmath26 . in the mott insulating phase
, the above model could be realized with almost arbitrary model parameters @xcite . before solving the model ,
we introduce a bond operator representation that will be very useful for elaborating the phenomena we found in this system .
there are four states on each rung , one singlet and three triplets , @xmath27 , @xmath28 , @xmath29 , @xmath30 , which create the singlet state and triplet states at @xmath6-th rung with the constraint , @xmath31 .
then the original hamiltonian eq.([h ] ) can be rewritten in terms of these bond operators as @xmath32 + \nonumber \\ & & \frac{j^{z}}{2}\sum_{j}[(s_{j}^{\dag } s_{j+1}^{\dag } t_{0,j}t_{0,j+1}+s_{j}^{\dag } s_{j+1}t_{0,j}t_{0,j+1}^{\dag } + h.c)+ \nonumber \\ & & ( n_{+,j}-n_{-,j})(n_{+,j+1}-n_{-,j+1})]-h\sum_{j}(n_{+,j}-n_{-,j } ) , \label{hb}\end{aligned}\ ] ] where , @xmath33 are the particle number operators of @xmath34 .
we have employed both stochastic series expansion ( sse ) quantum monte carlo ( qmc ) @xcite and exact diagonalization methods to investigate numerically the system s low energy properties .
analytical methods have also been applied , including the bond - operator mean - field ( bomf ) theory and bosonization method for a comparison and comprehension of the underline physics . in this work
, the sse qmc simulation is performed on a ladder of length up to @xmath35 .
the temperature is taken as @xmath36 and periodic boundary condition is imposed .
thus the lowest temperature @xmath37 is reached , which is sufficient to obtain the ground state observables .
we present two slices of phase diagrams to exemplify our results : ( i ) @xmath38 ( figure 1 ) ; ( ii ) @xmath39 ( figure 2 ) .
the properties of the ground states and the low energy excitations are elaborated as follows .
first , let us see two simplest limits .
the first is the full polarized ( fp ) state in large enough field @xmath12 .
the second occurs at @xmath40 , where the hamiltonian ( [ h ] ) reduces to an array of disjoint dimers . in the second limit
, the ground state and excitations are determined by individual dimers .
the energies of singlet and triplets are @xmath41 , @xmath42 , @xmath43 , @xmath44 . when @xmath45 , the ground state is an array of singlets ( we call it singlet dimer ( sd ) phase ) with an excitation gap @xmath46 . with the applied field @xmath12 increasing , the system s ground state experiences a transition from sd to fp state at @xmath47 .
if we fix @xmath38 and raise @xmath48 ( @xmath49 ) , a triplon solid ( ts ) phase intervenes between sd and fp phases ( figure 1 ) .
the ts state is a paving of singlets and triplets on the ladders alternatively . to detect the ts order in a general case
, one can define a triplon creation operator @xmath50 acting in spin s hilbert space @xmath51 we have @xmath52 , @xmath53 , and @xmath54 .
it is worth noting that @xmath55 is not a perfect hardcore boson @xcite . but with appropriate model parameters , the hardcore condition could be well fulfilled .
the diagonal real space correlation function of triplon is defined as @xmath56 where the thermodynamic limit @xmath57 is taken . the static structure factor for detecting ts order
is given by @xmath58 from the point of view of renormalization , the @xmath59 and @xmath60 , so that the hamiltonian ( [ hb ] ) at low energy sector can be replaced by @xmath61 where we have defined the pseudo - spin operators in a schwinger boson representation , @xmath62 .
the effective hamiltonian ( [ ts ] ) is nothing but a `` classical ising model in a field '' .
the ts phase is a consequence of the first antiferromagnetic interaction term .
if the second term prevails the first one , sd and fp phases could be reached if @xmath63 and @xmath64 respectively .
we also get the phase boundaries by comparing the ground state energies of the three states : ( i ) @xmath65 for the boundary between sd and ts and ( ii ) @xmath66 between ts and fp .
it is worth to point out that the effective hamiltonian ( [ ts ] ) also gives a good description for ts and fp phase in the phase diagram where @xmath67 .
the consistency between the analytical and numerical results are quite well , as shown in figure 3 .
if @xmath68 , the area of ts phase shrinks to the top right corner of the phase diagram ( figure 2 ) and there emerges a vast gapless phase called luttinger liquid ( ll ) . in isotropic case ,
i.e. , @xmath69 , there have been a large amount of literatures devoted to the study of the ground state and low excitations @xcite .
here we focus on the anisotropic case , @xmath70 , and answer the question whether there is still a region described by ll , which , to our knowledge , is not been discussed before .
we give a positive answer here .
interestingly , the low energy properties in the ll phase are also governed by the states @xmath71 and @xmath72 . but the effective hamiltonian includes a part of `` quantum fluctuations '' besides the `` classical '' part , @xmath73 , @xmath74 .
\label{ll}\ ] ] this model is exactly the form of the @xmath0 chain whose properties are well known owing to many reliable methods , such as bethe ansatz and bosonization . we know that its ground state is antiferromagnet or ll when @xmath75 or @xmath76 respectively @xcite . in ll ,
both diagonal and off - diagonal triplon correlations are important .
the former is defined in ( [ spi ] ) .
the latter can be defined as @xmath77 from the effective model ( [ ll ] ) , we can get the equal - time spin correlation by the standard abelian bosonization techniques @xcite , @xmath78 @xmath79 where @xmath80 denotes the lattice distance , @xmath81 are constants , @xmath82 is the uniform magnetization , @xmath83 is the so - called ll parameter with typical value less than @xmath84 that governs the spin correlations at long wavelength limit . for a comparison , we perform the qmc simulation on the original hamiltonian ( [ h ] ) .
we found the spin ladder is sensitive to the value of @xmath85 in ts region .
however , comparing with spin @xmath0 chain , the ts and ll still exist in the phase diagram of the spin ladder only with quantitative changes of phase boundaries .
the densities of four rung states @xmath86 shown in figure 4(b ) confirm the goodness of the effective model ( [ ll ] ) . from the nodes of structure factor @xmath87 in figure 4(a ) , we can see a clear phase transition from ts to ll .
the long range diagonal order in ts also could be seen in the correlation function @xmath88 of figure 4(d ) . in figure 4(c ) we show the power law decay behavior of correlation @xmath89 in ll region , which coincides with the relation ( [ cpm ] ) perfectly .
now we turn to the last phase .
first , we focus on the thermodynamic limit , and the case of @xmath38 ( see figure 1 ) , since the qualitative properties do not change if @xmath85 is not large enough .
we show the ground state phase is a gapped mott insulator of triplon that can be understood quite good by an effective ising model in a transverse field ( t - ising model )
. then we will work on finite system to figure out the low excitations of this phase .
we will show the lowest two states for finite systems are nearly degenerate and tend to be maximally entangled with @xmath90 increasing .
when @xmath12 is small and @xmath91 is large enough , the low energy sector of the system splits into two channels .
the first one ( channel a ) is a classical ising hamiltonian @xmath92 with @xmath93 and constraint @xmath94 .
when @xmath45 , the gound state is an alternative paving of @xmath72 and @xmath95 states along the ladders .
its lowest energy per rung reads @xmath96 the second one ( channel b ) is obtained by introducing a set of pseudo - spin operators in another schwinger boson representation , @xmath97 , @xmath98 , @xmath99 , with constraint @xmath100 , which reads @xmath101 .
\label{tising}\ ] ] this is exactly the t - ising model , whose properties are also well - known .
one can solve it exactly by applying the jordan - wigner transformation @xcite , and get the lowest energy per rung , @xmath102 where @xmath103 is the complete elliptic integral of the second kind . because the field @xmath12 is coupled with @xmath104 and @xmath105
, it does not appear in channel b. in fact , this is the reason why the phase boundary between sd and af phases is a straight line ( see figure 1 and 2 ) that almost does not change with the field increasing for a given @xmath85 until runs into another phase . in the following , combining numerical approaches , we emphasize several important consequences of the above effective hamiltonians .
it is easy to see that the difference of the lowest energies between channels a and b is @xmath106 so we see that channel b truly reflects the ground state of the original ladders system . in the whole parameter region
, channel b gives quite good description of the ground state energy of original system .
the goodness of effective @xmath107 for the ground state has been confirmed by qmc simulations on the original ladders system shown in figure 3 , 4 , and 5 . in thermodynamic limit and when @xmath108 , the effective hamiltonian @xmath107 exhibits an af phase with double degeneracy , @xmath109 this result needs an assumption of spontaneous symmetry breaking nevertheless .
we will address this issue later .
the staggered magnetization of this phase can be detected by the order parameter @xmath110 there is a quantum phase transition between the sd and af phases at the gapless point @xmath111 , which can be worked out from the divergence point of the second - order derivative of @xmath112 on @xmath91 , as well as the first - order derivative of the weight of singlets @xmath113 where @xmath114 is the lowest excitation spectrum of @xmath107 .
the staggered magnetization @xmath115 is deduced from the correlation function @xmath116 formally , which is too tedious to be presented here .
the smooth curve in figure 5(a ) shows @xmath115 that is obtained by the correlation function with @xmath117 .
comparing it with qmc results , we confirm the validity of the effective hamiltonian @xmath118 .
the phase transition between those two states is of second order , which can be seen from the staggered magnetic susceptibility @xmath119 shown in figure 5(b ) .
is the effective hamiltonian
@xmath107 good enough for low excitations ?
how do the translational and trs breaking happen in the af phase for the original ladders system ? to answer these questions , we performed an exact diagonalization on the original system up to @xmath120 ( i.e. @xmath121 sites ) to reveal the aspect of the original system that may be ignored .
we found , although the effective @xmath107 reflects the ground state of the original system perfectly , the low excitations above the ground state is largely restricted by channel a. the excitation spectrum ( [ wb ] ) of channel b gives an energy gap above its ground state , @xmath122 but the true gap of the original system should not exceed @xmath123 ( see ( [ delta ] ) ) , so in large @xmath124 we have @xmath125 ( labelled by colored markers ) obtained by exact diagonalization and analytical results for a ladder with @xmath57 ( shown by lines ) obtained by effective theories .
the dashed line represents the true gap at large @xmath124 limit . here
@xmath126 and @xmath127 are taken .
please see more details in the text.,width=421 ] , of the lowest two states , @xmath128 and @xmath129 .
( b ) the percentages , @xmath130 and @xmath131 , of dominant parts in the lowest two states .
please see more in the text.,width=421 ] the numerical result for @xmath126 is exemplified in figure 6 .
we see the excitations of the ladder system is much different from the effective @xmath107 . in large @xmath91
, we observe no extra energy levels lying between @xmath132 and @xmath112 , which ensures a true gap greater than @xmath133 .
another observation is the lowest two levels of finite systems are not exactly degenerate and not even the ones with af order , @xmath134 and @xmath135 .
instead , they are @xmath136 with dominant parts prevail the residuals .
the numerical results are shown in figure 7 .
we observed that @xmath137 has slightly lower energy than @xmath138 , but their energy difference @xmath139 rapidly reduces to zero with @xmath124 increasing . and ,
at the same time , the two states approach the so - called ghz states @xcite asymptotically , latexmath:[\[\begin{aligned } \left\vert \phi _ { 1}\right\rangle & \approx & \frac{1}{\sqrt{2}}\left ( \left\vert \phi _ { 2}\right\rangle & \approx & \frac{1}{\sqrt{2}}\left ( the result implies the entangled ghz states could be purified if we increase @xmath90 adiabatically for finite system . for long enough system ,
spontaneous symmetry breaking may occur and the entanglement will vanish .
nonetheless , the two ghz states do not exhibit a non - zero order parameter @xmath115 , while they do have the same antiferromagnetic spin correlation functions as @xmath134 and @xmath141 .
that is the reason why we name it a non - trivial af phase .
one can check that @xmath137 and @xmath142 are eigenstates of the time reversal operator @xmath143 where @xmath144 is the complex conjugation , if noticing that @xmath145 thus @xmath137 and @xmath146 preserve the exact trs as the same as the original hamiltonian ( [ h ] ) . from the point of view of the effective @xmath107 , the assumption of spontaneous symmetry breaking in thermodynamic limit is based an unpredictable disturbance from environment @xcite .
but here we have one favorable factor to avoid this - the destructive field should couple to local variables , @xmath147 , delicately since the system with true gap is antiferromangetic .
thus we think there is the chance to realize such entangled states that do not break the trs .
the dimer @xmath0 spin system discussed in this paper might be realized in a mott insulating phase of cold bosonic atoms in near future @xcite . we can estimate the typical energy scales . for @xmath20rb atoms with a lattice constant @xmath148 nm and about @xmath149 atoms in a bose - einstein condensate , we can chose @xmath150 khz ( corresponding to a time scale of 10 ms ) with a conservative choice of @xmath151 @xmath152 khz and @xmath153 , where @xmath18 is the hopping strength of cold atoms between two nearest neighbor minimums of laser potential and @xmath154 is on site interaction originating from the @xmath155 wave scattering .
these energy scales are clearly compatible with current experiments @xcite and make the system in a mott insulating area . in experiment , the density of condensates in momentum space @xmath156 can be measured by noise correlations which can be linked to spin - spin correlations @xcite .
we can use bragg scattering of light , which gives rise to the spin structure factor , to detect @xmath87 @xcite .
an alternative technique for imaging spin states in optical lattices has been put forward @xcite .
thus , all the phases discussed in this paper are detectable in experiment .
in summary , combining analytical and numerical methods , we have investigated the ground state phase diagrams and low excitations of the dimer @xmath0 spin ladder system .
we show that most features of the phases could be understood within the frame of bond operator theory and have proven it by using quantum monte carlo method .
we present the rich ground state phase diagram which can be detected in optical lattice by future experiment .
this work was supported by the nsfc under grants no . 11074177 , srf for rocs sem ( 20111139 - 10 - 2 )
10 dagotto e and rice t m 1996 _ science _ * 271 * 5249
|
we study the dimer @xmath0 spin model on two - leg ladders with isotropic heisenberg interactions on the rung and anisotropic @xmath0 interactions along the rail in an external field . combining both analytical and numerical methods
, we set up the ground state phase diagram and investigate the quantum phase transitions and the properties of rich phases , including the full polarized , singlet dimer , luttinger liquid , triplon solid , and a non - trivial antiferromagnetic phases with gap .
the analytical analyses based on solvable effective hamiltonians are presented for clear view of the phases and transitions .
quantum monte carlo and exact diagonalization methods are employed on finite system to verify the exact nature of the phases and transitions .
of all the phases , we pay a special attention to the gapped antiferromagnetic phase , which is disclosed to be a non - trivial one that exhibits the time - reversal symmetry .
we also discuss how our findings could be detected in experiment in the light of ultracold atoms technology advances .
|
the problem of a quantum particle confined to a two - dimensional tube consisting of two straight regions connected by a region of arbitrary curvature but constant width , @xmath2 , has aroused interest in recent years because of its applications in both the physics of small devices and electromagnetic waveguides.@xcite related geometries such as crossed wires@xcite and elbows@xcite have also been studied extensively .
the primary focus has been on the existence of unanticipated bound states and on the `` adiabatic '' limit where the radius of curvature is always much greater than the width of the tube.@xcite goldstone and jaffe proved that such a tube always has a bound state , with energy below the continuum threshold at @xmath3.@xcite in this paper , we study the specific case where the curved region is an arc of a circle .
this case can be solved numerically with relative ease , providing a laboratory in which to explore phenomena we believe to be quite general but difficult to demonstrate in arbitrary geometries .
we take the tube to have unit width .
we use the arc length , @xmath4 , along the outer boundary of the tube as one coordinate , and the distance @xmath5 , along the normal from the outer boundary toward the center of the circle as the other coordinate .
we mark the boundaries of the curved region as @xmath6 .
the angle subtended by the arc of the curved region is defined to be @xmath7 , and it satisfies the relation @xmath8 , where @xmath9 is the curvature ( here a constant ) .
figure [ tube ] shows two examples of such a tube .
notice that we allow ourselves to consider tubes with @xmath10 although such configurations require excursions from strictly planar geometry . also , we choose units such that @xmath11 making @xmath12 dimensionless . scattering in a bent tube is a relatively simple example of a multichannel problem .
incoming and outgoing waves are labelled by an integer , @xmath13 , the number of non - trivial nodes in the transverse wavefunction , @xmath14 .
the scattering is described by transmission and reflection matrices @xmath15 and @xmath16 whose dimension grows as each successive channel threshold ( at @xmath17 ) is passed . in ref .
@xcite it was pointed out that the same argument which demonstrates the existence of a bound state , also proves that there is a `` quasibound '' state just below each channel threshold .
the quasibound state below the @xmath18 channel threshold would be stable and normalizable if all lower channels ( @xmath19 ) were artifically closed .
such quasibound states should appear as resonances in the open channel scattering amplitudes .
our aim in considering this solvable special case is to explore the systematics of the bound states as they depend on @xmath9 and @xmath7 , and to elucidate the character of the channel threshold resonances .
other workers have studied quantum wires with circular bends . lent studied scattering in tubes with @xmath20 and showed that even for the largest curvature tractible there is very little reflection except at values of the energy just above the threshold.@xcite sols and macucci have confirmed this result and discovered the presence of resonances associated with quasi - bound states just below propagation thresholds.@xcite moreover , they found that at large angles , more than one bound state may develop , and that the binding energy of these bound states increases as the curvature or angle is increased . in
ii , we establish a formalism for solving for both bound and scattering states in tubes with circular bends .
we begin traditionally by representing the solution in the curved region in terms of bessel functions and imposing the standard continuity conditions at @xmath21 .
normally one would match the open channels to incoming and outgoing waves and the closed channels to _
falling exponentials_. the matching conditions generate a linear algebra problem , which leads eventually to an eigenvalue problem ( for the bound states ) and a solution for @xmath15 and @xmath16 ( for the scattering problem ) .
however , we find that this approach leads to serious convergence problems at high curvature . at this point
we abandon the traditional approach and formuate a novel method ( as far as we are aware ) where we match the interior wavefunction to _ both rising and falling exponentials _ in closed channels .
we then seek solutions by minimizing the coefficients of the rising exponentials .
we argue that this approach should converge more rapidly and reliably .
we believe this method has wider application than the problem at hand .
also in this section we derive a small @xmath9 approximation to the transmission amplitude in regions far from thresholds and resonances .
in iii we study the convergence of our method of calculating both scattering and bound states .
we show that it is more convergent and more efficient than the traditional method ( of allowing only falling exponentials in closed channels ) . in iv , we present results for the bound state eigenenergies and the scattering coefficients as functions of curvature and angle .
we show that at fixed curvature , further bound states appear with increasing @xmath7 . as @xmath22 , so does the number of bound states .
the binding energy of the ground state approaches a limit determined by the root of a simple bessel equation as @xmath22 .
for @xmath23 the limit is the square of the first zero of @xmath24 ( @xmath25 ) .
the higher bound state energies approach the same limit as the ground state , but more slowly .
we also find the angle at which the first excited state appears , as a function of @xmath9 , and give a theoretical explanation why this angle approaches @xmath26 as @xmath9 approaches zero .
we also plot the scattering coefficients @xmath27 and @xmath28 as functions of the energy and verify the presence of resonances associated with quasi - bound states just below the propagation threshold for each channel .
finally , we explore the behavior of @xmath28 and @xmath27 in the complex plane .
we explain why the elastic transmission amplitude @xmath29 drops to zero at the energy of the quasibound state just below the second threshold , @xmath30 , an effect which is echoed in other components of @xmath27 at higher thresholds . at the same time
we verify the predictions of the small @xmath9 approximation for low curvature .
in this section we summarize the basic formalism common to both the bound and scattering state problems .
we then specialize to the bound state case and finally , to the scattering case .
we are interested in the solutions to the wave equation , @xmath31 within a `` tube '' coordinatized by @xmath4 and @xmath5 .
@xmath32 vanishes on the boundaries , @xmath33 and @xmath34 , and we require that @xmath32 and and its normal derivative be continuous at @xmath21 where the curvature of the tube jumps abruptly from zero to @xmath9 .
we treat the curved and straight regions at the same time , taking @xmath35 for the straight region .
the coordinate differential is given by @xmath36 , and the differential area element is @xmath37 .
thus , the wave equation becomes @xmath38 we now make the standard change of variable , @xmath39 so that eq .
( [ se ] ) becomes @xmath40\phi = k^2\phi , \label{modse}\ ] ] if we define @xmath41 , and separate variables we recognize the radial equation as bessel s equation , with solutions of the form , @xmath42 where @xmath43 \label{radial}\ ] ] and @xmath44 here @xmath45 and @xmath46 are bessel and neumann functions respectively .
the linear combination of bessel functions in eq .
( [ radial ] ) satisfies @xmath47 at @xmath48 .
the order , @xmath49 , must be chosen to satisfy @xmath47 at @xmath50 .
@xmath49 may be either real or imaginary , as we discuss below .
the label @xmath51 denotes the number of non - trivial nodes ( @xmath52 ) in @xmath53 . to study the character of @xmath49 , we substitute eqs .
( [ radial ] ) and ( [ s_l ] ) back into the wave equation .
we find that the momentum satisfies the relation , @xmath54 where @xmath55 is a positive term that grows with @xmath56 .
thus , for fixed @xmath57 only a finite number of eigenvalues , @xmath58 , can be positive .
for any @xmath57 greater than @xmath59 there is at least one positive @xmath60 a fact which is closely related to the existence of a bound state . as @xmath61 increases , more positive @xmath60 values appear associated with the channel threshold resonances discussed below .
let @xmath62 correspond to real @xmath63 and @xmath64 give imaginary @xmath63 , where we define @xmath65 .
then , @xmath66 we have added the subscript `` @xmath67 '' to @xmath32 to denote the solution inside the circular region .
depending on @xmath57 , the exterior solutions , @xmath68 , either oscillate or vary exponentially with @xmath4 .
the general exterior solution is of the form , @xmath69 where @xmath70 when @xmath71 becomes imaginary and the solutions become oscillatory . in that case
we define @xmath72 with @xmath73 .
we now apply continuity conditions at @xmath74 , recalling that in the interior region the normal derivative of @xmath75 is @xmath76 and we find @xmath77 where @xmath78 and @xmath79 .
( [ match ] ) give an infinite dimensional matrix linear algebra problem to solve for the energies of bound states or the @xmath80matrix for scattering states . in practice
we must truncate the sum over internal and external channels .
we superpose @xmath81 internal solutions and attempt to match to @xmath82 external channels . when @xmath83 the @xmath84 with @xmath85 are coefficients of exponentially rising solutions .
because these solutions are not normalizable , standard practice is to dismiss them as unphysical and set @xmath84 for @xmath85 to zero from the beginning .
this leads to an eigenvalue condition for @xmath61 which has unique solutions only when @xmath86 , _ i.e. _ one is forced to use channel spaces of the same dimension both inside and outside the region of curvature . as the dimension of this channel space @xmath87 is increased , the accuracy of the solution should improve .
after all , it corresponds to taking more terms in the approximation to the wavefunction .
however , we have found that this straightforward approach fails at high curvature ( @xmath88 ) . as the channel space dimension
is increased , the solution does not seem to converge .
we believe that this is due to poor matching at the boundary @xmath89 : we expect that an interior function with @xmath13 nodes should match to a superposition of exterior functions dominated by terms with @xmath13 , @xmath90 , and @xmath91 , _ etc . _ nodes .
however , the @xmath92 restriction forbids the introduction of any exterior functions with higher frequencies than the interior functions , so that the @xmath93 , @xmath94 node terms are not present .
clearly it would be advantageous to allow @xmath95 .
however , a simple count of the number of degrees of freedom shows that the there is no solution under these circumstances . to solve this problem ,
we have implemented a novel approach : we retain the rising exponentials and seek first the set of @xmath96 and next the value of @xmath61 for which @xmath97 is minimized .
as we shall see , this approach removes the @xmath86 restriction , allowing us to take as many terms in either sum as we like .
of course , in the limit @xmath98 , @xmath99 in the closed channels for all physically interesting solutions , whether bound or scattering .
bound states correspond to solutions of eqs .
( [ match ] ) with @xmath100 .
our strategy will be to allow the @xmath84 to be non - zero and seek solutions which minimize @xmath97 .
we anticipate that the parity of the ground state will be even , so we choose internal solutions which are superpositions of @xmath101 or @xmath102 .
if there are several bound states , this symmetric _ ansatz _ will find the even ones .
a corresponding _ ansatz _ odd under @xmath103 finds the odd states .
even and odd bound states alternate as a function of the energy .
we use the method of lagrange multipliers to minimize @xmath104 subject to the constraint that @xmath105 .
there is some arbitrariness in this prescription .
the @xmath96 must be constrained in some way , otherwise the trivial solution @xmath106 will follow . the sensitivity of our method to the form of the constraint is discussed in the following section .
we define @xmath107 and seek to minimize @xmath108 first with respect to the @xmath96 and then with respect to k. the conditions @xmath109 , lead to an algebraic problem , with a solution @xmath110 .
substituting back into eq .
( [ delta ] ) , @xmath108 reduces to a function of @xmath61 , whose minima we locate numerically .
first we combine eqs .
( [ match ] ) to obtain @xmath111 as a function of the @xmath96 , @xmath112 next we substitute for @xmath111 in eq .
( [ delta ] ) , and differentiate with respect to @xmath113 to obtain , @xmath114 this is an @xmath81 dimensional , matrix eigenvalue problem .
the external channel space dimension , @xmath82 , appears in the definition of @xmath115 , but does not complicate eq .
( [ xidef ] ) . since @xmath115 is a positive definite matrix , it has real , positive eigenvalues , @xmath116 , @xmath117 _ etc .
_ , ordered from smallest to largest , and associated eigenvectors , @xmath118 , @xmath119 , _
etc_. each eigenvector gives a stationary point of the functional @xmath108 .
normalizing @xmath120 as required by the constraint it is easy to see that @xmath121 so the smallest eigenvalue gives the minimum @xmath122 at fixed @xmath61 .
to find the best approximation to the bound states for fixed @xmath81 and @xmath82 , we search for a value of @xmath123 that minimizes the lowest eigenvalue , @xmath116 .
the magnitude of @xmath116 measures the squared sum of coefficients of rising exponentials , and is a measure of the accuracy of our solution .
there is always at least one minimum of @xmath116 for @xmath123 a consequence of the general theorem of ref .
@xcite . for small values of @xmath124 , there is only one minimum at @xmath125 , corresponding to a ground state energy of @xmath126 . as @xmath124 is increased ( at fixed @xmath9 ) , there will be multiple minima of @xmath116 which we denote by @xmath127 , @xmath128 , @xmath129 , _ etc .
_ , with @xmath130 .
( @xmath127 corresponds to the ground state , @xmath128 to the second excited state , and so on . )
the odd - numbered excited states ( @xmath131 even ) are antisymmetric in @xmath4 and are obtained from an antisymmetric _
we must set the number of terms in the sums in eqs .
( [ xidef ] ) to values such that the computations are tractable while still retaining reasonable accuracy .
studies of convergence and accuracy are presented in iii .
we emphasize , though , that the convergence is not a trivial issue for this problem : the default assumption , @xmath92 , fails to converge as @xmath132 for interesting values of @xmath9 , forcing us to take a different approach . anticipating the results of iii , we note that the results presented in iv were obtained with @xmath133 and @xmath134 , which generated accuracy to about five decimal places .
we now extend our formalism to handle scattering solutions . at an energy @xmath12 ,
the `` longitudinal '' momentum in the @xmath135 channel is @xmath136 . if @xmath137 , then @xmath138 channels are `` open '' : for @xmath139 , @xmath140 is real and wavefunctions in the external regions are oscillatory . for @xmath141
, @xmath140 is imaginary .
these channels are `` closed '' : we define @xmath142 ; the solutions are exponential as in the bound state problem . since internal and external channel sums cover different ranges we will use different indices as follows : @xmath143 , @xmath144 , @xmath61 , and @xmath56 label internal ( curved region ) solutions ; and @xmath145 , @xmath13 , @xmath131 , @xmath146 , and @xmath147 label external ( straight region ) channels .
@xmath138 will always denote the number of open channels .
an incident wave in the @xmath148 channel will couple to all channels .
however , the solutions in the closed channels die exponentially , so that at a distance ( s @xmath149 ) , we see reflection and transmission in only the open channels .
we wish to solve for these @xmath28 and @xmath27 coefficients and explore their behavior as functions of the channel and energy of the incoming wave as well as the parameters of the tube .
unlike the bound state problem , this is not an eigenvalue problem .
there are many solutions ( in fact @xmath150 ) for each value of the energy .
our aim is to find the solutions and parametrize them in terms of the matrices @xmath28 and @xmath27 .
since our internal solutions are naturally constructed as either symmetric or antisymmetric ( in @xmath4 ) , it is convenient to do likewise with the external solutions .
we consider the symmetric case explicitly from which we will obtain @xmath151 .
the antisymmetric case proceeds along identical lines and yields @xmath152 .
the solutions are labelled by the indices @xmath146 , the channel of the incoming wave , and @xmath131 , the channel under consideration .
the symmetric solution is the sum of the left- and right - incoming solutions : @xmath153 unitarity requires @xmath154 , and from the antisymmetric solution , @xmath155 , or @xmath156 these constraints can be summarized by the requirement that the matrix @xmath157 be unitary , @xmath158 this is a special case ( for hamiltonians symmetric under @xmath159 ) of the general definition of the @xmath160matrix , @xmath161 where @xmath28 and @xmath27 are reflection and transmission coefficients for waves incident from the left and @xmath162 and @xmath163 are the analogous coefficients for right incident waves .
time reversal invariance requires @xmath164 , which requires @xmath165 and @xmath166 for symmetric hamiltonians.@xcite to solve for the matrices @xmath27 and @xmath28 , we return to the basis wavefunctions for the bound state problem , modified to accommodate the open channels and the factor of @xmath167 . note that this factor is present only in the open channels where @xmath139 . because our solution is symmetric , we take @xmath168 to be @xmath169 , @xmath170 by comparing eqs .
( [ trarray ] ) and ( [ psiarray ] ) we can identify @xmath151 in terms of @xmath171 and @xmath172 for open channels , @xmath173 for @xmath174 . by analysis similar to the bound state problem , in the case where @xmath49 is imaginary . instead of @xmath175
we use @xmath176 $ ] .
this amounts to multiplying by a constant @xmath177 , which will not change our result but keeps the numbers tractably small when @xmath124 is large .
we must now examine _ where _ @xmath49 is actually imaginary . for real @xmath63
the s - dependence of @xmath32 takes the form @xmath178 . by strict analogy with the straight regions
, we would then expect real @xmath63 for open channels @xmath179 and imaginary @xmath63 for @xmath180 .
however , the energy thresholds for propagation are actually _ lower _ in the curved region @xcite , so that we will find , for example , @xmath181 real beginning not at @xmath182 , but at @xmath183 , where @xmath184 is a number that we can find numerically because we expect @xmath185 at the threshold . ]
continuity at @xmath74 shows that @xmath186 and @xmath187 where @xmath188 and @xmath189 as in the bound state case , tradition would dictate setting the coefficients of growing exponentials in closed channels to zero _ ab initio_. we seek to improve convergence and accuracy by instead keeping the growing exponentials and minimizing their contributions .
we have found that the best convergence for high @xmath9 is obtained by minimizing the _ ratio _ of the growing contributions relative to the falling contributions , _
i.e. _ minimizing @xmath190 subject to the constraint that @xmath191 . because scattering boundary conditions are complex ( _ viz . _
( [ bcdef ] ) ) , we must be more careful about the reality properties of the variables we consider .
certainly @xmath192 and @xmath193 are complex in open channels . likewise , the matrices @xmath194 and @xmath195 are complex , although they are clearly real for @xmath141 .
our manipulations will be simplified considerably by using matrix notation .
even though @xmath195 and @xmath194 are not square matrices , there should be no ambiguity in the definition of quantities such as @xmath196 , _
etc._. we define a matrix projection onto the closed channels : @xmath197 , except @xmath198 when @xmath199 .
note that @xmath200 is an @xmath82 dimensional matrix in the external channel space .
for an incoming wave in the @xmath148 channel ( suppressing the @xmath146 superscript ) , we can once again define the projection onto rising exponentials , and implement the constraint @xmath191 by a lagrange multiplier , @xmath201 where @xmath202 and @xmath203 .
note that @xmath115 and @xmath204 are both real , symmetric matrices in the internal ( @xmath81dimensional ) space . in practical calculations , @xmath81 is small ( compared to the external channel space dimension , @xmath82 ) , so these matrices are relatively easy to manipulate .
it is convenient to work in the basis of eigenvectors of @xmath204 .
let @xmath205 be the orthonormal eigenvectors of @xmath204 corresponding to ( real , positive ) eigenvalues @xmath206 , @xmath207 we rewrite @xmath208 in the basis of the @xmath209 , using @xmath210 as a metric , @xmath211 inserting this definition back into eq .
( [ scatterdelta ] ) , we obtain @xmath212 where @xmath213 now , the condition @xmath214 becomes @xmath215 an @xmath81 dimensional eigenvalue problem with solutions @xmath216 . as in the bound state case , the eigenvalue @xmath217 is the value of @xmath108 at the stationary point .
small @xmath217 therefore correspond to solutions to the wave equation with negligible projection on exponentially growing solutions in the closed channels and are therefore physically interesting .
the number of physically significant solutions to this eigenvalue problem depends on the energy , @xmath57 . below the lowest threshold , @xmath100
, we do not expect ( and did not find ) any solutions with small @xmath108 except at the energies of bound states .
when one channel is open ( @xmath218 ) , there should be a single symmetric solution to the wave equation for each energy .
so we expect to find one eigenvalue , @xmath219 , much smaller than the rest .
when a second channel opens , there are now two symmetric solutions , so we expect to find two small eigenvalues , @xmath219 and @xmath220 .
we conclude that we must expand the vector @xmath208 as a linear combination of the @xmath138 eigenvectors associated with the lowest @xmath138 eigenvalues when @xmath138 channels are open .
we now expand the vector @xmath208 in terms of the lowest @xmath138 eigenvectors @xmath221 of @xmath222 : @xmath223 where @xmath224 is the expansion coefficient and the sum over @xmath61 corresponds to the expansion in terms of @xmath225 as defined in eq .
( [ a - expansion ] ) .
finally , we define the @xmath226 matrix @xmath227 so that we may write compactly @xmath228 we are now prepared to return to our identifications of @xmath229 and @xmath230 from eq .
( [ bcdef ] ) . inserting our above expression for @xmath231 into eq .
( [ b_n - def ] ) , we find that @xmath232 where @xmath233 is an @xmath234 matrix defined by @xmath235 substituting eqs .
( [ final - a ] ) and ( [ d - exp ] ) back into eq .
( [ c_n - def ] ) and recalling our identification with @xmath236 in eq .
( [ bcdef ] ) , we obtain at last @xmath237 note that in the end the matrix manipulations which are calculationally intensive finding eigenvalues , eigenvectors and inverses are only carried out on matrices of dimension @xmath81 . the external channel index , @xmath82 , can therefore be taken very much greater than @xmath81 without increasing the computation time significantly .
the antisymmetric scattering states can be analyzed in the same fashion , beginning with the parameterization of the scattering states , @xmath238 obtained by subtracting the left- and right - incoming wave solutions .
this allows us to solve for the quantity @xmath239 , which we then combine with eq .
( [ t+r ] ) to yield the desired matrices @xmath27 and @xmath28 .
when the curvature , @xmath9 , is small , the solutions to our problem simplify somewhat .
there is a quite well developed theory of the adiabatic approximation for a quantum wire with small and slowly varying curvature @xmath240 and @xmath241 but it does not apply to a circular region joining two straight regions because @xmath9 jumps abruptly to zero at @xmath124 . also ,
most of the results are restricted to @xmath242 .
nevertheless we can use the fact that @xmath9 is small to simplify the calculation of the _ elastic _ scattering amplitudes @xmath243 and @xmath244 in regions away from resonances and bound states .
the result is @xmath245 \exp\left(-ik_m\theta\over 2 + { \cal o}(k_m{\kappa}\theta)\right)\nonumber\\ { { \cal r}}_{mm } & = & { i{\kappa}\over 2}\sin(2k_ms_0-k_m\theta/2)\left[1+{\cal o}({\kappa})\right ] \exp\left(-ik_m\theta\over 2+{\cal o}(k_m{\kappa}\theta)\right ) , \label{smallkappa}\end{aligned}\ ] ] where @xmath246 and these expressions are only valid when @xmath247 .
the derivation of these results begins from eq .
( [ modse ] ) , which we expand in a fourier series , @xmath248 , @xmath249 where
@xmath250 since @xmath251 is of order @xmath9 when @xmath252 , it is easy to see that the effect of the @xmath135 channel on a wave incident in the @xmath253 channel begins at order @xmath254 .
this confirms our decision _ to work only to order @xmath9_. to this order @xmath255 , which makes this system unusual , since the perturbation appears in lowest order only in the kinetic term . in the end
this will lead to the result that an arbitrarily small curvature can result in a large scattering phase provided the product @xmath256 is large . working to order @xmath9
we can ignore the `` potential '' @xmath257 , with the result that @xmath258 , where @xmath259 .
we now construct symmetric and antisymmetric combinations of these interior solutions , and match them and their normal derivatives to the appropriate external scattering states eqs .
( [ trarray ] ) and ( [ asymin ] ) .
since channel coupling is of higher order in @xmath9 , @xmath260 are both unitary , so we parametrize them by phases , @xmath261 .
the result of the matching calculation is @xmath262 note that @xmath263 and @xmath264 differ by @xmath265 which need not be small even when @xmath266 because @xmath267 can be held fixed in this limit .
so the leading effect on @xmath243 and @xmath244 arises from this accumulating phase along the bending section .
the prefactor @xmath268 determines the modulus of @xmath243 and @xmath269 .
when we solve for @xmath270 , we find that effects of order @xmath9 cancel , which they must do since @xmath271 and @xmath272 must give unity to order @xmath254 .
this completes our derivation .
the results , eqs .
( [ smallkappa ] ) , give a useful approximation to the elastic amplitudes over much of their range when @xmath273 .
the straightforward approach to solving the wave equation by matching interior to exterior solutions failed for physically interesting values of curvature and angle , forcing us to develop a different method of calculation .
we are not aware of any studies of this method in the literature , and therefore present some details of our studies of convergence here .
readers not interested in calculational methods should skip directly to the results presented in
the traditional method of matching @xmath81 independent internal solutions to @xmath81 independent external solutions does not converge as @xmath81 increases at large values of @xmath9 .
this is illustrated graphically in figure [ badconv ] , where we plot the calculated ground state energy @xmath127 as a function of @xmath81 for @xmath274 and @xmath275 .
we have attributed this failure to converge to the absence of small wavelength ( in @xmath5 ) states on the outside , necessary to accurately expand the internal states . in order to demonstrate that this is indeed the case
, we have studied the convergence as a function of @xmath81 and @xmath82 .
table [ conv ] shows the effect on the calculated ground state energy @xmath127 of varying the channel spaces @xmath81 and @xmath82 in the case where @xmath276 and @xmath277 .
the best accuracy for the least amount of computation time is obtained by setting @xmath95 .
we can see from the lower left corner of the table that setting @xmath278 does not produce favorable results , because as higher - frequency interior functions enter the picture , we require even more high - frequency exterior functions for a good match at the boundary @xmath74 .
.tabulation of @xmath279 for @xmath280 [ cols="<,<,<,<,<",options="header " , ] \(b ) @xmath281 using the constraint @xmath282 , our actual method . as with the bound state problem
, we find that we must take @xmath283 to achieve convergence . to illustrate this further , figure [ rvsl ] plots @xmath284 for the same parameters , using the method outlined in
ii.c , as a function of @xmath81 for @xmath285 and for @xmath286 . from this
, we can see clearly that the @xmath286 case is unsatisfactory , as expected .
although the speed of convergence varies with the region of interest it is significantly slower when the energy is near a resonance we have taken parameters which give accuracy to at least a part in 100 .
we first examine the dependence of the bound state energies on the parameter @xmath124 , the length of the curved region of the tube .
figure [ eigenenergies ] plots the momenta @xmath287 corresponding to bound state eigenenergies against the angle of the circular arc @xmath7 @xmath288 for curvature @xmath289 .
eigenenergies for both symmetric states and antisymmetric states are shown .
the curved region introduces a term that behaves like an attractive finite square well potential with depth @xmath257 . increasing the length of this curved region
both lowers the energy of the ground state and allows for more excited states .
figure [ multmin ] shows plots of the lowest eigenvalue of the matrix @xmath115 as defined in
ii.b against @xmath61 , for curvature @xmath290 and different values of the angle @xmath7 .
the minima on these graphs are candidate ( symmetric ) bound states provided @xmath291 , and correspond to points in figure [ eigenenergies ] along a line of fixed angle .
as the angle increases , the number of eigenstates also increases . in the limit @xmath292 ,
all the @xmath287 approach the same value , which we will denote by @xmath293 .
the value of @xmath293 is that which corresponds to @xmath294 for any given curvature . to see this
, note that in the limit @xmath295 , the external straight region becomes negligible ( for bound states ) , and the internal curved region becomes very long . in this limit
it becomes possible to accomodate wavefunctions with any number of nodes ( in @xmath4 ) without introducing significant energy .
the longitudinal wavefunction , @xmath296 , ( the antisymmetric case can be handled analogously ) can oscillate any number of times over the range @xmath297 contributing only @xmath298 to the energy . by choosing @xmath63 very small this contribution to the energy
can be made negligible .
a minute tuning of @xmath63 allows one to match the wavefunction to a falling exponential at @xmath74 .
thus as @xmath299 there are an infinite number of bound states accumulating at the value of @xmath61 corresponding to @xmath185 . at any
fixed very large @xmath124 there are also many other bound states at energies between @xmath293 and @xmath300 .
the value of @xmath293 is determined by solving the transcendental equation , @xmath301 = 0 . \label{condition}\ ] ] this provides us with @xmath302 . as @xmath303 , @xmath304 becomes singular in eq .
( [ condition ] ) .
a careful analysis of the behavior at small @xmath305 shows that the equation reduces to @xmath306 , which gives @xmath307.@xcite figure [ kk ] plots the solution of eq .
( [ condition ] ) , @xmath302 against @xmath308 . as @xmath266 @xmath309 ,
the tube is straight and so @xmath294 corresponds to @xmath310 , as expected . and
as @xmath303 , @xmath311 as expected .
it should be noted that the amount of binding increases dramatically as @xmath303 , a fact which necessitated our choice of @xmath308 as the independent variable in fig .
[ kk ] .
next we turn our attention to determining the angle at which the first ( antisymmetric ) bound excited state appears .
we label the angles at which new bound states appear as @xmath312 , @xmath313 , _ etc . _ where @xmath312 corresponds to the ground state and @xmath314 @xmath315 to the @xmath316 excited state . for any @xmath9 ,
@xmath312 is zero , since there exists a bound state for any tube with a curved region , no matter how short .
figure [ antisym ] shows a plot of @xmath317 .
as can be seen from the graph , @xmath313 approaches @xmath26 as @xmath9 approaches zero .
we can see that this must be the case by considering the one - channel approximation to the interior wavefunction , namely it is sufficient to consider this approximation because in the limit that @xmath266 , the tube is very close to straight and thus the contributions from higher channels are negligible . in this limit
, the problem reduces to that of a square well potential .
the schrdinger equation then requires at the angle where the bound state first appears , @xmath320 so eq . ( [ bscond ] ) implies @xmath321 .
for a bound state to exist , the interior wavefunction ( @xmath322 ) must be able to match to a falling exponential at the boundary @xmath74 .
the smallest value of @xmath124 at which this can occur is that which satisfies the condition @xmath323 . recalling that @xmath324
, we then have we note that as @xmath9 is increased , the value of @xmath313 also increases .
this seems somewhat counter - intuitive , as we expect from the schrdinger equation that higher curvature implies greater binding , so that the length of the curved region necessary to bind an excited state should _ decrease _ with large @xmath9 .
this is indeed the case .
however , we must recall that @xmath7 depends on @xmath9 as well as @xmath124 . for a fixed value of @xmath7
, @xmath124 is much larger for small @xmath9 than for large @xmath9 .
thus , the attractive potential acts over a much longer distance for small @xmath9 , and this effect overshadows the difference in binding energy due to differences in @xmath9 alone .
previous workers have noted many features of the scattering problem , and we do not have any qualitatively new phenomena to add to their treatments.@xcite the primary features of interest to us are the rapid rise of @xmath326 from threshold , the steady increase of the phase of @xmath243 in regions of energy away from thresholds , and finally , the dramatic fluctuations in @xmath27 and @xmath28 at energies just below thresholds . our interest lies in understanding better the physical origins of these effects .
these features are due to the presence of a bound state just below threshold , a steadily growing phase over a wide range of @xmath61 , and weakly coupled quasibound states just below each channel threshold , respectively .
the last of these provides an elegant example of a resonance manifesting itself as total destructive interference a phenomenon familiar to particle physicists in the form of the sudden drop in the @xmath327 @xmath328-wave cross section just below @xmath329 threshold , due to the presence of the @xmath330 resonance.@xcite we begin by examining @xmath331 and @xmath284 as functions of @xmath61 .
figure [ tr ] plots @xmath332 and @xmath333 , the measures of transmitted and reflected flux , for @xmath334 at fixed @xmath335 and @xmath20 . note that in the region where @xmath336 ( @xmath337 ) , @xmath338 .
this is required by unitarity and serves as a check on our numerical calculations .
we have checked that our @xmath160-matrix remains unitary in the @xmath234 case as well .
as @xmath61 crosses the second threshold at @xmath26 there is significant inelasticity into the newly open channel evidenced by the fact that @xmath339 .
we also observe that at energies just below @xmath340 , the transmission amplitude in the highest open channel ( @xmath341 ) drops sharply almost to zero .
we shall show that this and other rapid variations in @xmath27 and @xmath28 are manifestations of resonances related to the presence of quasibound states just below each new channel threshold .
they are the energies at which we would find a bound state if all open channels were artificially closed .
it is difficult to figure out what is happening by considering @xmath342 and @xmath343 alone .
we will learn much more by looking at the argand diagrams for @xmath27 and @xmath28 .
the general characteristics will be illustrated by considering plots of @xmath29 and @xmath344 in the complex plane for @xmath345 as shown in figure [ argand ] .
@xmath346 and @xmath347 are not independent when only one channel is open .
unitarity ( @xmath348 ) requires both @xmath349 and @xmath350 .
we can parameterize both by the modulus ( @xmath82 ) and phase ( @xmath351 ) of @xmath347 , @xmath352 so it suffices to discuss the behavior of @xmath29 . according to fig .
( [ argand ] ) , @xmath29 vanishes at @xmath320 . as @xmath61 increases
, @xmath347 rapidly executes a clockwise circle of radius @xmath353 in the complex plane until @xmath354 at a @xmath61 only slightly larger than @xmath300
. then @xmath347 moves clockwise on the unit circle until suddenly , at a @xmath61 just below @xmath26 , @xmath347 suddenly executes an almost complete counterclockwise circle of radius @xmath353 . @xmath347
briefly resumes its steady phase growth , until the second channel opens at @xmath355 and @xmath347 moves off the unit circle .
first consider the behavior near @xmath320 .
we know on general grounds that @xmath356 and @xmath357 at threshold .
furthermore , we know that there is a symmetric bound state at an energy @xmath358 .
this must appear as a pole in @xmath359 at this energy . since @xmath160 must be unitary for any real @xmath57 ,
@xmath360 must be of the form , @xmath361 where @xmath362 , and @xmath363 is a smoothly varying function of @xmath61 near @xmath320 .
there is no nearby antisymmetric bound state , so @xmath364 is a smooth function of @xmath61 , @xmath365 near @xmath320 . comparing with the threshold behavior , @xmath356 and @xmath357
, we see that @xmath366 and @xmath367 . combining all these results
we obtain , @xmath368 at threshold , @xmath369 , and as @xmath61 increases , @xmath351 drops rapidly to zero over an interval scaled by @xmath370 .
( [ thresh ] ) parameterize a semicircle of radius @xmath353 centered at @xmath371 , which @xmath347 executes as @xmath372 grows from zero to a value greater than @xmath370 .
the speed at which @xmath347 moves measures @xmath370 , by the formula @xmath373 for @xmath374 the conditions of the small @xmath9 approximation discussed in ii are satisfied , provided @xmath9 is small .
thus we expect to find @xmath375 and @xmath376 .
we verify this by comparing the rate of change of the argument of @xmath347 with the small @xmath9 prediction , @xmath377 .
so in the range where the approximation is valid , we expect that the graph of the phase angle @xmath351 of @xmath27 will be linear with respect to @xmath372 , with a slope equal to half the angle subtended by the curved region of the tube .
computing @xmath29 for @xmath378 with @xmath379 and @xmath380 , a range and curvature for which the small @xmath9 approximation is valid , we have found @xmath381 , which confirms that this relation is indeed satisfied .
finally we turn to the dramatic behavior of @xmath27 and @xmath28 near channel thresholds .
it is clear from fig .
( [ argand ] ) that @xmath347 resonates ( _ i.e. _ it rapidly executes a counterclockwise circle in the complex plane@xcite ) at the energy where @xmath382 drops precipitously to zero .
this behavior is indicative of a pole in the complex @xmath57 plane just below the real axis .
the pole appears in @xmath360 because the quasibound state is symmetric in @xmath4 .
@xmath364 is smooth over this region in @xmath61 .
if we denote the pole location as @xmath383 , where @xmath384 and @xmath385 is small and positive ( by causality ) . then in the neighborhood of the pole , @xmath386 where @xmath387 denotes other , smooth terms in @xmath360 .
combining this with a smooth form for @xmath364 we see that @xmath347 executes a counterclockwise circle of radius @xmath353 in the complex @xmath61-plane as @xmath61 passes over the pole .
just before the vicinity of the pole , @xmath347 lay nearly on the unitarity circle , @xmath354 .
the only way that it could rapidly execute a counterclockwise circular path consistent with unitarity is if the resonant circle is _ tangent _ to the unitarity circle .
[ otherwise @xmath347 would have to pass outside the unit circle at some point on its circular path . ]
since the radius of the resonance circle is @xmath353 , the ( tangent ) resonant circle _ must pass through the origin_. thus @xmath382 must vanish in the near vicinity of the quasibound state as observed by ref .
@xcite , and shown in fig .
( [ argand ] ) .
the basic structure of @xmath29 and @xmath344 can be seen in the scattering coefficients for higher channels . as we observe in figure [ argand2 ] for @xmath388 , @xmath389 and @xmath390 follow the same pattern .
for @xmath391 just above zero , @xmath27 executes rapid clockwise motion as a result of the pole located near @xmath392 .
this is followed by smooth phase growth until @xmath61 reaches the quasi - bound state just under the @xmath393 threshold , at which time it executes the rapid counterclockwise circle in the complex plane characteristic of an elastic resonance .
our conclusion from this exercise is that the behavior observed in fig .
[ argand ] is characteristic of scattering in bent tubes .
the bound states , quasibound states and regions of smooth phase growth are robust properties of these systems .
they do not depend specially on our choice of a circular geometry which facilitated our numerical calculations .
perhaps the most interesting observation is the appearance of violent fluctuations in transmission and reflection properties of bent tubes in the vicinity of channel thresholds .
the fluctuations occur over very narrow intervals in energy ( the resonances are narrow ) , so low resolution experiments would quite likely fail to detect them . on the other hand , careful experiments using , for example , microwave radiation in waveguides , should see these rapid fluctuations in response associated with the opening of channel thresholds in which quasibound states occur . f. lenz , t. j. londergan , e. moniz , r. rosenfelder , m. stingl , and k. yazaki , ann .
( ny ) * 170 * ( 1986 ) 65 ; j. p. carini , j. t. londergan , k. muller , and d. p. murdock , phys . rev .
* b48 * ( 1993 ) 4503 .
|
we study the solutions to the wave equation in a two - dimensional tube of unit width comprised of two straight regions connected by a region of constant curvature .
we introduce a numerical method which permits high accuracy at high curvature .
we determine the bound state energies as well as the transmission and reflection matrices , @xmath0 and @xmath1 and focus on the nature of the resonances which occur in the vicinity of channel thresholds .
we explore the dependence of these solutions on the curvature of the tube and angle of the bend and discuss several limiting cases where our numerical results confirm analytic predictions .
= 10000 = 10000 = 10000 = 5000 = 10000 = 100 = 1000 and
|
the existence of the electric dipole moments ( edms ) of fermions depends on the cp violating interactions .
the complex cabibo kobayashi maskawa ( ckm ) matrix elements cause the cp violation in the standard model ( sm ) , however , the estimated fermion edms are negligibly small , if the above complex phases are considered .
this stimulates one to investigate these physical quantities in the framework of the new models beyond the sm , such as multi higgs doublet models ( mhdm ) , supersymmetric model ( susy ) , @xcite , ... , etc .. in the literature there exist the experimental results on the fermion edms , @xmath2 @xcite , @xmath3 @xcite and @xmath4 @xcite respectively and on the neutron edm h @xmath5 @xcite .
there is an extensive theoretical work done on the edms of fermions .
the quark edms have been estimated in several models @xcite and the edms of nuclei , deutron , neutron and some atoms have been studied extensively @xcite .
the lepton electric dipole moments have been predicted in various studies @xcite-@xcite . in @xcite
the lepton electric dipole moments has been analyzed in the framework of the seesaw model .
@xcite was devoted to the edms of the leptons in the model iii version of the 2hdm and @xmath6 has been predicted at the order of the magnitude of @xmath7 .
the work @xcite was related to the lepton edm moments in the framework of the sm with the inclusion of non - commutative geometry .
furthermore , the effects of non - universal extra dimensions on the electric dipole moments of fermions in the two higgs doublet model have been estimated in @xcite .
this work is devoted to the prediction of the lepton edms in the two higgs doublet model in which the flavor changing ( fc ) neutral current vertices in the tree level is permitted and the cp violating interactions are carried by complex yukawa couplings .
furthermore , we respect the split fermion scenario where the hierarchy of fermion masses are coming from the overlap of the fermion gaussian profiles in the extra dimensions .
the split fermion scenario has been studied in several works in the literature @xcite-@xcite . in @xcite
the alternative view of the fermion mass hierarchy has been introduced by assuming that the fermions were located at different points in the extra dimensions and this geometric interpretation resulted in the exponentially small overlaps of their wavefunctions .
the separation of fermions in the extra dimension forbids the local couplings between quarks and leptons and this can ensure a solution also to the proton stability .
@xcite was devoted to the locations of left and right handed components of fermions in the extra dimensions and their roles on the mechanism of yukawa hierarchies .
the constraint on the split fermions in the extra dimensions has been obtained by considering leptonic w decays and the lepton violating processes in @xcite .
the discussion of cp violation in the quark sector in the split fermion model was done in @xcite .
@xcite was related to the new configuration of split fermion positions in a single extra dimension and the physics of kaon , neutron and b / d mesons to find stringent bounds on the size of the compactification scale 1/r .
the contributions due to the split fermion scenario on the rare processes has been examined in @xcite and the shapes and overlaps of the fermion wave functions in the split fermion model have been studied in @xcite . in the present work ,
we consider the edms of charged leptons by assuming that they have gaussian profiles in the extra dimensions .
first , we study the edms of charged leptons in a single extra dimension , using the estimated location of them .
then , we assume that the number of extra dimensions is two and the charged leptons are restricted to the fifth extra dimension , with non - zero gaussian profiles . as a final analysis , we predict the edms of charged leptons by taking non - zero gaussian profiles in both extra dimensions . in the numerical calculations
, we observe that the electron edm @xmath6 is at the order of @xmath8 e - cm and it is too small to study the additional effects due to the extra dimensions .
the numerical value of @xmath9 ( @xmath10 ) is at most at the order of the magnitude of @xmath11 e - cm in the case that the leptons have non - zero gaussian profiles in the first extra dimension , for two extra dimensions , for the large values of the compactification scale @xmath12 and for the intermediate values of yukawa couplings .
the paper is organized as follows : in section 2 , we present edms of charged leptons in the split fermion scenario , in the two higgs doublet model .
section 3 is devoted to discussion and our conclusions .
the fermion edm is carried by the cp violating fermion - fermion - photon effective interaction and , for quarks ( for charged leptons ) , the complex ckm matrix ( possible lepton mixing matrix ) elements is the possible source of this violation , in the framework of the sm .
the estimated tiny numerical values of edms of fermions in the sm stimulates one to go beyond and the model iii version of the 2hdm is one of the candidate to get relatively large edm values , since the fc neutral currents ( fcnc ) are permitted at tree level and the new yukawa couplings can be complex in general .
furthermore , the addition of the spatial extra dimensions brings additional contributions sensitive to the compactification scale @xmath12 where @xmath13 is the radius of the compactification .
here we take the effects of extra dimensions into account and we follow the idea that the hierarchy of lepton masses are coming from the lepton gaussian profiles in the extra dimensions .
the yukawa lagrangian responsible for the lepton edm in a single extra dimension , respecting the split fermion scenario , reads : @xmath14 where @xmath15 and @xmath13 denote chiral projections @xmath16 , @xmath17 is the new scalar doublet . here
@xmath18 ( @xmath19 ) , with family indices @xmath20 , are the zero mode lepton doublets ( singlets ) with gaussian profiles in the extra dimension @xmath21 and they read @xmath22 with the normalization factor @xmath23 .
@xmath24 ( @xmath25 ) are the lepton doublets ( singlets ) in four dimension .
the parameter @xmath26 is the gaussian width of the leptons with the property @xmath27 , @xmath28 are the fixed position of @xmath29 left ( right ) handed lepton in the fifth dimension .
the positions of left handed and right handed leptons are obtained by taking the observed masses into account @xcite .
the idea is that the lepton mass hierarchy is due to the the relative positions of the gaussian peaks of the wave functions located in the extra dimension @xcite . by assuming that the lepton mass matrix is diagonal , one possible set of locations for the lepton fields read ( see @xcite for details )
@xmath30 here we choose the higgs doublets @xmath31 and @xmath17 as @xmath32\ , ; \phi_{2}=\frac{1}{\sqrt{2}}\left(\begin{array}{c c } \sqrt{2 } h^{+}\\ h_1+i h_2 \end{array}\right ) \,\ , . \label{choice}\end{aligned}\ ] ] with the vacuum expectation values , @xmath33 and collect sm ( new ) particles in the first ( second ) doublet .
notice that @xmath34 and @xmath35 are the mass eigenstates @xmath36 and @xmath37 respectively since no mixing occurs between two cp - even neutral bosons @xmath38 and @xmath36 at tree level , in our case .
the new higgs field @xmath17 , playing the main role in the existence of the charged lepton edm , is accessible to extra dimension and after the compactification on the orbifold @xmath39 , it is expanded as @xmath40 where @xmath41 is the 4-dimensional higgs doublet which contains the charged higgs boson @xmath42 , the neutral cp even ( odd ) higgs bosons @xmath36 ( @xmath37 ) and @xmath43 are the kk modes of @xmath17 .
the non - zero kk mode of higgs doublet @xmath17 includes a charged higgs of mass @xmath44 , a neutral cp even higgs of mass @xmath45 , a neutral cp odd higgs of mass @xmath46 where @xmath47 is the mass of @xmath48th level kk particle .
notice that the gauge field kk modes do not bring new contributions to the edms of charged leptons since they do not exist in the one loop diagrams + ( see fig .
[ fig1 ] ) .
now , we present the vertices existing in the diagrams , after the integration over the fifth dimension . the integration of the combination @xmath49 , ( @xmath50 ) , appearing in the part of the lagrangian ( eq .
( [ lagrangian ] ) ) , over the fifth dimension reads @xmath51 where the factor @xmath52 is @xmath53 \,\ ,
. \label{vij1}\end{aligned}\ ] ] here the fields @xmath54 , @xmath55 are the four dimensional lepton fields
. therefore we can define the yukawa couplings in four dimensions as @xmath56 where @xmath57 are yukawa couplings in five dimensions ( see eq .
( [ lagrangian ] ) ) . with the definition @xmath58 where n denotes the word `` neutral '' . ] in the case of two extra dimensions , after the compactification on the orbifold @xmath59 , the new higgs field @xmath17 can be expanded as @xmath60 where @xmath61 are the kk modes of @xmath17 .
notice that the mass of kk modes of the charged ( neutral cp even , neutral cp odd ) higgs is @xmath62 , ( @xmath63 , + @xmath64 ) where @xmath65 is the mass of @xmath66th level kk particle . in the case that the leptons are restricted only the fifth dimension , the vertex factor @xmath67 is the same as the one in eq .
( [ vij1 ] ) .
on the other hand , if we assume that the leptons are accessible two both dimensions with gaussian profiles as @xmath68 with the normalization factor @xmath69 , the integration of the part of the lagrangian + @xmath70 over the fifth and sixth extra dimensions results in the vertex factor @xmath71 \ , .
\label{vij2}\end{aligned}\ ] ] similar to a single extra dimension case , we define the yukawa couplings in four dimension as @xmath72 where @xmath73 . here , we present a possible positions of left handed and right handed leptons in the two extra dimensions , by respecting the observed masses .
similar to the previous discussion , we assume that the lepton mass matrix is diagonal and one of the possible set of locations for the gaussian peaks of the lepton fields in the two extra dimensions read @xmath74 where the numbers in the parenthesis denote the y and z coordinates of the location of the gaussian peak in the extra dimensions . here
we choose the same numbers for the y and z locations of the gaussian peaks .
now , we would like to present the edms of charged leptons with the addition of a single extra dimension where the localized leptons have gaussian profiles .
the effective edm interaction for a charged lepton @xmath75 is given by @xmath76 where @xmath77 is the electromagnetic field tensor , @xmath78 is edm of the charged lepton and it is a real number by hermiticity . with the assumption that there is no ckm type lepton mixing matrix and ignoring possible cp violating lfv interactions due to the lepton - lepton kk mode - higgs kk mode vertices , only the new neutral higgs part gives a contribution to their edms and @xmath75-lepton edm @xmath79 @xmath80 can be calculated as a sum of contributions coming from neutral higgs bosons @xmath81 and @xmath82 , @xmath83 for @xmath84 and @xmath85 where @xmath86\ , \ , , \nonumber \\
l}+y_{\tau r})}{2\ , r}]\ , , \label{coeff}\end{aligned}\ ] ] for @xmath87 and the functions @xmath88 , @xmath89 read @xmath90 with @xmath91 , @xmath92 and @xmath93 , @xmath94 , @xmath95 and @xmath96 , @xmath97 , @xmath98 are charges of @xmath99 and @xmath100 leptons respectively . in eq .
( [ emuedm ] ) we take into account only internal @xmath99-lepton contribution respecting our assumption that the yukawa couplings @xmath101 , are small compared to @xmath102 due to the possible proportionality of the yukawa couplings to the masses of leptons in the vertices . in eq . ( [ tauedm ] ) we present also the internal @xmath100-lepton contribution , which can be neglected numerically .
notice that , we make our calculations in arbitrary photon four momentum square @xmath103 and take @xmath104 at the end .
now , we present the edms of charged leptons with the addition of two extra dimensions for the case that the leptons are accessible to both extra dimensions : @xmath105 for @xmath84 and @xmath106 where @xmath107\ , \ , , \nonumber \\ c'_{2\,(n , s)}(f,\tau)&=&\cos[\frac{n\,(y_{f l}+y_{\tau r } ) + s\,(z_{f l}+z_{\tau r})}{2\ , r}]\ , , \label{coeff2}\end{aligned}\ ] ] for @xmath87 . in eqs .
( [ emuedm2 ] ) and ( [ tauedm2 ] ) , the parameters @xmath108 , @xmath109 and @xmath110 are defined as , @xmath111 , @xmath112 and @xmath113 . in eqs .
( [ emuedm2 ] ) and ( [ tauedm2 ] ) the summation would be done over @xmath114 except @xmath115 . finally ,
in our calculations , we choose the yukawa couplings complex and we used the parametrization @xmath116 therefore , the yukawa factors in eqs .
( [ emuedm ] ) , ( [ tauedm ] ) , ( [ emuedm2 ] ) and ( [ tauedm2 ] ) can be written as @xmath117 where @xmath118 . here
@xmath119 is the cp violating parameter which is the source of the lepton edm .
this work is devoted to the analysis of the effects extra dimensions on the edms of fermions in the case that the hierarchy of lepton masses are due to the lepton gaussian profiles in the extra dimensions .
the cp violating nature of the edm interactions needs the cp violating phases . here , for the complex phases
, we consider the complex yukawa couplings appearing in the fcnc at tree level in the framework of the 2hdm .
notice that we do not take the internal lepton kk modes into account ( see for example @xcite for the calculation of the kk modes of leptons ) .
the lepton - lepton kk mode - higgs zero mode and lepton - lepton kk mode - higgs kk mode vertices carry the possible lfv and cp violating interactions .
we ignored these contributions because of the difficulty arising during the summations .
we expect that , for the large values of the compactification scale , their effects on the physical parameters are suppressed . the four dimensional leptonic complex couplings @xmath120 are the free parameters of 2hdm .
we consider the yukawa couplings @xmath121 , as smaller compared to @xmath122 and we assume that @xmath123 is symmetric with respect to the indices @xmath124 and @xmath125 . in the case that no extra dimension exists , the upper limit of @xmath126 is predicted as @xmath127 ( see @xcite and references therein ) by using the experimental uncertainty , @xmath128 , in the measurement of the muon anomalous magnetic moment @xcite and assuming that the new physics effects can not exceed this uncertainty
. using this upper limit and the experimental upper bound of br of @xmath129 decay , br @xmath130 , the coupling @xmath131 can be restricted in the range , @xmath132 ( see @xcite ) . in our calculations
we choose the numerical values of the couplings @xmath133 ( @xmath134 ) around @xmath127 ( @xmath135 ) . for the coupling @xmath136
, we use the numerical values which are greater than @xmath133 , since we have no explicit restriction region . here , we respect the split fermion scenario where the hierarchy of lepton masses are due to the lepton gaussian profiles in the extra dimensions .
the sm scalar @xmath38 is a constant profile in the extra dimensions and the mass term , which is modulated by the mutual overlap of lepton wavefunctions , is obtained by integrating the operator @xmath137 over extra dimensions .
this idea is the main point to fix the position of left ( right ) handed lepton in the extra dimensions ( see @xcite for details ) . since the leptons are located in the extra dimensions with gaussian profiles , the parameter @xmath138 , where @xmath26 is the gaussian width of the fermions , is the free parameter of the model .
the locations of lepton fields in the extra dimensions are obtained in terms of gaussian width @xmath26 . in the present work
we take split leptons in a single and two extra dimensions and use a possible set of locations to calculate the strength of the lepton - lepton - new higgs scalars vertices , which play the main role in the calculation of the charged lepton edm .
first , we take a single extra dimension and use the estimated location of the leptons ( see eq .
( [ location ] ) ) to calculate the corresponding vertices ( eq .
( [ intvij1 ] ) ) .
after that , we assume that the number of extra dimensions is two and take that the leptons are restricted to the fifth extra dimension , with non - zero gaussian profiles . in this case the abundance of new scalar kk modes causes to increase the edm of charged leptons , especially the @xmath99 lepton edm
. however , the exponential suppression factor ( see eq . ( [ vij1 ] ) ) appearing in the summation of the kk modes causes that the sum does not have a large contribution . finally , we assume that the leptons have non - zero gaussian profiles also in the sixth dimension and using a possible set of locations in the fifth and sixth extra dimensions ( see eq .
( [ location2 ] ) ) , we calculated the edm of charged leptons . in this case
the additional the exponential factor appearing in the second summation further suppresses the lepton edm especially for the muon case .
now , we start to estimate the charged lepton edms and to study the parameter @xmath139 and the compactification scale @xmath12 dependencies of these measurable quantities . in fig .
[ edmmu500ro ] , we plot edm @xmath9 with respect to the parameter @xmath139 for @xmath140 , @xmath141 , @xmath142 and the intermediate value of @xmath143 . here the lower - upper solid ( dashed , small dashed ) line represents the edm for a single - two extra dimensions , for @xmath144 .
the edm is slightly larger for the case that the leptons have non - zero gaussian profiles in the first extra dimension , compared to the one where the leptons have non - zero gaussian profiles in both extra dimensions .
it is observed that @xmath145 is weakly sensitive to the parameter @xmath139 in the given interval , for the chosen value of compactification scale @xmath12 . in the two extra dimensions the numerical value of @xmath9 is larger compared to the one single extra dimension since there is a crowd of kk modes .
however the suppression exponential factor appearing in the summations causes that the contributions do not increase extremely .
the numerical value of @xmath9 is at the order of the magnitude of @xmath146 for @xmath147 , in the case that the leptons have non - zero gaussian profiles in the first extra dimension .
[ edmmu001rr ] is devoted to the edm @xmath9 with respect to the compactification scale @xmath12 , for @xmath148 , @xmath141 , @xmath142 and the intermediate value of @xmath143 . here the lower - upper solid ( dashed , small dashed ) line represents the edm for a single - two extra dimensions , where the leptons have non - zero gaussian profiles in the first extra dimension , for @xmath149 .
this figure shows that @xmath145 is weakly sensitive to the compactification scale @xmath12 , especially for @xmath150 in fig .
[ edmtau500ro ] , we present edm @xmath10 with respect to the parameter @xmath139 for @xmath140 , @xmath141 , @xmath142 and the intermediate value of @xmath151 . here
the lower - upper solid ( dashed , small dashed ) line represents the edm for a single - two extra dimensions , where the leptons have non - zero gaussian profiles in the first extra dimension , for @xmath152 .
for the case where the leptons have non - zero gaussian profiles in both extra dimensions , the numerical value of @xmath10 is almost the same as the one where the leptons have non - zero gaussian profiles in only one extra dimension , for two extra dimension scenario . it is shown that @xmath153 is weakly sensitive to the parameter @xmath139 in the given interval . due to the crowd of kk modes , in the two extra dimensions ,
the numerical value of @xmath10 is almost five times larger compared to the one in the single extra dimensions .
the numerical value of @xmath10 is at the order of the magnitude of @xmath1 for @xmath154 , in the two extra dimensions .
[ edmtau001rr ] represents the compactification scale @xmath12 dependence of the edm @xmath10 , for @xmath148 , @xmath155 , @xmath142 and the intermediate value of @xmath151 . here
the lower - upper solid ( dashed , small dashed ) line represents the edm for a single - two extra dimensions , where the leptons have non - zero gaussian profiles in the first extra dimension , for @xmath152 . for the case where the leptons have non - zero gaussian profiles in both extra dimensions the numerical values of @xmath10
is almost the same as the one where the leptons have non - zero gaussian profiles in only one extra dimension .
similar to the @xmath100 edm case @xmath153 is weakly sensitive to the compactification scale @xmath12 , especially for @xmath150 . + + now we would like to summarize our results : * @xmath145 is weakly sensitive to the parameter @xmath139 for @xmath156 and the compactification scale @xmath157 .
due to the abundance of kk modes , in the two extra dimensions the numerical value of @xmath9 is slightly larger compared to the one in the single extra dimension .
the numerical value of @xmath9 is at most at the order of the magnitude of @xmath158 for @xmath159 , in the case that the leptons have non - zero gaussian profiles in the first extra dimension .
* @xmath153 is weakly sensitive to the parameter @xmath139 for @xmath156 and the compactification scale @xmath157 .
the crowd of kk modes in the two extra dimensions bring additional contributions which enhance @xmath10 almost five times compared to the one in a single extra dimension .
the numerical value of @xmath10 is at the order of the magnitude of @xmath1 for @xmath154 , in the two extra dimensions .
* the addition of the effects of the internal lepton kk modes brings an extra dependence of the physical parameters to the scale @xmath12 , however , for the large values of the parameter @xmath12 , hopefully , these contributions do not affect the scale @xmath12 dependence of the physical parameters .
with the help of the forthcoming most accurate experimental measurements , the valuable information can be obtained about the existence of extra dimensions and the possibilities of gaussian profiles of the leptons .
this work has been supported by the turkish academy of sciences in the framework of the young scientist award program .
( eoi - tuba - gebip/2001 - 1 - 8 ) 1 c. r. schmidt and m. e. peskin , _ phys
* 69 * ( 1992 ) 410 .
e. d. commins et.al .
_ phys . rev .
a _ * 50 * ( 1994 ) 2960 . j. bailey et al , _ journ .
* g4 * ( 1978 ) 345 ; particle data group , d. e. groom et.al .
_ european phys .
journ . _ * c15 * ( 2000 ) 1 . k. f. smith et.al , _ phys .
lett . _ * b234 * ( 1990 ) 191 , 2885 ; i. s. altarev et .
al , _ phys .
* b276 * ( 1992 ) 242 .
e. p. sahabalin , _ sov .
_ * 28 * ( 1978 ) 75 ; j. f. donoghue , _ phy . rev . _ * d18 * ( 1978 ) 1632 ; a. czarnecki and b. krause , _ phys .
* 78 * ( 1997 ) 4339 ; i. b. khriplovich , _ yad .
* 44 * ( 1986 ) 1019 ( _ sov .
* 44 * ( 1986 ) 659 ) ; _ phys .
_ * b173 * ( 1986 ) 193 ; a. czarnecki and b. krause , _ acta phys .
_ * b28 * ( 1997 ) 829 ; _ phys .
* 78 * ( 1997 ) 4339 ; s , weinberg , _ phys .
* 58 * ( 1976 ) 657 ; n. g. deshpande and e. ma , _ phys .
rev._*d 16 * ( 1977 ) 1583 ; for a recent review , see h. y. cheng , _ int
phys._*a 7 * ( 1992 ) 1059 ; g. c. branco and m. n. rebelo,_phys .
* b 160 * ( 1985 ) 117 ; j. liu and l. wolfenstein , _ nucl .
phys .. _*b 289 * ( 1987 ) 1 ; c. h. albright , j. smith and s. h. h. tye , _ phys .
d. _ * d 21 * ( 1980 ) 711 ; d. atwood , s. bar - shalom , a. soni , _ phys .
rev . _ * d 51 * ( 1995 ) 1034 ; e. iltan , _ j. phys .
_ * g27 * ( 2001 ) 1723 ; e. iltan ,
_ phys.rev . _ * d65 * ( 2002 ) 073013 ; y. liao and x. li , _ phys .
rev . _ * d60 * ( 1999 ) 073004 ; d. g. dumm and g. a. g sprinberg , , _ eur - phys . j. _ * c 11 * ( 1999 ) 293 ; a. soni and r. m. xu , _ phys .
* 69 * ( 1992 ) 33 .
v. baryshevsky , hep - ph/0402209 ; o. lebedev , k. a. olive , m. pospelov , a. ritz , _ phys . rev . _ * d70 * ( 2004 ) 016003 ; a. g. rodriguez , m. a. h. ruiz , a. santiago , _ phys .
rev . _ * d69 * ( 2004 ) 073008 ; a. gutierrez - rodriguez , m.a.hernandez-ruiz and l.n .
luis - noriega , _ mod .
* a19 * ( 2004 ) 2227 ; r. e. mischke , aip conf.proc.675:246-250,2003 ; d. a. demir , oleg lebedev , k. a. olive , m. pospelov , a. ritz ,
_ nucl.phys .
_ * b680 * , ( 2004 ) 339 ; m. endo , m. kakizaki , m. yamaguchi , _ phys.lett . _
* b 583 * ( 2004 ) 186 ; d. kawall , f. bay , s. bickman , y. jiang , d. demille , _ phys .
* 92 * ( 2004 ) 133007 ; i. masina , c. savoy , _ phys.lett . _
* b579 * ( 2004 ) 99 ; j. hisano , y. shimizu , _ phys.lett . _ * b581 * ( 2004 ) 224 - 230 ; a. syntfeld , h. mach , w. kurcewicz , b. fogelberg , w. pociennik , e. ruchowska _ phys.rev . _ * c68 * ( 2003 ) 024304 ; t. suzuki , h. sagawa , k. hagino , _ phys.rev . _ * c68 * ( 2003 ) 014317
. b. dutta , r. n. mohapatra , _ phys .
* d68 * ( 2003 ) 113008 .
e. iltan , _ phys .
* d64 * ( 2001 ) 013013 .
e. iltan , _ jhep _ * 065 * ( 2003 ) 0305 .
e. iltan , _ jhep _ * 0404 * ( 2004 ) 018 .
n. arkani - hamed and m. schmaltz , _ phys .
rev . _ * d61 * , 033005 ( 2000 ) ; n. arkani - hamed , y. grossman and m. schmaltz , _ phys .
_ * d61 * ( 2000 ) 115004 .
e. a. mirabelli , schmaltz , _ phys .
rev . _ * d61 * ( 2000 ) 113011
. w. f. chang , i. l. ho and j. n. ng , _ phys .
rev . _ * d66 * 076004 ( 2002 ) . g. c. branco , a .
gouvea , m. n. rebelo , _ phys .
_ * b506 * ( 2001 ) 115 .
w. f. cang , j. n. ng , _ jhep _ * 0212 * ( 2002 ) 077 .
b. lillie , j. l. hewett , _ phys .
* 68 * ( 2003 ) 116002 .
y. grossman and g. perez , _ phys .
rev . _ * d67 * ( 2003 ) 015011 ; hep - ph/0303243 ( 2003 ) .
a. delgado , a. pomarol , m. quiros , _ jhep _ * 0001 * ( 2000 ) 030 ; y. grossman , _ int .
_ * a15 * ( 2000 ) 2419 ; d. e. kaplan and t. m. tait , _ jhep _ * 0111 * ( 2001 ) 051 ; g. barenboim , _ et .
_ , _ phys .
_ * d64 * ( 2001 ) 073005 ; w. f. chang , i. l. ho and j. n. ng , _ phys .
_ * d66 * ( 2002 ) 076004 ; w. skiba and d. smith , _ phys . rev .
_ * d65 * ( 2002 ) 095002 ; y. grossman , r. harnik , g. perez , m. d. schwartz and z. surujon , hep - ph/0407260 ( 2004 ) ; p. dey , g. bhattacharya , _ phys .
_ * d70 * ( 2004 ) 116012 ; y. nagatani , g. perez , hep - ph/0401070 ( 2004 ) ; r. harnik , g. perez , m. d. schwartz , y. shirman , hep - ph/0411132 ( 2004 ) .
e. iltan , h. sundu , _
acta phys.slov._ * 53 * ( 2003 ) 17 .
m. deile , et .
al . , muon g-2 collaboration , , _ nucl.phys.proc.suppl .
_ * 116 * ( 2003 ) 215 .
|
we predict the charged lepton electric dipole moments in the split fermion scenario in the framework of the two higgs doublet model .
we observe that the numerical value of the muon ( tau ) electric dipole moment is at the order of the magnitude of @xmath0 ( @xmath1 ) and there is an enhancement in the case of two extra dimensions , especially for the tau lepton electric dipole moment .
|
quantum memory is considered as one of the major challenge for quantum communication and a key element for quantum computing / simulation schemes based on linear optics during a past decade @xcite .
the main objective of quantum memory is to write , store and read out on demand a light pulse without destroying its quantum state . to implement the quantum memory ,
many different protocols were proposed ( see reviews @xcite ) , based on these a number of key experimental advances have been achieved ( see for instance @xcite ) .
recently the interests of community shift to the protocols ensured not only high efficiency of quantum state storage and retrieval , but also fit for storage of the signals with many degrees of freedom @xcite .
broadband or spatially multimode schemes provide this multiplexing . addressing to quantum protocols of long - distant communication , for which memory cells are the key tool ,
it is clear that the information capacity of the channel strongly depends on its spectral bandwidth , so that the cell memory as a spectral filter should satisfy some requirements . on the other hand ,
the involvement of transverse spatial degrees of freedom of the cell and `` parallelization '' of information stream is also an important resource of spatial multiplexing . assessing the benefits of quantum memories for production of multiphoton states
, authors @xcite have introduced an parameter @xmath2 , where @xmath3 and @xmath4 are the efficiency and time - bandwidth product of the memories , respectively ( and @xmath4 is the product of the acceptance bandwidth of the memories and their coherence lifetime ) .
thus , in addition to the high efficiency , we have to be concerned about the spectral properties of the proposed memory protocols . here
we analyze and compare in details the efficiency along with the spectral bandwidths of two memory protocols .
well known that if you consider off - resonant interaction between atomic ensemble ( prepared in the ground state ) and two fields - signal and driving , than the interaction hamiltonian for such a system , after elimination of the excited state , can be cast in the form @xmath5 in quantum optics this hamiltonian is well - known as the beamsplitter hamiltonian .
the interface mixes the input atom and light states as a `` beamsplitter '' and the `` reflection coefficient '' of unity corresponds to a perfect state swapping between light and atoms .
however , this ideology not always may be applicable .
if for some reason the upper excited state can not be excluded from consideration , such schemes are not beamsplitter - like , and therefore results proved for beamsplitter - like schemes can not be used to them without further justification . moreover , even for the off - resonant models of memory , when one takes into account the spatial aspects of light - matter interaction and considers the propagation of the fields in thick atomic layer , the direct beamsplitter analogy is also inapplicable .
we consider here two quantum memory protocols based on the resonant interaction between an atomic ensemble and fields - signal and driving .
the schemes differ in duration of interaction that leads to differences in formation of the atomic coherence , on which state of the signal field is mapped .
one of our main issues here is to analyze the quantum efficiency .
we want to find out whether this characteristic of memory is as universal for non beamsplitter - like protocols as for beamsplitter - like ones , i.e. could we predict how well a certain quantum state will be survived in the memory , if we know only the efficiency of this memory . to demonstrate our arguments
, we consider the storage of squeezing in the schemes of adiabatic and high - speed quantum memory .
thereto we analyze how the squeezed light from a particular source with the desired properties is mapped on the atomic ensemble and then read out . squeezing storage
was analyzed in @xcite , but as opposed to our consideration the authors there do not touch any spatial aspects of interaction and solved the problem for the flat input spectrum . we will also solve the eigenfunction problem for two considered memory protocols .
a similar problem was solved for raman - type protocols ( @xcite ) , when the transformation is unitary .
based on the analysis of eigenfunctions , we compare the spectral bandwidth of the memories .
bs an ideal quantum memory ( qm ) maps the quantum state of a light pulse into a long lived coherence in the atomic ground state using a 2-photon transition , with the help of a strong classical control field .
the signal retrieval , obtained through the interaction of the atomic coherence with the control field should provide a light pulse with the same classical and quantum properties as the input pulse .
several criteria have been proposed to measure the quality of this quantum memory .
the first one is the efficiency , which measures the ratio of the input and output field intensities and uses classical quantities .
the second one is the fidelity , @xmath6 , with @xmath7 , where @xmath8 is the initial wave vector ( pure state ) of the light , and @xmath9 is the density matrix describing the state of light after retrieval .
the fidelity is a measure of the conservation of the quantum features of the signal , and allows characterizing the memory as compared to a classical memory , for which the fidelity is limited to some maximal value .
however , the fidelity depends on the stored state and can not be considered as a characterization of the memory device itself .
a third benchmark the `` tv criterion '' was introduced in ref .
@xcite and involves a classical feature , the transmission ( efficiency ) but also quantum characterization with the conditional variance v , and allows to establish quantum bounds , while evaluating the classical characteristics . in this paper
, we will be interested in both the classical and the quantum benchmarks and we will consider the specific example of squeezed light storage , where several factors can impact the retrieved field . an intuitive understanding of the connection between the storage of squeezing and the efficiency can be gained from the `` beamsplitter '' model of a quantum memory , where the a full writing and reading cycle can be modelled by the transmission through a beamsplitter . indeed , let us write the relation between input signal @xmath10 and output signal @xmath11 in the form of the beamsplitter transformation : & & a_out ( t)= a_in ( t)- a_vac ( t).eff1 here , the `` transmission '' coefficient @xmath12 models the efficiency of the quantum memory process .
let us rewrite this in the fourier domain : & & a_out,= a_in,- a_vac , .
eff2 going from the field amplitudes @xmath13 to the fluctuations of the quadrature components @xmath14 and @xmath15 we get & & x_out , x_out,-= t x_in , x_in,-+ ( 1-t ) x_vac , x_vac,- .
eff3 introducing the squeezing parameters for the input field @xmath16 and for the output field @xmath17 , we obtain & & x_out , x_out,-=e^-r_out(),x_in , x_in,-=e^-r_in().eff4 for the vacuum field amplitude & & x_vac , x_vac,-=.eff5 as a result we get the equation & & 1-e^-r_out()[email protected] we can conclude that the efficiency allows a full characterization of the quantum memory for squeezing transfer if the beamsplitter model is applicable . if the efficiency is close to unity , then the light squeezing is well preserved in the read - out pulse . in this article , we will consider two cases of resonant quantum memory , namely the adiabatic qm @xcite and high - speed qm @xcite .
they are based on the ensemble of three - level atoms in @xmath19-configuration ( see fig .
1 ) . and
signal field @xmath20.,height=132 ] comparing with the non - resonant schemes , here the population of the upper level is non negligible , which leads to losses in the signal field during the light - matter interaction as well as during the storage time . as a result , the analogy between quantum memory and beamsplitter vanishes , and the ratio ( [ eff6 ] ) can be not satisfied .
nca in this paper , we consider an ensemble of three - level atoms in a @xmath21-configuration ( fig .
1 ) that will be used to store temporal and spatial multimode quantum fields .
the atoms resonantly interact with two electromagnetic fields , a signal field @xmath22 and a driving field @xmath23 , that connect the two atomic ground states to the excited state .
the driving field is a strong classical field propagating as a plane wave , while the signal field is a weak quantum field with a transverse structure . during the writing process
both fields interact with the medium simultaneously .
the dynamics of such a system is described by a set of equations for connecting the signal field amplitude @xmath24 that is going to be recorded in the atomic medium , operator @xmath25 , associated with the atomic coherence on the transition @xmath26 , which arises in the writing process , and operator @xmath27 , associated with an atomic coherence on the transition @xmath28 @xcite : & & a(z , t;q)=- gn c(z , t;q), 1 + & & c(z , t;q)=-c(z ,
t;q)+ gn a(z , t;q ) + b(z , t;q)+(z , t;q),2 + & & b(z , t;q)= - ^c(z , t;q)3 , here @xmath29 is a mean density of atoms , @xmath30 is the coupling constant between atoms and field on the @xmath31 is the spontaneous relaxation rate of upper level .
these equations are written in the fourier domain relative to the transverse co - ordinates ( @xmath32 and @xmath33 ) .
the langevin source @xmath34 is a due to the decay of optical atomic coherence and is assumed to fulfill the following conditions ( derived in the same approximations as the equations ( [ 1])-([3 ] ) ) : & & ( z , t;q)=0,(z , t;q)^(z^,t^;q^)=(t - t^)(z - z^)^2(q - q^),^(z , t;q)(z^,t^;q^)=0.4 amplitudes @xmath24 , @xmath25 , and @xmath27 have to obey the canonical commutation relations : & & @xmath35=(t - t^)^2(q - q^ ) , + & & @xmath36=@xmath37=(z - z^)^2(q - q^ ) .
we will compare the dynamics of the atomic and field variables in two limits : first , when the signal and driving field assumed to be very short pulses , much shorter than the excited state lifetime @xmath38 , so that we can neglect the spontaneous emission during the writing process ( high - speed quantum memory ) , and , vice versa , when @xmath39 and we can neglect the time derivative of the optical coherence @xmath27 ( adiabatic quantum memory ) . nca_ad the adiabatic approximation can be applied if the evolution time of the system greatly exceeds the life - time of atomic state @xmath40 . in this case
we can set @xmath41 in ( [ 2 ] )
. then equations ( [ 1])-([3 ] ) can be written as @xcite : & & a(z , t;q)=- c_1 a(z , t;q ) -c b(z , t;q)-2g(z , t;q ) , ad1 + & & b(z , t;q)= -c_2 b(z , t;q)- c^a(z , t;q)-2^/(z , t;q ) ad2 , in eqs . ( [ ad1])-([ad2 ] )
we have introduced the following notation for the constants : & & c_1=,c_2=,c=@xmath42,=||e^i.ad3 for simplicity , in the following we will assume @xmath43 .
let us now introduce dimensionless variables for space and time : & & z = c_1z , t = c_2t.ad4 as shown in @xcite using laplace transformation , one can get a general solution of this system in the form & & ( z , t;q)= _ 0^t dt^(0,t^;q)g_aa(z , t - t^)-p_0^z dz^(z^,0;q)g_ba(z - z^,t)+d_a(z , t;q),ad5 + & & ( z , t;q)= _
0^z dz^(z^,0;q ) g_bb(z - z^,t)-_0^t dt^ ( 0 , t^;q)g_ab(z , t - t^)+d_b(z , t;q).ad6 + & & p==.ad7 the kernels are expressed with the modified bessel functions @xmath44 and @xmath45 as & & g_aa(z , t)= @xmath46(t),ad8 + & & g_bb(z , t)= @xmath47(t),ad9 + & & g_ba(z , t)=g_ab ( z , t)=e^- t- z i_0@xmath48(t).ad10 here @xmath49 is a window function : @xmath50 for @xmath51 and equals zero otherwise , where @xmath52 is the interaction time ( @xmath53 for the writing stage and @xmath54 for the read out ) . the operators @xmath55 and @xmath56 are functions of of the langevine source operator @xmath57 : & & d_a(z , t;q)=_0^t_0^z dt^dz^(z^,t^;q)@xmath58 , + & & d_b(z , t;q)=_0^t_0^z dt^dz^(z^,t^;q)@xmath59 .
it is important to note that the noise operators @xmath60 and @xmath61 are independent of the initial conditions .
the solutions given by eqs .
( [ ad5 ] ) and ( [ ad6 ] ) can be first used to model the writing process . for this
we should take into account the initial conditions for operator @xmath24 : @xmath62 , which means that a signal field in some specific quantum state enters the cell at @xmath63 . before the writing process ,
all the atoms are in the lower level @xmath64 , and the initial state of the coherence operator @xmath65 is the vacuum state .
equations ( [ ad5 ] ) and ( [ ad6 ] ) can then be written as & & ^w(z , t;q)= _ 0^t dt^_in(t^;q)g_aa(z , t - t^)-p_0^z dz^^w(z^,0;q)g_ba(z - z^,t)+d_a(z , t;q),ad11 + & & ^w(z , t;q)= _ 0^z
dz^^w(z^,0;q ) g_bb(z - z^,t)- _ 0^t dt^_in(t^;q)g_ab(z , t - t^)+ d_b(z , t;q).ad12 in eq .
( [ ad12 ] ) , the second term is the memory term ; it shows the creation of the atomic coherence @xmath66 due to the interaction with the input signal @xmath67 and driving field @xmath68 , and corresponds to the storage of the input field in the atomic coherence , while the first term corresponds to the evolution of the initial atomic coherence @xmath69 interacting with the driving field and the third term to added noise . for @xmath70 , equation ( [ ad11 ] ) gives the output signal field , the first term corresponds to the leakage of the signal field , the second one to the generation of a field in the mode of the signal field from the interaction of the initial atomic coherence @xmath71 with the driving field and the third one to added noise .
since we use the heisenberg formalism , we deal with operators evolving in time and space and acting on fixed state vector .
the state vector of the full system is determined by the states of subsystems - the signal field , the atomic coherence , and langevin noise source @xmath72 . at the initial time , i.e. before the beginning of the writing procedure , all of these subsystems are independent , and the state vector @xmath73 is factorized : & & |^w=|^w_a|^w_b|^w_= |^w_a|vac_b |vac_,ad13 where @xmath74 is the initially prepared signal field state , @xmath75 is the initial state of the atomic coherence ( vacuum state ) , and @xmath76 is the state of the langevin noise source @xmath77 ( vacuum state ) . at the end of the writing stage ( @xmath78 ) ,
the signal field is vacuum , the driving field is zero , and we assume that the coherence @xmath79 does not decay with time .
turning to the analysis of the reading stage , we use again the heisenberg - langevin equations ( [ ad5 ] ) and ( [ ad6 ] ) , with different initial conditions and different initial state vector .
we assume that there is no input field at the entrance of the cell and the coherence is @xmath80 .
the state vector of the full system before reading is also factorized : & & |^r=|^r_a|^r_b|^r_= |vac_a|^r_b|vac_,ad14 let us derive the signal field operator during the reading process : & & ^r(z , t;q)= _
0^t dt^^r(0,t^;q)g_aa(z , t - t^)-p_0^z dz^^w(z^,t_w;q)g_ba(z - z^,t)+d_a(z , t;q).ad15 in this equation , the second term shows to the generation of a signal field through the interaction of the atomic coherence @xmath81 with the control field and corresponds to the memory read - out .
the first term gives the interaction of the input ( vacuum ) field with the atoms and the control field .
the last term corresponds to added noise from the langevine reservoir .
substitute ( [ ad12 ] ) into ( [ ad15 ] ) we obtain : & & ^r(z , t;q)= _ 0^t dt^^r(0,t^;q)g_aa(z , t - t^)+_0^t_w_0^z dt^dz^_in(t^;q)g_ab(z^,t_w - t^)g_ba(z - z^,t ) + & & -p_0^z _ 0^z^ dz^dz^ ^w(z^,0 ; q)g_bb(z^-z^,t_w)g_ba(z - z^,t ) + & & -p_0^z dz^d_b(z^,t_w;q)g_ba(z - z^,t ) + d_a(z , t;q),ad16 now let us make averaging of this expression over the vacuum subsystems . then instead of the canonical amplitude @xmath82 the new one arises : & & ( z , t;q)^r|^w| ^r(z , t;q)|vac^w|vac^r , ad17 where @xmath83 are vacuum state vectors before writing and @xmath84 the _ non - canonical amplitude _ ( nca ) , because the corresponding commutation relations have acquired the non - canonical form .
let us emphasize that we do not average over the complete wave function , but only over those subsystems that are in the vacuum state . in this case
, the quantum nature of the input signal is taken into account , and we can keep some of the quantum properties of the output signal field . as a result ,
all the terms except the second one in ( [ ad16 ] ) turn out equal to 0 : & & ^r(l , t;q)= _ 0^t_w
dt^_in(t_w - t^;q)g(t , t^).ad20 we thus cast an integral transformation of the input field operator into the operator of the output field retrieved in the reading process .
for the forward read - out ( when the reading driving field propagates in the same direction as the writing driving field ) , the kernel @xmath85 reads : & & g(t , t^)=_0^l dz^g_ab(z^,t^)g_ba(l - z^,t).ad21 for the backward read - out ( when these two driving fields propagate in opposite directions ) the transformation keeps the same form ( [ ad20 ] ) with the kernel & & g(t , t^)=_0^l dz^g_ab(z^,t^)g_ba(z^,t ) , ad22 where @xmath86 and @xmath87 are given by eq .
( [ ad10 ] ) although the introduced partially averaged operators do not convey all the quantum properties of the field ( because , for example , they do not obey canonical commutation relations that can be easily seen from comparison of the canonical amplitude ( [ ad16 ] ) with non - canonical one ( [ ad20 ] ) ) , they will be very useful in the following .
for example , for the @xmath88-quadrature of the retrieved field @xmath89 , & & : x_a(z , t;q)x_a(z^ ,
t^;q):= : x_a(z , t;q)x_a(z^ , t^;q ) : , a^r = x_a+i y_a , a^r = x_a+i y_a , where notation @xmath90 stands for normal ordering of the operators . in the high - speed model of quantum memory
the signal and driving pulses are assumed to be much shorter than the excited state lifetime @xmath91 , so that we can neglect spontaneous emission during the interaction processes , which eliminates a source of fluctuations and dissipation in ( [ 2 ] ) .
then the equations for signal field amplitude @xmath92 , atomic coherence between low levels @xmath93 , and optical coherence between levels @xmath64 and @xmath40 @xmath94 read @xcite : & & a(z , t;q)=- gn c(z , t;q),hs24 + & & c(z , t;q)= gn a(z , t;q ) + b(z , t;q),hs25 + & & b(z , t;q)= - ^c(z , t;q)hs26 , the notation here is the same as in ( [ 1])-([3 ] ) .
( [ hs24])-([hs26 ] ) can be solved in the general form by using the laplace transform .
details of the procedures are discussed in appendix a. as a result , applying the same technique of averaging over the vacuum subsystems , we can derive the solution for retrieval field formally in the same form as for adiabatic case : & & ^r(l , t;q)= _ 0^t_w dt^_in(t_w - t^;q)g(t , t^),hs20 here the kernel @xmath85 is given by & & g(t , t^)=_0^l dz^g_ab(z^,t^ ) g_ba(l - z^,t)hs21 + & & g(t , t^)=_0^l dz^g_ab(z^,t^)g_ba(z^,t),hs22 and functions @xmath86 and @xmath87 now read & & g_ab(z , t)=g_ba(z , t)=_0^tdt^g_ab(z , t - t^,)g_ab^(z , t^),hs27 + & & g_ab(z , t)=e^-i t j_0@xmath95 ( t),hs28 where @xmath96 is the bessel functions of the first kind of the zero order . here and below when we discuss high - speed qm we use dimensionless variables @xmath97 and @xmath98 given by & & t = t , z= 2g^2nz/.hs29
in this section we want to compare the results of two mental experiments obtained for both adiabatic and high - speed memory models : an experiment to study the efficiency of qm and experiment for the storage of squeezed light .
as shown in section [ bs ] the results of such experiments are uniquely related for the case of the beamsplitter - type interaction
. we will consider whether they related in some way in our models .
the study of the quantum efficiency of the qm implies the following scheme : signal and control pulses with some defined profiles both are incident on the input face of the atomic cell . propagating through the cell ,
these fields generate a coherence between the two lower levels with a certain spatial distribution , which depends both on the nature of light - matter interaction and on the duration of this interaction .
after some storage time , the retrieved control pulse is sent to the input or output cell face and triggers the emission at the signal frequency , which is collected by a photodetector .
one compares the total photon numbers in reconstructed signal and in the input pulse .
analytically , the quantum efficiency can be read as the following ratio of the averages : e_q=_0^t_r dta^_out(t;q)a_out(t;q)@xmath99 ^ -1 .ex1 then taking into account ( [ ad20 ] ) , one can obtain for the arbitrary input pulse & & e_q= 4.2 + & & = @xmath100@xmath101 ^ -1.choosing a rectangular signal profile , we have to neglect the time dependencies of the input correlation functions & & _ in^(t_w - t^;q)_in ( t_w - t^;q)=_in^(t;q)_in(t;q)=_in^(q)_in(q ) .
then the formula for the efficiency turns out to be essentially simpler & & e_q= _ 0^t_rdt@xmath102 ^ 2.ex2 one can see , for different components of @xmath103 the efficiencies are equal . in the second case , we assume that the input pulse is squeezed .
let us assume that pulsed light with given quantum properties is obtained from a steady source of squeezed light in which a pulse has been cut .
it is this kind of light source that was used in the experiments on quantum memory for squeezed states @xcite . here , we consider two examples of pulsed quantum light generation when the single pulse is obtained by cutting the time of radiation 1 ) of single - mode phase - locked sub - poissonian laser , which emits a squeezed light in a close to coherent state @xcite , and 2 ) of the degenerate optical parametric oscillator ( dopo ) , operating in the subthreshold regime @xcite . to assess the conservation of the quantum light properties , we will compare the quantum properties of the read - out pulse with the original one .
let us introduce the _ degree of light squeezing _
@xmath104 and @xmath105 for the input and output pulses , respectively : & & s_out,,q = e^-r_out(,q),_in,,q= e^-r_in(,q),=/.ex4 we assess the ratio of these parameters to compare squeezing in the output pulse with squeezing in the input one . taking into account eq .
( [ eff4 ] ) this ratio can be derived in the form & & = .4.4 according to ( [ hs20 ] ) input and output correlation functions in time domain are connected with the expression : & & : x^r(l , t , q)x^r(l , t^,q^):=_0^t_w dt_1 _ 0^t_w dt_2:x_in(t_w - t_1,q)x_in(t_w - t_2,q^):g(t , t_1)g(t^,t_2 ) . passing to the fourier domain
one can obtain & & : x_out , , qx_out , ^,q^:=_0^t_w dt_1 _ 0^t_w dt_2:x_in(t_w - t_1,q)x_in(t_w - t_2,q^):g(,t_1)g(^,t_2),4.6 where the fourier transformation is given by & & x_out,,q=_0^t_rdt x^r(l , t , q)e^it .
as for the input correlation function @xmath106 in the denominator of ( [ 4.4 ] ) , one can find it in the explicit form in appendix b for two squeezed light sources - the sub - poissonian phase - locked laser and degenerated optical parametric oscillator . before we start to compare the input and output squeezing , it should be noted that the squeezing degree in the input pulse is already less than in the steady beam @xcite .
this fact is important , because the mechanism of qm depends essentially on the pulsed nature of the recording light , and we can not neglect by this feature .
however , in @xcite it has been shown that if the condition @xmath107 for the pulse duration @xmath108 and the spectral width of the light @xmath109 is fulfilled then the pulse squeezing degree does not depend on its duration and equals to one in stationary radiation .
we will assume that this condition is verified .
all the necessary details specifying the input radiation parameters one can find in appendix b. graphically , the input pulse spectrum is shown in fig .
[ fig2 ] . .
, height=188 ] bellow we are interested in squeezing of the pulse as a whole that corresponds to the zero frequency component in the spectrum of squeezing @xmath110 , i.e. , in value @xmath111 , it characterizes the maximum achievable degree of squeezing .
it should be noted that the properties of input radiation in two above - described mental experiments are quite different .
in fact , studying the efficiency we are restricted by the analysis of the essentially narrow - band light , because for given profile of the input it can be characterized by a single time - independent amplitude . that is , one is discussing the storage for single mode ( rectangular in our description ) .
input - output transformation converts the profile , resulting in the other field profile at the output , but input and output profiles are connected by unique integral transformation and the single mode at the input is converted into a single ( different from the original ) mode at the output .
when we study the degree of squeezing we are dealing with a broadband signal : the input field can not be characterized by the single amplitude @xmath67 , but is a function of time .
this is the time dependence is the subject of the study in this experiment .
the question of the legitimacy for comparing the results of two such a different mental experiments , we will discuss below . fig . [ fig3 ] depicts the results of the numerical calculations of the efficiency @xmath112 and pulse squeezing @xmath111 ( more precisely , @xmath113 ) as a function of read - out time @xmath114 for forward retrieval in two models : high - speed qm ( fig . [ fig3]a ) and adiabatic qm ( fig . [
fig3]b ) . in two models : high - speed qm (
a ) and adiabatic qm ( b ) for forward retrieval.,height=188 ] the parameters for numerical calculation have been chosen as follows . in the scheme of high - speed qm for given dimensionless length of the medium @xmath115
, we found the duration of the input pulse , resulting in the lowest signal losses in writing ( see @xcite ) , @xmath116 meets this requirement .
the efficiency in fig .
[ fig3]a is plotted for these parameters . to calculate the pulse squeezing the same values of @xmath117 and @xmath118 have been chosen , and in addition , we used the input field parameters @xmath119 , to provide knowingly a good squeezing of the input pulse ( below we will demonstrate how to optimize the choice of the spectral width of the input signal ) .
for the calculation in the adiabatic limit we have chosen @xmath120 , providing approximately the same efficiency as for the high - speed qm .
the value of @xmath121 is chosen to satisfy the condition @xmath122 .
plot in fig .
[ fig3]b corresponds to @xmath123 , but it should be noted that the increase of @xmath109 has no effect on the behavior of the curves .
first , let us note that the efficiency and pulse squeezing in fig .
[ fig3 ] do not always coincide .
the efficiency increases monotonically and comes to saturation , while the pulse squeezing decreases after a certain time @xmath124 .
such a behavior of the curves is a clear demonstration of the non - beamsplitter type of interaction in these models .
this result is in line with our knowledge of the role of noises in these models .
it is not hard to believe that we have a high efficiency of qm , but lose the squeezing of radiation due to the contribution of vacuum noise when the read - out go on too long time . on the other hand ,
the figure shows the range of values @xmath114 where pulse squeezing higher than the efficiency .
the same situation takes place for curves in the case of backward retrieval ( see fig.[fig4 ] ) . in two models :
high - speed qm ( a ) and adiabatic qm ( b ) for backward retrieval.,height=188 ] on the face of it , these curves seem paradoxical : they appeared a range of values @xmath114 such that under sufficiently low efficiency ( 50% only ) , we stored a good squeezing in the output .
could it be ?
a clue to this paradox lies in the multimode nature of the interaction . in the next section
we discuss the mode properties in both adiabatic and high - speed qm models .
it is not difficult to see that for both considered here models of memory the kernels @xmath125 in eqs .
( [ ad20 ] ) and ( [ hs20 ] ) , which couple the input and output signals , are symmetrical with respect to permutation of the arguments @xmath126 .
this means we have a right to derive the equation for the eigenfunctions @xmath127 and eigenvalues @xmath128 of these kernels in the form & & _ i(t)=_0^t_wdt^_i(t^)g(t , t^).5.1 here the functions @xmath127 form the complete ortonormal set : & & _ 0^t_wdt^_i(t)_j(t)=_ij,_i^_i(t)_i(t^)=(t - t^).5.2 the equation ( [ 5.1 ] ) can be solved for each of the four kernels of the integral transformations ( [ ad20 ] ) and ( [ hs20 ] ) . in this article
we will analyze the eigenfunctions and eigenvalues only for backward read - out , as this is the case for both models when the `` abnormally '' high pulse squeezing at the relatively low efficiency appears .
we will carry out the numerical calculations at the same values of parameters as before ( in fig .
[ fig3 ] and [ fig4 ] ) .
first , let us note that the eigenvalues in both cases decrease rapidly ( see fig .
[ fig_ev ] ) . especially sharp decrease of eigenvalues
is observed for the model of high - speed qm , where , in fact , only the first two eigenvalues differ from zero .
this behavior shows that this scheme is a good filter for the first two eigenmodes .
figures [ fig5 ] and [ fig6 ] shows the first three eigenmodes ( the first row ) for the high - speed qm and adiabatic qm models , correspondingly , as well as the squares of these functions ( the second row ) .
, dimensionless frequency is @xmath129.,height=377 ] , dimensionless frequency is @xmath130.,height=377 ] the distinguishing feature of the high - speed qm model is not only the filtering in this scheme of only two modes , but also the localization of these modes in different intervals of the timeline ( the first mode is actually localized in the range of @xmath131 $ ] , and the second one - in @xmath132 $ ] ) .
draw attention that the boundaries of intervals coincide with the locations of two peaks in the curve in fig .
[ fig4]a . for the adiabatic memory such a mode `` separation ''
is not observed .
now let us see how squeezing is distributed over the modes , that means we want to follow whether all eigenmodes of the cell are squeezed .
for this aim we plotted the spectra of the first three modes ( see the third row in figs .
[ fig5 ] and [ fig6 ] ) .
recall that the input field has the wide spectrum of squeezing ( see fig .
[ fig2 ] ) .
one can see that all of our spectra of eigenmodes are concentrated in the area of good squeezing , that means , they are well squeezed . however , not all of them contribute to the squeezing of pulse as whole ( squeezing at the zero frequency ) .
going back to the curves in fig .
[ fig4]a , we see that at time @xmath133 the first mode of the field is mostly read - out , the second mode is localized on the other time interval , and the rest is filtered out by the cell .
the zero spectral component of the first mode is high and squeezed well ( that is the first mode as whole is squeezed well ) .
thus , although the half part of the photons still is not read - out - they will be retrieved from the second mode , one can obtain a high squeezing in the retrieved pulse .
it should be noted that the two qm models are very different spectrally . to estimate the typical spectral widths for eigenmodes in both cases
, we should revert from the dimensionless variables to dimensional ones . recall that the dimensionless procedures for the models are different . for time we got the relations ( [ hs29 ] ) and ( [ ad4 ] ) , respectively : _
hs t t_hs , t t_ad , where we now added the below indexes _ `` hs '' _ and _ `` ad '' _ to indicate the high - speed or adiabatic model , respectively .
besides , the requirements imposed on the rabi frequency , are opposite : _
hs,_ad , whence it follows @xmath134 , and the ratio of this frequencies has to be at least 2 orders . based on these inequalities and noting that the ratio of dimensionless spectral widths in figs .
[ fig5 ] and [ fig6 ] is @xmath135 , we find that the width of eigenmodes in the adiabatic model is at least 4 orders of magnitude smaller than the corresponding width in the high - speed model . at the same time , one can see that although the dimensionless values of @xmath136 are different , we can assume that both calculations were carried out at the same optical depth @xmath137 ( supposing the relation @xmath138 to be fulfilled ) .
let us note that the relation between the parameters typical for the high - speed qm , can be easily reached in experiments with resonant @xmath21-atoms , especially for cold atomic ensemble , than for the adiabatic condition .
figures [ fig5 ] and [ fig6 ] shows that the spectral bandwidths of memories are much less than the chosen spectral width of the input signal .
we can say that this choice was not optimal , and it is necessary to match the input signal to the memory capacity . in this sense , the model of high - speed qm looks more attractive because it provides a much wider spectral bandwidth of the signal as mentioned above .
the decrease in the value of the spectral width @xmath109 , while maintaining the duration of the input pulse results in a loss of input pulse squeezing ( see fig .
[ fig8 ] ) .
however , it is seen that for @xmath139 for the high - speed qm and for @xmath140 for the adiabatic qm these losses are relatively small ( about 10% ) , and the ratio @xmath141 is still satisfied . in fig .
[ fig8 ] the input and retrieval spectra of squeezing are plotted for the case when spectral bandwidth of squeezing of the input signal is matched to the memory bandwidth .
since considered memory cells are the mode filters , i.e. they are sensitive to only a few the first modes of input signal , then the presence of other modes at the input does not affect to the pulse squeezing . from the fig .
[ fig5 ] one can see that the zero spectral component is large enough only in the spectra of the first and second eigenmodes .
at the same time , the square input pulse is well approximated by the superposition of the first two modes .
that is why the comparison of the two mental experiments ( estimation of the efficiency and of the storage of squeezing ) with the different input signals seems to us justified .
for example , in the scheme of high - speed qm if knowing the properties of our cell as a mode filter , we form an input signal as a superposition of the first two modes , then the curves in fig .
[ fig4 ] remain almost unchanged .
in raman memory models based on three - level atoms , two quantum sub - systems ( the signal field and atomic coherence ) are interacted with each other .
the writing and read - out processes are formally realized by means of the interaction hamiltonian of the kind of @xmath142 .
this hamiltonian ensures the beamsplitter - like interaction for the thin atomic layer and as a result one can derive the input - output equation ( [ eff1 ] ) and conclude that squeezing and efficiency are coupled by the formula ( [ eff6 ] ) .
this means that the efficiency turns out to be sufficient to assess the quality of memory @xcite . in this article
we consider more complicated situation , which takes place in the cases of adiabatical and high - speed memories in the strongly resonant limit . in each of these models
there are three quantum sub - systems : the signal field and two atomic coherences for the high - speed memory and the signal field , atomic coherence , and the vacuum reservoir , forming the langevine source , for the adiabatic memory .
the presence of more than two active subsystems makes impossible direct interpretation of interaction as a beamsplitter - like one .
however maybe we still can apply such an interpretation in some sense ?
let s consider again the equation ( [ ad22 ] ) coupling the signal field at the output of the memory cell after the read - out with the signal at the front of the cell before the writing process .
let s rewrite it in the form of the non - averaged over vacuum sub - systems & & a_out ( t , q)=_0^t_w g(t , t^)a_in(t_w - t^,q)dt^+f_vac ( t , q).6.1 remind the indexes `` in '' and `` out '' mean @xmath143 and @xmath144 , respectively .
the second term on the right @xmath145 concentrates contributions of all sub - systems , which initially were in the vacuum state .
it is not difficult to see that the input - output amplitudes obey the correct commutation relations & & @xmath146=@xmath147=(t_1-t_2)^2(q_1-q_2). provided that & & @xmath148=@xmath149 ^2(q_1-q_2 ) 6.3 . as mentioned above the kernel @xmath85 is symmetrical and we have a right to derive eq .
( [ 5.1 ] ) for its eigenfunctions , which obey the requirements ( [ 5.2 ] ) .
let us decompose input and output amplitudes in the form & & a_in(t_w - t , q)=_ie_in;i , q_i ( t),a_out(t , q)=_ie_out;i , q_i ( t),6.6 where & & @xmath150=@xmath151=_ii^^2(q - q^ ) . the similar decomposition can be used for function @xmath152 : & & f_vac(t , q)=_if_vac;i , q_i ( t),f_vac;i , q=_0^t_w dt_i ( t)f_vac(t , q ) .
6.7 taking into account the equation ( [ 6.3 ] ) one can obtain & & @xmath153=(1-_i)_ii^^2(q - q^ ) or making normalization & & @xmath154=_ii^^2(q - q^),f_vac;i , q = e_vac;i , q .
6.8 substituting ( [ 6.6 ] ) , ( [ 6.7 ] ) and ( [ 6.8 ] ) into ( [ 6.1 ] ) one can obtain for each of modes & & e_out;i , q= e_in;i , q - e_vac;i , q.6.11 thus , in this sense , we can consider our model as a beamsplitter - like ones .
however such a classification associated with choice of the input signal as one ( and only one ) of the memory eigenfunction .
the question arises whether we can correctly determine the efficiency for an arbitrary input signal and whether it will be as comprehensive characteristic of the memory as the efficiency associated with the transfer of eigenmode .
if we consider the arbitrary input signal as a single mode from some complete orthonormal set of functions @xmath155 , then the efficiency can be expressed analytically . in this case
, one can obtain from the formula ( [ 4.2 ] ) & & e= _ 0^t_r dt @xmath156 ^ 2.e4 to assess how squeezing is changed under the full memory process , we can use ratio ( [ 4.4 ] ) putting there @xmath157 and @xmath158 & & e_squuzing=. according ( [ 4.6 ] ) the expression in numerator is given by & & : x_out , = 0,q=0x_out , -=0,-q=0:= + & & = _ 0^t_w
dt_1 _ 0^t_w dt_2:x_in(t_w - t_1,q=0)x_in(t_w - t_2,-q=0):g(=0,t_1)g(-=0,t_2 ) . now taking into account that & & : x_in , = 0,q=0x_in , -=0,-q=0:=_0^t_w dt_1 _ 0^t_w dt_2:x_in(t_1,q=0)x_in(t_2,-q=0 ) : . and
@xmath159 one can obtain & & e_squeezing=^2.e6 it is clear that the expressions ( [ e4 ] ) and ( [ e6 ] ) coincide only if the profile @xmath160 is an eigenfunction of the kernel @xmath161 .
thus , the efficiency determined for a single non - eigenmode provides an answer to the question about the ability of the memory to store the total number of photons , but does not characterize the preservation of other quantum properties of light , such as squeezing .
beamsplitter relation can be generalized in the form @xcite : & & r= s - e_vac , 6.15 where operator @xmath162 is a quantum amplitude of single input mode of the signal , and @xmath163 is an amplitude of different mode of output signal , which can be defined as a projection of input mode on the full set of the eigenmodes @xmath164 of the kernel of integral transformation ( [ 6.1 ] ) .
it is easy to see that this equation turns into eq .
( [ 6.11 ] ) when the input field is chosen in the form of one of the modes @xmath165 .
( [ 6.15 ] ) can be used for comparison of the degree of squeezing in the input and output modes .
however it does not provide us with any information about the ratio of squeezing on given ( for example , zero ) frequency that is the most important parameter for multimode squeezing .
we should note that eq .
( [ 6.15 ] ) is valid only when writing time is equal to reading time .
moreover , it links two single modes - one mode at the input and one mode at the output - and is inapplicable for multimode stochastic fields considered here .
such a generalization can be a subject for further study .
addressing once again to the comparison of two memory models presented in this paper , we note that different factors lead to a vanishing of the role of the langevin noises in this cases . in high - speed qm model
the noises can be excluded from consideration in very beginning , when we construct equations , since on the typical interaction time the spontaneous decay of the excited state is negligible , and hence the noise associated with this decay is also small .
the situation is different in adiabatic qm model . here , the rate of the spontaneous relaxation is large that yields to the presence of langevin noise sources , which associated with population decay of the excited state as well as with decay of the coherences on the transitions coupled with the excited state .
however , due to the initial assumption that the population of level @xmath166 is small ( and almost all atoms remain on the level @xmath167 ) the additional noise terms turn out to be small . in this study , we demonstrated that both memory models considered here allow us to achieve high efficiency for writing and restore of the light pulse , and are suitable for storage of the squeezed states .
we have shown that the quantum efficiency of whole writing - reading cycle for a pulse with arbitrary time profile is not a universal characteristic of the memory , through which the other properties of the memory ( for example , the ability to store a squeezing ) can be expressed .
only the efficiency defined in relation to a single eigenmode of the memory is such a comprehensive parameter .
thus for a pulse with an arbitrary shape it is not enough to provide the high efficiency of the memory to conclude about preserving the quantum state of light .
we have analyzed the eigenmodes of the two memory models discussed here and demonstrated interesting feature connected with the ability to manipulate separately by two first ( high - efficient ) eigenmodes in the high - speed qm scheme .
since the quantum memory operates as a mode filter , its important characteristic is a spectral bandwidth . estimating the spectral bandwidth of the adiabatic and high - speed qm schemes
, we have shown that it is significantly different for these schemes .
the spectral bandwidth of the high - speed qm is at least 4 orders of magnitude higher than of the adiabatic qm .
this demonstrate the benefits of the high - speed memory for storage of wideband signals .
we thank elisabeth giacobino and ivan sokolov for productive discussions . the reported study was supported by rfbr ( grants
12 - 02 - 00181a and 13 - 02 - 00254a ) and by the era.net rus project nanoquint .
a the general solutions of the eqs .
( [ hs24])-([hs26 ] ) under the arbitrary initial and boundary conditions can be written over the non - canonical amplitudes , introduced analogously to our consideration in section [ nca_ad ] : & & a^w ( z , t;q)=_0^ td t^a_in ( t- t^;q ) g_aa(z , t^),a1 + & & b^w(z , t;q)=-_0^ td t^a_in ( t- t^;q)g_ab(z , t^),a2 + & & c^w(z , t;q)=_0^ td t^a_in ( t- t^;q)g_ac ( t^ , z),a3 where @xmath168 . & & a^r ( z , t;q)=-_0^ zd z^b^r(z - z^ , 0;q)g_ba(z^ , t ) , b^r(z , 0;q)=b^w(z , t_w;q),a4 + & & b^r(z , t;q)=_0^ zd z^b^r(z - z^ , 0;q)g_bb(z^,t ) , a5 + & & c^r ( z , t;q)=_0^ zd z^b^r(z - z^,0;q)g_bc(z^ , t ) , a6 where upper indexes `` w '' and `` r '' indicate the writing and read - out stage , respectively
. here kernels @xmath169 are bilinear combinations of the expressions depending on the n - th bessel functions of the first kind denoted by @xmath170 & & g_aa(z , t)= ( t)-e^-i t j_1@xmath95 ( t),a7 + & & g_bb(z , t)=e^-i tj_1@xmath95 ( t),a8 + & & g_ab(z , t)=e^-i t j_0@xmath95 ( t).a9 the kernels in eqs .
( [ a1])-([a6 ] ) read & & g_aa(z , t)=_0^tdt^g_aa(z , t - t^ , ) g_aa^(z , t^),a10 + & & g_ab(z , t)=_0^tdt^g_ab(z , t - t^ , ) g_ab^(z , t^),a11 + & & g_ac(z , t)=_0^tdt^g_aa(z , t - t^ , ) g_ab^(z , t^)+c.c.,a12 + & & g_ba(z , t)=g_ab(z , t),a13 + & & g_bb(z , t)=2 ( z)t+ _ 0^tdt^g_bb(z , t - t^,)g_bb^(z , t^),a14 + & & g_bc(z , t)= ( z)t-_0^tdt^g_aa(z , t - t^,)g_ab^(z , t^)+c.c.a15
according to @xcite , the mean - square intracavity quadrature fluctuations of a single - mode laser are determined as & & ( : x_^2:)=-,(:y_^2:)=,=1.16 here , the parameter @xmath171 reflects the degree of ordering of the upper laser level excitation : @xmath172 for a completely random poissonian pumping , and @xmath173 for a completely regular one . here
, @xmath109 is the spectral width of the laser mode , @xmath174 is the average number of photons in the oscillation mode in the steady - state regime of oscillation and @xmath175 is the average number of photons appearing in an empty cavity under the influence of an external locking field .
this locking is necessary for suppression of phase diffusion in the oscillation , which prohibits efficient quadrature squeezing .
parameter @xmath176 determining laser locking to an external field is assumed to be small compared to unity . however , despite small value of @xmath176 , ignoring this quantity would violate the heisenberg uncertainty relation at small frequencies .
therefore , this quantity should be handled carefully .
similarly , we can write the corresponding expressions for the case of a degenerate optical parametric generation @xcite : & & ( : x_^2:)=,(:y_^2:)=-,s=<1.17 here , @xmath177 is the average steady - state number of photons stored in the pump mode , @xmath175 is the threshold number of photons in the pump mode , and @xmath109 is the spectral width of the signal mode . here , we analyze the operation of a dopo in the below - threshold regime and the parameter @xmath178 characterizes the degree of approaching the generation threshold . applying inverse fourier transformation to ( [ 16 ] ) , we have & & : x(t)x(t^):=-e^-(1-/2)|t - t^|,18 + & & : y(t)y(t^):=e^-/2|t - t^|.19 when describing a pulse of light in the quantum theory , it should be remembered that the field outside of the pulse is in a vacuum state , which can be explicitly expressed by writing the extracavity field amplitude in the form & & a(t)^t(t)a(t)+@xmath179a_vac ( t).20 here , the index _
`` vac '' _ in front of the heisenberg operator means that the operator must be averaged over the vacuum state when calculating measurable quantities .
the function @xmath180 is zero outside of the @xmath181 interval and is equal to unity within this interval . to characterize squeezing , along with the field quadratures for the intracavity field @xmath182 and @xmath183 , the quadratures for the field leaving the cavity for free - space propagation & & x(t)=@xmath184,y(t)=@xmath185 .
are introduced .
quadrature squeezing for stationary generation is determined based on the equalities relating the correlator of the quadrature operators with the corresponding average of the normal - ordered operators : & & x(t ) x(t^ ) = ( t - t^)+ : x(t ) x(t^ ) : .7 & & y(t ) y(t^ ) = ( t - t^)+ : y(t ) y(t^ ) : .8 for the isolated pulse equations ( [ 7 ] ) and ( [ 8 ] ) can be written in the form & & ^t(t)x(t ) ^t(t^)x(t^ ) = ^t(t)^t(t^)(t - t^)+ : ^t(t)x(t ) ^t(t^)x(t^ ) : , 21 + & & ^t(t)y(t ) ^t(t^)y(t^ ) = ^t(t)^t(t^)(t - t^)+ : ^t(t)y(t ) ^t(t^)y(t^ ) : .22 let us define the fourier transformation for function @xmath186 in limited time interval @xmath187 $ ] as & & f^t_=_0^t dt f(t)e^it.23 application of this transformation to ( [ 21 ] ) and ( [ 22 ] ) yields & & x^t _
x^t_^ = ^t(+ ^)+:x^t _ x^t_^ : , 24 + & & y^t _ y^t_^ = ^t(+ ^)+:y^t _ y^t_^ : , 25 where we introduced the notation & & ^t(+ ^)= e^/2.26 the second terms in ( [ 24])-([25 ] ) , which are related to calculation of the normally ordered average , are different for different sources . in the case of single - mode phase - locked sub - poissonian laser
, we readily find that & & : x^t _ x^t_^ : = @xmath188,27 where @xmath189 using definition ( [ eff4 ] ) and substituting it into ( [ 24 ] ) and ( [ 27 ] ) , we can express the squeezing parameter for the case of a sub - poissonian laser in the form & & e^-r_in ( ) = 1- + @xmath190.28 for parametric oscillation , the most interesting feature is the phase quadrature . formally acting similar to the laser case , i.e. , limiting analysis to time interval @xmath181 and applying fourier transformation , we have & & y^t
_ y^t_^ = @xmath191^t(+^)+ + & & + @xmath192- + & & + @xmath193.35 the obtained expression allows estimating the degree of squeezing of a pulse of light of arbitrary duration @xmath108 at arbitrary frequency .
100 l. m. duan , m. d. lukin , j. i. cirac , and p. zoller , long - distance quantum communication with atomic ensembles and linear optics , nature , 414(6862 ) , 413 - 418 ( 2001 ) sangouard n , simon c , zhao b et al .
a , 77(6 ) , 062301 ( 2008 ) nicolas sangouard , christoph simon , hugues de riedmatten , and nicolas gisin , quantum repeaters based on atomic ensembles and linear optics , rev . mod . phys .
, 83 , 33 - 80 ( 2011 ) e. knill , r. laflamme , and g. j. milburn , a scheme for efficient quantum computation with linear optics , nature , 409(6816 ) , 46 - 52 ( 2001 ) pieter kok , w. j. munro , kae nemoto , t. c. ralph , jonathan p. dowling , and g. j. milburn , linear optical quantum computing with photonic qubits , rev .
phys . , 79(1 ) , 135 ( 2007 ) f. bussires , n. sangouard , m afzelius , h. de riedmatten , c. simon , and w. tittel , prospective applications of optical quantum memories , arxiv:1306.6904 ( 2013 ) klemens hammerer , anders s. sorensen , and eugene s. polzik , rev .
phys . , 82 , 1041 - 1093 ( 2010 ) alexander i. lvovsky , barry c. sanders , and wolfgang tittel , optical quantum memory , nature photonics , 3(12 ) , 706 - 714 ( 2009 ) chien liu , zachary dutton , cyrus h. behroozi , and lene v. hau , observation of coherent optical information storage in an atomic medium using halted light pulses , nature , 409(6819 ) , 490 - 493 ( 2001 ) brian julsgaard , jacob sherson , j. ignacio cirac , jaromir fiurasek , and eugene s. polzik , experimental demonstration of quantum memory for light , nature , 432(7016 ) , 482 - 486 ( 2004 ) m. d. eisaman , a. andr , f. massou , m. fleischhauer , a. s. zibrov , and m. d. lukin , electromagnetically induced transparency with tunable single - photon pulses , nature , 438(7069 ) , 837 - 841 ( 2005 ) k. s. choi , h. deng , j. laurat , and h. j. kimble , mapping photonic entanglement into and out of a quantum memory , nature , 452(7183 ) , 67 - 71 ( 2008 ) bo zhao , yu - ao chen , xiao - hui bao , thorsten strassel , chih - sung chuu , xian - min jin , jorg schmiedmayer , zhen - sheng yuan , shuai chen , and jian - wei pan , a millisecond quantum memory for scalable quantum networks , nature physics , 5(2 ) , 95 - 99 ( 2008 ) k. f. reim , j. nunn , v. o. lorenz , b. j. sussman , k. c. lee , n. k. langford , d. jaksch , and i. a. walmsley , towards high - speed optical quantum memories , nature photonics , 4 , 218 - 221 ( 2010 ) u. schnorrberger , j. d. thompson , s. trotzky , r. pugatch , n. davidson , s. kuhr , and i. bloch , electromagnetically induced transparency and light storage in an at omic mott insulator , phys .
, 103 , 033003 ( 2009 ) b. kraus , w. tittel , n. gisin , m. nilsson , s. krll , and j. i. cirac , quantum memory for nonstationary light fields based on controlled reversible inhomogene ous broadening , phys .
a , 73 , 020302 ( 2006 ) j. nunn , n. k. langford , w. s. kolthammer , t. f. m. champion , m. r. sprague , p. s. michelberger , x .-
jin , d. g. england , i. a. walmsley , phys .
110(13 ) , 133601 ( 2013 ) nunn j , reim k , lee kc et al . , phys . rev .
lett . , 101(26 ) , 260502 ( 2008 ) simon c , de riedmatten h , afzelius m et al .
, 98 , 190503 ( 2007 ) valentina caprara vivoli , nicolas sangouard , mikael afzelius , nicolas gisin , new j. phys . , 15 , 095012 ( 2013 ) ivan iakoupov , anders s. sorensen , new j. phys . 15 , 085012 ( 2013 ) neil sinclair , erhan saglamyurek , hassan mallahzadeh , joshua a. slater , mathew george , raimund ricken , morgan p. hedges , daniel oblak , christoph simon , wolfgang sohler , wolfgang tittel , a solid - state memory for multiplexed quantum states of light with read - out on demand , arxiv:1309.3202 ( 2013 ) d.v .
vasilyev , i.v .
sokolov , e.s .
polzik , phys .
a 81 , 020302(r ) ( 2010 ) t.golubeva , yu.golubev , o.mishina , a. bramati , j. laurat , e.giacobino , high speed spatially multimode atomic memory , phys.rev .
a , 83(5 ) , 053810 ( 2011 ) a. dantan , j. cviklinski , m. pinard , and ph .
grangier , phys .
a 73 , 032338 ( 2006 ) j. nunn , i.a .
walmsley , m.g .
raymer , k. surmacz , f.c .
waldermann , z. wang , d. jaksch , phys .
a 75 , 011401 ( 2007 ) g. htet , a. peng , m. t. johnsson , j. j. hope , and p. k. lam , characterization of electromagnetically - induced - transparency - based continuous - variable quantum memories , phys .
a 77 , 012323 ( 2008 ) k. samburskaya , t. golubeva , yu .
golubev , and e. giacobino , quantum holography upon resonant adiabatic interaction of fields with an atomic medium in a lambda - configuration , opt .
110(5 ) , 775 ( 2011 ) t. golubeva , yu .
golubev , o. mishina , a. bramati , j. laurat , e. giacobino , high speed spatially multimode lambda - type atomic memory with arbitrary frequency detuning , eur .
j. d , 66 , 275 ( 2012 ) alexey v. gorshkov , axel andr , mikhail d. lukin , and anders s. sorensen , photon storage in @xmath21-type optically dense atomic media .
ii . free - space model , phys .
a 76 , 033805 ( 2007 ) j. appel , e. figueroa , d. korystov , m. lobino , and a. i. lvovsky , phys .
lett . , 100 , 093602 ( 2008 ) golubev yu .
, golubeva t. , ivanov d. broadband squeezed light from phase - locked single - mode sub - poissonian lasers , phys .
a. , 77 , 052316 ( 2008 ) y.m .
golubev , t.y .
golubeva , a.a .
gavrikov , c. fabre , pure and mixed stats in degenerate parametric oscillation , opt .
spectr . , 106(5 ) , 723 - 729 ( 2009 ) v. a. averchenko , t. yu .
golubeva , yu . m. golubev , c. fabre , broadband radiation of a degenerate optical parametric oscillator operating above threshold in information applications , opt .
spectr . , 105 ( 5 ) , 758 - 770 ( 2008 ) k. s. samburskaya , t. yu .
golubeva , v. a. averchenko , and yu .
m. golubev , quadrature squeezing in an isolated pulse of light , opt .
, 113(1 ) , 86 - 95 ( 2012 ) alexey v. gorshkov , axel andr , mikhail d. lukin , and anders s. sorensen , photon storage in @xmath21-type optically dense atomic media .
i. cavity model , phys . rev . a 76 , 033804 ( 2007 )
|
in this article the ability to record , store , and read out the quantum properties of light is studied .
the discussion is based on high - speed and adiabatic models of quantum memory in lambda - configuration and in the limit of strong resonance .
we show that in this case the equality between efficiency and squeezing ratio , predicted by the simple beamsplitter model , is broken .
the requirement of the maximum squeezing in the output pulse should not be accompanied by the requirement of maximum efficiency of memory , as in the beamsplitter model . we have demonstrated a high output pulse squeezing , when the efficiency reached only about 50% .
comprehension of this `` paradox '' is achieved on the basis of mode analysis .
the memories eigenmodes , which have an impact on the memory process , are found numerically .
also , the spectral analysis of modes was performed to match the spectral width of the input signal to the capacities of the memories .
@xmath0 ) @xmath1 ] ^
|
in @xcite @xcite asked himself _ how large must a universe be in order that it shall contain light emitted within it ? _
if we change the wording _ how large _ by _ how strong the gravity field must be _ and the word _
universe _ by the word _ object _ , then we have the history of _ black holes _ as follows : first known consideration of light not being able to escape from such an object ( not a universe ) was given by @xcite in @xcite .
eleven years later @xcite addressed the same question .
after the publication of einstein s general relativity @xcite published in @xcite a solution to general relativity advancing his well known radius ( later referred as the gravitational radius of a mass @xmath0 , @xmath6 ) .
about 50 years later it was recognized that black holes were predicted by general relativity .
was @xcite the known author of the name _ black hole _ in @xcite , referring to the _ continuous gravitational collapse _ of an over - compact mass .
it is evident that the first author to treat our universe as a black hole , five years before this name was used , was @xcite .
it is not necessary to think of a black hole as an over - compact mass .
one can have a black hole with any mass , as long as it also has the right size .
the biggest mass that one can consider is the mass of the _ seeable _ universe : a universe that contains all light emitted within it .
usually the word _ visible
_ universe refers to the visible light from the surface of last scattering , from the recombination epoch ( like the cmbr ) .
observable _ universe refers to the universe we could observe with light or some other means from the initial big - bang . as we will see , here the word _ seeable
_ refers to the universe considered as a black hole , like the surface of the event horizon , as noted by @xcite .
now , an average density of the observable universe is of the order of @xmath13 , about the critical density .
the age of the universe t is now believed to be about @xmath14 ( @xcite @xcite ) . then multiplying this by the speed of light we get the number @xmath15 .
if we look deep in space we are looking back in time , much more back the deeper we look .
we are limited in going back in time by the amount of @xmath14 , as we will see .
we are then limited to see deeper than about @xmath16 .
we may roughly calculate a mass for the observable universe by computing the mass from the density inside a sphere of size @xmath15 .
this gives a mass of about @xmath17 using the general relativity coefficient for volume @xmath18 .
using this value for the mass one gets the value of the gravitational radius of the universe as @xmath19 we have called this gravitational radius as the radius of the sphere of the _ seeable _ universe .
it has a deep physical meaning .
this wording comes from the calculation of the size of a universe that has all its light trapped , as we will see ( identified here with the gravitational radius @xmath7 ) .
it is then a maximum that can not be observed in the sense that any light getting close to this limit just turns around and goes back to its origin .
it is like being able to see your own back looking deep in space .
but the observational size of the universe is less than that today . when calculating it as the hubble radius , @xmath20 , a conventional way to define an observational universe ,
it is smaller than the _ seeable _
size @xmath7 .
we will prove that the radius of the seeable universe increases with time as @xmath21 , while the size of observable universe increases as ct ( today about @xmath16 ) . taking the speed of light as constant
this means a linear relation with time .
the speed of expansion for the gravitational radius is @xmath22 and its acceleration is therefore constant , an accelerated universe as has been observed for the last ten years type ia supernova , @xcite ; and cmb anisotropies , @xcite. clearly the gravitational radius is larger than the observable radius ( about a factor of @xmath23 today ) and its expansion going faster drags the far away galaxies out of sight . looking and identifying a far away galaxy , and looking again after some time long enough
, no galaxy would be there : it just disappears .
this is a proposed test that may be technically possible in the near future . with the expansion due to the dark energy background of the universe going as @xmath24
, it may be that this happens earlier than thought ( @xcite @xcite , @xcite @xcite ) .
in @xcite @xcite found an upper bound for the ratio of the entropy @xmath25 to the energy @xmath26 of any bounded system with effective size @xmath27 : @xmath28 about ten years later ( @xcite @xcite , @xcite @xcite ) a holographic principle was proposed giving a bound for the entropy @xmath29 of a bounded system of effective size @xmath27 as @xmath30 if the two bounds turn out to be identical @xmath0 is proportional to @xmath27 and the system strictly obeys the schwarzschild condition for a black hole .
also , for the case of a black hole we find that the hawking and unruh temperatures are the same .
identifying the bounds of equations ( [ e2 ] ) and ( [ e3 ] ) @xmath31 simplifying and using @xmath32 , we get @xmath33 this is the condition for the system @xmath34 to be a black hole .
then , its entropy is given by the @xcite relation @xmath35 that coincides with the two bounds of equations ( [ e2 ] ) and ( [ e3 ] ) .
the mass of the universe @xmath36 is a maximum . and
so is its size @xmath27 .
a bounded system implies a finite value for both . using present values for @xmath37 and @xmath38 they very closely fulfill the schwarzschild condition ( [ e4 ] ) .
this has been taken as an evidence for the universe to be considered as a black hole ( @xcite @xcite , @xcite @xcite ) .
the fact that a black hole has a temperature , and therefore entropy ( established by @xcite @xcite ) , implies that an observer at its surface , or event horizon , sees a perfect blackbody radiation , a thermal radiation with temperature @xmath39 given by @xmath40 where @xmath41 is the planck s constant , @xmath5 the speed of light , @xmath3 the gravitational constant , @xmath42 boltzmann s constant and @xmath0 the mass of the black hole .
this observer feels a surface gravitational acceleration @xmath43 . according to the @xcite effect
an accelerated observer also sees a thermal radiation at a temperature @xmath44 , proportional to the acceleration @xmath43 and given by @xmath45 based upon the similarity between the mechanical and thermo dynamical properties of both effects , equations ( [ e6 ] ) and ( [ e7 ] ) , we identify both temperatures and find the relation : @xmath46 identifying the unruh acceleration to the surface gravitational acceleration : @xmath47 and substituting in equation ( [ e8 ] ) we finally get @xmath48 this is the condition for a black hole .
since the hawking temperature refers to a black hole this result confirms the validity of the identification of the two temperatures , equations ( [ e6 ] ) and ( [ e7 ] ) , as well as the interpretation of the unruh acceleration in equation ( [ e9 ] ) .
from the equivalence principle , one gets the strength of the gravitational effect of a mass @xmath0 on the metric at position @xmath4 from it ( @xcite @xcite ) .
it is proportional to the dimensionless parameter @xmath1 , which normally is @xmath2 . here
@xmath3 is the gravitational constant , @xmath0 the mass of the gravitating body , @xmath4 the position of the metric from the gravitating body and @xmath5 the speed of light .
seeable _ universe is the sphere , with center at the observer , having a size such that it shall contain all light emitted within it . for this to occur
, one can impose that the gravitational effect on the velocity of light at @xmath4 give a zero value for the radial component , and non zero for the tangential one ( @xcite @xcite ) .
then , light is always inside the seeable universe , trapped within it .
the condition is given by the equality @xmath6 , where @xmath7 represents the radius of the _ seeable _ universe .
it is the gravitational radius of the mass @xmath0 .
we will prove in the next section that the cosmological equation of state @xmath49 , with the present observed value of @xmath50 , implies that the gravitational radius @xmath7 varies as @xmath51 .
the einstein tensor @xmath52 has a 4d zero divergence by construction : @xmath53 the tensor @xmath52 is a geometrical object and is the left hand side of the field equations .
then , the covariant divergence of the second term , i. e. , @xmath54 , @xmath55 the energy - momentum tensor , must be zero .
the resultant equation has been found under the full generalization of considering @xmath3 , @xmath5 and the cosmological @xmath56 as time variables ( @xcite @xcite ) .
we repeat it here : @xmath57 where @xmath58 is here energy density , @xmath59 the hubble parameter @xmath60 , @xmath61 the cosmological scale factor as is defined in the robertson - walker metric , and @xmath62 defines the equation of state . considering the cosmological constant @xmath56 as a true constant
one can integrate equation ( [ e12 ] ) giving , @xmath63 using the conversion @xmath64 $ ] , since @xmath65 is mass density , i.e. @xmath66 , the result ( [ e13 ] ) is transformed to @xmath67 and using the definition of the gravitational radius @xmath7 as the schwarzschild radius we get from equation ( [ e14 ] ) @xmath68 hence , the gravitational radius , the _
seeable _ size of the universe as a black hole , depends only on the cosmological scale factor @xmath61 and the equation of state parameter @xmath62 . with the present value for the parameter @xmath69 we have the solution from equation ( [ e15 ] ) as @xmath70 we use now the friedmann equation that , together with equation ( [ e15 ] )
, completely defines the cosmological model ( the integration of the two einstein cosmological equations ) : @xmath71 ^ 2 = 8 \pi g \rho / 3 = g m / [ a(t)]^3 \times \hbox{constant } \label{e17}\ ] ] whatever the time dependence of @xmath3 and @xmath0 may be , we prove now that their product @xmath72 must be a constant .
this comes from the consideration that the entropies in equations ( [ e2 ] ) , ( [ e3 ] ) and ( [ e5 ] ) must be an extensive property , proportional to mass @xmath0 .
we know that @xmath3 is always associated with a mass .
then the constancy of mass can be substituted by the constancy of @xmath72 . the friedman equation ( [ e17 ] )
is then , @xmath73^{2 } = \hbox{constant } \label{e18}\ ] ] which is the well known einstein - de sitter model with solution : @xmath74 = ( t / t_0)^{2/3 } \label{e19}\ ] ] recently ( @xcite @xcite ) it has been shown that the probability that the universe emerges right after a _ big trip _ , thus avoiding a _ big rip _ , in a state of initial constant @xmath75 , will be maximal if and only if one has @xmath76 . from equation ( [ e15 ] ) we see that this value corresponds to a gravitational radius for the universe expanding as the cosmological factor @xmath61 given in equation ( [ e18 ] ) , an initial einstein - de sitter state of the universe expanding as in equation ( [ e19 ] ) . after this
we have today strong evidence that the value of @xmath62 is @xmath77 , at least as far back as redshift @xmath78 . again , going to equation ( [ e15 ] ) we see that the gravitational radius of the universe expands now as , @xmath79^{-3 \omega } = [ a(t ) / a(t_0)]^{3 } = ( t / t_0)^2 \label{e20}\ ] ] then , the gravitational radius of the universe starts with a time dependence as given by the einstein - de sitter universe , equation ( [ e19 ] ) with the equation of state @xmath80 and evolves in such a way that the dark matter and energy components make the equation very close to @xmath9 as of today .
the observable size of the universe going as @xmath11 implies that all contents in the universe are dragged faster than this , @xmath51 as in equation ( [ e20 ] ) .
in our approach galaxies are physically dragged by expansion as @xmath51 which means that they are disappearing when they are dragged beyond the distance @xmath81 . by looking at a particular identifiable galaxy deep in space
, we could check its disappearance after some finite time .
this implies a repetition of observations of the same galaxy deep in space hoping to find at some later time that the pinpointed galaxy is no longer there .
figures [ f1 ] and [ f2 ] present this situation .
@xcite presented in @xcite arguments to expect objects to become out of causal contact . later
@xcite , as well as @xcite , pointed out this possibility about ten years ago . in this direction an out of sight possibility for far away galaxies has been pointed out by @xcite .
the calculations made , as @xcite presented , must be halved in our case due to the @xmath51 dependence of the dragging background presented here .
we also present it under a different point of view , with the same basic idea of getting out of sight galaxies in time due to accelerated expansion .
our conclusion on the dragging effect going as @xmath51 halves the calculations presented by these authors .
position of galaxy a at @xmath82 in the observable universe.,width=8,height=8 ] position of galaxy a at @xmath83 out of the observable universe.,width=8,height=8 ]
the universe can be seen as a black hole . from each observer in it
we can interpret that he is at the centre of a sphere , with the hubble radius , that would be the observable universe , but there is the gravitational radius that is larger than the observable one .
the isotropic acceleration present at each point in the universe , and given by equation ( [ e9 ] ) , implies that there is no distortion for the spherically distributed acceleration if we consider the strict conditions of the cosmological principle .
however , the presence of a nearby important mass , like the sun , in any point in interplanetary space will distort this spherically symmetric picture .
with respect to the probes pioneer 10/11 , that detected an anomalous extra acceleration towards the sun of value @xmath84 ( @xcite @xcite ) , we can see that this value is only a bit higher than the one predicted by equation ( [ e9 ] ) , which is about @xmath85 .
this difference is an effect that can be explained by the presence of the sun converting the isotropic acceleration to an anisotropic one .
for about ten years we have good evidence that the universe is expanding in an accelerated way .
we have found here that the present observations , that favours the value @xmath69 for the equation of state , imply an expanding law of the type @xmath21 giving a constant acceleration .
thus , the gravitational radius rg , that can be interpreted as the size of the universe as a black hole , trapping all light gives the maximum _ seeable _
size for the universe .
the hubble radius , @xmath20 , can be interpreted as the maximum observable size for the universe since it is limited by its age , the distance travelled by light going backwards in time up to the initial stages .
the background _
is then expanding at a faster rate than it can be observed .
and this _ vacuum _ is the frame for all cosmological structures , like the galaxies .
then , galaxies are physically dragged by expansion as @xmath51 which means that they are disappearing when they are dragged beyond the distance @xmath20 . by looking at a particular identifiable galaxy very deep in space
, we could check its disappearance after some finite time .
the past calculations for this to occur are here halved by the expansion law @xmath86 .
alfonso - faus , a. : astrophys .
space sci . * 325 * , 113 ( 2010 ) anderson , j.d .
, et al . :
prl * 81 * , 2858 ( 1998 ) bekenstein , j.d . : prd * 23 ( 2 ) * , 287 ( 1981 ) .
belinchn , j.a .
, alfonso - faus , a. : int . j. mod .
d * 10 * , 299 ( 2001 ) bennett , c.l . , et al .
: apj suppl * 148 * , 1 ( 2003 ) bennett , c.l . , et al .
: _ nine - year wilkinson microwave anisotropy probe ( wmap ) observations : final maps and results _ , preprint .
arxiv : astro - ph.co/1212.5225v3 ( 2013 ) clocchiatti , a. , et al : apj * 642 * , 1 ( 2006 ) de bernardis , p. , et al . :
nature * 404 * , 955 ( 2000 ) m. j. fullana i alfonso , m.j .
, alfonso - faus , a. : astrophys .
space sci .
* 337 * , 19 ( 2012 ) gudmundsson , e.h . ,
bjornsson , g. : apj * 565 * , 1 ( 2000 ) hanany , s. , et al .
: apj * 545 * , l5 ( 2000 ) hawking , s.w . :
nature * 248 * , 30 ( 1974 ) laplace , p.s . : 1795 , english version _ the system of the world _ , w. flint , london ( 1809 ) lineweaver , c.h . ,
davis , t.m . : misconceptions about the big bang , scientific american , march 2005 , 36 - 45 ( 2005 ) loeb , a. : prd * 65 * , 047301 ( 2002 ) michell , j. : philosophical transactions of the royal society * 74 * , 35 ( 1784 ) ney , e.p .
: electromagnetism & relativity , harper & row , new york ( 1962 ) perlmutter , s. et al .
, apj * 517 * , 565 ( 1999 ) riess , a.g .
, et al . :
aj * 116 * , 1009 ( 1998 ) riess , a.g .
, et al . :
apj * 607 * , 665 ( 2004 ) schwarzschild , k. : sitzber . dent .
berlin , kl , math .- phys .- tech .
189 - 196 and 424 - 434 ( 1916 ) starkman , g. , trodden . ,
m , vachaspati , t. : prl * 83 * , 1510 ( 1999 ) susskind , l. : math phys * 36 * , 6377 ( 1995 ) t hooft , g. : _ dimensional reduction in quantum gravity _ , preprint .
arxiv : gr - qc/9310026 ( 1993 ) tonry , j.l . , et al .
: apj * 594 * , 1 ( 2003 ) unruh , w.g .
: prd * 14 ( 4 ) * , 870 ( 1976 ) wheeler , .a . : jaaas invited lectures , friday december 29 , new york hilton ( 1967 ) .
yurov , a.v .
astashenok , a.v . , v. a. yurov , v.a .
: astrophys .
space sci , * 342 * , 1 ( 2012 )
|
from the equivalence principle , one gets the strength of the gravitational effect of a mass @xmath0 on the metric at position r from it .
it is proportional to the dimensionless parameter @xmath1 , which normally is @xmath2 . here
@xmath3 is the gravitational constant , @xmath0 the mass of the gravitating body , @xmath4 the position of the metric from the gravitating body and @xmath5 the speed of light .
the seeable universe is the sphere , with center at the observer , having a size such that it shall contain all light emitted within it . for this to occur one can impose that the gravitational effect on the velocity of light at @xmath4 is zero for the radial component , and non zero for the tangential one .
light is then trapped .
the condition is given by the equality @xmath6 , where @xmath7 represents the radius of the _ seeable _ universe .
it is the gravitational radius of the mass @xmath0 .
the result has been presented elsewhere as the condition for the universe to be treated as a black hole . according to present observations , for the case of our universe taken as flat ( @xmath8 ) , and the equation of state as @xmath9
, we prove here from the einstein s cosmological equations that the universe is expanding in an accelerated way as @xmath10 , a constant acceleration as has been observed .
this implies that the gravitational radius of the universe ( at the event horizon ) expands as @xmath10 .
taking @xmath5 as constant , observing the galaxies deep in space this means deep in time as @xmath11 , linear .
then , far away galaxies from the observer that we see today will disappear in time as they get out of the distance ct that is @xmath12 .
the accelerated expanding vacuum will drag them out of sight .
this may be a valid test for the present ideas in cosmology .
previous calculations are here halved by our results .
|
it is conventionally believed that the structure of accretion environment of a strongly magnetized neutron star in an x - ray pulsar , as well as its emission , are actually governed by a few key parameters such as the neutron - star spin period and magnetic field , the binary orbital period and separation of the companions , and of course the accretion rate @xmath0 .
sporadic outbursts occurring in some x - ray pulsars when the bolometric x - ray luminosity @xmath1 increases due to an enhanced accretion rate provide for the possibility of probing the observable response of such a complex system as an accreting x - ray pulsar to varying @xmath0 and getting more insight into the accretion onto a strongly magnetized neutron star and the accompanying high - energy emission . in particular , it is known since ginga observations of the x - ray pulsar 4u0115 + 63 in 19901991 @xcite that the energies of cyclotron absorption features observed in x - ray pulsar spectra can vary with their luminosities .
such variations are conventionally interpreted in terms of the luminosity dependence of the height of the line - forming region above the neutron - star surface @xcite .
the correspondent negative correlation between the cyclotron - line energy and the luminosity was observed to date for several bright x - ray pulsars ( e.g. @xcite ) .
note that the opposite case of the positive correlation was also registered at least for one or two x - ray pulsars ( e.g. see @xcite for her x-1 and @xcite for gx301 - 2 ) .
it should be emphasized that , to date , cyclotron - line parameters for all sources were mostly studied during fading phases of outbursts because of time lags between the beginning of an outburst and the start of observations .
for example , it was shown by @xcite ( hereafter referred to as paper i ) that the energy of the cyclotron line observed in spectra of the x - ray pulsar v0332 + 53 increases almost linearly with decreasing luminosity during the fading phase of an outburst in 20042005 .
this paper presents a comprehensive analysis of temporal and spectral parameters of v0332 + 53 over a nearly complete powerful outburst of 20042005 , including the brightening phase studied in detail for the first time .
this work employs the rxte data @xcite acquired during the observations with the following ids : 90014-xx - xx - xx , 90089-xx - xx - xx and 90427-xx - xx - xx .
note that paper i was based only on the i d 90427-xx - xx - xx data covering the fading phase of the outburst .
since the aims of the present work and paper i are slightly different , these data were reanalyzed to ensure the uniformity with the ids 90089-xx - xx - xx and 90014-xx - xx - xx corresponding to the brightening and fading phases of the outburst , respectively . in general ,
the data - analysis procedure was identical to that used in paper i. the luminosity was calculated assuming a source distance of 7 kpc @xcite .
note that the correct account of the dead time is crucial since the source count rate reached very high values . using improved software
, we recalculated the luminosities for the observations in the i d 90427 set and found that the luminosities for a few brightest epochs of this set are to be lowered by about 1520% in comparison with the values published in paper i. note , these corrections are not too large to affect the conclusions of paper i. the time dependence of the bolometric luminosity in the photon energy range @xmath2 kev during the outburst of 20042005 is shown in fig .
[ lcurve ] .
all the data are naturally grouped into the brightening ( filled green squares ) and fading phases ( open blue circles ) of the outburst . for comparison , the result of integral observations
are also plotted by the triangles in this figure ( see paper i ) .
the solid line is the rescaled light curve obtained with the asm monitor in the photon energy range @xmath3kev .
further data reduction for both pca and hexte spectrometers was done using standard tools of ftools / lheasoft 6.3.2 package . since the source was bright and
the outburst was nearly completely covered with the rxte observations ( see fig . [ lcurve ] ) , it is possible to obtain the high - quality broadband energy spectra of the source at different stages of the outburst and study the evolution of source parameters with the luminosity and the outburst phase . it was noted in paper i that the best - fit cutoff energy @xmath4kev for the xspec _
powerlaw@xmath5highecut _ model @xcite , which is usually used for approximation of x - ray pulsars spectra , is close to the lower limit ( @xmath63 kev ) of the pca energy range , so that the photon index can not be accurately retrieved from such a continuum model .
hence , the continuum spectrum of the source was fitted with the simpler _ cutoffpl _ model which describes the source spectrum similarly well , but has fewer parameters and is free of the above limitations .
the continuum spectrum was modified by one or few cyclotron - harmonic line features depending on the source luminosity .
a cyclotron absorption line was fitted with the lorentz profile @xcite @xmath7 where @xmath8 , @xmath9 , and @xmath10 are the line central energy , width , and depth , respectively .
the fundamental cyclotron line ( below we ll call it the cyclotron line for a simplisity ) is clearly detected in all observations , whereas the second - harmonic line was confidently observed not in all data sets . in some cases ,
inclusion of the third harmonic in the model affected the parameters of the second - harmonic line .
the third harmonic itself was registered and its parameters can be more or less reliably retrieved only for several brightest states ( see paper i for details ) .
we examined possible effects of including the second harmonic in the fit on the parameters of the cyclotron line and the continuum . in these tests ,
the central energy of the second - harmonic line was either a free parameter or fixed equal to two times the central energy of the cyclotron line .
it was found that the cyclotron line and continuum parameters do not depend within errors on the higher harmonics . .
this section presents the results of spectral analysis during almost entire outburst of v0332 + 53 .
the analysis is focused on the source spectral parameters as functions of the luminosity and time .
v0332 + 53 is one of two x - ray pulsars whose spectra exhibit at least three cyclotron - harmonic absorption line features @xcite . in general
, this source has a fairly complex , multi - component spectrum .
therefore , we applied the above - described spectral model including a power law with an exponential cutoff , up to three absorption lines depending on the source brightness and detectability , and the 6.4-kev iron emission line which was clearly observed in all states of the source . since the iron line was narrow and the energy resolution of the pca spectrometer at @xmath11 kev was limited ( @xmath12% ) , the actual line width can not be accurately determined
thus , a fixed value of @xmath13 kev was used in the xspec data analysis .
variations in the main spectral parameters with the time and the luminosity are described in detail below .
it was mentioned above that variations in the cyclotron - line energy and their possible dependence on the luminosity were discovered and reported by @xcite for several bright x - ray pulsars . the cyclotron - lineenergy
sourceluminosity relation was studied in detail for the first time by @xcite for v0332 + 53 in a wide luminosity range .
using integral and rxte data obtained during the fading phase of the 20042005 outburst , they discovered the strong negative correlation between the cyclotron line energy and the luminosity which can be described fairly well by a linear function . including both the fading and brightening phases of the outburst into an analysis makes it possible to study the possible difference in the cyclotron - line behavior during these outburst stages to verify the hypothesis on the hysteretic dependence of the source parameters on its luminosity .
figure [ cycvst ] shows the temporal evolution of the cyclotron - line energy during the entire outburst .
distinct asymmetry of this time profile can obviously be attributed to the asymmetry of the outburst light curve .
the existence of hysteresis the cyclotron - line energy as a function of the source was checked by constructing this dependence for the entire outburst ( see fig . [ cycvslx ] ) .
the resulting plot unambiguously rules out the hysteresis hypothesis .
this is valid for luminosity above @xmath14 erg s@xmath15 ( the lower - luminosity data are absent in the brightening phase ) .
hence , the physics of accretion column and its emission is essentially same at both these stages . .
symbols are the same as in fig.[lcurve ] . ] a formal linear fit of this dependence ( all available measurements have been used ) , @xmath16}\!\simeq\!-\!0.143(\pm0.002)l_{37}\!+\!29.56(\pm0.03)$ ] ( here , @xmath17 is the luminosity in the units of @xmath18 erg s@xmath15 , uncertainties correspond to a 90% confidence level ) , is practically the same as the one obtained in paper i for only the fading phase of the outburst and even takes into account the above - mentioned luminosity corrections for several highest - count - rate observations .
note , however , that allowance for these corrections makes the behavior of the cyclotron energy at very high luminosities @xmath19 erg s@xmath15 slightly different from this fit : the rate of the cyclotron - energy variations is higher . such a difference
( if any ) may be connected with possible changes in the accretion - flow structure and emission at very high luminosities then the source can transfer into other accretion regime . on the other hand ,
poor statistics impedes finalizing conclusions on the behavior of the cyclotron - line energy at low luminosities below @xmath20 . some evidence is seen in fig . [ cycvslx ] for steepening of the cyclotronenergy
luminosity slope below @xmath21 erg s@xmath15 , but its significance is not enough for the final conclusion .
note , that the behavior of the cyclotron energy at low luminosities is very interesting from the viewpoint of determining the so - called critical luminosity at which the shock wave at the top of the accretion column disappears and/or detecting the same reverse dependence ( if any ) of the cyclotron - line energy with the luminosity as was observed for her x-1 @xcite .
meanwhile , the lowest observed value of the cyclotron energy in the v0332 + 53 spectrum provides for an estimate for the magnetic field on the neutron - star surface .
unfortunately , the lowest - luminosity data point in fig.[cycvslx ] has fairly large uncertainty .
hence , it seems more expedient to use the result from the linear fit , that is 29.56 kev .
this yields @xmath22 where @xmath23 is the gravitational redshift near the surface of a neutron star with a radius of 10 km and a mass of @xmath24 .
( open circles and filled squares ) and the depth @xmath10 ( filled circles ) of the cyclotron line in v0332 + 53 spectra on the pulsar luminosity @xmath17 .
symbols are the same as in fig.[lcurve ] . ]
it is well known from the classical stellar spectral analysis that the equivalent width of a spectral line is a very important quantity for retrieving physical parameters and conditions in the line - forming region .
hence , in addition to such cyclotron - line model parameters as @xmath9 and @xmath10 , it seems expedient to have an observable quantity playing the role of cyclotron - line equivalent width in x - ray astrophysics and quantifying the total number of photons withdrawn from the continuum spectrum of a source due to resonance scattering in the cyclotron line(s ) .
the main problem in calculating the cyclotron - line equivalent width is its strong dependence on the adopted continuum model to be multiplied by the line models typically used to fit cyclotron - harmonic lines .
we used the following simple technique to calculate the cyclotron - line equivalent width @xmath25 : ( 1 ) all energy channels affected by the cyclotron line ( typically @xmath26 kev ) are excluded from the source spectrum ; ( 2 ) the remaining energy bins are fitted as a spectrum of the continuum ; ( 3 ) the energy bins excluded at step 1 are added again to the spectrum , their deviations from the continuum fit obtained at step 2 is calculated , and , upon normalization to the continuum spectrum , are integrated over energy ; ( 4 ) the resulting energy - integrated deviation of the normalized cyclotron - line energy bins from the continuum fit is adopted as a measure for the equivalent width @xmath25 of the cyclotron line .
it was found that the quantity @xmath25 calculated in such a way is not constant , but varies with the cyclotron - line energy changes during the outburst .
the @xmath25 dependence on the luminosity is shown in fig .
[ eqwcycvslx ] . as in fig .
[ cycvslx ] , different symbols such as filled squares and open circles correspond to the brightening and fading phases of the outburst , respectively .
the cyclotron - line equivalent width decreases almost linearly with the luminosity repeating the behavior of the cyclotron - line energy itself , including the increase of the `` rate '' of the equivalent width changes at high luminosities .
but it is nessesary to note , that the slope of this formal linear relation @xmath27}\!\propto\!-\!0.25l_{37}$ ] is different from that in fig.3 .
it is also worthy to note the similar behaviour of the cyclotron line optical depth @xmath10 ( see filled circles in fig .
[ eqwcycvslx ] ) , that can be understood from the revealed linear correlation between @xmath10 and @xmath25 . versus cyclotron energy @xmath8 for all observed luminosities .
symbols are the same as in fig.[lcurve ] . ]
the correlation between the cyclotron - line energy @xmath8 and its width @xmath9 has been noted earlier in several works @xcite for a number of x - ray pulsars with different cyclotron line energies and viewing angles .
the last one can play an important diagnostic role .
in particular , it was used by @xcite to explain nondetection of cyclotron lines in certain objects .
note , that the relation between @xmath9 and @xmath8 stems from both the doppler broadening due to thermal motion of electrons along the magnetic field in the line - forming region and the natural cyclotron - line broadening giving rise to broad lorentz - shaped wings of the line .
the observed doppler width of the cyclotron line is @xcite @xmath28 where @xmath29 denotes averaging over the duration of one observation which is typically about a few kiloseconds , i.e. , much longer than the x - ray pulsar period , @xmath30\ ] ] is the angle between the magnetic field in the line - forming region and the observer line of sight , @xmath31 is the angle between the neutron - star spin axis and the line of sight , @xmath32 is the angle between the neutron - star spin and magnetic axes , @xmath33 is the angle between the neutron - star magnetic axis and the plane comprising the neutron - star spin axis and the line of sight , and @xmath34 is the electron temperature . in the absence of neutron - star precession effects , @xmath35 over the binary orbit .
hence , @xmath36 and @xmath37 .
note , however , that the applied fit of the cyclotron line implies that the natural broadening implying the lorentz profile of the cyclotron line is assumed dominant . in this case , the observed relative cyclotron - line width @xmath38 is independent of the system geometry and the electron temperature in the source .
in addition , the ratio @xmath39 can also be contributed by other factors , e.g. , nonuniformity of the magnetic field over the line - forming region .
we examined the processed data for the presence of linear correlation between @xmath9 and @xmath8 .
the cyclotron line width as a function of its energy for v0332 + 53 during its 20042005 outburst is shown in fig .
[ sigma_ecyc ] .
no pronounced linear increase in @xmath9 with increasing @xmath8 is seen .
the typical width of the cyclotron line is about @xmath40 kev .
the fairly strong scatter of the data points for large cyclotron energies is probably caused by spectrum - statistics worsening at the low luminosities .
this is confirmed indirectly by an increase in the error - bar sizes .
an additional factor which affect to the observable broadening of the cyclotron line can be the finite energy resolution of the hexte spectrometer ( fwhm is about 5.5 kev at 30 kev ) . finally note that according to the theory of cyclotron - radiation transfer under neutron - star conditions ( see , e.g. , @xcite )
, the optical depth at the cyclotron line varies only slightly with @xmath41 and can hardly exclude the formation of a resonance scattering feature at fairly large @xmath41 . hence , the cyclotron line nondetection should rather be related to specific viewing conditions of the line - forming region such as its small visible area or screening of this region by the neutron star and/or accretion disc / stream .
it is well known that pulse profiles of x - ray pulsars are strongly dependent on the photon energy and luminosity .
this is especially prominent in soft energy bands strongly affected by the intrinsic absorption in the binary .
as the photon energy increases toward the hard x - ray range , the observed emission becomes progressively less sensitive to the poorly known environment of the accreting neutron star and carries more and more imprints of actual physical and geometrical parameters in hot polar - cap regions on the neutron star where this radiation is generated .
although these regions are believed to be much simpler in structure and physics , strong changes of pulse profiles are observed in their hard x - ray emission , as well . in particular , this was demonstrated by @xcite who used integral data to reconstruct uniformly the pulse profiles of ten bright x - ray pulsars in hard x - rays for different luminosities and study their variability .
this section is aimed at answering the question on the absence or presence of pulse - profile hysteretic behavior during the brightening and fading phases of the outburst from v0332 + 53 .
it was shown in the previous sections that the source spectrum and its parameters are practically the same at the outburst stages with the same luminosity .
taking into account the similarity of pulse profiles at the same luminosities for the brightening and fading phases of an outburst from 4u0115 + 63 @xcite we can expect the similar behavior for v0332 + 53 .
according to paper i , the main mystery of the considered source is the drastic change of its pulse profile near the cyclotron line . in particular
, it was shown that the pulse profile in the high - luminosity state ( @xmath42 ) has a sinusoidal double - peaked shape at all energies .
the relative intensities of the pulses are slightly different .
very small variations in the positions of the peaks depending on the photon energy range were observed . as the pulsar luminosity decreases , the pulse profile changes significantly . in the soft energy bands ( at photon energies less
then @xmath43 kev ) it becomes nearly single - peaked .
one the intra - peak gap disappears and the double - peaked profile transforms into an asymmetrical single - peaked one at energies near the cyclotron line . then the pulse profile becomes double - peaked again with the further increase in the photon energy . here , all the data covering the outburst of 20042005 , including the i d 90427 used in paper i , are reanalyzed uniformly and a detailed comparison of pulse profiles corresponding to the same luminosities during the outburst brightening and fading phases is made .
it is found that they are similar to each other ( minor differences may only exist in the softest energy band @xmath44 kev ) and their behavior with the luminosity identical to the described above .
note , that for several highest - luminosity observations it was impossible to produce the pulse profiles using pca data due to the absence of data modes with sufficient energy and time resolution below 20 kev .
for illustrative purposes the pulse profiles in different energy bands for the observation 90089 - 11 - 02 - 03 ( the luminosity was @xmath45 erg s@xmath15 at the brightening phase ) are shown in fig.[pp_2d ] .
the pca channels correspond to @xmath46 , @xmath47 and @xmath48 kev to demonstrate changes at soft photon energies and the transition from single- to double - peak shape .
following paper i , the hard energy channels ( hexte data ) were chosen to divide the cyclotron line in equal parts . .
position of the cyclotron line center is shown by the dashed line .
red colour corresponds to higher and blue to lower relative intensities .
solid lines represent levels of equal intensity : 0.20 , 0.24 , 0.28 , ... , 0.96 , 1.00 . ]
it is seen that relatively far from the cyclotron energy equal to @xmath49 kev for this observation , the pulse profile is double - peaked with a domination of a second peak in the photon energy range @xmath50 kev . in the left ( soft ) wing of the cyclotron line the profile is become single - peaked with its center at the pulse phase of @xmath51 . in the right ( hard ) wing of the cyclotron line
the profile is also single - peaked , but the position of its maximum is shifted to the pulse phase @xmath52 . at the higher energies the pulse profile becomes double - peaked again .
note , that the observed shifting of the peak over the phase near the cyclotron line energy is occurred gradually with increasing photon energy .
these features can be seen clearly in fig.[pp3d ] plotted using the technique described in detail in paper i. note that the observed pulse profile is contributed by three main components : ( i ) direct radiation coming from one or two under - the - shock hot polar columns which produce(s ) the continuum and hosts the cyclotron - line - forming region being presumably the upper part of the column adjacent to the shock , ( ii ) forward - scattered radiation outgoing from the side of the accretion stream feeding the `` front '' ( directly visible ) polar column and ( iii ) backward - scattered radiation outgoing from the side of the accretion stream feeding the `` back '' ( neutron - star shaded ) polar column .
quantitative pulse - profile modeling is a complicated problem requiring good knowledge of the features of radiation beaming near the cyclotron frequency @xcite .
in particular , it was shown by @xcite that the accretion column beam function at the cyclotron energy is significantly different from the radiation beaming at other energies .
it was already observed that the pulsed fraction @xmath53 , where @xmath54 and @xmath55 are maximum and minimum intensities in the pulse profile of an x - ray pulsar , respectively , increases with energy in the hard x - rays ( @xmath56 kev ) for all studied sources ( see e.g. @xcite and a last review of @xcite and references therein ) . however , this increase in not monotonic for some sources and shows such local features as , e.g. , maxima near the cyclotron lines .
the pulsed fraction as a function of the photon energy was studied in detail in a wide energy band ( @xmath2 kev ) for the pulsar 4u0115 + 63 using the rxte data @xcite .
it was found in this paper that ( i ) the pulsed fraction increases with the photon energy and decreased with the luminosity and ( ii ) the local pf maxima ( hump - like features ) exist not only near the cyclotron line , but also near the higher harmonics .
note that @xcite analyzed archival bepposax data for 4u0115 + 63 and found a local decrease in the pulsed fraction near the second cyclotron harmonic in the middle of the fading phase .
erg s@xmath15 .
the green open and red filled squares correspond to @xmath57 and @xmath58 , respectively , during the brightening phase ( rxte observations 90089 - 11 - 02 - 03 and 90089 - 11 - 04 - 04 , respectively ) .
the magenta open circles and blue open triangles correspond to @xmath59 and @xmath60 , respectively , during the fading phase ( rxte observations 90014 - 01 - 02 - 00 and 90014 - 01 - 05 - 02 , respectively ) . ] in this work , the v0332 + 53 pulsed fraction is analyzed in various photon energy ranges and for different luminosity states using the same complex approach that was employed by @xcite for 4u0115 + 63 .
figure [ ppfr_pca_hex ] shows the energy dependences of the pulsed fraction for observations with various pulsar luminosities obtained in a wide photon - energy range for the brightening and fading phases of the outburst .
the corresponding luminosities in the units of @xmath18 erg s@xmath15 are indicated to the right of each plot .
it is seen that the behavior of the v0332 + 53 pulsed fraction at high energies ( @xmath56 kev ) is typical of x - ray pulsars ( see , e.g. , @xcite ) . in particular , the pulsed fraction increases with the photon energy and the form of the increasing function @xmath61 depends on the luminosity , but changes only slightly over a wide luminosity range and is roughly the same for similar luminosities during the brightening and fading phases of the outburst . on the contrary ,
these dependencies for soft energy bands are very different .
we should specially note the photon energies below @xmath62 kev for which an increase in the pulsed fraction with decreasing photon energy is observed for the first time .
such behavior observed in a wide luminosity range is the most prominent for luminosities @xmath63 erg s@xmath15 .
it is difficult to explain it within the framework of known models since , in particular , no significant absorption was observed in the source spectrum during the entire outburst .
further detailed studies are required to clarify this problem and we plan to perform such an analysis in a separate paper . here
, we only note a steep rise in the pulsed fraction from @xmath64% to @xmath65% with an increase in the luminosity from @xmath66 to @xmath67 erg s@xmath15 and changes in the pulsed fraction behavior in the soft energy bands .
it should also be noted that the v0332 + 53 pulsed fraction at soft energies is small in comparison with values observed for other x - ray pulsars .
this fact was mentioned previously by @xcite who attempted to explain this feature assuming that the angles @xmath31 and/or @xmath32 are small . however , both these assumptions can hardly be conformed with the observed high values of the high - energy and high - luminosity pulsed fraction and the observed energy and luminosity dependence of the pulsed fraction . the most interesting feature at the high energies is a local pulsed fraction maximum near the cyclotron - line energy . such a hump - like feature is observed for almost any luminosity during both brightening and fading phases of the outburst . to rule out its possible artefact origin caused either by the hexte data itself
, we obtained the pulsed fraction dependence on the energy using the integral data ( @xcite , see also http://hea.iki.rssi.ru/integral/pulsars/index.php ) .
figure [ ppfr_hex_int ] shows how the pulsed fraction changes with the energy for two values of the luminosity .
the rxte / hexte and integral / isgri results are shown by the open circles and filled triangles , respectively .
the upper and lower panel of the figure correspond to the high and low luminosities @xmath68 and @xmath69 , respectively
. it is clearly seen for both luminosities that the pulsed fraction dependencies are practically the same for the rxte and integral data and the
hump-like features are clearly detected in both data sets .
we analyzed all the available hexte data and found that the hump
energy ( @xmath70 kev ) is more or less stable over a wide luminosity range .
this can imply that such a pulsed fraction hump
is possibly related to some intrinsic properties of the neutron - star magnetosphere rather than to the local cyclotron energy corresponding to the current luminosity . finally , it is interesting to note that the energy of this
hump-like feature is about of the maximum measured value of the cyclotron energy for this source ( see section 3.1 ) corresponding to the magnetic field at the surface of the neutron star .
it was shown by @xcite that the pulsed fraction in a wide energy band usually decreases with increasing luminosity of an x - ray pulsar .
this fact was preliminarily explained within the framework of a geometrical model in which the pulsed fraction is determined by the luminosity - dependent visible areas of the accretion columns .
kev versus the v0332 + 53 bolometric luminosity . as before ,
squares , circles and triangles of different types represent the rxte / hexte data covering the brightening and fading phases of the outburst at rising and decay parts of the outburst and the integral / isgri data , respectively . ] the luminosity dependence of the pulsed fraction in the photon energy range @xmath71 kev is shown in fig .
[ ppfrvslx ] .
the observed behavior does not fully conform to the predictions of the above - mentioned model . whereas the pulsed fraction decreases with increasing luminosity below @xmath72 erg s@xmath15 in accordance with this model , its value remains almost constant at
a level of @xmath73% in the luminosity range @xmath74 erg s@xmath15 and rises almost linearly at higher luminosities .
note that the luminosity dependence of the pulsed fraction is the same for both brightening and fading phases of the outburst .
the main results of the present work can be summarized as follows : 1 .
the brightening phase of the 20042005 outburst in the x - ray pulsar v0332 + 53 is studied in detail for the first time , thereby completing an analysis of the evolution of the accreting x - ray pulsar parameters over an entire outburst . 2 .
it is shown that the equivalent width @xmath25 demonstrates the strong positive linear correlation with the cyclotron - line depth @xmath10 and , similar the cyclotron - line energy @xmath8 , decreases almost linearly with the bolometric luminosity @xmath1 .
the hypothesis of hysteretic behavior of the observed radiation during the brightening and fading phases of the outburst was checked and can be ruled out . in particular , the source spectrum and its parameters and the pulsed fraction behavior are practically the same at the outburst stages with the same luminosity . 4 . the presence of a local hump - like feature near the cyclotron - line energy in the pulsed fraction as a function of the photon energy is confirmed .
its position remains nearly constant at about @xmath75 .
an increase in the pulsed fraction with decreasing photon energy below @xmath62 kev is observed for the first time in several observations .
the dependence of the pulsed fraction in the energy range @xmath71 kev on the v0332 + 53 bolometric luminosity during the brightening phase of the 20042005 outburst is found and compared with the similar dependence for the fading phase of the same outburst .
no hysteretic behavior is observed .
it was revealed for the first time that the pulse fraction rises almost linearly at higher luminosities .
we thank the referee for his useful comments and suggestions .
this work was supported by the russian foundation for basic research ( project no.07 - 02 - 01051 and 08 - 08 - 13734 ) , the program `` origin , structure and evolution of the objects in the universe '' ( p07 ) by the presidium of the russian academy of sciences and grant no.nsh-5579.2008.2 from the president of russia .
we are grateful for the data to the high energy astrophysics science archive research center online service provided by the nasa / goddard space flight center .
|
analysis of the data obtained with the rxte observatory during a powerful outburst of the x - ray pulsar v0332 + 53 in 20042005 is presented .
observational data covering the outburst brightening phase are analysed in detail for the first time .
a comparison of source parameters and their evolution during the brightening and fading phases shows no evidence for any hysteresis behaviour .
it is found that the dependences of the energy of the cyclotron absorption line on the luminosity during the brightening and fading phases are almost identical .
the complete data sequence including the outburst brightening and fading phases makes it possible to impose the more stringent constraints on the magnetic field in the source .
the pulse profile and pulsed fraction are studied as functions of the luminosity and photon energy .
[ firstpage ] x - ray : binaries ( stars:)pulsars : individual v0332 + 53
|
state determination for a physical system hinges on being able to non - invasively measure its relevant parameters . however , a measurement performed on a quantum system is by definition invasive , and hence quantum state estimation relies on measurements made on an ensemble of identically prepared systems .
quantum state estimation process is thus intrinsically statistical in nature , with its accompanying ambiguities and uncertainties .
there is no direct measurement possible for the quantum state of a single system and for the estimation of a state , we are required to determine the expectation values of a set of incompatible observables .
the accuracy of such a determination depends upon the size of the ensemble and ideally we need an infinite size ensemble to obtain the precise values of these expectations . however , if we are given a small ensemble with a fixed number of identically prepared states , how can we effectively extract information from it ?
the wave function collapse associated with projective measurements limits us from using each member of the ensemble more than once .
the problem of state estimation has been investigated by physicists since the inception of quantum information theory @xcite .
apart from the direct use of projective measurements , there exist other methods of state tomography which try to extract information from a system in different ways .
some prescriptions use the information gained about the system from one measurement to decide on the next measurement @xcite .
others employ maximum likelihood technique and numerical optimization @xcite , or use repeated weak measurements on a single system @xcite , to maximize the information gain . weak or unsharp measurements potentially hold promise for state estimation because of their non - invasiveness which allows state recycling . on the one hand
the disturbance caused by a weak measurement is less , however on the other hand it also gives us limited information .
the challenge therefore is to find a balance i.e. an intermediate regime of weakness , which leads to optimal information gain .
the idea has been explored recently in the context of a qubit with a small ensemble size @xcite .
weak measurements coupled with postselection have also been employed in the problem of state estimation where a complete characterization of the postselected quantum statistics @xcite or the direct measurement of the quantum wavefunction is used @xcite .
state measurement schemes based on weak measurement tomography have also been recently proposed @xcite .
there have also been critial analysis of these schemes @xcite . for continuous variable systems , quasi - probability distributions including the wigner distributions
can be tomographed by measuring the rotated quadrature components @xcite .
there are homodyne and heterodyne schemes for estimation of squeezed gaussian states @xcite and squeezed thermal gaussian states @xcite .
schemes using a single photon detector instead of a homodyne detector to characterize gaussian states have been proposed @xcite .
the possible advantage offered by an entangled gaussian probe to estimate the displacement of a continuous variable state has also been explored @xcite .
the importance of maximum likelihood methods has been emphasized @xcite and state reconstruction has been described @xcite . in a different direction , arthurs and kelly aimed to simultaneously measure position and momentum of a general wavefunction by coupling two different apparatuses with the system @xcite .
since @xmath0 and @xmath1 are noncommuting observables , this leads to an unsharp measurement .
symplectic tomography has been used to estimate the master equation parameters in an open setting for a single mode system @xcite .
alternatively , one can reuse each member of the ensemble if the first measurement is done weakly enough such that the disturbance induced is very small @xcite .
a similar idea has been utilized , albeit in a different direction , in the construction of loophole free hybrid bell - leggett- garg inequalities @xcite .
weak or unsharp measurements are performed by weakly coupling the device to the quantum system @xcite .
although the noise produced in such measurements is small , which should serve our purpose well , the information obtained is also very low .
therefore , there is a tradeoff between the disturbance and information gain . to effectively use weak measurements for state estimation , we need to optimize the process . in this work
we restrict ourselves to the realm of a special class of states of one continuous variable quantum systems called gaussian states .
these are states with gaussian - wigner quasiprobability distributions and include coherent states , squeezed states and thermal squeezed states .
the gaussian states are determined by the first and second moments of position and momentum .
we explore the advantage of the scheme involving weak measurements in estimating the gaussian states over projective measurements .
we show that with an optimal strength of the weak measurement , our technique is more powerful in the determination of the wigner quasiprobability distribution when the ensemble size available is small .
we have chosen the meter state in the form of a minimum uncertainty squeezed gaussian state and the tuning of the weakness of a measurement is achieved by changing the squeezing of the position quadrature .
this scheme is tested for average performance over a large number of states and over a large number of runs to kill statistical fluctuations .
we also take gaussian states at different temperatures and check whether the efficacy of our method depends in any way on temperature .
the paper is arranged as follows : in section [ tools ] we collate all the background material necessary for the problem .
we describe continuous variable states and wigner s quasiprobability distributions in brief .
symplectic methods are described briefly .
section [ scheme ] gives a description of weak measurement in quantum mechanics when applied to gaussian states . in section [ tomo ]
we describe how to perform tomography of gaussian states using our method .
we provide conclusions in section [ conc ] .
let us consider a one - dimensional quantum system with the position and momentum operators @xmath2 and @xmath3 , satisfying the commutation relation @xmath4=\iota \quad \rm{(}\hbar=1\rm { ) } \label{commutation}\ ] ] the corresponding eigenkets are @xmath5 and @xmath6 for a complete set @xcite , and are defined as @xmath7 an arbitrary mixed density operator in position and momentum bases can be represented by @xmath8 where @xmath9 is a function of real variables @xmath0 and @xmath10 while @xmath11 is a function of the real variables @xmath1 and @xmath12 .
an alternative and equivalent way of representing the system state @xcite is via its wigner distribution @xmath13 given by @xmath14 the probability distributions corresponding to position and momentum can be obtained by computing the marginals @xmath15 and can be used to calculate expectation values of arbitrary observables via the symmetric ordering rule @xcite .
the second order moments of position and momentum corresponding to a quantum state are given by @xmath16 and they obey the schrdinger uncertainty principle given by @xmath17\rangle\vert^2\ ] ] a compact way to represent the second order moments is via the variance matrix @xmath18 given by @xmath19 and the uncertainty condition re - expressed in terms of the variance matrix takes the elegant form @xcite @xmath20 the subclass of states for which the wigner distribution is a gaussian function are called gaussian states and play an important role in quantum optics and quantum information @xcite .
all states with gaussian wave functions are gaussian , however the class of gaussian states is a much bigger class and includes mixed states .
the wigner representation corresponding to a general gaussian state centered at the origin of the phase space is given by , @xmath21 where @xmath22 the matrix @xmath23 is related to the variance matrix @xmath18 as @xmath24 if the center of the gaussian is located at a point @xmath25 , this can be achieved by action of a displacement operator @xmath26 which acts on canonical operators as @xmath27 and leads to a point transformation of the winger function . for the gaussian - wigner function
this amounts to shifting the center to the location @xmath25 .
the matrix @xmath23 can be written as : @xmath28 where @xmath29 is a diagonal symplectic matrix belonging to the group @xmath30 , @xmath31 is a rotation matrix and @xmath32 is proportional to identity .
a total of three real parameters are involved in describing @xmath23 .
we further restrict ourselves to a special class of gaussians where @xmath23 is described by two parameters , namely , the temperature and squeezing @xcite . setting @xmath33 with @xmath34 representing the temperature @xmath35 , @xmath36 being the boltzmann s constant and @xmath37 having the units of frequency .
the corresponding variance matrix in this case is diagonal @xmath38 at @xmath39 or @xmath40 if we have @xmath41 , the gaussian state is a coherent state .
a coherent state is represented by a circle of radius @xmath42 in phase space .
now at @xmath43 , if @xmath44 and @xmath45 happen to be unequal , the state is a squeezed state .
such a gaussian state is represented by an ellipse .
the center of a general coherent or a general squeezed state may not be at the origin of phase space and in such cases it is said to be a displaced coherent or a displaced squeezed state .
the displacement of a wigner function @xmath46 , centered at the origin is achieved by means of a displacement operator @xmath26 which takes the center to @xmath25 . for a non - zero temperature ,
the value of the product @xmath47 is greater than @xmath48 ( for coherent states it is equal to half ) and it increases with a rise of temperature .
pictorially , the class of states can be represented by an ellipse in phase space with center at @xmath25 and semimajor axis oriented along either @xmath0 or @xmath1 and is depicted in figure [ spread ] . and with spreads @xmath44 and @xmath45 .
the estimated state is represented by another ellipse , bounded by a broken line , centered at @xmath49 and has spreads @xmath50 and @xmath51 . [ spread ] ]
consider a system represented by a displaced gaussian - wigner distribution function represented by @xmath52 and characterized by two real displacement parameters : one temperature parameter and one squeezing parameter .
such a wigner function can be obtained by starting with the centered gaussian - wigner function given in equation ( [ gaucen ] ) with parameters chosen as per equations ( [ theg ] ) and ( [ squeezing ] ) and applying a displacement operator @xmath26 as described in the previous section .
given that the system is in such a state , our goal is to estimate the state . for the purposes of measurement , consider a meter which is a macroscopic pointer with position and momentum variables @xmath53 and @xmath54 .
the meter is chosen to be in a squeezed coherent state represented by a wigner distribution @xmath55 such that @xmath56 and @xmath57 . here
we take temperature to be zero .
as we shall see , when we employ this meter to measure the position , the strength of the measurement can be varied by changing the squeezing of the meter along the position quadrature .
the larger values of the squeezing parameter correspond to a stronger measurement while the smaller values of the squeezing parameter correspond to a weaker measurement .
similarly , if we are measuring momentum one can tune the measurement strength by varying the value of squeezing of the variable @xmath58 .
the system and the meter form a composite system and the joint wigner function representing this two degrees of freedom system can be obtained by multiplying the two individual wigner functions .
for such a system it is natural to define a four dimensional column vector of phase space variables as @xmath59 the phase space displacement of the system variables given in equation ( [ displacement ] ) acts on this column vector to give us a displaced vector @xmath60 in terms of the above column vector the joint wigner function becomes @xmath61 the matrix @xmath23 is a diagonal @xmath62 matrix @xmath63 @xmath64}}{4\pi^2 \delta q \delta p \delta q_{m_1 } \delta p_{m_1}}\ ] ] when we perform a measurement ( weak or strong ) of the position @xmath0 , we switch on the following interaction hamiltonian between the system and the meter degrees of freedom @xmath65 the corresponding unitary transformation on the composite system - meter is @xmath66 in the language of the wigner quasi - probability distribution a unitary operation @xmath31 is equivalent to a symplectic transformation @xmath67 @xmath68 satisfying @xmath69 the symplectic transformation corresponding to the interaction hamiltonian acts on the phase variables by multiplication @xmath70 leading to the computation of the transformed wigner function under this symplectic transformation @xmath71 } } { 4\pi^2 \delta q \delta p \delta q_{m_1 } \delta p_{m_1}}.\ ] ] the wigner function of the meter after the above interaction is obtained by integrating over the system variables @xmath0 and @xmath1 and is given by @xmath72}}{2\pi\delta q_{m_1 } \delta p_{m_1}\delta q \sqrt{\frac{1}{(\delta q_{m_1})^2}+\frac{1}{(\delta q)^2}}}. \label{wigsymp1}\ ] ] we can see at this point that the state of the meter has become correlated with the state of the system .
however , as the meter is a macroscopic entity , its very observation leads to the collapse of its wavefunction and gives us a definite value .
thus the probability density for the meter to show a reading @xmath53 @xmath73 } } { \sqrt{2\pi } \delta q_{m_1 } \delta q\sqrt{\frac{1}{(\delta q_{m_1})^2 } + \frac{1}{(\delta q)^2}}}. \label{probweak1}\ ] ] on the other hand , the reduced state of the system after the measurement interaction represented by the symplectic transformation is obtained by integrating over the meter degrees of freedom leading to the wigner function for the system @xmath74 } } { 2 \pi \delta p_{m_1 } \delta p \delta q \sqrt{\frac{1}{(\delta p_{m_1})^2}+\frac{1}{(\delta p)^2 } } } \label{wigtrans}\ ] ] in the weak measurement limit @xmath75 is large ( i.e. the initial meter state is prepared in distributions wide in position ) .
since we have chosen the meter to be in a squeezed coherent state , this corresponds to a high degree of squeezing in the momentum quadrature of the initial meter state . in this limit
we have @xmath76 hence , weak measurement causes controllable disturbance to the state and the the disturbance vanishes in the limit of extremely weak measurement .
however , if we make the measurement too weak , the correlation between the meter state and the system state diminishes . in the limit of extremely weak measurement , where no disturbance is caused
, we do not learn anything about the system from observing the meter . in our scheme ,
the first measurement that we perform is a weak measurement of position @xmath0 with a tunable strength as described above .
subsequently , we carry out a projective measurement of momentum @xmath1 is on this system , then the probability density for obtaining any momentum as obtained from the modified system wigner function given in equation ( [ wigtrans ] ) is given by , @xmath77 } } { 2 \pi \delta p_{m_1 } \delta p \sqrt{\frac{1}{(\delta p_{m_1})^2}+\frac{1}{(\delta p)^2 } } } \label{probstrong1}\ ] ] in the reverse scenario where we do a weak measurement of momentum @xmath1 followed by a projective measurement of position @xmath0 , the composite system - meter system wigner function after the measurement interaction given by the hamiltonian @xmath78 is given by @xmath79}}{4 \pi ^2\delta p_{m_2 } \delta p \delta q_{m_2 } \delta q } \label{wigsymp2}\ ] ] where @xmath58 and @xmath80 denote the position and momentum coordinates of the meter measuring momentum @xmath1 of the system .
the wigner of the meter alone is given by @xmath81}}{2 \pi \delta p_{m_2 } \delta p \delta q_{m_2 } \sqrt{\frac{1}{(\delta p)^2}+\frac{1}{(\delta q_{m_2})^2}}}\ ] ] giving the probability density of the meter to show a reading @xmath58 being @xmath82 } } { \sqrt{2 \pi } \delta p \delta q_{m_2 } \sqrt{\frac{1}{(\delta p)^2}+\frac{1}{(\delta q_{m_2})^2 } } } \label{probweak2}\ ] ] the corresponding system wigner function becomes @xmath83}}{2 \pi \delta p_{m_2 } \delta p \delta q \sqrt{\frac{1}{(\delta p_{m_2})^2}+\frac{1}{(\delta q)^2}}}\ ] ] as before , in the weak measurement limit the disturbance caused in the system is limited and we have @xmath84 on this state we perform a projective measurement of position @xmath0 giving us the probability density for getting a result @xmath0 @xmath85}}{\sqrt{2 \pi } \delta p_{m_2 } \delta q \sqrt{\frac{1}{(\delta p_{m_2})^2}+\frac{1}{(\delta q)^2}}}. \label{probstrong2}\ ] ]
into two parts . for one half of the ensemble we measure position @xmath0
weakly ( the weakness being defined by the initial spread in position @xmath75 of the meter ) leading to a disturbed ensemble @xmath86 . on every member of the ensemble @xmath86
we carry out a projective measurement of momentum @xmath1 . with the other half of the initial ensemble @xmath87 , momentum @xmath1 is measured weakly leading to a disturbed ensemble @xmath88 on which a projective measurement of @xmath0 is carried out .
[ flowchart ] ] in order to perform complete state tomography of any gaussian state of the form discussed earlier , we are required to estimate the center of the gaussian wigner function , @xmath89 and the spreads @xmath44 and @xmath45 .
hence it is necessary to measure both position @xmath0 and momentum @xmath1 of the system as accurately as possible . to this end
, we divide the initial ensemble of identically prepared systems into two equal parts . on every member of one part
we perform a weak measurement of position @xmath0 .
the strength of the measurement is governed by the initial squeezing of the position quadrature of the meter determining the initial variance @xmath90 of the meter state .
the larger the value of @xmath75 , weaker is the measurement strength and vice versa .
the meter reading is recorded in each case and the final states of all the members are collected to generate a second ensemble .
the members of this ensemble are now used as the initial states of a second measurement , which is a projective measurement of momentum @xmath1 . as before the meter readings of this measurement are noted .
now the process is repeated with the members of the second part of the initial ensemble where we first measure @xmath1 weakly , with the strength of the measurement determined through @xmath91 and then carry out a projective measurement of @xmath0 . in all further analysis and discussions
we take @xmath92 .
a summary of the procedure is illustrated in figure [ flowchart ] .
the entire algorithm is repeated over many runs to rule out statistical fluctuations .
it is worth noting that the initial squeezing of the relevant quadrature which determines the strength of the measurement is a tunable parameter in our hand .
although we call certain measurements `` weak '' , we actually mean that it is not too strong to be projective and not too feeble to induce large errors to the measurement outcomes .
the main point is that the measurements are weak enough and do not cause the complete collapse of the state so that it can be used for subsequent measurements .
the expectation values obtained from the @xmath0 and @xmath1 measurements are used to estimate the values of @xmath89 , and the spreads @xmath44 and @xmath45 .
looking at the equations ( [ wigsymp1 ] ) and ( [ wigsymp2 ] ) reveals that the information about the system has flowed into the meter . in fact the meter is now centered over @xmath25 which is the center of the initial system state .
we carry out simulations using the meter reading probabilities given in equations ( [ probweak1 ] ) , ( [ probstrong2 ] ) , ( [ probweak2 ] ) and ( [ probstrong2 ] ) .
we take different ensemble sizes of member numbers @xmath93 , @xmath94 , @xmath95 and @xmath96 respectively with randomly generated gaussian states .
each virtual experiment is repeated over 1000 runs .
the quantities @xmath97 and @xmath98 for a state are estimated by taking the mean over the @xmath0 and @xmath1 measurements while @xmath44 and @xmath45 are estimated from the corresponding variances .
the order of measurement of @xmath0 and @xmath1 is reversed for the second part of the ensemble to rule out the possibility of preferential treatment of any of the observables . in the scheme involving projective measurements
only , we divide the original ensemble into two parts and perform @xmath0 and @xmath1 measurements independently on the individual members of these parts .
no sequential measurements are possible here because of the wavefunction collapses after the measurement . the accuracy of the state estimate is measured via the following distance measures @xmath99 where @xmath100 , @xmath101 , @xmath50 and @xmath51 are the estimated values of @xmath97 , @xmath98 , @xmath44 and @xmath45 , respectively .
the parameter @xmath102 is a measure of how well our method is able to estimate the center of the gaussian and @xmath103 gives a measure of how well the spreads of the gaussian have been estimated .
the two measures @xmath102 and @xmath103 represent closeness in position and width of the estimated wigner distribution from the original wigner distribution respectively .
we can immediately see that the lower these distances , the better the estimates .
for a perfect estimate the values should go to zero .
to study the average performance of our scheme for squeezed displaced thermal states , we begin by numerically generating 100 gaussian states at a particular temperature , with randomly chosen values of displacement and squeezing . to generate these states ,
the value of the parameter @xmath104 in equation ( [ squeezing ] ) is varied between @xmath105 and @xmath106 according to a uniform distribution .
similarly , the centers of the gaussians are also chosen randomly using uniform distributions between @xmath107 and @xmath108 for both @xmath97 and @xmath98 . with each of these 100 random states
, we numerically carry out the prescription given in subsection [ pres ] on a fixed number of identical copies of the state determining the ensemble size .
the simulation is carried out with the help of the results obtained in section [ scheme ] . the distance measures @xmath102 and
@xmath103 used to compare the efficacy of our method with projective measurements are computed .
each experiment involving one gaussian state is repeated 1000 times to reduce statistical fluctuations .
the process is carried out with ensembles of sizes @xmath96 , @xmath95 , @xmath94 and @xmath93 . for a given ensemble size
, the results for each member are averaged over 1000 runs and then the distance measures are averaged over the 100 states .
we show that there is a clear advantage of using our scheme when the ensemble size is small .
the test is carried out for three different sets of gaussian states corresponding to three different temperatures given by @xmath40 , @xmath109 and @xmath110 , respectively .
using averages over 100 random states further averaged over 1000 runs .
the behaviors of @xmath102 and @xmath103 are plotted with @xmath111 for ensemble sizes @xmath112 and @xmath113 .
the corresponding projective measurement results are plotted as dotted lines .
while the method performs well in estimating the position of the gaussian states for all ensemble sizes , as represented by @xmath102 , it provides a clear advantage for estimating the spreads represented by @xmath103 over projective measurements in the case of a small ensemble of size @xmath93 .
[ k_1 ] ] using averages over 100 random states further averaged over 1000 runs .
the behaviors of @xmath102 and @xmath103 are plotted with @xmath111 for ensemble sizes @xmath112 and @xmath113 .
the corresponding projective measurement results are plotted as dotted lines .
while the method performs well in estimating the position of the gaussian states for all ensemble sizes , as represented by @xmath102 , it provides a clear advantage for estimating the spreads represented by @xmath103 over projective measurements in the case of a small ensemble of size @xmath93 .
[ k_0.9 ] ] using averages over 100 random states further averaged over 1000 runs .
the behaviors of @xmath102 and @xmath103 are plotted with @xmath111 for ensemble sizes @xmath112 and @xmath113 .
the corresponding projective measurement results are plotted as dotted lines .
while the method performs well in estimating the position of the gaussian states for all ensemble sizes , as represented by @xmath102 , it provides a clear advantage for estimating the spreads represented by @xmath103 over projective measurements in the case of a small ensemble of size @xmath93 .
[ k_0.8 ] ] the performance of state estimation of gaussian states via our weak measurement protocol is compared to the corresponding performance of projective measurements .
this is done via plots of the distance measures @xmath102 and @xmath103 vs weakness parameter defined by the inverse of squeezing @xmath114 of the meter state , averaged over 100 such random states .
the process is carried out for four different small ensemble sizes @xmath115 and @xmath116 and three different absolute temperatures given by @xmath117 and @xmath118 .
let us first look at figure [ k_1 ] ( a ) .
in this case the distance measures @xmath102 and @xmath103 have been plotted with @xmath119 for an absolute temperature given by @xmath40 and ensemble size @xmath96 .
a low value of @xmath119 indicates the meter prepared as a wide gaussian in the position space .
this corresponds to the weak measurement limit .
a very weak measurement introduces a large amount of error in the measurement and this leads to a low quality of state estimation .
this can be seen from the fact that the values of both @xmath102 and @xmath103 , on the left hand side of the plot for the weak measurement method are much higher than those involving only projective measurements represented by the dotted line .
similarly , on the right side of the plot , the meter is prepared as a narrow gaussian .
the corresponding measurement limit for this side of the plot is that of strong projective measurements .
projective measurements destroy the state of the system and hence using the state for the second time leads to a low quality of state estimation .
only for an intermediate value of weakness , our method performs better than projective measurements .
this is seen from the plot of @xmath102 going below the dotted line representing the same distance measure for the projective measurement .
the plot of @xmath103 attains its minimum for the intermediate values of @xmath119 but remains above the dotted line .
it indicates that though our method has worked in giving a better estimation of the position of the gaussian state , it does not perform as well to provide an estimation of the spreads of the gaussian , in this particular case .
figure [ k_1 ] ( b ) shows the plot of the same parameters for the same absolute temperature but for a lower ensemble size of @xmath95 .
we find that here our method proves to be more effective than the projective measurements both for the estimations of the position and the spread of the gaussian wigner function .
moving on to figure [ k_1](c ) and ( d ) which are for the ensembles of sizes @xmath94 and @xmath93 we find that the relative efficacy of the estimation for position as well as the spread improves .
we repeat the same exercise with gaussian states with finite temperatures with @xmath120 and @xmath118 as indicated in figures [ k_0.9 ] and [ k_0.8 ] , respectively .
we observe the same trend as observed for the zero temperature in all these cases .
our method is not too effective in the extremely weak or extremely strong regimes .
it works in the intermediate regimes depending upon the size of the ensemble and its efficacy increases with the lowering of the ensemble size . in each of the plots
, it is observed that the distance measures attain small values for an optimal value of squeezing .
this is expected , as a very large value of squeezing ushers in too many errors into the `` weak measurement '' , while a small value causes a larger disturbance to the original state .
we observe from figures [ k_1 ] , [ k_0.9 ] and [ k_0.8 ] that for an optimal range of @xmath111 values , the weak distance measure curves go below the projective measurement line ( represented by broken straight lines ) . in this regime of @xmath111 values ,
our method is more effective than the projective measurement state estimation . the advantage is greater for smaller ensemble size .
in fact for the ensemble size of @xmath96 and @xmath40 , the performances of the optimal weak measurement method and projective measurement are almost equal as can be seen in figure [ k_1 ] .
however , as the ensemble size decreases , a clear advantage emerges for the proposed scheme .
there is no particular change in the advantage of our scheme relative to projective measurements , with change of temperature as is evident from plots with different temperature parameter @xmath121 .
in this paper , we have described our work on the estimation of gaussian states by a method employing weak or unsharp measurements .
we use phase space methods and the language of wigner distributions for state estimation .
we compare our results with state estimation based on projective measurements and show how one can do better in certain parameter regimes .
recycling of states , where one makes more than one measurement on a single copy before discarding it and tenability of the strength of the weak measurement are the two main ingredients of our scheme .
the strength of the measurement is directly related to the amount of squeezing in the initial pointer state and can be tuned at will and we optimize the performance of our scheme with respect to this weakness parameter .
the efficacy of the scheme is tested over a randomly chosen subset of gaussian states .
we demonstrate that the weak measurement based scheme produces a wigner distribution which is much closer to the original wigner distribution as compared to the scheme based on projective measurements , for small ensemble sizes . as the ensemble size increases , the relative advantage of our scheme decreases , as seen in the comparative results for varying ensemble sizes .
the behavior is repeated over the range of temperatures we have considered .
while in this work we have dealt with gaussian states with the maximum spread along the @xmath0 or @xmath1 axes it will be interesting to extend the scheme to general gaussian and non - gaussian states .
another interesting direction that we are following up is to compare our results with schemes similar to the arthurs and kelly setup where position and momentum are measured together .
39ifxundefined [ 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 _ _ ( , ) pp .
link:\doibase 10.1103/physreva.61.032306 [ * * , ( ) ] link:\doibase 10.1103/physreva.64.052312 [ * * , ( ) ] link:\doibase 10.1002/prop.200310009 [ * * , ( ) ] link:\doibase 10.1103/physreva.89.062121 [ * * , ( ) ] link:\doibase 10.18520/v109/i11/1939 - 1945 [ * * , ( ) ] link:\doibase 10.1103/physreva.81.012103 [ * * , ( ) ] link:\doibase 10.1038/nature10120 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.070402 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1038/srep01193 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/96/40002 [ * * , ( ) ] link:\doibase 10.1103/physreva.92.062133 [ * * , ( ) ] link:\doibase 10.1103/physreva.40.2847 [ * * , ( ) ] link:\doibase 10.1038/srep12289 [ * * , ( ) ] link:\doibase 10.1103/physreva.79.033834 [ * * , ( ) ] link:\doibase 10.1016/s0079 - 6638(08)70389 - 5 [ * * , ( ) ] link:\doibase 10.1103/physreva.70.053812 [ * * , ( ) ] link:\doibase 10.1103/physreva.87.012107 [ * * , ( ) ] link:\doibase 10.1103/physreva.55.r1561 [ * * , ( ) ] link:\doibase 10.1103/physreva.61.010304 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.81.299 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.220502 [ * * , ( ) ] link:\doibase 10.1002/j.1538 - 7305.1965.tb01684.x [ * * , ( ) ] link:\doibase 10.1103/physreva.80.052108 [ * * , ( ) ] link:\doibase 10.1119/1.1475328 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.040402 [ * * , ( ) ] link:\doibase 10.1103/physreva.89.012125 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.60.1351 [ * * , ( ) ] link:\doibase 10.1103/physrevd.40.2112 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.044103 [ * * , ( ) ] link:\doibase 10.1103/physrev.40.749 [ * * , ( ) ] link:\doibase 10.1007/bf02848172 [ * * , ( ) ] link:\doibase 10.1016/s0375 - 9601(99)00471 - 5 [ * * , ( ) ] link:\doibase 10.1140/epjst / e2012 - 01532 - 4 [ * * , ( ) ] link:\doibase 1983/11/10/print [ * * , ( ) ] link:\doibase 10.1119/1.2957889 [ * * , ( ) ]
|
we present a scheme to estimate gaussian states of one - dimensional continuous variable systems , based on weak ( unsharp ) quantum measurements .
the estimation of a gaussian state requires us to find position ( @xmath0 ) , momentum ( @xmath1 ) and their second order moments .
we measure @xmath0 weakly and follow it up with a projective measurement of @xmath1 on half of the ensemble , and on the other half we measure @xmath1 weakly followed by a projective measurement of @xmath0 . in each case
we use the state twice before discarding it .
we compare our results with projective measurements and demonstrate that under certain conditions such weak measurement - based estimation schemes , where recycling of the states is possible , can outperform projective measurement - based state estimation schemes .
|
most iterative decoders , e.g. turbo decoders @xcite , rely on knowledge of the signal - to - noise ratio ( snr ) or the channel reliability constant @xcite .
the snr is also required for other functionalities in the receiver .
many snr estimators have been proposed , both data - aided ( da ) that require pilot symbols or feedback from the decoders @xcite , and non - data - aided ( nda ) that are only based on the received observables @xcite .
a comparison of both da and nda snr estimators was performed in @xcite and compared to the cramr - rao lower bound ( crlb ) for da estimators .
the crlb for nda estimators was later derived in @xcite . the nda maximum likelihood ( ml )
estimator based on the expectation maximization ( em ) algorithm was proposed in @xcite and also compared to other nda estimators .
this nda ml estimator was found iteratively , but unfortunately requiring processing of all observables for each iteration , making it computationally complex .
the contributions in this paper are as follows .
to complement the nda crlb for snr in @xcite , we derive the nda crlb for the signal amplitude , the noise variance , the channel reliability constant , and the bit - error rate ( ber ) .
it is also shown how to estimate the _ a priori _ probability of the transmitted symbols , in the case when they are not equally likely .
furthermore , we provide a more direct , alternative derivation of the nda ml estimator and we propose a new , low complexity nda snr estimator .
the performance of the new estimator is compared to previously suggested nda estimators and found to be similar to that of the nda ml estimator .
this performance is achieved with significantly lower computational complexity than the ml estimator . only binary - phase - shift - keying ( bpsk ) transmission is considered here , but generalization to @xmath0-psk is straightforward @xcite .
let @xmath1 denote a binary random variable with equally likely symbols .
further , let @xmath2 represent a zero - mean gaussian random variable with unit variance . define a new random variable @xmath3 according to @xmath4 with probability density function ( pdf ) expressed as @xcite @xmath5 let @xmath6 , @xmath7 , and @xmath8 denote samples from @xmath9 , @xmath3 , and @xmath2 , respectively .
@xmath10 independent samples of @xmath3 is observed and collected in a column vector @xmath11}{^t}$ ] . if @xmath12 and @xmath13 , the model in represents bpsk transmission in additive white gaussian noise ( awgn ) @xmath14 where @xmath15 is the matched filter output , @xmath16}{^t}$ ] the transmitted data , and @xmath17}{^t}$ ] white gaussian noise .
@xmath18 is the transmitted energy and @xmath19 is the double - sided noise power spectral density .
define the snr as @xmath20 since all samples in @xmath15 are assumed independent , the logarithm of their joint pdf is given by @xcite @xmath21 the average ber can be expressed as @xmath22 where @xmath23 is the gaussian @xmath24-function .
the average mutual information ( mi ) @xcite between @xmath9 and @xmath3 in can be expressed as @xmath25 where @xmath26 is defined as @xcite @xmath27 the log - likelihood ratio ( llr ) for @xmath28 is defined as @xcite @xmath29 where @xmath30 is the channel reliability constant @xcite @xmath31 the instantaneous ber for a specific received symbol at position @xmath32 can be estimated by @xcite @xmath33 the corresponding instantaneous mi for a symbol at position @xmath32 can be estimated by @xcite @xmath34 with no knowledge of the transmitted symbols the average ber in or the average mi in depend solely on the snr , @xmath35 . also , in order to use an llr - based decoder ( basically all turbo - like decoders @xcite or soft decoders ) , to estimate the instantaneous ber in , or to estimate the instantaneous mi in , the channel reliability constant in needs to be known . however , as we show in section [ sec_estimators ] , the snr and the channel reliability constant are related through the second moment of the observables .
we therefore only need to estimate one of the two . here , we have chosen to estimate the snr , @xmath36 .
the crlb , here denoted by @xmath37 , provides a lower bound on the variance of any unbiased estimator @xcite .
let @xmath38 represent an arbitrary function of the parameters @xmath39 and @xmath40 , and define @xmath41 .
the normalized crlb ( ncrlb ) for @xmath42 can then be calculated as @xcite @xmath43 { \textbf{j}^{-\!1}}\left [ \begin{array}{cc } \!\!\!{\frac{\partial { g(\mu,{\sigma})}}{\partial \mu } } & \!{\frac{\partial { g(\mu,{\sigma})}}{\partial { \sigma } } } \\
\end{array } \!\!\!\right ] { ^t}\ ! , \label{eq_crlbd}\end{aligned}\ ] ] where @xmath44 is the fisher information matrix @xcite , defined as @xmath45 .
\label{eq_fm}\end{aligned}\ ] ] here @xmath46 denotes the expectation over @xmath3 .
a similar fisher information matrix as in has been derived in @xcite and the inverse can be expressed as @xmath47 , \label{eq_fminv}\end{aligned}\ ] ] where @xmath48 is a function of @xmath35 @xcite @xmath49 using , the crlb for @xmath35 can be calculated as reported in @xcite and @xcite .
the ncrlbs for @xmath39 , @xmath40 , @xmath35 , @xmath30 , and @xmath50 are @xmath51 the ncrlb for @xmath52 can also be found in a similar way by replacing @xmath50 in with @xmath53 .
unfortunately , there is no simple form to express @xmath54 .
note that @xmath55 for da estimation is found by setting @xmath56 in @xcite .
this implies that the ncrlbs for da estimation of @xmath39 , @xmath40 , @xmath35 , @xmath30 , and @xmath50 are easily found by letting @xmath56 in .
define moment @xmath57 of @xmath3 as @xmath58 which can be approximated by its sample average @xcite .
the second moment of @xmath3 is @xmath59 assume that an estimate of @xmath35 exists , denoted by @xmath60 . combining with and gives the following estimators for @xmath39 , @xmath40 , and @xmath30 @xmath61 the next sub - sections present different estimators for @xmath35 that can be used to estimate the above parameters .
the absolute moment ( am ) of @xmath3 is defined as @xcite @xmath62 and can also be approximated by the sample average @xcite @xmath63 for large @xmath39 or small @xmath40 , the am will tend to @xmath39 @xmath64 in other words , for high values of @xmath35 , @xmath60 can be closely approximated by @xmath65 using and .
this estimator was first introduced in @xcite and will here be referred to as the conventional method ( cm ) estimate . the ml estimator maximizes the joint pdf in . taking the partial derivatives of @xmath66
gives @xmath67 setting the derivatives in and to zero and solve for @xmath68 gives @xmath69 inserting in gives an expression depending only on @xmath15 , @xmath39 and @xmath70 , which can be solved iteratively by @xmath71 where @xmath72 denotes the estimate of @xmath39 after @xmath57 iterations .
the iteration in is identical to the iteration in the em algorithm presented in @xcite , but here derived in a different way .
a good starting point for the iterative estimator is the cm estimate of @xmath39 , @xmath73 .
after @xmath74 iterations the snr can be estimated by @xmath75 which will be referred to as the ml estimator .
the approach of estimating a parameter based on the moments of the observables is known as the method of moments ( mm ) @xcite .
the fourth moment of @xmath3 is @xcite @xmath76 combining with gives the mm estimator , e.g. @xcite , @xmath77 where @xmath70 and @xmath78 are approximated by the sample average .
if @xmath79 , is no longer real and @xmath60 is set to zero .
this will be referred to as the mm estimator .
an estimator based on the second moment and the am can be found by combining with .
unfortunately , there is no closed - form analytical solution for @xmath39 and @xmath40 as for the mm estimator .
however , dividing the square of @xmath80 by @xmath70 gives an expression that only depends on @xmath35 , @xmath81 an estimator for @xmath35 can therefore be stated as @xmath82 since there is no closed - form solution to , alternative methods must be explored . in @xcite , a table - lookup for @xmath83 is suggested .
a different approach is to approximate @xmath83 with a simple closed - form function , which was done in @xcite as , @xmath84 the estimator in , using the approximation in is referred to as the second - order polynomial ( p2 ) estimator . from it is easy to verify that @xmath85 therefore , we suggest the following approximation of @xmath86 and its inverse @xmath87 numerical optimization , using the nelder - mead simplex method @xcite to minimize the mean squared difference between and gives @xmath88 , @xmath89 , and @xmath90 . the estimator in , using the novel approximation in , with @xmath91 whenever @xmath92 , is a new approach we propose and is here referred to as the am estimator .
( 94,84)(-1,-2 ) ( -3,0 ) .,title="fig:",width=321 ] ( -2,46)(0,0)[r ] ( 49,-1)(0,0)[t]@xmath93 fig .
[ fig_h_approx ] shows the analytical expression from together with its indistinguishable approximation in .
since the cm estimator in depends only on @xmath93 it is also shown in fig .
[ fig_h_approx ] .
it is clear that the function used by the cm estimator converges to the analytical one for large @xmath35 , but differs for small @xmath35 .
the same figure also shows the approximation in .
since this polynomial approximation was only optimized between @xmath94 to @xmath95 db ( @xmath96@xmath97 ) @xcite it differs from the analytical expression outside this region .
define @xmath98 to be the _ a priori _ probability of @xmath9 in @xmath99 the ml estimator is invariant to non - equiprobable symbols and gives the same results even if @xmath100 @xcite .
it is straightforward to show that @xmath80 , @xmath70 , and @xmath78 are independent of @xmath98 @xcite .
since the cm , the mm , the p2 , and the am estimators are based only on these quantities , they will give the same results independent of @xmath98 . however , when @xmath101 , the odd moments are non - zero , @xmath102 this means that the _ a priori _
probability @xmath98 can be estimated using @xmath103 , by combining with and @xmath104
the performance of the snr estimators is evaluated based on their normalized mean squared error ( nmse ) @xmath105 and their normalized bias ( nb ) @xmath106 where @xmath107 is estimated based on @xmath10 samples .
the number of trials was chosen to @xmath108 .
the best estimator is an unbiased estimator with minimum nmse @xcite .
[ fig_nmse_g_gs ] shows the nmse and fig .
[ fig_nb_g_gs ] shows the nb , both for @xmath109 observables .
the nda and the da ncrlb are also included as a reference , even though they are only bounds for unbiased estimators . all the estimators presented here are biased when @xmath109 , even for high @xmath35 which is evident from fig .
[ fig_nb_g_gs ] .
different approaches to reduce the bias has been suggested , e.g. , @xcite .
the cm estimator has a large nb ( and therefore also a large nmse ) for low @xmath35 .
for large @xmath35 the cm estimator approaches the ml estimator , which was shown analytically in @xcite .
in fact , figs . show that all estimators , except the p2 estimator converges to the same constant nmse and constant nb for high @xmath35 ( the nb is around 5% above the true @xmath35 ) .
the p2 estimator only works well between -3 to 3 db , the interval for which it was optimized .
the mm estimator has the second highest nmse for low @xmath35 .
the ml estimator after @xmath110 iterations has the lowest nmse at -6 db , but fig .
[ fig_nb_g_gs ] shows that it at the same time has the second highest nb .
finally , the suggested am estimator has almost identical performance ( both in nmse and nb ) as the ml estimator for all @xmath35 , even though it has a computationally complexity that is less than the first ml iteration
. figs .
show the nmse and the nb for different @xmath10 at -2 db .
this corresponds to an @xmath111 around 1 db for a half - rate code , e.g. the original turbo code @xcite .
at this low snr , the cm estimator has bad performance .
the nb saturates around 60% above the true value ( not shown here ) , which gives the high nmse in fig .
[ fig_nmse_g_k ] .
the p2 estimator has a negative nb for large @xmath10 at this snr .
the ml estimator after @xmath112 iterations and the am estimator have a small positive nb ( around 1% ) for large @xmath10 .
the ml estimator and the mm estimator are the only two estimators that are unbiased for large @xmath10 , but only the ml estimator approaches the ncrlb in fig .
[ fig_nmse_g_k ] which makes it asymptotically optimal @xcite .
the second best estimator , after the ml estimator , over the whole range of @xmath10 is the suggested am estimator .
in this paper we have derived the nda ncrlb for the signal amplitude , the noise variance , the channel reliability constant , and the ber in an awgn channel with bpsk modulated transmission .
it was also shown that these parameters , as well as the _ a priori _ probability of the transmitted symbols and the instantaneous mi can all be estimated based on the snr estimate .
a novel snr estimator with low computationally complexity was introduced and shown to be surpassed in performance only by the iterative ml estimator among previously suggested estimators .
the proposed estimator performs close to the performance of the iterative ml estimator at significantly lower computationally complexity . c. berrou , a. glavieux , and p. thitimajshima , `` near shannon limit error - correcting coding and decoding : turbo - codes , '' in _ proc .
ieee int .
( icc 93 ) _ , vol . 2 ,
geneva , switzerland , may 1993 , pp .
10641070 .
p. hoeher , i. land , and u. sorger , `` log - likelihood values and monte carlo simulation some fundamental results , '' in _ proc
. int .
symp . on turbo codes and rel .
topics _ , brest , france , sept . 2000 , pp . 4346 .
|
non - data - aided ( nda ) parameter estimation is considered for binary - phase - shift - keying transmission in an additive white gaussian noise channel .
cramr - rao lower bounds ( crlbs ) for signal amplitude , noise variance , channel reliability constant and bit - error rate are derived and it is shown how these parameters relate to the signal - to - noise ratio ( snr ) .
an alternative derivation of the iterative maximum likelihood ( ml ) snr estimator is presented together with a novel , low complexity nda snr estimator .
the performance of the proposed estimator is compared to previously suggested estimators and the crlb .
the results show that the proposed estimator performs close to the iterative ml estimator at significantly lower computational complexity .
|
recently , much attention has been focused on the topic of scale - free networks which characterize many social , information , technological and biological systems.@xcite the qualitative properties of many interesting real - world examples , such as the internet network , the power grid network and the protein interaction network , are as following : ( 1 ) : : the degree distribution has power - law tail ; ( 2 ) : : local clustering of edges : graph is not locally tree - like ; ( 3 ) : : small average distance .
the networks can be visualized by nodes representing individuals , organizations , computers and by links between them representing their interactions . for the purpose of analyzing topology
, we ignore the variation in the type of links .
robustness of the network topology comes from the presence of alternate paths , which ensures the communication remains possible in spite of the damages to the network .
designers of the networks must assume that networks have random failures or might be attacked , and some of these attacks can result in damage .
the robust networks will continue functioning in spite of such damages .
although many literatures have discussed what the optimal network topology would be , many real - world networks present the power - law degree distribution .
when the scale - free networks are subjected to random breakdowns , with a fraction @xmath0 of the nodes and their connections are removed randomly , the network s integrity might be compromised : when the exponent of the power - law degree distribution @xmath1 , there exists a critical threshold @xmath2 , such that for @xmath3 , the network would break into smaller and disconnected parts , but the networks with @xmath4 are more resilient to random breakdowns .
cohen _ et .
al_@xcite presented a criterion to calculate the percolation critical threshold to random failures to scale - free networks .
if we attack the scale - free networks intentionally : the removal of sites is not random , but rather sites with the highest connectivity are targeted first .
the numerical simulations suggest that scale - free networks are highly sensitive to this kind of attack.@xcite cohen _ et . al_@xcite studied the exact value of the critical fraction needed for disruption .
thus scale - free networks are highly robust against random failures of nodes and hypersensitive to intentional attacks against the system s largest nodes .
so a randomly chosen node has low degree with high probability , but removal of a highly connected node might produce large effect to the network .
this situation is often compared to that of the classical random graph presented by erd@xmath5s and rnyi.@xcite such graphs have a poisson degree distribution .
this makes the random graphs less robust to random failures than comparable networks with power - law degree distribution , but much more robust against attacks on hubs . in this paper , we specifically focus on the robustness of the network topology to random failures .
we use the percolation theory and the optimization method to investigate the guideline which can maximize the robustness of the scale - free networks to random failures of nodes with the constrained condition that the average connectivity of per node in the network is constant .
the percolation theory provides the measures of distribution which are possible ways for measuring robustness .
we examine the relationship between the average connectivity per node and the network robustness to random failures .
then , we investigate the trend of the network robustness to random failures with the network size @xmath6 .
the work may provide the theoretical evidence that if the minimal connectivity and the exponent of the power - law degree distribution take in more optimal way , the robustness of the scale - free networks can be optimized .
if we construct and maintain a network with a given number of nodes as being proportional to the average number of links @xmath7 per node in the network , our goal then becomes how to maximize the robustness of a network with @xmath6 nodes to random failures with the constraint that the number of links remains constant but the nodes are connected in a different and more optimal way .
our goal is to maximize the threshold for random removal with the condition that the average degree @xmath7 per node is constant .
we construct the following model .
@xmath8 { \rm s.t .
} & \langle k \rangle={\rm constant . }
\end{array } \right.\ ] ] for any degree distribution @xmath9 , the threshold for random removal of nodes is@xmath10 @xmath11 where @xmath12 is calculated from the original connectivity distribution .
a wide range of networks have the power - law degree distribution : @xmath13 where @xmath14 is the minimal connectivity and @xmath15 is an effective connectivity cutoff presented in finite networks . to the power - law degree distribution , the average @xmath7 can be given with the usual continuous approximation , this yields @xmath16}{2-\alpha}. % \end{array}\ ] ] from ( [ f2.2 ] ) , @xmath17 can be calculated as @xmath18}{[k^{(3-\alpha)}-m^{(3-\alpha)}]}.\ ] ] in a finite network , the largest connectivity @xmath19 can be estimated from@xcite @xmath20 where @xmath6 is the number of the network nodes .
then we have that @xmath21 to the power - law degree distribution @xmath9 , we have @xmath22 & = & c[k^{(1-\alpha)}-m^{(1-\alpha)}]/(1-\alpha ) , \\[8pt ] \end{array}\ ] ] this yields @xmath23}.\ ] ] in the real world , there always exists the relation @xmath24 , so we have @xmath25 combining ( [ f2.3 ] ) and ( [ f2.4 ] ) , we have that @xmath26.\ ] ] from ( [ f2.5 ] ) , we have the following numerical results .
it can been seen from table 1 that the distribution exponent @xmath27 increases when the minimal connectivity @xmath28 increases . combining ( [ f2.6 ] ) and ( [ f2.5 ] )
, we have that @xmath29 from table 1 , we can get the following relationship : ( 1 ) : : when the average connectivity @xmath7 per node is constant , the exponent @xmath27 increases when the minimum connectivity @xmath28 increases ; ( 2 ) : : to the minimum connectivity @xmath30 , the exponent @xmath27 decreases when the average connectivity @xmath31 of the network increases . using the results obtained above
we construct the following model .
@xmath32 s.t . &
\frac{(\alpha-1)}{(\alpha-2)}m[1-n^{-\frac{\alpha-2}{\alpha-1}}]=\langle k \rangle\\[10pt ] & m\in z^{+ } , \end{array } \right.\ ] ] where @xmath33}{[k^{(3-\alpha)}-m^{(3-\alpha)}]}$ ] .
the numerical results suggest that whether the network size @xmath6 is very large or not , @xmath2 reaches its maximum value when @xmath30 .
the numerical results are presented in table 2 . from table 2
, we can get the following three conclusions : ( 1 ) : : if the average connectivity @xmath7 per node and the exponent @xmath27 of the power - law degree distribution is constant , the robustness of the scale - free networks will decrease when the network size becomes larger .
( 2 ) : : if the network size @xmath6 is constant , the robustness of the scale - free networks increase when the average connectivity @xmath7 becomes larger .
( 3 ) : : to the random failures , we have to take several times cost to increase the robustness of the scale - free networks about 1% .
it is well known that the networks with power - law degree distribution are resilient to random failures . but this conclusion do nt answer the following three questions : ( i ) to a constant average connectivity @xmath7 , how to determine the distribution exponent of the scale - free networks so that the networks are more robust to the random failures .
( ii ) to different network size , how many edges we need to add to the network to satisfy the robustness level to random failures .
( iii ) to an exist network with power - law degree distribution , what we need to do to improve the network robustness . in this paper
, we use the percolation theory and the mathematic programme method to optimize the robustness of the scale - free networks for random failures and give the numerical results .
finally , we give the relationship between the threshold @xmath2 and the network size , the degree distribution exponent @xmath27 and the average connectivity @xmath7 per node . from fig . 1
, we can get the conclusion that if the scale - free networks size become large , the network robustness to random failures would become weak . to the internet and other growing scale - free networks the designers must add more links to the network to improve the average connectivity per node to random failures .
subjects for further study include ( i ) an analysis of the robustness of the scale - free networks to the intentional attack to the highest connectivity nodes .
( ii ) the optimization of complex network under both random failures and intentional attack . ( iii ) the topology structure to improving the robustness of existing scale - free networks .
the authors are grateful to dr .
qiang guo for her valuable comments and suggestions , which have led to a better presentation of this paper .
this research was supported by chinese natural science foundation grant nos .
70431001 and 70271046 .
|
it has been found that the networks with scale - free distribution are very resilient to random failures .
the purpose of this work is to determine the network design guideline which maximize the network robustness to random failures with the average number of links per node of the network is constant .
the optimal value of the distribution exponent and the minimum connectivity to different network size are given in this paper .
finally , the optimization strategy how to improve the evolving network robustness is given .
|
the discovery of superconductivity in ferromagnetic ( fm ) uge@xmath4 under pressure gave great impact for condensed - matter community @xcite , since ferromagnetism and superconductivity have been considered to be mutually exclusive phenomena .
after the discovery , superconductivity has been found in several u - based fm compounds @xcite .
ucoge is one of the compounds among the fm superconductors discovered so far which can be readily explored in experiments , because of its high superconducting ( sc ) transition temperature ( @xmath0 ) and low curie temperature ( @xmath5 ) at ambient pressure .
actually , important and intriguing experimental results have been reported for ucoge @xcite , and some of them are summarized as follows .
1 . microscopic measurements of @xmath6sr [ fig.1(a ) ] @xcite and co - nqr [ fig.1(b ) ] @xcite have shown that the superconductivity occurs within the fm region , resulting in homogeneous microscopic coexistence of the ferromagnetism and superconductivity .
the superconductivity is also found in the paramagnetic state and @xmath0 shows a broad maximum around a critical pressure of ferromagnetism [ fig.1(c ) ] @xcite , which is in sharp contrast with uge@xmath4 where the sc is seen only in the fm phase @xcite .
+ ( a)temperature variation of the normalized asymmtries ( @xmath7 ) of the fm and paramagnetic phases measured with zero - field muon - spin rotation on ucoge @xcite . below @xmath8 k
, magnetism is observed in the whole sample volume although superconductivity occurs at @xmath9 0.8 k , indicative of coexistence of ferromagnetism and superconductivity on a microscopic scale .
( b ) temperature dependence of @xmath10co @xmath11 in the single - crystal sample @xcite .
@xmath11 above 2.3 k , shown by the blue dots , was measured at the paramagnetic ( pm)-frequency ( 8.3 mhz ) peak arising from the @xmath12 transitions , and @xmath11 below 2.3 k , shown by the red dots , was measured at the fm - frequency ( 8.1 mhz ) peak shifted from the pm peak due to the appearance of the internal field at the co site .
the inset shows the nqr signal arising from the @xmath12 transitions obtained at 4.2 k ( pm ) and 0.9 k ( fm ) .
two @xmath11 components were observed in the sc state , the faster ( slower ) component denoted by red open ( closed ) squares ; the red broken curve below @xmath0 represents the temperature dependence calculated assuming a line - node gap with @xmath13 .
the sc anomaly observed from the fm signal is clear evidence that superconductivity occurs in the fm region .
( c ) pressure - temperature phase diagram of ucoge @xcite .
ferromagnetism is denoted as blue area , superconductivity is as yellow area .
@xmath0(@xmath14 ) extrapolates to a fm quantum critical point at the critical pressure @xmath15 .
superconductivity coexists with ferromagnetism below @xmath15 as blue - yellow hatched area .
inset : amplitude of @xmath16 at @xmath17 as a function of pressure .
the data follow a linear @xmath14 dependence and extrapolate to @xmath18 gpa . ]
the temperature variation of the @xmath19co - nqr spectra shown in fig .
[ fig2 ] indicates that the fm region grows continuously around @xmath5 coexisting with the pm region , and that the fm moment in the fm region appears discontinuously , whereas the hysteresis behavior was not observed in the temperature variation of the spectra and @xmath11 in the both states below @xmath5 is almost identical @xcite . since the divergence of @xmath11 and the clear anomaly in the specific heat
were observed at @xmath5 @xcite and the fm ordered moment is small ( @xmath20 ) at low temperatures @xcite , the fm phase transition in ucoge is considered to possess weak first - order character with critical fm fluctuations , which is close to second - order one . although we need further investigation of sample and pressure dependence on the fm transition to understand the transition character thoroughly , the presence of the critical fluctuations implies the validity of the scenarios of spin fluctuation - induced superconductivity . + ( a ) @xmath19co - nqr spectra ( @xmath21 ) at 4.2 k in the pm state@xcite .
( b ) temperature dependence of the nqr spectrum from the @xmath22 transitions ( @xmath23)@xcite . the pm and fm signals coexist between 1 and 2.7 k. the blue ( red ) broken lines represent gaussian fits to the pm ( fm ) peaks ; the solid lines are guides to the eye . ]
although transport properties show rather three dimensional isotropic characters @xcite , studies of the sc upper critical magnetic field ( @xmath3 ) [ fig .
2 ( a ) ] and angle dependence of @xmath3 [ fig .
2 ( b ) ] have shown that sc properties are highly anisotropic @xcite ; the superconductivity survives in fields exceeding 15 t along the @xmath24 and @xmath25 axes , whereas @xmath3 for fields along the @xmath26 direction ( @xmath27 ) is as small as 0.8 t. + ( a ) temperature dependence of the upper critical fields for @xmath28 and @xmath26-axis @xcite . the temperature is normalized by the sc critical temperature @xmath0 at zero field .
( b ) temperature dependence of the upper critical field @xmath29 close to @xmath30 @xmath24-axis @xcite .
the field direction is tilted from the @xmath24 to @xmath26-axis .
the inset shows the angular dependence of @xmath29 at 90 mk .
the values of @xmath29 greater than 16 t are the results of linear extrapolations to 90 mk . ]
the static magnetic susceptibility is strongly anisotropic @xcite ; the magnetic anisotropy is ising like with the @xmath26 axis as a magnetic easy axis [ fig .
[ fig4 ] ( a ) ] .
in addition , direction - dependent nuclear - spin lattice relaxation rate ( @xmath11 ) measurements on a single crystalline sample have revealed the magnetic fluctuations in ucoge to be ising - type fm ones along the @xmath26 axis ( longitudinal - mode fluctuations ) as shown in fig .
[ fig4 ] ( b ) @xcite .
the ising - type fluctuations are also suggested from the anisotropic correlation length measured by the inelastic neutron measurements @xcite .
+ nmr shift @xmath31 measured in field along the @xmath24 , @xmath25 and @xmath26 directions @xcite .
remarkable anisotropy was observed . in the inset , @xmath31 is plotted against the bulk susceptibility @xmath32 measured in 2 t. the nearly identified slopes for all the directions indicate that the hyperfine coupling constants are isotropic .
( b ) direction - decomposed dynamic susceptibility @xmath33 and static susceptibility @xmath34 along each direction @xcite .
identical ising anisotropy for both quantities found above 8 k suggests that the longitudinal mode dominates these fluctuations .
it is noteworthy that in this @xmath35 range , the knight shift scales with @xmath33 , which is predicted by the self - consistent renormalization theory for a 3-dimensional nearly fm metals @xcite . ]
5 . from detailed angle - resolved nmr and meissner measurements
@xcite , it was found that magnetic fields along the @xmath26 axis ( @xmath36 ) is a tuning parameter of the longitudinal fm fluctuations and strongly suppresses the fm fluctuations , as shown in fig . 5 ( a ) and 5 ( b ) .
this sharp angle dependence of @xmath11 was also confirmed at 600 mk just above @xmath0 @xcite .
it was also shown that the superconductivity is observed in the limited magnetic - field region where the longitudinal fm spin fluctuations are active [ fig .
4 ( c ) ] .
these results suggest that the superconductivity in ucoge is tightly coupled with the longitudinal fm spin fluctuations along the @xmath26-axis .
+ ( a ) angle dependence of @xmath11 in the @xmath37 plane measured in three different magnetic fields at 1.7 k @xcite .
( b ) plot of the @xmath11 against @xmath36 .
the @xmath11 data collapse onto a single curve when plotted against @xmath36 .
( c ) @xmath36 dependence of magnetic fluctuations along the @xmath26 axis @xmath38 at 1.7 k , extracted using @xmath39 .
angle dependence of @xmath29 , determined by the onset of the meissner signal , is plotted against @xmath40 .
it is found that the superconductivity is observed in the limited field region ( yellow slashed area ) , where the longitudinal fluctuations are active .
inset : plot of @xmath38 against @xmath41 .
the relation of @xmath42 is shown by dashed lines in ( c ) and the inset . ] from a theoretical point of view , when an itinerant fm superconductor possesses the large energy splitting between the majority and minority spin fermi surfaces , exotic spin - triplet superconductivity is anticipated , in which pairing is between parallel spins within each spin fermi surface .
in addition , it has been argued that critical fm fluctuations near a quantum phase transition could mediate spin - triplet superconductivity @xcite , and actually , ( iv ) in the above indicates that the superconductivity in ucoge is intimately coupled with the ising fm spin fluctuations .
however , as far as we know , there have been no reports identifying magnetic fluctuations inducing unconventional superconductivity . here ,
we show , from model calculations on the basis of the scenario of fm spin fluctuation - mediated superconductivity in ucoge , that unconventional temperature and angle dependence of @xmath3 is consistently understood by adopting the characteristic longitudinal fm fluctuations revealed experimentally . besides , by comparing calculations for possible pairing states , we also discuss candidates of sc gap symmetries in ucoge .
in this section , we introduce a simple model to study @xmath3 in ucoge at ambient pressure where the ferromagnetism and superconductivity coexist . although ucoge has a complicated multi - band structure , since the spin fluctuations dominate the low energy physics in ucoge , we can use a simplified model to study its superconductivity with appropriately taking into account the critical fm spin fluctuations observed in the experiments @xcite .
the model relevant to ucoge in the fm phase with an applied magnetic field @xmath43 is @xmath44c_k(\tau),\\ s_{\rm ex}&=-({\mbox{\boldmath $ h$}}_{ex}+\mu_b{\mbox{\boldmath $ h$}})\cdot \sum_k\int_0^{1/t}d\tau c^{\dagger}_k(\tau){\mbox{\boldmath $ \sigma$}}c_k(\tau),\\ s_{\rm int}&=-\frac{2g^2}{3}\sum_q \int_0^{1/t}d\tau\int_0^{1/t}d\tau^{\prime } s_q^c(\tau)\chi^c({\mbox{\boldmath $ q$ } } , \tau-\tau^{\prime})s_{-q}^c(\tau^{\prime}),\end{aligned}\ ] ] and effects of the vector potential on electron motions are taken into account in a semiclassical way .
@xmath45 is an annihilation operator of low energy quasiparticles which are formed through hybridizations between conduction electrons and @xmath46-electrons .
a typical energy scale of the quasiparticles is @xmath47 ( k ) estimated from the specific heat @xcite and the coherence temperature @xcite . since the anisotropy in the resistivity in ucoge is small as discussed in the introduction , we use an isotropic dispersion @xmath48 where @xmath49 is the lattice constant and @xmath6 is the chemical potential .
we fix the electron density per site as @xmath50 which leads to a small but nearly spherical fermi surface .
as discussed in ref.@xcite , the anisotropic mass model can not explain the anomalous angle dependence of @xmath3 . in our model ,
the anisotropy in @xmath3 arises from the exchange term and the pairing interaction term in eq.([eq : action ] ) .
we use an energy unit @xmath51 and a length unit @xmath52 . the second term of eq.([eq : action ] )
includes the exchange splitting of the fermi surface in the fm phase which is assumed to be large compared to the sc transition temperature at zero - field @xmath53 .
the exchange field is parallel to the @xmath26-axis and fixed to be @xmath54 , which would lead to suppression of the pauli depairing effect against in - plane magnetic fields in the equal spin pairing states of the superconductivity @xcite .
this is a key assumption to understand the large anisotropy in @xmath3 in the present study , because the equal spin pairing states are not robust against the in - plane zeeman field while the observed @xmath55 are huge in ucoge .
however , we put a remark on this assumption .
ucoge has a multi - band structure with a strong spin - orbit interaction and 5@xmath56 states are located near the fermi level . in such a system , it is possible that , even if the magnetization in the fm state is small , the exchange splitting for each fermi surface is large . actually , in the band calculations , the orbital magnetizations and the spin polarizations cancel to result in a small total magnetic moment .
the present assumption can be checked by nmr knight shift measurements in the sc states . finally , the last term in eq .
( [ eq : action ] ) describes the interaction between the quasiparticles through the ising spin fluctuations , and @xmath57 .
the dynamical susceptibility @xmath58 strongly depends on the @xmath26-axis magnetic field as revealed in the nmr experiments @xcite .
@xmath59 where @xmath60 and @xmath61 is approximately the fermi velocity . here
, we have assumed a simple landau damping in @xmath58 , although existence of non - landau damping was reported @xcite .
the coefficient @xmath62 is determined by the nmr experiments , and a normalization factor @xmath63 t is introduced as @xmath64 so that the dimensionless parameter @xmath26 is @xmath65 .
although the origin of this magnetic field dependence in the susceptibility is not so clear , we have a possible explanation for it . according to the band calculation ,
the 5@xmath56 bands in ucoge have some semi - metallic like structures around several @xmath66-points @xcite . for such structures ,
the density of states would be @xmath67 , which results in @xmath68 near the criticality . in this scenario ,
detailed structures in the bands are important , and therefore , the spin fluctuations in ucoge under a magnetic field can depend on sample qualities .
this problem will be studied elsewhere .
we note that the coupling constant @xmath69 should be regarded as a renormalized one including the vertex corrections which enhance the sc transition temperatures .
although we have postulated a simple analytic @xmath70-dependence in @xmath71 , there might exist non - analytic corrections @xmath72 in 3-dimensional isotropic systems @xcite .
the non - analytic corrections are important in connection with fluctuation - induced first order phase transitions . however , since the first order character of the fm transition is weak , and critical fm fluctuations are well developed in ucoge as discussed in sec.[sec : intro ] , it would be legitimate to neglect the first order character , and assume that @xmath71 is analytic in @xmath70 as long as calculations of the superconducting properties are concerned .
in addition , if exist in the ising - like system , the non - analytic corrections would not change the present study qualitatively as discussed in @xcite . in this study , we focus on the anisotropy between the @xmath24 and @xmath26-axes upper critical fields ( @xmath73 and @xmath27 ) .
however , it was reported that the @xmath25-axis upper critical field @xmath74 shows a `` s - shaped '' curve at high fields , and relations to the reentrant superconductivity in urhge were discussed @xcite . in urhge , the magnetization flips at a critical @xmath75 and the spin fluctuations at @xmath76 are considered to be different from those of @xmath77 @xcite .
similar scenarios are expected to hold also in ucoge where the s - shaped @xmath74 would indicate a crossover of the superconductivities of different mechanisms under magnetic field along the @xmath25 axis .
the problem of @xmath74 in ucoge is an another issue and will be studied elsewhere . to determine the upper critical field @xmath3 in the @xmath78-plane , we solve the linearized eliashberg equation in the presence of a vector potential @xcite .
the eliashberg equation for an equal - spin pairing state in the fm phase is @xmath79 \delta_{\sigma^{\prime}\sigma^{\prime}}(k^{\prime}),\end{aligned}\ ] ] where @xmath80 and @xmath81 .
a short notation @xmath82 is used hereafter .
here we have neglected the effect of the magnetization in the orbital motion , because the observed moment is very small in ucoge @xcite .
note that the green s function @xmath83 has off - diagonal elements when @xmath84 .
the gap function depends both on the momentum in the relative coordinate and the center - of - mass coordinate , and is expanded as @xmath85 , where @xmath86 is the lowest level landau function .
we solve the linearized eliashberg equation for two cases corresponding to the classification by the magnetic point group @xmath87 for ucoge which has an orthorhombic crystal structure @xcite .
one is the so - called a state which corresponds to point node symmetry with a @xmath88-vector @xmath89 near the @xmath90 point .
the other one is the so - called b state corresponding to horizontal line node symmetry with @xmath91 . here ,
@xmath92 and @xmath93 are real coefficients .
both of the gap functions are determined by self consistent calculations , and the resulting sc states are non - unitary because of the presence of the exchange field @xmath94 .
we note that , under the applied magnetic field @xmath43 , the calculated gap functions are distorted and has lower symmetry than that of the cubic lattice in the present model .
the pairing interaction is evaluated at the first order in @xmath95 , @xmath96 for the selfenergy , we include only the term @xmath97,\end{aligned}\ ] ] where @xmath98 is the non - interacting green s function .
other neglected terms in the self energy are much smaller than @xmath99 in magnitude .
as noted in the previous section , the coupling constant @xmath69 is a renormalized one .
in the present study , the coupling constant is fixed as @xmath100 which is rather large to give a high transition temperature so that the numerical costs are reduced .
however , qualitatively , our main results do not depend on the value of @xmath95 .
for this value of @xmath95 , the mass enhancement factor @xmath101 is @xmath102 , which is not so large because the density of states at the fermi surface is small in the present parameters .
the transition temperature for the a state without magnetic field is @xmath103 which we regard as 1.0 k according to the experiments for ucoge .
then , the hopping integral is fixed as @xmath104 k throughout in our study , and the order of this value is consistent with the coefficient of the linear specific heat @xmath105 mj / k@xmath106mol @xcite and the coherence temperature @xmath107 k @xcite .
we also fix the lattice constant , @xmath108 which leads to an effective mass for the orbital motion of the cooper pairs @xmath109 where @xmath110 is the bare electron mass .
correspondingly , we fix @xmath111 t , which is consistent with the nmr experiments @xcite . before discussing numerical results of @xmath3 of the present model
, we note that the orbital limit @xmath112 depends on positions of gap nodes in general @xcite .
a simple explanation is as follows .
let us consider a superconductor with isotropic fermi surface which has nodes in a gap function .
the effective velocity for the cyclotron motion of the orbital depairing effect can be of the form of @xmath113 for the basis function @xmath114 corresponding to the sc gap symmetry , where @xmath115 is a velocity of electrons . for @xmath114 with horizontal line nodes , when the magnetic field is parallel to the @xmath26-axis , the cyclotron motion is suppressed at the nodes where in - plane @xmath116 is zero , resulting in a large orbital limiting field .
the cyclotron motion is not suppressed for @xmath117 because @xmath118 is non - zero on the fermi surface , which leads to the anisotropy of the orbital limiting field @xmath119 .
on the other hand , for the point node gap function , @xmath116 can not be zero except for the poles of the fermi surface @xmath120 which are realized as cross points of two line nodes @xmath121 and @xmath122 .
therefore , for point nodes at the poles of the fermi surface , the order is turned over , @xmath123 . following this consideration ,
if there is no suppression of the pairing interaction and the pauli depairing effect is negligible , @xmath124 is expected for the a ( b ) state .
we now turn to discussions of calculation results .
temperature dependence of @xmath125 with @xmath126 for the a state is shown in fig .
[ fig : a1 ] . field angle dependence of @xmath3 at @xmath127 in the a state . ] field angle dependence of @xmath3 at @xmath127 in the a state . ]
it is found that the qualitative behaviors of @xmath3 are independent of @xmath26 .
@xmath73 exhibits a strong coupling behavior and reaches @xmath128 .
although the precise value of @xmath129 depends on the choice of @xmath130 and @xmath95 , strong coupling calculations of @xmath3 naturally explain the upward curvature in @xmath73 and its large value observed in the experiments , which is a key to understand the anomalous anisotropic behavior in @xmath3 .
the calculated anisotropy ratio @xmath131 becomes @xmath132 in consistent with the experiments , although the fermi surface of our model is merely spherical .
we emphasize that , for @xmath3 to have a large anisotropy between the @xmath24 and @xmath26-axes , the experimentally observed @xmath133-dependence in the susceptibility is crucially important .
for a comparison , we calculate @xmath27 in the case that @xmath71 has weaker @xmath134-dependence , mean - field like dependence of @xmath135 , for which results are shown with purple symbols in fig .
[ fig : a1 ] . in this case ,
suppression of @xmath3 for the @xmath26-axis is only moderate , and we can not obtain anomalously large anisotropy in @xmath3 . in fig . [
fig : a2 ] , we show field angle dependence of @xmath136 at @xmath127 , where @xmath137 is the angle from the @xmath24-axis to the @xmath26-axis . it is seen that @xmath136 is strongly suppressed when the field direction is slightly tilted from the @xmath24-axis to the @xmath26 axis .
this strong angle dependence is again due to the @xmath133 suppression of the pairing interaction .
the calculation results well explain the unusual angle dependence of @xmath3 observed experimentally .
this good agreement is strong evidence that the superconductivity in ucoge is actually mediated by the critical ising spin fluctuations .
we note that the relation between @xmath36 and @xmath3 is considered to be an analog of the isotope effect in conventional superconductors where ion mass @xmath138 controls transition temperature @xmath0 . in both cases , suppression of the pairing interaction by increasing @xmath36 or @xmath138 leads to reduction of @xmath0 .
one important reason why we can establish such a relation between @xmath36 and @xmath3 in the @xmath46-electron system ucoge is that the critical spin fluctuations dominate the low energy physics .
details of the system and complexity of multi - band are sub - dominant factors for the superconductivity , and therefore , the sc properties are well understood once the spin fluctuations are appropriately taken into account .
the transition temperature at zero - field for the b state is @xmath139 which differs from that for the a state , because of the presence of the exchange field along the @xmath26-axis .
however , the difference is not important in our study .
temperature dependence of @xmath125 for the b state is shown in fig .
[ fig : b1 ] . field angle dependence of @xmath3 at @xmath140 in the b state . ] field angle dependence of @xmath3 at @xmath140 in the b state . ] for the b state , strong - coupling behaviors for the @xmath24-axis are not so significant , because the orbital limiting fields depend on the positions of gap nodes as discussed in the previous section , and the enhancement of @xmath73 for the b state is weaker than that for the a state . for the present value of @xmath141
, the pauli depairing effect can not be negligibly small at high magnetic fields , which leads to the suppression of @xmath73 in that region . for the @xmath26-axis , detailed behaviors depend on the value of the parameter @xmath26 . in the case of rather weak suppression of the ising fluctuations ,
i.e. @xmath142 , @xmath3 is enhanced at low temperatures .
if the suppression is sufficiently strong , for example @xmath143 , such enhancement can not be seen .
we also show @xmath27 calculated with the mean field @xmath144 by the purple curve as in the a state . in the b state ,
@xmath27 is not strongly suppressed and @xmath27 is larger than @xmath73 in contrast to the experiments .
the relation @xmath145 with the mean - field @xmath144 is directly understood from the general relation @xmath146 for horizontal line node gap functions in simple systems .
behaviors of @xmath3 as a function of the field angle also depend on the value of @xmath26 as shown in fig .
[ fig : b2 ] . for @xmath142 , @xmath136 shows a non - monotonic behavior .
this is because , for the b state , the relation @xmath147 holds in the orbital limit if there is no suppression of the ising fluctuations , while in fact the suppression in @xmath148 gets stronger as @xmath149 . on the other hand , for @xmath143
, @xmath3 monotonically decreases . comparing these results with those for the a state which are robust against the parameter value
, one may conclude that the a state is a promising candidate for the sc realized in the fm state in ucoge , although the b state is not ruled out .
in this paper , we have studied the upper critical field @xmath3 for the fm superconductor ucoge in the fm phase . in the introduction
, we discussed several experimental observations which are key clues to understand relations between the fm and sc .
to clarify the relationship , we introduced the simple model which includes the critical fm spin fluctuations with anomalous @xmath133 dependence revealed by the nmr experiments .
the linearized eliashberg equations are solved within the first order in the spin fluctuations . in the a state with point nodes ,
temperature dependence of @xmath73 shows an upward curvature while @xmath27 is suppressed . when the field angle is tilted from the @xmath24 to the @xmath26 axis
, @xmath3 is strongly suppressed .
these behaviors are totally due to the characteristic suppression of the fm spin fluctuations by @xmath36 , and in nice agreement with the experimentally observed @xmath3 in ucoge .
this is strong evidence of the scenario that the spin - triplet superconductivity is actually mediated by the critical ising fm spin fluctuations in ucoge .
the relation between @xmath36 and @xmath3 is an analog of the isotope effect in conventional superconductors . in the b state with a horizontal line node ,
both enhancement of @xmath73 and suppression of @xmath27 are moderate compared to those in the a state .
therefore , the a state is a promising candidate for the pairing symmetry in ucoge , although the b state is not ruled out . finally , we put remarks on some remaining problems related to the present study .
firstly , in the present study , we have simply adopted the experimentally observed @xmath133 dependence in @xmath148 . although we suggested a possible explanation , the origin of this anomalous dependence is not clear .
secondly , we have assumed that the exchange splitting is large enough to suppress the pauli depairing effect .
however , this assumption should be checked experimentally by e.g. nmr knight shift measurements .
there is also a related problem .
we have focused on the coexistence region of the superconductivity and the ferromagnetism where the exchange splitting can be large . under pressure
@xcite , @xmath5 is suppressed and the exchange splitting becomes small . in this case and in the paramagnetic phase , the pauli depairing effect would not be suppressed .
actually , it was reported that , around the critical pressure where @xmath5 is extrapolated to zero , @xmath150 is suppressed while @xmath151 is increasing @xcite .
these behaviors could be understood as a result of two competing effects ; although the pauli depairing effect would not be suppressed by the exchange field near the criticality , the fm spin fluctuations are enhanced .
investigations of these issues are needed for a comprehensive understanding of ucoge , and they would be studied in the future .
the authors thank s. yonezawa , and y. maeno for experimental support and valuable discussions , and d. aoki , j. flouquet , a. de visser , a. huxley , h. harima , and h. ikeda for valuable discussions .
this work was partially supported by kyoto univ .
ltm centre , yukawa institute , the `` heavy electrons '' grant - in - aid for scientific research on innovative areas ( no . 20102006 , no . 21102510 , no .
20102008 , no .
23102714 , and no .
23840009 ) from the ministry of education , culture , sports , science , and technology ( mext ) of japan , a grant - in - aid for the global coe program `` the next generation of physics , spun from universality and emergence '' from mext of japan , a grant - in - aid for scientific research from japan society for promotion of science ( jsps ) , kakenhi ( s ) ( no .
20224015 ) , kakenhi ( c ) ( no .
23540406 ) and first program from jsps .
99 saxena s s _ et al .
_ 2000 _ nature _ * 406 * 587 belitz d , kirkpatrick t r and vojta t 1997 _ phys .
b _ * 55 * 9452 ; belitz d , kirkpatrick t r and vojta t1999 _ phys .
lett . _ * 82 * 4707 ; belitz d , kirkpatrick t r and rollbhler j 2005 _ phys .
lett . _ * 94 * 247205 ; kirkpatrick t r and belitz d 2012 _ phys .
b _ * 85 * 134451 .
chubukov a v , ppin c and rech j 2004 _ phys .
lett . _ * 92 * 147003 ; efremov
d v , betouras j j and a. v. chubukov 2008 _ phys .
b _ * 77 * 220401(r ) ; maslov d l , chubukov a v and saha r 2006 _ phys . rev .
b _ * 74 * 220402(r ) .
|
identification of pairing mechanisms leading to the unconventional superconductivity realized in copper - oxide , heavy - fermions , and organic compounds is one of the most challenging issues in condensed - matter physics .
clear evidence for an electron - phonon mechanism in conventional superconductors is seen by the isotope effect on the superconducting transition temperatures @xmath0 , since isotopic substitution varies the phonon frequency without affecting the electronic states . in unconventional superconductors ,
magnetic fluctuations have been proposed to mediate superconductivity , and considerable efforts have been made to unravel relationships between normal - state magnetic fluctuations and superconductivity . here , we show that characteristic experimental results on the ferromagnetic ( fm ) superconductor ucoge ( @xmath1 k and @xmath2 k ) can be understood consistently within a scenario of the spin - triplet superconductivity induced by fm spin fluctuations .
temperature and angle dependencies of the upper critical magnetic field of the superconductivity ( @xmath3 ) are calculated on the basis of the above scenario by solving the eliashberg equation .
calculated @xmath3 well agrees with the characteristic experimental results observed in ucoge .
this is a first example that fm fluctuations are shown to be a pairing glue of superconductivity .
|
the description of the measurement process has been a topic debated from the early developments of quantum mechanics @xcite . besides being a basic issue in the interpretation of the quantum formalism
it has also a practical interest in predicting the results of experiments pointing out some paradoxical aspects of the quantum laws when compared to the classical ones .
the development of devices whose noise figures are close to the quantum limit makes these experiments within the technological feasibility and demands for a systematization of the theory of quantum measurement with deeper understanding of its experimental implications @xcite . in ordinary quantum mechanics measurements are taken into account by postulating the wave function collapse @xcite .
this approach has the unpleasant feature of introducing an extra assumption in the theory which , moreover , regards only instantaneous and perfect measurements . in the last twenty years
relevant steps have been made in upgrading the von neumann postulate with more realistic and satisfactory approaches .
these approaches essentially recognize that a measured system is not isolated but in interaction with a measurement apparatus .
the way this basic fact is taken into account in the evolution law of the measured system has been developed according to different languages and points of view .
a group algebra approach to the problem of an open quantum system was proposed by postulating ( completely positive ) semigroup properties for the dynamical law of a system in interaction with a markovian environment @xcite . in this case
the open system is described by a semigroup master equation later successfully used in modeling quantum optics experiments @xcite .
the semigroup master equation , still preserving positivity and trace of the density matrix operator , introduces decoherentization by dynamically quenching the off - diagonal elements of the density matrix , a key property used to explain the absence of superposition states in a measurement apparatus @xcite or in a general macroscopic system @xcite thought as systems interacting with an environment .
open quantum systems were gathered to measured systems by obtaining an all alike semigroup master equation for a continuous measurement process @xcite modeled by repeated instantaneous effect - valued measurements @xcite , i.e. , partial localization or decoherentization kicks @xcite given to the density matrix at random values of the measured observable .
in the model @xcite the semigroup master equation comes out only after averaging the instantaneous results of the measured observable with probability distribution in agreement with standard quantum mechanics @xcite . in absence of this
average , i.e. , for a particular _ selection _ of the measurement outcomes , a nonlinear stochastic differential equation was proposed to describe the evolution of the density matrix during a continuous measurement process @xcite . since in this case during the measurement process the density matrix coincides with its square , a nonlinear stochastic differential equation could be derived for the corresponding wave function @xcite crowning previous attempts to include the lders postulate @xcite , a generalization of the von neumann one to _ selective _ measurements , into a stochastic schrdinger equation @xcite .
this equation , obtained also in the framework of the quantum filtering theory @xcite , is a special case of the more general quantum state diffusion equation for open systems @xcite .
its nonlinearity can be avoided only renouncing to make predictions on the outcome of the measured observable .
indeed , a linear stochastic differential equation was proposed for the so called _ a posteriori _ states @xcite , i.e. , those unnormalized states which describe the quantum system when the measurement result is already known .
as originally suggested by feynman @xcite the path integral approach to quantum mechanics is a quite appropriate framework for incorporating the effect of a measurement .
the hint was picked up by modeling a continuous measurement of position having a certain result ( selective measurement ) by a restriction of the feynman path integral @xcite .
the method was extended to measurements of a generic observable , function of momentum and position , by a restriction of the quantum propagator in the phase - space formulation @xcite .
the restriction was originally proposed as a gaussian functional damping the paths proportionally to the time - averaged squared difference between the value taken by the observable on the paths and the measurement result , normalized to a given variance @xcite . however , the corresponding quantum propagator has desirable dynamical semigroup properties if the gaussian restriction is chosen linear in time @xcite which is equivalent to make the previously proposed variance scaling with the inverse of the measurement time @xcite .
the modifications introduced in the quantum propagator by the restricted path integral approach can be incorporated into an effective hamiltonian depending on the selected measurement result .
an effective wave equation was then derived which is linear in the wave function if the selected measurement result is considered known @xcite . on the other hand ,
if the selected measurement result is to be determined according to the evolution of the measured system the effective wave equation is read as a nonlinear stochastic differential equation .
equivalence to the previously proposed stochastic equations occurs @xcite .
the selective constraint imposed to the restricted path integral approach can be removed by summing over all possible measurement outcomes . in this way a nonselective process described by a density matrix was obtained @xcite .
the effect of the interaction with the measurement apparatus modifies the corresponding quantum propagator through an influence functional @xcite which turns out to be equivalent to those obtained with other methods @xcite . from the above incomplete list of approaches ,
apart from an evident difference in the languages , a rather unified picture of the problem of quantum measurement emerges .
it deserves uprising to the systematic theory of a quantum system evolving under the effect of a measurement process .
such theory , named , for brevity , measurement quantum mechanics , is formally presented in section ii . a less formal introduction to measurement quantum mechanics is attainable through the analysis of a model of measurement device which is rather general .
this is important not only as a justification of the formal approach to the theory but , above all , for clearly defining the meaning of a measurement in relation to the observer .
as noted by cini @xcite paradoxical features arise from not considering the objective role of the observer in a measurement process .
the evolution of a measured system in interaction with a measurement apparatus can not depend on the observer looking or not at the pointer of the apparatus . in order to take into account such objectivity
the measurement apparatus must be classical with respect to the observer @xcite . in section iii
we propose a simple model of measurement apparatus shaped as an environment of particles linearly interacting with the measured system and in contact with a heat reservoir at a fixed temperature @xcite . in the high temperature limit
the particles behave as an informational environment which extracts information in objective way . noticeably
, the formal structure of measurement quantum mechanics is obtained in the same limit .
a further analysis of measurement quantum mechanics is given in section iv in connection to a recent optical experiment @xcite showing quantum zeno effect .
comparison of the experimental results with the theoretical predictions indicate that the experiment @xcite was performed in a regime of very strong coupling of the measurement apparatus with the measured system .
repetition of this experiment or similar ones in a weaker coupling regime is desirable for investigating interesting quantum features .
some final remarks are given in section v.
in this section we show the mathematical equivalence of the five approaches to the problem of measurement in quantum systems briefly described in the introduction .
the demonstration is organized in steps relating neighboring pairs of approaches and giving rise to the equivalence loop sketched in fig .
no one of these equivalence steps is novel .
however , we reconsider all them in a unified framework and language with naturally emerging definitions for concepts as selective and nonselective measurements as well as _ a priori _ and _ a posteriori _ analysis of a measurement process . as a result
we get a theory which takes into account the effect of a measurement process in ordinary quantum dynamics through a phenomenological parameter coupling the measured system to the measurement apparatus .
we enter the loop of fig . 1 at the group algebra approach to the master equation and go on in the clockwise sense . ( 314,314)(-40,-260 ) = 0.75pt ( 97.0,-45.0)(120,90 ) ( -53.3,-153.5)(120,90 ) ( 4.7,-329.0)(120,90 ) ( 189.3,-329.0)(120,90 ) ( 246.3,-153.5)(120,90 ) ( 64.7,-30.0)(4,3)14.4 ( 64.7,-30.0)(-4,-3)14.4 ( 249.3,-30.0)(-4,3)14.4 ( 249.3,-30.0)(4,-3)14.4 ( 7.7,-205.5)(-1,3)5.7 ( 7.7,-205.5)(1,-3)5.7 ( 306.3,-205.5)(1,3)5.7 ( 306.3,-205.5)(-1,-3)5.7 ( 157.0,-314.0)(1,0)18.0 ( 157.0,-314.0)(-1,0)18.0 ( 157.0,-157.0)(0,0)*mqm * the dynamics of a system interacting with an environment
is conveniently described in terms of a reduced density matrix operator @xmath0 obtained by tracing out the environment variables from the density matrix operator of the whole system + environment .
the unitary evolution of @xmath0 for the isolated system is modified to an irreversible one by the interaction with the environment . in the limit of a markovian environment
a dynamical law described by a completely positive semigroup has been postulated @xcite resulting in the following master equation for the reduced density matrix operator @xmath1 + { 1\over2 } \sum_\nu \left ( \left [ \hat l_\nu(t ) \hat \rho(t ) , \hat l_\nu(t)^{\dag } \right ] + \left [ \hat l_\nu(t ) , \hat \rho(t ) \hat l_\nu(t)^{\dag } \right ] \right ) \label{lindbeq}\ ] ] where @xmath2 is the hamiltonian operator for a general nonautonomous system and @xmath3 are the lindblad operators representing the influence of the environment on the system .
the above equation is thought to describe the general case of an open quantum system .
the evolution of a quantum system subjected to a measurement process is a particular case where the environment is the measurement apparatus and the lindblad operators are proportional to the measured quantities . if , for simplicity , we consider the measurement of a single observable represented by the hermitian operator @xmath4 , the corresponding lindblad operator can be chosen as @xmath5 .
note that we allow for an explicit time dependence of @xmath6 in order to include the case of general observables such as the continuous quantum nondemolition ones @xcite .
the function @xmath7 has dimensions @xmath8=[t^{-1 } a^{-2}]$ ] and represents , as we shall see in the next section , the coupling of the measured system to the measurement apparatus .
the measurement process is continuous in time and the measurement coupling is time dependent as requested in the description of a general experimental situation . due to the presence of commutators in the right hand side of ( [ lindbeq ] ) the trace of the reduced density matrix operator is a conserved quantity which we assume to be unity
@xmath9 we shall see later , when introducing selective measurements , that the reduced density matrix operator @xmath10 corresponds to an incoherent mixture of pure states associated to selective processes .
in antithesis , the process described by @xmath10 is called nonselective and eq .
( [ trace ] ) can be interpreted as a normalization relation for the probability distribution over the selective processes . according to this interpretation
, the result of a nonselective measurement , @xmath11 , and its associated variance , @xmath12 , can be evaluated by the trace rules @xmath13 \label{avens}\ ] ] and @xmath14 , \label{var}\ ] ] respectively . here overlining is used to denote the statistical average over the selective measurement results @xmath15 to be defined in the following together with the corresponding probability distribution .
the equivalence of the master equation ( [ lindbeq ] ) to a density matrix propagator expressed in terms of an influence functional is our first step in the loop of fig .
1 . by moving to the coordinate representation of the reduced density matrix operator @xmath16
the corresponding evolution equation becomes the partial differential equation @xmath17 ^ 2 \bigg ] \rho(q_1,q_2,t ) \end{aligned}\ ] ] which can be transformed into the integral equation @xmath18 the two - point green function ( density matrix propagator ) @xmath19 has a phase - space path - integral representation which can be derived by standard methods @xcite .
for a small time interval @xmath20 one has @xmath21 ) and inserting at the appropriate places the four identities @xmath22 where @xmath23
we get @xmath24 \delta t \right .
\nonumber \\ & & ~-{1\over2 } \kappa(t '' - \delta t ) \left [ a \left(p_1^{(1)},q_1^{(1)},t '' - \delta t \right ) - a \left(p_2^{(1)},q_2^{(1)},t '' - \delta t \right ) \right]^2 \delta t \nonumber \\ & & ~\left . - { i\over \hbar } \left [ p_2^{(1 ) } { q_2 '' - q_2^{(1 ) } \over \delta t } -h \left(p_2^{(1)},q_2^{(1)},t '' - \delta t \right ) \right ] \delta t \right ) \rho(q_1^{(1)},q_2^{(1)},t '' - \delta t ) + { \cal o}(\delta t^2 ) .
\end{aligned}\ ] ] by iterating this relation @xmath25 times with @xmath26 and then taking the limit @xmath27 , a functional measure arises @xmath28~ d[q]_{q',t'}^{q'',t '' } = \lim_{n \to \infty } \prod_{n=1}^{n } { dp^{(n ) } \over 2 \pi \hbar } \prod_{n=1}^{n-1}dq^{(n)}\ ] ] with boundary conditions imposed only to the @xmath29 paths . by comparison with eq .
( [ inteq ] ) we conclude @xmath30~ d[q_1]_{q_1',t'}^{q_1'',t '' } \int d[p_2]~ d[q_2]_{q_2',t'}^{q_2'',t''}~ \exp \left ( { i\over \hbar } s[p_1,q_1 ] - { i\over \hbar } s[p_2,q_2 ] ~- z [ p_1,q_1,p_2,q_2 ] \right ) \label{green2}\end{aligned}\ ] ] where @xmath31 = \int_{t'}^{t '' } dt~ [ p \dot q - h(p , q , t ) ] \label{s}\ ] ] @xmath32 = { 1 \over 2 } \int_{t'}^{t '' } dt~ \kappa(t ) \left [ a(p_1,q_1,t ) - a(p_2,q_2,t ) \right]^2 . \label{iphase}\ ] ] the effect of the measurement in the two - point green function is represented by the functional @xmath33 which reduces to the identity for @xmath34 , i.e. , in absence of a measurement process .
the functional @xmath33 is the feynman - vernon influence functional @xcite evaluated by tracing out the degrees of freedom of the environment in the phase - space path - integral formulation of quantum mechanics .
as we shall see in the next section , the influence functional approach allows us to give an enlightening interpretation of the nonselective measurement processes with an explicit expression of the parameter @xmath7 . the second step in the equivalence loop of fig .
1 is accomplished by a formal manipulation of the influence functional @xcite . by using the identity @xmath35
^ 2 \right ) = \nonumber \\ & & ~ \int d[a]~ \exp \left ( - \int_{t'}^{t '' } dt~ \kappa(t ) [ a(p_1,q_1,t ) - a(t)]^2 - \int_{t'}^{t '' } dt~ \kappa(t ) [ a(q_2,p_2,t ) - a(t)]^2 \right ) \label{identity}\end{aligned}\ ] ] where the functional measure arises by slicing the interval @xmath36 $ ] into @xmath27 subintervals at times @xmath37 with @xmath26 , i.e. , @xmath38 = \lim_{n \to \infty } \prod_{n=1}^{n } da^{(n ) } \sqrt{2 \kappa(t^{(n ) } ) \delta t \over \pi } , \label{d[a]}\ ] ] the two - point green function @xmath19 can be decomposed into a couple of one - point green functions @xmath39}$ ]
@xmath40~ ~g_{[a]}(q_1'',t'';q_1',t ' ) ~g_{[a]}(q_2'',t'';q_2',t')^ * \label{g2ing1a}\ ] ] where @xmath41}(q'',t'';q',t ' ) = \int d[p ] d[q]_{q',t'}^{q'',t '' } \exp \left ( { i \over \hbar } s[p , q ] - \int_{t'}^{t '' } dt~ \kappa(t ) [ a(p , q , t ) - a(t)]^2 \right ) .
\label{g[a]}\ ] ] this one - point green function is the phase - space generalization of the restricted path - integral approach originally proposed on physical grounds as a damping of the feynman paths incompatible with the measurement result @xcite . here
the measurement result is the function @xmath15 which , according to the interpretation of the functional measure @xmath42 $ ] , is , in general , continuous but not differentiable .
if we assume that the measured system is in a pure state @xmath43 at the beginning of the measurement process , i.e. , @xmath44 eq .
( [ g2ing1a ] ) allows us to introduce a pure state @xmath45}(t ) \rangle$ ] which is the evolution at time @xmath46 of the initial one @xmath43 .
indeed , the reduced density matrix can be thought as the functional integral @xmath47 ~\psi_{[a]}(q_1,t ) \psi_{[a]}(q_2,t)^ * \label{decoa}\ ] ] with the wave function of the pure state defined by the one - point green function @xmath39}$ ] @xmath48}(q'',t '' ) = \int dq'~ g_{[a]}(q'',t'';q',t ' ) ~ \psi(q',t ' ) .
\label{psiag}\ ] ] the decomposition ( [ g2ing1a ] ) introduces the announced statistical description of a nonselective measurement process in terms of selective processes corresponding to different measurement results @xmath15 and characterized by a certain probability distribution . due to
( [ decoa ] ) the conservation of the trace of the reduced density matrix operator gives @xmath49~ \parallel \psi_{[a]}(t ) \parallel^2 .
\label{proba}\ ] ] therefore , @xmath50}(t ) \parallel^2 $ ] is the probability distribution for the selective process with measurement result @xmath51 $ ] .
once the probability distribution is known , the nonselective measurement result ( [ avens ] ) and variance ( [ var ] ) which , according to ( [ decoa ] ) , are written also as @xmath52 ~ \langle \psi_{[a]}(t ) | \hat a(t ) |\psi_{[a]}(t ) \rangle
\label{avea1}\ ] ] and @xmath53 ~ \langle \psi_{[a]}(t ) | \left ( \hat a(t ) - \overline{a(t ) } \right)^2 |\psi_{[a]}(t ) \rangle , \label{vara1}\ ] ] can be expressed in terms of the selective results @xmath15 by @xmath52~ \parallel \psi_{[a]}(t ) \parallel^2 a(t ) \label{avea2}\ ] ] and @xmath54~ \parallel \psi_{[a]}(t ) \parallel^2 \left ( a(t ) - \overline{a(t ) } \right ) \left ( a(t-\tau ) - \overline{a(t ) } \right ) , \label{var2}\ ] ] respectively .
the equivalence between ( [ avea1 ] ) and ( [ avea2 ] ) as well as that one between ( [ vara1 ] ) and ( [ var2 ] ) can be demonstrated explicitly by using the definition of the functional measure @xmath42 $ ] and the functional expression of the wave function @xmath55}(q , t)$ ] .
the prescription @xmath56 in ( [ var2 ] ) is used to avoid the divergence which occurs by integrating the term @xmath57 .
the third step of our equivalence loop , namely the existence of a differential equation for the wave function @xmath55}(q , t)$ ] , follows from the same standard method used for the reduced density matrix .
the explicit representation of the functional integral giving the one - point green function @xmath39}$ ] allows us to write the difference of the wave function between two close times @xmath58 and @xmath59 up to terms @xmath60 and find the differential equation @xmath61}(q , t ) = \left [ -{i \over \hbar } h \left ( -i \hbar { \partial \over \partial q},q , t \right ) - \kappa(t ) \left [ a \left ( -i \hbar { \partial \over \partial q},q , t \right ) - a(t ) \right]^2 \right
] \psi_{[a]}(q , t ) .
\label{psia}\ ] ] the mathematical properties of this differential equation crucially depend on two possible ways of analyzing a selective measurement process . a first possibility is to consider an analysis _ a posteriori_. the result of the measurement @xmath15
is already known and one wants to complete the quantum mechanical information on the measured system by evaluating the associated wave function @xmath55}(q , t)$ ] . in this case
( [ psia ] ) is an effective wave equation linear in the wave function .
this equation is deterministic but not regular , in general , depending upon the nature of the function @xmath15 .
the effect of the measurement appears as an anti - hermitian term added to the hamiltonian of the unmeasured system . due to this term the norm of the wave function @xmath55}(q , t)$ ]
is not conserved in agreement with the probabilistic normalization ( [ proba ] ) .
a second possibility is to consider an analysis _ a priori_.
the result of the measurement @xmath15 is to be predicted not deterministically but randomly with probability distribution @xmath50}(t ) \parallel^2 $ ] .
equation ( [ psia ] ) becomes a stochastic differential equation nonlinear in @xmath55}(q , t)$ ] .
this can be seen more explicitly by introducing an appropriate noise @xmath62}(t)$ ] which relates the selective result @xmath15 to the expectation value of the operator @xmath6 in the state @xmath63}(t ) \rangle$ ] .
comparison of ( [ avea1 ] ) and ( [ avea2 ] ) allows us to write @xmath64}(t ) | \hat a(t ) |\psi_{[a]}(t ) \rangle
\over \langle \psi_{[a]}(t ) | \psi_{[a]}(t ) \rangle } + \eta_{[a]}(t ) \label{aintoa}\ ] ] where @xmath62}(t)$ ] is defined by @xmath65~ \parallel \psi_{[a]}(t ) \parallel^2 \eta_{[a]}(t ) = 0 .\ ] ] the nonlinearity of the stochastic equation for @xmath55}(q , t)$ ] is made clear by inserting ( [ aintoa ] ) into ( [ psia ] ) .
this stochastic equation is also unconventional in the sense that it contains the square of the noise term @xmath62}(t)$ ] .
it is evident that eq .
( [ psia ] ) is fully appropriate in the _ a posteriori _ analysis of a selective measurement process .
however , in the _ a priori _ analysis one has to deal with an unconventional stochastic equation containing a complicated noise function defined in terms of a time - dependent probability distribution
. it would be more desirable to have a standard stochastic equation with a white noise .
this can be achieved , together with the fourth step in our equivalence loop , by a change of variable in the measurement result .
analogously to eq .
( [ aintoa ] ) we write @xmath66}(t ) + { \xi(t ) \over 2 \sqrt{\kappa(t ) } } \label{aintoaxi}\ ] ] where @xmath67}(t)$ ] is considered known for the moment and @xmath68 is a noise whose characterization is obtained through the following considerations . the two - point green function already decomposed in ( [ g2ing1a ] ) into a couple of one - point green functions @xmath39}$ ] with functional measure @xmath42 $ ] can alternatively be decomposed into another couple of one - point green functions @xmath69}$ ] @xmath70 ~g_{[\xi]}(q_1'',t'';q_1',t ' ) ~g_{[\xi]}(q_2'',t'';q_2',t')^ * \label{g2ing1xi}\ ] ] with gaussian measure @xmath71 = \lim_{n \to \infty } \prod_{n=1}^{n } d\xi^{(n ) } \sqrt{\delta t \over 2 \pi } \exp \left ( -{1\over2 } { \xi^{(n)}}^2 \delta t \right ) \label{d[xi]}\ ] ] which makes @xmath69}$ ] exponentially linear in @xmath72 @xmath73}(q'',t'';q',t ' ) = \exp \left(\int_{t'}^{t '' } dt~ { 1\over4 } \xi(t)^2 \right ) g_{[a]}(q'',t'';q',t ' ) = \int d[p ] d[q]_{q',t'}^{q'',t '' } \nonumber \\ & & ~ \times \exp \left ( { i \over \hbar } s[p , q ] - \int_{t'}^{t '' } dt~ \kappa(t ) \left [ [ a(p , q , t)- { a}_{[\xi]}(t)]^2 - 2 [ a(p , q , t)- { a}_{[\xi]}(t ) ] { \xi(t ) \over 2 \sqrt{\kappa(t ) } } \right ] \right ) .\end{aligned}\ ] ] with respect to the gaussian measure @xmath74 $ ] , the noise @xmath68 has average @xmath75~\xi(t ) = 0 \label{ximedio}\ ] ] and covariance @xmath76~ \xi(t_1)~\xi(t_2 ) = \delta(t_1-t_2 ) , \ ] ] i.e. , it is a white noise .
as in the previous case , the reduced density matrix can be written as a functional integral @xmath77 ~\psi_{[\xi]}(q_1,t ) \psi_{[\xi]}(q_2,t)^ * \label{decoxi}\ ] ] over pure states with wave functions @xmath78}(q , t)$ ] defined by @xmath79}(q'',t '' ) = \int dq'~g_{[\xi]}(q'',t'';q',t ' ) ~ \psi(q',t ' ) .\ ] ] according to the expression of the green function @xmath69}$ ] , the integral statement for @xmath78}(q , t)$ ] can be transformed into an operatorial expression by the usual slicing procedure @xmath80}(t ) \right]^2
\nonumber \\ & + & \sqrt{\kappa(t ) } \left [ \hat a(\hat p,\hat q , t ) - { a}_{[\xi]}(t ) \right ] \xi(t ) \bigg ] \bigg ) ~ |\psi_{[\xi]}(t ' ) \rangle \label{stocsol}\end{aligned}\ ] ] where @xmath81 means chronological ordering . by introducing the wiener process @xmath82 with ito algebra @xmath83 and @xmath84 ( @xmath85 ) , eq .
( [ stocsol ] ) is recognized as the formal solution of the following ito stochastic differential equation @xcite @xmath86}(t ) \rangle & = & \left [ - {
i \over \hbar } \hat h(\hat p,\hat q , t ) - { 1 \over 2 } \kappa(t ) \left [ \hat a(\hat p,\hat q , t)- { a}_{[\xi]}(t ) \right]^2 \right]~ |\psi_{[\xi]}(t ) \rangle ~dt \nonumber \\ & & ~ + \sqrt{\kappa(t ) } \left [ \hat a(\hat p,\hat q , t ) - { a}_{[\xi]}(t ) \right ] ~|\psi_{[\xi]}(t ) \rangle~dw(t )
. \label{stoceq}\end{aligned}\ ] ] in ( [ stoceq ] ) we still have to specify the function @xmath67}(t)$ ] .
according to the decomposition ( [ decoxi ] ) , the conservation of the trace of the reduced density matrix operator gives @xmath87 ~\langle \psi_{[\xi]}(t ) | \psi_{[\xi]}(t ) \rangle .\ ] ] since the gaussian measure @xmath74 $ ] is normalized , up to a zero - average fluctuation we have @xmath88}(t ) | \psi_{[\xi]}(t ) \rangle = 1 $ ] .
the normalization of the state @xmath89}(t ) \rangle$ ] uniquely determines the expression of @xmath67}(t)$ ] . indeed , by evaluating the ito differential @xmath90}(t )
| \psi_{[\xi]}(t ) \rangle = 2 \sqrt{\kappa(t ) } ~ \langle \psi_{[\xi]}(t ) dw(t ) + { \cal o}(dt^{3/2})\ ] ] we see that the norm of the state @xmath89}(t ) \rangle$ ] is conserved only if we impose @xmath91}(t ) = \langle \psi_{[\xi]}(t ) | \hat a(t ) |\psi_{[\xi]}(t ) \rangle .
\label{nlxi}\ ] ] in this case the stochastic differential equation ( [ stoceq ] ) becomes the nonlinear quantum state diffusion equation proposed in @xcite .
unlike the effective wave equation ( [ psia ] ) , the quantum state diffusion equation ( [ stoceq ] ) is suitable for the _ a priori _ analysis of the selective processes but awkward when dealing with the _ a posteriori _ analysis .
indeed , in the latter case one should guess the noise realization @xmath68 giving rise , according to ( [ aintoaxi ] ) and ( [ nlxi ] ) , to the known value of the measurement result @xmath15 .
as in the decomposition of the reduced density matrix in terms of the pure states @xmath45}(t ) \rangle$ ] , the nonselective measurement result ( [ avens ] ) and variance ( [ var ] ) can be expressed in terms of the pure states @xmath89}(t ) \rangle$ ] by @xmath92 ~ \langle \psi_{[\xi]}(t ) | \hat a(t ) |\psi_{[\xi]}(t ) \rangle \label{avexi}\ ] ] and @xmath93 ~ \langle \psi_{[\xi]}(t ) | \left ( \hat a(t ) - \overline{a(t ) } \right)^2 |\psi_{[\xi]}(t ) \rangle , \label{varxi}\ ] ] respectively . as fifth and final step , the equivalence loop of fig
1 is closed by regaining the starting differential equation for the reduced density matrix operator latexmath:[\[\hat \rho(t ) = \overline { overlining means functional integration with measure @xmath74 $ ] . by using ito algebra
we get @xmath95}(t ) \rangle ~ \langle \psi_{[\xi]}(t ) | + d | \psi_{[\xi]}(t ) \rangle ~ d \langle \psi_{[\xi]}(t ) | } \nonumber \\ & = & \left ( - { i \over \hbar } \left [ \hat h(t ) , \hat \rho(t ) \right ] - { 1 \over 2 } \kappa(t ) \left [ \hat a(t ) , \left [ \hat a(t ) , \hat \rho(t ) \right ] \right ] \right ) dt + { \cal o}(dt^{3/2})\end{aligned}\ ] ] which is eq .
( [ lindbeq ] ) with the previously discussed choice for the lindblad operators .
deeper insight into the problem of measurement in quantum mechanics can be gained by completing the formal discussion of the previous section with general but explicit models for a measurement process . for this purpose
the influence functional approach is a quite appropriate one and some attempts have already been done in this direction @xcite .
here we discuss the measurement process in terms of an informational environment made of particles linearly interacting with the measured system and in contact with a heat reservoir at fixed temperature .
each particle selectively measures the evolution of the system while the collection of them represents an apparatus performing a nonselective measurement .
this model has two nice features .
firstly , it is simple from a mathematical point of view and the corresponding influence functional can be evaluated exactly @xcite .
secondly , very clear and controllable approximations can be made on the influence functional in order to reduce it to the formal expression given in the previous section .
these approximations are physically related to the fact that the informational environment can be defined as a many - body system in which each particle , still interacting quantum mechanically with the measured system , behaves classically from the point of view of the observer . only in this case
the decision of the observer to look or not to look at the pointer of the instrument does not influence the result of the measurement itself @xcite .
we start by considering the hamiltonian of a measured system interacting with an environment @xmath96 where @xmath97 is the hamiltonian of the measured system with phase - space coordinates @xmath98 and @xmath99 and @xmath100 is the hamiltonian of the interacting environment .
the environment is made by different sets of particles with mass @xmath101 and phase - space coordinates @xmath102 and @xmath103 , where @xmath104 labels the sets and @xmath105 the particles in each set .
the particles of each set @xmath104 interact linearly with an observable @xmath106 of the system through a function @xmath107 which transduces a displacement of the observables @xmath108 into displacements of the coordinates @xmath102 @xmath109 ^ 2 \right ) .
\label{henv}\ ] ] the interaction of the @xmath105th particle of the environment with the @xmath104th observable is characterized by a proper angular frequency @xmath110 . in terms of spectral response
we can say that each particle measures the fourier component of @xmath111 at its proper angular frequency @xmath110 .
we assume that at time @xmath112 the system is described by the density matrix @xmath113 and each particle of the environment is at thermal equilibrium with temperature @xmath114 . since the equilibrium value for the coordinates @xmath102 depends on the coordinates @xmath98 and @xmath99 , the assumption of thermal equilibrium at time @xmath112 implies a correlation between the environment and the system at the same instant of time @xcite . the density matrix for the total system at time @xmath112 can be written as @xmath115 where @xmath116 and @xmath117 .\ ] ] the initial displacements @xmath118
are uniquely expressed in terms of the values assumed by the measured observable at the initial points @xmath119 and @xmath120 by requiring translational and time reversal invariance @xcite . the steps required to get the expression of the influence functional are standard @xcite .
at time @xmath121 the reduced density matrix of the system , obtained by tracing out the coordinates of the environment in the total density matrix @xmath122 can be obtained by propagating the initial density matrix @xmath113 @xmath123 by the two - point green function @xmath124 d[q_1]_{q_1',t'}^{q_1'',t '' } \int d[p_2 ] d[q_2]_{q_2',t'}^{q_2'',t '' } \exp \left ( \frac{i}{\hbar } s[p_1,q_1 ] - \frac{i}{\hbar } s[p_2,q_2 ] \right ) f[p_1,q_1,p_2,q_2]\end{aligned}\ ] ] where @xmath125 $ ] is given by ( [ s ] ) and the influence functional @xmath126 is @xmath127 = \int dq '' \int dq_1 ' dq_2 ' \int d[p_1 ] d[q_1]_{q_1',t'}^{q'',t '' } \int d[p_2 ] d[q_2]_{q_2',t'}^{q'',t '' } \nonumber \\&&~\times \exp \left ( \frac{i}{\hbar } s_{env}[p_1,q_1,a_1 ] - \frac{i}{\hbar } s_{env}[p_2,q_2,a_2 ] \right ) \rho_{env } ( \delta q_1',\delta q_2',t ' ) \label{inf}\end{aligned}\ ] ] with @xmath128 = \int_{t'}^{t '' } dt~ \left ( \sum_{\nu n } p_{\nu n } \dot q_{\nu n } - h_{env}(p , q-\lambda a , t ) \right ) .\ ] ] in the above formulas the notation @xmath129 , @xmath130 , @xmath131 is a shortening for @xmath132 , @xmath133 , @xmath134 and @xmath135 , @xmath136 for @xmath137 , @xmath138 .
the integrations in ( [ inf ] ) are gaussian and can be performed giving @xmath139 = \prod_{\nu } \exp \bigg ( - \int_{t'}^{t '' } dt \int_{t'}^{t } ds ~ \lambda_\nu(t ) \left [ a_{\nu}(p_1,q_1,t ) - a_{\nu}(p_2,q_2,t ) \right ] \big ( \alpha_\nu(t - s ) \lambda_\nu(s ) \nonumber \\ & & \times \left [ a_{\nu}(p_1,q_1,s ) -
a_{\nu}(p_2,q_2,s ) \right ] + i \beta_\nu(t - s ) { d \over ds } \left ( \lambda_\nu(s ) \left [ a_{\nu}(p_1,q_1,s ) + a_{\nu}(p_2,q_2,s ) \right ] \right ) \big ) \bigg ) \label{gfi}\end{aligned}\ ] ] where the two kernels @xmath140 \label{flu1}\ ] ] and @xmath141 \label{dis}\ ] ] describe fluctuation and dissipation phenomena , respectively @xcite .
now we turn the attention to the requirement that the environment is informational , i.e. , classical with respect to the observer so that the readout of information through the coordinates @xmath130 has an objective value .
since a classical system is one whose quantized structure can not be appreciated we must impose the thermal fluctuations to be large in comparison to the quanta of the environment @xmath142 in this case we have @xmath143 .
\label{flu2}\ ] ] memory effects in the fluctuation and dissipation kernels are an inessential complication which can be avoided by assuming an ensemble of particles in the environment with a continuous spectrum of frequencies . if , for simplicity , we choose the same frequency density @xmath144 for each set @xmath104 we get @xmath145 \simeq \int_{0}^{\infty } d\omega { d n_\nu \over d \omega } \omega^2 \cos[\omega ( t - s ) ] = \omega \delta ( t - s ) .\ ] ] of course , the above assumption of continuous frequency spectrum implies that the condition ( [ classic ] ) can not be satisfied in the whole frequency range @xmath146 by a finite temperature .
therefore , we should assume the condition ( [ classic ] ) to be valid only in a finite range @xmath147 $ ] which contains the most significant part of the fourier spectrum of @xmath148 and is relevant in monitoring the measured observables .
in this case the condition ( [ classic ] ) allows us to neglect the dissipation term , proportional to ( [ dis ] ) , with respect to the fluctuation one , proportional to ( [ flu2 ] ) . in conclusion
, the influence functional for an informational environment can be written as @xmath149 = \prod_{\nu } \exp \left ( - \frac12 \int_{t'}^{t '' } dt ~\kappa_\nu(t ) \left [ a_{\nu}(p_1,q_1,t ) - a_{\nu}(p_2,q_2,t ) \right]^2
\right ) \label{fvernon}\ ] ] with @xmath150 equations ( [ fvernon ] ) and ( [ kappone ] ) generalize to more than one monitored quantity the expression of the functional already introduced in ( [ iphase ] ) for nonselective measurements .
note that @xmath151 can be written as @xmath152 where @xmath153 and @xmath154 .
the two parameters @xmath155 and @xmath156 can be made indefinitely small by increasing the density of meters ( particles ) in the informational environment proportional to @xmath157 , or their temperature @xmath114 , respectively .
these characteristic length and time play the role of similar parameters introduced _ ad hoc _ in the spontaneous localization @xcite and in the restricted path - integral @xcite approaches .
we close this section by evaluating the measurement result arising as a readout from the informational environment and its corresponding variance .
the observer reads the result of the nonselective measurement of the @xmath104th observable by looking at a pointer which responds to the coordinates of the measurement apparatus .
this response is a weighted sum over the proper frequencies of the informational environment and is characterized by a normalized response function @xmath158 such that the pointer displacement and its variance are @xmath159 @xmath160 where @xmath161 \label{tr1}\ ] ] @xmath162 .
\label{tr2}\ ] ] due to the condition ( [ classic ] ) the thermal relaxation time of the measurement apparatus turns out to be much smaller than the characteristic timescales of the measured system .
this allows us to use an adiabatic approximation for the total density matrix operator and factorize it as @xmath163 where @xmath10 is the reduced density matrix operator of the system which takes into account the influence of the informational environment and @xmath164 is the density matrix operator of the informational environment at thermal equilibrium around the instantaneous value of the measured observables . in the representation of the environment coordinates @xmath164 has matrix elements @xmath165 which are operators with respect to the system coordinates through the displacements @xmath166 .\ ] ] in this case the traces in ( [ tr1 ] ) and ( [ tr2 ] ) contain gaussian integrals over the environment coordinates which can be performed giving @xmath167 = \lambda_\nu(t ) { \rm tr } \left [ \hat a_\nu(t ) \hat \rho(t ) \right]\ ] ] @xmath168 = { \hbar \over 2 m
\omega_{\nu n } } \coth \left ( { \hbar \omega_{\nu n } \over 2 k_b t } \right ) + \lambda_\nu(t)^2 { \rm tr } \left [ \hat a_\nu(t)^2 \hat \rho(t ) \right ] .\ ] ] in the limit ( [ classic ] ) and for a continuous spectrum of frequencies with the previously chosen density we have @xmath169 @xmath170 where @xmath171 and @xmath172 are given by ( [ avens ] ) and ( [ var ] ) , respectively . up to the transduction factor @xmath173 and @xmath174 ,
respectively , eq.s ( [ p ] ) and ( [ dp ] ) are the result of the measured observable @xmath175 and its variance read from the pointer .
the result of the measurement is the nonselective outcome of the observable @xmath175 obtained from measurement quantum mechanics .
the measurement variance is the sum of a classical variance associated to the measurement apparatus and the quantum variance associated to the measured system .
the nonselective result and variance ( [ p ] ) and ( [ dp ] ) can be expressed in terms of selective processes .
the decomposition of @xmath171 and @xmath172 has already been discussed in the previous section . concerning the decomposition of the variance associated to the measurement apparatus we note that the nonselective measurement performed by the informational environment is achieved by summing over the frequencies of the particles in the environment .
therefore , the contribution to the measurement variance given by the selective process at frequency @xmath176 is @xmath177 as expected from the equipartition theorem . equation ( [ dp ] ) sets the definition of classical and quantum measurements .
the first ones are those characterized by a dominance of the variance associated to the measurement apparatus , i.e. , by a dominance of the thermal ( or brownian ) noise . in the second ones
the quantum variance of the measured system is larger than the thermal noise .
this second case promises a richer variety of experimental scenarios because the measurement variance depends on the preparation of the measured system and/or the strength of the coupling to the measurement apparatus .
in many experimental situations one is dealing with an average of measurements on an individual quantum system each time prepared in the same initial state or a single measurement on an ensemble of independent identical quantum systems with the same initial conditions . in both cases or in a combination of them averaged measurement results , instead of individual measurement results , are actually registered as outcome of the experiment . according to the discussion of section ii we have two ways for theoretically reproducing the outcome of such an experiment .
we can consider _ a priori _ selective measurements and obtain the experiment outcome by averaging the corresponding selective results in the sense of eq .
( [ avexi ] ) . we can also consider a nonselective measurement and evaluate the experiment outcome directly from eq .
( [ avens ] ) .
the choice between the two methods may depend on the particular problem one is faced to .
the method based on a nonselective process is more direct but it is based upon the solution of a master equation for the reduced density matrix operator which could be much more difficult than solving many times the the diffusion state equation ( [ stoceq ] ) and averaging the selective measurement results . in this section
we will deal with an experimental situation which can be described by a simple 2 by 2 density matrix .
the choice of the direct method based on nonselective measurements is , therefore , the natural one .
let us consider a system with time - independent hamiltonian and discrete energy spectrum @xmath178 the evolution of the system subjected both to an external time - dependent perturbation @xmath179 and to a continuous nonselective measurement of the observable represented by the operator @xmath180 is given by @xmath181 - { 1 \over 2 } \kappa(t ) \left [ \hat a , \left [ \hat a , \hat \rho(t ) \right ] \right ] .\ ] ] an interesting situation is attained in the case the external perturbation @xmath182 stimulates the system to make transitions among the unperturbed levels @xmath105 and the occupancy of some of these levels is measured . at what extent does the measurement disturb the stimulated transitions ? eventually , inhibition of the stimulated transitions due to the occupancy measurement occurs and one has an example of what is called quantum zeno effect @xcite . from a theoretical point of view
a two - level system is the simplest one for studying the interplay between the effects of the measurement process and of the external perturbation .
this situation has been also experimentally investigated @xcite .
an ensemble of about 5000 @xmath183 ions was stored in a penning trap .
two hyperfine levels of the ground state of @xmath183 , created by a static magnetic field and hereafter called levels 1 and 2 , were driven by a radiofrequency , resonant between levels 1 and 2 , turned on for @xmath184 ms .
the amplitude of the radiofrequency was adjusted to make the initially vanishing occupancy of level 2 to be unity at time @xmath114 in absence of other disturbances . during the radiofrequency pulse , @xmath105 optical pulses of length @xmath185 ms and frequency equal to the transition frequency between level 1 and a third level 3
were also applied .
the number of photons emitted in the spontaneous transition @xmath186 , the transition @xmath187 being forbidden , was roughly proportional to the occupancy of level 1 .
the optical pulses acted , therefore , as a measurement of the occupancy of level 1 .
the occupancy at time @xmath114 of level 1 was observed to be frozen near its initial unity value , i.e. , the stimulated transition @xmath188 in the period @xmath114 to be inhibited , proportionally to the number @xmath105 of optical pulses . before trying a direct interpretation of the experiment @xcite , let us consider the case of a continuous measurement of the occupancy of level 1 in a two - level system subjected to stimulated transitions . in the representation of the unperturbed hamiltonian @xmath189 where @xmath190 we assume a perturbation @xmath182 with matrix elements @xmath191 and @xmath192 with @xmath193 real and @xmath194 . in the same representation the matrix elements of the measured occupancy of level 1 are @xmath195 and @xmath196 .
the master equation for the reduced density matrix operator then gives @xmath197 \nonumber \\ \dot{\rho}_{22}(t)&= & - { i\over \hbar } \left [ v_{21 } \rho_{12}(t ) - \rho_{21}(t ) v_{12 } \right ] \\
\dot{\rho}_{12}(t)&= & \left [ -{i\over\hbar } ( e_1-e_2 ) - { \kappa \over 2 } \right ] \rho_{12}(t ) - { i\over \hbar } \left [ v_{12 } \rho_{22}(t ) - \rho_{11}(t ) v_{12 } \right ] \nonumber\end{aligned}\ ] ] and @xmath198 .
the measurement coupling @xmath199 is assumed constant . by summing and subtracting the first two equations we find @xmath200 \rho_{12}(t ) - { i\over \hbar } v_{12 } \left ( \rho_{22}(t ) - \rho_{11}(t ) \right ) .
\nonumber\end{aligned}\ ] ] the first equation is the conservation of the trace of the reduced density matrix operator in the case of a nonselective measurement , @xmath201 . by defining @xmath202 \qquad
\rho_{22}(t ) - \rho_{11}(t ) = \gamma(t)\ ] ] with @xmath203 , @xmath204 and @xmath205 real , the other two equations give @xmath206 where we have introduced the rabi angular frequency @xmath207 the above system has a simple solution for @xmath208 ( resonance ) and in this case we get @xmath209 \label{rho11 } \\
\rho_{12}(t)&= & e^ { -\frac{i}{\hbar}(e_1-e_2)t -{1\over4 } \kappa t -i\omega t_0 } \bigg [ { \rm re } \left(\rho_{12}(0 ) e^{i \omega t_0}\right ) e^ { -{1\over4 } \kappa t } \nonumber \\ & & + i { \rm i m } \left ( \rho_{12}(0 ) e^{i \omega t_0}\right ) \left ( \cos(wt ) - \frac{\kappa}{4w } \sin(wt ) \right ) - i \left ( \rho_{22}(0)-\rho_{11}(0)\right ) \frac{\omega_r}{2w}\sin(wt ) \bigg ] \label{rho12}\end{aligned}\ ] ] where @xmath210 the angular frequency @xmath211 coincides with the rabi angular frequency @xmath212 when @xmath213 . in this case
the effect of the measurement disappears and the system oscillates between levels 1 and 2 with angular frequency @xmath212 . in the opposite limit of strong measurement coupling the frequency @xmath211 is imaginary and an overdamped regime is achieved in which transitions are inhibited .
the border between the two regimes is at @xmath214 corresponding to a critical measurement coupling @xmath215 the behavior of the measured occupancy of level 1 which , according to eq .
( [ avens ] ) , turns out to be @xmath216 is shown in fig . 2 for different values of the measurement coupling starting from initial conditions @xmath217 and @xmath218 . the transition between the rabi - like regime and the zeno - like regime is marked by the disappearing of the oscillatory behavior at @xmath219 .
the quantum variance associated to the nonselective measurement of level 1 can be evaluated according to eq .
( [ var ] ) and is @xmath220 .
it vanishes in the limit of strong measurement coupling and oscillates with angular frequency @xmath212 between 0 and @xmath221 in the opposite limit @xmath222 . at the critical measurement coupling @xmath219 , after a short transient of the order of @xmath223 , the quantum variance approaches the constant and maximum value @xmath224 .
it is worth to note that an identical behavior of the reduced density matrix operator is obtained if we consider @xmath189 as the measured quantity . in this case eq.s
( [ rho11 ] ) and ( [ rho12 ] ) still hold with the substitution @xmath225 where @xmath226 is the measurement coupling associated to the measurement of @xmath189 .
this observation allows us to compare fig .
2 with fig . 1 of ref .
@xcite where the quantum zeno effect for the same two - level system investigated here was analyzed _ a posteriori _ in the case of a single selective measurement of energy . as expected , quantitative differences between nonselective measurements and single selective measurements are obtained but in both cases zeno inhibition occurs for strong measurement coupling . the behavior shown in fig .
2 agrees also with that found in @xcite where the quantum zeno effect is analyzed within the quantum trajectory approach @xcite .
the discussion relative to fig .
2 can be made general . in the limit
@xmath227 the influence of the measurement disappears and ordinary quantum mechanics is recovered . in the limit
@xmath228 the influence of the measurement dominates and the average measurement result ( [ avens ] ) as well as the corresponding variance ( [ var ] ) are frozen into the values assumed at the beginning of the measurement . in the case of fig .
2 these values are @xmath217 and @xmath229 , respectively . moreover , in the limit @xmath228 the off - diagonal elements of the reduced density matrix operator vanish and a classical behavior is obtained . in order to recover the results of the experiment @xcite the previous analysis for a continuous nonselective measurement with @xmath199 constant
must be generalized to a series of measurement pulses spaced by intervals of no measurement . according to the experimental procedure we consider @xmath105 nonselective measurements of the occupancy of level 1 with coupling @xmath199 during the intervals
@xmath230 $ ] , @xmath231 , with @xmath185 ms and @xmath184 ms . during the remaining part of the interval @xmath232 $ ] the system is subjected only to the radiofrequency perturbation at @xmath233 mhz whose amplitude @xmath193 is fixed by the condition @xmath234 that gives , in absence of measurement , @xmath235 if @xmath236 . since the radiofrequency is resonant @xmath237 .
the time evolution of the reduced density matrix operator corresponding to the described process is obtained by successive iterations of ( [ rho11 ] ) and ( [ rho12 ] ) with the same formulas evaluated for @xmath213 . the theoretical probability for the transition @xmath188 at the end of the interval @xmath114 @xmath238 can be compared with the corresponding experimental data @xmath239 available for @xmath240 . the uncertainty for the experimental transition probabilities is estimated to be @xmath241 @xcite .
3 we show the sum of the squared differences between the theoretical and experimental transition probabilities normalized to the experimental uncertainties , the @xmath242 , as a function the phenomenological parameter @xmath243 . in the same figure
we show also the probability @xmath244 that the observed value for the chi - square should exceed the value @xmath245 by chance . in our case
the number of degrees of freedom is @xmath246 .
the two arrows indicate the chi - square value obtained by the same authors of @xcite using a theoretical model based on the instantaneous von neumann collapse without ( higher value ) or with ( lower value ) corrections for the finite duration of the measurement pulses , the effect of optical pumping from level 2 to level 1 and the measured value of @xmath212 . depending upon the statistical confidence level we adopt , from fig .
3 we see that values of @xmath247 are required to fit properly the experimental data .
the minimum value of @xmath242 obtained from measurement quantum mechanics slightly differs from the @xmath242 obtained on the basis of the von neumann postulate including experimental corrections .
it is neither surprising nor exciting that measurement quantum mechanics is able to explain the experimental results of ref .
@xcite in terms of a strong zeno inhibition .
not surprising because measurement quantum mechanics contains more naive approaches to the problem of quantum measurements as the von neumann postulate , for instance , which was already shown to reproduce the experimental results .
not exciting because strong zeno inhibition as well as full rabi oscillations are two trivial extreme regimes .
however , measurement quantum mechanics tells us that another interesting and unexplored regime exists .
it is the regime which occurs when the measurement coupling is comparable to the critical value . in this case a strong competition between stimulated transitions and measurement inhibition takes place as clearly recognized from fig . 2 and also from the analysis of the corresponding quantum variance , @xmath248 , which gets maximum at @xmath249 .
measurement quantum mechanics suggests us also how to explore this regime . in order to make @xmath250
we should decrease @xmath199 and/or increase @xmath251 with respect to the values used in @xcite . according to the discussion of section iii @xmath199
can be decreased by reducing the density of particles in the informational environment . in the experiment
@xcite the `` particles '' of the informational environment are the excited modes of the photon vacuum . their density can be reduced by lowering the intensity of the optical radiation acting as a measurement probe . on the other hand , eq .
( [ kcrit ] ) shows that @xmath251 can be increased by means of the amplitude @xmath193 of the perturbation which stimulates the rabi oscillations . in the experiment @xcite
this amounts to increase the intensity of the radiofrequency radiation . of course
, by increasing @xmath193 the rabi angular frequency increases too so that the condition @xmath234 used in @xcite can be maintained only by decreasing the period @xmath114 and , consequently , the duration of the measurement pulses .
the possibility to vary the strength of the coupling is also related to the opportunity of looking for decoherence effects in the same experiment .
indeed , from eq .
( [ rho12 ] ) we see that the off - diagonal elements of the density matrix vanish exponentially with time constant @xmath252 . for an appropriate measurement coupling ,
i.e. , a proper intensity of the optical radiation , @xmath253 could be experimentally accessible by optical homodyne tomography techniques @xcite and decoherence phenomena could be investigated .
measurement quantum mechanics has been introduced by showing the equivalence among formalisms developed with the aim of including the effect of the measurement on the dynamics of a quantum system .
the theory contains one parameter , the measurement coupling , for each measured observable .
the measurement couplings can be expressed in terms of the detailed properties of a measurement apparatus which extracts information from the measured system in objective way , i.e. , independently of the presence of an observer looking at the pointer of the apparatus .
the definition of quantum and classical measurements then emerges naturally according to the dominance , in the uncertainty of the pointer , of the classical fluctuations due to the apparatus or of the quantum fluctuations due to the measured system .
ordinary quantum mechanics is obtained as a limit case for vanishing measurement couplings . in the opposite limit of infinite measurement couplings
the dynamics of the measured system is frozen and classical .
both limits are not interesting , the first one because there is no measurement and the second one because there is no quantum dynamics to measure .
interesting physics is in between and measurement quantum mechanics is a tool for designing experiments in this intermediate regime .
|
measurement quantum mechanics , the theory of a quantum system which undergoes a measurement process , is introduced by a loop of mathematical equivalencies connecting previously proposed approaches . the unique phenomenological parameter of the theory is linked to the physical properties of an informational environment acting as a measurement apparatus which allows for an objective role of the observer .
comparison with a recently reported experiment suggests how to investigate novel interesting regimes for the quantum zeno effect .
|
the game _ toads and frogs _ , invented by richard guy , is extensively discussed in `` winning ways''[1 ] , the famous classic by elwyn berlekemp , john conway , and richard guy , that is the _ bible _ of combinatorial game theory .
+ this game got so much coverage because of the simplicity and elegance of its rules , the beauty of its analysis , and as an example of a combinatorial game whose positions do not always have values that are numbers .
+ the game is played on a @xmath2 strip with either toad(t ) , frog(f ) or @xmath3 on the squares .
left plays t and right plays f. t may move to the immediate square on its right , if it happens to be empty , and f moves to the next empty square on the left , if it is empty .
if t and f are next to each other , they have an option to jump over one another , in their designated directions , provided they land on an empty square .
( see [ 1 ] page 14 ) .
+ in symbols : the following moves are legal for toad : + @xmath4 , + @xmath5 , + and the following moves are legal for frog : + @xmath6 , + @xmath7 . + already in `` winning ways''[1 ] there is some analysis of toads and frogs positions , but on _ specific _ , small boards , such as @xmath8 . in 1996 , erickson[2 ] analyzed more general positions . at the end he made five conjectures about the values of some families of positions .
all of them are starting positions ( positions where all t are rightmost and all f are leftmost ) .
+ in [ 3 ] , we discussed the symbolic finite - state approach to prove the value of the positions in class a and b. however , the patterns of the value of toads and frogs game are not limited to only class a or class b. there are also patterns in the positions where the variables are on both t and f , for example @xmath9 . in this paper , we will analyze some of these positions .
+ to be able to understand the present article , reader need a minimum knowledge of combinatorial game theory , that can be found in [ 1 ] .
in particular , readers should be familiar with the notions of _ value _ of a game and sum games .
+ let s recall the bypass reversible move rule , dominated options rule and inequality of two games .
( see [ 1 ] page 33 , 62 - 64 ) . + * the bypassing right s reversible move rule * + [ ht ] @xmath10 if @xmath11 + * the dominated options rule * + let @xmath12 + if @xmath13 and @xmath14 then @xmath15 + * sum and inequality of two games * + let @xmath16 + let @xmath17 + @xmath18 + @xmath19 if right goes first and left wins .
+ @xmath20 if left goes first and right wins .
+ @xmath21 is equivalent to @xmath22 + the only notations we use are @xmath23 and @xmath24 .
we will not use any shorthand notation like @xmath25 , @xmath26 , etc .
* definition * : + general class ai : all positions with ( numeric ) i number of @xmath3 ( with symbols on both t and f ) .
+ general class bi : all positions with ( numeric ) i number of f ( with symbols on both t and @xmath3 ) .
+ the general classes a and b are the more general positions of the classes a and b that we talked about in [ 3 ] . + for the general class
, we can not apply the finite state method that we use in [ 3 ] anymore since we now have infinitely many positions that come from the combination of the two letters with symbols on them .
however we manage to categorize all positions in the general class a1 , the class of all positions with exactly on @xmath3 .
it is in fact the only general class that we manage to do it for .
+ many positions in these classes do not have a nice compact formula ; for example in a2 , @xmath27 . on the other hand
, many positions have a nice formula .
we will prove some of the starting positions like @xmath28 later on in the appendix of this paper .
+ once we detect the patterns of the positions , the proof is quite routine .
we now do the proof of each specific position by hand with the help of a computer .
we hope to see the computer playing more roles in assisting with the proofs in the future .
we present the values of some starting positions in this section .
we have a fast program written in java to calculate the outcome of the sum of two given positions ( = , @xmath29,@xmath30,@xmath31 ) .
this program does not calculate the value of the sum of two games .
it only gives the outcome .
it works well with the positions that have a simple value .
the author s brother and the author wrote this program originally to check the value of the game of the form @xmath32 for which so far the values of the game are 0 or * except the column b=2 which will be proved to be infinitesimal when @xmath33 .
the program is on the author s web site and we present the tables here .
+ + [ htbp ] [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ] note 1 ) @xmath34 means `` can not be compared to g '' .
+ note 2 ) we drop the values of the first two columns where @xmath35 since they will all be proved in the appendix .
l means long .
we are not writing it out here .
+ erickson s conjecture 4 is false since @xmath36 .
+ ( * erickson s conjecture 4 * : @xmath37 or @xmath38 for all @xmath39 . ) + the author believes there are no patterns in @xmath40 for positions of the form @xmath41 ; for any fixed @xmath42 .
in [ 2 ] , jeff erickson made 6 conjectures .
jesse hull proved conjecture 6 ( toads and frogs is np - hard ) in 2000 .
doron zeilberger and the author of this paper proved conjecture 1 ( later on in this paper ) , 2 ( in [ 3 ] ) , 3 ( later on in this paper ) and disprove conjecture 4 .
conjecture 5 is still open .
we restate conjecture 5 here .
+ * erickson s conjecture 5 * : + @xmath32 is an infinitesimal for all a , b except ( a , b ) = ( 3,2 ) + this conjecture seems very interesting and hard but not impossible to prove .
we split erickson s fifth conjecture into 2 stronger conjectures which are conjectures 3 and 4 here .
+ we believe there are still a lot of nice patterns and conjectures in this game that we overlooked .
once ram gets cheaper and maple gets faster , we will have more information . + * conjecture 1 ) * assume @xmath43 , @xmath39 , @xmath44 and @xmath45 + 1.1)@xmath46 + + 1.2 ) for @xmath47 , @xmath48 + + 1.3 ) for @xmath49 + * conjecture 2 ) * for @xmath50 , @xmath51 + * conjecture 3 ) * @xmath52 for any @xmath53 , except for @xmath54 . + * conjecture 4 ) * @xmath55 or @xmath56 for any @xmath57 . +
* conjecture 5 ) * for a fixed integer @xmath58 , @xmath59 such that @xmath60 for all @xmath61 . +
* future work * + 1 ) categorize all the positions that have exactly one f ( general class b1 ) ( conjecture 1 might be a good start ) .
i want to thank my twin brother thotsaphon thanatipanonda for writing the fast program in java to find the counterexample of conjecture 4 of erickson .
i also want to thank my advisor doron zeilberger for always supporting and motivating .
there will not be this paper without him .
in this section we classify all the positions that have one @xmath3 .
the lemmas below give a recurrence to the positions .
we can compute the values of the positions in polynomial time .
we omit the proofs here .
+ * notation * + @xmath62 + @xmath63 times .
+ @xmath64 = any combination of @xmath65 and @xmath66 that has @xmath66 as its right most entry .
for example @xmath67 .
+ @xmath68 = any combination of @xmath65 and @xmath66 that has @xmath65 as its left most entry . for example @xmath69 . + death leaps principle(dlp ) : the position with one empty square for which the only possible move for both sides is a jump has value 0 . for example @xmath70 .
@xmath71 @xmath72 @xmath73 @xmath74 + * example 1 : * @xmath75 + * example 2 : * @xmath76 + * example 3 : * @xmath77 + * note ) * we get the implicit value of example 2 from example 1 and implicit value of example 3 from example 2 .
we will get the value recursively this way .
+ @xmath78 @xmath79 @xmath80 when ( @xmath81 is even and @xmath82 ) or ( @xmath40 is odd and @xmath83 ) .
+ * note : * + 1 ) when @xmath40 is even , @xmath84 and @xmath81 is odd , @xmath85 , the recursive is going to bounce back and forth between positions .
+ we will refer to @xmath68 if @xmath86 .
we will refer to @xmath64 if @xmath87 .
then the positions will start over again .
+ 2 ) when @xmath40 is even , @xmath81 is even and @xmath88 then we will refer to @xmath64 . +
when @xmath40 is odd , @xmath81 is odd and @xmath89 then we will refer to @xmath68 .
we will refer to the lemma below a lot .
we state it here .
+ * one side death leap principle * ( one side dlp ) : if x is the position where the only possible move of left is a jump and there is no two or more consecutive empty square in x then x @xmath90 .
* proof : * we have to show that when left moves first and two players take turn playing , left will lost(left will run out of the legal move first ) .
this is true since after left jumps over one of the @xmath66 , right can response by moving to the empty square where the @xmath66 was jumped over .
+ * * example 1)**@xmath91 + * example 2 ) * @xmath92 + * convention * : + in the following sections of the appendix , we will prove the positions using the `` shorthand '' notation .
we explain by using an example .
+ * example * : to show : @xmath93 + we will have to show @xmath94 + that is to show @xmath95 + to show @xmath20 we need to show that when left moves first and two players take turn playing , right will win .
( on the other hand , to show @xmath19 we need to show that when right moves first and two players take turn playing , left will win . )
+ we will show that in these two sum games , for all the possible choices of left move , right can find a response to the move so that he will win at the end .
+ we will do some case analysis here . in the above position left
has three choices .
+ in the proof we will see + @xmath96 + we will write the response of right immediately
. you will see + * case 1 : * @xmath97 + * explanation : * right responses by picking the left option of @xmath98 on the right hand side . + * case 2 : * @xmath99 + * explanation : * right responses by moving the left most f. + * case 3 : * @xmath100 + * explanation : * left picks the right option of @xmath98 on the right hand side .
right responses by moving the right most f. + * note : * + 1 ) when the position simplify to the known one in [ 3 ] or [ 5 ] ( which is mostly positions in class a ) , we will claim such result without proving them again .
+ 2 ) the positions we considered in toads and frogs are a hot game which means the players select a good move and fight for an advantage
. we will not consider the possible move in a cold game which is the whole integer .
we show @xmath101 is an infinitesimal , @xmath102 the observation comes from figure 2 when @xmath54 . for any fixed integer @xmath103 , @xmath104 , @xmath105 * prove : * by induction on @xmath40 . + * base case : @xmath106 * , @xmath107 . + * case 1 : * @xmath108 , true by one side dlp . + * case 2 : * @xmath109 .
the left hand side is @xmath90 by one side dlp . +
* induction step : * @xmath110 . + * case 1 : * @xmath111 , true by one side dlp . + * case 2 : * @xmath112 , true by induction . + * case 3 : * @xmath113 .
the left hand side is @xmath90 by one side dlp .
+ @xmath101 is an infinitesimal , @xmath102 * by symmetry we only need to show : * for any fixed integer @xmath114 .
+ @xmath115 . + * i ) * @xmath116 + * ii ) * @xmath117 + * i ) * case 1 : @xmath118 + case 1.1 : @xmath119 + @xmath120 , true by lemma c.1
. + @xmath121 + @xmath122 + @xmath123 .
the left hand side is @xmath124 .
+ @xmath125 , true by lemma c.1 . + @xmath126
. the left hand side is @xmath90 by one side dlp .
+ @xmath127 + @xmath128 + @xmath129 the statement is true by one side dlp . + @xmath130 .
the left hand side is @xmath90 by one side dlp .
+ @xmath131 .
the left hand side is @xmath90 by one side dlp
. + @xmath132 + @xmath133 .
the left hand side is @xmath90 by one side dlp .
+ @xmath134 .
this is the case 1.1.2 + @xmath135 , true by lemma c.1 .
+ case 2 : @xmath136 .
the left hand side is 0 .
+ case 3 : @xmath137 + @xmath138 this is clearly true .
+ @xmath139 .
the left hand side is @xmath90 . +
* ii ) * case 1 : @xmath140 .
this is i ) case 3 .
+ case2 : @xmath141 + @xmath139 .
this is i ) case 3.2 .
+ @xmath142 .
the left hand side is @xmath90 by one side dlp . +
the theorem is proved .
@xmath143 , @xmath144 elwyn berlekamp , john conway , and richard guy , _ winning ways for your mathematical plays _ , academic press , new york , 1982 .
+ jeff erickson , _ new toads and frogs results _ , game of no chance , 1996 .
+ thotsaporn `` aek '' thanatipanonda , doron zeilberger , _ a symbolic finite - state approach for automated proving of theorems in combinatorial game theory _ ,
preprint ( available from the author s website ) .
+ thotsaporn `` aek '' thanatipanonda , _ three more results of combinatorial game toads and frogs _ , preprint ( available from the author s website ) . + thotsaporn `` aek '' thanatipanonda , _ library values for toads and frogs _ , preprint ( available from the author s website ) .
|
we show the value of positions of the combinatorial game `` toads and frogs '' .
we present new values of starting positions .
moreover , we discuss the values of all positions with exactly one @xmath0 , @xmath1 . at the end
, we post five new conjectures and discuss the possible future work .
|
in various medical studies , outcomes of interest include the time to death or time to tumor recurrence .
when observing survival times , censoring is common because of partially known information .
thus , survival data consist of survival time as an outcome , censoring status , and many covariates as risk factors . the relationship between survival time and
the covariates has been studied extensively . in survival data analysis ,
customized statistical methodologies are employed because of non - normal distributions and censoring .
the surveillance , epidemiology and end results ( seer ) program , a premier source for cancer statistics in the united states , contains information on incidence , prevalence , and survival from specific geographic areas representing 28 percent of the us population @xcite .
survival as an endpoint is one of the important outcomes of the seer database ; hence , it is often of interest to determine the relationship between survival and covariates . when analysing seer data , it is important to screen for outlying observations or outliers that significantly deviate from other measurements because they may distort the conclusions of the data analysis @xcite . therefore , the development of outlier detection methods is essential to obtain reliable results .
outlier detection has been studied in various types of data including normal data @xcite , multivariate normal data @xcite , censored data @xcite , incomplete survey data @xcite , time series data @xcite , gene expression data @xcite , proteomics data @xcite , functional data @xcite , spatial data @xcite and circular data @xcite . for more details , see @xcite , @xcite , and @xcite .
the algorithm of @xcite was developed to identify outlying observations based on cox linear regression for censored data .
it can be more effective to utilize quantile regression because of it is robust to outliers . @xcite and
@xcite proposed to use quantile regression for outlier detection in proteomics data .
the most algorithms focus on determining whether observations are outliers according to a threshold , which should be specified in advance .
the dichotomous algorithms , which depend solely on a pre - specified cut - off , may often be unsatisfactory .
thus , a function for providing scores and flexibly determining a threshold can be helpful . in this paper , we present three outlier detection algorithms for censored data : the residual - based , boxplot , and scoring algorithms .
the residual - based and boxplot algorithms were developed by modifying existing algorithms @xcite , and the scoring algorithm was developed to provide the outlying magnitude of each point from the distribution of observations and enable the determination of a threshold by visualizing the scores .
the presented algorithms are based on quantile regression , which is robust to outliers .
the algorithms were investigated by a simulation study , and their characteristics were summarized at the conclusion of that study .
we implemented the three algorithms customized for censored survival data in an package called outlierdc , which can be conveniently employed in the environment and freely obtained from the comprehensive r archive network ( cran ) website ( http://cran.r-project.org/web/packages/outlierdc/index.html ) .
we demonstrate its use with real data from the seer database ( http://seer.cancer.gov ) , which contains a number of data sets related to various cancers . the remainder of this paper is organized as follows . in sections
[ theory ] and [ package ] , we describe three algorithms using censored quantile regression for identifying outlying observations in censored data and then implement them into an package outlierdc . in section [ simul ] ,
simulation studies are conducted to investigate the performance of the outlier detection algorithms . in section [ ex ] , we illustrate the application of the algorithm using outlierdc with a real example .
we present our conclusions in section [ dis ] .
in this section , we describe three outlier detection algorithms based on censored quantile regression .
we here focus only on detecting too large observations , because too small observations can be generated by censoring .
we first define the notation used to explain the algorithms .
let @xmath0 be an uncensored dependent variable of interest , such as survival time or some transformation of it , and let @xmath1 and @xmath2 be a censoring variable and a @xmath3-dimensional covariate vector for the @xmath4th observation , respectively .
we observe the triples @xmath5 and define @xmath6 which represent the observed response variable and the censoring indicator , respectively .
we consider the quantile regression model @xmath7 where @xmath8 for some @xmath9 is a @xmath3-dimensional quantile coefficient vector and @xmath10 is a random error whose @xmath11th conditional quantile equals zero .
the conditional quantile function is defined as @xmath12 where @xmath13 is the @xmath11th conditional quantile of @xmath14 given @xmath15 , and @xmath16 is the conditional cumulative distribution function of the survival time @xmath17 given @xmath15 .
several approaches such as @xcite , @xcite , and @xcite can be used to estimate the conditional quantile coefficients @xmath18 .
for instance , let us consider @xcite as the basis of quantile regression for the outlier detection algorithms .
previous methods have stringent assumptions , such as unconditional independence of the survival time and the censoring variable , or global linearity at all quantile levels . to alleviate the assumptions , @xcite proposed locally weighted censored quantile regression based on the local kaplan - meier estimator with nadaraya - watsons type weights and a biquadratic kernel function . the local kaplan - meier estimates of the distribution function @xmath19
are obtained by @xmath20 where @xmath21 and @xmath22 is a sequence of non - negative weights adding up to 1 . here ,
the nadaraya - watson type weight is employed by @xmath23 where @xmath24 is a density kernel function and @xmath25 is a bandwidth converging to zero as @xmath26 . by plugging the estimator ( [ local ] ) into the weight function ( [ weight ] ) , the estimated local weights @xmath27 are obtained .
a weight function @xmath28 is used for each censored observation as follows .
@xmath29 where @xmath30 is the conditional cumulative distribution function of the censoring time @xmath31 given @xmath15 .
the regression coefficient estimates @xmath32 can be obtained by minimizing the objective function @xmath33.\ ] ] it is natural to consider the distance from each observation to the center to identify outliers .
utilizing the residuals from fitting quantile regression , the quantreg procedure in provides an outlier detection algorithm for uncensored data .
@xcite proposed the residual - based outlier detection algorithm based on cox linear regression for censored data .
it can be more effective to utilize quantile regression because it is robust to outliers .
thus , we modify the residual - based algorithm for censored data by utilizing the residuals from fitting censored quantile regression in the following manner .
let @xmath34 be the @xmath4th residual defined as @xmath35 where @xmath36 is the @xmath37th conditional quantile for the @xmath4th observation by censored quantile regression .
the outlier indicator for the @xmath4th observation , @xmath38 , is defined as @xmath39 where @xmath40 is a resistant parameter for controlling the tightness of cut - offs and @xmath41 is the corrected median of the absolute residuals .
that is , @xmath42 where @xmath43 is the inverse cumulative distribution function ( cdf ) of gaussian density with the @xmath3th quantile . as default values
, we consider @xmath44 and @xmath45 like in the procedure . the @xmath4th observation is declared an outlier if @xmath46 . in our package outlierdc , this algorithm is implemented in the function with the argument . the algorithm is summarized as follows : + * algorithm 1 * : residual - based outlier detection 1 . fit a censored quantile regression model with @xmath47 to the data .
2 . calculate the residuals @xmath48 .
3 . compute the scale parameter estimate @xmath41 by the residuals and the inverse cdf .
4 . declare each observation an outlier if its corresponding residual is larger than @xmath49 . a simple outlier detection approach based on a boxplot @xcite has widely been used for uncensored data .
@xcite and @xcite proposed to use the boxplot algorithm based on quantile regression for high - throughput high - dimensional data .
we modify the boxplot algorithm using quantile regression for censored data in the following manner .
a censored quantile regression model is fitted to obtain the @xmath50th and @xmath51th conditional quantile estimates , @xmath52 and @xmath53 , respectively .
the inter - quantile range ( iqr ) for the @xmath4th observation can be obtained by @xmath54 the outlier indicator for the @xmath4th observation , @xmath55 , is defined as @xmath56 where the upper fence
is defined as @xmath57 and @xmath58 is to control the tightness of cut - offs with a default value of 1.5 .
if an observation is located above the fence , we declare it an outlier .
the algorithm is powerful , particularly when the variability of data is heterogeneous .
we implement the algorithm in the function with the argument . it can be summarized as follows .
+ * algorithm 2 * : boxplot outlier detection 1 .
fit censored quantile regression models with @xmath59 and @xmath60 to the data .
2 . obtain the @xmath50th and @xmath51th conditional quantile estimates .
calculate @xmath61 4 .
construct the fence , @xmath62 5 .
declare each observation an outlier if it is located above the fence .
the residual - based and boxplot algorithms described in the previous sections focus on determining whether observations are outliers according to a threshold , which should be specified in advance .
these dichotomous algorithms , which depend solely on a pre - specified cut - off , may often be unsatisfactory .
moreover , the boxplot algorithm can be applicable when a single covariate exists .
thus , we developed the scoring algorithm , which provides the outlying degree that indicates the magnitude of deviation from the distribution of observations given the covariates .
visualizing the scores enables the flexible determination of a threshold for outlier detection .
the resulting scores are free from the levels of the covariates even though the variability of the data is heterogeneous .
the outlying score is based on the relative measure of conditional quantiles . the outlying score for the @xmath4th observation
is defined as @xmath63 the score is the difference between the distance of the observation from the @xmath37th quantile relative and that of the @xmath51th quantile , conditional on its corresponding covariates .
larger scores indicate higher outlying possibilities .
the normal qq plot of the scores enables identification of outlying observations .
when the scores are visualised , a threshold @xmath64 can be determined , and an observation @xmath4 is declared an outlier if @xmath65 .
the algorithm is implemented in the function with the argument and summarised as follows .
+ * algorithm 3 * : scoring outlier detection 1 .
fit censored quantile regression models with @xmath66 and @xmath60 to the data .
2 . obtain the @xmath50th , @xmath37th , and @xmath51th conditional quantile estimates .
calculate the outlying score @xmath67 .
4 . generate the normal qq plot of the outlying scores .
5 . determine a threshold @xmath64 to identify outlying observations that are outside the distribution of the majority of observations .
declare each observation an outlier if its corresponding score is greater than the threshold .
we develop an package outlierdc , which is designed to detect outliers in censored data under the environment system @xcite .
the outlierdc package utilizes existing packages , including methods @xcite , formula @xcite , survival @xcite , and quantreg @xcite .
the package methods is adopted to provide formal structures for object - oriented programming .
formula is used to manipulate the design matrix on the object .
the package survival enables the handling of survival objects by the function .
the package quantreg provides typical censored quantile regressions .
the function plays a pivotal role in outlier detection .
the usage and input arguments are as follows : odc < - function(formula , data , method = c(``score '' , `` boxplot'',``residual '' ) , rq.model = c(``wang '' , `` penghuang '' , `` portnoy '' ) , k_r = 1.5 , k_b = 1.5 , h = .05 ) * [ formula ] : a type of object with a object on the left - hand side of the @xmath68 operator and covariate terms on the right - hand side .
the survival object with its survival time and censoring status is constructed by the function . * [ data.frame ] : a data frame with variables used in the .
it needs at least three variables , including survival time , censoring status , and covariates . * [ character ] : an outlier detection method to be used .
the options , , and implement the scoring , boxplot , and residual - based algorithms , respectively
. the default is . *
[ character ] : a type of censored quantile regression to be used for fitting .
the options , , and conduct wang and wang s , portnoy s , and peng and huang s censored quantile regression approaches , respectively .
the default is . *
[ numeric ] : a value to control the tightness of cut - offs having a default value of 1.5 for the residual - based algorithm . * [ numeric ] : a value to control the tightness of cut - offs having a default value of 1.5 for the boxplot algorithm . * [ numeric ] : bandwidth for locally weighted censored quantile regression with a default value of 0.05 .
.output slots for the class [ cols="<,<,<",options="header " , ]
in this section , we illustrate the use of outlierdc for the detection of outlying observations using real data from the us seer database system @xcite .
the data for patients with extrahepatic cholangiocarcinoma can be obtained from the seer website ( http://seer.cancer.gov ) . to call the package and the data set , > library(outlierdc ) >
data(ebd ) >
dim(ebd ) [ 1 ] 402 6 the data consist of 402 observations with six variables . to take a glance at the data , display the first six observations as follows : > head(ebd ) i d meta exam status time ratio 1787 55468952 0 12 1 26 0.0000000 1788 8883016 0 12 1 11 0.0000000 1789 10647194 0 12 0 134 0.0000000 1790 16033679 2 12 1 1 0.1666667 1791 19519884 0 12 0 111 0.0000000 1792 19574077 0 12 1 8 0.0000000 to illustrate the outlier detection algorithms , we utilized the number of metastatic lymph nodes ( called ) as a covariate .
the response variable is the survival time in months ( ) , and its censoring status is denoted by , where 0 means censored .
the outlier detection algorithm can be run by the function as follows : > fit < - odc(formula = surv(log(time ) , status ) meta , data = ebd ) this command with the essential arguments and runs the scoring outlier detection algorithm with wang and wang s censored quantile regression to create the object .
the arguments method and can be omitted when the defaults are used . the argument for a threshold does not need to be specified in advance for the scoring algorithm .
its full command is fit < - odc(formula = surv(log(time ) , status ) meta , data = ebd , + method = `` score '' , h = 0.05 ) that is , outlier detection is performed by the scoring algorithm ( ) based on the locally weighted censored quantile regression with the bandwidth ( ) for selecting outliers .
the command provides a summary for the class . to use it ,
type the object name on the command line : > fit outlier detection for censored data call : odc(formula = surv(log(time ) , status ) meta , data = ebd ) algorithm : scoring algorithm ( score ) model : locally weighted censored quantile regression ( wang ) value for cut - off k_s : # of outliers detected : 0 top 6 outlying scores : times delta ( intercept ) meta score outlier 346 4.48 0 1 9 4.59 327 2.71 1 1 13 4.54 326 2.08 1 1 14 2.52 296 4.86 1 1 4 2.35 354 3.09 1 1 10 2.11 233 5.29 0 1 1 1.95 the output via consists of two parts : basic model information and top outlying scores .
the first part shows the overall information such as the formula used ( ) , the algorithm ( ) , the fitted quantile regression model ( ) , the threshold value to be applied ( ) , and the number of outliers detected ( ) . the command displays the model formula with input arguments and the used outlier detection algorithm .
next , the top six scores are displayed with the original data in decreasing order .
the number of outliers detected ( ) is zero because a threshold @xmath64 ( ) has not been provided thus far .
the decision is postponed until the result is updated by the function .
a threshold can be determined by visualizing the scores . to visualize the scores ,
plot(fit ) the function draws a normal quantile - quantile ( qq ) plot of the outlying scores , as shown in figure [ fig : qqplot ] .
the qq plot of outlying scores in figure [ fig : qqplot ] shows that the two points in the top right lie away from the line that passes through the first and third quartiles .
a threshold is added by to this plot .
thus , the result can be updated by >
fit1 < - update(fit , k_s = 4 ) > plot(fit1 ) > fit1 outlier detection for censored data call : odc(formula = surv(log(time ) , status ) meta , data = ebd ) algorithm : scoring algorithm ( score ) model : locally weighted censored quantile regression ( wang ) value for cut - off k_s : 4 # of outliers detected : 2 top 6 outlying scores : times delta ( intercept ) meta score outlier 346 4.48 0 1 9 4.59 * 327 2.71 1 1 13 4.54 * 326 2.08 1 1 14 2.52 296 4.86 1 1 4 2.35 354 3.09 1 1 10 2.11 233 5.29 0 1 1 1.95 the two points with scores greater than the cut - off ( = 4 ) were the 346th and 327th observations , which are marked by an asterisk .
the residual - based algorithm with a coefficient @xmath69 of 1.5 can be applied using with as follows : > fit2 < - odc(surv(log(time ) , status ) meta , data = ebd , method = `` residual '' , k_r = 1.5 ) > plot(fit2 ) > fit2 outlier detection for censored data call : odc(formula = surv(log(time ) , status ) meta , data = ebd , method = `` residual '' , k_r = 1.5 ) algorithm : residual - based algorithm ( residual ) model : locally weighted censored quantile regression ( wang ) value for cut - off k_r : 1.5 # of outliers detected : 9 outliers detected : times delta ( intercept ) meta residual sigma outlier 57 4.80 0 1 2 1.63 1.6 * 80 5.04 1 1 0 1.64 1.6 * 189 5.38 0 1 0 1.98 1.6 * 191 5.20 0 1 0 1.80 1.6 * 233 5.29 0 1 1 2.00 1.6 * 296 4.86 1 1 4 1.90 1.6 * 6 of all 9 outliers were displayed .
nine observations by were selected as outliers , six of which are shown in the above output .
all the outliers detected can be displayed by running .
the boxplot algorithm with a coefficient @xmath70 of 1.5 can be applied using with , as follows : >
fit3 < - odc(surv(log(time ) , status ) meta , data = ebd , method = `` boxplot '' , k_b = 1.5 ) > plot(fit3 ) > fit3 outlier detection for censored data call : odc(formula = surv(log(time ) , status ) meta , data = ebd , method = `` boxplot '' , k_b = 1.5 ) algorithm : boxplot algorithm ( boxplot ) model : locally weighted censored quantile regression ( wang ) value for cut - off k_b : 1.5 # of outliers detected : 1 outliers detected : times delta ( intercept ) meta ub outlier 346 4.48 0 1 9 4.32 * 1 of all 1 outliers were displayed .
the boxplot algorithm with a coefficient @xmath70 of 1.5 detected only one outlying point .
the 346th observation detected was also detected by both the scoring and residual - based algorithms .
the boxplot algorithm with a coefficient @xmath70 of 1.0 yielded the same result as the scoring algorithm with a threshold @xmath64 of 4.0 ; that is , the 346th and 327th observations were detected .
lastly , the function can be used to give the estimated 10th , 25th , 50th , 75th , and 90th quantile coefficients as follows : > coef(fit ) q10 q25 q50 q75 q90 ( intercept ) 1.609 2.549 3.332
4.190 5.037 meta -0.039 -0.064 -0.091 -0.121 -0.138
in this paper , we proposed three algorithms to detect outlying observations on the basis of censored quantile regression .
the outlier detection algorithms were implemented for censored survival data : residual - based , boxplot , and scoring algorithms .
the residual - based algorithm detects outlying observations using constant scale estimates , and therefore , it tends to select relatively many observations to achieve a high level of sensitivity in identifying outliers .
thus , this algorithm is effective when high sensitivity is essential .
the results of our simulation study imply that the boxplot and scoring algorithms with censored quantile regression are more effective than the residual - based algorithm when considering sensitivity and specificity together .
the residual - based and boxplot algorithms require a pre - specified cut - off to determine whether observations are outliers .
thus , these two algorithms are useful if a cut - off can be provided in advance .
moreover , the boxplot algorithm can be applicable when a single covariate exists .
the scoring algorithm is more practical in that it provides the outlying magnitude or deviation of each point from the distribution of observations and enables the determination of a threshold by visualizing the scores ; thus , this scoring algorithm is assigned as the default in our package .
all the algorithms were implemented into our developed package outlierdc , which is freely available via comprehensive r archive network ( cran ) .
the function yields the result of outlier detection by the residual - based , boxplot , or scoring algorithm .
the resulting object can be used for generic functions such as , , , and .
the help page for the function contains several examples for use in algorithms .
these can be easily accessed by the command . in our package , there are several options that users need to choose . for convenience
, the most effective and practical choice is assigned as the default for each option .
thus , first - time users can run our package easily by following the illustration without a deep understanding of the presented algorithms .
this research was supported by the basic science research program through the national research foundation of korea ( nrf ) funded by the ministry of education , science and technology ( 2010 - 0007936 ) .
bguin c , hulliger b ( 2004 ) multivariate outlier detection in incomplete survey data : the epidemic algorithm and transformed rank correlations .
journal of the royal statistical society : series a ( statistics in society ) 167(2):275294
|
outlying observations , which significantly deviate from other measurements , may distort the conclusions of data analysis .
therefore , identifying outliers is one of the important problems that should be solved to obtain reliable results . while there are many statistical outlier detection algorithms and software programs for uncensored data , few are available for censored data . in this article
, we propose three outlier detection algorithms based on censored quantile regression , two of which are modified versions of existing algorithms for uncensored or censored data , while the third is a newly developed algorithm to overcome the demerits of previous approaches .
the performance of the three algorithms was investigated in simulation studies .
in addition , real data from seer database , which contains a variety of data sets related to various cancers , is illustrated to show the usefulness of our methodology .
the algorithms are implemented into an package outlierdc which can be conveniently employed in the environment and freely obtained from cran .
keywords : outlier detection , quantile regression , censored data , survival data
|
basic ideas on instabilities and dissipative pattern formation in open spatially extended nonlinear systems @xcite have gained substantial interest in many fields of science in the last two decades .
( see , e.g. , refs .
the investigations become rather intensive also in the field of solid - state physics and electronics @xcite where a large variety of different nonlinear nonequilibrium phenomena occurs , which can cause spontaneously arising spatio - temporal patterns in materials and devices .
together with highly developed device technology , these circumstances suggest promising opportunities for future research and developements .
recently , attention has been called to thyristor - like semiconductor structures with large active areas @xcite , as these nonlinear systems with bistable properties show several spatial and spatio - temporal current - density patterns .
such semiconductor structures could potentially be used as multi - stable elements for integrated circuits , self - organizing devices for image recognition and image processing , etc .
one of the most striking results of former considerations based on phenomenological models @xcite is that the instability mechanisms posess some features very similar to those studied in biology or chemical media , which may show an extremly manifold selforganizing behavior .
however , at present the understanding of instability mechanisms and pattern formation in multilayer structures is far from being complete .
apparently , there is a lack of systematic investigations concerning self - organization aspects for this family of devices . in a very early theoretical paper
@xcite the process of single current filament formation has only been analyzed for the simplest case of @xmath5-@xmath2-@xmath5-@xmath2 structures , that according to ridley s terminology @xcite belong to the class of extended electrical systems with current controlled negative differential resistance .
recently , gate - driven @xmath5-@xmath2-@xmath5-@xmath2 structures have been used to study current filamentation and the dynamics of switching fronts under the constraints of two global couplings @xcite . in specially designed multilayer structures different types of filament oscillation have been observed @xcite .
very few theoretical papers have been devoted to instabilities in standard thyristors ; they all aimed merely on a safe operation in power applications @xcite .
furthermore , it is not an easy task to adapt the advanced experience of power device technology directly to the specifics of devices showing phenomena of self - organization .
one of the main reasons for this is that the operation of semiconductor devices and especially thyristors depends very sensitively on many physical and design parameters . on the one hand
, this gives numerous opportunities for a proper choice of those parameters responsible for the intended function of these devices .
but on the other hand the high - dimensional parameter space makes a systematic analysis rather difficult , because the use of numerical methods and the performance of experiments are usually time - consuming and/or expensive .
therefore detailed studies of the underlying physical processes in conjunction with the material and the design parameters are necessary and should lead to the deepest possible analytical description of the device physics . following this motivation and aiming at further development of concepts introduced earlier in refs .
@xcite we suggest in this paper an analytical approach for instability mechanisms in a family of devices based on thyristor - like @xmath0-@xmath1-@xmath2-@xmath0-@xmath3 structures , in which several modes of pattern formation have already been found experimentally @xcite .
the outline of the paper is as follows : in section 2 the investigated semiconductor structure and its basic physics with emphasis on the nonlinear properties are described .
a non - stationary 1@xmath42-dimensional model rendering possible an analytical investigation of two - dimensional current - density and potential distributions transversal to the main current flow direction is derived in section 3 for two cases characterized by different nonlinear electrical mechanisms . in section 4 ,
stationary homogeneous states are determined from the model equations and used to derive the current - voltage characteristics for systems with uniform current - density distributions .
finally , the stability of stationary homogeneous states is studied in section 5 , including a discussion of important parameter dependencies .
the system under consideration consists of a semiconductor structure with a vertical width @xmath6 of several hundreds of @xmath7 m and with the two other ( transversal ) widths @xmath8 . along the vertical direction
the structure is characterized by a @xmath0-@xmath1-@xmath2-@xmath0-@xmath3 doping profile as schematically shown in fig .
the outer @xmath0 and @xmath3 layer of the structure are provided with metal contacts .
the internal design of the structure possesses the following peculiarities : \(i ) along the ( vertical ) @xmath9-direction the @xmath0-@xmath1-@xmath2-@xmath0-@xmath3 structure can be divided into an active ( triggering ) subsystem consisting of a four - layer thyristor - like @xmath1-@xmath2-@xmath0-@xmath3 structure and a subsystem of a distributed resistive layer consisting essentially of the remainder @xmath1 layer .
note that the thyristor - like regeneration mechanism in the active subsystem causes an autocatalytic increase of the current density in a certain current interval which is counteracted by the resistive layer .
the width of the @xmath2-@xmath0-@xmath3 part is only a few tens of @xmath7 m which is more than an order less than the whole width @xmath6 of the device .
the lifetime @xmath10 of the excess carriers in the bulk material is supposed to be so small that conductivity modulation of the bulk due to the injection of electrons and holes from the outer @xmath3 and @xmath0 layers is prevented .
this can be achieved by reducing the lifetime in the main part of the bulk in a certain distance to the anode contact leaving a relatively thin part of the @xmath1 layer with large @xmath10 nearby the inner @xmath2 layer .
\(ii ) the doping concentrations @xmath11 and @xmath12 in the layers forming the cathode emitter @xmath0-@xmath3 junction [ fig .
[ fig1](b ) ] are similar to those used in high frequency bipolar transistors .
the electron injection efficiency of this junction increases from nearly zero at very low current densities to a saturation value of approximately 0.8 - 0.9 at rather small current densities of the order of 1 ma/@xmath13 .
\(iii ) the doping concentration @xmath14 of the @xmath1 layer is much smaller than the doping concentration @xmath15 of the @xmath2 base .
hence , for low currents , current transport through this junction is dominated by a leakage current of electrons from the @xmath2 base into the @xmath1 emitter and , consequently , the injection efficiency of the @xmath1-@xmath2 junction is very small .
\(iv ) the widths of the @xmath0 and @xmath2 base , @xmath16 and @xmath17 , in the thyristor - like subsystem are much smaller than the diffusion lengths @xmath18 and @xmath19 of the minority carriers in these layers : @xmath20 , @xmath21 .
consequently , both base transport factors are about one .
when a dc voltage is applied to the device such that the outer @xmath3 layer is biased negatively with respect to the outer @xmath0 layer , the anode and cathode emitter junction are forward biased , while the collector junction is reverse - biased .
thus , double injection occurs inside the thyristor - like subsystem accompanied by a characteristic nonequilibrium plasma - field stratification along the vertical direction @xmath9 as shown in fig .
[ fig2](a ) . due to the chosen doping proportion at the @xmath1-@xmath2 junction , which is just inverse to that in standard thyristors
, there is a significant electron leakage current from the @xmath2 base into the @xmath1 layer which obviously exceeds the electron injection current .
therefore , a rather large excess charge accumulates in a thin plasma layer ( @xmath22 layer ) . if the carrier lifetime in this plasma layer is sufficiently large , the local dynamics of this charge at non - stationary conditions may be considered to be the most inertial process , thus determining the temporal evolution of the whole device .
the derivation of the model equations in section 3 is based essentially on these physical properties .
the nonlinear feedback mechanism between the electron and hole injection currents in the @xmath1-@xmath2-@xmath0-@xmath3 subsystem is similar to that in standard thyristors .
however , there are remarkable peculiarities and it is useful to distinguish between two cases , corresponding to certain current intervals , which are characterized by a particular nonlinear mechanism , respectively , causing the appearance of negative differential resistance in the current - voltage characteristic of the thyristor - like subsystem . in analogy to the two - transistor approach for thyristors
we consider the thyristor - like subsystem to be composed of an @xmath2-@xmath0-@xmath3 and a @xmath1-@xmath2-@xmath0 transistor to specify the two different regenerative mechanisms .
current regeneration @xcite in thyristors is possible for current intervals , in which the condition @xmath23 holds , where @xmath24 and @xmath25 denote the thyristor voltage and current , and @xmath26 is the differential resistance .
@xmath27 is implicitly given by the steady state condition @xmath28 where @xmath29 and @xmath30 denote the current gains of the @xmath2-@xmath0-@xmath3 and the @xmath1-@xmath2-@xmath0 transistor , respectively [ fig .
[ fig2](a ) ] .
charge carrier recombination in the space charge region of the emitter junction of a transistor influences the current gain significantly @xcite and typically leads to a monotonic increase of the current gain with increasing current for not too large currents .
consider the device parameters to be of such kind that the current gain @xmath29 approaches a value close to one at relatively low currents ( @xmath31a/@xmath13 ) . then eq .
( [ eq : alpha_sum ] ) can be fulfilled at a very low injection efficiency of the @xmath1-@xmath2 emitter junction .
thus , this case subsequently called _ low current range _ is roughly characterized by @xmath32 and @xmath33 . if the doping profile and other parameters are chosen such that @xmath29 saturates in the low current range before eq .
( [ eq : alpha_sum ] ) is true , regeneration does not occur at low currents .
nevertheless , at larger currents regeneration might become possible again when @xmath30 increases monotonously with the current .
this can occur at sufficiently large current , when a plasma layer @xmath22 evolves in the @xmath1 layer near the @xmath1-@xmath2 junction and the injection of excess holes from this plasma layer into the @xmath2 base increases superlinearly as proposed by fletcher @xcite .
consequently , the @xmath22 layer plays the role of a plasma emitter , the efficiency of which rises strongly with the current so that eq .
( [ eq : alpha_sum ] ) can be fulfilled .
this case is called _ moderate current range _ subsequently and is roughly characterized by @xmath34 and @xmath35 .
before we start with a detailed analysis of these two cases let us turn to a short discussion on important transversal processes which have to be taken into account for an analytic description , too . to this end , consider a so - called differential element of the whole @xmath0-@xmath1-@xmath2-@xmath0-@xmath3 structure , which is small in @xmath36 and @xmath37 direction . fig . [
fig2](b ) shows schematically contours of possible current fluctuations in such an element which may lead to a destabilization of the homogeneous current flow . obviously , there are three channels along which transversal currents effect a transversal spatial coupling in the thyristor - like subsystem : the @xmath0 base , the @xmath2 base , and the possibly conductivity modulated @xmath22 layer in the @xmath1 bulk , in which the corresponding transversal fluctuations @xmath38 , @xmath39 , and @xmath40 may arise . for the resistor subsystem transversal current fluctuations
@xmath41 are taken into account ; they are distributed over the whole bulk due to three - dimensional deformations of the potential distribution @xmath42 ; these deformations are caused by potential fluctuations at the interface between the two subsystems , which in turn result from local current fluctuations in the thyristor - like subsystem .
the metallic layers forming the anode and cathode contact on the top and bottom of the device represent further channels denoted by @xmath43 and @xmath44 in fig .
[ fig2](b ) and render possible the formation of closed loops of current fluctuations inside the system without any interaction with elements of the external circuit .
thus , all differential elements of the system are connected to each other by both the vertical feedback mechanism and transversal couplings which have to be considered in a self - consistent description as outlined in detail in the following sections .
as mentioned above the excess plasma of the @xmath45 layer is considered to be the most inertial part of the system .
consequently , all other nonstationary transport effects are neglected .
we start with a derivation of the basic equations for the low current range . based on these results
the peculiarities concerning the case of moderate currents are discussed in a second subsection . in the low current case
the dynamics of the plasma emitter can be described in terms of the excess charge density @xmath46 per unit square in the plasma layer @xmath22 : @xmath47 where @xmath48 is the sheet electron current - density in the plasma layer , @xmath49 the two - dimensional @xmath50-operator , @xmath51 the diffusion coefficient of electrons , @xmath10 the lifetime of excess charge carriers at low injection , @xmath52 the @xmath9 component of the electron current - density at the boundary between the plasma layer @xmath22 and the @xmath2 base , @xmath53 the @xmath9 component of the electron current - density leaving the thyristor - like subsystem and entering the resistor subsystem .
refer to the @xmath9 component of the corresponding current - density vector ; for a better readability the supplement _
@xmath9 component _ is omitted in the following explanation .
accordingly , quantities denoted by @xmath54 refer to pure sheet current densities per unit length in a certain layer of the structure .
physical constants as well as material and design parameters are listed separately in tables [ table1 ] and [ table2 ] . ]
@xmath24 denotes the voltage drop across the thyristor - like subsystem , @xmath14 the acceptor concentration of the @xmath1 layer , @xmath55 the conductivity of the plasma layer @xmath56 , which is equal to @xmath57 in the low current range , @xmath58 the elementary charge , @xmath59 , @xmath60 and @xmath61 the electron and hole mobility , and @xmath62 the average electron density per unit square in the plasma layer with width @xmath63 . following the charge - control model of transistors and thyristors ( see , e.g. , ref .
@xcite ) the variable @xmath53 can be expressed by @xmath64 thus , the charge equation can be re - written as @xmath65 with @xmath66 .
the variable @xmath67 can be further approximated in the following way by means of the maximum value of the excess charge carrier concentration @xmath68 in the plasma layer : @xmath69 relations connecting the concentrations of the minority charge carriers at both sides of the @xmath1-@xmath2 junction , @xmath68 and @xmath70 , and the hole current - density @xmath71 at the interface between the plasma layer and the @xmath2 base are provided by the classical theory of @xmath5-@xmath2 junctions ( see , e.g. , ref .
@xcite ) : @xmath72 where @xmath15 is the donor concentration in the @xmath2 base , @xmath73 the effective width of the @xmath2 base , which depends on the voltage drop @xmath74 across the @xmath2-@xmath0 collector junction , and @xmath75 the diffusion coefficient of holes .
note that due to the assumption @xmath76 [ sect .
2 , assumption ( iv ) ] the hole current - density @xmath77 at the collector is equal to @xmath71 .
the dependence of the base width @xmath73 on the collector voltage drop @xmath78 can be approximated by ( see ref .
@xcite ) : @xmath79,\\ { \mathrm with } \hspace{2em } v_{pth}&=&\frac{qn_d w_{20}^2}{2\epsilon \epsilon_0 } , \nonumber\end{aligned}\ ] ] where @xmath17 denotes the whole width of the @xmath2 base , @xmath80 the punch - through voltage of the @xmath2-@xmath0 junction , @xmath81 and @xmath82 the vacuum permittivity and the dielectric constant of silicon .
equations ( [ eq : approx ] ) - ( [ eq : rel1 ] ) yield an expression linking the hole current - density at the collector junction , @xmath83 , the charge @xmath67 and the collector voltage @xmath84 : @xmath85 in the @xmath2-@xmath0-@xmath3 transistor the local current balance for a differential element of the @xmath0 base is given by : @xmath86 where @xmath87 and @xmath88 denote the total current - densities at the @xmath2-@xmath0 collector and @xmath0-@xmath3 emitter junction , @xmath89 the conductivity of the @xmath0 base , and @xmath90 the voltage drop across the emitter junction . because of @xmath20
, the electron current - density at the collector @xmath91 is practically equal to the electron current - density @xmath92 injected by the @xmath3 layer .
nevertheless , since the hole leakage from the @xmath0 base into the @xmath3 emitter is strongly nonlinear we have to take into account a @xmath25-dependence ( or bias - dependence ) of the parameter @xmath29 .
the parametric dependence between the total emitter current - density @xmath88 and its electron injection component @xmath92 is given as an implicit function coupling the emitter voltage @xmath90 with the emitter current - density @xmath88 : @xmath93 here @xmath94 , @xmath95 and @xmath96 are saturation values of the electron injection current - density , and the linear and nonlinear hole leakage current - densities , respectively .
the parameter @xmath97 depends on the concrete mechanism of the hole leakage in the @xmath0-@xmath3 junction @xcite . for the case under study
we assume @xmath98 , i.e. sah - noyce - shockley recombination in the space charge region . in the @xmath22-@xmath2-@xmath0 transistor the local current balance for a differential element of the @xmath2 base and of the plasma layer @xmath22
is given by : @xmath99 , \\
j_t - j_2&=&\mbox{\boldmath $ \nabla_{\perp } \cdot i_{\cal p}$ } \nonumber \\ \label{eq : div2'a } & = & -\sigma_{\cal p } w_{\cal p}\delta_{\perp } v_t + d_n \delta_{\perp } q , \\\end{aligned}\ ] ] where @xmath100 is the conductivity of the @xmath2 base .
note , that low injection conditions are assumed for the plasma emitter . combining eqs .
( [ eq : div2 ] ) and ( [ eq : div2a ] ) yields : @xmath101 + \sigma_{\cal p } w_{\cal p } \delta_{\perp } v_t - d_n\delta_{\perp}q,\\ j_{cp}&= & j_c - j_{cn}. \ ] ] the voltage drop @xmath24 across the thyristor - like subsystem is given by @xmath102 , where @xmath103 is the voltage drop across the @xmath1-@xmath2 junction .
using @xmath104 and eqs .
( [ eq : approx ] ) and ( [ eq : rel ] ) the following relation between @xmath67 and @xmath103 can be derived : @xmath105 where @xmath106 and @xmath107 denote the equilibrium minority carrier concentrations of the @xmath2 base and @xmath1 bulk , respectively , and @xmath108 is the intrinsic carrier concentration .
now we indicate the peculiarities occuring at moderate currents , which case is characterized by a high injection level in the @xmath1 bulk . in this case
ambipolar effects in the plasma layer should be included into consideration . in the frame of the charge - control model
the following relations between the current densities @xmath53 and @xmath109 , the sheet current densities @xmath110 and @xmath111 and the charge @xmath67 in the plasma layer @xmath22 are valid : @xmath112 here @xmath113 is the transversal conductance of the plasma layer , @xmath114 the transit time through the plasma layer , and @xmath115 the ambipolar diffusion coefficient . note that eqs .
( [ eq : tet1 ] ) and ( [ eq : tet3 ] ) substitute eqs .
( [ eq : tet ] ) and ( [ eq : i_pl_low ] ) , respectively .
the charge balance equation for the plasma layer should be re - written as @xmath116 where @xmath117 is the high - injection lifetime . by inserting eqs . ( [ eq : tet1 ] ) and ( [ eq : tet3 ] ) we obtain @xmath118\end{aligned}\ ] ] [ compare with eq .
( [ eq : char-dgl ] ) ] .
the relation between the excess carrier concentrations at the borders of the space charge region of the @xmath1-@xmath2 junction changes from shockley s to fletcher s @xcite form : @xmath119 [ compare with eq .
( [ eq : rel ] ) ] .
taking into account eqs .
( [ eq : approx ] ) , ( [ eq : rel1 ] ) , and ( [ eq : rel ] ) we obtain in analogy to eq .
( [ eq : qjc ] ) an expression linking the charge @xmath67 in the plasma layer , the voltage drop @xmath120 across the @xmath22-@xmath2-@xmath0-@xmath3 subsystem and the hole current - density @xmath77 of the collector : @xmath121^{1/2 } \left[\frac{j_{cp}}{(1-\alpha_1)j_m}\right]^{1/2 } , \end{aligned}\ ] ] where the normalizing quantities @xmath122 and @xmath123 are connected by @xmath124^{1/2}. \end{aligned}\ ] ] as already mentioned above , the current gain @xmath29 of the @xmath2-@xmath0-@xmath3 transistor is approximately current - independent in the moderate current range .
then , the electron current - density @xmath125 of the collector is given by @xmath126 in the @xmath22-@xmath2-@xmath0 transistor conductivity modulation in the plasma layer @xmath22 due to excess carrier concentrations should be taken into account at high injection level ; consequently eq .
( [ eq : div2a ] ) is substituted by : @xmath127.\end{aligned}\ ] ] on account of vanishing field and diffusion currents at the transversal borders of the sample ( @xmath128 ) the following boundary conditions for the @xmath22-@xmath2-@xmath0-@xmath3 subsystem have been chosen for both the low - injection and the high - injection case : @xmath129 here @xmath130 is a unit vector normal to the transversal surfaces of the sample . in the resistor - like part of the @xmath1 bulk displacement currents are neglected .
the three - dimensional potential distribution in this region is described by laplace s equation : @xmath131 with the boundary conditions @xmath132 where @xmath133 denotes the voltage drop across the whole sample , @xmath134 the width of the resistive layer , and @xmath135 the three - dimensional laplace - operator .
the current - density vector @xmath136 inside the @xmath1 bulk is connected with the local potential @xmath42 according to @xmath137 and its @xmath9 component @xmath138 at the boundary @xmath139 `` feeds '' the thyristor - like subsystem locally so that @xmath140 the @xmath0-@xmath1-@xmath2-@xmath0-@xmath3 sample is driven by an ideal voltage source @xmath141 via an external load resistor @xmath142 .
this yields the following load line equation : @xmath143 the sets of equations ( [ eq : q-dgl ] ) - ( [ eq : phi_2 ] ) , ( [ eq : t_conds ] ) - ( [ eq : load ] ) and ( [ eq : u ] ) , ( [ eq : div1 ] ) , ( [ eq : div2 ] ) , ( [ eq : phi_2 ] ) - ( [ eq : load ] ) present a self - consistent non - stationary @xmath144-dimensional description of the system under consideration for the low and moderate current range , respectively .
to describe the homogeneous stationary states eqs .
( [ eq : char-dgl ] ) , ( [ eq : div1 ] ) , ( [ eq : div2 ] ) , ( [ eq : p_em ] ) , and ( [ eq : lapl - dgl ] ) for the low current range and eqs .
( [ eq : div1 ] ) , ( [ eq : div2 ] ) , ( [ eq : char - high ] ) , ( [ eq : div2 ] ) , and ( [ eq : lapl - dgl ] ) for the moderate current range with all time and space derivatives set to zero have to be solved : @xmath145 .
this leads to @xmath146 .
the current gain @xmath29 of the @xmath2-@xmath0-@xmath3 transistor can be written as @xmath147 combining the charge balance equation ( [ eq : char-dgl ] ) for the stationary homogeneous state with @xmath148 and eq .
( [ eq : al ] ) yields @xmath149 according to eq .
( [ eq : qjc ] ) the variable @xmath67 is proportional to @xmath77 , which is equal to @xmath150 ; together with eq .
( [ eq : u ] ) and @xmath120 one obtains : @xmath151 \left(1-\alpha_1 \right ) j.\end{aligned}\ ] ] combining eqs .
( [ eq : stat ] ) and ( [ eq : statq ] ) leads to the following parametrically determined @xmath152-characteristic of the thyristor - like @xmath22-@xmath2-@xmath0-@xmath3 subsystem for the low current range : @xmath153}\right]^2 , \end{aligned}\ ] ] where the voltage drop @xmath90 across the @xmath0-@xmath3 emitter and the current density @xmath25 are coupled by @xmath154 the differential resistance of the thyristor - like subsystem for the homogeneous state under low current conditions is then given by : @xmath155^{-1 } .
\nonumber\end{aligned}\ ] ] using the charge balance equation ( [ eq : char - high ] ) and @xmath156 we obtain the following relation between @xmath67 and @xmath25 : @xmath157 on the other hand , for the homogeneous case the variable @xmath67 depends on @xmath25 according to the relation @xmath158^{1/2 } \left(\frac{j}{j_m}\right)^{1/2},\end{aligned}\ ] ] which follows from eqs .
( [ eq : u ] ) , ( [ eq : qjc ] ) together with @xmath159 and @xmath120 . choosing @xmath160 and equalizing eqs .
( [ eq : stat ] ) and ( [ eq : statq ] ) yields together with eq .
( [ eq : q_m ] ) the following expression for the @xmath152-characteristic of the thyristor - like subsystem for the case of moderate currents : @xmath161,\\\end{aligned}\ ] ] where @xmath123 and @xmath80 are given by eqs .
( [ eq : jm ] ) and ( [ eq : u ] ) .
the specific differential resistance of the thyristor - like subsystem in the homogeneous state for this case is : @xmath162 the global @xmath163-characteristic of the complete @xmath0-@xmath1-@xmath2-@xmath0-@xmath3 structure ensues from @xmath164 where @xmath165 denotes the voltage drop across the resistive layer , the resistance of which is determined by @xmath166 . by properly adjusting the design parameters of the @xmath3 layer it is possible to realize the appearance of a negative differential resistance for both the low and moderate current case . in the calculations presented below the saturation current - densities @xmath94 and @xmath95 have been used to adjust the two cases .
figure [ fig3](a ) shows stationary current - density vs. voltage characteristics for the low current case . besides the characteristic @xmath152 of the thyristor - like subsystem and the characteristic of the total structure @xmath163 , which are both calculated for the set of standard parameters ,
two further @xmath163-characteristics for reduced values of the conductivity of the @xmath1 bulk are plotted . for standard
parameter values the bulk resistance of the @xmath1 bulk is not sufficiently large to compensate the negative differential resistance of the thyristor - like subsystem at low currents .
for this reason the two curves @xmath152 and @xmath163 are nearly identical and can not be distinguished in the diagram .
in fact , the current - density @xmath167 , at which the negative differential resistance transforms to a positive one , are nearly equal and approximately 1 ma/@xmath13 for the two cases .
@xmath167 can be reduced by a considerable amount only , if the resistance of the @xmath1 bulk is enlarged at least by a factor of ten , as illustrated by the two other curves in fig . [ fig3](a ) .
figure [ fig3](b ) shows current - density vs. voltage characteristics for the case of moderate currents and the standard value of the conductivity of the @xmath1 bulk .
the current - density @xmath167 , at which the differential resistance changes from negative to positive values , is 0.45 a/@xmath13 for the total structure , i.e. considerably lower than the corresponding value ( @xmath168 1 a/@xmath13 ) for the thyristor - like subsystem .
the blocking voltage is approximately equal to the punch - through voltage @xmath80 for both the low and moderate current case .
this is because the emitter efficiencies are very small at low currents , so that a hole injection current from the anode emitter , that is sufficiently large to initiate the regenerative process , becomes possible only , when the effective width @xmath169 of the @xmath2 base is approximately zero , i.e. under punch - through condition .
in this section we analyze the stability of stationary homogeneous states along the transversal dimensions with respect to small 1@xmath42-dimensional fluctuations of the space - dependent variables @xmath170 . the standard approach for the linearization procedure implies to make the following ansatz for all variables in the vicinity of a stationary state @xmath171 @xmath172 the small variations @xmath173 introduced here have to satisfy the same boundary conditions as the variables @xmath174 [ see eqs .
( [ eq : t_conds])-([eq : t_conds4 ] ) and ( [ eq : condphi ] ) , ( [ eq : condphi_2 ] ) ] .
the distribution of the potential in the resistive layer , which is essentially three - dimensional , can be described as solution of the laplace equation ( [ eq : lapl - dgl ] ) with the boundary conditions ( [ eq : condphi ] ) and ( [ eq : condphi_2 ] ) . a general solution of this equation is a superposition of a uniform excitation mode and various @xmath175-modes describing spatially periodic potential perturbations : @xmath176 with @xmath177 , @xmath178 , and @xmath179 denoting integers .
figure [ fig4 ] illustrates the potential fluctuations @xmath180 as well as the potential @xmath181 in a reduced two - dimensional space . for a given potential fluctuation @xmath182 at the boundary @xmath134 between the resistive layer and the thyristor - like subsystem the @xmath175-dependence of the potential variations @xmath183 in the resistive layer
is given by : @xmath184 the @xmath9 component @xmath138 of the current density @xmath136 at the border between the thyristor - like subsystem and the resistive layer is equal to the local anode current - density @xmath109 of the thyristor - like subsystem [ see eqs .
( [ eq : ohm ] ) , ( [ eq : j_t ] ) ] , so that the following relation is valid : @xmath185 from eqs .
( [ eq : varphi ] ) and ( [ eq : jt ] ) one obtains the following variational derivative by which fluctuctions of @xmath186 and @xmath187 are coupled for each wave - number @xmath175 : @xmath188 the quantity @xmath189 can be considered as an effective specific resistance of the bulk material with respect to a potential fluctuation with wave - number @xmath175 .
it differs from the specific resistance @xmath190 of the @xmath1 bulk because of three - dimensional deformations of the potential relief which arise due to a non - uniform current flow in three dimensions .
now we apply the linear stability analysis to the charge dynamic equation making a perturbation ansatz for all the variables @xmath191 subjected to two - dimensional transversal fluctuations , @xmath192 and then study the stability for the both cases mentioned above . for the low current range the following set of variational relations can be derived from eqs .
( [ eq : q-dgl])-([eq : phi_2 ] ) and ( [ eq : deriv_r ] ) : @xmath193\delta j_t + \left[a_c^k k^2 d_n - ( 1- a_c^k)\frac{kt}{q } \frac{1}{r_2^k}\frac{1}{q_{stat}}\right ] \delta q , \\
\delta q & = & -\frac{\partial q}{\partial v_t } \bigg \vert_{stat } r_r^k\delta j_t + \frac{\partial q}{\partial j_{cp } } \bigg \vert_{stat } \delta j_{cp } , \\ \label{eq : derivs'_4 } \delta j_{cp } & = & ( 1-a_c^k)\left[1 + \left(\frac{1}{r_{\cal p}^k } + \frac{1}{r_2^k } \right)r_r^k\right]\delta j_t + ( 1-a_c^k ) \left[k^2 d_n + \frac{kt}{q}\frac{1}{r_2^k } \frac{1}{q_{stat}}\right ] \delta q , \ ] ] where @xmath194 and the differential resistance of the thyristor - like subsystem for the homogeneous state is given by : @xmath195 the coefficient @xmath196 is defined as ratio of the fluctuations of the electron current - density @xmath197 and the total current density @xmath198 at the collector , which in turn can be written as : @xmath199 the differential resistances @xmath200 and @xmath201 of the @xmath0-@xmath3 emitter junction are defined as @xmath202 the @xmath175-dependent quantities @xmath203 , @xmath204 , and @xmath205 can be considered as effective resistances of the @xmath0 base , the @xmath2 base , and the @xmath206 layer , respectively , for a current - density fluctuation with wave - number @xmath175 . @xmath207 using these dependencies we obtain from eqs .
( [ eq : derivs])-([eq : derivs_4 ] ) the final dispersion relation for the increment @xmath208 as a sum of four components : @xmath209 where the partial components are : @xmath210 \times \frac{kt}{q j_t } \times \frac{1}{r_2^k } , \nonumber \\
\zeta_{\cal p } & = & \zeta_{\cal p } ' + \zeta_{\cal p } '' , \\ \zeta_{\cal p } ' & = & - k^2 d_n \frac{(1-a_c^k)}{{\cal k } ( k ) } \left [ \alpha_1 \left(1 + \frac { r_r^k}{r_{\cal p}^k}\right ) - ( a_c^0-\alpha_1)\frac{r_r^k}{r_t^0 } \right ] , \nonumber \\
\zeta_{\cal p}''&= & \frac{(a_c^k - \alpha_1 ) r_r^k } { \tau _ * { \cal k}(k ) } \times \frac{1}{r_{\cal p}^k}. \nonumber\ ] ] the function @xmath211 is given by the expression : @xmath212 - ( a_c^0-\alpha_1)\frac{r_r^k}{r_t^0}. \ ] ] note that all partial components depend on the current - density dependent coeffficients @xmath213 , @xmath214 and @xmath29 .
besides , the contribution of the terms @xmath215 , @xmath216 , and @xmath217 is controlled by the effective resistances of the @xmath0 base , the @xmath2 base , and the @xmath206 layer , respectively . for @xmath217 also
the diffusion properties of the plasma charge play an important role . for the moderate current range a corresponding set of variational relations
can be obtained by applying the linearization procedure to eqs .
( [ eq : tet1])-([eq : div2 ] ) with taking into account the relations eqs .
( [ eq : phi_2 ] ) and ( [ eq : deriv_r ] ) : @xmath218\delta j_t + \left [ k^2 a_c^k(b-1)d_p - ( 1-a_c^k)\frac{kt}{qq_{stat}}\frac{1}{r_2^k } \right]\delta q , \\
\delta q & = & - \frac{\partial q}{\partial v_t } \bigg\vert_{stat } r_r^k \delta j_t + \frac{\partial q}{\partial j_{cp}}\bigg\vert_{stat } \delta j_{cp } , \\
\label{eq : derivs_4 } \delta j_{cp } & = & ( 1-a_c^k)\left[1 + \left(\frac{1}{r_2^k}+\frac{1}{r_{\cal p}^k } \right)r_r^k\right]\delta j_t + \left [ ( 1-a_c^k)(b-1)k^2d_p + ( 1-a_c^k)\frac{kt}{qq_{stat}}\frac{1}{r_2^k } \right]\delta q,\ ] ] where @xmath219 the parameter @xmath205 can again be considered as an effective resistance of the plasma layer with respect to current - density and potential fluctuations with a wave - number @xmath175 . however , in the case of moderate currents @xmath205 is conductivity modulated : @xmath220 the final dispersion relation for the increment @xmath208 results from eqs .
( [ eq : derivs])-([eq : derivs_4 ] ) : @xmath221 with @xmath222 \frac{1-a_c^k}{\tau _ * g_0 { \cal k}(k ) } \frac{kt}{qj_t } \times \frac{1}{r_2^k } , \nonumber \\
\zeta_{\cal p } & = & \zeta_{\cal p } ' + \zeta_{\cal p } '' , \\
\zeta_{\cal p}'&= & -k^2 d_p \left\{b(1-a_c^k ) + a_c^k + \frac{(b-1)(1-a_c^k)}{{\cal k}(k ) } \left [ g_k\left(1 + \frac{r_r^k}{r_{\cal p}^k } \right)- ( 1 - a_c^k)\frac{r_r^k}{r_2^k } \right ] \right\ } , \nonumber
\\ \zeta_{\cal p}''&= & \left[2 g_k(1-\alpha_1 ) - g_0(1-a_c^k ) \right ] \frac{r_r^k}{\tau _ * g_0 { \cal k}(k ) } \times \frac{1}{r_{\cal p}^k } , \nonumber\end{aligned}\ ] ] and @xmath223 @xmath224 thereby the differential resistance @xmath26 of the thyristor - like subsystem is determined by eq .
( [ eq : r_t ] ) . for both the low and the moderate current case
the increment @xmath225 is proportional to the sum of the differential resistance of the thyristor - like subsystem , @xmath26 , and that of the resistive layer , @xmath226 , in the limit of perturbations with small wave - numbers : @xmath227 this obviously corresponds to a one - dimensional loading of the thyristor - like subsystem by the resistive layer . from eq .
( [ eq : limit ] ) follows the well - known criterium for stability with respect to homogeneous fluctuations : @xmath228 figure [ fig5 ] shows dispersion relations @xmath229 for the low current case .
the calculations have been performed with two reduced values of the @xmath0 base conductivity @xmath230 in order to illustrate two qualitatively different instabilities that can occur in the semiconductor structure . if the @xmath0 base conductivity is sufficiently low , dispersion relations like the curves denoted by ( 2 ) can occur .
they are characterized by a pronounced maximum at a critical wave - number that is definitely larger than zero .
curve ( 2a ) represents the case that the increment @xmath208 of this critical wave - number is zero , while all other increments are still smaller than zero , i.e. , the uniform state becomes destabilized with respect to current - density fluctuations with the critical wave - number . when the sample current is decreased , the curve is shifted up [ curve ( 2b ) ] , and the increments of a certain band of wave - numbers are larger than zero . this behavior can be interpreted as a turing - like instability @xcite .
when the transversal coupling inside the thyristor - like subsystem is strengthened , e.g. by reducing the transversal resistance of the @xmath0 base , the maxima of @xmath229 shift to smaller @xmath175 values leading to dispersion relations like those denoted by ( 1 ) .
they are characterized by a monotonous decrease of @xmath229 with growing @xmath175 , which can be considered as an instability of ridley s type @xcite .
in such a case , the uniform state is typically destabilized via a saddle node bifurcation by fluctuations with the shortest possible wave - number @xmath231 , if uniform fluctuations are suppressed by a sufficiently large load resistor in the external circuit .
similar to the turing - like case , the curve is shifted upwards , when the sample current is decreased .
note that the parameter @xmath25 is the same for the curves ( 1a ) and ( 2b ) and has been adjusted such that the differential resistance of the global current - density vs. voltage characteristic vanishes , i.e. , @xmath232 .
dispersion relations @xmath229 for the moderate current case are shown in fig .
similar to the low current case the homogeneous state is destabilized at a critical current level either by an instability of ridley s ( curves 1 ) or turing s type ( curves 2 ) .
the bifurcation type can be controlled by a proper design .
the most important parameters are the conductivities of the resistive layer , the @xmath0 and the @xmath2 base .
the former essentially governs the transversal spreading of the potential drop across the resistive layer and tends to damp out current - density fluctuations . the @xmath0 and @xmath2 base conductivity control the transversal hole and electron current densities in the two base layers .
as these base currents in turn regulate the injection of electrons and holes from the @xmath233 emitter and the plasma layer @xmath22 , respectively , they play an important role concerning the transversal spreading of a region with enhanced current density . the competition between the damping and the activating process , is the essential physical mechanism allowing - under properly chosen design parameters - the destabilization of a homogeneous current - density distribution by spatially periodic current - density fluctuations strongly reminding of the generic mechanism proposed by turing .
in conclusion we point out that the suggested theory , which is based on a subdivision of the semiconductor structure into an active and a passive subsystem , supplies a self - consistent quantitative description of the nonlinear mechanisms , which control the longitudinal current flow , and the transversal current - density instabilities in multilayer semiconductor systems .
the nonlinear longitudinal electrical properties in the active layer are based on a thyristor - like regenerative mechanism , while the passive subsystem acts as a simple resistive layer .
depending on the system parameters two different types of regenerative mechanisms can be distinguished leading to specific model equations .
a stability analysis of the derived equations reveals that in each case two basic bifurcation types can be expected when the total sample current is varied .
the bifurcations are caused by instability mechanisms that are closely related to those studied by turing @xcite and ridley @xcite .
it turned out that in particular the conductivity ratios of the @xmath0 base , the plasma layer , and the resistive layer are important parameters to control the instability mechanism .
thus , adjusting the respective transversal conductivities , e.g. , by irradiating the sample with protons or ions , using compensated substrate material , or adapting the sample geometry , it is possible to design a multilayer system with the aspired ( in)stability features .
finally , we emphasize that the suggested approach presents a basis for further numerical analysis of pattern formation processes .
the present paper was prepared in the frame of the international project `` dissipative pattern formations in semiconductors and semiconductor devices '' supported by the deutsche forschungsgemeinschaft and the russian academy of science .
99 a. turing , phil .
b * 237 * , 37 ( 1952 ) .
g. nicolis and i. prigogine , _ self - organization in non - equilibrium systems _
( wiley , new - york , 1977 ) .
h. haken , _ advanced synergetics _ , 2nd ed .
( springer , berlin , 1987 ) . b. kerner and v.v .
osipov , _ autosolitons _ ( kluwer , dordrecht , 1994 ) .
m. mikhailov , _ foundations of synergetics _ , 2nd ed .
( springer , heidelberg , 1994 ) . , ed
. by h. engel et al .
( w & t - verlag , berlin , 1996 ) .
d. walgraef , _ spatio - temporal pattern formation _
( springer , new york 1997 ) . , ed . by f.h .
busse and s.c .
mller ( springer , berlin , heidelberg 1998 ) .
e. schll , _ nonequilibrium phase transitions in semiconductors _
( springer , berlin , 1987 ) . , ed . by f .-
niedernostheide ( springer , berlin , 1995 ) .
niedernostheide , m. arps , r. dohmen , h. willebrand and h .- g .
purwins , phys .
solidi b * 172 * , 247 ( 1992 ) .
gorbatyuk , i.a .
linijchuk and a.v .
svirin , sov .
phys . lett . *
15 * , 224 ( 1989 ) [ pisma zh .
. fiz . * 15 * , 42 ( 1989 ) ] .
a.v . gorbatyuk and p.b .
rodin , j. comm .
electron . *
40 * , 49 ( 1995 ) [ radiotekhnika
i lektronika * 40 * , 1876 ( 1994 ) ] .
a.v . gorbatyuk and f .- j .
niedernostheide , physica d * 99 * , 339 ( 1996 ) . m. bode and h .-
purwins , physica d * 86 * , 53 ( 1995 ) .
d. ruwisch , m. bode , h .- j . schulze and f
niedernostheide , in : _ nonlinear physics of complex systems _ , _ lecture notes in physics _ , vol .
476 , ed . by j. parisi , s. mller , and w. zimmermann ( springer , berlin , heidelberg , 1996 ) , pp . 194 - 212 .
varlamov and v.v .
osipov , sov .
phys . semicond .
* 3 * , 803 ( 1969 ) [ fiz .
poluprovodn . * 3 * , 950 ( 1969 ) ] .
ridley , phys .
82 * , 954 ( 1963 ) .
a.v . gorbatyuk and p.b .
rodin , z. phys .
b * 104 * , 45 ( 1997 ) .
m. meixner , p. rodin , and e. schll , phys .
e * 58 * , 2796 ( 1998 ) , and phys .
e * 58 * , 5586 ( 1998 ) .
niedernostheide , b.s . kerner und h .-
purwins , phys .
b * 46 * , 7559 ( 1992 ) .
niedernostheide , m. ardes , m.or-guil und h .-
purwins , phys .
b * 49 * , 7370 ( 1994 ) .
niedernostheide , m. or - guil , m. kleinkes und h .- g .
purwins , phys .
e * 55 * , 4107 ( 1997 ) .
gorbatyuk , _ dynamics and stability of high speed retroaction processes in structures of power thyristors _ , a.f .
ioffe institute preprint 962 , leningrad ( 1985 ) ( in russian ) .
wachutka , ieee trans .
( ed ) * 38 * , 1516 ( 1991 ) .
jaecklin , ieee trans .
( ed ) * 39 * , 1507 ( 1992 ) .
a. wierschem , f .- j .
niedernostheide , a.v .
gorbatyuk and h .-
purwins , scanning * 17 * , 106 ( 1995 ) .
w. gerlach , _ thyristoren _
( springer , heidelberg , 1981 ) ( in german ) .
s. m. sze , _ physics of semiconductor devices _
( wiley , new - york , 1981 ) .
fletcher , proc .
ire * 45 * , 863 ( 1957 ) .
@xmath234 & @xmath235 & 2.8 + @xmath51 & diffusion coefficient of electrons & @xmath236 + @xmath75 & diffusion coefficient of holes & @xmath237 + @xmath175 & boltzmann constant & @xmath238v / k + @xmath108 & intrinsic carrier concentration & @xmath239 @xmath240 + @xmath58 & elementary charge & @xmath241 as + @xmath81 & vaccuum permittivity & @xmath242 f / cm + @xmath82 & dielectric constant of si & 11.7 + @xmath60 & electron mobility & @xmath243 + @xmath61 & hole mobility & 480 @xmath13/vs + @xmath94 & saturation value of the electron injection current - density at the @xmath0-@xmath3 junction & @xmath244 a@xmath245 + @xmath95 & saturation value of the hole injection current - density at the @xmath0-@xmath3 junction & @xmath246 a@xmath245 + @xmath96 & saturation value of the recombination - generation current - density at the @xmath0-@xmath3 junction & @xmath247 a@xmath245 + @xmath248 & acceptor concentration of the @xmath0 emitter & @xmath249 @xmath240 + @xmath14 & acceptor concentration of the @xmath1 bulk & @xmath250 @xmath240 + @xmath11 & acceptor concentration of the @xmath0 base & @xmath251 @xmath240 + @xmath15 & donor concentration of the @xmath2 base & @xmath252 @xmath240 + @xmath253 & donor concentration of the @xmath3 emitter & @xmath254 @xmath240 + @xmath16 & width of the @xmath0 base & @xmath255 cm + @xmath17 & width of the @xmath2 base for @xmath256 & @xmath257 cm + @xmath63 & width of the plasma layer & @xmath258 cm + @xmath134 & width of the resistive layer & 1 cm + @xmath89 & conductivity of the @xmath0 base & @xmath259 + @xmath260 & conductivity of the @xmath2 base & @xmath261 + @xmath262 & conductivity of the @xmath56 layer & + @xmath263 & conductivity of the @xmath1 bulk & @xmath264 + @xmath10 & lifetime in the @xmath56 layer for low currents & @xmath265s + @xmath117 & lifetime in the @xmath56 layer for moderate currents & @xmath266s +
|
semiconductor @xmath0-@xmath1-@xmath2-@xmath0-@xmath3 structures with large junction and contact areas are treated as 1@xmath42-dimensional active media , in which self - organized pattern formation can be expected .
the local bistable behavior of the structures may emanate from two different mechanisms both governed by a nonlinear current feedback - loop between the electrons and holes injected from the outer layers . by considering the device to be composed of an active subsystem with negative differential resistance and a passive resistive layer with positive differential resistance an analytical approach is suggested to understand and describe the corresponding physical mechanisms in a self - consistent way .
analytical solutions of the derived model equations allow a description of homogeneous stationary states and yield explicit expressions of the current - density vs. voltage characteristics of the whole structure and its subsystems .
a stability analysis of the homogeneous states with respect to two - dimensional transversal harmonic fluctuations is performed for the two cases under study .
the resulting dispersion relations allow two different types of instability .
while the first one is of ridley s type which is characteristic for any spatially extended electrical system with negative differential resistance , the second type can be considered as a solid - state analogue of turing s instability known as a generic instability mechanism which may lead , e.g. , to the formation of periodic patterns .
16.5 truecm 0 truecm
|
_ acknowledgements _ gt acknowledges support from fapesp proc .
2014/13120 - 7 and cnpq bolsa de produtividade 301996/2014 - 8 .
lt was supported by the u.s .
department of energy , office of science , office of nuclear physics under award no .
de - sc0004286 and polish national science center grant dec-2012/06/a / st2/00390 .
dm would like to acknowledge cnpq graduate fellowship n. 147435/2014 - 5 parts of this work were done when lt visited campinas on faepex fellowship number 2020/16 , as well as when gt participated in the int workshop `` exploring the qcd phase diagram through energy scans '' we thank faepex and the int organizers for the support provided .
we wish to thank miklos gyulassy for enlightening discussions which posed the conceptual challenges that eventually led to this work , and mike lisa for showing us experimental literature and useful discussions .
m. a. stephanov and y. yin , phys .
lett . * 109 * , 162001 ( 2012 ) doi:10.1103/physrevlett.109.162001 [ arxiv:1207.0747 [ hep - th ] ] . j. y. chen , d. t. son and m. a. stephanov , phys .
* 115 * , no . 2 , 021601 ( 2015 ) doi:10.1103/physrevlett.115.021601 [ arxiv:1502.06966 [ hep - th ] ] .
d. t. son and p. surowka , phys .
* 103 * , 191601 ( 2009 ) doi:10.1103/physrevlett.103.191601 [ arxiv:0906.5044 [ hep - th ] ] .
e. v. gorbar , d. o. rybalka and i. a. shovkovy , arxiv:1702.07791 [ hep - th ] .
b. betz , m. gyulassy and g. torrieri , phys .
c * 76 * , 044901 ( 2007 ) doi:10.1103/physrevc.76.044901 [ arxiv:0708.0035 [ nucl - th ] ] . z. t. liang and x. n. wang , phys .
lett . * 94 * , 102301 ( 2005 ) [ arxiv : nucl - th/0410079 ] .
f. becattini , l. csernai and d. j. wang , phys .
c * 88 * , no . 3 , 034905 ( 2013 ) erratum : [ phys .
c * 93 * , no . 6 , 069901 ( 2016 ) ] doi:10.1103/physrevc.93.069901 , 10.1103/physrevc.88.034905 [ arxiv:1304.4427 [ nucl - th ] ] .
f. becattini , f. piccinini and j. rizzo , phys .
c * 77 * , 024906 ( 2008 ) doi:10.1103/physrevc.77.024906 [ arxiv:0711.1253 [ nucl - th ] ] . y. cheng , l. p. csernai , v. k. magas , b. r. schlei and d. strottman , phys .
c * 81 * , 064910 ( 2010 ) [ arxiv:1006.5820 [ nucl - th ] ] .
lifshitz and landau , fluid dynamics t. burch and g. torrieri , phys .
d * 92 * , no . 1 , 016009 ( 2015 ) [ arxiv:1502.05421 [ hep - lat ] ] .
d. montenegro and g. torrieri , arxiv:1604.05291 [ hep - th ] .
m. crossley , p. glorioso and h. liu , arxiv:1511.03646 [ hep - th ] .
g. s. denicol , h. niemi , e. molnar and d. h. rischke , phys .
d * 85 * , 114047 ( 2012 ) erratum : [ phys .
d * 91 * , no . 3 , 039902 ( 2015 ) ] doi:10.1103/physrevd.85.114047 , 10.1103/physrevd.91.039902 [ arxiv:1202.4551 [ nucl - th ] ] .
b. gripaios and d. sutherland , phys .
* 114 * , no . 7 ,
071601 ( 2015 ) doi:10.1103/physrevlett.114.071601 [ arxiv:1406.4422 [ hep - th ] ] .
|
we extend the effective theory approach to the ideal fluid limit where the polarization of the fluid is non - zero . after describing and motivating the equations of motion @xcite ,
we expand them around the hydrostatic limit , obtaining the sound wave and vortex degrees of freedom .
we discuss how the presence of polarization affects the stability and causality of the ideal fluid limit .
relativistic hydrodynamics is an area which currently enjoys a high level of both theoretical and phenomenological activity @xcite .
heavy ion collision experiments have been shown to be well described by hydrodynamic simulations .
the connection of hydrodynamics to microscopic theory , and the development of hydrodynamics as an effective field theory , are however still not well understood . the direct observation of polarization of @xmath0 particles in heavy ion collisions @xcite and microscopic strongly coupled systems @xcite , as well as phenomena such as chiral transport and anomalies
@xcite has motivated the phenomenological study of vorticiity in relativistic fluids @xcite and its relationship to microscopic spin polarization .
a `` realistic '' effective theory for describing the relationship between vorticity and polarization in the ideal fluid limit is however still in development .
vorticity does not emerge in the transport limit , but rather close to the thermodynamic / hydrodynamic regime .
the former was studied within the usual thermodynamic techniques , updated with the inclusion of angular momentum @xcite , but this is not a realistic setup for a strongly coupled dynamical system , where equilibrium is to a good approximation local rather than global .
the authors in @xcite used a cooper - frye formula extended to spinor observables , assuming polarization is zero before freezeout .
this assumption however violates detailed balance across the freezeout hypersurface @xcite . if hadrons after freeze - out carry both vorticity and spin polarization , then so must the microscopic constituents during the evolution of the fluid @xcite and the backreaction needs to be taken into account .
the problem here is a conceptual one : if one defines the ideal fluid limit in terms of isotropy of the comoving frame @xcite , the transfer of angular momentum to spin is forbidden , or at least dictated by the coarse - graining scale @xcite .
however , the local hydrostatic limit in the presence of angular momentum and spin explicitly necessitates of such an anisotropy @xcite , for the maximum entropy state for a medium where microscopic degres of freedom have spin generally has polarization @xcite . and , of course , vorticity , the conserved current of the diffeomorphism invariance within perfect fluid dynamics @xcite , is be broken @xcite if angular momentum can be transferred to local polarization .
usual insights from transport theory can be misleading since polarization is a breakdown of molecular chaos , since it is a two - particle microscopic phase space correlation @xcite . to overcome this , one needs the effective field theory approach , succesfully used to both define ideal @xcite and dissipative @xcite hydrodynamics : one writes down hydrodynamics in lagrangian form , and treats the microscopic scale as an effective planck constant and develops fluctuation - driven terms within the effective field theory .
in addition , however , polarization requires the inroduction of additional degrees of freedom , which explicitly break isotropy . in @xcite
we constructed such a theory .
we shall briefly introduce it and examine its behavior close to hydrostatic equilibrium . in this
approach perfect fluid without polarization can be described by three fields , representing the comoving ( lagrangian ) coordinate systems @xmath1 .
fluidity can be defined via symmetries , namely the invariance of the lagrangian under a volume - preserving diffeomorphism invariance @xcite . in general @xmath2=1\ ] ] it directly follows @xcite that the lagrangian is of the form , at the lowest order in gradients @xmath3}{\phantom{a},\phantom{a}}b_{ij}=\partial_\mu \phi_i \partial^\mu \phi_j\ ] ] the lagrangian above straight - forwardly yields the general energy momentum tensor whose conservation yields euler s equations @xcite , @xmath4 if one substitutes , for energy density @xmath5 , pressure @xmath6 and chemical potential @xmath7 @xmath8 the latter , defined in terms of an internal @xmath9 conserved current @xmath10 and the chemical shift symmetry @xcite @xmath11 to be @xmath12 the lagrangian in general is related to the free energy by a legendre transformation w.r.t .
the chemical potential .
the diffeomorphism symmetries and the chemical shift allow for a uniquely defined flow velocity @xmath13 with the comoving projector being @xmath14 if the system has intrinsic polarization , the coordinates @xmath15 are not enough because they do not contain information about polarization . moreover , because of the conceptual difficulties described earlier , we need to define precisely what we mean by the ideal hydrodynamic limit , especially because of the polarizations breaking of local isotropy and vorticity conservation .
we take the ideal hydrodynamic limit to mean that dynamics is dependent only on the flow @xmath16 and comoving variables , the comoving variables are defined entirely by local entropy maximization , and that only sound waves and vortices arise as local excitations@xcite . regarding the polarization effective degrees of freedom @xmath17 ,
they have been defined in ref @xcite as the ( infinitesimal ) volume integral of the total angular momentum of a fluid cell in the local comoving frame , normalized using the volum of the cell itself .
this can be interpreted as a way to keep track of the angular momentum structure of the fluid during a coarse graining procedure .
angular momentum conservation has always been considered in relativistic hydrodynamics , therefore no new symmetry has to be added , only the new polarization degrees of freedom : the lack of a spin tensor and the symmetry of the stress - energy tensor imply local angular momentum conservation both for ideal and viscous hydrodynamics .
if the polarization were `` passively transported along the fluid '' , in other worlds , if polarization were conserved , then one wold need an additional symmetry on the effective lagrangian .
this situation is similar to the _ actual _ @xmath9 symmetry considered in ref @xcite , where the phase invariance , represented by an explicit dependence of the lagrangian on the gradient of the local phase , was not enough to recover ideal ( non - polarized ) hydrodynamics , because the electric current is not guaranteed to be proportional to the energy flux . to enforce this requirement
one can introduce the concept of shift symmetry , which can be interpreted as an invariance of physics under local rescaling of the electric charge units .
if one uses the same formalism to describe the flow of the spin current ( @xmath18 in the comoving frame ) , one would have @xmath19 which would represent , not only the invariance of physics under rescaling of the polarization units , but also invariance under an arbitrary addition of an extra polarization which is constant along the flux lines .
following @xcite , the invariant that should enter the effective lagrangian is @xmath20 if we apply the `` revised '' chemical shift ( [ rev_chem_shift ] ) , @xmath21 is invariant since @xmath22 will enter in the derivative and annihillate with @xmath23 is vanishing by construction .
conservation of angular momentum , even in the ideal limit , does not imply that polarization should correspond to an even approximatel conserved current .
we therefore treat @xmath18 as an auxiliary field distinct from @xmath24 , and impose the interaction between orbital and spin angular momentum via the equation of state . in order for local equilibrium to be well - defined , we chose to write directly the polarization degrees as proportional to the local vorticity
, @xmath25 a violation of this condition will inevitably result in goldstone modes and topological constraints which create long - term correlations which make a hydrodynamic limit impossible @xcite .
note that in an external field such as ( but not only ) magnetohydrodynamics @xcite , this symmetry would be broken and the lagrangian would depend directly on polarization .
in general one expects a lagrangian which depends on @xmath21 , @xmath17 and mixed terms all together . by counting gradients and enforcing symmetries ,
the lowest order term which respects the internal diffeomorphism symmetry is @xmath26 .
for example @xmath27 $ ] is a higher term in gradient ( @xmath28 being proportional to the vorticity , which is a gradient ) , @xmath29 would violate parity and @xmath30 be higher order in gradient , again .
parity violating terms would of course be permitted in the context of anomalous hydrodynamics ) . considering that polarization introduces a correlation between microstates ,
the presence of polarization at a given entropy @xmath31 should change the free energy , to leading order in gradient , as @xmath32 where @xmath33 is a dimensionful constant representing polarizeability ( it can be positive for a ferromagnetic material and negative for an antiferromagnetic one ) . for dimensional reasons ,
@xmath34 given this , a physically reasonable way to introduce polarization that we adopted is @xmath35 for instance , the ideal gas without polarization with density @xmath36 of particles of which a fraction @xmath37 are polarized , assuming nearest - neighbor interactions @xmath38 and @xmath39 , making @xmath40 .
we can now make the link between the lagrangian formulation and usual thermodynamics using the methods of @xcite but with the angular momentum in lieu of chemical potential ( note that the collinearity between angular momentum and polarization is what makes this analogy possible ) . while physically , because of lack of isotropy it is non - trivial to relate the usual derivatives of the free energy to what we know as pressure and energy density , the existence of some free energy @xmath41 to be minimized will lead to a constraint of the type @xmath42 which allows us to link the constant of proportionality in eq .
[ prevortspin ] to @xmath41 via legendre transfroms @xmath43
if we had @xmath41 in terms of a partition function we could find @xmath44 given a local temperature and vorticity explicitly , via a derivation along the lines of @xcite .
note that the qcd equation of state dependence on angular momentum , necessary to calculate @xmath44 , is as yet unknown , but it should be strictly related to the equation of state at finite magnetic field by the amount of angular momentum carried by that magnetic field .
the equation of motion , using the standard euler - lagrange equations , is @xmath45 where @xmath46\right\ } -\ ] ] @xmath47 - 2 c \left ( \chi + 2 \ , \omega^2 \ , \partial _
{ \omega^2 } \chi \right)f^{\prime } \times\ ] ] @xmath48 p^{\sigma\mu}_i - \right.\ ] ] @xmath49 with @xmath50 , @xmath51 and @xmath52,{ ... }$ ] corresponding to , respectively , antisymmetrization and symmetrization of the indices , as done in @xcite .
in addition to generally breaking homogeneity and vorticity conservation , unlike non - polarized hydrodynamics the higher gradient in velocity dependence these equations will be higher than 2nd order in gradient . to understand the consequences of this , we linearized the hydrostatic limit with `` temperature '' @xmath53 , @xmath54 not that this is not the actual temperature of the background , it is proportional to that only in the conformal limit . using the notation in @xcite , where @xmath55 is shorthand for @xmath56 and @xmath57 $ ] for its trace and time derivatives are denoted by dots , the non - polarized hydrodynamics gives the usual wave equation for sound waves , the stationary vortex state polarization terms which will increase the gradients at each order by one unit .
the free part of the equation will be @xmath58 -\frac{1}{2}[\partial\pi\cdot\partial\pi ] -\frac{1}{2 } \dot\pi^2 \right\ } + \ ] ] @xmath59 \right\ } + \left(\frac{1}{2}a + c\right)[\partial\pi]^2.\ ] ] where the constants @xmath60 are @xmath61 giving an equation of motion @xmath62 note that this will _ always _ make these excitations susceptible to ostrogradski s instabilities , from lowest order @xcite . to go further
, we decompose the perturbation into an irrotational and rotational parts @xmath63 , which , in the linearized limit , have their own fourier modes @xmath64\ ] ] for the longitudinal ones we have @xmath65 if this is well defined we recognize already that there is a hydrodynamic solution ( that is , the frequency goes to zero when the momentum vanishes ) . in order to be well defined we need @xmath66 it is alwways well defined for small frequencies , but , depending on the sign of @xmath67 .
the two solutions for @xmath68 are @xmath69 note that , depending on the parameters and on the momentum scale , this can be complex .
an imaginary part to the frequency corresponds to damped modes and instabilities . applying the machinery to the transverse modes we get @xmath70
there is always a couple of non hydrodynamic modes associated withe the vorticous excitations .
the solutions for @xmath71 are @xmath72 we have calculated the maximum possible wave velocity for @xmath73 numerically , and it is shown as a function of the coefficients in fig .
[ dispersion ] on the one hand , as one can see , polarization has important consequences on vortex non - propagation @xcite . for a general equation of state , a change in temperature due to the propagation of a sound wave will also change the susceptibility .
because of conservation of angular momentum , this will also change the vorticity .
this mixing between sound waves and vortices due to the sensitivity of the equation of state to polarization will allow vortices to propagate and create an effective hydrostatic limit even in the hydrodynamic limit in the presence of fluctuations , which could lead to stabilization @xcite for small vorticities , the `` fluctuation - driven turbulence '' phase conjectured in @xcite . on the other hand , the ideal hydrodynamic limit also leads to a non - causal propagation for both sound and group velocity , as shown in fig .
[ dispersion ] where @xmath74 and @xmath75 are systematically varied , and @xmath76 is maximized . for all significant values ( @xmath77 ) of these parameters , @xmath76 goes significantly above the causal limit .
this is actually a reasonable result , as it is a direct consequence of the fact that the free energy @xmath41 , and hence the local dynamics , is sensitive to an accelleration ( similar non - causal dissipative violations of propagation were found at higher gradient order in @xcite ) .
ostrogradski s theorem guarantees such lagrangians are unstable ( and one needs the lagrangian picture to see this instability ) , and hence hydrodynamic systems including a response to angular momentum in the eos will not have a stable ideal hydrodynamic limit .
additional terms in the lagrangian depending on @xmath24 only generally do not change this conclusion .
one would need then to take into account the polarization as a separate dynamical degree of freedom from the vorticity and include @xmath21 ( and @xmath17 since polarization is not a conserved quantity ) .
note that in this way it is possible to obtain an israel - stewart type equation @xmath78 which could resolve this issue , with an appropriate relaxation time @xmath79 , like it did for first order viscous hydrodynamics .
the procedure to use is similar to what has already been done in ref @xcite to insert a relaxation time equation for the pressure correction , however it is important to remind that the current formalism used is incompatible with dissipative processes ( indeed any time irreversible process ) , and it must be extended as it was done in ref @xcite itself .
it is mandatory to double the degrees of freedom ( or double the time dimension , with the average of the two time coordinates as the physical time ) in order to theoretically justify an effective lagrangian as the leftover of the underlying ( full ) lagrangian , after `` integrating out '' the microscopic degrees of freedom . in this extended formalism angular momentum ( and
in fact even four - momentum ) conservation has to be enforced on the with additional conditions , fore more informations on the subject see , for instance ref .
@xcite .
this is very different from ideal spinless hydrodynamics , where navier - stokes equations necessitate of relaxation dynamics but euler equations are well - defined . in hydrodynamics with polarization , there seems to be a general conflict between causality and the non - dissipative regime , one which might lead to a quantitative lower limit for dissipation ( the @xmath80 needed to restore causality ) in systems whose fundamental constituents have spin .
this will be quantitatively examined in a subsequent work . in conclusion
, we took the effective theory for ideal hydrodynamics in the presence of polarization , linearized it under hydrostatic equilibrium and examined the behavior of the lowest - lying modes , sound wave and vortical . unlike in normal ideal hydrodynamics , the two mix . on the one hand , this is likely to stabilize the vortex mode by adding an effective soft energy gap to it . on the other , causality of sound propagation
is no longer guaranteed even in the ideal limit unless polarization effects are soft enough . fixing this
most likely requires introducing relaxation ( israel - stewart ) type dynamics already from the ideal fluid limit , and this has implications for the minimum viscosity for fluids whose microscopic degrees of freedom have non - zero spin .
|
since the discovery of fe - based high-@xmath5 superconductors by kamihara _ et al_.@xcite , their many - body electronic properties have been studied very intensively .
figure [ fig : phasediagram ] shows a typical phase diagram of fe - based superconductors . in the under - doped regime , the second - order orthorhombic ( o ) structure transition occurs at @xmath6 , and the stripe - type magnetic order is realized at @xmath7 . in the o phase , the orbital polarization @xmath8 is realized , where @xmath9 is the filling of @xmath10 orbital @xcite .
also , sizable softening of shear modulus @xmath11 @xcite and the renormalization of phonon velocity @xcite indicate the development of orbital fluctuations near the orthorhombic phase .
strong spin fluctuations are also observed near the magnetic ordered phase .
these main characters of the phase diagram should be understood in order to clarify the mechanism of superconductivity .
theoretical studies of orbital polarization had been proposed in refs .
@xcite . however , _
non - magnetic _ structure transition can not be explained based on the hubbard model , once we apply the mean - field approximation or random - phase - approximation ( rpa ) . to solve this problem ,
we had recently improved the rpa by including the vertex correction ( vc ) for the susceptibility that is dropped in the rpa @xcite : by applying this self - consistent vc ( sc - vc ) method to the hubbard model for fe - based superconductors , both spin and orbital fluctuations mutually develop , and both the o structure transition and the softening of @xmath11 can be explained .
note that the `` electronic nematic state '' with large in - plane anisotropy of resistivity or magnetization well above @xmath12 @xcite also indicates the occurrence of the ( local ) orbital order @xcite .
is the structure transition temperature , which is expected to be induced by orbital polarization @xmath13 according to the arpes measurements and the sizable softening of @xmath11 .
@xmath14 is the magnetic transition temperature . ]
the phase diagram in fig .
[ fig : phasediagram ] indicates that both spin and orbital fluctuations could be closely related to the mechanism of high-@xmath5 . up to now , the spin - fluctuation - mediated @xmath2-wave state @xcite and orbital - fluctuation - mediated @xmath3-wave state @xcite had been studied based on the multiorbital models .
the @xmath2-wave state has the sign reversal of the gap between hole - like fermi surfaces ( h - fss ) and electron - like fermi surfaces ( e - fss ) , whereas the @xmath3-wave state has no sign reversal .
experimentally , the robustness of @xmath5 against impurities in many fe - based superconductors @xcite indicates the realization of the @xmath3-wave state , at least in `` dirty '' compounds with high residual resistivity @xcite .
also , the `` resonance - like '' hump structure in the neutron inelastic scattering @xcite can be explained by considering the energy dependence of the inelastic scattering if the @xmath3-wave state is realized .
although fully - gapped @xmath1-wave state is realized in many optimally - doped high-@xmath5 compounds , nodal @xmath1-wave state ( accidental node is not protected by symmetry ) is also observed in some compounds with lower @xmath5 @xcite .
the appearance of the accidental node strongly indicates the presence of `` competing pairing interactions '' @xcite . in the spin fluctuation scenario ,
if the @xmath15-orbital hole - pocket is under the fermi level , spin fluctuations in @xmath16-orbitals develop at @xmath17 while those in @xmath15-orbital develop at @xmath18 , and this frustration gives a nodal gap structure around the @xmath15-orbital part on the e - fs @xcite .
this mechanism was studied in detail in ref .
@xcite by introducing phenomenological pairing interaction .
however , the @xmath15-orbital h - fs presents in real compounds . in this case
, fully - gapped @xmath2-wave state is obtained by the rpa since spin fluctuations develop at @xmath17 in all @xmath19-orbitals , consistently with neutron experiments @xcite . very interestingly , in optimally - doped bafe@xmath0(as ,
p)@xmath0 , nodal gap structure with high-@xmath5 ( @xmath20k ) is realized .
the superconducting ( sc ) gaps on the three h - fss are fully - gapped and almost orbital - independent both in the @xmath21 plane @xcite and in the @xmath22 plane @xcite , consistently with the orbital fluctuation scenario in ref . @xcite .
also , loop - shape nodes on the e - fss are observed by angle - resolved thermal conductivity measurement in the vortex state @xcite and arpes measurements @xcite .
these results indicate the existence of competing pairing interactions , and the study of these facts would be significant to understand the mechanism of high-@xmath5 superconductivity . on the other hand , the arpes measurement by ref .
@xcite reported the horizontal node on the @xmath23-orbital e - fs around the z point in bafe@xmath0(as , p)@xmath0 , contrary to the reports by refs .
this result is consistent with the prediction of the theory of the spin - fluctuation - mediated @xmath2-wave state in ref .
however , the existence of the horizontal node would be inconsistent with the large in - plane field angle dependence of the thermal conductivity reported in ref . @xcite
also , very small @xmath24-linear term in the specific heat in the sc state would not be compatible to the presence of nodes on heavy hole - like fss @xcite .
in this paper , we theoretically study the origin of the nodal gap structure in bafe@xmath0(as , p)@xmath0 , in order to obtain a significant information of the pairing mechanism of fe - based superconductors . for this purpose
, we construct the three - dimensional ( 3d ) ten - orbital tight - binding model for bafe@xmath0(as , p)@xmath0 , and calculate the dynamical spin and orbital susceptibilities due to the combination of coulomb and quadrupole interactions . by solving the eliashberg gap equation
, we obtain the fully - gapped @xmath25-wave ( @xmath3-wave ) state due to strong spin ( orbital ) fluctuations .
when both spin and orbital fluctuations strongly develop , which will be realized near the o phase , nodal @xmath1-wave state with loop - shape nodes on the e - fss is realized due to the competition between attractive and repulsive interactions .
it is realized during a smooth crossover between @xmath3- and @xmath2-wave states @xcite .
in contrast , the sc gaps on the h - fss are fully - gapped due to orbital fluctuations .
thus , the present study explains the main characters of the gap structure in bafe@xmath0(as , p)@xmath0 . in refs .
@xcite present authors have shown that small quadrupole interaction induced by fe - ion oscillations gives rise to the large antiferro- and ferro - orbital fluctuations .
in addition , we had developed the _
spin+orbital _ fluctuation theory in multiorbital hubbard model by including the vcs to the susceptibilities , which are neglected in the rpa @xcite .
it was found that the aslamazov - larkin type vc due to coulomb interaction produces large effective quadrupole interaction .
the emergence of the orbital fluctuations due to the vc is also recognized in a simple two - orbital model , using the self - consistent vc method @xcite as well as newly developed two - dimensional renormalization group method ( rg+crpa method ) @xcite .
plane and ( b)@xmath26 plane of the present ten orbital model for the filling @xmath27 .
the green , red , blue and black lines correspond to @xmath28 , @xmath29 , @xmath15 and @xmath23 orbitals , respectively . in ( b ) , there are three h - fss ( fs1 , fs2 and fs3 ) and four e - fss ( fs4 and fs5 ) . ] in sec .
[ sec : formulatoin ] , we introduce the three - dimensional ten - orbital tight - binding model , which contains two fe - sites in each unit cell . we analyze this model based on the rpa , by taking both the coulomb and quadrupole interactions into account .
the latter interaction originates from the coulomb interaction beyond the rpa , described by the vertex corrections . in sec .
[ sec : sc ] , we analyze the sc gap equation for various model parameters , and derive the loop - shape nodes on e - fss due to the competition between orbital and spin fluctuations .
some discussions and the summary are presented in secs .
[ sec : discussion ] and [ sec : summary ] , respectively . in appendix
a , we show the obtained orbital fluctuations that give the @xmath3-wave state with nearly isotropic gap functions on the three h - fss . in appendix b
, we discuss the sc state in heavily electron - doped systems when the @xmath15-orbital h - fs disappears .
in this paper , we set @xmath30 and @xmath31 axes parallel to the nearest fe - fe bonds , and the orbital @xmath23 , @xmath28 , @xmath29 , @xmath15 , and @xmath32 are denoted as 1 , 2 , 3 , 4 , and 5 respectively .
first , we perform the local - density - approximation ( lda ) band calculation for bafe@xmath0as@xmath0 and bafe@xmath0p@xmath0 using wien2k code based on the experienced crystal structure . next
, we derive the ten - orbital tight - binding model that reproduces the lda band structure and its orbital character using wannier90 code and wien2wannier interface . @xcite using the obtained two sets of tight - binding parameters ( hopping integrals and on - site energies ) , the parameters of bafe@xmath0(as@xmath33p@xmath34)@xmath0 are well approximated by making a linear combination of them with a ratio of @xmath35 @xcite . in this paper , we use the tight - binding parameters for @xmath36 . the obtained kinetic term is given as @xmath37 where @xmath38 represent the unit cell , @xmath39 ( @xmath40 , @xmath41 ) represent the two fe sites , @xmath42 represent the @xmath19 orbital , and @xmath43 is the spin index .
@xmath44 is the position of fe - site , @xmath45 is the creation operator of the @xmath19 electron , and @xmath46 with @xmath47 and @xmath48 ( @xmath49 or @xmath50 ) is the local potential ( hopping integral ) .
however , the @xmath15-orbital h - fs given by the lda is too small compared to the experimental results by arpes measurements . in order to increase the size of @xmath15-orbital h - fs
, we introduce the following orbital - dependent potential term around the @xmath51-point : @xmath52 c^{\dagger}_{l \alpha , \sigma}(\bm{k } ) c_{l \alpha , \sigma } ( \bm{k}),\end{gathered}\ ] ] where @xmath53 is the energy shift of the orbital-@xmath54 at @xmath51-point .
we put @xmath55 ev , @xmath56 ev and the others are 0 .
the fss in this model are composed of three h - fss around @xmath51-point and four e - fss around @xmath57- and @xmath58-points .
figure [ fig : fermi ] show the obtained fss in the ( a ) @xmath22 and ( b ) @xmath26 planes , respectively .
the electron filling per fe - site is @xmath27 . in fig .
[ fig : fermi ] ( b ) , there are three h - fss ( fs1 , fs2 and fs3 ) and two e - fss ( fs4 and fs5 ) .
we call fs4 ( fs5 ) the outer ( inner ) e - fs .
next , we explain the interaction term .
we introduce both the coulomb interaction ( @xmath59 , @xmath60 , @xmath61 ) and quadrupole interaction .
the latter are induced by the electron - phonon ( e - ph ) interaction due to fe ion oscillations as follows,@xcite @xmath62 where @xmath63 , and @xmath64 is the quadrupole interaction at @xmath65 .
@xmath66 is the cutoff energy of the quadrupole interaction .
@xmath67 is the quadrupole operator introduced in ref .
@xcite , which will be shown in appendix a. @xmath68 has many non - zero off - diagonal elements as explained in refs.@xcite . as explained in ref .
@xcite , @xmath69 ( @xmath70 ) is induced by in - plane ( out - of - plane ) fe - ion oscillations . in this paper
, we put @xmath71 unless otherwise noted .
also , the aslamazov - larkin type vc due to coulomb interaction produces large effective quadrupole interaction @xmath69 @xcite .
thus , the quadrupole interaction in eq .
( [ eqn : hint ] ) is derived from both the vc and e - ph interaction .
now , we perform the rpa for the present model , by using @xmath72 meshes .
the irreducible susceptibility in the ten orbital model is given by @xmath73 where @xmath74 and @xmath75 .
@xmath76 and @xmath77 are the fermion and boson matsubara frequencies .
@xmath78^{-1}$ ] is the _ d _ electron green function in the orbital basis , where @xmath79 is the matrix elements of @xmath80 and @xmath81 is the chemical potential .
then , the susceptibilities for spin and charge sectors in the rpa are given by @xcite @xmath82 where @xmath83 @xmath84 @xmath85 where @xmath86 .
in the rpa , the enhancement of the spin susceptibility @xmath87 is mainly caused by the intra - orbital coulomb interaction @xmath59 , using the `` intra - orbital nesting '' of the fss .
on the other hand , the enhancement of @xmath88 in the present model is caused by the quadrupole - quadrupole interaction in eq .
( [ eqn : hint ] ) , utilizing the `` inter - orbital nesting '' of the fss .
the magnetic ( orbital ) order is realized when the spin ( charge ) stoner factor @xmath89 , which is the maximum eigenvalue of @xmath90 , is unity . when @xmath27 , the critical value of @xmath59 is @xmath91 ev , and the critical value of @xmath92 is @xmath93 ev for @xmath94 .
hereafter , we set the unit of energy as ev .
next , we explain the linearized eliashberg equation . in order to obtain the fine momentum dependence of the sc gap , we concentrate on the gap functions only on the fss as done in ref .
@xcite : we used @xmath95 points for each fermi surface sheet . in the presence of
dilute impurities ( @xmath96 ) , the linearized eliashberg equation is given as @xcite @xmath97 where @xmath98 is the eigenvalue that reaches unity at @xmath99 . @xmath100 and @xmath101 denote the fss , and @xmath102 is the gap function on the @xmath100-th fs ( fs@xmath100 ) at the fermi momentum @xmath103 .
the integral in eq .
( [ eqn : eliasheq ] ) means the surface integral on fs@xmath101 . the paring interaction @xmath104 in eq . ( [ eqn : eliasheq ] )
is @xmath105 @xmath106 where @xmath107 is the transformation unitary matrix between the band and the orbital representations . in eq .
( [ eqn : eliasheq ] ) , @xmath108 is given as @xmath109 where @xmath110 is the impurity induced quasiparticle damping rate .
here , we calculate the damping rate using @xmath24-matrix approximation .
we consider the case of fe - site substitution , where the impurity potential @xmath111 is diagonal in the @xmath19-orbital basis .
the @xmath24-matrix for an impurity at @xmath112 site is given as @xmath113^{-1 } \hat{i}^{\alpha } , \ ] ] which is @xmath4-independent in the orbital basis .
here , @xmath114 is the impurity potential , and @xmath115 is the local green function given as @xmath116^{\a}_{l l ' } ( \e_n ) = \frac{1}{n } \sum_{\bm{k } ' } g_{l l'}^{\a } ( \bm{k } ' , \e_n ) \\
= -s_n \frac{i \pi } { ( 2 \pi)^3 } \sum_{j } \int_{\mathrm{fs}j } \frac{d \bm{k}_{\mathrm{fs}j}'}{v^{j } ( \bm{k } ' ) } u_{l \a , j } ( \bm{k } ' ) u_{l ' \a , j}^ { * } ( \bm{k } ' ) , \end{gathered}\ ] ] where @xmath117 . in the @xmath24-matrix approximation , which is exact for @xmath118
, the normal self - energy in the band diagonal basis is given as @xmath119 where @xmath120 is the impurity concentration ratio .
then , the quasiparticle damping rate is given as @xmath121 also , @xmath122 is the impurity - induced anomalous self - energy given as @xmath123 where @xmath124 in this calculation , we simplify the energy dependence of @xmath125 .
we assume that @xmath126 ( @xmath127 ) can be separated into the momentum and orbital dependent part @xmath128 and energy dependent part @xmath129 : @xmath130 we calculated @xmath131 without approximation . on the other hand , @xmath129 is determined approximately as @xmath132 , \ ] ] where @xmath133 is the largest value of @xmath134 for any @xmath135 , and @xmath103 .
it is verified that this simplification affects the momentum dependence of the sc gap functions only quantitatively , although the obtained @xmath98 is quantitatively underestimated .
thus , this approximation would be appropriate for the present purpose , that is , the analysis of the anisotropy of the sc gap .
in this section , we analyze the linearized eliashberg equation , eq.([eqn : eliasheq ] ) , using the 3d model of bafe@xmath0(as , p)@xmath0 for @xmath27 .
hereafter , we use @xmath136 meshes for calculating charge and spin susceptibilities .
we assume that @xmath137 and @xmath138 , and fix the ratio @xmath139 . in solving the eliashberg equation , we used @xmath140 points for each fermi surface sheet and 512 matsubara frequencies . in this paper
, we perform the calculation @xmath141 and @xmath142 . and @xmath143 .
( a),(b ) sc gap functions on the h - fss in @xmath22 and @xmath144 planes .
the green , red , blue and black lines correspond to @xmath28 , @xmath29 , @xmath15 and @xmath23 orbitals , respectively .
( c ) @xmath145 dependence of the sc gaps on the h - fss in @xmath26 planes .
horizontal node appears on the fs3 around @xmath146 .
( d),(e ) sc gap functions on the e - fss in @xmath22 and @xmath144 planes .
( f),(g ) 3d gap functions on the outer and inner e - fss . ]
first , we study the spin - fluctuation - mediated @xmath25-wave superconducting state for @xmath147 by putting @xmath143 and @xmath148 . here
, we put @xmath149 ( @xmath150 ) , and the obtained eigenvalue is @xmath151 .
the obtained gap structure is almost independent of @xmath152 .
first , we discuss the sc gaps on the h - fss .
figures [ fig : s+-gap](a ) and ( b ) show the obtained gap functions on the h - fss in the @xmath22 and @xmath153 planes , respectively .
the definitions of @xmath154 and fs1 - 5 are shown in fig .
[ fig : fermi ] . in the @xmath22 plane ,
the sc gap size weakly depends on the orbital character of the fss
. however , in the @xmath155 plane , the sc gap size strongly depends on the @xmath19-orbital . especially , the sc gap on the @xmath23-orbital fs is almost zero and negative , reflecting the small spin fluctuations in the @xmath23-orbital because of the absence of the intra @xmath23-orbital nesting .
( note that @xmath23-orbital is absent on the e - fss . )
the horizontal node is clearly recognized in the sc gap in the @xmath156 plane shown in fig .
[ fig : s+-gap ] ( c ) .
the obtained horizontal node on fs3 near @xmath21 is consistent with the previous rpa calculation by suzuki _
et al _ @xcite .
the obtained horizontal node would contradict to the four - fold symmetry of the thermal conductivity @xcite and the small volovik effect in the specific heat measurement @xcite . according to arpes measurements ,
the horizontal - node was reported in ref .
@xcite , whereas it was not observed in refs .
@xcite .
next , we discuss the sc gaps on the e - fss
. figures [ fig : s+-gap ] ( d ) and ( e ) show the obtained gap functions on the e - fss in the @xmath22 and @xmath153 planes , respectively . as we can see , line nodes do not appear on the e - fss .
this result is consistent with the analysis in ref .
@xcite , that is , the @xmath25-wave gap on the e - fss is fully - gapped if the h - fs made of @xmath15-orbital appears .
note that the sc gaps for @xmath144 in fig .
[ fig : s+-gap ] ( e ) are obtained by rotating the gaps in the @xmath157 plane in ( d ) by @xmath158 . also , fig .
[ fig : s+-gap ] ( f ) and ( g ) show 3d gap functions on the outer and inner fss ( fs4 and fs5 ) , respectively . on both e - fss ,
the sc gap on the `` flat part '' is larger than that on the `` high curvature part '' .
next , we study the orbital - fluctuation - mediated @xmath3-state superconducting state for @xmath159 by putting @xmath94 and @xmath148 . here
, we put @xmath160 ( @xmath161 ) , and the obtained eigenvalue is @xmath162 .
the obtained gap structure is almost independent of @xmath163 .
figures [ fig : s++gap ] ( a ) and ( b ) show the obtained gaps on the h - fss in the @xmath22 and @xmath153 planes , respectively . in highly contrast to the spin fluctuation scenario ,
the gap size on @xmath23-orbital fs is comparable with that on the other fss , since strong orbital correlations are developed in all @xmath19-orbitals : note that the quadrupole interaction possesses many non - zero interorbital matrix elements .
the present numerical result is consistent with our previous calculation using the 2d 5-orbital model.@xcite and @xmath164 .
( a),(b ) sc gap functions on the h - fss in @xmath22 and @xmath144 planes .
the green , red , blue and black lines correspond to @xmath28 , @xmath29 , @xmath15 and @xmath23 orbitals , respectively .
( c ) @xmath145 dependence of the sc gaps on the h - fss in @xmath26 plane .
used colors are same as ( a ) and ( b ) .
( d),(e ) sc gap functions on the e - fss in @xmath22 and @xmath144 planes .
( f),(g ) 3d gap functions on the outer and inner e - fss . ]
figure [ fig : s++gap ] ( c ) shows that the sc gap size of each h - fss is approximately independent on @xmath145 , which is consistent with the small orbital dependence of the sc gap in ( ba , k)fe@xmath0as@xmath0 and bafe@xmath0(as , p)@xmath0 observed in refs .
@xcite . figures [ fig : s++gap](d ) and
( e ) show the obtained gaps on the e - fss in the @xmath22 and @xmath153 planes , respectively .
figure [ fig : s++gap ] ( f ) and ( g ) show the 3d sc gap functions on the outer and inner e - fss ( fs4 and fs5 ) , respectively .
thus , the obtained sc gaps on the e - fss is isotropic for any @xmath145 .
we also discuss the sc gap functions in the case of @xmath165 and @xmath166 in eq .
( [ eqn : hint ] ) .
figure [ fig : nooxy ] shows @xmath145 dependence of the sc gaps on the h - fss for @xmath167 ( @xmath168 ) and @xmath94 . in this case , the gap function on the @xmath23-orbital h - fs is smaller compared to the case of @xmath169 in fig .
[ fig : s++gap ] . on the other hand ,
the obtained sc gaps on the e - fss are almost isotropic , similarly to the results of @xmath169 .
dependence of the sc gaps on the h - fss in @xmath26 plane .
used parameters are @xmath170 , @xmath166 and @xmath94 . ]
recently , several measurements observed the nodal gap structure in bafe@xmath0(as@xmath33p@xmath34)@xmath0 @xcite .
this compound is very clean , and very accurate measurements of gap structure have been performed .
they present a significant challenge for theories to reproduce the observed gap structure .
as discussed in subsection iii - a and b , we can not reproduce the line - nodes on the electron fss when either spin or orbital fluctuations solely develop .
-wave gap ( shown in fig . [ fig : nodalgap1 ] ) due to the competition of orbital fluctuations ( = inter - orbital attraction ) and spin fluctuations ( = intra - orbital repulsion ) .
green , red , blue lines correspond to @xmath28 , @xmath29 , @xmath15 , and @xmath23-orbitals , respectively . ] here , we study the emergence of highly anisotropic @xmath1-wave state due to the strong orbital and spin fluctuations . in the phase diagram of bafe@xmath0(as@xmath33p@xmath34)@xmath0 , both @xmath14 and @xmath12 decrease to zero at almost the same critical point @xmath171 .
this fact means that both spin and orbital fluctuations become comparable in magnitude at @xmath172 . here
, we consider the case that the @xmath3-wave state is realized by stronger orbital fluctuations . as increasing the spin fluctuation with momentum @xmath173 , @xmath174 and @xmath175 are suppressed when both @xmath103 and @xmath176 are on the fss with the same orbital character , and finally the sign change @xmath177 could be achieved .
such strong anisotropy originates from the competition between the attractive interaction of @xmath178 and repulsive interaction of @xmath179 in eq .
( [ eqn : worb - cs ] ) . as shown in fig .
[ fig : node ] , strong spin fluctuations on the @xmath15-orbital ( due to intra @xmath15-orbital nesting ) produce the loop - shape node on the e - fs . similar anisotropic @xmath1-wave gap modified by the spin fluctuations is considered to be realized in ( y , lu)ni@xmath0b@xmath0c.@xcite , @xmath180 and @xmath181 .
( a),(b ) sc gap functions on the h - fss in @xmath22 and @xmath144 planes .
( c ) @xmath145 dependence of sc gaps on the hole fss in @xmath26 plane .
( d),(e ) sc gap functions on the e - fss in @xmath22 and @xmath144 planes .
( f),(g ) 3d gap functions on the outer and inner e - fss .
the green lines represent the gap nodes . ]
hereafter , we present numerical results in the presence of small amount of impurities ( @xmath111=1 and @xmath182 ) , just to make the sc gap functions smoother .
figure [ fig : nodalgap1 ] show the results of nearly @xmath3-wave state with nodal structure on outer e - fs .
we put @xmath183 and @xmath184 , ( @xmath185 , @xmath186 ) , and the eigenvalue is @xmath187 .
figures [ fig : nodalgap1 ] ( a)-(c ) show the obtained sc gaps on the h - fss in the @xmath22 plane , @xmath21 plane , and @xmath156 plane , respectively .
the obtained sc gaps on the h - fss are nearly isotropic and orbital - independent , similarly to the results in fig . [
fig : s++gap ] . especially , the gap size of the @xmath23-orbital h - fs is large even in the presence of loop - shape nodes on e - fss .
figures [ fig : nodalgap1 ] ( d ) and ( e ) show the obtained sc gaps on the e - fss in the @xmath22 and @xmath153 planes , respectively .
the sc gap on the inner e - fs is fully opened , and its sign is same as that on the h - fss . on the outer e - fs , in contrast , the sc gap shows the sign change near @xmath188 ( @xmath189 ) in the @xmath22 plane ( @xmath21 plane ) .
this sign change is caused by strong spin fluctuations in the @xmath15-orbital , as we have explained in fig .
[ fig : node ] . in this case
, the sc gaps on the h - fss remains fully - gapped , due to the fact that the band - mass of h - fss is larger than that of the e - fss . as results ,
closed loop - shape nodes appear in the flat part on the outer e - fs , as recognized in figs . [
fig : nodalgap1 ] ( f ) and ( g ) .
this gap structure is consistent with the prediction given by the angle - resolved thermal conductivity under the magnetic field @xcite . and @xmath190 .
( a),(b ) sc gap functions on the e - fss in @xmath22 and @xmath144 planes .
( c),(d ) 3d gap functions on the outer and inner e - fss .
the green lines represent the gap nodes . ] as increasing the value of @xmath59 ( or reducing @xmath120 ) slightly , the area of the sign reversed part on the outer e - fs increases , and the sc gap on the inner e - fs also shows the sign reversal .
figures [ fig : nodalgap3 ] ( a ) and ( b ) show the sc gap functions on the e - fss for @xmath183 and @xmath191 ( @xmath185 and @xmath192 ) .
the obtained eigenvalue is @xmath193 .
the obtained gap functions are approximately given by shifting the gaps in figs .
[ fig : nodalgap1 ] ( d ) and ( e ) downwards , and line nodes appear on both the inner and outer e - fss . as described in figs .
[ fig : nodalgap3 ] ( c ) and ( d ) , closed nodal loops appear in the flat part on the outer e - fs and in the high curvature part on the inner e - fs .
the sc gaps on the h - fss are almost the same as those shown in fig [ fig : nodalgap1 ] ( a)-(c ) , so we do not show them . and @xmath194 .
( a),(b ) sc gap functions in the e - fss on @xmath22 and @xmath144 planes .
( c),(d ) 3d gap functions on the outer and inner e - fss .
the green lines represent the gap nodes . ] as increasing the value of @xmath59 ( or reducing @xmath120 ) further , the sign of the sc gap on the outer e - fs is completely reversed , and small closed loop - nodes appear only on the inner e - fs .
the obtained sc gaps are nearly @xmath2-wave state .
figures [ fig : nodalgap2 ] ( a ) and ( b ) show the obtained gap functions on the e - fss for @xmath183 and @xmath195 ( @xmath185 , @xmath196 ) .
they are approximately given by shifting the gaps in figs .
[ fig : nodalgap3 ] ( a ) and ( b ) downwards .
the obtained eigenvalue is @xmath193 .
figures [ fig : nodalgap2 ] ( c ) and ( d ) show the obtained 3d gap functions on the outer and inner e - fss , respectively . apparently , closed nodal loops appear in the high curvature part on the inner e - fs , whereas no nodes appear on the outer e - fs .
this numerical result is consistent with the recent arpes measurement by yoshida _
et al _ @xcite .
on the other hand , the sc gaps on the h - fss are similar to those in figs .
[ fig : nodalgap1 ] ( a)-(c ) . in figs .
[ fig : nodalgap1]-[fig : nodalgap2 ] , we fixed the impurity parameters as @xmath182 and @xmath197 .
now , we discuss the sc gap functions for general impurity parameters .
figure [ fig : unimp ] ( a ) shows the @xmath59-@xmath198 phase diagram for both @xmath197 and @xmath199 . the solid ( dashed ) lines represent the boundaries between @xmath3 wave and nodal-@xmath1 wave , or nodal-@xmath1 wave and @xmath25 wave for @xmath197 ( @xmath199 ) . as decreasing @xmath200 or increasing @xmath201 ,
the following crossover would be realized : ( i ) full gap @xmath25-wave @xmath202 ( ii ) nodal @xmath1-wave @xmath202 ( iii ) full gap @xmath3-wave .
when both @xmath59 and @xmath92 are fixed , the same crossover occurs when @xmath198 increases .
the residual resistivity for @xmath197 derived from the linear response theory is about @xmath203 per @xmath204 .
we note that , in the present numerical calculation using 3d model , line - nodes can appear even if @xmath148 as shown in fig .
[ fig : unimp ] .
in contrast , in the previous calculation using 2d model @xcite , we could not obtain the line - nodes for @xmath148 , since the sc state changes from the @xmath3-wave to @xmath2-wave discontinuously as @xmath59 increases .
-@xmath198 phase diagram for @xmath168 obtained for ( a ) @xmath197 and @xmath205 with @xmath206 and for ( b ) @xmath197 with @xmath207 and @xmath208 . ] in the present study based on the rpa , the @xmath209 crossover is realized in case of @xmath210 for @xmath211 .
one of the main reasons would be the factor @xmath212 in front of @xmath213 in eq .
( [ eqn : worb - cs ] ) , reflecting the su(2 ) symmetry of the spin space .
however , this factor @xmath212 might be overestimated since the recent polarized neutron scattering measurements indicates the relation @xmath214 above @xmath5 due to the spin - orbit interaction @xmath215 @xcite .
moreover , we have recently improve the rpa by including the vc , and found that orbital fluctuations strongly develop in the hubbard model @xcite .
then , the orbital susceptibility is @xmath216 , where @xmath217 is the charge vc for the irreducible susceptibility . according to ref .
@xcite , the magnitude of the three - point vertex is estimated as @xmath218 , and then eq . ( [ eqn : worb - cs ] ) would be replaced with @xmath219 .
figure [ fig : unimp ] ( b ) shows the @xmath59-@xmath198 phase diagram for @xmath207 and @xmath208 , in case of @xmath197 .
we find that the @xmath3-wave region is widely extended , and the nodal-@xmath1-wave region is also widen .
the obtained gap structure in the crossover regime for @xmath220 is the loop - shape nodes shown in figs .
[ fig : nodalgap1]-[fig : nodalgap2 ] .
in previous sections , we analyzed the gap equations based on the three - dimensional ten orbital model for bafe@xmath0(as , p)@xmath0 .
when orbital fluctuations are solely developed , fully - gapped @xmath3-wave state is realized . on the other hand , when spin fluctuations are solely developed
, we obtain the @xmath2-wave state with horizontal node on a h - fs . during the crossover between @xmath3-wave and @xmath25-wave states due to the competition between orbital and spin fluctuations ,
the loop - shape nodes appear on the outer ( inner ) e - fs when the spin fluctuations are slightly weaker ( stronger ) than the orbital fluctuations .
the obtained phase - diagram is shown in fig .
[ fig : unimp ] .
we stress that all three h - fss are fully - gapped during the crossover , since the sc gap on the @xmath23-orbital originates from the inter - orbital nesting between different h - fss .
the crossover from the @xmath2-wave state to the @xmath3-wave state is also induced by increasing the impurity concentration . in this study , we considered the orbital - diagonal on - site impurity potential at fe @xmath100-site , considering the fe - site substitution by other elements . in this case ,
inter - band impurity scattering is always comparable to intra - band one , as shown by the @xmath24-matrix approximation in ref . @xcite .
for this reason , when the spin fluctuations are solely developed , the realized @xmath25-wave state with @xmath221k is suppressed by small amount of impurities , with small residual resistivity @xmath222 cm ( @xmath223 is the mass - enhancement factor ) . since @xmath224
, we can safely expect that the sc state in _ dirty fe - based superconductors _ ( say @xmath225 cm ) would be the @xmath3-wave state due to orbital fluctuations .
there are many important future issues . as we discussed in sec .
[ sec : loop - shape - node ] , one of our important future problems is to study the present 3d ten - orbital hubbard model ( @xmath226 ) using the sc - vc method developed in ref . @xcite . using this method ,
we have recently shown that the @xmath3-wave state is realized in the 2d five - orbital hubbard model ( @xmath226 ) .
note that the ferro - orbital fluctuations induced by the vc , which explain the orthorhombic structure transition in fig .
[ fig : phasediagram ] , enlarge @xmath5 further @xcite .
it was recently confirmed that the mechanism of orbital fluctuations due to vc had been realized even in a simple two - orbital model , using the sc - vc method as well as the recently developed two - dimensional renormalization group analysis ( rg+crpa method ) @xcite .
another important future issue is to include the self - energy due to orbital and spin fluctuations , @xmath227 and @xmath228 , into the gap equation .
they are given as @xmath229 where positive ( negative ) sign corresponds to @xmath230 ( @xmath231 ) .
the real and imaginary parts of the total self - energy @xmath232 represent the mass enhancement and quasiparticle inelastic scattering , respectively .
both effects suppress @xmath5 .
moreover , orbital and momentum dependence of @xmath233 would strongly modify the anisotropy of the sc gap functions .
thus , the self - energy correction in the gap equation will be important for the quantitative analysis of the sc gaps .
in this paper , we studied sc gap structure using ten orbital model for bafe@xmath0(as , p)@xmath0 . when the orbital fluctuations due to inter - orbital quadrupole interaction ( [ eqn : hint ] ) are strong , the @xmath3-wave state is realized .
in contrast , the @xmath2-wave state is formed by strong spin fluctuations , mainly due to intra - orbital coulomb interaction @xmath59 .
both spin and orbital fluctuations would strongly develop in the optimally - doped regime near the o phase . in this case , we find that a smooth crossover between @xmath3- and @xmath2-wave states is realized by changing the interactions or impurity concentration , without large suppression in @xmath5 . during this @xmath234 crossover ,
the loop - shape nodes are universally formed on the e - fss , as a result of the competition between inter - orbital attractive interaction and intra - orbital repulsive interaction .
this result is consistent with recent angle - resolved thermal conductivity measurement @xcite and arpes measurement @xcite . during the crossover ,
the sc gaps on the h - fss are fully - gapped and almost orbital independent due to orbital fluctuations , consistently with recent arpes measurements @xcite .
we are grateful to y. matsuda , t. shibauchi , a. fujimori , t. yoshida , s. shin , t. shimojima , d. s. hirashima and y. yamakawa for valuable discussions . this study has been supported by grants - in - aid for scientific research from mext of japan .
numerical calculations were partially performed using the yukawa institute computer facility .
in this paper , we discussed the development of orbital fluctuations due to the quadrupole interaction in eq .
( [ eqn : hint ] ) . here
, we consider the quadrupole operator at @xmath235 , where @xmath236 and @xmath237 represent the unit cell and the fe - site ( a or b ) , respectively .
then , the operator @xmath238 ( @xmath239 ) is given as @xmath240 where @xmath54 and @xmath241 represents the @xmath19-orbital , @xmath242 , and the coefficient is defined as @xmath243 for @xmath244 , where @xmath245 and so on .
the non - zero coefficients are given as @xcite @xmath246 where 1,2,3,4,5 respectively correspond to @xmath23 , @xmath28 , @xmath29 , @xmath15 , @xmath32 .
-channel quadrupole fluctuations @xmath247 for @xmath27 and @xmath248 , in the ( a ) @xmath249 plane and
( b ) @xmath250 plane .
the obtained @xmath251 dependence of @xmath247 is rather weak . ] in the presence of quadrupole interaction in eq .
( [ eqn : hint ] ) , @xmath252 given in eq .
( [ eqn : chic ] ) is strongly enhanced .
then , it is convenient to introduce the quadrupole interaction , by @xmath253 , where @xmath254a or b. in the present study , the channel @xmath255 is the most strongly enhanced , due to the good inter - orbital nesting in fe - based superconductors @xcite .
figure [ fig : chiq ] shows the obtained @xmath256 in the @xmath249 and @xmath250 planes @xcite .
the large peak at @xmath257 originates from the interorbital ( @xmath258 ) nesting between e - fs and h - fs , and the small peak at @xmath259 originates from the interorbital ( @xmath260 ) nesting between two h - fss .
therefore , the obtained development of @xmath261 and @xmath262 means the existence of strong orbital fluctuaitions on the @xmath28- , @xmath29- , @xmath15- and @xmath23-orbitals , of which the fss of bafe@xmath0(as , p)@xmath0 are composed . for this reason
, we obtain the orbital - fluctuaition - mediated @xmath3-wave state with approximately isotropic sc gap on the three h - fss , consistently with the arpes measurements in refs .
@xcite .
in addition to the af orbital fluctuation , the ferro - orbital fluctuations with respect to @xmath263 are induced by the `` two - orbiton process '' in ref @xcite as well as `` two - magnon process '' in ref .
these processes are given by the azlamasov - larkin ( al ) type vertex correction , since the al term describes the interference between ferro- and af - fluctuations that is neglected in the rpa .
the ferro - orbital fluctuations induce the softening of @xmath11 as well as the orthorhombic structure transition .
although the ferro - orbital fluctuations also contribute to the @xmath1-wave sc state , we consider that they are not the major mechanism of the sc : first , the relation @xmath264 is expected in the ferro - orbital fluctuation mediated @xmath265 wave state , since these fluctuations develop only on the @xmath28 and @xmath29 orbitals .
this relation is inconsistent with experiments .
second , we have recently solved the eliashberg gap equation based on the sc - vc method @xcite and found that the main pairing interaction is actually af orbital fluctuations with respect to @xmath266 , since the peak of the ferro - orbital susceptibility is very narrow in the momentum space .
therefore , the present analysis by taking only af orbital fluctuations would be justified .
in this paper , we have studied the gap functions for @xmath27 . in this appendix
, we discuss the heavily electron doped case ( @xmath267 ) .
figure [ fig : gap_n61_s++ ] ( a ) shows the h - fss on the @xmath26 plane . as one can see , the @xmath15-orbital h - fs disappears . ) for @xmath268 , @xmath164 and @xmath148 .
( a ) @xmath145-dependence of the h - fss , and ( b ) the obtained sc gap functions on the h - fss .
we also show the 3d gap functions on the ( c ) outer e - fs and ( d ) inner e - fs . ] now , we consider the orbital - fluctuation - mediated @xmath3-wave sc state . here
, we put @xmath268 ( @xmath161 ) and @xmath94 and obtained eigenvalue @xmath98 is 0.87 .
first , we discuss the sc gaps on the h - fss .
figure [ fig : gap_n61_s++ ] ( b ) shows the @xmath145 dependence of the sc gaps on the h - fss in the @xmath26 plane .
in contrast to the case of @xmath27 , the sc gap on the @xmath23-orbital hole fs becomes smaller , while each sc gap on the h - fss on any plane perpendicular to the @xmath145-axis is almost isotropic .
next , we focus on the sc gaps on the e - fss
. figures [ fig : gap_n61_s++ ] ( c ) and ( d ) show the 3d gap functions on the outer and inner e - fss , respectively . as with the case of @xmath27 ,
the sc gaps on the e - fss are nearly isotropic on any plane perpendicular to the @xmath145-axis . ) for @xmath143 , @xmath269 and @xmath148 .
( a),(b ) sc gap functions on the h - fss in the @xmath22 and @xmath144 planes .
( c ) @xmath145 dependence of sc gap functions on the h - fss for @xmath26 .
( d),(e ) 3d gap functions on the outer e - fs and inner e - fs . ] next , we discuss the spin - fluctuation - mediated @xmath25-wave sc state . here
, we put @xmath269 ( @xmath150 ) and @xmath143 .
the obtained eigenvalue @xmath98 is 1.68 .
figures [ fig : gap_n61_s+- ] ( a ) and ( b ) show the gap functions on the h - fss in the @xmath22 and @xmath144 planes , respectively . in common with the results of @xmath27 ,
the sc gap function on the @xmath23-orbital h - fs is almost zero since the spin fluctuations in the @xmath23-orbital do not develop .
figures [ fig : gap_n61_s+- ] ( c ) shows the @xmath145 dependence on the sc gaps on the h - fss in the @xmath26 plane .
it clearly shows the horizontal line node on fs3 near @xmath144 .
next , we discuss on the e - fss .
figures [ fig : gap_n61_s+- ] ( d ) and ( e ) show the 3d gap functions on the outer and inner e - fss , respectively . unlike the case of @xmath27 , line nodes appear on the e - fss even when @xmath226 .
this result is consistent with the analysis in ref .
@xcite : nodal gap appears on the e - fss when the @xmath15-orbital h - fs disappears because of the competition of different spin fluctuations ; @xmath270 on @xmath16-orbitals and @xmath271 on @xmath15-orbital . in the presence of @xmath15-orbital h - fs for @xmath272 , on the other hand , spin fluctuations develop at @xmath273 in all @xmath19-orbitals .
then , fully - gapped @xmath25-wave state is realized . in this case ,
the competition of orbital and spin fluctuations induce the loop - shape nodes as discussed in sec .
iii - c .
we comment that khodas and chubukov had discussed the emergence of the loop - shape nodes on the e - fss in the `` folded model '' with two fe - atoms in each unit - cell @xcite : when the `` vertical nodes '' of e - pockets are realized in the `` unfolded model '' with one fe - atoms in each unit - cell , the loop - shape nodes are realized in the folded model by taking the finite hybridization between two e - fss inherent in 122 systems @xcite into account . within the rpa ,
the vertical nodes appear only in the absence of the xy - like hole pocket at @xmath274 as discussed by kuroki et al .
@xcite , and therefore the mechanism of the `` loop - shape nodes '' by khodas and chubukov requires the absence of xy - like hole pocket , at least within the rpa .
the calculation in this appendix gives a numerical verification of the theory of khodas and chubukov .
m. yi , d. h. lu , j .- h .
chu , j. g. analytis , a. p. sorini , a. f. kemper , s .- k .
mo , r. g. moore , m. hashimoto , w. s. lee , z. hussain , t. p. devereaux , i. r. fisher , z .- x .
shen , pnas * 108 * , 6878 ( 2011 ) .
fernandes , l. h. vanbebber , s. bhattacharya , p. chandra , v. keppens , d. mandrus , m.a .
mcguire , b.c .
sales , a.s .
sefat , and j. schmalian , phys .
lett . * 105 * , 157003 ( 2010 ) a. kawabata , s. c. lee , t. moyoshi , y. kobayashi , and m. sato , j. phys .
. jpn . * 77 * ( 2008 ) suppl .
c 103704 ; m. sato , y. kobayashi , s. c. lee , h. takahashi , e. satomi , and y. miura , j. phys .
* 79 * ( 2009 ) 014710 ; s. c. lee , e. satomi , y. kobayashi , and m. sato , j. phys .
* 79 * ( 2010 ) 023702 .
s. graser , a. f. kemper , t. a. maier , h .-
cheng , p. j. hirschfeld , and d. j. scalapino , new journal of physics .
* 12 * , 073030 ( 2010 ) ; s. graser , a. f. kemper , t. a. maier , h .-
cheng , p. j. hirschfeld , and d. j. scalapino , phys .
b * 81 * 214503 ( 2010 ) .
s.maiti , m.m .
korshunov , t.a .
maier , p.j .
hirschfeld and a.v .
chubukov , phys .
lett . * 107 * , 147002 ( 2011 ) ; s. maiti , m.m .
korshunov , t.a .
maier , p.j .
hirschfeld and a.v .
chubukov , phys .
b * 84 * , 224505 ( 2011 ) t. shimojima , f. sakaguchi , k. ishizaka , y. ishida , t. kiss , m. okawa , t. togashi , c .- t . chen , s. watanabe , m. arita , k. shimada , h. namatame , m. taniguchi , k. ohgushi , s. kasahara , t. terashima , t. shibauchi , y. matsuda , a. chainani , and s. shin , sccience * 332 * , 564 ( 2011 ) .
t. yoshida , s. ideta , t. shimojima , w. malaeb , k. shinada , h. suzuki , i. nishi , a. fujimori , k. ishizaka , s. shin , y. nakashima , h. anzai , m. arita , a. ino , h. namatame , m. taniguchi , h. kumigashira , k. ono , s. kasahara , t. shibauchi , t. terashima , y. matsuda , m. nakajima , s. uchida , y. tomioka , t. ito , k. kihou , c. h. lee , a. iyo , h. eisaki , h. ikeda , r. arita , t. saito , s. onari and h. kontani , arxiv:1301.4818
|
to understand the origin of the nodal gap structure realized in bafe@xmath0(as , p)@xmath0 , we study the three - dimensional gap structure based on the three - dimensional ten - orbital hubbard model with quadrupole interaction . in this model ,
strong spin and orbital fluctuations develop by using the random - phase - approximation . by solving the eliashberg gap equation
, we obtain the fully - gapped @xmath1-wave state with ( without ) sign reversal between hole - like and electron - like fermi surfaces due to strong spin ( orbital ) fluctuations , so called the @xmath2-wave ( @xmath3-wave ) state .
when both spin and orbital fluctuations strongly develop , which will be realized near the orthorhombic phase , we obtain the nodal @xmath1-wave state in the crossover region between @xmath3-wave and @xmath2-wave states .
the obtained nodal @xmath1-wave state possesses the loop - shape nodes on electron - like fermi surfaces , due to the competition between attractive and repulsive interactions in @xmath4-space .
in contrast , the sc gaps on the hole - like fermi surfaces are fully - gapped due to orbital fluctuations .
the present study explains the main characters of the anisotropic gap structure in bafe@xmath0(as , p)@xmath0 observed experimentally .
|
four - positions theory studies the geometry of four proper euclidean displacements , see ( * ? ? ?
* chapter 5 , 89 ) .
often , it is considered as a discretization of differential properties of order three of a smooth one - parameter motion . in this article
we take a different point of view and regard the four displacements as an elementary cell in a quadrilateral net of positions , that is , the discretization of a two- or more - parameter motion .
motivated by applications in the rising field of discrete differential geometry @xcite we intend to study quadrilateral nets of positions such that the relative displacement between neighbouring positions is a pure rotation .
the present text shall provide foundations for this project .
thus , the questions we are interested in are induced from discrete differential geometry .
nonetheless , our contributions seem to be of a certain interest in their own right .
denote the four displacements by @xmath0 , @xmath1 , @xmath2 , @xmath3 and the relative displacements of consecutive positions by @xmath4 ( indices modulo four ) .
we require that every relative displacement @xmath5 is a _ pure rotation . _ in this case we call the quadruple @xmath6 a _ rotation quadrangle .
_ rotation quadrangles with additional properties ( the displacements @xmath7 and @xmath8 are also rotations ) have been studied in ( * ? ? ?
* chapter 5 , 9 ) .
our contribution is more general and we provide answers to questions not asked in @xcite . in section [ sec : construction - rotation - quadrilaterals ] we address the problem of constructing rotation quadrilaterals .
kinematic mapping turns out to be a versatile tools for this task .
subsequently , we study points @xmath9 whose homologous images @xmath10 lie on a circle ( section [ sec : homologous - points ] ; figure [ fig : homologous - points - planes - lines ] , left ) , oriented planes @xmath11 whose homologous images @xmath12 are tangent to a cone of revolution ( section [ sec : homologous - planes ] ; figure [ fig : homologous - points - planes - lines ] , center ) , and oriented lines @xmath13 whose homologous images @xmath14 form a skew quadrilateral on a hyperboloid of revolution ( section [ sec : homologous - lines ] ; figure [ fig : homologous - points - planes - lines ] , right ) .
one position in a rotation quadrilateral , say @xmath0 , can be chosen arbitrarily without affecting the quadrilateral s geometry .
the remaining positions can be constructed inductively by suitable rotations but the final position @xmath3 must be such that both @xmath15 and @xmath16 are rotations .
this is certainly possible , albeit a little tedious , in euclidean three space .
the quadric model of euclidean displacements , obtained by study s kinematic mapping @xcite seems to be a more appropriate setting for this task .
it transforms two positions corresponding in a relative rotation to conjugate points on the study quadric @xmath17 .
more precisely , the composition of all rotations about a fixed axis with a fixed displacement is a straight line in the study quadric .
thus , a rotation quadrilateral corresponds , via study s kinematic mapping , to a skew quadrilateral @xmath18 in the study quadric .
we require here and in the following that @xmath18 spans a three - space which is not completely contained in the study quadric @xmath17 .
this genericity assumption is not fulfilled in the special configuration studied in ( * ? ? ?
* chapter 5 , 9 ) . the quadrilateral @xmath18 is uniquely determined by three suitable chosen vertices @xmath19 , @xmath20 , @xmath21 and the connecting line of @xmath21 and the missing vertex @xmath22 .
alternatively , @xmath18 is determined by two opposite vertices , say @xmath19 and @xmath21 , and two opposite edges . in any case
the input data defines a three - space whose intersection with the study quadric is a hyperboloid @xmath23 containing @xmath18 .
the completion of the missing data is easy . denoting the relative revolute axes in the moving space by @xmath24 the translation of these considerations into the language of kinematics
reads : [ th : rotation - quadrangle-1 ] a rotation quadrangle is uniquely determined by 1 .
a fixed displacement @xmath25 , three revolute axis @xmath24 , @xmath26 , @xmath27 , and two rotation angles @xmath28 and @xmath29 ( indices modulo four ) or 2 . by two displacement @xmath25 , @xmath30 and two relative revolute axes @xmath24 , @xmath27 ( indices modulo four ) .
computing the actual rotation angles and line coordinates of the revolute axes from the quadrilateral @xmath18 is straightforward .
the necessary formulas are , for example , found in ( * ? ? ?
* satz 13 ) .
beware that @xcite uses conventions for the study parameters that slightly differ from modern authors , for example @xcite or @xcite .
therefore minor adaptions to the formulas may be necessary .
now we turn to the study of the locus of points in the moving space whose homologous images lie on a circle ( figure [ fig : homologous - points - planes - lines ] , left ) . as usual
we consider straight lines as special circles of infinite radius . if the four relative displacements are general screws , the sought locus is known to be an algebraic curve of degree six @xcite . in case of rotation quadrilaterals this is still true but the curve is highly reducible : [ th : homologous - points ] the locus of points in the moving space whose homologous images lie on a circle consists of the four relative revolute axis @xmath31 , @xmath32 , @xmath33 , @xmath34 and their two transversals @xmath35 and @xmath36 . by the general result of @xcite
the sought locus can consist of six lines at most .
we show that all points on the six lines mentioned in the theorem have indeed homologous images on circles . for points
@xmath37 this is trivially true since we have @xmath38 . since any transversal carries four points of this type , lemma [ lem : coplanar - points ] below , implies the same property for all points on @xmath35 and @xmath36 .
( at least the first part of this lemma is well - known .
an old reference is @xcite . )
[ lem : coplanar - points ] consider a straight line @xmath13 in the moving space and four affine displacements .
if @xmath13 contains four points whose homologous images are co - planar than this is the case for all points of @xmath13 .
if , in addition , the homologous images of three points lie on circle , this is the case for all points of @xmath13 .
we can parametrize the homologous images of @xmath13 as @xmath39 such that points of the same parameter value @xmath40 are homologous .
the co - planarity condition reads @xmath41 ( we omit the argument @xmath40 for sake of readability ) . since it is at most of degree three in @xmath40 , it vanishes identically if four zeros exist .
this proves the first assertion of the lemma .
four homologous points @xmath42 , @xmath43 , @xmath44 , @xmath45 lie on a circle ( or a straight line ) , if three independent bisector planes @xmath46 of @xmath47 and @xmath48 have a line , possibly at infinity , in common .
this is equivalent to @xmath49 corresponding points are assumed to be coplanar .
therefore there exist scalars @xmath50 , @xmath51 and @xmath52 , not all of them zero , such that @xmath53 hence , the circularity condition becomes @xmath54 it is at most quadratic in @xmath40 .
therefore , existence of three zeros implies its vanishing and the second assertion of the lemma is proved .
the set of circles through quadruples of homologous points originating from a relative revolute axis @xmath24 is not very interesting in the context of this article . of more relevance
is the set of circles generated by points of the transversal lines @xmath35 and @xmath36 .
[ th : hyperboloid ] the circles through homologous images of points of a transversal line ( @xmath35 or @xmath36 ) lie on a hyperboloid of revolution . ]
we consider only points on the transversal @xmath35 .
any two of its successive homologous images @xmath55 , @xmath56 intersect in a point @xmath57 , the intersection point of @xmath58 and @xmath59 .
therefore the lines @xmath60 , @xmath61 , @xmath62 , @xmath63 are the edges of a skew quadrilateral and lie on a pencil of hyperboloids .
the intersection points with circles through homologous points form congruent ranges of points on the lines @xmath60 , @xmath61 , @xmath62 , @xmath63 and the intersection points @xmath57 correspond in this relation .
consider now a circle @xmath64 such that every intersections point @xmath65 with @xmath55 lies outside the segment @xmath66 $ ] for @xmath67 . setting @xmath68 ( figure [ fig : trajectory - circles ] ) we see that @xmath69 and @xmath70 .
this implies @xmath71 . in other words ,
both pairs of opposite edges in the skew quadrilateral @xmath72 have the same sum of lengths .
this is a well - known criterion for the existence of a hyperboloid of revolution through @xmath60 , @xmath61 , @xmath62 , and @xmath63 .
in this section we study the locus of planes @xmath11 in the moving space such that its homologous images @xmath73 , @xmath74 , @xmath75 , @xmath76 are tangent contact to a cone of revolution ( figure [ fig : homologous - points - planes - lines ] , center ) .
actually we will not discuss this in full generality but restrict ourselves to the case of _ oriented _ planes in _ oriented _ contact with a cone of revolution .
this means the cone axis is the intersection line of the _ unique _ bisector planes of @xmath77 and @xmath78 .
having in mind applications in discrete differential geometry ( a kinematic generation of conical nets , see @xcite ) this additional assumption is justified .
since the offset surface of a cone of revolution is again a cone of revolution , every plane parallel to a solution plane @xmath11 is a solution plane as well .
we infer that the sought locus of planes consists of _ pencils of parallel planes_. this can also be confirmed by direct computation .
denote by @xmath79 the plane coordinate vector of @xmath11 .
if @xmath80 is the homogeneous transformation matrix describing @xmath25 , the coordinate vectors @xmath81 of the homologous planes are found from @xmath82 the condition that they have a point in common reads @xmath83 and is of degree four in @xmath84 , @xmath85 , @xmath86 , and @xmath87 . assuming that holds
, the condition on the conical position can be stated in terms of the oriented normal vectors @xmath88 which are obtained from @xmath81 by dropping the first ( homogenizing ) coordinate .
the planes are tangent to a cone of revolution if the four points with coordinate vectors @xmath88 lie on a circle ( note that these vectors have the same length ) .
this condition results in a homogeneous cubic equation @xmath89 which is nothing but the circle - cone condition known from four - position synthesis of spherical four - bar linkages ( see @xcite ) .
a closer inspection shows that there exists a homogeneous polynomial @xmath90 of degree four such that becomes @xmath91 hence , solution planes are characterized by the simultaneous vanishing of @xmath92 and @xmath93 .
since these equations are independent of @xmath84 , the value of the homogenizing coordinate can be chosen arbitrarily .
this is the algebraic manifestation of our observation that the solution consists of pencils of parallel planes . since the two homogeneous equations @xmath92 and @xmath93 are of respective degrees four and three there exist at most twelve pencils of parallel planes whose homologous images are tangent to a cone of revolution .
but not all of them are valid : the circle - cone equation @xmath93 also identifies vectors @xmath94 , @xmath95 , @xmath96 , @xmath97 that span a two - dimensional sub - space . the corresponding planes @xmath73 , @xmath74 , @xmath75 , @xmath76 are all parallel to a fixed direction but do not qualify as solution planes and have to be singled out . the complete description of all solution planes is given in [ th : homologous - planes ] the locus of planes in the moving space whose homologous images are tangent to a cone of revolution consists of six pencils of parallel planes : 1 .
the planes orthogonal to one of the four revolute axes @xmath31 , @xmath32 , @xmath33 , @xmath34 and 2 .
the planes orthogonal to one of the two transversals @xmath35 or @xmath36 of these axes .
we already know that there exist at most twelve parallel pencils of solution planes , obtained by solving the quartic equation @xmath98 and the cubic equation @xmath89 . by lemma [ lem : spherical - coplanar ] below
( which actually has been proved in a more general context in @xcite ) , six of these pencils of planes are spurious and have to be subtracted from the totality of all solution planes .
thus , we only have to show that the planes enumerated in the theorem are really solutions .
this is obvious in case of a plane @xmath11 orthogonal to axis @xmath24 since @xmath77 and @xmath78 are identical . in case of a plane orthogonal to @xmath35 or @xmath36 it is an elementary consequence of theorem [ th : hyperboloid ] .
[ lem : spherical - coplanar ] consider four spherical displacements @xmath99 , @xmath100 , @xmath101 , and @xmath102 . then there exist in general six lines ( possibly complex or coinciding ) through the origin of the moving coordinate frame , whose homologous lines are co - planar
. the co - planarity condition applied to a vector @xmath103 reads @xmath104 we compute classes of proportional solution vectors by solving the two cubic equations @xmath105 it has nine solutions .
but the three fixed directions ( one real , two conjugate imaginary ) of the relative rotation @xmath106 are spurious so that only six valid solutions remain .
finally , we also investigate the locus of lines @xmath13 in the moving frame whose homologous images @xmath107 , @xmath108 , @xmath109 , @xmath110 form a skew quadrilateral on a hyperboloid of revolution ( figure [ fig : homologous - points - planes - lines ] , right ) . as usual , we restrict ourselves to generic configurations .
moreover , we make an additional assumption on the line s orientation : setting @xmath111 ( indices modulo four ) we require that , when walking around the edges of the skew quadrilateral @xmath112 we follow the line s orientation only on every second edge ( figure [ fig : homologous - points - planes - lines ] , right ) .
the manifold of lines in euclidean three - space depends on four parameters .
the mentioned condition on the four homologous images @xmath107 , @xmath108 , @xmath109 , @xmath110 of a line @xmath13 imposes , however , five conditions : four intersection conditions for the skew quadrilateral and one further condition for the hyperboloid of revolution .
therefore , we expect no solution for a general screw quadrangle . on the other hand we already know ( theorem [ th : hyperboloid ] ) that for rotation quadrangles the homologous images of two transversals @xmath35 and @xmath36 form the required configuration
. we will show that these are the only solutions . considering the skew quadrilateral condition only , it is obvious that either * @xmath113 and @xmath114 intersect @xmath58 , possibly at a point at infinity , or * @xmath113 and @xmath114 are orthogonal to @xmath58 .
only the first condition is compatible with our requirement on the orientation of the homologous lines .
therefore , it is necessary that @xmath13 intersects all lines @xmath24 ( indices modulo four ) .
hence @xmath115 or @xmath116 .
combining this result with theorem [ th : hyperboloid ] we arrive at [ th : homologous - lines ] the two transversals @xmath35 and @xmath36 of the four relative revolute axes @xmath31 , @xmath32 , @xmath33 , and @xmath34 are the only lines whose homologous images form a skew quadrilateral ( with proper orientation ) on a hyperboloid of revolution .
we investigated geometric structures related to four displacements @xmath25 ( @xmath67 ) such that the relative displacements @xmath4 are pure rotations .
in particular , we completely characterized the points , planes and lines whose homologous images lie on a circle , are tangent to a cone of revolution or form a skew quadrilateral on a hyperboloid of revolution ( with proper orientation ) . in a next step
we plan to study quadrilateral nets of proper euclidean displacements such that neighbouring positions correspond in a pure rotation . using the results of this article
it is possible to show that independent pairs of principal contact element nets @xcite can occur as trajectories of such nets .
since mechanically generated motions are often composed of pure rotations the theory might also be useful in more applied settings .
m. husty and h .-
schrcker . algebraic geometry and kinematics . in ioannis
z. emiris , frank sottile , and thorsten theobald , editors , _ nonlinear computational geometry _ , volume 151 of _ the i m a volumes in mathematics and its applications _ , chapter algebraic geometry and kinematics .
springer , 2009 .
|
we consider the geometry of four spatial displacements , arranged in cyclic order , such that the relative motion between neighbouring displacements is a pure rotation .
we compute the locus of points whose homologous images lie on a circle , the locus of oriented planes whose homologous images are tangent to a cone of revolution , and the locus of oriented lines whose homologous images form a skew quadrilateral on a hyperboloid of revolution .
keywords : four - positions theory , rotation quadrilateral , homologous points , homologous planes , homologous lines , study quadric . + msc 2010 : 70b10
|
observations have discovered eight small but spatially resolved thermal coronae of early - type galaxies in hot cluster ( kt@xmath7 3 kev ) environments ( vikhlinin et al .
2001 ; yamasaki , ohashi & furusho 2002 ; sun et al .
2005 , s05 hereafter ) .
as summarized in s05 , the coronae of massive galaxies can survive in spite of the evaporation and gas stripping by the hot and dense icm , as well as the energy imparted by the central agn outbursts .
thus , the properties of these `` survivors '' can put constraints on important physics , e.g. , microscopic transport processes , gas stripping , stellar mass injection and agn heating ( e.g. vikhlinin et al .
2001 ; s05 ) .
the presence of a dense gas component ( m@xmath8 m@xmath9 with a high central density of 0.1 - 0.3 @xmath10 ) potentially provides the fuel for the central supermassive black hole ( smbh ) , in environments where the amount of the galactic cold gas is at a minimum . moreover , their properties shed light on the evolution history of their hosts . for the previous galaxy coronae discovered in rich clusters ,
there is no strong evidence for a high velocity of the host galaxy relative to the surrounding icm or for a strong nuclear activity of the host galaxy . in this paper
, we show an extreme example for which the galaxy is known to be moving very fast ( @xmath11 2500 km / s ) through a hot cluster and to host a powerful agn with strong radio emission .
ngc 1265 ( 3c 83.1b , 4c 41.06 ) is the prototype of narrow - angled tail ( nat ) radio sources ( ryle & windram 1968 ) .
ngc 1265 is in the perseus cluster , 27@xmath12 northwest of the cluster center ( fig .
its radial velocity ( 7536 km / s , huchra , vogeley & geller 1999 ) is @xmath0 2300 km / s higher than that of the perseus cd galaxy ngc 1275 ( 5264 km / s , huchra et al .
1999 ) , indicating a large infalling velocity for ngc 1265 .
highly complicated radio morphologies are revealed from 2 cm to 92 cm ( owen , burns & rudnick 1978 ; odea & owen 1986 ; sijbring & de bruyn 1998 ) .
two narrow east - west radio jets emerge from the nucleus , but are quickly bent towards the north to form two tails , which are further merged 4@xmath12 north of the nucleus .
radio images at 49 cm and 92 cm ( sijbring & de bruyn 1998 ) further reveal a low surface brightness extension of the tail that bends at least three times over an angle of almost 360 .
the projected total length of the tail is at least 40@xmath12 ( or 0.88 mpc ) .
this complex morphology was tentatively explained as a result of the galaxy s infalling orbit and the bulk flow of the icm gas to the east ( sijbring & de bruyn 1998 ) . despite the unique appearance of the galaxy in the radio , optically it is a rather normal elliptical .
a small dust lane is revealed by the image ( fig .
1 ) , which is almost perpendicular to the jet direction and aligned with the galaxy major axis .
the galaxy is not a strong x - ray emitter from the data ( rhee , burns & kowalski 1994 ) .
we obtained a deep ( 94 ksec ) observation on this galaxy and the x - ray source is fully resolved .
the velocity of the perseus cluster is 5366 km / s ( struble & rood 1999 ) .
we use a redshift of 0.018 , to calculate the luminosity distance of ngc 1265 , assuming h@xmath13 = 70 km s@xmath14 mpc@xmath14 , @xmath15@xmath16=0.3 , and @xmath17=0.7 .
the luminosity distance is 78.4 mpc and 1@xmath1 corresponds to 0.366 kpc .
we use the galactic absorption of 1.46@xmath410@xmath18 @xmath6 , which is consistent with the absorption column derived from the x - ray analysis .
the solar photospheric abundance table by anders & grevesse ( 1989 ) is used in the spectral fits .
uncertainties quoted are 1 @xmath19 .
a 93.9 ksec observation was performed with the advanced ccd imaging spectrometer ( acis ) on march 15 - 16 , 2003 .
the optical axis lies on the ccd s3 and the data were telemetered in very faint mode .
standard data analysis is performed which includes the correction for the slow gain change .
since the ccd s1 is off , we investigated the background light curve from the source - free region on the ccd s3 . excluding time intervals with significant background flares results in a total exposure of 65.7 ksec .
although the cluster emission is still significant across the whole field , any weak background flares do not affect our analysis on small scales , since a local background is used .
for the analysis of the cluster diffuse emission , we used the period d background file , `` aciss_d_7_bg_evt_271103.fits . ''
the particle background level ( measured in pha channels 2500 - 3000 adu ) was 1.1% higher than that of the period d background data .
thus , we increased the background normalization by 1.1% .
we corrected for the acis low energy quantum efficiency ( qe ) degradation , which increases with time and is positionally dependent .
the calibration files used correspond to calibration database 3.0.0 from the x - ray center . in the spectral analysis , a low energy limit of 0.5 kev is used to minimize the calibration uncertainties in low energy . despite the gigantic extent ( @xmath20 ) of ngc 1265 in the radio ( odea & owen 1986 ; sijbring & de bruyn 1998 ) ,
its x - ray extent is tiny ( @xmath21 ) , as shown in fig .
the x - ray emission is composed of a central point source and diffuse emission .
the central point source is located within 0.1@xmath1 of the nucleus determined from the observation ( fig .
( http://cxc.harvard.edu/cal/aspect/celmon/ ) ] the soft x - ray emission ( 0.5 - 1.5 kev ) is asymmetric , with a sharp edge 3@xmath1 south of the nucleus and an extension to the north . the asymmetry is quantitatively shown by the 0.5 - 1.5 kev surface brightness profiles ( exposure - corrected ) along the south , north and east - west sectors ( fig .
the sectors are defined in the caption of fig .
the 0.5 - 1.5 kev surface brightness profiles in the east and west differ very little so we combine them . the surface brightness to the south decreases rapidly at @xmath22 and reaches the background level beyond @xmath23 , while the brightness profile to the north show excess above the local background to at least 25@xmath1 .
this morphology implies the motion of the x - ray source towards the south and the action of ram pressure by the surrounding icm .
the azimuthally averaged profile is flat beyond 30@xmath1 , which marks the local background level .
the nuclear point source dominates the hard x - ray emission ( fig .
as there is no statistically important difference on the 2 - 6 kev surface brightness profiles in the south and north , only the azimuthally averaged profile is shown ( fig .
the point spread function ( psf ) of the central point source was derived with the ray tracer ( chart ) , assuming the best - fit absorbed powerlaw spectrum derived from the spectral analysis ( discussed later in @xmath242.3 ) .
since the x - ray source is only 7.1@xmath1 from the optical axis of the observation , the psf is symmetric . in the 0.5 - 1.5 kev band ,
the psf size is much smaller than the source size ( fig .
2 ) . in the 2 - 6 kev band , a model with the psf plus
the local background matches the data well within the inner @xmath25 , but under - estimates the observed brightness of 2@xmath26 at a level of 5.2 @xmath19 .
this hard x - ray excess is indicative of low mass x - ray binary ( lmxb ) emission in ngc 1265 , which will be discussed next .
we assume that the lmxb light profile follows the stellar light profile which is derived from the observation .
a bright star 3@xmath1 east of the nucleus distorts the galaxy image and affects the ground photometry of the galaxy .
the center of the star is saturated in the hst f702w image , but not in the f673n image .
we subtract the star light and derive the light profile of the galaxy .
the galaxy contributes 70% of the total light in the f702w image .
the stellar type is unknown , so we simply adopt the same percentage for the b band .
the total b magnitude of the galaxy measured from the ground is 12.45 mag after extinction correction and k - correction .
thus , l@xmath27 l@xmath9 for ngc 1265 . from the l@xmath28 - l@xmath29 scaling relation derived by sarazin , irwin & bregman ( 2001 ) , the expected lmxb light distributions in the 0.5 - 1.5 kev and 2 - 6 kev band
are shown in fig .
2 . in the hard band , a model with a central agn ( psf ) , local background and lmxb light roughly re - produces the data , but over - estimate the observed emission by @xmath0 15% . in the soft band ,
the lmxb emission is able to account for the emission beyond 9@xmath1 , where the surface brightness profile is much flatter than that within 9@xmath1 .
however , as the lmxb light does not blur the southern sharp edge ( fig .
2 ) , the lmxb emission must be discrete and dominated by those brightest binaries . in the following , we constrain the spectral properties of the x - ray source and the gas properties of the corona . the spectral properties of the ngc 1265 x - ray source were measured . in view of the asymmetry
, we extracted the global spectrum from a 10@xmath1 radius circle centered 4.6@xmath1 north of the nucleus .
the region is chosen to enclose all x - ray emission of ngc 1265 above the local background ( fig .
2 ) , although the x - ray bright region is considerably smaller ( fig .
the background spectrum was extracted from the 17@xmath30 annulus centered on the nucleus .
we first used a model with two components ( powerlaw + mekal ) . if only galactic absorption is applied , the fit is acceptable but the nuclear source has an unusually flat spectrum ( table 1 ) .
the presence of a dust lane on the image may imply extra absorption around the central source .
thus , we allowed additional absorption for the powerlaw component . as shown in table 1 ,
the fit is much improved and the derived powerlaw index ( @xmath0 2.0 ) is typical for an agn .
an absorbed nucleus is also supported by the flat 0.5 - 1.5 kev surface brightness profile within the central 1@xmath1 ( fig .
2 ) . in any case
, the corona temperature measurement is robust ( 0.60 - 0.65 kev ) .
however , in this fit , the lmxb emission ( not suffering the excess absorption column ) can not be separated from the nuclear emission . to better
separate the thermal corona , the central agn and the lmxb emission , we extracted spectra in two regions , within 1.6@xmath1 radius of the nucleus ( including nearly 90% of the nuclear emission , fig .
2 ) , and the global region excluding the central 1.6@xmath1 ( 1.6@xmath31 , but note the different centers ) .
separate spectral fits were performed as shown in table 1 . for the inner region ,
we apply a model with a mekal plus an absorbed powerlaw . for the outer region
, we apply a model with a mekal plus thermal bremsstrahlung emission from lmxb .
the background - subtracted spectra are shown in fig .
3 with the best fits of the two - component model .
abundances can not be well constrained ( particularly the upper bound ) and are fixed at the solar value .
the spectral fits to both regions give an abundance lower limit of 0.5 solar ( 90% confidence level ) . the allowed abundance change and
the inclusion of the hard component do not affect the robustness of the temperature measurement .
we derived the deprojected temperature of the inner region ( the fifth row in table 1 ) with the standard non - parametric `` onion peeling '' technique ) .
the spectral properties of the 1.6@xmath32 annulus , where the x - ray emission is still rather symmetric , were also examined ( the sixth row in table 1 ) .
gas temperature decreases from 0.7 kev at the outskirts to 0.45 kev within the central 0.6 kpc high - density core , implying the action of radiative cooling . as shown in fig .
2 , the central point source also accounts for some emission in the outer region ( @xmath0 29% ) and there may be lmxb emission within 1.6@xmath1 at the level of @xmath0 5%
. however , these extra components bring little change to the fits and computed parameters ( including luminosities ) .
the best - fit absorption excess for the agn is @xmath3310@xmath5 @xmath6 .
the 2 - 10 kev and bolometric luminosities of the central agn are 5.5@xmath410@xmath34 ergs s@xmath14 and 3.0@xmath410@xmath35 ergs s@xmath14 respectively .
the 0.5 - 2 kev and bolometric luminosities of the corona are 3.3@xmath410@xmath34 ergs s@xmath14 and 4.8@xmath410@xmath34 ergs s@xmath14 respectively .
the 0.3 - 10 kev luminosity of the 1.6@xmath31 lmxb component is 1.8@xmath410@xmath34 ergs s@xmath14 .
we also examined the spectral properties of the north extension .
its spectrum was extracted from the region defined in table 1 .
the result confirms that the north extension is from the emission of thermal gas , while the lmxb component only contributes to @xmath0 5% of the 0.5 - 1.5 kev counts there .
the soft coronal emission dominates in the 0.5 - 1.5 kev band . within 1.6@xmath1 radius
, the nuclear emission only contributes to 6% of the total 0.5 - 1.5 kev brightness .
although the lmxb emission may contribute to @xmath0 24% of the total 0.5 - 1.5 kev emission in the 1.6@xmath31 region , the contribution drops to 10% in a smaller region enclosing all soft x - ray bright region ( a 6.5@xmath1 radius circle centered at 3.2@xmath1 north of the nucleus ) .
thus , the 0.5 - 1.5 kev surface brightness well delineates the morphology of the soft corona .
we need to know the properties of the surrounding icm to study the corona - icm interaction .
if ngc 1265 is on the plane of the sky ( 27@xmath12 from the cluster center ) , the electron density of the surrounding icm is 5.9 @xmath4 10@xmath36 @xmath10 from the data ( ettori , fabian & white 1998 ) .
if the galaxy is within 20@xmath12 ( or 439 kpc ) of the plane of the sky along the line of sight , the icm density is @xmath37 @xmath10 .
we take a typical icm electron density of 5 @xmath38 @xmath10 in this work , but keep the uncertainty in mind .
the spectrum of the surrounding icm ( 16@xmath39 ) is studied , as well as that of the whole icm emission on the s3 ccd .
the surrounding icm ( projected ) is 10 times hotter ( @xmath0 6.7 kev , table 1 ) than the small galaxy corona ( @xmath0 0.6 kev ) .
the icm temperature is consistent with the value for this region ( furusho et al .
2001 ) .
we want to characterize the gas distribution of the corona in different directions with the derived 0.5 - 1.5 kev surface brightness profiles .
the contribution of the nuclear emission to the soft band and the lmxb emission have to be subtracted .
the first component is derived from the psf and the spectral fits in @xmath242.3 , while the lmxb component is estimated from the optical light .
both components are small contributions to the total x - ray emission within 8@xmath1 .
the lmxb component in the south sector is not subtracted , as the scaled lmxb light from the optical light largely over - estimates the actual lmxb light beyond 3@xmath1 in the south .
nevertheless , any lmxb contribution to the bright core within 2@xmath1 is small .
since the x - ray source is so small , the psf has to be considered in the fits .
we applied a model composed of a @xmath40 model + a constant background for profiles in the south , north and east - west , as well as the azimuthally averaged profile .
we also tried to put a truncation radius to the @xmath40 model , especially for the southern profile .
the results are shown in table 2 . without a truncation radius ,
the fit to the southern profile yields an unphysical @xmath40 ( @xmath41 10 ) , which is thus not shown .
the fits to the profiles in the south and north are shown in fig .
the southern profile is clearly truncated at the sharp edge ( 2.2@xmath42 south the nucleus ) with a large jump of surface brightness across the edge ( @xmath0 106 ) , while the truncations in other directions are not significant with a small surface brightness jump of @xmath43 3 - 5 . despite the large uncertainties , gas distributions ( assuming the same emissivity ) in all directions are consistent with the same one , but with different truncation radii ( or no truncation ) .
we also constrain the width of southern edge by smearing the truncated @xmath40 model with a gaussian exp(-r@xmath44/2@xmath45 ) .
the best - fit model requires @xmath19 = 0 and the 95% upper limit of @xmath19 is only 0.2 kpc .
we can compare the ngc 1265 corona with the two large coronae in a1367 ( s05 ) .
both luminous coronae in s05 are largely symmetric and the surface brightness jump across the presumably pressure - confined boundary is at most 4 - 9 .
thus , the sharp edge only 2.2@xmath1 ( or 0.81 kpc ) south of the ngc 1265 s nucleus is certainly unique and can not be explained by the confinement of the still icm pressure .
the gas density and mass can be derived from the surface brightness distribution .
the best - fit thermal models of inner and outer regions give almost identical emissivity , if the metallicity is chosen the same . for simplicity , we assume a constant metallicity and a constant emissivity for the whole corona . in view of the unsymmetric shape of the ngc 1265 corona
, we applied two methods to derive the density profile and the total gas mass .
the first is to use the fit to the azimuthally averaged 0.5 - 1.5 kev profile derived above . assuming a metallicity of 0.5 - 1.5 solar , the central electron density is 0.28 - 0.48 @xmath10 , and the total gas mass is 3.2 - 5.4 @xmath46 m@xmath9 ( within r@xmath47 of 7.7 kpc ) .
the second method is to assume that the corona is composed of two hemispheres characterized by the surface brightness profiles of the south and the north respectively . assuming a metallicity of 0.5 - 1.5 solar , the central electron density is 0.22 - 0.37 @xmath10 , and the total gas mass is 2.4 - 4.2 @xmath48 m@xmath9 ( within r@xmath47 ) .
thus , the central electron density of the corona is 0.22 - 0.48 @xmath10 and the total gas mass is 2.4 - 5.4 @xmath48 m@xmath9 . assuming a central abundance of 0.5 - 1.5 solar , the central gas cooling time is only 5 - 15 myr . with the derived gas density profile , the x - ray ism - icm thermal pressure ratios at the southern edge and the northern tail can be estimated .
we assume an ism temperature of [email protected] kev ( table 1 ) and applied monte - carlo simulations to calculate the errors .
the derived x - ray ism - icm pressure ratios are 24@xmath4916 at the southern edge and 1.8@xmath50 at the end of the northern tail .
although the errors are large , the values indicate that the static icm pressure is not enough to produce the southern sharp edge .
x - ray emission , at the significance levels of 2.2 - 3.6 @xmath19 , is found at the positions of three radio inner knots in the jets ( e1 , e2 and w1 from odea & owen 1986 ) .
a total of @xmath0 28 counts are collected from these three knots . assuming a power law spectrum with a photon index of 2.0 , the 0.5 - 10 kev luminosities are @xmath51 ergs s@xmath14 , 6@xmath52 ergs s@xmath14 and 5@xmath52 ergs s@xmath14 for e1 , e2 and w1 respectively .
recent observations show that x - ray jet is common in galaxies with radio jets ( e.g. , worrall , birkinshaw & hardcastle 2001 ; sambruna et al .
however , this deep observation of ngc 1265 implies that x - ray jet can be very faint in some cases ( at least 40 times fainter than the faintest jet in worrall et al .
we estimate the 1 kev x - ray to 8-ghz radio flux density ratio over the x - ray detected regions of the jet , @xmath53 , which is smaller than what is generally found for low - power radio galaxies where the jet x - ray emission is considered as synchrotron emission ( @xmath54 , see the list in evans et al .
non - detection of ngc 1265 s optical jet is also against the synchrotron interpretation . in spite of the non - detection of the continous x - ray jets ,
there are spatial indentations east and west of the corona ( fig .
1a ) that are aligned with the radio jets .
these features imply the interaction between the narrow jets and the x - ray corona .
the widths of the indentations are comparable to the widths of the radio jets .
the sharp x - ray edge is a strong indication of the ism - icm interaction . the mean free paths of particles near the edge are : @xmath55 = 27 kpc , @xmath56 = 1.5 ( n@xmath57/0.1 @xmath10)@xmath14 pc , @xmath58 = 60 ( @xmath56/1.5 pc ) pc , and @xmath59 = 5.4 kpc , assuming temperatures of 6.7 kev and 0.7 kev for the icm and the ism near the edge respectively . as @xmath55 is much larger than the size of the corona ( @xmath0 4 kpc in diameter ) , the icm may not be treated as fluid relative to the corona
furthermore , as the galaxy with only stars is rather porous and the corona is so small , it is questionable whether a bow shock can be generated ahead of the corona to change the fluid pattern there .
a collisionless bow shock may still be generated as implied by the weak 92 cm emission detected ahead of the x - ray edge ( sijbring & de bruyn 1998 ) , but this depends on the magnetic field structure around the corona .
the magneto - hydrodynamic ( mhd ) interaction between the galactic magnetic field and the icm particles ( and the magnetic field frozen into them ) is however poorly understood .
@xmath59 derived above is much larger than the width of the edge ( @xmath43 0.2 kpc ) , which can only be understood if the actual mean free paths of particles are largely suppressed by the magnetic field at the icm - ism boundary .
recent simulations of gas stripping and the evolution of x - ray coronae in clusters ( e.g. , toniazzo & schindler 2003 ; acreman et al . 2003 ) tried to study the problem with only fluid dynamics , but did not address the small sizes of x - ray coronae , nor the inclusion of mhd .
the observed properties of the ngc 1265 corona demonstrate that it is necessary to include magnetic fields , especially for small coronae moving in the hot clusters as the mean free path of particles is sensitive to temperature ( @xmath60 t@xmath44/n ) .
our knowledge of the magnetic field in elliptical galaxies is poor ( reviewed by widrow 2002 ) .
moss & shukurov ( 1996 ) proposed two types of seed fields in elliptical galaxies : stellar magnetic field ejected by sne and stellar winds , and magnetic remnants that arise if elliptical galaxy forms from mergers of spirals .
these seed fields have to be amplified by dynamos .
they proposed two types of dynamos : acoustic turbulence driven by snia and stellar winds , and vortical turbulence driven by stellar motion .
since the magnetic reynolds number is large , the amplified magnetic field should be frozen in the gas ( or the hot gas in elliptical galaxies ) .
when the galaxy falls into the dense icm , the galactic magnetic field may be compressed and stretched at the boundary , responsible for the suppression of transport processes . however
, this simple picture may be too naive as there are a lot of details unclear .
moreover , the small corona of ngc 1265 is surrounded by the hot icm filling the ngc 1265 galaxy . how the field within the corona can be magnetically isolated from the field in the surroundings ( still in the galaxy )
is unclear , especially if the turbulent diffusion of magnetic field is important ( e.g. , lesch & bender 1990 ) .
assuming the radial velocity difference of ngc 1265 and the perseus cluster is ngc 1265 s radial velocity relative to the cluster , the velocity of ngc 1265 ( v@xmath61 ) is 2170/cos@xmath62 km / s , where @xmath62 ( between 0 and 90 degree ) is the angle between the line of sight and the opposite of the moving direction .
the radial velocity dispersion of the perseus cluster is 1324 km / s ( struble & rood 1999 ) , which corresponds to a 3d velocity dispersion of 2293 km / s ( @xmath19 ) .
ngc 1265 s velocity can not be too high ( e.g. , 3@xmath19 , or 6880 km / s , @xmath62 = 71.6 ) for the galaxy to remain bound .
the sharpness of the x - ray edge implies that @xmath62 is not small .
otherwise , the edge will be smeared by projection . in principle
, we can use the sharp x - ray edge to constrain the velocity of the galaxy .
this first requires a good understanding of projection .
we performed a number of simulations .
the shape of the corona is assumed to be prolate with the nucleus at the center .
since the front of the corona is ablated and compressed by ram pressure , the front is approximated by a large sphere cut by the prolate ellipsoid , while the center of the sphere lies north of the nucleus .
the small corona is assumed to be axially symmetric to the direction of the motion . the coronal density profile is represented by a @xmath40 model with a truncation at the boundary of the corona ( see @xmath242.5 ) .
the projected morphology of the simulated corona matches the observed morphology well .
we compare the projected profiles with the observed ones and find that the x - ray sharp edge can not be preserved for @xmath63 45 .
as @xmath62 decreases from 90 , the front edge has to be shifted towards the nucleus to allow the projected edge to match the observation .
thus , the best - fit velocity increases as @xmath62 decreases , opposite from the dependence of velocity on @xmath62 from the radial velocity constraint .
therefore , the velocity can be in principle constrained well by these two methods .
however , several uncertainties prevent a tight constraint on the velocity .
first , there is a degeneracy between the icm density and the galaxy velocity .
second , the coronal density can vary 70% for an abundance change of 0.5 - 1.5 solar .
lastly , the mhd properties of the icm are not known . for an ambient density of 5@xmath64 @xmath10 ,
the best - fit angle is 50 - 65 and the velocity is 3500 - 5000 km / s if we treat the icm as fluid .
two interesting facts are : 1 ) the surrounding icm density is @xmath65 @xmath10 if ngc 1265 s velocity is @xmath43 5000 km / s and the abundance of the corona is @xmath43 1.2 solar ; 2 ) the abundance of the corona has to be @xmath43 3 solar ( a tighter upper limit than that derived from the spectral analysis ) for the ism gas to be dense enough to withstand the ram pressure . combining these constraints ,
we conclude ( conservatively ) : 45 @xmath66 67 and 3100 km / s @xmath43 v@xmath67 5500 km / s ( note the local sound speed is 1.3@xmath68 km / s ) .
as the corona moves through the icm with v@xmath61 of @xmath11 3100 km / s , it is subject to ram pressure stripping and other stripping processes by the surrounding icm ( e.g. , nulsen 1982 ) .
the analysis of ram pressure stripping in s05 ignores the high ism pressure .
the survival of ngc 1265 s corona in the face of high speed ram pressure stripping and the presence of a sharp edge indicate that high thermal pressure of the inner cooling core is the key to preserve the corona from ram pressure stripping , although the outskirts of the corona are very vulnerable to stripping .
as was discussed in the last section , the physics around the small corona should be best described by mhd .
the kelvin - helmholtz ( k - h ) instability may or may not be suppressed . lacking a good understanding of the relevant physics , only a qualitative discussion is presented .
no matter how we treat the surrounding icm , as fluid or as particles , the typical mass - loss rate of the corona by k - h instability or by the bombardment of the icm particles is approximately ( e.g. , nulsen 1982 ) : @xmath69 this process acts at the boundary of the corona .
we calculate the stellar mass injection rate in these two regions , @xmath70 and 1.6@xmath31 . using the generally adopted mass injection rate @xmath71 = 0.15 m@xmath9 yr @xmath72 l@xmath73 ( faber & gallagher 1976 ) and the optical luminosity in the chosen region ( from the data )
, we estimate 0.04 m@xmath9/yr and 0.15 m@xmath9/yr for the inner and outer region respectively . the mass deposition rate by cooling , @xmath74m@xmath75l@xmath76/5kt , is calculated in each region to be 0.19 m@xmath9/yr and 0.16 m@xmath9/yr .
thus , in the inner region , the mass deposition rate is higher than the expected stellar mass injection rate , while in the outer region , the expected mass loss rate by stripping is comparable ( if not higher , because of the uncertainties of equation 1 ) to the expected stellar mass injection rate . although the mass loss by stripping decreases rapidly to be @xmath77 the stellar injection at radius @xmath43 1 - 1.6@xmath1 , the cooling rate changes little , if there is no feedback on this small scale .
the estimated life time of the current corona is then only 0.1 - 0.2 gyr .
there are several factors that can reduce the effects of stripping and offset cooling to help the corona to survive much longer .
first , the stripping by transport processes can be suppressed by magnetic field .
second , stellar mass injection is a key ( e.g. , shown in the simulations by acreman et al .
while the faber & gallagher ( 1976 ) result is an average , the host galaxies with surviving coronae may have systematically larger stellar mass injection rates than the average .
stellar mass injection outside the corona can also affect the icm flow pattern there , although the injected ism gas is not likely to sink into the corona before being stripped and mixed with the icm . for ngc 1265 ,
if we compare the mass flow rate of the icm and the expected stellar mass injection rate outside the corona upstream , on the cross - section of the corona , the latter is only @xmath0 10% of the icm mass flux , and is not likely to be a significant factor .
however , combined with the galactic potential , magnetic field and stellar motion , the icm flow inside the galaxy ( but outside the corona ) should be much more complicated than the free flow far away from the galaxy
. this may significantly affect the stripping rate estimated by equation 1 .
third , snia in principle can provide enough heating energy . using the sn rate estimates from cappellaro , evans & turatto ( 1999 ) ,
the snia rate within the size of the ngc 1265 corona is @xmath0 0.2 / 100 yr .
assuming an energy release of 6 @xmath78 ergs per snia event , the total energy release of sne ia is 3.8 @xmath79 ergs s@xmath14 , which is comparable to the x - ray luminosity of the corona .
however , it has been long known that most of sn energy does not couple to the hot gas from the observed l@xmath80 - l@xmath29 relation of the x - ray coronae of early - type galaxies ( e.g. , brown & bregman 1998 ) .
thus , the question is how snia dissipates its kinetic energy into the hot gas .
lastly , vikhlinin et al .
( 2001 ) proposed that largely reduced heat flux from the surrounding icm can offset cooling inside the corona if the heat conductivity in the corona is close to the spitzer value . as shown in vikhlinin et al .
( 2001 ) and s05 , the large temperature gradient from the icm to the ism implies that heat conduction must be suppressed . assuming a corona radius of 2 kpc and an ambient density of 5@xmath64 @xmath10
, we estimate for ngc 1265 that heat conduction has to be suppressed by at least a factor of @xmath0 60 at the icm - ism boundary , using the method presented in s05 .
we notice that ngc 1265 s ism must be multi - phase inside the small corona , as implied by the presence of the dust lane .
dust with mass @xmath81 m@xmath9 was detected in @xmath0 80% of nearby large elliptical galaxies ( e.g. , van dokkum & franx 1995 ) .
dust detection in radio galaxies is @xmath0 2 times higher than in radio - quiet galaxies , and the dust lane in radio galaxies is usually perpendicular to the radio axis ( van dokkum & franx 1995 ) , just as is found in ngc 1265 .
it has been suggested that the central dust in elliptical galaxies is the product of stellar mass loss from red giant stars ( mathews & brighenti 2003 ) .
this scenario links dust with the 10@xmath82 k coronal gas ( also thought to have a stellar origin ) , although it is unknown what the cooling product of the hot gas is .
both cold dust and the cooling product of the high - density x - ray gas ( related or not ) can provide fuel to the smbh . ngc 1265 s nucleus is detected in both radio and x - ray ( fig .
the radio observations show that the central radio source is a compact flat - spectrum source ( @xmath43 0.1@xmath1 , owen et al .
1978 ) , with a total radio luminosity of 7.3@xmath52 ergs s@xmath14 ( 10 mhz - 5 ghz ) .
the x - ray nuclear source is absorbed ( n@xmath3 = 1.5 - 3@xmath410@xmath5 @xmath6 ) with a 2 - 10 kev luminosity of @xmath8310@xmath34 ergs s@xmath14 .
this high amount of absorption is unusual for fr i radio galaxies , as only ngc 4261 ( 3c 270 ) has a similar high absorption column in a systematic study of 25 fr i radio galaxies ( donato , sambruna & gliozzi 2004 ) .
the conclusion of donato et al .
( 2004 ) is that fr i radio galaxies generally lack a standard optically thick torus , but the exception of ngc 1265 implies that its nuclear x - ray emission comes from somewhere close to the nucleus and some optically thick clouds exist between the nucleus and our line of sight .
the central agn can have significant impacts on the small corona through its jets , especially in outbursts .
ngc 1265 is luminous in radio ( 3c 83.1b ) , 8.9 jy at 1.4 ghz . the radio flux index is 0.67 ( from nasa / ipac extragalactic database ) . from 10 mhz
to 5 ghz , the total radio luminosity is 3.7@xmath84 ergs s@xmath14 , while the luminosity of the two jets is 2.6@xmath85 ergs s@xmath14 ( odea & owen 1987 , distance adjusted to our value ) .
the jet power is known to be much larger than the radio luminosity of the source .
the jet power is distributed in relativistic electrons , ions and magnetic field .
the power in ions usually dominates .
as the jets of ngc 1265 are bent by the icm ram pressure , the kinetic power of its jets can be estimated ( e.g. , jones & owen 1979 ) .
we assume that the jet flow is non - relativistic , while the power of relativistic jets is much larger . from euler s equation
, we have : @xmath86 v@xmath87 p@xmath88 r / r@xmath89 , where @xmath86 is the particle density in the jet , v@xmath89 is the velocity of the jet flow , r@xmath89 is the radius of the jet , r is the radius of curvature of the bent jet , and p@xmath88 is the icm ram pressure . the kinetic power of jets is : l@xmath90 @xmath91
r@xmath92 @xmath86 v@xmath93 . with the relation from the bending of jets and adopting r@xmath89 = 1 kpc , r = 12 kpc ( from the radio image ) , v@xmath89 = 10@xmath94 km / s , n@xmath95 = 5@xmath64 @xmath10 and v@xmath61 = 3500 km / s , l@xmath96 = 4.3 @xmath97 ergs s@xmath14 .
r@xmath89 could be smaller but v@xmath61 may be larger .
their product should not change a lot based on the relation from the bending of jets .
the energy in magnetic field is only @xmath0 10% - 25% of l@xmath96 if the minimum pressure magnetic field is adopted ( 20 - 30 @xmath98 g from odea & owen 1987 ) .
the agn has been active for @xmath99 yr from the length of the tail .
we estimate that the total jet power over the active phase of the nucleus is at least 1600 times the thermal energy in the current corona . in view of the great vulnerability of the current corona to agn feedback ,
we conclude that the jets carry great amount of energy without dissipation through the small corona , as we previously suggested in s05 for weaker radio sources associated with the coronae in ngc 3842 and ngc 4874 .
the interaction of jets and x - ray plasma is highly suggestive from the small x - ray indentations 2.5@xmath1 east and west of the nucleus ( fig .
the radio emission of the jets is anti - correlated with the x - ray emission of the corona ( fig .
1 ) , as the radio emission of the jets turn on after they leave the dense corona . the jets may undergo a transition after they leave the dense corona , possibly because of the change in the external pressure .
the coronal thermal pressure , 1.9@xmath100 ( n@xmath101/0.1 @xmath10 ) ( kt/0.6 kev ) dynes @xmath6 , is larger than the minimum pressure in the inner knots of the jets , 6 - 9@xmath102 dynes @xmath6 ( odea & owen 1987 ) , while the icm thermal pressure is only 10@xmath103 dynes @xmath6 .
thus , the ism thermal pressure may be enough to confine the jets , while the icm pressure may not .
the edge - brightening of the jets from the first knots implies the action of ram - pressure ( odea & owen 1986 ) .
the jets are clearly bent by the icm ram pressure , rather than the pressure gradient in the ism , as originally suggested by jones & owen ( 1979 ) .
if the dense core of the corona cools , the cooling product may feed the central smbh .
the stellar velocity dispersion of ngc 1265 is unknown . using the faber - jackson relation and m@xmath104-@xmath19 relation ( tremaine et al .
2002 ) , the mass of the smbh at the nucleus of ngc 1265 is @xmath105 m@xmath9 . assuming this mass and a central gas temperature of 0.45 kev ,
the bondi accretion radius of the central smbh is 36 pc ( or 0.1@xmath1 ) if the surrounding coronal gas has little residual bulk motion relative to the smbh . the bondi accretion rate and the accretion luminosity
are 0.0031 m@xmath9/yr and 1.8@xmath106 ergs s@xmath14 ( assuming a radiation efficiency of 0.1 ) for a central electron density of 0.3 @xmath10 .
the bondi accretion luminosity is known to largely over - estimate the luminosity of many low luminosity agn ( e.g. , loewenstein et al . 2001 ) .
the possible explanations include that the accretion process is radiation inefficient , and other processes ( e.g. , feedback ) reduce the bondi accretion rate significantly ( e.g. , gliozzi , sambruna & brandt 2003 ) .
the main results of this deep observation of the prototype nat galaxy ngc 1265 are : \1 . a small ( @xmath0 4 kpc ) and asymmetric thermal corona
is detected around the nuclear region of ngc 1265 .
a sharp edge is found 2.2@xmath1 ( 0.8 kpc ) south of the nucleus , while a @xmath107 extension is present to the north .
this indicates a high velocity motion of ngc 1265 towards the south on the plane of sky , which is consistent with the long radio tail to the north .
the southern edge is very sharp , with a jump of surface brightness of @xmath0 106 and a width of @xmath43 0.2 kpc . as the mean free paths of particles in the icm ( 27 kpc ) and from the corona to the icm ( 5.4 kpc )
are much larger than the width of the edge , magnetic field around the edge is required to significantly reduce the particle diffusion .
the interaction between the small coronae and the icm must be studied by mhd .
the temperature of the coronal gas decreases from 0.7 kev at the outskirts to 0.45 kev at the center , while the temperature of the surrounding perseus icm ( projected ) is @xmath0 6.7 kev .
the coronal abundance is consistent with solar .
the 0.5 - 2 kev luminosity of the corona is 3.3@xmath85 ergs s@xmath14 .
the central electron density is @xmath0 0.22 - 0.48 @xmath10 , while the total gas mass is @xmath108 m@xmath9 .
\4 . an absorbed ( n@xmath3 = 1.5 - 3 @xmath109 @xmath6 ) low luminosity x - ray nucleus is detected , with a 2 - 10 kev luminosity of @xmath110 .
this amount of high nuclear absorption column is not usual for fr i radio galaxies .
three inner radio knots are detected in x - rays .
indentations east and west of the corona indicate the interaction between jets and hot gas .
there is an anti - correlation between the emission of the radio jets and the x - ray corona ( fig .
this implies that jets have undergone a transition as they leave the dense corona .
. the great vulnerability of ngc 1265 s corona to the large power carried by jets from the central smbh implies that agn deposit all energy outside the current corona .
this also may imply that agn heating can not quench the cooling very close to the nucleus ( several kpc ) in `` cooling - flow '' clusters .
\8 . if the stripping , cooling and evaporation are not suppressed , the current corona can only survive for 0.1 - 0.2 gyr .
however , enhanced stellar mass injection , suppressed stripping and snia heating ( if energy coupled to the hot gas ) can significantly increase the lifetime of the corona .
for survival of ngc 1265 s corona , heat conduction has to be suppressed by at least a factor of @xmath0 60 at the icm - ism interface .
the survival of ngc 1265 s corona from ram pressure stripping implies that coronae of similar or more massive galaxies can survive @xmath41 1000 km / s stripping in environments like the core of the coma cluster .
a systematic study based on data is underway , which will better understand the evolution of these small corona in rich clusters , including their fates , fueling of the smbh , interaction with the environment , and the relation with large cooling cores .
we thank alexey vikhlinin , bill forman & paul nulsen for inspiring discussions .
we are grateful to c. p. odea and ylva pihlstrom for providing us the vla images of ngc 1265 shown in fig .
1 . we thank the referee for helpful comments .
we acknowledge support from the nasa contracts g02 - 3184x and smithsonian institution .
acreman , d. m. , stevens , i. r. , ponman , t. j. , sakelliou , i. 2003 , mnras , 341 , 1333 anders , e. , & grevesse n. 1989 , geochimica et cosmochimica acta , 53 , 197 brown , b. a. , bregman , j. n. 1998 , apj , 495 , l75 cappellaro , e. , evans , r. , turatto , m. 1999 , a&a , 351 , 459 donato , d. , sambruna , r. m. , gliozzi , m. 2004 , apj , 617 , 915 ettori , s. , fabian , a. c. , white , d. a. 1998 , mnras , 300 , 837 evans , d. a. , hardcastle , m. j. , croston , j. h. , worrall , d. m. , birkinshaw , m. 2005 , mnras , in press ( astro - ph/0502183 ) faber , s. & gallagher , j. 1976 , apj , 204 , 365 furusho , t. , yamasaki , n. y. , ohashi , t. , shibata , r. , ezawa , h. 2001 , apj , 561 , l165 gliozzi , m. , sambruna , r. m. , & brandt , w. n. 2003 , a&a , 408 , 949 huchra , j. p. , vogeley , m. s. , geller , m. j. 1999 , apjs , 121 , 287 jones , t. w. , owen , f. n. 1979 , apj , 234 , 818 lesch , h. , & bender , r. 1990 , a&a , 233 , 417 loewenstein , m. , mushotzky , r. f. , angelini , l. , arnaud , k. a. , quataert , e. 2001 , apj , 555 , 21l mathews , w. g. , brighenti , f. 2003 , apj , 590 , 5l moss , d. , & shukurov , a. 1996 , mnras , 279 , 229 nulsen , p. e. j. 1982 , mnras , 198 , 1007 odea , c. p. , & owen , f. n. 1986 , apj , 301 , 841 odea , c. p. , & owen , f. n. 1987 , apj , 316 , 95 owen , f. n. , burn , j. o. , rudnick , l. 1978 , apj , 226 , l119 rhee , g. , burns , j. o. , kowalski , m. p. 1994 , aj , 108 , 1137 ryle , m. , windram , m. d. 1968 , mnras , 138 , 1 sambruna , r. m. et al .
2004 , apj , 608 , 698 sarazin , c. l. , irwin , j. a. , bregman , j. n. 2001 , apj , 556 , 533 sijbring , d. , & de bruyn , a. g. 1998 , a&a , 331 , 901 struble , m. f. , & rood , h. j. 1999 , apjs , 125 , 35 sun , m. , vikhlinin , a. , forman , w. , jones , c. , murray , s. s. 2005 , apj , 619 , 169 ( s05 ) toniazzo , t. , & schindler , s. 2001 , mnras , 325 , 509 tremaine , s. et al . 2002 , apj , 574 , 740 van dokkum , p. g. , franx , m. 1995 , aj , 110 , 2027 vikhlinin , a. , markevitch , m. , forman , w. , jones , c. 2001 , apj , 555 , l87 widrow , l. m. 2002 , rvmp , 74 , 775 worrall , d. m. , birkinshaw , m. , hardcastle , m. j. 2001 , mnras , 326 , 7l yamasaki , n. y. , ohashi , t. , & furusho , t. 2002 , apj , 578 , 833 note : the @xmath40-model fits to the 0.5 - 1.5 kev s@xmath80 in different directions ( truncated or not ) .
the local psf is included in the fits .
f@xmath111 is the flux of the local background , which slightly decreases from the south ( closer to the cluster center ) to the north .
cts s@xmath14 arcsec@xmath112 , separated by a factor of 2 ) superposed on its own image ( exposure corrected , smoothed by one acis pixel ) .
the 2 - 6 kev contours ( green , from 0.25 to 8@xmath64 cts s@xmath14 arcsec@xmath112 , separated by a factor of 2 ) are also shown ( exposure corrected , smoothed by one acis pixel ) .
the sharp edge @xmath23 south of the nucleus is prominent , as well as the extension to the north .
a 20 cm vla image of the jets is shown as blue contours ( image from ylva pihlstrom , 1.29@xmath113 beam size ) .
the radio jet emission is anti - correlated with the x - ray emission of the corona .
x - ray emission from three radio knots e1 , e2 and w1 ( fig . 1 of odea & owen 1986 ) , is also detected .
there are indentations in the east and west of the corona along the jet direction , aligned with the knots .
this indicates an interaction between the jets and the x - ray gas .
b ) the 0.5 - 1.5 kev contours superposed on the wfpc2 image ( f702w filter ) .
a dust lane , nearly perpendicular to the jet direction and aligned with the major axis of the galaxy , is detected within the central 1@xmath1 radius . a nearby bright star ( @xmath23 east of the nucleus ) distorts the optical light of the galaxy .
the outer contours of the optical light are also plotted .
the x - ray light is asymmetric compared to the optical light .
c ) contours ( red ) of the perseus cluster superposed on the dss i image .
the cluster center ( ngc 1275 ) is on the lower left .
a low resolution vla image of 3c 83.1b ( blue contours ) is shown ( image from chris odea , fig .
11 of odea & owen 1986 ) , with ngc 1265 at the head of the nat source .
the tail of the radio galaxy is at least 40@xmath12 in projection and bends sharply at least 4 times ( see more figures in odea & owen 1986 and sijbring & de bruyn 1998 ) . the region around the galaxy ( 2.3@xmath114 ) is also enlarged on the left insert , where contours of a 6 cm vla image ( from chris odea ) are superposed on the dss ii image . the optical light of ngc 1265 can be traced to the edge of this subfield .
[ fig : img : smo ] ] and 120@xmath1 .
the dotted line is the optical light profile determined from the observations and is normalized to match the predicted lmxb luminosity from the scaling relation by sarazin et al .
( 2001 ) , with the assumption that the lmxb distribution follows the optical light .
the solid line is the sum of the previous two components .
the southern surface brightness profile shows a sharp edge at @xmath23 ( smoothed by the psf ) from the nucleus , while the extension in the north can be traced to at least 8@xmath1 ( excluding the possible lmxb component ) . *
right * : the 2 - 6 kev surface brightness profile .
the yellow line , dashed line and the dotted line represent the same components as those in the left panel , but in the 2 - 6 kev band .
the black solid line is the sum of the psf and the local background , which clearly underestimates the data from 2@xmath1 to 8@xmath1 .
the red line is the sum of the psf , the local background and the lmxb component scaled from the optical light , which roughly reproduces the data .
[ fig : img : smo ] ] ) and outer ( 1.6@xmath31 ) regions of the ngc 1265 x - ray source , with the two - component models shown as the fourth and the seventh rows in table 1 ( dot - dash line : corona emission ; dashed line : absorbed nuclear emission or lmxb emission ; solid line : the sum ) .
the blend of iron l lines is significant in both spectra .
the shapes of the spectra at e @xmath11 1.5 kev are different .
the absorbed nucleus dominates in the hard x - rays for the inner region , while the lmxb is a significant component for the outer region .
[ fig : img : smo ] ] model with a constant background .
the estimated nuclear emission at the 0.5 - 1.5 kev band is subtracted from the profile .
the best - fit model has been convolved with the local psf , which smoothes the sharp edge .
the surface brightness jump across the edge is @xmath0 106 .
* lower * : the 0.5 - 1.5 kev surface brightness profile of the north sector with the best - fit model , a @xmath40 model with a constant background .
the estimated nuclear emission and the lmxb emission is subtracted from the profile .
the best - fit model has been convolved with the local psf .
the truncation of the @xmath40 model is not significant . despite the not so small errors , the gas distributions of these two sectors are consistent with the same one , but with different truncation radii ( table 2 ) .
[ fig : img : smo ] ]
|
a deep observation of ngc 1265 ( 3c 83.1b ) , the prototype for the narrow - angled - tailed ( nat ) radio galaxy , reveals a small cool x - ray thermal corona ( @xmath0 0.6 kev ) embedded in the hot icm of the perseus cluster ( @xmath0 6.7 kev ) .
the corona is asymmetric with a sharp edge ( @xmath02.2@xmath1 , or 0.8 kpc from the nucleus ) to the south and an extension to the north ( at least @xmath2 from the nucleus ) , which is interpreted as the action of ram pressure while solely the static icm confinement is unable to explain .
we estimate that the corona is moving with a velocity of @xmath0 2.4 - 4.2 times the local sound speed to the south .
the presence of the sharp edge for this small corona indicates that the transport processes are largely suppressed by the magnetic field there .
the magnetic field around the corona also suppresses heat conduction by at least a factor of @xmath0 60 across the corona boundary .
we conclude that it is unrealistic to study the interaction of the small x - ray coronae with the hot icm without the consideration of the roles that magnetic field plays , a factor not included in current simulations .
an absorbed ( n@xmath3 = 1.5 - 3 @xmath4 10@xmath5 @xmath6 ) nucleus is also detected , which is not usual for fr i radio galaxies .
weak x - ray emission from three inner radio knots in the jets is also detected .
indentations at the east and west of the corona indicate interaction between the jets and the x - ray corona .
narrow jets carry great amounts of energy out of the central agn and release the energy outside the corona , preserving the tiny and vulnerable corona .
this case reveals that the inner kpc core of the corona of massive galaxies can survive both high - speed stripping and powerful agn feedback .
thus , the cooling of the x - ray coronae potentially provides fuel to the central smbh in rich environments where the amount of the galactic cold gas is at a minimum .
|
babar s discovery of the @xmath0 state@xcite generated strong interest in heavy meson spectroscopy , chiefly due to its surprisingly low mass with respect to expectations .
these expectations are based on quark models or lattice gauge theory .
unfortunately , at present large lattice systematic errors do not allow a determination of the @xmath4 mass with a precision better than several hundred mev . and ,
although quark models appear to be exceptionally accurate in describing charmonia , they are less constrained by experiment and on a weaker theoretical footing in the open charm sector .
it is therefore imperative to examine reasonable alternative descriptions of the open charm sector .
the @xmath0 was produced in @xmath5 scattering and discovered in the isospin violating @xmath6 decay mode in @xmath7 and @xmath8 mass distributions .
its width is less than 10 mev and it is likely that the quantum numbers are @xmath9@xcite .
finally , if the @xmath10 mode dominates the width of the @xmath0 then the measured product of branching ratios@xcite br(b^0 d_s(2317)k ) br(d_s(2317 ) d_s ^0 ) = ( 4.4 0.8 1.1)10 ^ -5 implies that @xmath11 , consistent with the @xmath0 being a canonical @xmath12 @xmath1 meson . in view of this , cahn and jackson
have examined the feasibility of describing the masses and decay widths of the low lying @xmath3 and @xmath4 states within the constituent quark model@xcite .
they assume a standard spin - dependent structure for the quark - antiquark interaction ( see below ) and allow general vector and scalar potentials .
their conclusion is that it is very difficult to describe the data in this scenario .
indeed , the @xmath0 lies some 160 mev below most model predictions ( see ref.@xcite for a summary ) , leading to speculation that the state could be a @xmath13 molecule@xcite or a tetraquark@xcite .
such speculation is supported by the isospin violating discovery mode of the @xmath0 and the proximity of the s - wave @xmath13 threshold at 2358 - 2367 mev .
other studies have been made with qcd sum rules@xcite , using heavy quark symmetry to examine decay models@xcite , or in unitarised chiral models@xcite .
although these proposals have several attractive features , it is important to exhaust possible canonical @xmath1 descriptions of the @xmath0 before resorting to more exotic models .
here we propose a simple modification to the standard vector coulomb+scalar linear quark potential model that maintains good agreement with the charmonium spectrum and agrees remarkably well with the @xmath3 and @xmath4 spectra .
possible experimental tests of this scenario are discussed .
the spectra we seek to explain are summarised in table [ ddsspectratab ] . unfortunately
, the masses of the @xmath14 ( labelled @xmath15 ) and @xmath16 ( labelled @xmath17 ) are poorly determined .
belle have observed@xcite the @xmath14 in @xmath18 decays and claim a mass of @xmath19 mev with a width of @xmath20 , while focus@xcite find @xmath21 mev with a width @xmath22 . while some authors choose to average these values , we regard them as incompatible and consider the cases separately below . finally , there is an older mass determination from belle@xcite of @xmath23 mev with a width of @xmath24 .
the @xmath25 has been seen in @xmath18 decays to @xmath26 and @xmath27 by belle@xcite .
a breit - wigner fit yields a mass of @xmath28 mev and a width of @xmath29 mev .
alternatively , a preliminary report from cleo@xcite cites a mass of @xmath30 mev and a width of @xmath31 mev .
finally , focus@xcite obtain a lower neutral @xmath25 mass of @xmath21 mev .
other masses in table [ ddsspectratab ] are obtained from the pdg compilation@xcite .
the bottom flavoured meson spectra of table [ btab ] have been obtained with the ` average ' extended model parameters and @xmath32 gev . as with the open charm spectra , a flavour - dependent constant was fit to each pseudoscalar .
the second row reports recently measured p - wave @xmath18 meson masses@xcite ; these are in reasonable agreement with the predictions of the first row
. when these results are ( perhaps incorrectly ) extrapolated to light quark masses , light scalar mesons are possible .
thus a simple @xmath33 interpretation of the enigmatic @xmath34 and @xmath35 mesons becomes feasible .
finally , the work presented here may explain the difficulty in accurately computing the mass of the @xmath36 in lattice simulations .
if the extended quark model is correct , it implies that important mass and spin - dependent interactions are present in the one - loop level one - gluon - exchange quark interaction .
it is possible that current lattice computations are not sufficiently sensitive to the ultraviolet behaviour of qcd to capture this physics .
the problem is exacerbated by the nearby , and presumably strongly coupled , @xmath13 continuum ; which requires simulations sensitive to the infrared behaviour of qcd .
thus heavy - light mesons probe a range of qcd scales and make an ideal laboratory for improving our understanding of the strong interaction . for recent reviews of open charm states see e. s. swanson , phys .
rept . * 429 * , 243 ( 2006 ) [ arxiv : hep - ph/0601110 ] ; p. colangelo , f. de fazio and r. ferrandes , mod .
a * 19 * , 2083 ( 2004 ) [ arxiv : hep - ph/0407137 ] .
a. drutskoy _
[ belle collaboration ] , phys .
lett . * 94 * , 061802 ( 2004 ) ; k. abe _ et al .
_ , `` improved measurements of anti - b0 @xmath37 d / sj+ k- decays '' , arxiv : hep - ex/0507064 . r. cahn and j.d .
jackson , phys .
* d68 * , 037502 ( 2003 ) .
t. barnes , f. e. close and h. j. lipkin , phys .
d * 68 * , 054006 ( 2003 ) [ arxiv : hep - ph/0305025 ] .
see also a. p. szczepaniak , phys .
b * 567 * , 23 ( 2003 ) [ arxiv : hep - ph/0305060 ] .
e. van beveren and g. rupp , phys .
lett . * 91 * , 012003 ( 2003 ) [ arxiv : hep - ph/0305035 ] ; k. terasaki , phys . rev .
d * 68 * , 011501 ( 2003 ) [ arxiv : hep - ph/0305213 ] ; t. e. browder , s. pakvasa , and a. a. petrov , phys . lett .
b * 578 * , 365 ( 2004 ) [ arxiv : hep - ph/0307054 ] ; l. maiani , f. piccinini , a. d. polosa and v. riquer , phys .
lett . * 93 * , 212002 ( 2004 ) [ arxiv : hep - ph/0407017 ] ; y. q. chen and x. q. li , phys .
* 93 * , 232001 ( 2004 ) [ arxiv : hep - ph/0407062 ] ; d. s. hwang and d. w. kim , phys .
b * 601 * , 137 ( 2004 ) [ arxiv : hep - ph/0408154 ] ; l. maiani , f. piccinini , a. d. polosa and v. riquer , phys .
d * 71 * , 014028 ( 2005 ) ; u. dmitrasinovic , phys .
lett . * 94 * , 162002 ( 2005 ) .
s. narison , phys .
b * 605 * , 319 ( 2005 ) ; phys .
b * 210 * , 238 ( 1988 ) .
colangelo and f. de fazio , phys .
b * 570 * , 180 ( 2003 ) [ arxiv : hep - ph/0305140 ] .
f. k. guo , p. n. shen , h. c. chiang and r. g. ping , arxiv : hep - ph/0603072 .
k. abe _ et al . _ [ belle collaboration ] , phys .
d69 , 112002 ( 2004 ) ; s. n. gupta and s. f. radford , phys .
d * 24 * , 2309 ( 1981 ) .
j. t. pantaleone , s. h. h. tye and y. j. ng , phys .
d * 33 * , 777 ( 1986 ) .
s. n. gupta , s. f. radford and w. w. repko , phys .
d * 28 * , 1716 ( 1983 ) .
t. barnes , s. godfrey and e. s. swanson , phys .
d * 72 * , 054026 ( 2005 ) [ arxiv : hep - ph/0505002 ] .
s. godfrey , phys .
b * 568 * , 254 ( 2003 ) [ arxiv : hep - ph/0305122 ] .
f. e. close and e. s. swanson , phys .
d * 72 * , 094004 ( 2005 ) [ arxiv : hep - ph/0505206 ] .
s. godfrey and r. kokoski , phys .
d * 43 * , 1679 ( 1991 ) .
|
.3 truecm it is shown that quark mass dependence induced by one loop corrections to the breit - fermi spin - dependent one gluon exchange potential permits an accurate determination of heavy - light meson masses .
thus the @xmath0 is a canonical @xmath1 meson in this scenario .
the multiplet splitting relationship of chiral doublet models , @xmath2 , holds to good accuracy in the @xmath3 and @xmath4 systems , but is accidental .
radiative transitions and bottom flavoured meson masses are discussed .
= 0.3 true in
|
consider the case where @xmath33 parties perform a nonlocality experiment .
the input and outcome of the @xmath65-th party will be denoted by @xmath66 and @xmath67 respectively .
we consider the case where non - detection events can occur , they will be denoted by @xmath67 taking the value @xmath8 .
we will denote by @xmath68 the outcome of party @xmath65 after postselecting on having a detection . as discussed in the main text
, we make the assumption of limited detection locality and show that these correlations form a polytope .
the theorem stated here is more general than needed for the main text , where we only consider the case of 2 parties .
+ _ definitions : _ let @xmath69 , @xmath70 , @xmath71 be sets of random variables with alphabets @xmath72 , @xmath73 and @xmath74 respectively , @xmath75 . in the following , the corresponding lower case letters will denote values in the respective alphabet
. we will denote probability distributions over a random variable @xmath76 by @xmath77 , the value of this distribution for a given value of @xmath76 by @xmath78 .
for ease of notation , we will often omit the random variable and just write @xmath79 .
we will denote conditional probability distributions over a random variable @xmath76 conditioned on a random variable @xmath80 by @xmath81 . in the case of continuous random variable
we denote the probability density by @xmath82 . in the following
we assume that all the probability distributions are well defined .
the set of all probability distributions over @xmath76 will be denoted by @xmath83 and of all conditional probability distributions over @xmath76 conditioned on @xmath80 by @xmath84 .
+ we define the following sets : * the sets of 1-party distributions with limited detection : @xmath85 * the set of @xmath33-party limited detection local distributions : @xmath86 * the set of @xmath33-party postselected limited detection local distributions : @xmath87 we further define the following two sets , which we will prove to be the vertices of @xmath88 and @xmath89 : @xmath90 @xmath91 with these definitions , we can now state the theorem .
it refers to polytopes , which for the purposes of this work are simply seen as a convex structure with a finite set of vertices .
equivalently they can be defined by a finite set of inequalities . +
* theorem : * for fixed @xmath33 , @xmath92 and @xmath93 , @xmath89 is a polytope whose vertices are a subset of @xmath94 . furthermore , for fixed @xmath95 , @xmath96 is also a polytope .
+ _ proof : _ to ease notation we will omit writing @xmath33 , @xmath92 and @xmath93 from now on . the first part of the theorem follows from the following two lemmas .
+ * lemma 1 : * @xmath97 is a polytope with vertices given by @xmath98 .
+ _ proof : _ due to the normalisation of probability distributions , _
i.e. _ , @xmath99 we have that @xmath100 and we can therefore work in the lowerdimensional subspace given by @xmath101 . in this subspace
, we are then left with the polytope defined by the inequalities @xmath102 .
this is the definition of a hypercube whose vertices are defined by the corresponding part of @xmath103 .
+ the lemma follows .
@xmath104 + * lemma 2 : * let @xmath105 and @xmath106 be polytopes with vertices @xmath107 and @xmath108 respectively .
let @xmath109 .
let @xmath110 .
then @xmath111 is a polytope whose vertices are a subset of @xmath112 .
+ _ proof : _ by definition , @xmath111 is convex and @xmath113 .
let @xmath114 , then by definition we have : @xmath115 in the first step we use the fact that any element of @xmath105 and @xmath106 can be written as a convex combination of their vertices and we define @xmath116 , @xmath117 and @xmath118 , @xmath119 . in the last step we defined @xmath120 , which fulfils @xmath121 and @xmath122 and also used the definition of @xmath123 .
+ this proves the lemma.@xmath104 + using these two lemmas in conjunction ( and using lemma 2 iteratively ) proves that @xmath124 is a polytope and that @xmath125 contains its vertices . to finalize the proof of the theorem we need to show that @xmath126 is a polytope as well
this can be seen directly since the set is obtained by slicing @xmath124 with the hyperplanes defined by @xmath127 . cutting a polytope with hyperplanes results in another polytope .
the final step is a simple rescaling of the entries ( equivalent to rescaling the axes ) and therefore the set remains again a polytope .
+ this proves the theorem .
it is possible to impose any desired @xmath14 at the price of adding some noise to the system .
assume that alice and bob set their detection systems , which we assume to have an efficiency @xmath128 , such that any time a non - detection event occurs , the system still gives an outcome with the probability @xmath14 .
in this way , alice and bob impose that their detection systems have a limited detection efficiency given by the chosen @xmath14 and @xmath129 and they can treat the resulting correlations by the tools presented above .
this however comes at the price of adding local noise to their correlations .
in fact , assume that alice and bob would share the nonlocal correlation @xmath130 if the detectors were perfect and there were no losses ( _ e.g. _ , given by projective measurements on a pure quantum state ) and denote by @xmath131 and @xmath132 the marginal distributions of alice and bob , respectively .
in the non - detection cases , the detection systems are set up such that they give with probability @xmath14 an outcome given by the local distributions @xmath133 and @xmath134 , respectively .
then , by postselecting on the cases where the detection systems gave an outcome , alice and bob share the correlation : @xmath135 they can then analyze this correlation using the tools of limited detection efficiency presented above .
one possibility of dealing with losses and detector inefficiencies is to assign the non - detection events to an additional outcome and treat the resulting correlations using the usual tools of nonlocality .
this corresponds exactly to the strategy we just presented with @xmath136 .
however , our approach is more general , allowing to assign only a fraction of the non - detection events to an outcome and postselecting on the rest . for a fixed detection efficiency @xmath128
, our method therefore encompasses both of the previous strategies , full postselection and full assignment to an additional outcome , and additionally allows for an arbitrary mixture of the two .
it is at this point not obvious to us that for a given experiment ( meaning for a given @xmath130 and a given @xmath128 ) , all of these strategies would yield the same result .
we leave it up for future works to analyze this question in more detail .
in this section we prove the link between limited detection local and measurement dependent local distributions .
this can be proven more generally , here we only present the 2-party version .
+ _ definitions : _ we introduce the random variables @xmath137 and @xmath138 with alphabet @xmath139 such that @xmath140 if and only if @xmath141 .
we define the set of limited detection local distributions allowing for measurement dependence : @xmath142 we also define the set of measurement dependent local correlations : @xmath143 _ theorem : _ if @xmath144 then @xmath145 _ proof : _ we have 1 .
@xmath146 2 .
@xmath147 3 .
@xmath148 4 .
@xmath149 .
let us prove a few implications : * if @xmath150 and @xmath151 are local , meaning that they fulfil condition 2 , then @xmath152 and @xmath153 are also local : @xmath154 * if @xmath150 and @xmath151 are local , then @xmath155 and @xmath156 are also local : @xmath157 * knowing less can not result in knowing more , meaning that upper and lower bounds on @xmath158 also hold for @xmath159 : assume @xmath160 , then @xmath161 where we used that @xmath162 .
the same holds for lower bounds @xmath163 . due to this ,
condition 3 implies @xmath164 using the implications above , we can show that the conditions imply bounds on @xmath165 : @xmath166 we can now prove the theorem : @xmath167 this is by definition an mdl - correlation : @xmath168
to produce the twin - photon partially polarisation entangled state , we employ the source depicted in the top - left hand side of .
a pump laser at 404 nm is sent to two cascaded type - i bbo crystals to produce photon pairs at 808 nm in a coherent superposition of two orthogonal polarisations via the process of spontaneous parametric down - conversion . to generate the twin - photon state of eq .
( 10 ) in the maintext , we set the desired weights of the coherent superposition and the phase between @xmath169 and @xmath170 by placing a half wave - plate and a tilted birefringent crystal in the path of the pump laser .
the generation process introduces some distinguishability between the two components of the state which is erased by an adapted compensation crystal .
moreover , to remove the discernibility in terms of spatial modes ( k - vectors ) , the paired photons are coupled , each , into a single mode fibre .
thanks to this configuration , we can produce the desired state with a fidelity of 99% ( see also ref .
@xcite for more details ) . to perform the measurements , alice and bob have each two polarization analysers with fixed settings , defining two bases , \{a@xmath171,a@xmath172 } and \{b@xmath171,b@xmath172 } , respectively .
these analysers are made of a half wave - plate and a wollaston prism .
each prism of the setup has an extinction ratio of 6od .
they are connected to the source via a 50/50 beam - splitter . to choose the measurement basis at both locations ,
two independent random bit sequences are utilized .
this way , only the two randomly selected outputs associated with the given basis are considered , the other two being disregarded @xcite .
this configuration therefore requires the use of 8 single photon detectors , being all silicon avalanche photodiodes .
alice employs 4 excelitas spcm - aqr-16-fc featuring @xmath17360% detection efficiency and dark count rates of @xmath173100hz , while bob has 4 qutools featuring @xmath17350% detection efficiency and dark count levels of @xmath1732000 hz .
note that the detectors configuration is chosen in such a way that it can satisfy , as much as possible , the assumption @xmath174 for all @xmath1 , @xmath2 , @xmath175 , and @xmath176 with an overall efficiency of about 5% ( fibre coupling and propagation losses included ) .
the detection arrival times of all the detection events are recorded by a time - to - digital converter ( not shown ) and the random choice of the basis is applied during the data post - processing thanks to two different quantum random number generators .
2ifxundefined [ 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 link:\doibase 10.1103/physrevlett.114.220404 [ * * , ( ) ] link:\doibase 10.1016/s0375 - 9601(99)00807 - 5 [ * * , ( ) ]
|
the demonstration and use of nonlocality , as defined by bell s theorem , rely strongly on dealing with non - detection events due to losses and detectors inefficiencies .
otherwise , the so - called detection loophole could be exploited . the only way to avoid this is to have detection efficiencies that are above a certain threshold .
we introduce the intermediate assumption of limited detection efficiency , that is , in each run of the experiment , the overall detection efficiency is lower bounded by @xmath0 .
hence , in an adversarial scenario , the adversaries have arbitrary large but not full control over the inefficiencies .
we analyse the set of possible correlations that fulfill limited detection locality ( ldl ) and show that they necessarily satisfy some linear bell - like inequalities .
we prove that quantum theory predicts the violation of one of these inequalities for all @xmath0 .
hence , nonlocality can be demonstrated with arbitrarily small limited detection efficiencies .
we validate this assumption experimentally via a twin - photon implementation in which two users are provided with one photon each out of a partially entangled pair .
we exploit on each side a passive switch followed by two measurement devices with fixed settings . assuming the switches are not fully controlled by an adversary , nor by hypothetical local variables , we reveal the nonlocality of the established correlations despite a low overall detection efficiency .
_ introduction _ when studying the discoveries in fundamental physics of the past century , one can not help but come across bell s seminal work on the nonlocal nature of quantum theory @xcite .
it implies that quantum physics can produce correlations which can not be explained by a common past with local variables propagating contiguously .
this has not only proven fascinating from a foundational point of view , but also given rise to applications in device independent quantum information processing ( diqip ) @xcite , such as quantum key distribution @xcite , randomness generation @xcite , or entanglement certification @xcite . for a semi broad - audience presentation of these concepts , see @xcite .
let us briefly recall the concept of local and nonlocal correlations .
assume that a source emits pairs of particles that travel to two distant stations , say alice and bob .
as depicted in , the two experimentalists perform one out of several possible measurements on the individual particles they receive and record the associated outcomes .
we denote alice s and bob s measurement choices by @xmath1 and @xmath2 and their recorded outcomes by @xmath3 and @xmath4 , respectively .
they can then compute the correlation @xmath5 . given the setup
, it seems natural to think that any correlations that alice and bob can observe in this way are due to particles having a common past , as they come from the same source .
we refer to this common past by @xmath6 .
correlations that can be explained by the existence of such a parameter are called _ local _ : @xmath7 bell s work shows that there are quantum correlations that can not be reproduced by such a local model , proving that quantum physics is inherently _
nonlocal_. nonlocality has since been demonstrated many times @xcite , and finds repercussion in a variety of applications @xcite . and
@xmath2 , which we depict here as the setting of an active switch , and return outputs @xmath3 and @xmath4 , respectively .
there is the possibility for non - detection events , in which case the corresponding output variable takes the value @xmath8 .
since these losses can be seen as happening inside the box , they can depend on the inputs @xmath1 and @xmath2 , respectively .
we analyze the limited detection local case of this scenario , meaning that a local hidden variable @xmath6 not only fully describes the state of the particle , but can also influence whether or not a non - detection event occurs.,scaledwidth=50.0% ] however , when demonstrating quantum nonlocality , several issues have to be addressed . here , we are specifically interested in one of them : what happens if the particles can be lost on the way to or inside the measurement stations , including the possibility that they reach the detectors but are simply not registered . in this case , we say that @xmath9 or @xmath10 .
one immediate idea is of course to carefully analyze why the particles get lost and , if the mechanisms are well understood , to simply discard these cases .
this means that alice and bob postselect on the cases for which they both register a detection event : @xmath11 and @xmath12 .
notably , this opens up the possibility that fully local correlations appear nonlocal if our understanding of the cause of the non - detection events is incorrect , a situation that we wish to avoid @xcite .
this is especially relevant in the case of active adversaries in diqip applications .
another option is to consider the non - detection events as an additional possible outcome and simply check if the resulting correlation is nonlocal . in this case
, one will never mistake a local correlation for a nonlocal one .
the drawback , however , is that even highly nonlocal distributions may now appear local . in the end ,
the only way address this issue , usually called the detection loophole , consists of not only producing highly nonlocal correlations but also having a high enough detection efficiency @xcite . if the latter is not satisfied , even a perfect state preparation and perfectly calibrated measurement apparatus do not help and one is left with an inconclusive experiment unless the detection loophole is assumed to be not exploited .
it is only recently that all the potential loopholes in a bell test have been closed in a single experiment for the first time @xcite .
however , closing the detection loophole in long - range experiments and applications remains difficult .
namely , after a transmission through @xmath13 km of the currently best optical fiber , the losses already exceed the tolerable threshold .
we therefore take a different route in this paper . here , we introduce the concept of limited detection locality .
it consists of an intermediate assumption between neglecting the detection loophole and closing it completely .
we show that this assumption , even when arbitrarily weak , allows demonstrating nonlocality by postselection with arbitrarily low overall detection efficiency .
we then connect the concepts of limited detection locality and measurement dependent locality @xcite .
finally , we discuss a dedicated twin - photon experiment that we performed , and show how quantum nonlocality can be revealed despite low detection efficiencies without having to make the fair - sampling assumption . _ limited detection locality ( ldl) _ assume that there exist a fixed @xmath14 and @xmath15 with @xmath16\subsetneqq [ 0,1]$ ] such that @xmath17 and similarly for bob .
this corresponds to the assumption that , for any input @xmath1 and any common local variable @xmath6 , there is a probability of at least @xmath14 and at most @xmath15 of having a detection .
consider for example a world in which the polarization degree of freedom of photons was as of yet undiscovered .
it is nowadays well known that the detection efficiency of most photonic detectors is indeed sensitive to the polarization degree of freedom of light .
however the detection efficiency never goes up to @xmath18 or down to @xmath19 , which corresponds to our assumption of limited detection efficiency with nontrivial @xmath14 and @xmath15 .
it is also worth noting that the assumption can be imposed at the cost of weakening the established correlations by a partial postselection .
we do not discuss this further here , but refer to @xcite for an explanation of this idea .
we refer to correlations fulfilling conditions ( [ locality ] ) and ( [ lde ] ) as _ limited detection local_. note that technically , the case of @xmath16=[0,1]$ ] can still be analyzed with our model , making it possible to retrieve the results discussed in @xcite . in an experiment
, one can additionally determine the actual observed detection efficiencies , which may of course be different for the different sets of inputs .
it is necessary that some detections occurred for all possible sets of inputs , which reads @xmath20 since @xmath21 we have that @xmath22 .
all of this holds analogously for bob s side . to ease notation
, we define @xmath23 .
we can now focus on the postselected limited detection local distributions given by @xmath24 similarly to local correlations , these postselected limited detection local correlations fulfill certain conditions .
more precisely , they form a convex polytope and therefore respect a set of linear bell - like inequalities @xcite . making the additional assumption that @xmath25 for all @xmath26 , one of these inequalities is , for example , given by @xmath27 in an experiment with given losses
, the experimentalists can check for which values of @xmath14 and @xmath15 their observed correlations violate this inequality .
they can then conclude that no limited detection local model with these parameters could have reproduced them .
interestingly , there are quantum correlations that do not fulfill this inequality independently of the observed detection efficiencies @xmath28 and for any upper bound @xmath15 as long as @xmath29 .
in fact , this is achieved by all quantum correlations violating hardy s paradox @xcite and can therefore be realised using any sufficiently pure partially entangled 2-qubit state with the right set of projective measurements
. this may be quite surprising since it is well - known that without the assumption of limited detection efficiency ( [ lde ] ) , a minimal observed detection efficiency of @xmath30 is required to demonstrate nonlocality for 2 parties using binary inputs and outputs .
however , making the arbitrarily weak additional assumption that @xmath31 allows demonstrating quantum nonlocality despite arbitrarily large losses and detection inefficiencies .
_ link to measurement dependent locality ( mdl) _ another way to counterfeit nonlocal correlations using only local resources is if the common history @xmath6 is correlated with the inputs @xmath1 and @xmath2 . if the correlation can be arbitrary , then any nonlocal correlation can be counter - fitted in this way , so limitations have to be imposed to be able to make any conclusions . together with some co - authors
, we recently studied the case of measurement dependent local correlations @xcite that are defined in the following way : @xmath32 note that if alice and bob each have @xmath33 inputs , then @xmath34 due to the normalization of probability distributions .
similarly to this paper , we showed that the set of mdl - correlations for fixed @xmath35 and @xmath36 can be analysed using bell - like inequalities .
it turns out that there exists a strong link between the concepts of limited detection locality and measurement dependent locality .
we make this connection explicit by the following theorem , whose demonstration is detailed in the supplemental material @xcite : * theorem : * assume that we have a correlation that can be produced by using a combination of postselected limited detection ( [ lde ] ) and measurement dependent ( [ mdlcond ] ) local ( [ mdlcorr ] ) resources , with parameters @xmath37 and @xmath38 , respectively .
then , this correlation can also be reproduced using only measurement dependent local resources with @xmath39 and @xmath40 .
intuitively , the link comes from the fact that an adversary that is allowed , via postselection , to discard unsatisfying inputs is in fact influencing the choice of inputs .
a consequence of this theorem is that whenever a correlation can not be reproduced by a measurement dependent local model with bounds @xmath41 and @xmath42 , then it can also not be realized using limited detection efficiencies with @xmath43 and @xmath44 , where @xmath33 is the number of inputs for each of the two parties .
this allows us to use any result derived for the mdl - scenario and apply them to ldl - correlations .
even more interestingly , we are now able to deal with the problems of losses and measurement dependence in a straightforward way since we can simply focus exclusively on measurement dependence . or @xmath18 .
the boxes are given inputs @xmath1 and @xmath2 , which decide which pairs of detectors are switched on , and return outputs @xmath3 and @xmath4 depending on which detector fires .
if the photon goes towards the inactive pair of detectors it is simply considered as lost and not registered .
we compare this scenario to the scenario where a local hidden variable @xmath6 that can influence the losses and settings try to emulate our observed correlations.,scaledwidth=50.0% ] _ experiment : _ we now discuss the twin - photon based experimental realization whose principle is depicted in .
each user s station comprises 2 measurement devices with fixed settings , _
i.e. _ , 4 single - photon detectors ( 8 in total ) . the detection events at alice s are space - like separated from the corresponding ones at bob s . in order to emulate active switches between different analyzis settings at each user s location without being limited by driving speeds
, we exploit instead a simple and elegant solution for passive implementation , referred to as zbinden switches @xcite .
first , standard 50/50 beam - splitters connect the incoming photons to the measurement devices .
then , two quantum random number generators ( qrng ) provide the inputs @xmath1 for alice and @xmath2 for bob to randomly select one pair of detectors associated with one measurement . the other two detectors of each box remain inactive .
consequently , when , in any run of the experiment , a photon hits an inactive detector , it is merely disregarded .
hence , with perfect devices , the maximum attainable detection efficiency is @xmath45 .
our source produces partially entangled photon pairs close to the state @xmath46 and the 8 detectors are set up to perform measurements close to the ideal projective measurements @xmath47 , @xmath48 , @xmath49 and @xmath50 , with @xmath51 .
further experimental details are given in @xcite as well as in the supplemental material @xcite . without any additional noise
, the resulting correlations would violate inequality ( [ ldlineq ] ) , @xmath52 and @xmath53 .
obviously , due to noise and imperfections even in the postselected case this is not fully realized .
our postselected correlations , meaning that we discard all the non - detection events , can be found in table [ probexp ] . while we did not precisely determine the overall detection efficiency in our experiment , it is worth to note that it stands clearly below the required @xmath54 efficiency that is necessary to close the detection loophole in the case of 2 parties with binary inputs and outputs @xcite .
however , assuming that the losses are the same for all input pairs , _
i.e. _ , @xmath25 , we can still conclude that we reveal quantum nonlocality for @xmath55 .
this follows from inequality ( [ ldlineq ] ) .
there are two ways to illustrate this result .
first , assume that a local hidden variable guides , on each side , the particle through the beam - splitter and influences the detection efficiency .
such a local hidden variable would need to have the power to reduce the overall detection efficiency down to below @xmath56 in order to mimic nonlocal correlations .
next , consider an adversary partially controlling the detectors efficiencies . assuming standard 50/50 beam - splitters
, we have that @xmath57 and one gets @xmath58 , _ i.e. _ , such an adversary would have to lower the detector efficiencies corresponding to settings they do not like down to below @xmath59 in order to have local correlations appear nonlocal . in our experiment , due to the beam - splitters , the losses in the fibres and the difficulties of perfectly aligning 8 detectors , we can safely consider that @xmath60 , leading to a required @xmath61 in order to reproduce the correlations using a local hidden variable model .
.measured unconditional probabilities @xmath62 , including the distribution of the inputs .
[ cols="^,^,^ , < ,
< , < , < " , ] additionally , the way our experiment was built and carried out clearly shows the link between limited detection efficiency and measurement dependence . at both alice s and
bob s stations , the respective beam - splitter sends the photon to either the detectors corresponding to input @xmath19 or to the ones corresponding to input @xmath18 .
the inputs @xmath1 and @xmath2 determine which pair of detectors we consider as switched on each side .
if the photon goes towards the detectors for input @xmath18 while we input @xmath19 , then the detection is not registered and the photon is lost .
the non - detection events thereby allow the photons to choose the input they will reply to .
the limited detection assumption then corresponds to limiting this choice , _
i.e. _ , while the action of the beam - splitter can be biased , it can not be fully biased .
the link to measurement dependence is here obvious .
we can use theorem 1 to extend our analysis to the case where we consider our random number generators providing the inputs @xmath1 and @xmath2 to be biased .
let us assume that there is some measurement dependence as defined in ( [ mdlcond ] ) , given by parameters @xmath38 , as well as some limited detection efficiency , given by parameters @xmath37 .
using an mdl inequality provided in @xcite and theorem 1 , we then conclude that our experiment shows quantum nonlocality for @xmath63 _ conclusion _ losses and detection inefficiencies have been serious obstacles in all experiments dedicated to test bell s theorem .
this is why a loop - hole free bell - test took so long to be successfully demonstrated .
they are also part of the main weak points that an adversary may attack in any task whose security relies on quantum nonlocality .
to help dealing with both of these issues from a theoretical point of view , we have introduced the additional assumption of limited detection efficiency ( [ lde ] ) .
the assumption is that the inefficiencies in the setup are only partially exploited , an idea that we consider very intuitive in itself especially in the case of a fundamental bell test , provided nature is non - malicious .
the case of an adversary is also interesting .
if one assumes an adversary with full control over alice s and/or bob s devices , then one needs to resort to full device - independent quantum information processing ( diqip ) @xcite .
however , semi - diqip is extremely appealing and timely as it allows one to lower the extremely high requirement of full diqip .
for example , considering diqip as a way to check the proper functioning of our system , _
i.e. _ , going back to the original idea of self - testing @xcite , it is very natural to assume that possible defects do not set @xmath64 .
in such a case , our approach provides much simpler self - tests .
even against an active adversary , there are cases in which it is reasonable to assume that they have only partial control on the detection efficiencies .
making the fair - sampling assumption when it is not warranted can lead to disastrously wrong conclusions @xcite .
the same holds for the ldl assumption in cases where it is not justified .
however , due to the fact the ldl assumption is strictly weaker than assuming fair - sampling , it can clearly hold and be justified more often .
in other words , if an adversary is not able to fully suppress the detection events , _
i.e. _ , @xmath29 , nonlocality can be demonstrated and exploited .
we connected the ideas of limited detection locality and measurement dependent locality .
we showed that results from studying measurement dependent local correlations can be applied to the case of limited detection locality .
moreover , it is possible to deal with detection efficiencies and lack of measurement independence at the same time , which we hope will be of use for future bell experiments .
finally , we carried out an experiment to bring these theoretical ideas to the real world .
our twin - photon based demonstration , having detection efficiencies too low to close the detection loophole @xcite , remains conclusive as long as we assume a limited detection efficiency such that @xmath55 .
it furthermore reveals very clearly the link between measurement dependence and limited detection .
we believe that the methods depicted here will be of use to any frontier experiments or applications that do not manage to close the detection loophole completely .
_ note added _ during the review process , we became aware , in addition to ref .
@xcite , of two more experimental demonstrations related to loophole - free bell tests @xcite .
_ acknowledgments _ we acknowledge alex may who wrote down an early version of a proof similar to the one given in appendix 2 showing a connection between limited detection and measurement dependence , as well as roger colbeck , who mentioned the idea of limited detection in private discussions .
we also thank valerio scarani for discussions and laurent labont for his help in the data acquisition process .
we further acknowledge financial support from the universit nice sophia antipolis , the european project chist - era diqip , the swiss nccr - qsit , the fondation ixcore pour la recherche , as well as the fondation simone & cino del duca ( institut de france ) .
24ifxundefined [ 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 * * , ( ) link:\doibase 10.1103/physrevlett.67.661 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.010503 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.120405 [ * * , ( ) ] _
_ , @noop ph.d .
thesis , ( ) , link:\doibase 10.1038/nature09008 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.250404 [ * * , ( ) ] link:\doibase 10.1038/nphys2705 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1103/physreva.83.032123 [ * * , ( ) ] http://stacks.iop.org/1367-2630/13/i=6/a=063031 [ * * , ( ) ] link:\doibase 10.1038/nature15759 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.190402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.220404 [ * * , ( ) ] @noop `` '' link:\doibase 10.1103/physrevlett.71.1665 [ * * , ( ) ] link:\doibase 10.1016/s0375 - 9601(99)00807 - 5 [ * * , ( ) ] http://stacks.iop.org/1367-2630/11/i=4/a=045021 [ * * , ( ) ] @noop * * ( ) link:\doibase 10.1103/physrevlett.107.170404 [ * * , ( ) ] http://stacks.iop.org/1367-2630/15/i=8/a=083047 [ * * , ( ) ] @noop ( ) @noop ( )
|
plenty of tools in astrophysics are developed using system programming languages such as fortran , c or c++ .
these languages are known to provide high performance and fast executions but they rely heavily on the developer for concurrency and memory control , which may lead to common errors as shown in fig.[fig1 ] : a ) access to invalid memory regions , b ) dangling pointers and attempts to free already freed memory , c ) memory leaks and , d ) race conditions .
this can produce random behaviors and affect the scientific interpretation of the results .
the recently created language rust prevents such problems and fields like bioinformatics @xcite have already started to take advantage of it .
astroinformatics can benefit from it too .
we first discuss the general principles behind this new language and what makes it attractive when compared to more traditional languages such as c or fortran .
we then show that this language can reach the same performance as a fortran n - body simulator , mercury - t @xcite , designed for the study of the tidal evolution of multi - planet systems .
mozilla research , motivated by the development of a new web browser engine @xcite , released in 2015 the first stable version of a new open source programming language named rust .
it uses patterns coming from functional programming languages @xcite and it is designed not only for performance and concurrency , but also for safety .
rust introduces concepts like ownership , borrowing and variable lifetime , which : 1 . facilitates the automatic control of the lifetime of objects during compilation time .
there is no need for manually freeing resources or for an automated garbage collector like in java or go ; 2 . prevents the access to invalid memory regions ; 3 .
enforces thread - safety ( race conditions can not occur ) .
these zero - cost abstraction features make rust very attractive for the scientific community with high performance needs . in rust , variables are non - mutable by default ( unless the mutable keyword is used ) and they are bound to their content ( i.e , they own it or they have ownership of it ) .
when you assign one variable to another ( case a in fig.[fig2 ] ) , you are not copying the content but transferring the ownership , so that the previous variable does not have any content ( like when we give a book to a friend , we stop having access to it ) .
this transfer takes also place when we call functions ( case b in fig.[fig2 ] ) , and it is important to note that rust will free the bound resource when the variable binding goes out of scope ( at the end of the function call for case b ) .
hence , we do not have to worry about freeing memory and the compiler will validate for us that we are not accessing a memory region that has already been freed ( errors are caught at compilation time , before execution time ) . additionally , apart from transferring ownership
, we can borrow the content of a variable ( case c in fig.[fig2 ] ) . in this case ,
two variables have the same content but none of them can be modified , thus protecting us from race conditions .
alternatively , we can borrow in a more traditional way ( like when we borrow a book from a friend , he is expecting to get it back when we stop using it ) like in case d in fig.[fig2 ] , where the function borrows the content of a variable , operates with it ( in this case , it could modify its content ) and returns it to the original owner ( not destroying it as shown in case b ) . exceptionally , all these rules can be violated if we make use of unsafe blocks , which is strongly discouraged but necessary in certain situation ( e.g. , dealing with external libraries written in fortran , c or c++ ) .
if present , unsafe blocks allow us to clearly identify parts of the code which should be carefully audited , keeping it isolated and not making the whole program unsafe by default like in fortran , c or c++ .
we explored the advantages and drawbacks of rust for astrophysics by re - implementing the fundamental parts of mercury - t @xcite , a fortran code that simulates the dynamical and tidal evolution of multi - planet systems .
we developed a simple n - body dynamical simulator ( without tidal effects ) based on a leapfrog integrator in rust , fortran , c and go ( which provide a garbage collector for memory management ) .
the software design and implementation does not include any language - specific optimization that a developer with basic knowledge would not do .
we compiled the four implementations with an optimization level 3 ( rustc / gfortran / gcc compiler ) and the standard compilation arguments in the case of go .
we selected the best execution time out of five for an integration of 1 million years and the results are in table [ table ] .
for this particular problem , rust is as efficient as fortran , and both surpass c and go implementations [ cols="^,^,^,^",options="header " , ] [ table ] time might be improved if language - specific optimizations were implemented .
based on mercury - t , we implemented the additional acceleration produced by tidal forces between the star and its planets into our rust and fortran leapfrog integrators . to test the codes , we ran a simulation of 100 million years with the same initial conditions as the case 3 described in the mercury - t article @xcite , hence a single planet with a rotation period of 24 hours , orbiting a brown dwarf ( 0.08 @xmath0 ) at 0.018 au with an eccentricity of 0.1 .
the results are shown in fig .
[ fig3 ] , the rust and fortran code are practically identical and they reproduce a similar behavior to what is shown in the mercury - t article . nevertheless , leapfrog is a very simple integrator and not very accurate .
this can be seen in the eccentricity evolution , which is slightly different from the mercury - t article and appears noisy .
as an additional exercise , we implemented the whfast integrator @xcite in rust ( black line in fig .
[ fig3 ] ) .
this better integrator leads to a better agreement with mercury - t thus demonstrating that a high level of accuracy can also be achieved with rust .
we have shown the reliability of rust as a programming language as opposed to fortran , c or even go .
rust allows the user to avoid common mistakes such as the access to invalid memory regions and race conditions .
we have also shown that it is a competitive language in terms of speed and accuracy .
the main challenge we experienced was the initial learning curve , it was necessary to really understand and get used to the ownership and borrowing concepts .
once the paradigm shift is done , the benefits are immediate .
we therefore encourage the community to consider rust as a language that will help us produce good quality , memory safe , concurrent and high - performance scientific code .
|
the astrophysics community uses different tools for computational tasks such as complex systems simulations , radiative transfer calculations or big data .
programming languages like fortran , c or c++ are commonly present in these tools and , generally , the language choice was made based on the need for performance . however , this comes at a cost : safety . for instance , a common source of error is the access to invalid memory regions , which produces random execution behaviors and affects the scientific interpretation of the results . in 2015 , mozilla research released the first stable version of a new programming language named rust .
many features make this new language attractive for the scientific community , it is open source and it guarantees memory safety while offering zero - cost abstraction .
we explore the advantages and drawbacks of rust for astrophysics by re - implementing the fundamental parts of mercury - t , a fortran code that simulates the dynamical and tidal evolution of multi - planet systems .
|
let @xmath0 be a domain in the complex plane containing 0 .
we define the _ harmonic measure distribution function _ @xmath1 by the formula @xmath9 where @xmath10 denotes the harmonic measure of the set @xmath11 from the basepoint 0 in @xmath0 . for @xmath12 ,
the number @xmath1 gives the probability that a brownian particle released at 0 first exits @xmath0 within a distance @xmath2 of @xmath13 .
two questions regarding such functions have been investigated in a number of papers @xcite .
first , what functions can be constructed as @xmath3 for some @xmath0 ?
second , what can be determined about @xmath0 from @xmath3 ?
these questions originally arose from brannan and hayman s 1989 paper @xcite , which described the current state of some problems in complex analysis and listed some new ones .
specifically , these questions arose from problem 6.116 , proposed by stephenson .
problem 6.116 poses the questions listed above for the related function @xmath14 , defined to be the harmonic measure of @xmath15 in the connected component of @xmath16 containing 0 .
the following informally stated theorem ( restated as below ) summarizes the results in this paper .
this existence result gives sufficient conditions on a function @xmath4 for @xmath4 to arise as the harmonic measure distribution function of a bounded simply connected domain .
[ thm : intro ] let @xmath4 be a nondecreasing , right - continuous function that is zero on @xmath17 and @xmath18 on @xmath19 , for some numbers @xmath20 .
suppose that the slopes of the secant lines to the graph of @xmath21 for @xmath22 are bounded from below by a positive number .
if @xmath21 is discontinuous at @xmath23 , and if @xmath24 is small enough , then there exists a domain @xmath0 that is bounded and simply connected , and is symmetric in the sense of , such that @xmath25 .
furthermore , up to sets of harmonic capacity zero , this domain @xmath0 is unique among bounded symmetric domains .
the precise condition on @xmath24 is given in the statement of .
the requirement that @xmath4 increase from @xmath13 to @xmath18 is necessary for @xmath4 to arise as the harmonic measure distribution function of _ any _ bounded domain .
a set of less restrictive but more technical sufficient conditions is given in .
our proof of relies on an earlier existence result . in (
* theorem 2 ) , snipes and ward proved that if @xmath4 is a right - continuous step function with finitely many jumps , and increases from 0 to 1 , then @xmath4 arises as the harmonic measure distribution function of some multiply connected domain .
they were able to describe this multiply connected domain fairly precisely .
much of the work in the current paper is done to pass from these multiply connected domains to simply connected domains .
we review some results concerning harmonic measure distribution functions of simply connected domains . in @xcite and @xcite
, it was shown that sequences of simply connected domains that converge in certain senses have convergent harmonic measure distribution functions .
we will use these results ; see .
in @xcite , @xcite , and @xcite , walden , ward , betsakos and solynin investigated the behavior of the harmonic measure distribution function @xmath1 of a simply connected domain for @xmath2 near @xmath26 . if @xmath27 and @xmath0 is simply connected , it is easy to see that @xmath28 for all @xmath29 , and @xmath30 for all @xmath31 .
these papers established that for every number @xmath32 with @xmath33 , there is some simply connected domain @xmath0 such that @xmath34 exists and is positive .
it was also shown that not all harmonic measure distribution functions of simply connected domains have this property ( or even the weaker property @xmath35 for some numbers @xmath36 ) .
these papers show that for some properties ( asymptotic behavior of @xmath4 at @xmath37 ) there is a domain whose harmonic measure distribution function has those properties .
they do not specify the whole function @xmath4 ; that is , they do not provide sufficient conditions for a function to arise as the harmonic measure distribution function of some domain .
in this paper , we will provide three progressively more restrictive , but easier to check , conditions on @xmath4 , such that if @xmath4 meets any of those conditions then @xmath4 must be the harmonic measure distribution function of a simply connected domain . informally states the most transparent , and thus most restrictive , of these sufficient conditions .
[ exa : jump ] the function @xmath38 graphed in satisfies the sufficient conditions of ( see ) ; thus , we know that there exists a simply connected domain @xmath0 such that @xmath25 . before the current paper , it was not known whether this function , or indeed any function with a nonhorizontal linear segment , was the harmonic measure distribution function of any domain . ,
defined by , that arises as the @xmath39-function of some simply connected domain.,title="fig : " ] the organization of this paper is as follows . in , we define our terms and give some background information about harmonic measure and harmonic measure distribution functions . in sections [ sec :
convergence ] , [ sec : sufficient ] and [ sec : functions ] , we describe our sufficient conditions .
specifically , in , we show that if @xmath40 , where @xmath41 is a sequence of bounded , simply connected domains with uniformly locally connected complements , then @xmath25 for some simply connected domain @xmath0 . in
we provide a candidate sequence @xmath41 .
since monotonic functions can be approximated by step functions , by @xcite ( mentioned above ) any function @xmath4 that increases from zero to one is the limit of a sequence of harmonic measure distribution functions of multiply connected domains .
we provide a condition that lets us approximate these domains with well - behaved simply connected domains , yielding a sufficient condition on @xmath4 that may be checked by examining one sequence of domains rather than all sequences of simply connected domains . in , we show that functions that increase fast enough , in some senses that are easy to check , satisfy the sufficient condition of and thus are harmonic measure distribution functions of simply connected domains . in particular , and the existence part of are proven in . in
we consider uniqueness . from
* remark 1 ) , a step function arises as the harmonic measure distribution function of many different domains .
however , if the domain is assumed to be bounded and satisfy a symmetry condition , then the domain is unique
. the symmetry condition can be generalized to arbitrary domains .
we show that for bounded simply connected domains , the symmetry condition again implies uniqueness .
we also show that the domains produced in sections [ sec : sufficient ] and [ sec : functions ] satisfy this symmetry condition .
thus , if a function @xmath4 satisfies the necessary conditions of section [ sec : sufficient ] or [ sec : functions ] , then @xmath4 is the harmonic measure distribution function of a unique bounded simply connected symmetric domain .
finally , in , we prove some technical lemmas , involving the harmonic measure of circle domains and related domains , that we use in sections [ sec : sufficient ] and [ sec : functions ] .
it should be emphasized that this paper provides only _ sufficient _ conditions for a function @xmath4 to be realizable as the harmonic measure distribution function of a simply connected domain @xmath0 .
these conditions are not necessary .
let @xmath42 is simply connected@xmath43 be the set of all harmonic measure distribution functions of simply connected domains , and let @xmath44 be the set of all functions @xmath4 that satisfy the sufficient conditions of section @xmath45
. then @xmath46 .
some of these inclusions are proper .
it is easy to exhibit harmonic measure distribution functions that do not satisfy the conditions of .
in particular , no continuous function satisfies these conditions , and there are many well - behaved domains with continuous @xmath39-functions . for example , any disk containing but not centered at the origin has a continuous @xmath39-function .
by considering a constant sequence of domains , we can in fact exhibit functions @xmath25 that satisfy the sufficient conditions of but not . in other words , @xmath47 .
it is not known whether such functions satisfy the conditions of .
also , @xmath48 .
the condition in is sufficient for a function @xmath4 to arise as the harmonic measure distribution function of a bounded simply connected symmetric domain whose complement is locally connected . by uniqueness , if @xmath25 for some bounded simply connected symmetric domain @xmath0 whose complement is _ not _ locally connected , then @xmath4 can not satisfy the conditions of . an early version of this paper @xcite appeared as the first author s undergraduate senior thesis , advised by the second author , while both authors were at harvey mudd college .
some of our later work on this paper was done while the first author was at the university of chicago and at purdue university .
the first author would like to thank henry krieger for acting as second reader of @xcite and for useful advice regarding functional analysis .
in this section , we will provide definitions for many terms and concepts used throughout this paper .
let @xmath0 be a connected domain , let @xmath49 be a point in @xmath0 , and let @xmath11 be a measurable subset of @xmath50 .
let @xmath51 denote the usual harmonic measure of @xmath11 in @xmath0 with basepoint @xmath49 .
the harmonic measure can be calculated by solving the dirichlet problem : @xmath52 from @xcite , @xmath51 is also the probability that a brownian particle , released from @xmath49 , first exits @xmath0 through @xmath11 . notice that if @xmath53 is a conformal mapping that extends continuously to @xmath54 , and if @xmath55 and @xmath56 , then @xmath57 in particular , if @xmath58 is one - to - one on @xmath54 then @xmath59 for every @xmath60 . also , if @xmath61 , @xmath62 , and @xmath63 , then @xmath64 this property is referred to as the _ monotonicity in the domain property _ of harmonic measure . finally ,
if @xmath61 and @xmath65 , then @xmath66 we may define the harmonic measure distribution function , or @xmath39-function , of @xmath0 by @xmath67 recall that @xmath68 where @xmath69 is the connected component of @xmath16 that contains @xmath13 . then by monotonicity in the domain , @xmath70 .
[ rmk : hnesc ] all harmonic measure distribution functions have a number of common properties . for any domain @xmath0 , @xmath3 is right - continuous , nondecreasing , and if @xmath27 , then @xmath28 for all @xmath29 .
furthermore , if @xmath50 is bounded , then there is some @xmath71 such that @xmath72 for all @xmath73 .
these observations give a set of necessary conditions for a function @xmath4 to arise as the harmonic measure distribution function of any domain .
here is another necessary condition stemming from the relation between @xmath3 and @xmath74 .
this necessary condition applies only to the @xmath39-functions of simply connected domains . as noted in @xcite
, it follows from beurling s solution to milloux s problem @xcite that , if @xmath27 and @xmath0 is simply connected , then @xmath75 , where @xmath76 .
the domain @xmath77 is simple enough that @xmath78 may be computed directly via the riemann map ; we find that @xmath79 for @xmath2 large we have the simpler inequality @xmath80 for some constant @xmath81 ; this inequality is known as beurling s lemma . as noted in @xcite
, this inequality shows that the condition @xmath25 for some simply connected domain is more restrictive than the condition @xmath25 for some arbitrary domain .
( see for specific counterexamples . )
we will not discuss necessary conditions further in this paper .
[ dfn : circle ] we say that @xmath82 is a _ circle domain _ if there exist numbers @xmath83 such that @xmath84 where @xmath85 is a connected closed proper subset of @xmath86 that is symmetric about the real axis and contains a positive real .
we call the arcs @xmath85 the _ boundary arcs _ of @xmath87 .
we may regard the boundary circle @xmath88 as the @xmath89th boundary arc @xmath90 ; then @xmath90 is a full circle and @xmath85 , @xmath91 , is not .
we say that @xmath92 is a _ blocked circle domain _
if @xmath0 is simply connected , symmetric about the real axis , @xmath93 for some circle domain @xmath87 , and if @xmath94 consists of a number of _ gates _ , that is , radial segments whose endpoints are at distances @xmath95 and @xmath96 from 0 .
see . and
a blocked circle domain @xmath0.,title="fig : " ] circle domains have historically been of interest because any step function satisfying the necessary conditions of arises as the @xmath39-function of a circle domain .
see ( * ? ? ? * theorem 2 ) , quoted in this paper as . in the current paper , we will use sequences of blocked circle domains to construct domains with prescribed @xmath39-functions . if @xmath97 is a connected arc of a circle centered at the origin , for notational convenience we refer to the angle at the origin , subtended by @xmath97 , as the _ arclength _ of @xmath97 .
[ dfn : symmetric ] throughout this paper , when we say that a domain @xmath0 is _ symmetric _ , we mean that @xmath98 , that @xmath0 is symmetric about the real axis , and that for every @xmath12 , @xmath99 is either empty or a connected closed set that contains a point on the positive real axis .
this notion of symmetry will be needed for the uniqueness results in .
we observe that a connected subset of @xmath100 is necessarily path - connected .
all circle domains and blocked circle domains are symmetric in this sense .
let @xmath0 be symmetric , and let @xmath2 , @xmath101 be real numbers with @xmath12 .
we remark that if @xmath102 then @xmath103 , and if @xmath104 then @xmath105 . if @xmath0 is also bounded and connected , let @xmath37 and @xmath71 be the largest and smallest numbers , respectively , such that @xmath106 ; then @xmath107 and @xmath108 .
if in addition @xmath0 or @xmath109 is simply connected , then @xmath110=\emptyset$ ] .
[ dfn : ulconnected ] a path - connected closed set @xmath97 is _ locally connected _ if , for every @xmath111 , there exists a @xmath112 such that any two points @xmath113 , @xmath114 with @xmath115 can be joined by a continuum @xmath116 of diameter at most @xmath117 . a sequence of closed sets @xmath118 is _
uniformly locally connected _ if , for every @xmath111 , there exists a @xmath112 independent of @xmath89 such that any two points @xmath113 , @xmath119 with @xmath115 can be joined by a continuum @xmath120 of diameter at most @xmath117 .
[ dfn : equicontinuous ] a sequence of maps @xmath121 on a domain @xmath0 is _ equicontinuous _ if , for all @xmath111 , there exists a @xmath112 , depending only on @xmath117 , such that if @xmath122 , @xmath123 and @xmath124 , then @xmath125 for all @xmath89 . in particular , if @xmath126 is part of a uniformly locally connected sequence then @xmath126 is locally connected , and if @xmath127 is part of an equicontinuous sequence then @xmath127 is uniformly continuous .
the important result of this section is . this theorem states that if @xmath4 is the limit of a sequence of harmonic measure distribution functions of well - behaved domains , then @xmath4 is itself the harmonic measure distribution function of a well - behaved domain .
this result may be viewed as a sufficient condition on @xmath4 that forces @xmath25 for some simply connected domain @xmath0 .
[ thm : snipesbarton ] let @xmath4 be a right - continuous , nondecreasing function .
suppose that there exists a sequence of domains @xmath41 such that [ item : snipesbarton : first ] @xmath126 is simply connected , there exist numbers @xmath71 , @xmath128 such that for all @xmath89 , @xmath129 , [ item : snipesbarton : localconn ] @xmath130 is uniformly locally connected , and [ item : snipesbarton : conv ] @xmath131 pointwise at all points of continuity of @xmath4 .
then there exists a domain @xmath0 that is bounded and simply connected , contains @xmath13 , and has locally connected complement , such that @xmath25 .
furthermore , there is a subsequence of @xmath41 whose riemann maps converge uniformly to the riemann map of @xmath0
. follows immediately from the following two theorems .
( will be proven below ; was proven in ( * ? ? ?
* theorem 1 ) . )
let @xmath132 denote the unit disk @xmath133 .
[ thm : sccdconv ] suppose that @xmath41 is a sequence of domains in the complex plane such that the following conditions hold .
@xmath126 is simply connected , there exist numbers @xmath71 , @xmath128 such that for all @xmath89 , @xmath129 , and @xmath130 is uniformly locally connected
. then the riemann maps @xmath134 are equicontinuous , and so they have continuous extensions to the closed unit disk @xmath135 . if @xmath136 , @xmath137 for all @xmath89 , then the sequence @xmath121 contains a subsequence @xmath138 that converges uniformly on @xmath135 to some uniformly continuous map @xmath58 .
furthermore , the limit @xmath58 is itself the riemann map of some domain @xmath0 with locally connected complement .
[ thm : snipes : conv ] let @xmath0 and @xmath126 , @xmath139 , be simply connected domains containing the point @xmath140 , with @xmath141 and @xmath142 , and with harmonic measure distribution functions @xmath39 and @xmath143 , respectively .
suppose that the normalized riemann mappings @xmath144 , @xmath145 , @xmath146 and @xmath134 , @xmath136 , @xmath137 have continuous extensions to the closed unit disk @xmath135 .
if @xmath147 pointwise on the boundary @xmath148 , then @xmath149 pointwise at all points of continuity of @xmath39 . because we wish to use , is an appropriate convergence result to use to prove .
however , is not the only known result of its type . as noted in @xcite , if the domains @xmath126 are simply connected and @xmath150 in the sense of carathodory , then @xmath151 for almost all @xmath2 . to prove
, we will need the following three theorems ( taken from ( * ? ? ?
* theorem 2.1 ) , ( * ? ? ?
* propositon 2.3 ) and @xcite , respectively ) .
[ thm : pommer : localconn ] let @xmath58 map @xmath132 conformally onto the bounded domain @xmath0
. then @xmath58 has a continuous extension to @xmath135 if and only if @xmath152 is locally connected .
[ thm : pommer : equic ] let @xmath127 map @xmath132 conformally onto @xmath126 with @xmath136 .
suppose that @xmath153 for all @xmath89 .
if @xmath130 is uniformly locally connected , then @xmath121 is equicontinuous on @xmath135 .
[ thm : arzela ] suppose that @xmath121 is a sequence of pointwise bounded , equicontinuous complex functions on @xmath135 .
then there is a subsequence of the @xmath127 that converges uniformly on @xmath135 .
theorems [ thm : pommer : equic ] and [ thm : arzela ] imply that a subsequence of @xmath121 converges uniformly on @xmath135 to some map @xmath58 . because @xmath121 is equicontinuous , @xmath58 must be uniformly continuous .
we need only show that @xmath58 is the riemann map of some domain @xmath0 ; will then imply that @xmath152 is locally connected . to show that @xmath58 is a riemann map , we must show that @xmath58 is analytic and injective and that @xmath154 is open and simply connected . by @xcite ,
if @xmath155 uniformly in some domain and @xmath127 is analytic , then @xmath58 is also analytic and @xmath156 uniformly on compact subsets .
( this fact may be easily seen from the cauchy integral formulas . ) by @xcite , a nonconstant analytic function maps open sets onto open sets .
also , if @xmath58 is one - to - one , the preimage of any loop in @xmath157 must be a loop in @xmath132 ; since @xmath132 is simply connected , loops are contractible , and so @xmath154 must be simply connected as well .
so we need only show that @xmath58 is one - to - one . because @xmath158 and @xmath147 uniformly , @xmath58 can not be a constant
. thus , for any fixed @xmath159 , the zeros of @xmath160 must be isolated .
let @xmath161 be any jordan curve in @xmath132 with @xmath49 in its interior , and with @xmath162 on @xmath161 for some @xmath117 .
consider @xmath163 this quantity is equal to the number of zeros of @xmath160 in the interior of @xmath161 .
but @xmath164 for all @xmath89 , and so since @xmath147 and @xmath165 uniformly on @xmath161 , we must have that @xmath166
so @xmath58 is one - to - one .
in this section , we will provide a more restrictive , but easier to check , condition on a function @xmath4 that will force it to arise as the harmonic measure distribution function of some domain .
we begin by quoting a known existence result .
[ thm : snipes : circle ] let @xmath21 be a right - continuous step function , increasing from @xmath13 to @xmath18 , with its discontinuities at @xmath167 , @xmath168 , where @xmath83 .
then there exists a circle domain @xmath87 with @xmath89 arcs whose harmonic measure distribution function @xmath169 is equal to @xmath21 .
the radii of the @xmath89 arcs and of the boundary circle in @xmath87 are given by @xmath167 , @xmath170 , @xmath171 and by @xmath172 respectively .
monotonic functions can be approximated by step functions .
therefore , if @xmath4 is a nondecreasing function , then there is some sequence of circle domains @xmath173 such that @xmath174 .
this sequence does not satisfy the conditions of .
however , we will find conditions ( ) on the sequence @xmath173 such that we may construct a sequence @xmath41 of blocked circle domains that does .
informally , the conditions are that the circle domains @xmath5 do not have too many short arcs , and that the function @xmath4 increases fast enough . provides a sufficient condition for to hold , which may be checked by examining @xmath173 rather than all sequences of simply connected domains @xmath41 that satisfy @xmath131 .
our proof of relies on lemmas [ lem : deriv ] , [ lem : radial ] and [ lem : unifcctd ] .
we defer their precise statements and proofs to .
we begin by fixing some terminology .
suppose that we have a sequence @xmath173 of circle domains , and a sequence @xmath41 of blocked circle domains such that @xmath175 and @xmath176 is a union of gates . in this section and in we will use the following symbols to describe such sequences of domains .
( in we will often discuss single domains rather than sequences of domains ; when doing so we will use the same notation without the @xmath89 subscript . ) * @xmath5 denotes the @xmath89th circle domain .
* @xmath126 denotes the @xmath89th blocked circle domain . *
@xmath177 denotes a boundary arc of @xmath5 ; the innermost arc is @xmath178 , the next innermost arc is @xmath179 , and so on .
* @xmath180 denotes the radius of @xmath177 . *
@xmath181 denotes half the arclength of @xmath177 . *
@xmath182 denotes the angle ( from the real axis ) of a gate of @xmath126 lying between @xmath177 and @xmath183 .
* @xmath184 denotes the inset angle of that gate .
* @xmath185 measures the depth of the shortest arc between @xmath186 and @xmath177 . * @xmath187 measures the depth of the deepest gate between @xmath186 and @xmath177 .
we emphasize that the numbering of boundary arcs and gates is to begin at zero rather than one .
see for an illustration of @xmath180 , @xmath181 and @xmath182 , @xmath184 .
we define @xmath184 by the equation @xmath188 we require that @xmath182 and @xmath184 both be nonnegative ; this implies that @xmath189 and @xmath190 . , @xmath181 , and @xmath182 , @xmath184 , used to describe circle domains @xmath5 and blocked circle domains @xmath126 . here
@xmath191 and @xmath192.,title="fig : " ] we now clarify and make precise the definition of @xmath185 .
let @xmath186 and @xmath177 be any two boundary arcs of @xmath5 . consider the arcs @xmath193 , @xmath194 , lying between @xmath186 and @xmath177 .
in many of the theorems to come , we will need to consider the shortest of these arcs
. the important value will usually not be the arclength of the shortest arc @xmath193 , but its depth in the channel outlined by @xmath186 and @xmath177 .
we define @xmath195 this number is zero if none of the inner arcs are shorter than the outer arcs , and otherwise is half the difference in arclength between the shorter of the outer arcs and the shortest of the inner arcs .
( we divide the arc length difference by two because there are two ends to the channel and the shortest arc is inset in both of them . )
similarly , @xmath196 measures the depth of the deepest gate in @xmath126 between @xmath186 and @xmath177 .
see for an illustration of @xmath185 and @xmath187 .
we observe that we can bound @xmath187 by @xmath185 and @xmath197 for @xmath198 . by , if @xmath198 then @xmath199 by , @xmath200 combining these formulas we have that @xmath201 by , the right - hand side is equal to @xmath202 , and so @xmath203 and @xmath185 used to describe circle domains @xmath5 and blocked circle domains @xmath126 . here
@xmath204 , @xmath205 , and @xmath206.,title="fig : " ] we now consider functions .
suppose that @xmath4 is a _ candidate for a harmonic measure distribution function _
, meaning that @xmath4 is a right - continuous function defined on @xmath207 that is 0 on @xmath17 , is nondecreasing on @xmath208 $ ] , is 1 on @xmath19 , and satisfies @xmath209 on @xmath210 , for some numbers @xmath211 .
define the _ minimal secant slope _
@xmath212 of @xmath4 by @xmath213 the infimum in this definition is the infimum of the slopes of secant lines to the graph of @xmath4 between @xmath37 and @xmath71 ; multiplying by @xmath214 normalizes this number so that @xmath215 .
we fix @xmath216 so that @xmath217 and @xmath218 .
define @xmath219 by @xmath220 thus , @xmath221 is a right - continuous nondecreasing step function and approximates @xmath4 from below .
let @xmath5 be a circle domain with boundary arcs at radii @xmath180 and with @xmath222 ; by such an @xmath5 exists .
then @xmath223 pointwise at all points of continuity of @xmath4 .
we remark that @xmath224 , that @xmath225 if @xmath4 is continuous at @xmath37 , and that if @xmath226 then @xmath227 for all @xmath228 .
we now show that if the domains @xmath5 do not have too many short arcs , and if the function @xmath4 increases fast enough , then @xmath4 is the harmonic measure distribution function of some bounded , simply connected domain @xmath0 .
[ thm : main ] suppose that the function @xmath4 is a candidate for a harmonic measure distribution function .
let @xmath212 be the minimal secant slope given by , and define the circle domains @xmath5 as above .
let @xmath229 be any sequence of numbers , for example @xmath230 , such that 1 .
[ item : kappasmall ] @xmath231 , and 2 .
[ item : kappabig ] @xmath232 suppose that 1 .
[ item : derivpos ] the minimal secant slope @xmath226 , and 2 .
[ item : fewshortarcs ] @xmath233 , where @xmath234 is the number of boundary arcs of @xmath5 with arclength at most @xmath235 .
then there exists a sequence @xmath41 of blocked circle domains , satisfying the conditions of , such that @xmath131 pointwise at all points of continuity of @xmath4 .
so by , @xmath4 is the harmonic measure distribution function of some bounded , simply connected domain @xmath0 . for each domain @xmath5
, we will construct a suitable blocked circle domain @xmath175 . by construction , @xmath174 pointwise at points of continuity of @xmath4 .
we want to show that if @xmath236 and @xmath226 , then we may choose @xmath126 such that @xmath41 and @xmath4 satisfy the conditions of .
the only difficult conditions are and , that is , the requirement that @xmath130 be uniformly locally connected and the requirement that @xmath131 . since @xmath174 , we may replace by a requirement that @xmath237 .
recall that @xmath238 .
the circle domains @xmath5 , and thus @xmath181 and @xmath185 , are then determined by the fact that @xmath222 . given the inset angles @xmath184 of the gates , the domains @xmath126 and the parameters @xmath182 , @xmath187 are determined . we are free to choose @xmath184 provided @xmath239 .
we set the inset angles @xmath184 of the gates in @xmath240 to be @xmath241 the number @xmath242 thus determines the size and shape of ( most of ) the ends of the channels . since @xmath243 , the channel ends must eventually become short , but since @xmath244 and the width of the channels is @xmath245 , the channel ends must become thin much faster than they become short .
we remark that if the inset angle @xmath246 , then @xmath247 .
recall that @xmath248 , where @xmath182 is the gate angle .
therefore , if @xmath246 then @xmath249 ; there is only one gate between @xmath177 and @xmath183 , and it lies along the positive real axis . furthermore , if @xmath249 then @xmath247 . since @xmath250 this implies that @xmath251 or @xmath252 ; thus , because there are at most @xmath234 arcs @xmath253 with @xmath251 , there are at most @xmath254 gates that lie along the real axis , one on each side of each short arc .
recall that we imposed two conditions on @xmath229 .
the reasons are as follows .
a necessary ( not sufficient ! ) condition for uniform local connectivity of the sequence @xmath41 is that the inset angle @xmath184 of the gates be small whenever @xmath255 is small .
( see . ) since @xmath256 , and @xmath181 need not be small , we must have that @xmath243 as @xmath257 .
however , @xmath258 is controlled by the harmonic measure of the gates ( see ) , which is small if @xmath184 is large compared with @xmath255 ( see ) ; thus , while @xmath243 , @xmath242 can not go to zero too quickly .
we make our requirements precise . in
, we will establish a bound on @xmath259 ; it is @xmath260 recall that there are at most @xmath254 gates that lie along the real axis ; thus , there are at most @xmath254 numbers @xmath45 that satisfy @xmath249 . since @xmath261 , and @xmath262 ,
the second sum is at most @xmath263 .
there are @xmath89 boundary arcs ( and so at most @xmath89 pairs of gates with @xmath264 ) .
recall that if @xmath264 then @xmath265 .
again by our choice of @xmath180 , the first sum is at most @xmath266 .
so @xmath267 by our assumptions on @xmath234 and @xmath242 , both terms go to zero , and so holds : @xmath131 at points of continuity of @xmath4 .
we turn to .
assume that the minimal secant slope @xmath212 is greater than @xmath13 .
recall the definitions of @xmath185 and @xmath187 ( the depths of the shortest arc and deepest gate , respectively , between @xmath186 and @xmath177 ; see ) .
states that @xmath268 .
since @xmath269 for all @xmath89 and @xmath270 , this implies that @xmath271 in , we will establish sufficient conditions for @xmath130 to be uniformly locally connected .
these conditions are that , for every @xmath111 , there exist positive numbers @xmath272 , @xmath273 such that * if @xmath274 and @xmath275 , then @xmath276 , and * if @xmath277 and @xmath278 then @xmath279 . informally
, these conditions say that if two arcs in a blocked circle domain are sufficiently close , then the gates between them are not too deep , and that if an arc is long enough to be almost a full circle , then it is close to the outer boundary circle . if @xmath280 , then @xmath281 .
when @xmath226 , we can use this lower bound on the harmonic measure of each arc to ensure that the conditions of hold .
we do so as follows . in
, we will show that if @xmath226 , then for each @xmath111 there exist numbers @xmath282 , @xmath283 ( depending on @xmath212 , @xmath71 , @xmath37 ) such that * if @xmath274 and @xmath284 , then @xmath276 , and * if @xmath277 and @xmath285 , then @xmath286 .
fix @xmath111 .
let @xmath287 .
since @xmath243 as @xmath257 , there is some @xmath288 such that if @xmath289 then @xmath290 .
let @xmath291 suppose @xmath292 and @xmath293 . then because @xmath294 , the depth @xmath185 of the shortest arc between @xmath85 and @xmath253 satisfies @xmath295 .
furthermore , since @xmath296 we have that @xmath297 and so @xmath298 ; thus @xmath290
. then @xmath299 .
so if @xmath226 then the conditions of hold , as desired .
in this section , we exhibit a family of functions that satisfy the conditions of , and are thus the harmonic measure distribution functions of simply connected bounded domains ( ) . loosely speaking ,
these are functions that begin with a jump and reach 1 quickly after that . our proof of relies on lemmas [ lem : fjump ] and [ lem : deriv ] ; we defer their statements and proofs to .
[ thm : fjump ] suppose that @xmath4 is right - continuous , @xmath13 on @xmath300 , strictly increasing on @xmath301 $ ] , and @xmath18 on @xmath19 , with @xmath211 and @xmath209 on @xmath210 .
let @xmath212 be the minimal secant slope of @xmath4 given by , so @xmath302 for all @xmath303 , @xmath304 with @xmath305 .
let @xmath306 . notice that @xmath307 and @xmath308 . if @xmath226 and @xmath309 , then there is a number @xmath310 depending only on @xmath212 and @xmath32 such that , if @xmath311 , then @xmath4 satisfies the conditions of .
hence there exists a simply connected bounded domain @xmath0 , arising from a sequence of circle domains , such that @xmath25 .
furthermore , @xmath312 , where @xmath313 we remark that for each fixed @xmath212 , @xmath314 , the function @xmath315 is continuous and strictly increasing on @xmath316 , and satisfies @xmath317 , @xmath318 ; thus , there is a unique ( small ) positive number @xmath319 that solves .
choose some such function @xmath4 .
let @xmath5 , @xmath177 , @xmath181 be as in . by our construction of @xmath5 , the harmonic measure of the innermost arc @xmath178 of @xmath320
is given by @xmath321 for all @xmath89 .
so by , if @xmath322 , then @xmath323 for fixed @xmath89 and @xmath45 , @xmath177 is a boundary arc of @xmath5 that lies between @xmath178 and @xmath324 .
thus , by , we have that @xmath325 thus , if @xmath326 and @xmath274 , then @xmath327 if @xmath328 , then @xmath326 .
if @xmath329 , then @xmath330 suppose that @xmath331 ; by monotonicity of @xmath332 we have that @xmath333 thus , if @xmath334 , then there is some positive constant @xmath335 such that @xmath336 for all @xmath89 and @xmath45 ; that is , the arclength of every boundary arc is at least @xmath337 .
let @xmath229 be any sequence that satisfies conditions and of
. by assumption on @xmath4 , condition holds . since @xmath338 ,
if @xmath89 is large enough then _ no _ boundary arcs of @xmath5 have arclength less than @xmath235 , and so condition holds .
thus the conditions of hold , and so a bounded simply connected domain @xmath0 exists such that @xmath25 .
[ rmk : fjumpexample ] if @xmath339 , and the numbers @xmath340 are defined as above , then @xmath341 , @xmath342 , and @xmath343 .
recall the function @xmath344 of .
we remark that for this function , @xmath339 . also , @xmath345 and @xmath346 , so @xmath347 .
thus , we see that @xmath4 satisfies the conditions of , and so @xmath25 for some simply connected domain @xmath0 .
let @xmath4 be a function .
under certain conditions ( theorems [ thm : main ] and [ thm : fjump ] ) , there exists a simply connected domain @xmath0 such that @xmath25 .
we are interested in whether this domain @xmath0 is unique ; that is , if @xmath0 and @xmath348 are domains and @xmath349 , what else must be true of @xmath0 and @xmath348 to allow us to conclude that @xmath350 ?
it is clear that some conditions must be imposed .
if a domain @xmath0 is rotated around the point 0 , or reflected across a line through 0 , then the domain is changed but the harmonic measure distribution function remains the same .
if a single point is deleted from a domain @xmath0 , then the harmonic measure of all boundary sets in @xmath50 is unchanged , and so the harmonic measure distribution function is unchanged .
more generally , if @xmath351 has harmonic capacity zero and @xmath352 , then @xmath0 and @xmath353 have the same harmonic measure distribution function . in the following discussion
, we will disregard sets of harmonic capacity zero .
much more interesting examples of non - uniqueness exist .
consider domains whose harmonic measure distribution functions are step functions .
( such functions were studied extensively in @xcite .
the underlying domains of course are not simply connected . )
if @xmath354 is a step function with discontinuities at @xmath167 , @xmath355 , then @xmath356 is a subset of @xmath357 . the proof of ( * ?
* theorem 2 ) implies that there are uncountably many such domains @xmath87 with @xmath358 .
for example , @xmath359 may be taken to be an arc with arclength of any preassigned number @xmath335 , @xmath360 ; the sets @xmath361 may be taken to be connected arcs centered at any preassigned angles , or indeed to be disconnected sets .
thus , if @xmath362 is a step function , then @xmath363 for many domains @xmath364 .
however , if @xmath87 and @xmath365 are circle domains in the sense of , then @xmath363 implies @xmath366 .
( see . ) furthermore , if @xmath362 is a step function , then @xmath87 is a circle domain if and only if @xmath87 is bounded and symmetric in the sense of .
we conjecture that these conditions suffice to imply uniqueness .
[ conj : unique ] suppose that @xmath0 and @xmath348 are two domains , both of which are bounded and symmetric in the sense of .
suppose further that @xmath349 .
then @xmath350 up to a set of harmonic capacity zero .
we will prove only in the two special cases where @xmath0 is a circle domain ( ) or where @xmath0 is simply connected ( )
. we will also show ( ) that the simply connected domains @xmath0 produced by are symmetric ; since they are clearly bounded , will apply .
thus , if a function satisfies the conditions of , then it arises as the @xmath39-function of a unique bounded , simply connected symmetric domain . in , we will state this conclusion more precisely . throughout this section , by `` symmetric ''
we mean `` symmetric in the sense of . ''
[ lem : circleunique ] let @xmath87 and @xmath367 be two circle domains .
suppose that @xmath363 .
then @xmath368 except possibly for finitely many points on the positive real axis . in the proof of , it was convenient to allow finitely many boundary arcs of arclength 0 , that is , boundary `` arcs '' consisting of single points on the positive real axis .
such points have harmonic measure zero and can not be detected from the harmonic measure distribution function . for the remainder of this proof ,
we ignore such points ; that is , we assume that the boundary arcs of @xmath87 and @xmath367 have positive arclength .
since @xmath87 is a circle domain , @xmath362 must be a step function .
let its discontinuities be at @xmath167 , @xmath355 .
the boundary arcs of @xmath87 are circular arcs with midpoints lying on the positive real axis .
let @xmath253 and @xmath369 be the boundary arcs of @xmath87 and @xmath367 , respectively , and let @xmath370 , @xmath371 be half their arclengths .
we need only show @xmath372 for @xmath274 . since @xmath87 and @xmath367 are bounded , the outermost boundary component is a full circle , and so @xmath373 .
suppose that @xmath374 for at least one @xmath45 .
let @xmath11 be the union of all arcs @xmath253 such that @xmath375 , and let @xmath376 be the union of the corresponding arcs of @xmath367 .
then @xmath376 is nonempty and @xmath377 , and since @xmath378 , we have that @xmath379 . then @xmath380 because @xmath363 . but
@xmath381 by the property of monotonicity in the domain of harmonic measure .
since @xmath382 is a _ proper _ subset of both @xmath11 and @xmath383 , we have that @xmath384 .
but @xmath385 and @xmath386 .
thus , @xmath387 this is a contradiction ; thus @xmath388 for @xmath274 .
similarly , @xmath389 for @xmath274 and so @xmath368 except possibly for finitely many points on the real axis .
now , we consider simply connected domains . we begin with some general remarks .
[ rmk : sccdsymm ] let @xmath0 be a bounded simply connected symmetric domain .
let @xmath37 and @xmath71 be the largest and smallest numbers , respectively , such that @xmath390 . by the remarks after , @xmath391 .
let @xmath144 be the riemann map ; assume that @xmath58 is normalized such that @xmath145 and @xmath146 . since normalized riemann maps are unique , and since @xmath0 is symmetric about the real axis , @xmath392 .
thus , @xmath393 is real if and only if @xmath49 is real .
since @xmath145 , @xmath146 , and @xmath58 is continuous and one - to - one , we have that @xmath394 and @xmath395 .
[ lem : mainsymmetric ] let @xmath41 be a sequence of bounded , simply connected domains containing 0 that are symmetric in the sense of .
let @xmath134 be the riemann maps of the domains @xmath126 , normalized so that @xmath136 and @xmath137 .
suppose that @xmath147 uniformly on @xmath132 , where @xmath58 is the riemann map of some simply connected domain @xmath0 .
then @xmath0 is symmetric .
in particular , if @xmath4 satisfies the conditions of then @xmath25 for some bounded , simply connected , symmetric domain @xmath0 .
recall that a bounded domain @xmath0 is symmetric in the sense of if and only if @xmath396 is path - connected for all @xmath12 , @xmath0 is symmetric about the real axis , and @xmath397 for some @xmath398 . by , @xmath399 for all @xmath159 . since @xmath400 , we have that @xmath392 , implying that @xmath0 is also symmetric about the real axis .
let @xmath401 , @xmath71 be the smallest numbers such that @xmath402 , @xmath403 . since @xmath147 uniformly , @xmath404 . by , @xmath405 , and
. thus @xmath406 .
we need only show that the set @xmath396 is path - connected for all @xmath407 .
we remark that it is easier to show path - connectedness of open sets than of arbitrary sets .
thus , we begin by showing that if the two open sets @xmath16 and @xmath408 are path - connected , then so is @xmath396 .
let @xmath49 , @xmath409 .
then @xmath410 .
we want to show that some path - connected arc of @xmath100 , with endpoints @xmath49 and @xmath411 , is contained in @xmath0 .
because @xmath0 is open , there exists some @xmath111 such that @xmath412 and @xmath413 .
take @xmath414 , and take @xmath415 .
suppose that @xmath16 is path - connected .
then there exists a path @xmath416 that connects @xmath49 to @xmath417 to @xmath418 to @xmath411 , and that ( except for the endpoints ) lies entirely in @xmath16 .
similarly , if @xmath408 is path - connected , then there is a path @xmath419 that connects @xmath411 to @xmath49 and that lies in @xmath408 .
then @xmath420 is a simple closed curve lying in @xmath0 .
because @xmath0 is simply connected , the interior of @xmath420 must lie in @xmath0 , and so an arc of @xmath100 connecting @xmath49 and @xmath411 must lie in @xmath0 .
thus , if the two open sets @xmath16 and @xmath421 are both path - connected , then so is the set @xmath396 .
we wish to show that @xmath16 is path - connected .
we first show that @xmath422 is path - connected for all @xmath139 and all @xmath423 .
choose some @xmath139 and fix some @xmath423 .
since @xmath126 is symmetric , if @xmath49 , @xmath424 with @xmath425 , then @xmath49 and @xmath411 may be connected by a path lying along the two arcs @xmath426 and @xmath427 and along the segment @xmath428 $ ] lying on the negative real axis . in particular ,
if @xmath429 and @xmath430 then @xmath49 and @xmath411 may be connected by a path lying in @xmath422 .
we now pass to the domain @xmath0 .
let @xmath431 , @xmath432 .
then for @xmath192 , @xmath18 , we have @xmath433 for some @xmath434 , @xmath435 .
let @xmath436 , and let @xmath89 be large enough that @xmath437 for all @xmath438 .
so @xmath439 and thus @xmath440 .
choosing @xmath441 , we see that @xmath442 is path - connected , and so there is some continuous function @xmath443\mapsto\omega_n\cap b(0,r-\varepsilon)$ ] such that @xmath444 and @xmath445 .
consider @xmath446 .
this is a continuous path connecting @xmath431 and @xmath447 , and @xmath448)\subset\omega$ ] .
furthermore , @xmath449 and so @xmath448)\subset b(0,r)$ ] as well .
thus , @xmath16 is path - connected for all @xmath12 .
similarly , @xmath408 is connected .
this completes the proof that @xmath0 is symmetric .
we now show that among bounded simply connected symmetric domains , @xmath39-functions uniquely determine the domain .
[ thm : unique ] suppose that @xmath450 , for two planar domains @xmath0 and @xmath348 such that [ item : unique : bdd ] @xmath0 and @xmath348 are bounded and contain the point @xmath13 , [ item : unique : symm ] @xmath0 and @xmath348 are symmetric in the sense of , and [ item : unique : sccd ] @xmath0 and @xmath348 are simply connected . then @xmath350 .
let @xmath451 , and let @xmath452 . by the properties of @xmath39-functions , @xmath37 and @xmath71 are the largest and smallest numbers , respectively , such that @xmath390
; since @xmath453 we have that @xmath454 .
let @xmath455 , @xmath456 be the riemann maps ; we may assume that @xmath457 , and that @xmath146 , @xmath458 . then by , @xmath459 .
now , @xmath58 is a bounded harmonic function defined on a @xmath460 domain . as is well - known (
see for example ( * ? ? ?
* theorem 1.4.7 ) ) , it follows that the radial limit @xmath461 exists for almost every @xmath462 $ ] .
( in fact , a stronger notion of limit , called the _ non - tangential limit _ , exists for a.e . @xmath101 . ) suppose that for @xmath192 , @xmath18 , @xmath463 exists for some @xmath464 , @xmath465 , with @xmath466 .
consider the segments @xmath467 .
then @xmath468 is a path connecting @xmath13 to @xmath469 .
furthermore , @xmath470 and @xmath471 do not intersect , and they lie entirely in the ( closed ) upper half - plane . since riemann maps are orientation - preserving , and @xmath463 lies in @xmath50 , by the symmetry condition we must have that @xmath472 .
thus , @xmath473 is defined for a.e .
@xmath474 $ ] , and is nondecreasing . if @xmath475 , then let @xmath476 , and let @xmath477 ; recall from that @xmath478 is harmonic in @xmath0 .
then let @xmath479 .
suppose that @xmath480 , and @xmath481 .
then @xmath482 for a.e .
@xmath483 , and so @xmath484 .
but @xmath485 ; thus if @xmath481 and @xmath480 , then @xmath486 .
recall from that @xmath487 ; therefore , if @xmath488 and @xmath481 , then @xmath489 .
similarly , if @xmath490 , then @xmath491 . because @xmath0 is simply connected , @xmath3 is strictly increasing on @xmath208 $ ] .
therefore , we may extend @xmath492 to a continuous function @xmath493\mapsto [ \mu , m]$ ] .
we then have that @xmath494 for a.e
. @xmath101 .
similarly , @xmath495 .
but since @xmath496 , this means that @xmath497 for a.e .
@xmath101 .
consider @xmath498 .
since @xmath457 and @xmath58 , @xmath499 are injective on @xmath132 , @xmath500 is analytic with a removable singularity at @xmath13 , and is never 0 on @xmath135 .
recall that @xmath501 is continuous on @xmath502 and that its value does not depend on the choice of branch cut of @xmath503 .
thus , @xmath504 is continuous on @xmath132 . for
any given @xmath159 , we may take the branch cut of @xmath503 to avoid a neighborhood of @xmath505 .
thus @xmath506 is harmonic in a neighborhood of any point @xmath159 , and so @xmath506 is harmonic on all of @xmath132 .
furthermore , @xmath507 for a.e .
@xmath508 ; thus @xmath509 in @xmath132 , and so @xmath510 must be an imaginary constant on @xmath132 .
since @xmath511 , we must have that @xmath512 and so @xmath513 .
since @xmath157 and @xmath514 , this implies that @xmath350 .
it is possible to weaken slightly the conditions and .
suppose that @xmath0 is symmetric . if @xmath0 is simply connected , then @xmath0 is bounded if and only if @xmath50 is bounded , which in turn is true if and only if @xmath515 for some finite number @xmath71 .
conversely , suppose that @xmath0 is symmetric , bounded and connected .
we know that there exist numbers @xmath37 , @xmath71 such that @xmath516 on @xmath300 , @xmath517 on @xmath210 , and @xmath518 on @xmath19 . then @xmath519 up to a set of harmonic capacity zero . if @xmath519 and @xmath110 $ ] is empty , then by symmetry @xmath0 is simply connected .
again by symmetry , @xmath110 $ ] is empty if and only if @xmath520 is strictly increasing on @xmath208 $ ] .
thus , if @xmath348 is symmetric and either simply connected or bounded and connected , and if @xmath496 for some domain @xmath0 that satisfies all three conditions of , we have that up to a set of harmonic measure zero , @xmath348 must be bounded and simply connected and so the theorem holds .
however , we do need to require that @xmath348 be bounded or simply connected . in (
* example 1 ) , the authors showed that the bounded simply connected symmetric domain @xmath521 has the same harmonic measure distribution function as the unbounded , doubly connected symmetric domain @xmath522 .
as seen from the example of circle domains , we probably can not weaken the condition that both @xmath0 and @xmath348 be symmetric .
it is possible to weaken the requirement that @xmath0 be simply connected by generalizing , but removing this requirement entirely is beyond the scope of this paper .
we conclude this section by remarking that , if a function @xmath4 satisfies the conditions of or , then there exists a bounded simply connected domain @xmath0 such that @xmath25 . by
, @xmath0 is symmetric .
so gives uniqueness , and the following theorem is proven .
[ thm : mainunique ] let @xmath4 be a function that satisfies the conditions of or .
then there exists a domain @xmath0 that is bounded and simply connected , has locally connected complement and is symmetric in the sense of , such that @xmath25 . furthermore , this domain @xmath0 is unique up to sets of harmonic measure zero among bounded symmetric domains .
in this section we state and prove some lemmas , involving harmonic measure and circle domains , that we have used in this paper .
we begin with a fundamental estimate of harmonic measure .
next come our estimates for individual circle domains , and finally our estimates for sequences of blocked circle domains .
specifically , lemmas [ lem : channelstraight ] and [ lem : curvechannel ] let us bound the harmonic measure of a set at the bottom of a channel ( such as a short arc in a circle domain or a gate in a blocked circle domain ) . puts a lower bound on the length of an arc in terms of its harmonic measure , while controls the harmonic measure of an arc lying between two other arcs . puts a bound on the difference in harmonic measure distribution functions between a circle domain and the corresponding blocked circle domain ( using lemmas [ lem : curvechannel ] and [ lem : altgatebound ] ) .
finally , provides a sufficient condition for a sequence of blocked circle domains to be uniformly locally connected .
[ lem : channelstraight ] suppose @xmath523 is a finitely connected jordan domain and suppose @xmath524 let @xmath525 with @xmath526 and assume that for @xmath527 , @xmath528 separates @xmath431 from @xmath529 .
if the length @xmath530 of @xmath531 is measurable , then @xmath532 roughly speaking , if a brownian particle is released from a point @xmath431 inside a long , thin channel , the probability that the particle reaches the end of the channel before it hits the side increases with the width of the channel and decreases exponentially with the length of the channel .
but @xmath543 is a curved rectangle bounded away from zero . by applying the conformal map @xmath554
, we may transform @xmath543 to the straight rectangle @xmath555 . then @xmath556 .
implies that @xmath557 note that @xmath558 . since @xmath559 , this completes the proof .
we seek to understand the harmonic measure distribution of circle domains more precisely .
we begin by putting a lower bound on the arclength of an individual arc in terms of the harmonic measure of that arc .
let @xmath578 ; recall from that @xmath579 is harmonic in @xmath580 , @xmath581 on @xmath97 and @xmath582 on @xmath583 . by the property of monotonicity in the domain of harmonic measure , @xmath584 since @xmath579 is harmonic in @xmath585 , we have that @xmath586 we seek an upper bound on @xmath579 on @xmath100 .
consider the function @xmath587 where @xmath588 is a positive real number ( not necessarily an integer ) .
we take the branch cut of @xmath589 and of @xmath590 to lie along the negative real axis . since @xmath595
, @xmath596 is analytic in a neighborhood of @xmath431 . if @xmath591 then @xmath597 . if @xmath598 or @xmath599 then this inequality is strict , so @xmath600 . by continuity
this inequality holds in a neighborhood of @xmath431 .
so the function @xmath601 is analytic and bounded away from the negative real axis ( the branch cut of @xmath590 ) in a neighborhood of @xmath431 ; thus @xmath594 is harmonic in that neighborhood .
choose @xmath610 ; then @xmath611 since @xmath612 .
consider the domain @xmath613 then @xmath50 consists of three smooth pieces : the two straight lines from the origin to @xmath614 , and the arc ( of radius @xmath71 ) connecting these two points and passing through the negative real axis .
then @xmath579 is harmonic in @xmath615 and continuous on @xmath54 , and the function @xmath616 is harmonic in @xmath0 and continuous on @xmath54 except at the points @xmath617 .
we remark that the boundary values of @xmath500 are known ; @xmath618 on the straight segments and @xmath619 on the boundary circle .
@xmath620 is also zero on the boundary circle ; on the straight segments , @xmath620 is unknown but satisfies @xmath621 with equality holding only at @xmath622 .
so by the maximum principle @xmath623 in @xmath0 . recall that @xmath624 where @xmath2 is the radius of the arc @xmath97 .
since @xmath625 on @xmath626 , and @xmath627 for @xmath628 , we may rewrite this integral as @xmath629 setting @xmath637 , we see that @xmath638 for any @xmath639 with @xmath640 , where @xmath641 .
this inequality holds for any arc @xmath97 , of any radius or arclength , with @xmath33 .
unfortunately , we seek a lower bound on @xmath335 and not an upper bound on @xmath32 , and so we must choose @xmath639 and then solve for @xmath335 ; to do so , we assume @xmath642 and @xmath643 ( implying @xmath644 and that @xmath97 lies in the right half - plane ) .
suppose @xmath645 .
then since @xmath611 we have @xmath646 and so we may choose @xmath647 .
now , @xmath648 since @xmath644 we have @xmath649 ; by our assumption on @xmath2 we have @xmath650 .
the function @xmath651 is increasing on @xmath652 .
so @xmath653 solving for @xmath335 completes the proof .
next , we prove a lemma that controls the harmonic measure of arcs lying in between other arcs .
this lemma is somewhat technical , but very useful .
informally , states that , if the @xmath39-function of a circle domain @xmath5 approximates some function @xmath4 , and the minimal secant slope of @xmath4 is positive , then any boundary arc long enough to be almost a full circle must be close to the outer boundary circle , and if two boundary arcs are close together then the arcs between them can not be too short .
these conditions mirror the conditions of , which provides a sufficient condition for a sequence of blocked circle domains to be uniformly locally connected .
thus , and are vital to the proof of . also , in combination with provides sufficient conditions to ensure that none of the boundary arcs are very short , and so lemmas [ lem : fjump ] and [ lem : deriv ] combine to give . [ lem : deriv ] let @xmath4 be a function that is a candidate for a harmonic measure distribution function , as in , and define the numbers @xmath37 and @xmath71 and the circle domains @xmath5 as in that section .
let @xmath212 be the minimal secant slope of @xmath4 , as in .
fix some @xmath654 . for each @xmath655 and each @xmath274 , write @xmath656 , @xmath657 , @xmath658 , @xmath659 , @xmath660 .
the conclusion follows from because the function @xmath670 is increasing on @xmath671 $ ] and satisfies @xmath672 .
we remark that by the definition of @xmath662 , if @xmath85 and @xmath253 are two boundary arcs , then @xmath673 thus , if @xmath674 or @xmath370 is small then holds automatically ; is of interest mainly if both of the arcs @xmath85 and @xmath253 are relatively long .
recall that the @xmath679th boundary arc of @xmath87 is denoted @xmath85 , is a subset of the circle of radius @xmath95 centered at zero , and has arclength @xmath680 . if @xmath85 and @xmath253 are two boundary arcs with @xmath292 , then @xmath681 , and if another arc @xmath682 lies between @xmath85 and @xmath253 , then @xmath683 and so @xmath194 . *
claim : * let @xmath684 be an integer with @xmath685 .
then there is an increasing function @xmath686 , defined for @xmath687 , that satisfies the following condition .
suppose that @xmath85 and @xmath253 are two boundary arcs with @xmath681 , and that there are @xmath688 arcs @xmath682 between @xmath85 and @xmath253 .
then there are at most @xmath689 such arcs @xmath682 that satisfy @xmath690 the function @xmath691 depends only on @xmath684 , @xmath212 , @xmath37 and @xmath71 .
in particular , @xmath686 does not depend on @xmath89 , the number of boundary arcs .
suppose that the claim holds .
if @xmath85 and @xmath253 are any pair of arcs with @xmath292 , let @xmath688 be the number of arcs @xmath682 that lie between @xmath85 and @xmath253 .
the number of such arcs @xmath682 which also satisfy @xmath692 is an integer which is at most @xmath689 .
therefore , if @xmath693 , then none of the arcs @xmath682 between @xmath85 and @xmath253 satisfy @xmath692 .
that is , none of these arcs have arclength less than @xmath694 . recalling the of @xmath662 as the depth of the shortest arc between @xmath85 and @xmath253 ,
we see that @xmath695 .
we will complete the proof of by showing that we may choose the functions @xmath686 such that @xmath696 , where @xmath697 is given by and @xmath698 equals the quantity on the right - hand side of .
let @xmath85 and @xmath253 be any two boundary arcs of @xmath87 with @xmath681 .
let @xmath702 .
we remark that if @xmath703 then the claim holds for any nonnegative functions @xmath686 ; we will therefore assume @xmath612 .
fix some number @xmath707 with @xmath708 . if @xmath682 is between @xmath85 and @xmath253 , and @xmath709 , we call @xmath682 a _ @xmath707-short arc_. let @xmath710 be the set of @xmath707-short arcs between @xmath85 and @xmath253 .
we want to find a function @xmath699 such that @xmath711 .
we will do this by bounding the harmonic measure of these arcs .
let @xmath713 , and let @xmath714 and @xmath715 be radial line segments of inner radius @xmath95 , outer radius @xmath561 , at angle @xmath716 .
let @xmath717 be the connected component of @xmath718 containing 0 .
let @xmath719 be the union of the @xmath707-short arcs , and let @xmath720 .
the set @xmath712 is drawn in bold in
. then @xmath721 by , @xmath722 notice that by definition of @xmath212 and @xmath87 , @xmath723 provided @xmath724 .
therefore , @xmath725 .
so @xmath726 we wish to control the right - hand side .
let @xmath727 .
if @xmath728 , define @xmath699 to be the number that satisfies @xmath729 we may solve for @xmath699 to see that @xmath730 observe that @xmath731 , and so the function @xmath732 is continuous on @xmath733 $ ] . furthermore , if @xmath734 then the function @xmath735 is increasing on @xmath736 $ ] . since @xmath737 ,
we have that @xmath738 , and so @xmath699 is increasing for @xmath739 $ ] . if @xmath740 , then @xmath741 , and so none of the arcs @xmath682 between @xmath85 and @xmath253 satisfy @xmath742 .
in this case , the claim holds for @xmath700 .
otherwise , @xmath743 and we may apply with @xmath744 . for
the inductive step , suppose that the function @xmath686 exists and is finite .
we want to show that @xmath751 exists .
pick two boundary arcs @xmath85 and @xmath253 with @xmath681 .
there are @xmath704 arcs between @xmath85 and @xmath253 , not including @xmath85 and @xmath253 .
define @xmath702 . recall that if @xmath682 lies between @xmath85 and @xmath253 , and @xmath709 , then we call @xmath682 a _ @xmath707-short arc_. we wish to find a @xmath751 such that at most @xmath752 arcs are @xmath753-short arcs . by definition of @xmath686 , at most @xmath754 arcs are @xmath755-short arcs .
we refer to @xmath755-short arcs simply as _
short arcs_. if we choose the function @xmath756 such that @xmath757 for all @xmath758 , then all @xmath753-short arcs are short arcs .
it is this assumption that allows us to work by induction .
if none of the arcs @xmath682 between @xmath85 and @xmath253 are short arcs , then the claim holds for any @xmath757 .
in particular , if @xmath759 then the claim holds . therefore , we assume that at least one arc @xmath682 is a short arc
. * @xmath763 and @xmath764 are not short arcs . *
if @xmath682 lies between @xmath763 and @xmath764 , then @xmath682 is a short arc . *
conversely , if @xmath682 is a short arc between @xmath85 and @xmath253 , then there is exactly one number @xmath765 with @xmath762 such that @xmath682 lies between @xmath763 and @xmath764 .
we begin our analysis of the short arcs by bounding @xmath767 for each @xmath765 .
there are @xmath768 arcs lying between @xmath763 and @xmath764 , and they are all short arcs
. there are at most @xmath769 short arcs between @xmath85 and @xmath253 .
therefore , @xmath770 recall that @xmath771 .
therefore , @xmath772 since @xmath773 , and @xmath685 , we have that @xmath774 where the second inequality is chosen for the sake of simplicity .
thus , if @xmath762 then @xmath775 .
we now apply this bound .
let @xmath776 denote the number of arcs between @xmath763 and @xmath764 . by definition of @xmath699
, there are at most @xmath777 boundary arcs @xmath682 between @xmath763 and @xmath764 that satisfy @xmath778 recall that @xmath699 is an increasing function .
furthermore , since @xmath763 and @xmath764 are not short , we have that @xmath779 . therefore
, if @xmath762 then @xmath780 suppose that @xmath682 lies between @xmath85 and @xmath253 and satisfies @xmath781 then @xmath682 is a short arc , and so there is some number @xmath765 such that @xmath782 . by , @xmath783 .
for each @xmath765 there are at most @xmath777 such arcs .
therefore , there are at most @xmath784 arcs @xmath682 between @xmath85 and @xmath253 that satisfy . but
@xmath785 is equal to the number of short arcs between @xmath85 and @xmath253 , which by definition is at most @xmath754 . in our proof of
, we needed conditions guaranteeing that @xmath237 .
this means that we want estimates on the harmonic measure of the gates of blocked circle domains . in this section
we establish those conditions .
we follow the development originally given in @xcite , but adapted to our situation . throughout this section ,
let @xmath87 be a circle domain and @xmath0 a blocked circle domain with @xmath93 , as in .
let their boundary arcs be at radii @xmath561 and have arclength @xmath793 , and let the gates be at angles @xmath794 .
let @xmath795 .
that is , @xmath796 measures the depth of the gate in its channel .
define @xmath801 $ ] to be a disk minus a slit . then since @xmath802 , @xmath803,\psi).\ ] ] the harmonic measure on the right - hand side may be computed explicitly : we transform @xmath804 to the upper half - plane @xmath805 via the conformal map @xmath806 the point @xmath13 is mapped to a point @xmath807 on the positive imaginary axis , and @xmath808 $ ] to an interval @xmath809 $ ] in @xmath810 . here
@xmath811 using the harmonic function @xmath812 we may compute @xmath813,\psi ) & = \omega(it,[-r',r'],{\mathbb{u } } ) = \frac{2}{\pi}\arctan\frac{r'}{t } \\&=\frac{2}{\pi } \arctan \sqrt{\frac{(r_{k+1}-r_{k})(m^2-r_{k+1}r_{k})}{{r_{k}(r_{k+1}+m)^2 } } } \leq\frac{2}{\pi } \sqrt{\frac{r_{k+1}-r_k}{r_k}}.\end{aligned}\ ] ] we omit the details .
recall that , in , we needed @xmath814 .
that is , we needed the harmonic measure distribution functions of blocked circle domains @xmath126 to approach those of the corresponding circle domains @xmath5 . in the next lemma , we assemble the known results involving gates in order to achieve a uniform bound on @xmath815 .
[ lem : radial ] let @xmath87 be a circle domain , and let @xmath93 be a blocked circle domain such that @xmath94 is a union of gates .
define @xmath561 , @xmath370 , and @xmath816 , @xmath796 as in , so that @xmath561 is the radius of the @xmath45th arc , @xmath796 is the inset angle of the @xmath45th pair of gates , and so on . fix
@xmath817 $ ] .
we first show that @xmath819 is bounded by the harmonic measure of the gates in @xmath0 .
recall that @xmath820 let @xmath821 and @xmath822 be the portions of @xmath356 and @xmath50 , respectively , lying within a distance @xmath2 of zero . then @xmath823 and @xmath824 , and so we need only bound @xmath825
. we may transform @xmath87 to @xmath0 by adding the gates ( and deleting disconnected components ) .
let @xmath826 be the union of the gates , that is , the newly - added boundary .
then @xmath827 , and @xmath828 is empty .
let @xmath837 be a gate that lies between the arcs @xmath253 and @xmath838 .
we apply with @xmath839 , @xmath140 , @xmath840 , @xmath841 , @xmath842 , and @xmath843 , @xmath844 . by definition of @xmath370 and @xmath816 , the conditions of hold ; thus , @xmath845 finally , we move to sequences @xmath41 of blocked circle domains . in the proof of , we used a simple condition for such a sequence to have uniformly locally connected complements .
we now prove that implication .
fix some @xmath111 .
we wish to find some @xmath112 such that if @xmath49 , @xmath848 and @xmath849 , then there is a continuum of diameter at most @xmath117 contained in @xmath850 connecting @xmath49 and @xmath411 .
we need only consider the case where @xmath49 , @xmath851 .
in fact , we can go further and consider only the case where @xmath49 , @xmath411 lie on boundary arcs ( not gates ) , that is , where @xmath49 , @xmath852 .
we will use two elementary geometric estimates throughout this proof .
first , let @xmath855 be an annular sector with inner radius @xmath2 , outer radius @xmath856 , and subtending an angle of @xmath101 .
then its diameter is bounded by the equation @xmath857 furthermore , let @xmath49 and @xmath411 be complex numbers .
suppose that the angle from zero between @xmath49 and @xmath411 is @xmath101 , @xmath480 .
then @xmath858 take @xmath49 , @xmath852 with @xmath849 .
let @xmath859 , @xmath860 ; without loss of generality @xmath861 . then @xmath862 .
we wish to show that @xmath49 and @xmath411 may be connected by a continuum of diameter at most @xmath117 . and @xmath411 in and the subsets of @xmath50 connecting them . in the diagram
at left , @xmath49 and @xmath411 lie on the same side of the real axis ; in the diagram on the right , they lie on opposite sides of the real axis.,title="fig : " ] if they do , then they are connected by a continuum lying in @xmath850 that lies on the boundary of a sector of an annulus of inner radius @xmath863 , outer radius @xmath180 , and that subtends an angle at most @xmath864 .
( see a. ) by , the diameter of this continuum is at most @xmath865 since @xmath866 , we know by condition that @xmath867 . also , @xmath868 . by choosing @xmath869
, we see that @xmath49 , @xmath411 are connected by a continuum in @xmath850 of diameter at most @xmath117 .
if @xmath49 and @xmath411 lie on different sides of the real axis , then since @xmath870 , @xmath49 and @xmath411 must lie on the same side of the imaginary axis .
if they lie on the right - hand side , then they can be connected by a path contained in the union of their boundary arcs in @xmath320 with the positive real axis .
if the domain of the @xmath590 function is taken to be @xmath871 $ ] , then this path has length @xmath872 by , @xmath873 recall that @xmath874 , and so @xmath875 measures the inset angle of the deepest gate between the arc @xmath186 and the outer boundary circle .
let @xmath876 consider the annular sector , symmetric about the real axis and containing part of the negative real axis , with inner radius @xmath863 , outer radius @xmath874 , and lying between the angles @xmath877 .
( see b. ) this sector subtends an angle of @xmath878 at zero .
now , @xmath49 and @xmath411 can be connected by a continuum in @xmath50 lying in this sector . by
, this path has diameter at most @xmath879 lars v. ahlfors , _ complex analysis _
, third ed .
, mcgraw - hill book co. , new york , 1978 , an introduction to the theory of analytic functions of one complex variable , international series in pure and applied mathematics .
mr 510197 ( 80c:30001 ) carlos e. kenig , _ harmonic analysis techniques for second order elliptic boundary value problems _ , cbms regional conference series in mathematics , vol .
83 , published for the conference board of the mathematical sciences , washington , dc , 1994 .
mr 1282720 ( 96a:35040 ) christian pommerenke , _ boundary behaviour of conformal maps _ , grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences ] , vol .
299 , springer - verlag , berlin , 1992 .
mr 1217706 ( 95b:30008 ) byron l. walden and lesley a. ward , _ distributions of harmonic measure for planar domains _ , xvith rolf nevanlinna colloquium ( joensuu , 1995 ) , de gruyter , berlin , 1996 , pp .
mr 1427094 ( 98f:30026 )
|
let @xmath0 be a planar domain containing 0 .
let @xmath1 be the harmonic measure at 0 in @xmath0 of the part of the boundary of @xmath0 within distance @xmath2 of 0 .
the resulting function @xmath3 is called the harmonic measure distribution function of @xmath0 . in this paper
we address the inverse problem by establishing several sets of sufficient conditions on a function @xmath4 for @xmath4 to arise as a harmonic measure distribution function .
in particular , earlier work of snipes and ward shows that for each function @xmath4 that increases from zero to one , there is a sequence of multiply connected domains @xmath5 such that @xmath6 converges to @xmath4 pointwise almost everywhere .
we show that if @xmath4 satisfies our sufficient conditions , then @xmath7 , where @xmath0 is a subsequential limit of bounded simply connected domains that approximate the domains @xmath5 .
further , the limit domain is unique in a class of suitably symmetric domains .
thus @xmath8 for a unique symmetric bounded simply connected domain @xmath0 .
|
one - dimensional markov processes ( such as random walks or brownian motion ) conditioned not to intersect form an important class of models which arise in the study of random matrix theory , growth processes , directed polymers , tilings and certain problems in combinatorics and representation theory ( see the surveys @xcite ) .
the probability distribution for these collections of lines may be analysed using a technique which exploits the non - intersection property : in different guises this tool is the karlin - mcgregor formula , the lingstrom - gessel - viennot formula , and the physics method of free fermions .
these methods yield exact expressions ( as determinants ) for the statistics of such conditioned processes .
asymptotic analysis then gives rise to certain universal scaling limits in the sense of convergence of finite - dimensional distributions and accordingly provides exact expressions for these distributions .
striking examples of this are the works of @xcite which give limit theorems for the finite - dimensional distribution of the fluctuations of the height functions for the polynuclear growth ( png ) model in terms of fredholm determinants involving the extended @xmath15 kernel ( see also @xcite ) . using exactly solvable methods , @xcite further proved a functional limit theorem and established the existence of a continuous version of a stochastic process with the above family of finite - dimensional distributions which is known as the airy ( or sometimes @xmath15 ) process .
such scaling limits have often been studied by analysing exact ( and often determinantal ) formulas for finite - dimensional distributions and correlation functions .
it is natural to think that such canonical random processes might also be studied by probabilistic techniques ; and this study seems all the more warranted in light of the numerous significant questions about these limiting processes which have remained unanswered by the existing approaches . in this paper
, we begin such a probabilistic study and solve a number of these open questions .
the main tool which we will employ is the notion of a _
gibbs property_. let us illustrate this idea in an informal way , so as to avoid introducing unnecessary notation ( the specific _ brownian gibbs property _ which we will use is given explicitly in definition [ maindefbgp ] ) . in studying random processes ( or fields ) @xmath16 from @xmath17 ( where @xmath18 may be taken to be a discrete lattice , graph , or euclidean space ) , it is natural to consider the conditional distribution of @xmath16 inside a compact subset @xmath19 of @xmath18 , given its values on @xmath20 .
there is a class of such processes with the property that these distributions depend only on the values of @xmath16 on the boundary of @xmath19 ( in the discrete case , we mean the exterior boundary ) .
furthermore , given these values , the distribution of @xmath16 on @xmath19 is determined relative to a reference measure in terms of a radon nikodym derivative which is often written in terms of a hamiltonian , and which is measurable with respect to the sigma - field generated by @xmath16 on @xmath19 and its ( exterior ) boundary .
this property , which is often called the gibbs property , may be regarded as a spatial version of the markov property .
gibbs properties for random lattice indexed fields are ubiquitous in models of statistical physics ( for instance , in ising or potts models ) and have received extensive treatment since the seminal work of dobrushin @xcite .
we dispense with generalities and focus on a simple , though highly relevant , example of the gibbs property .
fix @xmath21 and let @xmath22 \to { \ensuremath{\mathbb{z}}}\}_{k=1}^{n}$ ] denote a system of @xmath0 simple symmetric random walks ( starting at time @xmath23 and ending at time @xmath24 ) whose starting and ending points satisfy @xmath25 , and which are conditioned on the non - intersection requirement that @xmath26 for all @xmath27 and @xmath28 .
fix two integers @xmath29 and a line index @xmath30 . then , conditioned on the values of @xmath31 , @xmath32 and @xmath33 for @xmath34 , the law of @xmath35 is uniform over the set of simple random walk paths of length @xmath36 between @xmath31 and @xmath32 that intersect neither @xmath37 nor @xmath38 .
( if @xmath39 , there is only one path to be avoided . ) in other words , the line ensemble measure is invariant under resampling according to the usual random walk measure subject to the non - intersection condition .
the continuum counterpart of this system is a variant of dyson s brownian motion given by several brownian bridges conditioned not to intersect .
the resampling property has an analogue as well : the rule is the same , except that the underlying measure is now an independent system of brownian bridges between the given endpoints , rather than the uniform law on simple random walk paths .
this property , which we call the _ brownian gibbs property _ and introduce formally in definition [ maindefbgp ] , is easily understood and in many ways unsurprising .
the main contribution of this article is the observation of theorem [ mainthm2 ] , that this brownian gibbs property is preserved in the edge scaling limits of non - intersecting line - ensembles ( such as the random walk and brownian bridge examples which we have seen ) . from this perspective
, the observation seems fairly intuitive ; however , the processes encountered in these scaling limits arise in many other perspectives from which this gibbs property is neither straightforward nor intuitive in fact , this observation has not previously appeared in the literature . before discussing the consequences of the brownian gibbs property
, we briefly discuss the two main difficulties in the proof of the result .
each difficulty derives from the edge scaling limit results for such systems having only been proved in terms of finite - dimensional distributions .
the first problem is whether the limiting consistent family of finite - dimensional distributions , which is called the multi - line airy process @xcite , has a version which is supported on continuous , non - intersecting curves .
restricting to just the top line , the existence of a continuous version was shown in @xcite by the exactly solvable systems approach involving asymptotic analysis of fredholm determinant expressions .
the second difficulty is to show convergence in distribution of the finite system of brownian bridges to this limiting ensemble .
once these two difficulties are settled , a coupling argument serves to prove that the brownian gibbs property is maintained in the limit .
in fact , we resolve both difficulties simultaneously by proving a functional limit theorem for the brownian bridge line ensembles ( which is in essence tightness for this family of curves ) , thus showing the existence of the _ airy line ensemble_. in proving such tightness ( as well as the non - intersecting nature of the limit ) , one can not appeal entirely to the finite - dimensional distributions , as they pertain only to a finite number of deterministic times and not random exceptional times at which there may be a small gap between lines or a large modulus of continuity for some line .
the proof must also handle the squeezing of space that occurs under edge scaling .
the determinantal approach seeming to fail here , we appeal to a probabilistic perspective .
we employ the brownian gibbs property of @xmath0 brownian bridges to show that , focusing on the top @xmath28 curves in a rescaled window of size @xmath40 $ ] , the following three events occur with high probability : ( i ) the collection of curves remains uniformly absolutely continuous with respect to @xmath28 brownian bridges ; ( ii ) the minimal gap between consecutive curves remains uniformly bounded from below ; ( iii ) the top curve and the @xmath28-th curve remain uniformly bounded from above and below . that the scaling limit has the brownian gibbs property has a number of significant consequences . in particular , in proposition [ propabscon ]
we show that the top line of the airy line ensemble ( often known as the airy process ) is absolutely continuous with respect to brownian motion ( with diffusion parameter 2 ) on any fixed interval ( this is also true for every other line ) .
this leads to a proof of conjecture 1.5 of @xcite ( stated here as theorem [ propjoh ] ) , a long - standing claim that the airy process minus a parabola achieves its maximum at a unique point . as we explain in section [ absconsec ] , this conjecture leads to theorem [ maxdist ] , which identifies and proves the convergence of the law of the endpoint of a geometrically weighted ground state directed polymer ( i.e. , of the point - to - line maximizing path in last passage percolation ) . appealing to the brownian
absolute continuity result we prove herein , @xcite has employed the airy process continuum statistics of @xcite to give and prove an exact formula for the law of this endpoint .
proposition [ propabscon ] also extends results of @xcite and proves a functional central limit theorem for the airy process on short time scales . in the process of proving our main theorem
, we record several results of independent interest ( see section [ absconsec ] ) , including the proof of conjecture 1.21 of @xcite .
the extensive literature on non - intersecting line ensembles suggests many directions in which to pursue the approach introduced here .
the non - intersecting random walkers model described above goes by the name _ vicious walkers _ and was introduced by de gennes @xcite as a model of directed fibrous structures in @xmath41 dimensions .
this model was then studied in @xcite .
this area has remained of interest due to its connections with symmetric function theory ( specifically schur functions and enumeration of young tableaux as in @xcite ) , discrete analogues of random matrix theory ( for instance @xcite ) , and 2d yang - mills theory @xcite .
the non - intersecting brownian line ensembles that we consider are variants of dyson brownian motion @xcite and can be considered as continuum versions of the vicious walkers model .
part of the attention which they attract is due to the description that they offer of random matrix eigenvalue processes @xcite . within the above literature
are a number of interesting and important scaling limits and perturbations to the basic model considered in this paper .
perturbations to the first few brownian particles result in perturbed airy - like limiting line ensembles @xcite ( see our section [ airylikesec ] ) . scaling limits in the vicinity of the bulk of the brownian bridges
lead to the dyson sine process , and large - scale perturbations ( such as separating half of the starting points and/or ending points ) lead to perturbations in the limiting processes such as the pearcy @xcite or tacnode processes @xcite . likewise , limiting processes
have been derived when the brownian bridges are replaced by alternative path measures such as brownian excursions @xcite , or bessel processes @xcite .
the approach developed in this paper may be extended to study line ensembles with a gibbs property which penalises but does not exclude crossing of curves with consecutive labels . in an upcoming article @xcite , we employ our probabilistic techniques to construct the _ kpz line ensemble _ ( related to the multi - layer extension of the stochastic heat equation constructed in @xcite ) and prove that is has such a `` soft '' brownian gibbs property .
the top labeled curve of this line ensemble is the fixed time hopf - cole solution to the kpz equation with narrow - wedge initial data ( see @xcite ) . as a result of the gibbs property we are able to prove that this solution to the kpz equation is locally absolutely continuous with respect to brownian motion ( analogously to proposition [ propabscon ] ) . in the present paper ,
non - intersecting brownian paths play the key role ; in @xcite , this role is assumed by diffusions associated to the quantum toda lattice hamiltonian .
oconnell @xcite discovered these diffusions and their relationship to directed polymers ( and hence also to the kpz equation
see @xcite ) .
returning to discrete line ensembles , there are a variety of examples coming from the study of growth processes ( such as the various png line ensembles @xcite ) which have the same large @xmath0 limits as in the brownian case @xcite .
measures on rhombus tilings are closely related to representation theory as well as to vicious walkers . in that setting
, @xcite relates gibbs properties of discrete sine line ensembles to the associated determinant kernel .
the papers @xcite construct infinite - dimensional gibbs line ensembles corresponding to the bulk scaling limits of these tiling related line ensembles .
rhombus tilings also represent perfect matchings for the honeycomb lattice . in @xcite
, it is shown that models based on perfect matchings ( on any weighted doubly - periodic bipartite graph @xmath42 in the plane ) are exactly solvable in a rather strong sense .
the authors of @xcite not only derive explicit formulas for the surface tension ; they classify the gibbs measures on tilings and explicitly compute the local probabilities in each of them .
these results are a generalization of @xcite where similar results for @xmath43 with constant edge weights were obtained .
subsection [ lineensembledef ] contains the definition of a line ensemble and the brownian gibbs property ; subsection [ twohelp ] contains the statement of the strong gibbs property and monotonicity properties for brownian gibbs line ensembles ; subsection [ airylineensemblesec ] contains the definitions of the edge - scaled dyson line ensemble .
the main results of this paper are contained in section [ mainresec ] .
subsection [ secunique ] contains a uniqueness conjecture regarding the airy line ensemble ; subsection [ hypsec ] contains a more general formulation of the main result and a set of general hypotheses under which it holds ; subsection [ airylikesec ] records how many other line ensembles satisfy these more general hypotheses .
section [ absconsec ] presents some interesting consequences of our main results .
subsection [ subsubabs ] includes a proof that any line of the airy line ensemble minus a parabola has increments which are absolutely continuous with respect to the brownian bridge ( with diffusion parameter 2 ) ; in subsection [ uniquenesssec ] , this absolute continuity is used to prove the uniqueness of the location at which the maximum of the top line of the airy line ensemble minus a parabola is attained . in subsection [ transec ]
this uniqueness result implies that this location describes the endpoint of a directed polymer model .
the section also contains the proof of a conjecture of adler and van moerbeke .
the main technical results of the paper are contained in the estimates of proposition [ propacceptprob ] which is proved in section [ technicalsec ] .
finally , section [ proofslemmas ] contains a proof of the strong gibbs property lemma [ stronggibbslemma ] and the monotonicity lemmas [ monotonicity ] and [ lemmonotonetwo ] .
this project was initiated at the 2010 clay mathematics institute summer school in buzios , brazil .
the authors also thank the mathematical science research institute , the fields institute and the mathematisches forschungsinstitut oberwolfach for their hospitality and support , as much of this work was completed during stays at these institutes .
we thank jinho baik , jeremy quastel and herbert spohn for their input and interest .
we also thank our referee for a thorough reading of this work and many useful comments .
a.h . would like to thank scott sheffield for drawing attention to a talk in 2006 in which andrei okounkov proposed problems closely related to the discussion in section [ secunique ] and for interesting ensuing conversations , and neil oconnell and jon warren for useful early discussions regarding approaches to proving the results in this article .
i.c . recognizes support and travel funding from the nsf through the pire grant oise-07 - 30136 and grant dms-1208998 ; as well as support from the clay mathematics institute through a clay research fellowship and microsoft research through the schramm memorial fellowship .
a.h . was supported principally by epsrc grant ep / i004378/1 .
we begin by providing the necessary definitions and notations to state our main results .
we also include the statement of our two primary tools utilized in proving the main result the strong gibbs property and the monotonicity properties of line ensembles .
we also include some basic facts about brownian motions and brownian bridges . in order to state our main result
we must introduce the concept of a line ensemble and the brownian gibbs property .
[ maindef ] let @xmath18 be a ( possibly infinite ) interval of @xmath44 , and let @xmath45 be an interval of @xmath46 .
consider the set @xmath16 of continuous functions @xmath47 endowed with the topology of uniform convergence on compact subsets of @xmath48 .
let @xmath49 denote the sigma - field generated by borel sets in @xmath16 .
@xmath18-indexed line ensemble _
@xmath50 is a random variable defined on a probability space @xmath51 , taking values in @xmath16 such that @xmath50 is a @xmath52-measurable function .
intuitively , @xmath50 is a collection of random continuous curves ( even though we use the word `` line '' we are referring to continuous curves ) , indexed by @xmath18 , each of which maps @xmath45 into @xmath46
. we will often slightly abuse notation and write @xmath53 , even though it is not @xmath50 which is such a function , but rather @xmath54 for each @xmath55 .
furthermore , we write @xmath56 for the line indexed by @xmath57 . given a @xmath18-indexed line ensemble @xmath50 , and a sequence of such ensembles @xmath58 , a natural notion of convergence is the weak- * convergence of the measure on @xmath59 induced by @xmath60 , to the measure induced by @xmath50 ; we call this notion _ weak convergence as a line ensemble _ and denote it by @xmath61 . in order words
, this means that for all bounded continuous functionals @xmath62 , @xmath63 as @xmath64 .
a line ensemble is _ non - intersecting _ if , for all @xmath65 , @xmath66 for all @xmath67 .
all statements are to be understood as being almost sure with respect to @xmath68 .
we turn now to formulating the brownian gibbs property . as a matter of convention
, all brownian bridges will have diffusion parameter 1 unless otherwise noted .
[ maindefbgp ] let @xmath69 .
a point @xmath70 is called a @xmath28-decreasing list if @xmath71 for @xmath72 .
we write @xmath73 for the set of @xmath28-decreasing lists .
let @xmath74 and @xmath75 be two @xmath28-decreasing lists .
let @xmath76 satisfy @xmath77 , and let @xmath78 \to { \ensuremath{\mathbb{r}}}^*$ ] ( where @xmath79 ) be two given continuous functions that satisfy @xmath80 for all @xmath81 $ ] as well as the boundary conditions @xmath82 , @xmath83 and @xmath84 , @xmath85 . the _ @xmath86-avoiding brownian line ensemble on the interval @xmath87 $ ] with entrance data @xmath88 and exit data @xmath89 _ is a line ensemble @xmath90 with @xmath91 , @xmath92 $ ] and with the law of @xmath90 equal to the law of @xmath28 independent brownian bridges @xmath93 \to { \ensuremath{\mathbb{r}}}\}_{i=1}^{k}$ ] from @xmath94 to @xmath95 conditioned on the event that @xmath96 for all @xmath97 $ ] .
note that any such line ensemble @xmath90 is necessarily non - intersecting .
now fix an interval @xmath98 and @xmath99 and let @xmath100 and @xmath101 , with @xmath102 .
set @xmath103 and @xmath104 with the convention that if @xmath105 then @xmath106 and likewise if @xmath107 then @xmath108 .
write @xmath109 and @xmath110 . a @xmath18-indexed line ensemble @xmath111
is said to have the _
brownian gibbs property _ if @xmath112 where @xmath113 and @xmath114 is the @xmath86-avoiding brownian line ensemble on @xmath87 $ ] with entrance data @xmath115 and exit data @xmath116 .
note that @xmath114 is introduced because , by definition , any such @xmath86-avoiding brownian line ensemble is indexed from @xmath117 to @xmath118 , but we want @xmath90 to be indexed from @xmath119 to @xmath120 . [ wbdef ] let @xmath121 , @xmath102 , and @xmath122 .
write @xmath123 for the law of @xmath28 independent brownian bridges @xmath124 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath125 , that satisfy @xmath94 and @xmath95 .
write @xmath126 for the expectation under @xmath123 .
let @xmath127 \to { \ensuremath{\mathbb{r}}}$ ] be a measurable function such that @xmath128 and @xmath129 .
define the non - crossing event on an interval @xmath130 $ ] by @xmath131 the conditional measure @xmath132 } \big)$ ] is the @xmath133-avoiding line ensemble on @xmath87 $ ] with entrance data @xmath134 and exit data @xmath135 ; it will be denoted by @xmath136 .
we define the _ acceptance probability _ as @xmath137}).\ ] ] note that this implies that @xmath138})}{{\mathfrak{a}}(a , b,\bar{x},\bar{y},f)}.\ ] ] the following two sets of lemmas will be essential to the proof of our main results .
the proofs of these results appear at the end of the paper , in section [ proofslemmas ] .
we end this subsection with a few useful tidbits about brownian bridges . in order to introduce the strong gibbs property for brownian gibbs line ensembles
, we first introduce the concept of a stopping domain .
[ defstopdom ] consider a line ensemble @xmath139\to { \ensuremath{\mathbb{r}}}$ ] .
for @xmath140 , and @xmath30 denote the sigma - field generated by @xmath50 outside @xmath141 $ ] by @xmath142\setminus(\ell , r ) , \textrm { and } \mathcal{l}_{k+1},\ldots,\mathcal{l}_n \textrm { on } [ a , b]\big\}.\ ] ] the random variable @xmath143 is called a _ stopping domain _ for lines @xmath144 if for all @xmath145 , @xmath146 in other words , the domain is determined by the information outside of it .
we will generally assume that , when discussing a stopping domain , it is for the top @xmath28 indexed line in the line ensemble ( and will not mention this ) .
we will make use of a version of the strong markov property where the concept of stopping domain introduced in definition [ defstopdom ] plays the role of stopping time .
let @xmath147 denote the set of continuous functions @xmath148 with each @xmath149\to { \ensuremath{\mathbb{r}}}$ ] .
define @xmath150 let @xmath151 denote the set of borel measurable functions from @xmath152 .
[ stronggibbslemma ] consider a line ensemble @xmath139\to { \ensuremath{\mathbb{r}}}$ ] which has the brownian gibbs property .
write @xmath68 and @xmath153 as the probability measure and expectation on @xmath50 .
fix @xmath154 .
for all random variables @xmath143 which are stopping domains for lines @xmath155 , the following _ strong brownian gibbs property _ holds : for all @xmath156 , @xmath68 almost surely , @xmath157 = { \mathcal{b}_{\bar{x},\bar{y},f}}^{\mathfrak{l},\mathfrak{r } } \big[f(\mathfrak{l},\mathfrak{r},b_1,\ldots , b_k ) \big],\ ] ] where @xmath158 , @xmath159 , @xmath160 ( or @xmath161 if @xmath162 ) , @xmath163 is given in definition [ wbdef ] .
the proof of this lemma is given in section [ proofslemmas ] .
the main message of the lemma is that the distribution of the value of a line ensemble inside a stopping domain is entirely determined by the boundary data and specified according to the non - intersecting brownian bridge measure .
the following lemmas demonstrate certain typies of monotonicity which exist between non - intersecting brownian bridge measures .
[ monotonicity ] fix @xmath121 , @xmath164 and two measurable functions @xmath78\rightarrow
{ \ensuremath{\mathbb{r}}}\cup\{-\infty\}$ ] such that for all @xmath165 $ ] , @xmath166 .
let @xmath122 be two k - decreasing lists such that @xmath167 and @xmath168 .
recalling definition [ wbdef ] , set @xmath169 } \big)$ ] , and likewise define @xmath170 .
then there exists a coupling of @xmath171 and @xmath170 such that almost surely @xmath172 for all @xmath173 and all @xmath165 $ ] .
[ lemmonotonetwo ] fix @xmath121 , @xmath164 , a measurable function @xmath127\rightarrow { \ensuremath{\mathbb{r}}}\cup\{-\infty\}$ ] and a measurable set @xmath174 $ ] .
consider two pairs of k - decreasing lists @xmath175 and @xmath176 such that @xmath177 , @xmath178 and @xmath179 and @xmath180 for each @xmath125 .
then the laws @xmath181 and @xmath182 may be coupled so that , denoting by @xmath183 and @xmath184 the curves defined under the respective measures , @xmath185 for each @xmath125 and for all @xmath186 $ ] . the following decomposition is closely related to the lvy - ciesielski construction of brownian motion discussed in @xcite and can be proved by checking that the covariance of the constructed process @xmath187 coincides with that of a brownian bridge .
[ browniandecomp ] fix @xmath188 , @xmath189 and consider a sequence of times @xmath190 .
define a sequence of independent centered gaussian random variables @xmath191 so that @xmath192 = \frac{(t_i - t_{i-1})(t - t_{i})}{(t - t_{i-1})}.\ ] ] for each @xmath193 define an interpolation function @xmath194 and let @xmath195 . define a sequence of independent brownian bridges @xmath196 such that @xmath197\rightarrow { \ensuremath{\mathbb{r}}}$ ] with the property that @xmath198 and let @xmath199 .
then the random function @xmath200\rightarrow { \ensuremath{\mathbb{r}}}$ ] , @xmath201 is equal in law to a brownian bridge @xmath202 on @xmath203 $ ] ( i.e. , with @xmath204 and @xmath205 = \tfrac{s(t - s)}{t}$ ] ) . .
a brownian bridge may be constructed by fixing its value at an intermediate time @xmath206 according to a suitably scaled gaussian random variable @xmath207 , and then inserting a brownian bridge ( plus linear shift ) between @xmath208 and @xmath209 and a second brownian bridge ( plus linear shift ) between @xmath209 and @xmath210 .
this construction can be performed iteratively and results in a decomposition involving a collection of independent gaussian random variables , and independent brownian bridges .
the case illustrated corresponds to two iterates of this construction . , scaledwidth=50.0% ] [ cor : touch ] fix a continuous function @xmath211\rightarrow { \ensuremath{\mathbb{r}}}$ ] such that @xmath212 and @xmath213 .
let @xmath214 be a standard brownian bridge on @xmath215 $ ] .
define two events : @xmath216 ( crossing ) and @xmath217 ( touching )
. then @xmath218 . from lemma [ browniandecomp ]
( with the choice @xmath219 , @xmath220 and @xmath221 ) , we can decompose @xmath214 into two independent brownian bridges @xmath222\rightarrow { \ensuremath{\mathbb{r}}}$ ] for @xmath223 , and an independent centered gaussian random variable @xmath207 with @xmath224=1/4 $ ] .
let @xmath225 .
then , conditioned on @xmath226 and @xmath227 , the event @xmath228 holds only for a particular ( though random ) value of @xmath207 . however , due to independence and since @xmath207 is gaussian , the probability it takes a given value is zero , thus proving the corollary .
although the following result may be considered standard , we include it for the reader s convenience .
[ coropen ] consider the space of continuous functions from @xmath215 $ ] to @xmath46 endowed with the uniform topology and let @xmath229 be an open subset which contains a function @xmath62 such that @xmath230 .
let @xmath231\rightarrow { \ensuremath{\mathbb{r}}}$ ] be a standard brownian bridge
. then @xmath232 \subseteq u)>0 $ ] . set @xmath233 \subseteq u\}$ ] .
the set @xmath229 being open , it contains a piecewise linear function @xmath62 such that @xmath230 .
moreover , there is a @xmath234 such that @xmath235 .
assume that @xmath62 has a discontinuous derivative at times @xmath236 for some @xmath237 ( also set @xmath238 and @xmath239 ) .
using the decomposition of lemma [ browniandecomp ] ( into @xmath240 and @xmath183 ) we have @xmath241 \geq { \ensuremath{\mathbb{p}}}\big ( \bigcap_{i=1}^{j-1 } \big\{|b(t_i)-f(t_i)|<\delta/2\big\}\big ) { \ensuremath{\mathbb{p}}}\big ( \max_{1\leq i\leq j}\max_{s\in [ 0,t_{i}-t_{i-1 } ] } |b_i(s)| \leq \delta/2 \big).\ ] ] since @xmath242 are jointly gaussian , the first term on the right can be bounded from below by some @xmath243 , while the independence of the bridges @xmath244 imply that the second term can be bounded below by @xmath245 } |b_i(s)| \leq \delta/2
\big ) > \eta'>0\ ] ] for some other @xmath246 .
this last fact follows because , by lemma [ lembridgemod ] , the maximum absolute value of a brownian bridge has a continuous distribution on @xmath247 . from the above two bounds
, it follows that @xmath248>\eta\eta'>0 $ ] , as desired .
we also record a fact about brownian bridge which will be useful on several occasions .
[ lembridgemod ] let @xmath200 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath249 , be a brownian bridge .
let @xmath250 .
then , for @xmath251 , @xmath252 the formula appears as @xmath253 in chapter @xmath254 of @xcite .
the following result is also used . recall that ( specializing the notation from definition [ wbdef ] ) we write the measure of one brownian bridge @xmath255 \to { \ensuremath{\mathbb{r}}}$ ] with constraints @xmath256 and @xmath257 as @xmath258 .
[ lembrownbridge ] let @xmath259 .
then @xmath260 \big ) \leq 4 \big(2/\pi\big)^{1/2 } \big ( \delta m \big)^{1/2}.\ ] ] note that @xmath231 \to { \ensuremath{\mathbb{r}}}$ ] under @xmath261 may be represented @xmath262 , where @xmath263 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath264 , is standard brownan bridge .
since @xmath265 for @xmath266 $ ] , @xmath267 \big ) \leq \mathcal{w}_{1;0,0}^{0,1 } \big ( b(s ) > - 2 \delta \ , \
, \forall \ , \ , s \in \big [ 0 , \delta m^{-1 } \big ] \big).\ ] ] we further write @xmath268 for brownian motion @xmath231 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath269 , ( the @xmath270 indicating that the right - hand endpoint value is unspecified ) .
note then that @xmath271 \big ) \nonumber \\ & \leq & \mathcal{w}_{1;0,*}^{0,1 } \big ( b(s ) > - 2 \delta \ , \ , \forall \ , \ , s \in \big [ 0 , \delta m^{-1 } \big ] \big\vert b(1 ) > 0 \big ) \nonumber \\ & \leq & 2 \mathcal{w}_{1;0,*}^{0,1 } \big ( b(s ) > - 2 \delta \ , \ , \forall \ , \
, s \in \big [ 0 , \delta m^{-1 } \big ] \big ) .
\label{eqlembbtwo}\end{aligned}\ ] ] the first inequality follows because standard brownian bridge is stochastically dominated by brownian motion conditioned to have positive end - value .
the second inequality follows from the fact that @xmath272 .
note further that @xmath273 \big ) \nonumber \\ & = & 2 \mathcal{w}_{1;0,*}^{0,1 } \big ( b \big ( \delta m^{-1 } \big ) \leq - 2 \delta \big ) \nonumber \\ & = & 2 \mathbb{p } \big ( n \geq 2 \big ( \delta m \big)^{1/2 } \big ) \geq 1 - 2 \big(2/\pi\big)^{1/2 } \big ( \delta m \big)^{1/2}. \label{eqlembbthree}\end{aligned}\ ] ] the first equality depended on the reflection principle . in the second
, @xmath0 under @xmath274 has the law of a standard normal random variable ; the inequality is due to the density of this distribution being uniformly bounded above by @xmath275 . combining ( [ eqlembbone ] ) , ( [ eqlembbtwo ] ) and ( [ eqlembbthree ] ) , we obtain the statement of the lemma .
we now define the _ dyson _ and _ edge - scaled dyson _ line ensembles which form the sequence of line ensembles whose limit we will consider . brownian bridges and the scaling window of width @xmath6 and height @xmath7 . scaling yields the edge - scaled dyson line ensemble depicted on the right . ]
[ dn ] for each @xmath276 , define a brownian bridge line ensemble @xmath277\rightarrow { \ensuremath{\mathbb{r}}}$ ] such that @xmath278 is equal in law to the limit ( as @xmath279 goes to zero ) of the law of @xmath0 brownian bridges @xmath280 with @xmath281 and conditioned on @xmath282 for each @xmath283 and @xmath284 $ ] .
clearly this line ensemble is continuous and non - intersecting , and furthermore it has the brownian gibbs property .
observe that as the starting and ending points go to zero , the distributions of the height of the @xmath0 lines at @xmath285 converge to a non - trivial limit which can be explicitly calculated via the karlin - mcgregor formula @xcite .
the resulting ensemble on the interval @xmath286 $ ] with this non - trivial entrance and exit law is continuous and non - intersecting and has the brownian gibbs property .
as @xmath287 goes to zero this procedure yields a consistent family of measures which one identifies as the desired line ensemble with starting and ending height all identically zero . ] we define the @xmath0-th _ edge - scaled dyson line ensemble _ @xmath288 ( on the same probability space as @xmath289 ) to be the ordered list @xmath290 where @xmath291 \to { \ensuremath{\mathbb{r}}}$ ] are rescaled dyson lines given by @xmath292 we write @xmath293 for the probability measure associated with the @xmath0-th edge - scaled dyson line ensemble .
we define @xmath294 for @xmath295 and @xmath296 $ ] to be the set of top @xmath28 lines of @xmath288 at times given by @xmath297 . in practice
, we will consider @xmath298 $ ] or @xmath299 .
the law @xmath300 represents the entire edge - scaled dyson line ensemble at times given by @xmath297 .
one should note that the scaling in ( [ ykn ] ) involves centering at @xmath301 .
this is due to the fact ( cf .
@xcite ) that for @xmath302 $ ] , @xmath303 if we fix @xmath304 then @xmath300 is a determinantal point process consisting of @xmath0 distinct points . it is natural to add a parabola from the edge - scaled dyson line ensemble because the limiting shape for the edge has non - zero concavity .
call @xmath305 and similarly extend the above introduced notation .
though we will not make extensive use of this parabolically shifted line ensemble , it is presently convenient . the point process given by @xmath306 may be written in terms of a correlation kernel @xmath307 ; as @xmath0 goes to infinity , this kernel converges in the trace - class norm ( see @xcite for a definition ) to a limiting kernel , known as the airy kernel @xmath308 .
the general theory of determinantal point processes ( cf .
@xcite ) implies that , for any fixed @xmath28 , and for @xmath309 , the probability measure on points in @xmath310 converges to a limiting measure on @xmath28 distinct points . likewise , for @xmath299 and for any fixed @xmath311 , the joint probability measure on points in @xmath310 converges to a limiting measure on @xmath28 distinct points at times in @xmath297 .
these measures are called the _ finite - dimensional distributions of the multi - line airy process_. as the finite set of times @xmath297 is augmented , and likewise as @xmath28 increases , these finite - dimensional distributions form a consistent family , so that kolmogorov s consistency theorem implies the existence of a stochastic process with these marginal distributions .
however , much as in the construction of brownian motion from its finite - dimensional distributions , this implies neither continuity nor any other regularity properties of the process thus constructed .
johansson @xcite considered the top line ( the case that @xmath312 ) and proved that there exists a _ continuous version _ of the above stochastic process on any interval @xmath298 $ ] .
this should be considered as analogous to proving that there exists a continuous version of the brownian motion , which is initially only specified by its finite - dimensional distributions . in the process of proving our main result ( that the brownian gibbs property for @xmath288 passes over to the @xmath4 limit ) , we will need to show also that there exists a version of the multi - line airy process supported on continuous , non - intersecting curves for any interval @xmath298 $ ] .
we will call this the _ airy line ensemble_. the exact nature of the finite - dimensional distributions of the multi - line airy process are , in fact , unnecessary for the statement and proof of our results .
our first result , theorem [ mainthm ] , proves the existence of a continuous , non - intersecting airy line ensemble which , after subtracting a parabola and scaling down by a factor of @xmath313 , has the brownian gibbs property and is the edge scaling limit for non - intersecting brownian bridges .
we then formulate some hypotheses on sequences of line ensembles under which we will prove existence of continuous , non - intersecting limiting line ensembles with brownian gibbs properties . finally , we discuss some of the more general airy - like line ensembles which satisfy these hypotheses . in section [ absconsec ]
we turn to some of the consequences of our results .
[ mainthm ] there exists a unique continuous non - intersecting @xmath314-indexed line ensemble which has finite - dimensional distributions given by the multi - line airy process .
we call this the _ @xmath314-indexed airy line ensemble _ and denote it by @xmath315
. moreover , the @xmath314-indexed line ensemble @xmath316 given @xmath317 for each @xmath318 , has the brownian gibbs property .
additionally , for any @xmath295 and @xmath189 the line ensemble @xmath319)$ ] converges weakly ( definition [ maindef ] ) as @xmath4 to the line ensemble given by @xmath50 restricted to @xmath320 $ ] .
this follows immediately from the more general results of section [ hypsec ] .
proposition [ hypsat ] shows that the edge - scaled dyson line ensemble satisfies the hypotheses of theorem [ mainthm2 ] which in turn proves the above result .
the weak convergence result follows from proposition [ proptightnessh ] which is shown along the way to proving theorem [ mainthm2 ] .
the above proof appeals to a more general set of results which we will soon present in theorem [ mainthm2 ] .
however , we now briefly explain the approach by which we will derive these results .
we start by considering the edge - scaled dyson line ensemble restricted to the top @xmath28 lines in an interval @xmath40 $ ] ( i.e. , @xmath319)$ ] )
. we will prove theorem [ mainthm ] ( and theorem [ mainthm2 ] ) by showing that , as @xmath0 tends to infinity , this sequence converges weakly as a line ensemble to a limiting line ensemble which is continuous , non - intersecting , has finite - dimensional distributions given by the multi - line airy process minus a parabola and scaled down by @xmath313 ( all of which is shown in proposition [ proptightnessh ] ) , and has the brownian gibbs property ( shown in proposition [ propbrowniangibbsh ] )
. once these facts are available , consistency of measures with respect to the interval @xmath40 $ ] and with respect to @xmath28 yields theorem [ mainthm ] .
the key to proving propositions [ proptightnessh ] and [ propbrowniangibbsh ] is proposition [ propacceptprob ] which should be considered to be the main technical component of this article .
this proposition shows that the @xmath28 lines on the interval @xmath40 $ ] remain sufficiently well - behaved uniformly in high @xmath0 .
this is measured in two ways .
the first measure is the _ acceptance probability _ ( see definition [ wbdef ] ) . roughly speaking
, the acceptance probability for a line ensemble on an interval @xmath87 $ ] is the probability that the following operation is accepted : remove the top @xmath28 curves on the entire interval @xmath87 $ ] , and redraw them with the same starting and ending points according to the law of independent brownian bridges .
the outcome is accepted if there is no point of intersection between the resampled curves or with the @xmath321-st curve .
the second measure is the minimal gap between any two of the top @xmath28 lines on an interval @xmath87 $ ] .
proposition [ propacceptprob ] shows that both the acceptance probability and the minimal gap stay bounded from below with high probability as @xmath4 , and that the top @xmath28 curves stay bounded between @xmath322 for @xmath323 large enough , with a uniformly high probability .
the first estimate shows that the top @xmath28 lines on @xmath40 $ ] are uniformly equicontinuous in @xmath0 with high probability ; alongside the known finite - dimensional convergence for @xmath324 ( see section [ airylikesec ] ) , this yields the necessary tightness to prove the weak convergence statement of proposition [ proptightnessh ] .
the minimal gap estimate shows that the limiting measure is supported on non - intersecting lines . that the limiting line ensemble has the brownian gibbs property is then shown by using the skorohod representation theorem to couple the line ensembles for all @xmath0 so as to have uniform convergence of all @xmath28 lines on the interval @xmath40 $ ] .
we reformulate the brownian gibbs property in terms of a resampling procedure and couple the brownian bridges used for this resampling .
given these two layers of coupling , we show that the limiting line ensemble inherits the invariance in law under resampling from the finite @xmath0 ensemble and thus has the brownian gibbs property hence proposition [ propbrowniangibbsh ] . we have not mentioned here how we will go about proving proposition [ propacceptprob ] .
the proof appears section [ technicalsec ] which begins by discussing the ideas involved .
recall that a gibbs measure is said to be extremal if it may not be written non - trivially as a sum of two gibbs measures .
the @xmath314-indexed line ensemble @xmath50 appearing in theorem [ mainthm ] is an example of a brownian gibbs measure which presumably is extremal , as are its affine shifts @xmath325 for @xmath326 .
each element of the family @xmath327 also enjoys the property of being statistically invariant under horizontal shifts in the argument once the parabolic shift is removed . to the best of our knowledge
, it was scott sheffield who first raised the possibility that , in fact , this family of measures exhausts the list of such extremal gibbs measures : [ conj ] we say that an @xmath314-indexed line ensemble @xmath328 is @xmath8-invariant if @xmath329 is equal in distribution to @xmath328 for each @xmath330 .
the set of extremal brownian gibbs @xmath314-indexed line ensembles @xmath50 which have the property that @xmath328 ( given by @xmath331 for @xmath318 ) is @xmath8-invariant is @xmath327 , where @xmath50 appears in theorem [ mainthm ] . beyond its intrinsic interest , this conjecture may be worth investigating in light of its possible use as an invariance principle for deriving convergence of systems to the airy line ensemble ; indeed , we understand that andrei okounkov suggested problems in this direction in a talk in 2006 . as such
, the characterization could serve as a route to universality results .
( see also @xcite for some partial universality results using a different approach involving the komls - major - tusndy coupling of random walks with brownian motion . )
we now formulate some general hypotheses under which we will prove existence of continuous non - intersecting limiting line ensembles with the brownian gibbs property .
[ hypdef ] fix @xmath295 and @xmath189 .
consider a sequence @xmath332 and @xmath333 such that there exists an @xmath334 such that for all @xmath335 , @xmath336 and @xmath337 .
a sequence of line ensembles @xmath338 , @xmath339\rightarrow { \ensuremath{\mathbb{r}}}$ ] satisfies hypothesis @xmath340 if it satisfies the following three hypotheses : * @xmath341 : for each @xmath0 , @xmath60 is a non - intersecting line ensemble with the brownian gibbs property . *
@xmath342 : for every finite set @xmath343 $ ] , the joint distribution of @xmath344 converges weakly .
* @xmath345 : for all @xmath346 and all @xmath347 $ ] there exists @xmath234 and @xmath348 such that for all @xmath349 , @xmath350 if one requires the convergence in @xmath342 to be only for singletons @xmath351 we write @xmath352 instead ; in that case we will refer to all three hypotheses @xmath341 , @xmath352 and @xmath345 as @xmath353 .
fix @xmath295 and @xmath189 .
consider a sequence of line ensembles @xmath338 satisfying hypothesis @xmath353 .
define the minimal gap between the first @xmath28 lines on the interval @xmath87\subseteq [ -t , t]$ ] to be @xmath354 } \big\vert \mathcal{l}^n_i(s ) - \mathcal{l}^n_{i+1}(s ) \big\vert.\ ] ] recall also the acceptance probability @xmath355 given in definition [ wbdef ] .
we can now state the article s main technical component .
note that the reason we require hypotheses @xmath356 is that we need an extra buffer region around @xmath40 $ ] in order to establish sufficient control over the lines in that interval .
[ propacceptprob ] fix @xmath295 and @xmath189 .
consider a sequence of line ensembles @xmath338 satisfying hypothesis @xmath357 .
there exists a stopping domain @xmath358 with @xmath359 and @xmath360 ( almost surely ) such that the following holds . for all @xmath361 ,
there exists @xmath362 and @xmath363 , such that , for @xmath335 , @xmath364 and @xmath365 additionally we have that for all @xmath366 there exists @xmath367 such that , for each @xmath368 , @xmath369 , 1\leq i\leq k \big ) \geq 1-\epsilon.\ ] ] the proof of this proposition is fairly involved and will be given in section [ proof : propacceptprob ] .
we mention in passing that the use of the stopping domain @xmath358 is a technical device needed for the proposition s proof . at least for line ensembles satisfying hypothesis @xmath357 ,
it is a consequence after the fact of the next proposition that ( [ appropeqn ] ) also holds when @xmath370 and @xmath371 are replaced by deterministic times .
given proposition [ propacceptprob ] , we can prove the following two propositions : [ proptightnessh ] fix @xmath295 and @xmath189 .
consider a sequence of line ensembles @xmath338 satisfying hypothesis @xmath356 .
then @xmath372 converges weakly as @xmath4 to a unique limit , which is the continuous non - intersecting line ensemble @xmath373\rightarrow { \ensuremath{\mathbb{r}}}$ ] whose finite - dimensional distributions coincide with the limiting distributions ensured by hypothesis @xmath374 .
[ propbrowniangibbsh ] the line ensemble @xmath375 specified in proposition [ proptightnessh ] has the brownian gibbs property .
combining these propositions yields the following general theorem : [ mainthm2 ] consider a sequence of line ensembles @xmath338 satisfying hypothesis @xmath340 for all @xmath295 and all @xmath189 .
then there exists a unique continuous non - intersecting @xmath314-indexed line ensemble @xmath376 which has finite - dimensional distributions given by those ensured by hypothesis @xmath342 and which has the brownian gibbs property .
we have proved the above result on any finite interval @xmath40 $ ] for the top @xmath28 lines .
we can conclude the existence and uniqueness of the infinite ensemble via consistency .
the brownian gibbs property depends on only a finite number of lines and a finite interval of times , so that it extends to the full @xmath314-indexed line ensemble . accepting proposition [ propacceptprob ] , we may now prove propositions [ proptightnessh ] and [ propbrowniangibbsh ] .
hypothesis @xmath374 ensures convergence of finite - dimensional distributions . to prove weak convergence of line ensembles , it suffices ( in light of theorem 8.1 of @xcite ) to prove tightness .
( we must then additionally prove the non - intersection property of the limiting ensemble . )
the basic idea is to use the estimates of proposition [ propacceptprob ] .
in particular , tightness relies on the fact that having a strictly positive acceptance probability implies that the actual lines in consideration look sufficiently like brownian bridges that their moduli of continuity go to zero in probability .
the proof must cope with the fact that our estimate for acceptance probability is not at the deterministic times @xmath377 but rather in a random stopping domain @xmath378 $ ] with @xmath379 and @xmath380 .
this complication is resolved through a careful application of the strong gibbs property which reduces the problem to a single claim which is then proved via a coupling argument .
the tightness criteria for @xmath28 continuous functions is the same as for a single function . for an interval
@xmath87 $ ] , define the @xmath28-line modulus of continuity @xmath381\\ |s - t|<r } } \big\vert f_i(s)-f_i(t)\big\vert.\ ] ] define the event @xmath382 as an immediate generalization of theorem 8.2 of @xcite , a sequence of probability measures @xmath383 on @xmath28 functions @xmath384 on the interval @xmath87 $ ] is tight if the one - point ( single @xmath385 ) distribution is tight and if for each positive @xmath386 and @xmath387 there exists a @xmath388 and an integer @xmath334 such that @xmath389 note that even when the line ensemble measures replacing @xmath383 are on more than @xmath28 lines , @xmath390 will still refer to the top @xmath28 lines on the interval @xmath87 $ ] . turning to our case ,
the one - point distribution is clearly tight because we have one - point convergence by means of hypothesis @xmath391 .
we wish to prove further that for all @xmath386 and @xmath387 positive , we can choose @xmath388 small enough and @xmath334 large enough so that @xmath392 for all @xmath335 .
observe that , as we are assuming hypothesis @xmath340 , we may apply proposition [ propacceptprob ] .
this ensures the almost sure existence of a stopping domain @xmath393 which contains @xmath40 $ ] , is contained in @xmath394 $ ] and is such that for all @xmath395 there exists @xmath234 and @xmath207 such that for all @xmath349 , @xmath396 where we have abbreviated @xmath397 observe that @xmath398 where @xmath399 we claim that for any @xmath400 there exists @xmath388 , @xmath401 , @xmath367 and @xmath334 such that , for @xmath335 , @xmath402 note that both @xmath403 and @xmath404 are measurable with respect to @xmath405 . with @xmath406 denoting the indicator function , we can thus write the left - hand side of ( [ lhsabove ] ) as @xmath407 \big],\ ] ] where the sigma - field @xmath405 is defined in ( [ defstopdom ] ) . observe that by applying the strong gibbs property given in lemma [ stronggibbslemma ] , @xmath68 almost surely @xmath408 = \mathcal{b}_{\bar{x},\bar{y},f}^{\mathfrak{l}^n,\mathfrak{r}^n}\big[w_{-t , t}(\rho , r ) \big]\ ] ] where , for @xmath409 , @xmath410 , @xmath411 , @xmath412 and where @xmath413 is a function of @xmath155 in the left - hand side and @xmath414 in the right - hand side ( i.e. , these correspond to the choice of the functions @xmath415 in the definition of @xmath413 ) .
* claim : * let @xmath416 , @xmath401 and @xmath367 .
there exists @xmath388 such that , for all @xmath417 $ ] , @xmath418 $ ] , and @xmath419 , @xmath127 \to { \ensuremath{\mathbb{r}}}$ ] satisfying @xmath420 , @xmath421 and @xmath422 , @xmath423 \geq 1-\eta/2.\ ] ] as noted before , @xmath424 is an event depending on the modulus of continuity of the random @xmath28 lines between @xmath40 $ ] specified by the measure @xmath123 ( recall @xmath123 and @xmath126 are given in definition [ wbdef ] and denote the measure and expectation of @xmath28 brownian bridges on @xmath87 $ ] with starting values @xmath134 and ending values @xmath135 ) .
let us assume the claim for the moment .
by choosing @xmath425 small enough ( depending on @xmath426 ) , and using ( [ strongapp ] ) we may bound ( [ condexp ] ) as at least @xmath427 = ( 1-\eta/2){\ensuremath{\mathbb{p}}}^n({\mathfrak{a}}(\mathfrak{l}^n,\mathfrak{r}^n)\geq \delta , s_{k , m}).\ ] ] since we may apply proposition [ propacceptprob ] it follows from ( [ revapeqn ] ) and ( [ maxmineqn ] ) that by choosing @xmath323 large enough and @xmath287 small enough , @xmath428 for these values of @xmath429 , by taking @xmath425 to be small enough so that the above bound for ( [ condexp ] ) holds , we can bound ( [ lhsabove ] ) by @xmath430 which is exactly as desired to prove tightness .
note that the tightness implies almost sure continuity of the limiting line ensemble .
that said , it does not provide the non - intersection condition .
however , this follows immediately from ( [ gappropeqn ] ) of proposition [ propacceptprob ] .
we now finish by proving the claim . first note that if we replace @xmath424 by @xmath390 in the statement of the claim , then by containment of events it follows that proving this modified version of the claim implies the stated claim .
we will make such a replacement and prove the modified version .
consider @xmath431 distributed as independent brownian bridges on @xmath215 $ ] such that @xmath432 and @xmath433 .
we will use these as the basis for a coupling proof of the claim . for each @xmath434 associate the random modulus of continuity @xmath435 .
the process @xmath436 having bounded sample paths , for each @xmath434 this random variable is supported on @xmath437 .
observe that from the standard brownian bridges we can construct the brownian bridges @xmath183 on @xmath87 $ ] of @xmath123 by setting @xmath438 the @xmath28 line modulus of continuity @xmath439 may then be bounded by @xmath440 by assumption @xmath441 . since @xmath189 and
we have assumed @xmath417 $ ] and @xmath418 $ ] it follows that @xmath442 for some constant @xmath443 .
set @xmath444 . by using ( [ modcontr ] ) and the fact that the modulus of continuity for @xmath445 is bounded above by the modulus of continuity for @xmath425 it follows that @xmath446 now observe that conditioning @xmath123 on any event @xmath228 such that @xmath447 is equivalent to conditioning the measure of the @xmath28 brownian bridges @xmath431 on some other event @xmath448 also of measure at least @xmath287 .
the random variables @xmath449 are supported on @xmath437 and converge to zero as @xmath425 goes to zero ( since the brownian bridges are continuous almost surely ) , so that , by choosing @xmath425 small enough ( with all of the other variables fixed ) we can be assured that , conditioned on the event @xmath448 , @xmath450 with probability at least @xmath451 .
since @xmath452 the above reasoning applies presently since we are conditioning on @xmath453 . by the uniformity over @xmath454 in the specified ranges , the claim follows . by proposition [ proptightnessh ] , we know that the limit of the probability measure on @xmath60 is supported on the space of @xmath28 continuous , non - intersecting curves on @xmath40 $ ] .
it follows ( since that space is separable ) that we may apply the skorohod representation theorem ( see @xcite for instance ) to conclude that there exists a probability space such that all of the @xmath455 as well as a limiting ensemble @xmath456 are defined on the space with the correct marginals , and with the property that for every @xmath457 and @xmath458 , @xmath459 in the uniform topology .
we will presently show that for any fixed line index @xmath27 and any two times @xmath460 $ ] with @xmath164 , the law of @xmath456 is unchanged if @xmath461 is resampled between times @xmath462 and @xmath463 according to a brownian bridge conditioned to avoid the @xmath464-st and @xmath465-st lines .
this property is equivalent to the brownian gibbs property .
the argument used for one line clearly works for several lines .
also , we assume ( for simplicity ) that @xmath466 ( i.e. , the curve in question is neither the highest nor the lowest line ) .
we show this by coupling the resampling procedure for all values of @xmath0 . in order to do this
, we perform the resampling in two steps .
we first choose a brownian bridge and paste it in to agree with the starting and ending points ; then we check if it intersects the lines above and below and , if it does not , we accept it . in order to couple this procedure over various values of @xmath0
we fix a single collection of sampling brownian bridges . towards this end , fix a sequence of brownian bridges @xmath467\rightarrow { \ensuremath{\mathbb{r}}}$ ] such that @xmath468 ; our probability space may be augmented to accommodate these independently of existing data .
define the @xmath469-th resampling of line @xmath27 to be @xmath470 where the purpose of the last two terms on the right is to add the necessary affine shift to the brownian bridge to make sure that @xmath471 and @xmath472 .
now define @xmath473 \right\},\ ] ] the index of the first accepted ( non - intersecting ) brownian bridge resampling .
write @xmath474 for the line ensemble with the @xmath475 line replaced by @xmath476 .
( here , @xmath477 stands for resampled . ) given this notation , another way of stating the brownian gibbs property for @xmath60 is that @xmath478 we wish to show that this same property holds for the limiting line ensemble @xmath456 , because that will imply the brownian gibbs property . similarly to ( [ eqavoid ] ) , we set @xmath479 \big\}$ ] . if we can show that almost surely @xmath480 then we are done .
this is because we know both that @xmath455 converges to @xmath481 and that @xmath474 converges to @xmath482 ( all in the uniform topology ) .
since by ( [ restarstar ] ) the laws of @xmath455 and @xmath474 coincide , then so must the laws of their almost sure uniform limits .
this proves that the laws of @xmath481 and @xmath482 coincide and hence proves the brownian gibbs property for @xmath481 .
thus , it remains to prove that @xmath483 converges almost surely to @xmath484 .
we prove two lemmas which together lead to the desired conclusion .
the sequence @xmath485 is bounded almost surely . by proposition [ proptightnessh ] , we know that @xmath481 is composed of non - intersecting continuous functions
. therefore , @xmath486 , j \in \ { i , i+1\ } } \big\vert \mathcal{l}_j^\infty - \mathcal{l}_{j-1}^\infty \big\vert$ ] is almost surely positive .
uniform convergence implies that there exists an @xmath334 such that , for all @xmath335 , the following conditions hold : @xmath487 } \big\vert\mathcal{l}_{i\pm 1}^n(t)- \mathcal{l}_{i\pm 1}^\infty(t)\big\vert<\delta/4.\ ] ] therefore , for @xmath335 , there exists a continuous curve @xmath62 from @xmath488 to @xmath489 which has a neighborhood of radius @xmath490 which does not intersect any curve @xmath491 for @xmath492 and @xmath493 or @xmath494 .
a basic property of wiener measure ( recorded in the present paper as corollary [ coropen ] ) is that any @xmath495-norm neighbourhood of a continuous function has positive wiener measure .
thus there is an almost surely finite @xmath469 such that @xmath496 stays in a neighborhood of @xmath497 about the function @xmath62 . by construction , for @xmath335 , @xmath496 will not intersect @xmath498 .
this shows that almost surely @xmath483 stays bounded as @xmath4 .
[ lemuniquel ] there exists a unique limit point for @xmath499 . clearly , since @xmath485 is bounded , it has some limit points .
let @xmath500 be the minimum of these ( finitely many ) limit points . to see uniqueness
observe that the only way that there could be an infinite sequence of values of @xmath0 for which @xmath501 would be if @xmath502 touches , but never crosses , one of its neighboring lines @xmath503 .
we claim that this event has probability zero .
a basic property of wiener measure ( recorded in the present paper as corollary [ cor : touch ] ) is that a brownian bridge and an independent almost surely continuous function which start and end a positive distance apart will almost surely either cross or never touch and hence will have probability zero of touching without crossing .
observe then that , up to affine shift , @xmath502 is just a brownian bridge that is independent of @xmath503 and , moreover , that it starts and ends a positive distance away from its neighboring lines .
thus the event of touching without crossing has probability zero , and so the lemma is proved . having established that @xmath504 exists , all that remains is to prove that @xmath505 . clearly , @xmath506 , because the only way that @xmath507 could intersect @xmath503 is if it touches this line but does not cross it . however , as we saw in the proof of lemma [ lemuniquel ] , this event happens with probability zero . finally , we argue that @xmath508 by establishing a contradiction otherwise .
assume for the moment that there is an @xmath509 for which @xmath510 does intersect @xmath503 .
then there is a positive distance between these lines .
this implies that , for some sufficiently large @xmath334 , @xmath511 intersects @xmath498 for all @xmath335 .
hence , @xmath512 , in contradiction to the definition of @xmath513 .
hence , we find that @xmath514 and so complete the proof of the proposition .
we now introduce a two - parameter family of brownian bridge line ensembles which satisfy the hypotheses laid out in definition [ hypdef ] .
we will mention shortly some of the contexts in which line ensembles obtained as scaling limits of this family arise .
fix @xmath515 and @xmath516 as well as two vectors @xmath517 and @xmath518 such that @xmath519 and @xmath520 for all @xmath65 .
for each @xmath521 define the @xmath522-perturbed brownian bridge line ensemble @xmath523\to { \ensuremath{\mathbb{r}}}$ ] such that @xmath524 are distributed as brownian bridges conditioned not to intersect with starting points @xmath525 and ending points @xmath526 as before when starting points coincide this is understood as the weak limit of the path measures as the starting points converge together .
clearly these line ensembles are continuous , non - intersecting and have the brownian gibbs property .
we define the @xmath0-th @xmath522-perturbed edge - scaled dyson line ensemble @xmath527 on this probability space to be the ordered list @xmath528 , where @xmath529\to { \ensuremath{\mathbb{r}}}$ ] is given by @xmath530 and , as in the unperturbed case , define the scaled and parabolically shifted line ensemble @xmath531 it is clear that , in the case where @xmath532 , @xmath527 is simply the edge - scaled dyson line ensemble given in definition [ dn ] . when only one of @xmath533 or @xmath534 is non - zero , the resulting ensembles have been studied in @xcite under the name of dyson s non - intersecting brownian motions with
_ a few outliers_. when both @xmath535 , these _ wanderer _ ensembles have been studied in @xcite . in the latter case
, @xcite requires the restriction on the values of @xmath536 and @xmath537 that the straight lines connecting @xmath538 to @xmath539 intersect the @xmath540 axis at points of the form @xmath541 where @xmath542 .
we say that the line ensembles @xmath543 are _ airy - like_. from these results , we find the following . [ hypsat ] consider @xmath544 and @xmath545 ( depending on @xmath0 ) such that either @xmath546 or @xmath547 , or @xmath535 and the values of @xmath536 and @xmath537 are such that the straight lines connecting @xmath538 to @xmath539 intersect points @xmath541 where @xmath542 .
then , for all @xmath295 and @xmath189 , the sequence of line ensembles @xmath527 satisfy hypotheses @xmath340 .
hypothesis @xmath341 follows immediately from the definition of the line ensembles .
the other two hypotheses utilize the determinantal nature of the finite dimensional marginals of the line ensembles , as developed in @xcite . in (
* proposition 3.3 ) , it is proved that for any set of times @xmath548 $ ] , the @xmath297-indexed collection of point processes @xmath549 are determinantal . likewise is true for @xmath550 and
since the two differ by a deterministic shift and scaling , all conclusions below ( which are stated with regards to the second ensemble , holds true for the first as well ) .
what this means is that all of the correlation functions correlation function is given essentially by the probability of finding points in small neighborhoods of @xmath551 for @xmath552 , @xmath553 and @xmath554 . ] for @xmath550 can be written in terms of determinants of a single symmetric kernel @xmath555 . in the proof of (
* theorem 1.2 ) , it is shown that the kernel @xmath556 converges to a perturbation of the airy kernel @xmath557 . by (
* theorem 5 ) , this implies that the point processes @xmath550 converge weakly on cylinder sets to a point process @xmath558 which is still determinantal . from the determinantal structure one
also shows that @xmath559 has a limiting joint distribution described via a fredholm determinant involving the kernel @xmath556 .
this means that @xmath560 has a top - most particle for each @xmath561 .
it is easy to check ( using ( * ? ? ?
* theorem 4.a ) ) that @xmath560 has an infinite number of particles .
this implies in fact that @xmath562 has a limit in distribution as @xmath0 goes to infinity , which shows that hypothesis @xmath342 holds .
in theorem 4.d of @xcite it is shown that in a determinantal point process , the probability of two particles coinciding is 0 . since @xcite proved that @xmath558 is such a point process and since @xmath563 converges weakly to @xmath564 , it follows that for all @xmath395 and @xmath347 $ ] , there exists @xmath234 and @xmath565 such that for all @xmath566 , @xmath567 < { \epsilon}.\ ] ] this proves hypothesis @xmath345 .
the ensemble of non - intersecting brownian exclusions @xcite also satisfies @xmath340 theorem [ mainthm2 ] therefore applies and shows , in particular , uniform convergence of the edge scaled line ensemble to the airy line ensemble .
there exists a variety of other finite line ensembles which do not display the brownian gibbs property , yet which ( under appropriate scaling ) converge in finite - dimensional distribution to airy - like line ensembles .
these include the line ensembles associated with the polynuclear growth ( png ) model , last passage percolation ( lpp ) and tiling problems .
the finite line ensembles do , however , display gibbs properties with respect to different underlying path measures . via an invariance principle , these gibbs properties converge , under the airy rescaling , to the brownian gibbs property .
in fact , if so inclined , one could carry out the program orchestrated in this paper in these non - brownian settings .
however , since the limiting objects ( the airy - like line ensembles ) are the same as the ones that we consider , the only benefit of doing so would be to prove tightness of the finite png , lpp or tiling line ensembles .
similarly to the brownian bridge case , there are simple ways to perturb the png and lpp models ( for instance by modifying boundary conditions or external sources ) which result in perturbations to the line ensembles and hence to their limits @xcite .
the above - mentioned modifications to our program provide a means to prove that the limiting processes associated with these perturbations can actually be realized as the marginals of continuous , non - intersecting line ensembles with the brownian gibbs property .
in fact , a number of these limiting processes coincide with the finite - dimensional distributions of the airy - like ensembles for which proposition [ hypsat ] applies hence by theorem [ mainthm2 ] the existence and brownian gibbs property for these limiting line ensembles is immediate .
theorem [ mainthm ] ( and its more general counterpart theorem [ mainthm2 ] ) has several interesting consequences which are stated here .
the next result holds for any line ensemble with the brownian gibbs property , although for concreteness we state it for the airy line ensemble .
[ propabscon ] for @xmath121 , let @xmath568 denote the @xmath569 line of the airy line ensemble . for any @xmath570 , @xmath571 , the measure on functions from @xmath572\to { \ensuremath{\mathbb{r}}}$ ] given by @xmath573 is absolutely continuous with respect to brownian motion ( of diffusion parameter 2 ) on @xmath572 $ ] . alternatively ,
the measure on functions from @xmath572\to { \ensuremath{\mathbb{r}}}$ ] given by @xmath574 is absolutely continuous with respect to brownian bridge ( of diffusion parameter 2 ) on @xmath572 $ ] .
the absolute continuity with respect to brownian bridge ( of diffusion parameter 2 ) follows immediately from the gibbs property .
the reason for the parameter 2 is due to the @xmath313 factor which arises in relating the airy line ensemble to the ensemble @xmath50 in theorem [ mainthm ] .
when the right endpoint is not fixed , then the absolute continuity is with respect to brownian motion ( of diffusion parameter 2 ) . to show this , we need
only show that the distribution of @xmath575 is absolutely continuous with respect to lebesgue measure .
this can be seen , however , by applying the above shown brownian bridge absolute continuity result for a slightly larger interval @xmath576 $ ] for some @xmath395 .
the brownian bridge absolute continuity on this larger interval then implies that @xmath575 has a density with respect to the lebesgue measure .
proposition [ propabscon ] was anticipated in the paper of prhofer and spohn @xcite ( page 23 ) where the multi - line airy process was introduced in terms of finite - dimensional distributions .
the focus of that paper was not on studying line ensembles in their own right , but rather on the relationship between these ensembles and particular solvable growth models .
shortly thereafter , johansson @xcite studied a closely related growth model and strengthened the convergence results of @xcite in that setting by showing weak convergence of the growth interface to the top line of the multi - line airy process ( called just the airy process ) .
prhofer and spohn s prediction that locally the airy process looks brownian was based on the observation that short time airy process increments had nearly linear variance .
more evidence for this claim was provided by @xcite which showed that the finite - dimensional distributions of the airy process converge , as the time differences go to zero and the process is scaled diffusively , to those of brownian motion .
an immediate corollary of our proposition [ propabscon ] is a stronger version of these results .
let @xmath577 and @xmath295 .
define @xmath578 \to { \ensuremath{\mathbb{r}}}$ ] as the process @xmath579 .
then , as @xmath279 goes to zero , @xmath580 converges weakly ( with respect to the uniform topology on @xmath581,{\ensuremath{\mathbb{r}}})$ ] ) to two - sided brownian motion ( of diffusion parameter 2 ) @xmath582 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath269 .
johansson conjectured that there exists a unique maximization point for the airy process minus a parabola and explained how such a result would lead to a characterization of the transversal fluctuations of directed polymers . as a corollary of our main theorem , we prove that conjecture .
[ propjoh ] let @xmath583 .
then , for each @xmath189 , @xmath584 attains its maximum at a unique point in @xmath40 $ ] , almost surely .
likewise , @xmath584 attains its maximum at a unique point on all of @xmath46 , almost surely .
the first statement follows immediately from three facts : ( 1 ) @xmath585 is absolutely continuous with respect to brownian motion ( of diffusion parameter 2 ) @xmath586 on @xmath40 $ ] ( starting at zero at time @xmath23 ) ; ( 2 ) the law of @xmath587 on @xmath40 $ ] is absolutely continuous with respect to the law of @xmath588 on @xmath40 $ ] ; ( 3 ) brownian motion attains its maximum at a unique point on any given interval .
( this uniqueness also plays an important role in williams decomposition of one - dimensional diffusions @xcite . )
fact @xmath117 follows from proposition [ propabscon ] and fact @xmath589 from girsanov s theorem .
we furnish a brief proof of fact @xmath590 .
observe that for any two non - overlapping deterministic intervals , the maximum of a brownian motion ( of diffusion parameter 2 ) @xmath214 on one interval almost surely differs from the maximum on the other .
this is because , conditioned on the values of @xmath214 at the endpoints of an interval , the maximum attained on the interval has a continuous distribution . for each @xmath591
let @xmath592 denote the partition of @xmath40 $ ] into @xmath593 non - overlapping intervals of equal length .
were the maximal value of @xmath214 on @xmath40 $ ] achieved more than once , there would exist some value of @xmath591 and two indices @xmath594 such that the maxima of @xmath214 on the two intervals @xmath595 and @xmath596 are equal . that this event has probability zero implies fact @xmath590 .
to prove the second statement in theorem [ propjoh ] , we must show that the argmax stays bounded as @xmath24 goes to infinity .
this result is proved below and recorded as corollary [ corneglim ] .
the remainder of this subsection is devoted to proving corollary [ corneglim ] below . recall that we write the airy line ensemble minus a parabola and scaled by @xmath313 as @xmath50 ( i.e. @xmath597 for all @xmath318 and @xmath598 ) .
note that in the notation above , @xmath599 . in theorem [ mainthm ]
we showed the existence of this line ensemble and that it has the brownian gibbs property .
the one - point distribution of the random variable @xmath600 has the tracy - widom ( gue ) distribution . in @xcite
precise bounds are proved for the tails of this probability distribution .
we do not need the full strength of these bounds and thus record a weaker version here .
for some @xmath601 , all @xmath602 and each @xmath603 , @xmath604 the factor of @xmath313 tags along throughout the following argument , but plays no particular consequence .
[ propargmax ] for all @xmath605 , there exists @xmath601 and @xmath606 such that for all @xmath607 and @xmath608 , @xmath609 } 2^{1/2}\mathcal{l}_1(s ) > x \big ) \leq \exp \big\ { -c ( t^2 + x)^{3/2 } \big\}.\ ] ] the proof follows the same strategy that we will use later in establishing lemma [ lemnobigmax ] : if the top curve is ever high , resampling shows that it is likely to be high at some certain fixed time ; but , in the present case , the top curve at any given time has the tracy - widom ( gue ) distribution . in this way , the upper tail in ( [ eqtracywidom ] ) applies not only to each @xmath610 but also to the supremum of this function over a bounded interval of @xmath385-values . for @xmath611 , @xmath612 ,
set @xmath613 .
. for @xmath318 , set @xmath615 .
for each @xmath603 , @xmath610 has the tracy - widom ( gue ) distribution ; hence , there exists @xmath601 such that , for each @xmath318 , @xmath616 let @xmath598 and fix @xmath318 ; take @xmath617 for convenience .
let @xmath618 : 2^{1/2}\mathcal{l}_1(t ) \geq x \big\}$ ] ; if the infimum is taken over the empty - set , we set @xmath619 .
of course , @xmath620 precisely when @xmath621 .
on the event @xmath620 , note that @xmath622 forms a stopping domain in the sense of definition [ defstopdom ] . by the strong gibbs property ( lemma [ stronggibbslemma ] ) ,
almost surely , if @xmath623 occurs , the conditional distribution of @xmath624 \to { \ensuremath{\mathbb{r}}}$ ] , with respect to @xmath625 , is given by brownian bridge ( of diffusion coefficient 2 ) @xmath626 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath627 , @xmath628 , conditioned to remain above the curve @xmath629 .
if @xmath623 occurs then @xmath630 and @xmath631 , and hence the monotonicity lemmas [ monotonicity ] and [ lemmonotonetwo ] imply that the conditional distribution of @xmath632 stochastically dominates the distribution of a brownian bridge on @xmath633 $ ] with endpoint values @xmath8 and @xmath634 . noting that @xmath635 implies that @xmath636 lies to the left of the midpoint of the interval @xmath633 $ ]
, we see that such a brownian bridge exceeds @xmath637 at @xmath636 with probability at least @xmath638 .
the conclusion of the argument which we have presented in this paragraph is that @xmath639 now set @xmath640 .
we require that @xmath641 in order to apply the tail bounds , so this translates into @xmath642 .
given this value of @xmath323 , @xmath643 and given that we have assumed @xmath644 for some fixed @xmath605 , it follows that the above expression exceeds 0 for @xmath27 larger than some explicit constant @xmath606 .
thus we can bound the right hand side in ( [ eqmaxonenew ] ) by using ( [ eqtracywidom ] ) . also using ( [ eqprobvinew ] ) and
the fact that @xmath323 is positive , we find that , for @xmath645 and @xmath646 , @xmath647 the identical bound for negative values of @xmath27 may be similarly obtained . summing ( [ eqmaxoneplusnew ] ) and its negative-@xmath27 counterpart over @xmath318 , @xmath648 , yields the statement of the proposition with a decrease in the value of @xmath601 .
the next result is immediate from proposition [ propargmax ] .
[ corneglim ] we have that @xmath649 @xmath68-almost surely .
write @xmath650 for the set at which @xmath651 attains its maximum value . by corollary [ corneglim ] , @xmath652^c = \emptyset \big ) \to 1 $ ] as @xmath653 .
hence , it suffices to show that the supremum of @xmath651 on each @xmath654 $ ] is attained at a unique point almost surely .
this statement is easy to see for brownian bridge @xmath655 \to { \ensuremath{\mathbb{r}}}$ ] pinned at arbitrary entrance and exit locations .
hence , the required statement follows from proposition [ propabscon ] .
let @xmath656 denotes the @xmath8 value at which the supremum of @xmath657 is attained .
[ ktailcor ] there exists @xmath601 such that , for all @xmath602 , @xmath658 note that the occurrence of @xmath659 } 2^{1/2}\mathcal{l}_1(t ) \leq - x^2/2 $ ] and @xmath660 are sufficient to ensure that @xmath661 .
however , each of these events has probability at least @xmath662 : in the first case and for @xmath8 sufficiently high , this follows from proposition [ propargmax ] ; in the second , from the lower bound in ( [ eqtracywidom ] ) . a decrease in @xmath601
renders the desired inequality valid for all @xmath602 .
we now explain the context and significance of theorem [ propjoh ] . in order to do this ,
we introduce directed polymers , last passage percolation and the png model @xcite . for simplicity
, we focus only on a certain lattice - based version of these models . a _ directed polymer in a random environment _
@xcite is a ( quenched ) probability measure @xmath663 on nearest - neighbor random walks @xmath664 of length @xmath665 with @xmath666 ( we call this set @xmath667 ) : @xmath668 where @xmath669 is the uniform measure on @xmath667 , @xmath670 is the inverse temperature , and for a given path @xmath664 , @xmath671 with @xmath672 i.i.d .
random variables .
the functional @xmath24 models the passage time of the path @xmath664 through a random environment , or , think of @xmath673 as a potential then and @xmath674 is the total potential energy of a path .
the factor @xmath675 is known as a boltzmann weight , and the necessary normalization @xmath676 is the partition function .
taking @xmath677 to infinity , @xmath678 concentrates all of its probability measure on the path @xmath664 with minimal energy or maximal @xmath679 .
this is called the last passage path and the resulting model is called _ last passage percolation_.
two exponents describe the energy and path properties of the directed polymer . defining the free energy by @xmath680 , the exponent @xmath681 represents the size of the free energy fluctuations , i.e. , @xmath682
; the exponent @xmath683 represents the size of transversal fluctuations of the path @xmath664 away from the line @xmath684 , i.e. , @xmath685 .
it has long been predicted in the physics literature that for a wide class of weights ( having sufficient moments ) , @xmath686 and @xmath687 , and in particular that @xmath688 .
the belief that the exponents , and also the limiting statistics for these fluctuations , are independent ( up to certain centering and scaling ) of temperature and of weight distribution is termed kpz universality , after kardar , parisi and zhang @xcite who used earlier ( highly non - rigorous ) calculations of forster , nelson , and stephen @xcite to predict the exponents @xmath689 and @xmath690 ; ( they did not predict the limiting statistics associated with these fluctuations see the review @xcite for more recent works ) .
the first set of mathematical steps towards proving kpz university was made regarding the zero - temperature ( @xmath691 ) polymer with specific weight distributions .
following @xcite , fix @xmath672 to be geometric with parameter @xmath537 ( so that @xmath692 = ( 1-q)q^m$ ] ) and consider the random variable @xmath693 which represents the point - to - point last passage time ( or ground - state energy ) over paths @xmath664 in @xmath667 which are pinned to end at @xmath694 .
define the random process @xmath695 by linearly interpolating the values set by the relation @xmath696 where @xmath697 johansson then showed that , for any fixed @xmath24 , the random function @xmath698\rightarrow { \ensuremath{\mathbb{r}}}$ ] converges weakly ( as a probability measure on @xmath581,{\ensuremath{\mathbb{r}}})$ ] under the uniform topology ) to the random function @xmath699 where @xmath700 is the top line of the airy line ensemble that we have defined in this paper .
( johansson showed that this top line was a well - defined continuous function , but did not treat other lines of the ensemble . )
the airy process had been introduced a year or so earlier by prhofer and spohn ( see also the independent and parallel work of okounkov and reshetikhin @xcite ) who considered a related model , for which they proved convergence of finite - dimensional distributions .
these results also extended the seminal works of baik , deift and johansson @xcite and johansson @xcite which only dealt with one - point convergence .
the upshot of this work is that ( at least for @xmath691 and geometric weights ) the point - to - point free energy of a directed polymer fluctuates on the scale @xmath701 for @xmath686 , and has a non - trivial transversal correlation @xmath702 for @xmath687 and moreover that this entire correlation structure can be described in terms of the airy process .
this picture is believed to hold for all @xmath703 and for all distributions with sufficiently many moments .
this has only been proved ( at the level of the exponent ) for one type of distribution @xcite or for @xmath704 as @xmath705 @xcite in which case one encounters the so - called _ continuum directed polymer _ and the related kpz stochastic pde ( see also @xcite ) .
returning to the original problem of the directed polymer ( with unconstrained endpoint ) , define @xmath706 which is in essence the location of the maximizing path up to time @xmath665 ( the infimum and supremum are due to the discreteness of our weights ) .
likewise , define @xmath656 according to @xmath707 where @xmath584 is as in ( [ hstarstar ] ) .
this is well - defined in light of theorem [ propjoh ] .
johansson was interested in showing that the limit of the law of @xmath708 would coincide with the law of @xmath709 ( a result that he phrased a little differently , because it was not proven that @xmath709 was well defined ) .
this convergence implies that the law of the endpoint of the directed polymer ( at least at zero temperature ) converges to the law of the point at which the maximum of the airy process minus a parabola is achieved . in this direction
, he showed that @xmath710 is a tight sequence and formulated a theorem which states that , under conjecture 1.5 of @xcite ( which is now proved as theorem [ propjoh ] ) , this convergence indeed holds .
thus , we have shown the next theorem .
[ maxdist ] the random variables @xmath708 converge to @xmath709 in distribution as @xmath705 . theorem [ maxdist ] shows that the polymer endpoint fluctuates in the transversal direction on the scale of @xmath711 ( i.e. @xmath687 ) and has a limiting location determined by the location of the maximum of the airy process minus a parabola .
this theorem holds also for lower lines in the airy line ensemble ( so that each curve has a unique maximizing point after a parabola is subtracted ) .
the interpretation of this result in terms of polymers is slightly more obtuse .
johansson considered a multi - line png line ensemble ( as introduced in @xcite ) .
the second line actually records information about the last passage time for two polymer paths which can not touch ( and the @xmath569 line corresponds to @xmath28 such paths ) .
thus one sees that these multi - path polymers have endpoints whose law can be described in terms of the airy line ensemble . given that @xmath709 is a well - defined random variable , one might ask what its distribution is , or , for that matter , what is the joint distribution of @xmath712 . from corollary [ ktailcor ]
one knows that as @xmath8 grows , the probability @xmath713 decays at least as fast as @xmath714 for some constant @xmath715 .
the study of the distribution of @xmath712 has received a lot of attention in particular with a number of new results coming after the present paper was posted . for @xmath0 non - intersecting brownian bridges ( as well as some of the other variants discussed earlier ) ,
the question had previously been studied in physics @xcite and more recently in @xcite and mathematically in @xcite .
a direct approach to computing this joint distribution for the airy line ensemble @xcite relies on taking frchet derivatives of the continuum statistics formulas for the top line of the airy line ensemble ( as developed in @xcite ) and also uses the brownian absolute continuity results provided in this paper in proposition [ propabscon ] .
these approaches yield different formulas which are shown to coincide in @xcite .
these exact formulas have led to exact values of the tail decay constant @xmath716 .
theorem 1.6 of @xcite computes the large time asymptotics for the joint distribution of the airy process ( to fourth order ) .
as the authors mention , the theorem was contingent upon a claimed , though not proved result
here stated as corollary [ propadmoconj ] .
in fact , a complete proof of this theorem was anyway furnished shortly afterwards by widom @xcite using fredholm determinants ; more recently , it has been extended to higher - order precision in @xcite .
[ propposass ] let @xmath577 and @xmath598 .
the conditional distribution of @xmath717 given that @xmath718 stochastically dominates the unconditioned distribution of @xmath717 .
recall from definition [ dn ] the edge - scaled dyson line ensembles @xmath719 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath3 .
let @xmath602 , and write @xmath720 for the event that @xmath721 .
we claim that @xmath288 , @xmath3 , given @xmath720 , converges weakly as a line ensemble to @xmath50 given the event that @xmath722 . to verify this , recall from the proof of proposition [ propbrowniangibbsh ] that a simultaneous construction of these line ensembles may be made so that , for each @xmath69 , @xmath723 converges to @xmath724 locally uniformly .
note that @xmath725 has limit @xmath726 whose value is given by the tracy - widom ( gue ) distribution .
therefore , for @xmath0 large enough ,
we know that the event @xmath720 on which we are conditioning has strictly positive probability ( with a non - zero limit ) . this means we may couple the conditioned systems as @xmath0 variables so as to enjoy the same local uniform convergence of the lines .
this in turn implies the weak convergence of the entire ensemble and verifies the claim made at the start of this paragraph .
note that @xmath288 given @xmath727 stochastically dominates @xmath288 without conditioning .
this is not a consequence of the monotonicity lemmas we have proved , but can easily be proved via the same method as those lemmas ( see section [ proofslemmas ] ) .
we forgo repeating the proof here .
the coupling of conditioned systems shows that @xmath50 given @xmath722 inherits this stochastic domination property .
hence , after putting this back into a statement about @xmath700 , the corollary follows .
we obtain the next result as an immediate consequence of corollary [ propposass ] by noting that the positive association implies @xmath728 [ propadmoconj ] for any @xmath577 , @xmath729 almost surely .
this section is devoted to the proof of proposition [ propacceptprob ] .
let us briefly outline how we proceed .
the claim of equation ( [ maxmineqn ] ) is the easy part of the proof . in order to prove the bound of equation ( [ maxmineqn ] )
, we prove lemmas [ lemnobigmax ] and [ lemnobigmin ] .
we then turn to the claims of equations ( [ appropeqn ] ) and ( [ gappropeqn ] ) which are much more substantial .
the proof strategy begins by noting that , in light of the uniform control provided by equation ( [ maxmineqn ] ) on the @xmath321-st curve on @xmath40 $ ] , with high probability the convex majorant of this curve may have high derivative only on short intervals close to the endpoints @xmath23 and @xmath24 . restricting attention to a slightly shorter domain
, we can be assured that with high probability this concave majorant does not have a large slope ; however , the two @xmath28-decreasing lists which describe the entrance and exit data of the top @xmath28 curves at the ends of this new interval could be very poorly behaved , with clustering of adjacent points , for example . in proposition
[ propconcave ] , we will show how the mutual avoidance constraint causes this entrance and exit data to improve under a further slight shortening of the domain on which the curves are considered . with well - spaced boundary data and a well - behaved concave majorant for the lower curve boundary data , we are then able to obtain the regularity claimed by equations ( [ appropeqn ] ) and ( [ gappropeqn ] ) . as an aside ,
we mention that the effect identified by proposition [ propconcave ] that in a system of mutually avoiding processes , low quality boundary data is rare because it entails that the avoidance condition is likely to be quickly violated is vaguely reminiscent of the separation of arms in planar critical percolation ( see the appendix of @xcite ) and separation of level set contours in two - dimensional random surfaces ( see section 3.5 of @xcite ) . in the proofs in this section , three devices
to which we will often appeal are ( i ) the strong gibbs property ( lemma [ stronggibbslemma ] ) , to argue that we may consider random intervals whose size is measurable with respect to the external sigma - fields ( called stopping domains ) as if there were deterministic ; ( ii ) monotone couplings of non - intersecting brownian bridge ensembles ( lemmas [ monotonicity ] and [ lemmonotonetwo ] ) , to understand the behavior of the full @xmath0 line ensembles in terms of that of @xmath28 line ensembles ; and ( iii ) the brownian gibbs property , to translate regularity of boundary data into regularity of the internal collection of lines .
the proof of the following lemma already illustrates in a small way the guiding theme of the paper : the brownian gibbs property is not merely an expression for the conditional distributions of the line ensemble ; rather , the property offers a valuable probabilistic resampling technique by which to prove regularity of line ensembles which enjoy it .
the method of proof is similar to proposition [ propargmax ] s , which we encountered earlier . in this section
, we will often suppress the superscript @xmath0 in the index of the line ensemble .
[ lemnobigmax ] fix @xmath295 and @xmath189 .
consider a sequence of line ensembles @xmath338 satisfying hypothesis @xmath353 .
write @xmath730 } \mathcal{l}^n_i(s)$ ]
. then for all @xmath395 , there exists an @xmath602 and @xmath731 such that , for all @xmath335 and @xmath173 , @xmath732 it suffices to prove the case @xmath733 . for @xmath734 ,
set @xmath735 .
we will use a resampling argument to show that the event @xmath736 is improbable .
if this event takes place , it means a large maximum is achieved at a random time ; but a resampling will show that in this case it is also likely that the top curve is high at one of a fixed number of deterministic times .
however , the distribution of @xmath737 at fixed times is controlled by hypothesis @xmath352 , which will show that this outcome is unlikely . for @xmath738 , @xmath739 , set @xmath740 by hypothesis @xmath352 ,
for @xmath346 , there exists @xmath741 such that , for all @xmath368 and for each such @xmath237 , @xmath742 now consider @xmath602 and restrict @xmath237 further so that @xmath743 .
let @xmath744 : \mathcal{l}^n_1(t ) \geq x \big\},\ ] ] where , if the infimum is taken over the empty - set , we set @xmath745 .
of course , @xmath746 precisely when @xmath747 .
on the event @xmath746 , note that @xmath748 forms a stopping domain in the sense of definition [ defstopdom ] . by the strong gibbs property of lemma [ stronggibbslemma ] ,
@xmath293-almost surely , if @xmath749 occurs , the conditional distribution of @xmath750 \to { \ensuremath{\mathbb{r}}}$ ] is given by brownian bridge @xmath751 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath752 , @xmath753 , conditioned to remain above the curve @xmath754 .
noting that @xmath755 and that @xmath756 , the monotonicity lemmas [ monotonicity ] and [ lemmonotonetwo ] imply that this conditional distribution stochastically dominates that of an independent brownian bridge on @xmath757 $ ] with endpoint values @xmath8 and @xmath758 . noting that @xmath759 implies that @xmath760 lies to the left of the midpoint of the interval @xmath757 $ ]
, we see that such a brownian bridge exceeds @xmath761 at @xmath760 with probability at least @xmath638 . the conclusion of the argument which we have presented in this paragraph is thus @xmath762 for all @xmath368 . invoking again hypothesis @xmath352
, we choose @xmath602 large enough that the right - hand term in ( [ eqmaxone ] ) is at most @xmath763 for each @xmath368 ; applying ( [ eqprobvi ] ) , we find that @xmath764 technically , we must also justify ( [ eqmaxoneplus ] ) in the case where @xmath765 is replaced by @xmath766 and by @xmath767 . treating @xmath23 and @xmath24 as deterministic times to which we apply hypothesis @xmath352 , the derivation of these two extra bounds proceeds very similarly to the argument that we have given . summing ( [ eqmaxoneplus ] ) over @xmath738 , @xmath743 , and adding its counterparts for the two miscellaneous cases , yields the statement of the lemma .
lemma [ lemnobigmax ] has an analog for the minumum value .
[ lemnobigmin ] fix @xmath295 and @xmath189 .
consider a sequence of line ensembles @xmath338 satisfying hypothesis @xmath353 .
write @xmath768 } \mathcal{l}^n_i(s)$ ]
. then for all @xmath395 , there exists an @xmath602 and @xmath731 such that , for all @xmath335 and @xmath173 , @xmath769 we prove the assertion by induction on @xmath28 .
let @xmath770 and assume the inductive hypothesis at @xmath771 .
( if @xmath312 , define a zeroth curve @xmath772 so that ( [ mindysonprob ] ) trivially holds . ) we must show that ( [ mindysonprob ] ) now holds for @xmath773 as well .
define events @xmath774 and @xmath775 as @xmath776\big\ } , \\
\tilde{e}_{\tilde{x } } & = & \big\ { \mathcal{l}^n_{k}(t ) \geq -\tilde x \ ,
t =- t , t\big\}.\end{aligned}\ ] ] there exists @xmath777 and @xmath778 such that for large @xmath0 , @xmath779 .
this fact follows from the inductive hypothesis , which shows that @xmath780 , and hypothesis @xmath352 , which shows that for @xmath781 large enough , @xmath782 . from now on , we will restrict attention to the portion of the probability space in which the event @xmath783 , @xmath784 , @xmath785 $ ] is distributed as a brownian bridge from @xmath786 to @xmath787 conditioned to not intersect @xmath788 and @xmath789 . if we remove the condition to avoid @xmath789 and call the resulting conditioned ( only on @xmath788-avoidance ) brownian bridge @xmath214
, then it follows from lemma [ monotonicity ] that @xmath790 and @xmath791 can be coupled so that @xmath792 .
if we further replace the starting and ending heights for @xmath586 by @xmath793 , then , since this lowers the starting and ending points , we may use lemma [ lemmonotonetwo ] to couple the resulting brownian bridge @xmath202 ( which is still conditioned to avoid @xmath788 but now starts and ends at @xmath781 ) so that @xmath794 . finally ,
if we replace the condition of avoiding @xmath788 by the condition of staying below @xmath795 , the resulting brownian bridge @xmath796 can be coupled ( again by lemma [ monotonicity ] ) so that @xmath797 .
the output of this string of deductions is that , on the event @xmath228 , there is a coupling of @xmath798\to { \ensuremath{\mathbb{r}}}$ ] with @xmath799\to { \ensuremath{\mathbb{r}}}$ ] such that @xmath800 where @xmath796 is distributed as brownian bridge with @xmath801 conditioned to stay below @xmath795 . by possibly choosing @xmath781 larger , we can ensure that @xmath802 .
given this , the conditioned brownian bridge @xmath796 is not likely to go low . in particular , using lemma [ lembridgemod ] , we readily show that @xmath803\big ) \leq e^{2 } e^{-2y^2/t}.\ ] ] by the monotonicity between @xmath724 and @xmath796 on the event @xmath228 , and the high probability of @xmath228 , we conclude that @xmath804\big ) \leq { \epsilon}/2 + e^{2 } e^{-2y^2/t}.\ ] ] choosing @xmath805 and setting @xmath806 , we find that @xmath807\big ) \leq \frac{{\epsilon}}{2}+\frac{{\epsilon}}{2},\ ] ] which is the result desired to prove the inductive step for @xmath773 . .
the left - hand side represents step 1 and the right - hand side represents step 2.,height=302 ] first observe that equation ( [ maxmineqn ] ) follows immediately from lemmas [ lemnobigmax ] and [ lemnobigmin ] . turning to the proofs of equations ( [ appropeqn ] ) and ( [ gappropeqn ] ) , beyond lemma [ lemnobigmax ] , the central element used here is proposition [ propconcave ] , a result which we will shortly state . in proposition
[ propconcave ] , we consider a @xmath28-curve @xmath133-avoiding line ensemble defined on an interval @xmath40 $ ] .
the result shows that , although the line ensemble may have poorly behaved entrance and exit data ( including points clustered together very closely ) , the curves in the ensemble typically become well - separated on the slightly shorter interval @xmath808 $ ] , provided only that the lower boundary condition @xmath62 has a well - behaved convex hull . a few definitions are needed . for @xmath809 and for @xmath145
, let @xmath810 be the set of functions @xmath811\to { \ensuremath{\mathbb{r}}}$ ] such that its least concave majorant @xmath812\to { \ensuremath{\mathbb{r}}}$ ] satisfies @xmath813 for all @xmath814 $ ] .
for @xmath741 , let @xmath815 be the set of pairs @xmath816 such that @xmath817 fix @xmath295 , @xmath818 and @xmath189 .
consider a line ensemble @xmath819\to { \ensuremath{\mathbb{r}}}$ ] .
define @xmath820 \to { \ensuremath{\mathbb{r}}}$ ] so that @xmath821 is the least concave majorant of @xmath789 restricted to @xmath40 $ ] .
write @xmath822 for the right derivative of @xmath823 \to { \ensuremath{\mathbb{r}}}$ ] . recalling definition [ defstopdom ] , for @xmath824 ,
define a stopping domain @xmath825 by @xmath826 : { \mathfrak{c}}'_{+}(t ) \leq k \big\ } , \\
\tilde{\mathfrak{r}}_k & = & \sup \big\ { t \in \big [ -t , t \big ] : { \mathfrak{c}}'_{+}(t ) \geq - k \big\},\end{aligned}\ ] ] adopting the convention that @xmath827 and @xmath828 .
note that @xmath825 is indeed a stopping domain for lines @xmath155 because this random variable is measurable with respect to @xmath789 .
[ lemellrbd ] fix @xmath295 , @xmath818 and @xmath189 .
consider a line ensemble @xmath819\to { \ensuremath{\mathbb{r}}}$ ] .
for @xmath367 , on the event @xmath829 } \mathcal{l}_1(t ) \leq m \big\ } \cap \big\ { \mathcal{l}_{k+1}(-t ) \geq
-m \big\ } \cap \big\ { \mathcal{l}_{k+1}(t ) \geq -m \big\},\ ] ] we have that , almost surely , @xmath830 the line segment that connects @xmath831 and @xmath832 has slope at least the left - derivative of @xmath833 , since @xmath834 is concave .
this left - derivative is at least @xmath709 , by the definition of @xmath835 .
hence , @xmath836 note that @xmath837 and @xmath838 , the latter since the concave region @xmath839 , y \leq { \mathfrak{c}}(t ) \big\}$ ] has an extreme point at @xmath840 .
hence , ( [ eqctk ] ) holds also when @xmath789 replaces @xmath834 .
applying the bounds provided by the occurrence of @xmath841 , we find that @xmath842 . likewise , the second bound .
recall that @xmath843 denotes the conditional distribution of @xmath28 brownian bridges on @xmath844 $ ] with entrance and exit data @xmath134 and @xmath135 conditioned to meet neither one another nor the curve @xmath62 .
[ propconcave ] fix @xmath845 and two finite intervals @xmath846 for which @xmath847 .
then , for all @xmath361 , there exists @xmath848 such that ,
for @xmath849 , @xmath850 and @xmath851 , @xmath852 and @xmath853 } \big\vert b_i(s ) - b_{i+1}(s ) \big\vert < \delta \big ) < \epsilon.\ ] ] this result is the key to completing our proof of proposition [ propacceptprob ] .
we will use it with @xmath854 replaced by stopping domains ( as is justified by the strong gibbs property ) .
the proof of this result , which will be given in section [ proof : propconcave ] , relies fundamentally on the resampling technique that is at the heart of the brownian gibbs definition .
returning to the task of finishing the proof of proposition [ propacceptprob ] , by a combination of lemma [ lemnobigmax ] and the one - point convergence ensured by hypothesis @xmath352 , we deduce the following : for all @xmath395 , there exists an @xmath731 and @xmath855 such that , for all @xmath335 and @xmath856 , @xmath857 where @xmath841 is given in equation ( [ eevent ] ) ( this also follows immediately from the now proved equation ( [ maxmineqn ] ) ) .
on this event , lemma [ lemellrbd ] ensures that by choosing @xmath858 we have @xmath859 by the strong gibbs property given in lemma [ stronggibbslemma ] , we may apply proposition [ propconcave ] with @xmath854 replaced by the stopping domain @xmath860 . also , replace @xmath279 by @xmath861 , and let @xmath709 and @xmath323 be specified as above , and @xmath862 $ ] , @xmath863 $ ] and @xmath864 . the conclusion ( [ gapeqn ] ) , along with ( [ lrbound ] ) , shows that there exists @xmath287 ( depending on @xmath279 through @xmath865 , and also depending on @xmath28 and @xmath24 ) such that @xmath866 we now obtain ( [ gappropeqn ] ) from ( [ eprob ] ) and ( [ condgap ] ) where @xmath24 is replaced by @xmath867 . to obtain
( [ appropeqn ] ) , set @xmath868 and @xmath869 ; certainly , @xmath870 forms a stopping domain .
taking @xmath871 , note that , by ( [ lrbound ] ) , the occurrence of @xmath841 ensures that @xmath872 and @xmath873 .
note also that , by the combination of lemma [ stronggibbslemma ] and proposition [ propconcave ] , for all @xmath395 , there exists @xmath287 ( with the same dependencies as before ) such that @xmath874 these last two observations combine with ( [ eprob ] ) to give ( [ appropeqn ] ) .
this completes the proof of proposition [ propacceptprob ] . in this section
we will state and prove two lemmas , [ lemellplusone ] and [ eqheps ] , which will , at the end of the section , serve as key inputs in our proof of proposition [ propconcave ] .
fix @xmath865 and intervals @xmath875 as in the statement of the proposition and let @xmath395 . to study the law @xmath876 } \big)$ ] specified in definition [ wbdef ]
, we will condition @xmath877 in two steps : first , on @xmath878 \cup [ r-1,r]}$ ] , and afterwards , also on @xmath879}$ ] .
the principal conclusions of the respective steps will be lemma [ lemellplusone ] and lemma [ eqheps ] .
we start by analysing the law @xmath877 given @xmath878 \cup [ r-1,r]}$ ] , where @xmath880 and @xmath881 .
although the entrance and exit data @xmath175 at times @xmath469 and @xmath425 may include pairs of adjacent points that are extremely close , we will explain how , under the measure in question , this data improves in the interior of @xmath844 $ ] ; we will show that , with a conditional probability which is positive uniformly in such @xmath134 , @xmath135 and @xmath62 , the top @xmath28 curves at times @xmath882 and @xmath883 have risen significantly from the underlying boundary condition and have separated from each other .
fix @xmath295 , @xmath824 , @xmath145 and @xmath850 .
a @xmath28-decreasing list @xmath884 will be called _
@xmath279-well - spaced _ at @xmath885 ( or at @xmath883 ) if @xmath886 ( or @xmath887 ) and if @xmath888 .
we write @xmath889 for the event that @xmath890 is @xmath891-well - spaced at both @xmath892 .
the main result of step 1 is the following .
[ lemellplusone ] there exists an @xmath893 such that for @xmath894 , @xmath895 , @xmath896 , @xmath850 and @xmath851 , @xmath897 \cup [ r-1,r ] } \big ) > \epsilon.\ ] ] inverting the relationship in lemma [ lemellr ] shows that for some @xmath898 , for all @xmath899 there exists @xmath900 ( which depends also on @xmath901 and @xmath902 ) such that the event of the minimal gap between consecutive lines being less than @xmath287 at times @xmath903 or @xmath883 , has probability at least @xmath904 .
note that as @xmath905 , so too can @xmath906 .
lemma [ lembkeps ] shows that there is an @xmath907 such that the event that @xmath908 and @xmath909 holds with probability at least @xmath910 .
now let @xmath911 be chosen such that for all @xmath912 , @xmath913 .
the existence of such an @xmath911 is guaranteed by the decay of @xmath906 as @xmath905 . combining the two bounds given above ( via the union bound )
shows that , for all @xmath912 , @xmath914 \cup [ r-1,r ] } \big ) > \tilde{e}_0-{\epsilon}(\delta ) > \delta.\ ] ] hence by replacing @xmath287 by @xmath279 , the above bound yields the claimed result .
the remainder of step 1 is devoted then to stating and proving lemmas [ lemellr ] and [ lembkeps ] .
this requires some initial understanding of the vectors @xmath890 under the conditioning in the first step .
the next result shows that at times @xmath885 and @xmath883 it is unlikely that the conditioned value of @xmath226 significantly exceeds a certain value . not only will it be useful in the proof of lemma [ lemellr ] but also in the ultimate proof of proposition [ propconcave ] .
[ lemconmax ] there exists a constant @xmath915 such that for all @xmath916 and all @xmath895 , @xmath896 , @xmath850 and @xmath851 , for @xmath917 , @xmath918 \cup [ r-1,r ] } \big ) < \exp \big\ { - c r^2 \big\}.\ ] ] for @xmath919 , set @xmath920 these points have been selected so that @xmath921 and @xmath922 for each @xmath919 and so that @xmath923 } f(t ) \leq x_k ' - \big(r-\ell\big)^{1/2 } = y_k ' - \big(r-\ell\big)^{1/2}.\ ] ] this bound is due to the fact that @xmath880 , @xmath924 and @xmath925 , and is explained in figure [ figconmax ] . for the case @xmath926 .
the bold lines emanating from @xmath927 and @xmath928 have slopes @xmath709 and @xmath929 , and thus lie above @xmath880 .
the lines point of intersection , should it have @xmath8-coordinate in @xmath844 $ ] , lies at most @xmath709 units above either of this pair of points . as such ,
the three dashed rectangular regions in the upper part of the figure are all disjoint from the curve @xmath62 ; each has positive probability of containing a brownian bridge such as those drawn here.,scaledwidth=70.0% ] we claim now that @xmath930 } \big ) \geq \big ( 1 - 2e^{-2 } \big)^k.\ ] ] to verify ( [ eqnclowerbd ] ) , note that ( [ fmaxprime ] ) implies that if the curve @xmath10 does not leave the interval of width @xmath931 centered at @xmath932 , then it will not meet @xmath62 .
similarly , if each curve @xmath933 \to { \ensuremath{\mathbb{r}}}$ ] does not leave the interval of width @xmath931 centered at @xmath934 , then mutual self - avoidance of these curves will be achieved .
calling @xmath935 the probability that the supremum of the modulus of standard brownian bridge on @xmath936 $ ] does exceed @xmath937 ( noter that @xmath935 is a constant by virtue of brownian scaling ) .
the left - hand side of ( [ eqnclowerbd ] ) is at least @xmath938 . by lemma [ lembridgemod ] , @xmath939 and
thus we establish the right - hand side of ( [ eqnclowerbd ] ) .
we now apply the monotonicity lemma [ lemmonotonetwo ] with the choice @xmath940 \cup [ r-1,r]$ ] to conclude the proof . indeed , due to the coupling provided by that lemma , @xmath941 however , @xmath942 to go from the second line to the third involves bounding the numerator from above and the denominator from below .
the denominator is bounded by noting that @xmath943})$ ] and then using ( [ eqnclowerbd ] ) . for the numerator , recall that under the measure @xmath944 , the path @xmath945\to { \ensuremath{\mathbb{r}}}$ ] satisfies @xmath946 and @xmath947 . noting that @xmath948 , the event in question may be realized only if the maximum value attained by @xmath226 exceeds its common endpoint value of @xmath949 by at least @xmath950 ;
hence , lemma [ lembridgemod ] provides the bound on the numerator .
the next lemma was one of the main pieces necessary for the proof of lemma [ lemellplusone ] .
[ lemellr ] for all @xmath346 , there exists @xmath951 such that , for all @xmath895 , @xmath896 , @xmath850 , @xmath851 , and @xmath952 , @xmath953 \cup [ r-1,r ] } \big ) > 1 - { \epsilon}.\ ] ] in the case that @xmath926 .
given the event @xmath954 \cup [ r-1,r]}$ ] and conditionally on all the data specifying the top three curves except for their values at @xmath882 , the scenario that two among these three values are very close to one another will be shown to be very unlikely .
we will establish this by showing that , for each possible way that two points in the triple may be very close , there is a much probable alternative where this is not the case .
this alternative will be found by pushing some elements in the triple upwards , as in the depicted transformation @xmath955 ; the pushing map will be chosen to dilate space , so that close - pair regions in the domain of the map are associated with regions in the range in such a way that the range is much more probable than the domain.,title="fig:",scaledwidth=70.0% ] + the proof of lemma [ lemellr ] will need some further notation , and another result ( lemma [ lemqn ] ) , but the main idea of the proof is straightforward to grasp ( and is illustrated in figure [ figwssketch ] for the case @xmath956 ) . using the decomposition in lemma [ browniandecomp ] , we will condition @xmath877 on all the information specifying the curves @xmath933 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath125 , except for the vector @xmath957 .
the conditioned data generates a sigma - field which we call @xmath958 .
this data consists of @xmath28 pairs of functions , @xmath959 \to { \ensuremath{\mathbb{r}}}$ ] and @xmath960 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath125 , with each of these @xmath961 functions equal to zero at each endpoint ; @xmath962 and @xmath963 are the affine ( and time - translated ) shifts of @xmath183 on @xmath964 $ ] and on @xmath965 $ ] chosen so that the endpoint values vanish ; as such , the data that remains obscure to a witness of @xmath958 is precisely the @xmath28-vector @xmath957 . under @xmath877 conditionally on @xmath958 , this @xmath28-vector is normally distributed .
we are interested in whether the event @xmath966 \cup [ r-1,r]}$ ] occurs .
there is a certain set ( which we will call @xmath967 and formally define in the actual proof of lemma [ lemellr ] ) of choices for this gaussian @xmath28-vector which realize @xmath966 \cup [ r-1,r]}$ ] ; @xmath968 is random and @xmath958-measurable .
we will prove lemma [ lemellr ] by arguing that , for typical choices of the information recorded by @xmath958 , the set @xmath968 does not usually contain pairs of points that are extremely close . to establish this
, we need a definition .
[ closed - disp ] write @xmath969 , and note that @xmath970 . a subset @xmath971 is _ closed under @xmath972-displacement _ if whenever @xmath973 and @xmath974 , then @xmath975 as well .
note that , given the bridge data @xmath962 and @xmath963 for @xmath976 , the set @xmath968 of acceptable choices for the vector @xmath957 is indeed closed under @xmath972-displacement . for this reason , the next lemma is the general statement about @xmath28-tuples of i.i.d .
gaussian random variables that we will invoke to show that clustering of the elements in @xmath968 does not usually occur .
[ lemqn ] let @xmath977 denote a gaussian random variable with independent entries each of variance @xmath978 , and let @xmath979 be closed under @xmath972-displacement . for @xmath980 ,
let @xmath981 denote the conditional distribution of @xmath0 given that @xmath982 .
fix @xmath367 and @xmath983 .
then for all @xmath395 , there exists @xmath984 such that , for @xmath980 satisfying @xmath985 , @xmath986 and @xmath979 closed under @xmath972-displacement , @xmath987 for @xmath988 define @xmath989 by @xmath990 where @xmath991 , with @xmath992 to be specified .
let @xmath993 denote the distribution of @xmath994 .
thus , for @xmath995 , @xmath996 for @xmath997 , set @xmath998 then @xmath999 .
hence , @xmath1000 we estimate the right - hand side of ( [ eqmubeps ] ) by finding a lower bound for @xmath1001 in terms of @xmath1002 . to do so ,
we write @xmath1003 for the density of the gaussian vector @xmath0 , and note that @xmath1004 where @xmath1005 is the jacobian of @xmath1006 at @xmath134 .
note that @xmath1007 .
observe that @xmath1008 note also that , if @xmath1009 , then @xmath1010 for each @xmath1011 and @xmath1012 . hence , for such @xmath134 , @xmath1013 we find from ( [ eqnu ] ) that @xmath1014 using ( [ eqmubeps ] ) , we obtain @xmath1015 we see that , by choosing @xmath1016 , ( [ mueqn ] ) holds as desired
. we will consider only the case @xmath1017 as that of @xmath1018 follows similarly .
we begin by making precise the notion of conditioning on the data in the top @xmath28 curves except for their values at @xmath882 .
recall that , under the law @xmath877 , the curve @xmath933 \to { \ensuremath{\mathbb{r}}}$ ] is distributed as a brownian bridge that travels from @xmath1019 to @xmath1020 .
write @xmath1021 \to { \ensuremath{\mathbb{r}}}$ ] for the affine shift of @xmath183 that is zero at the endpoints , given by the formula @xmath1022 we may apply the decomposition of lemma [ browniandecomp ] for @xmath436 with the choice @xmath1023 and @xmath1024 ( and @xmath238 ) .
as such , we write @xmath1025 here , @xmath1026 are independent gaussian random variables with @xmath1027 and the curves @xmath1028 , @xmath1029 are independent standard brownian bridges of respective durations @xmath117 and @xmath1030 .
let @xmath958 denote the sigma - field generated by @xmath1031 .
informally , the information in @xmath958 is the data described by @xmath933\to { \ensuremath{\mathbb{r}}}$ ] , @xmath976 , lacking the data which determines @xmath1032 , @xmath976 .
define the conditional measure @xmath1033 for an event @xmath297 as the conditional expectation of @xmath1034 with respect to the sigma - field @xmath958 .
with respect to the conditional measure @xmath1035 , define the @xmath958-measurable random subset @xmath1036 as the set of points @xmath1037 such that if @xmath1038 then the event @xmath966 \cup [ r-1,r]}$ ] occurs .
note that almost surely with respect to @xmath1035 , @xmath968 is closed under @xmath972-displacement ( definition [ closed - disp ] ) .
let @xmath1039 denote the successive values at @xmath882 of the linear interpolations between @xmath1040 and @xmath1041 for @xmath125 : @xmath1042 these values are such that @xmath1043 .
let @xmath1044 denote the measure defined in the statement of lemma [ lemqn ] with parameters specified by @xmath1045 in ( [ eqaxyi ] ) , @xmath1046 in ( [ sigmaeqn ] ) , and @xmath968 defined above .
note that @xmath1044 is a @xmath958-measurable random measure that specifies the set of values for the gaussian vector @xmath1047 which would cause the event @xmath966 \cup [ r-1,r]}$ ] to occur .
to summarise , the distribution of @xmath1048 under @xmath1049 \cup [ r-1,r ] } \big)$ ] may be constructed by a two - step procedure .
first , the random measure @xmath1044 is obtained by realizing the data that generates @xmath958 ; and then a random variable @xmath1050 having the law @xmath1044 is constructed .
the distribution of @xmath1048 under @xmath1049 \cup [ r-1,r ] } \big)$ ] has the law of @xmath1051 averaged over the two steps .
we will apply lemma [ lemqn ] to @xmath1044 .
we need to treat separately those parameter choices for this measure such that the measure assigns significant probability to @xmath28-vectors with high first component , because lemma [ lemqn ] is not useful for such measures . to do this ,
we introduce a variable @xmath1052 to be fixed later .
let @xmath1053 denote the @xmath958-measurable event @xmath1054 observe then that we can bound the probability of the event of interest in lemma [ lemellr ] as follows : @xmath1055 \cup [ r-1,r ] } \big ) \leq a_1 + a_2,\ ] ] where @xmath1056 \cup [ r-1,r ] } \big)\ ] ] and @xmath1057 \cup [ r-1,r ] } \big).\ ] ] let us first bound @xmath1058 . write @xmath1059 for the expectation operator of @xmath1060 \cup [ r-1,r ] } \big)$ ] .
then @xmath1061 represents the conditional expectation with respect to the sigma - field @xmath958 . as such
, we may write @xmath1062 \cup [ r-1,r ] } \big ) = \mathcal{\tilde e}\big({\bf 1}_{\tilde{n}_1 > m ' } \big ) = \mathcal{\tilde e } \big ( \mathcal{\tilde e}\big({\bf 1}_{\tilde{n}_1 > m ' } \vert { \sigma}\big ) \big).\ ] ] the random variable @xmath1063 is measurable with respect to @xmath958 and is bounded between @xmath1064 and @xmath117 . multiplying it by the indicator function of @xmath1065 serves only to decrease the full expectation .
however , from the definition of @xmath1044 and of @xmath1053 , we have that , @xmath1060 \cup [ r-1,r ] } \big)$ ] almost surely , @xmath1066 on the event @xmath1067 , @xmath1068 and hence @xmath1069 on the other hand , @xmath1070 \cup [ r-1,r ] } \big ) = { \mathcal{w}_{k;\bar{x},\bar{y}}^{\ell , r}}\big ( b_1(\ell + 1 ) > m ' + a_1 \ , \big\vert \ , { { \rm nc}}^f_{[\ell,\ell + 1 ] \cup [ r-1,r ] } \big).\ ] ] recalling that the absolute value of each component of @xmath134 and @xmath135 is assumed to be at most @xmath323 , it follows that the same is true of the components of @xmath1045 .
hence , lemma [ lemconmax ] implies that the right - hand side of ( [ eqwxyr ] ) is at most @xmath1071 , provided that @xmath1072 is chosen to be high enough ( depending on @xmath1073 ) .
hence , @xmath1074 or @xmath1075 .
note that the bound on which we insist for @xmath1076 enters into the definition of the event @xmath1053 , so that this choice is relevant as we now turn to show that @xmath1077 .
first observe that , in light of ( [ sigmaeqn ] ) and the conditions placed on the intervals @xmath1078 and @xmath902 by proposition [ propconcave ] , there exists a constant @xmath983 such that @xmath1079 for all @xmath895 and @xmath896 .
note also that @xmath851 implies that @xmath1080 .
thus we may apply lemma [ lemqn ] , with the choice @xmath1081 and with @xmath1045 and @xmath1046 as just specified , to find that , @xmath1035 almost surely , @xmath1082 the choice of @xmath287 here depends on @xmath1083 and hence depends on the parameters @xmath1084 as necessary . by the definition of @xmath1053
, it follows that likewise that , @xmath1035 almost surely , @xmath1085 summing over @xmath1086 shows that , restricted to @xmath1053 , @xmath1087 we again appeal to conditional expectations : @xmath1088 note then that @xmath1089 the equality holds due to the relationship between @xmath1090 and the distribution of @xmath1048 under @xmath1049 \cup [ r-1,r ] } \big)$ ] explained in the third paragraph after ( [ eqaxyi ] ) . in this way
, we see that @xmath1091 .
rescaling @xmath279 to absorb the constants , we complete the proof of lemma [ lemellr ] in the case that @xmath1017 ; the case @xmath1018 is analogous .
the other component of the proof of lemma [ lemellplusone ] is the following result .
[ lembkeps ] there exists an @xmath1092 such that for each @xmath895 , @xmath896 , @xmath850 and @xmath851 , @xmath1093 \cup [ r-1,r ] } \big ) > { \epsilon}_0.\ ] ] define @xmath1094 as @xmath1095 and @xmath1096 for @xmath125 .
note that @xmath1097 and @xmath1098 for such @xmath27 , because @xmath1099 ; hence , each @xmath1100 ( or @xmath1101 ) is less than its @xmath8- ( or @xmath540-)counterpart . depicted , lemma [ lembkeps ] analyses three independent brownian bridges on @xmath844 $ ] from @xmath1102 to @xmath1103 conditioned to avoid each other and the curve @xmath62 ; this boundary data is shown in the left - hand sketch .
( labels indicate @xmath540-coordinates except the four lowermost labels in each sketch . )
monotonicities reduce the problem to studying a system of such bridges from @xmath1100 to @xmath1101 , as shown in the right - hand sketch ; considering the event that these bridges remain in the respective corridors @xmath1104 then provides a lower bound on the probability in ( [ eqwxi]).,title="fig:",scaledwidth=80.0% ] + by applications of the monotone couplings of lemmas [ monotonicity ] and [ lemmonotonetwo ] , the probability in ( [ eqellb ] ) will not rise if we replace the function @xmath62 in the conditioning by @xmath161 , nor if we consider the law @xmath944 in place of @xmath877
. we must therefore bound below @xmath1105 \cup [ r-1,r]}^{-\infty } \big).\ ] ] we begin by noting that this quantity is at least @xmath1106}^{-\infty } \big).\ ] ] the desired lower bound for ( [ eqwxi ] ) is now easily seen by consulting figure [ figcorridorsagain ] .
the unit - width corridors @xmath1104 , @xmath125 have disjoint interiors , so that the event in ( [ eqwxi ] ) will happen if each brownian bridge @xmath933 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath1107 , @xmath1108 remains in the corridor @xmath1104 .
we claim that each bridge does so with probability at least @xmath1109 , for some @xmath715 whose only dependence on the parameters is on @xmath709 .
this follows readily by decomposing @xmath933 \to { \ensuremath{\mathbb{r}}}$ ] by splitting its domain @xmath964 \cup [ \ell+1,r-1 ] \cup [ r-1,r]$ ] ( using lemma [ browniandecomp ] with a coordinate change of the domain ) and applying a bound ( from lemma [ lembridgemod ] ) on the maximum modulus of standard brownian bridge ; the argument relies on each of the three slopes of either boundary curve of @xmath1104 being bounded above in absolute value by a constant which depends only on @xmath709 . the first and the third of these slopes are @xmath709 and @xmath929 ; the second , which is @xmath1110 , is in absolute value at most @xmath1111 , because @xmath1112 is implied by @xmath880 , while @xmath1113 by assumption on @xmath1078 and @xmath902 .
this means that the probability of interest ( [ eqwxi ] ) is at least @xmath1114 .
the difference @xmath1115 being at most a constant depending on the pair @xmath1116 , we see that ( [ eqwxi ] ) is bounded below by some @xmath1117 having dependencies as desired in the statement of the lemma we are presently proving .
this bound , in turn , proves the desired lower bound to complete the proof of this lemma .
we now begin the second step of the groundwork for proving proposition [ propconcave ] . in the first step , we learnt in lemma [ lemellplusone ] that there is a uniformly positive probability that the top @xmath28 curves at times @xmath882 and @xmath883 are @xmath279-well - spaced under the law @xmath1118 \cup [ r-1,r]}\big)$ ] . in the second step , we will exploit the as - yet - unused conditioning on @xmath1119}$ ] to improve this inference , finding that , under @xmath1120}\big)$ ] , these curves are in fact well - separated throughout @xmath1121 $ ] with high probability .
this result is achieved by first finding a lower bound ( valid with high probability ) on the acceptance probability of resampling the line ensemble on the interval @xmath1122 $ ] .
let @xmath1123 be the marginal distribution of @xmath1120}\big)$ ] on @xmath28 curves from @xmath1124\cup [ r-1,r]\to { \ensuremath{\mathbb{r}}}$ ] induced by setting @xmath1125 restricted to @xmath964\cup[r-1,r]$ ] .
likewise , let @xmath1126 be the marginal distribution of @xmath1127\cup[r-1,r]}\big)$ ] on @xmath28 curves @xmath1128\cup [ r-1,r]\to { \ensuremath{\mathbb{r}}}$ ] .
we prove lemmas [ lemradnik ] and [ lemfirststep ] which together lead to lemma [ eqheps ] the main result of step 2 .
[ lemradnik ] let @xmath419 and let @xmath811\to { \ensuremath{\mathbb{r}}}$ ] be measurable with @xmath1129 and @xmath1130 .
then @xmath1123 and @xmath1126 can both be considered as measures on the same probability space of continuous curves from @xmath964\cup [ r-1,r]\to { \ensuremath{\mathbb{r}}}$ ] and can be coupled in such a way that the radon - nikodym derivative of @xmath1123 with respect to @xmath1126 is given by @xmath1131 the normalization constant @xmath1132 is given by @xmath1133 this follows immediately from definitions .
[ lemfirststep ] there exists an @xmath1134 such that , for all @xmath1135 and all @xmath895 , @xmath896 , @xmath850 and @xmath851 , @xmath1136 we claim that , on the event @xmath889 , @xmath1137 for some constant @xmath1138 .
figure [ figcorridors ] illustrates the derivation of this claim . to bound below the acceptance probability in ( [ eqacprob ] )
, we will construct a disjoint collection of width-@xmath279 corridors @xmath1139 , @xmath125 , all of which lie above the curve @xmath62 .
set @xmath1140 to be the event that @xmath1139 contains the range of an independent brownian bridge starting at @xmath885 with value @xmath1032 and ending at @xmath883 with value @xmath1141 , for all @xmath409 .
we will find a lower bound on the probability of @xmath1140 valid for each @xmath125 . to define the corridor @xmath1139
, we consider two cases . for @xmath125 , the index @xmath27 is called _ close _ if @xmath1142 .
if an index @xmath27 is close , we will define the domain @xmath1139 to be a corridor with a single kink located so as to insure that it stays strictly above the function @xmath62 ; whereas if @xmath27 is not close , then @xmath1139 will be an unkinked corridor . for @xmath125 , we define the corridor - centre function @xmath1143 \to { \ensuremath{\mathbb{r}}}$ ] . if index @xmath27 is close , set @xmath1144 for @xmath1145 $ ] , and note that the range of @xmath1146 is a polygonal path comprising two segments in this case . if index @xmath27 is not close , then set @xmath1147 .
in either case , set @xmath1148 , @xmath1149 it is readily confirmed that , for @xmath1150 , the occurrence of @xmath1151 ensures that the @xmath28 corridors @xmath1139 are pairwise disjoint and that their union does not meet the curve @xmath62 . that @xmath62 is avoided is insured by the bound of @xmath709 on the convex majorant of @xmath62 and the kinking of corridors associated with close @xmath27 s .
if @xmath27 is not close , then the probability of @xmath1140 is simply the probability that the maximum modulus of a standard brownian bridge of duration @xmath1152 is at most @xmath279 .
if @xmath27 is close then , writing @xmath1153 $ ] for the @xmath8-coordinate of the kink in the path @xmath1146 , this lower bound may be expressed using the brownian decomposition lemma [ browniandecomp ] in terms of a normal random variable of variance @xmath1154 and two brownian bridges of durations @xmath1155 and @xmath1156 .
lemma [ lembridgemod ] is used to control the deviation of these bridges from linear motion . in either case
, we find that the probability of @xmath1140 is at least @xmath1157 , with @xmath1158 . in the case that @xmath926 , and each index @xmath1159 is close .
the bold lines indicate curves realizing the event @xmath889 .
the three corridors are of width @xmath891 , and , with conditional probability at least @xmath1160 , each will contain a brownian bridge between the midpoints of the intervals that form the ends of each corridor , as illustrated by the dotted curves in the figure.,title="fig:",scaledwidth=70.0% ] + to complete the argument , observe that , by lemma [ lemellplusone ] and the definition of @xmath1126 , there exists @xmath1161 such that for all @xmath1162 , @xmath1163 .
this implies that @xmath1164 and hence completes the proof .
we now state the main result of step 2 .
[ eqheps ] there exists an @xmath1134 such that for all @xmath1162 and all @xmath895 , @xmath896 , @xmath850 and @xmath851 , @xmath1165 the probability in question is the same as under @xmath1166 , because the acceptance probability in question is measurable with respect to the curves @xmath1167 \cup [ r-1,r ] \to { \ensuremath{\mathbb{r}}}$ ] , @xmath125 .
we use lemma [ lemradnik ] to find that @xmath1168 where @xmath1169 the inequality in ( [ eqzxyf ] ) is due to lemma [ lemfirststep ] .
hence , as desired , @xmath1170 we may now present the proof of proposition [ propconcave ] which , when shown , completes the proof of proposition [ propacceptprob ] .
choosing @xmath1171 in lemma [ eqheps ] , we obtain the assertion ( [ apeqn ] ) regarding acceptance probabilities . obtaining the assertion ( [ gapeqn ] ) about the minimal gap
requires a little more effort but is closely related to ( [ apeqn ] ) . roughly speaking
, we wish to show that , if the minimal gap becomes very small , so must the acceptance probability , so as to obtain ( [ gapeqn ] ) from ( [ apeqn ] ) .
to argue this , we will use lemma [ lembrownbridge ] to bound hitting probabilities for a brownian bridge .
note that the law @xmath1172 can be formed in two steps : first by realizing its marginal @xmath1123 on @xmath1173 \cup [ r-1,r]$ ] , and second by realizing its conditional marginal on @xmath1121 $ ] . for @xmath1174 , define the event @xmath1175 ( in terms of data available after the first step
i.e. , the sigma - field generated by the projection onto @xmath1173 \cup [ r-1,r]$ ] ) : @xmath1176 recalling that we have assumed that @xmath895 , @xmath1177 ( with @xmath847 ) , @xmath850 and @xmath851 , note that lemmas [ lemconmax ] and [ eqheps ] imply that , for any given @xmath346 , there exists @xmath1178 large enough such that @xmath1179 in fact , lemma [ lemconmax ] only bounds the probability of @xmath1180 .
we claim that the probability of the event @xmath1181 can easily be bounded as going to zero as @xmath1076 increases by appealing to the monotonicity result of lemma [ monotonicity ] . indeed ,
if the function @xmath62 were replaced by @xmath1182 , the claim would be immediately true , and by monotonicity , it is likewise true for any @xmath1183 .
let @xmath1184 denote the midpoint of the interval @xmath1185 $ ] .
we now adopt the shorthand that , on the probability space on which @xmath1172 is defined , @xmath1186 and @xmath1187 denote @xmath1188 and @xmath1189 .
suppose now that @xmath1186 and @xmath1187 are such that @xmath1175 is satisfied .
let @xmath992 be a parameter whose value will be set later .
for each @xmath1190 , let @xmath1191 $ ] denote the infimum of those @xmath1192 $ ] such that @xmath1193 ; if no such point exists , we formally set @xmath619 .
then on the event @xmath1194 , @xmath1195 is a stopping domain ( for the top @xmath28 curves ) under the independent brownian bridge law @xmath1196 on the curves @xmath1197 .
observe that , given @xmath620 and the trajectories @xmath1198 , the conditional probability that @xmath1199 for all @xmath1200 $ ] , is equal to @xmath1201 \big).\ ] ] this expression naturally arises by considering the difference @xmath1202 ; the factor of two appearing in the expression for the duration arises because taking the difference of brownian motions doubles the variance of the process . noting that @xmath1203 ( as we assumed @xmath1204 is such that @xmath297 can be satisfied ) and @xmath1205 by assumption , an evident stochastic domination shows that the probability ( [ eqwdelta ] ) is at most @xmath1206 \big).\ ] ] lemma [ lembrownbridge ] implies that this quantity is at most @xmath1207 .
combining these bounds and invoking the strong gibbs property of lemma [ stronggibbslemma ] , we learn that @xmath1208 by the definition of @xmath1209 , this means that @xmath1210 \textrm { such that $ b_i(t ) - b_{i+1}(t ) \leq \delta$ } \big ) \leq 8 \pi^{-1/2 } \big ( \delta m ' \big)^{1/2}.\ ] ] similarly , we find that @xmath1211 \textrm { such that $ b_i(t ) - b_{i+1}(t ) \leq \delta$ } \big ) \leq 8 \pi^{-1/2 } \big ( \delta m ' \big)^{1/2}.\ ] ] note that if @xmath1212 are three events in a probability space @xmath1213 such that @xmath1214 and @xmath1215 , then @xmath1216 .
this fact may be employed in concert with ( [ eqwxyone ] ) and ( [ eqwxytwo ] ) to yield that @xmath1217 \textrm { such that $ b_i(t ) - b_{i+1}(t ) \leq \delta$ } \big ) \leq 16 \pi^{-1/2 } \big ( \delta m ' \big)^{1/2}.\ ] ] on the event @xmath1175 , we have that ( from the acceptance probability bound ) @xmath1218 hence , @xmath1219 \textrm { such that $ b_i(t ) - b_{i+1}(t ) \leq \delta$ } \ , \big\vert \ , { { \rm nc}^f}_{[\ell + 1,r-1 ] } \big ) \leq 16 \pi^{-1/2 } \big ( \delta m ' \big)^{1/2 } \epsilon^{-3}.\ ] ] taking @xmath1220 and summing over @xmath1221 , we conclude that , should @xmath1175 occur , @xmath1222 } \big ( b_i(s ) - b_{i+1}(s ) \big ) \leq ( m')^{-1 } \tfrac{\epsilon^8}{2 ^ 8 \pi ( k-1)^2 } \ , \big\vert \ , { { \rm nc}^f}_{[\ell + 1,r-1 ] } \big ) \leq \epsilon.\ ] ] let us conclude the proof now .
the law @xmath1172 can be realized by first sampling from its marginal @xmath1166 on @xmath964\cup [ r-1,r]$ ] and then sampling the remaining path on @xmath1223 $ ] from the measure @xmath1224}\big)$ ] where @xmath1186 and @xmath1187 are determined from the first step as @xmath1225 and @xmath1226 . by ( [ eqprobaone ] )
, the event @xmath297 fails to occur under @xmath1166 with probability at most @xmath1227 . on the event
@xmath297 we may appeal to ( [ eqwxyprime ] ) . combining these two facts yields the bound @xmath1228 } \big ( b_i(s ) - b_{i+1}(s ) \big ) \leq ( m')^{-1 } \frac{\epsilon^8}{2 ^ 8 \pi ( k-1)^2 } \bigg ) \leq 3 \epsilon.\ ] ] we have obtained ( [ gapeqn ] ) and have thus completed the proof of proposition [ propconcave ] .
the proofs of the monotonicity lemmas are based on an analogous result for non - intersecting random walks which can be proved by coupling markov chains .
the proof of the strong gibbs property is a fairly straightforward extension of the usual proof of the strong markov property .
the approach used in proving these two lemmas is identical so we will demonstrate it in the case of lemma [ monotonicity ] and briefly explain how it applies equally well to lemma [ lemmonotonetwo ] .
we demonstrate this coupling for brownian motion by exhibiting an analogous coupling of constrained simple symmetric random walks which then converge to the brownian motions .
the result for the random walks follows from a coupling of monte - carlo markov chains which converge to the constrained path measures and hence demonstrates the coupling .
consider @xmath1229 and @xmath1230 such that @xmath1231 for all @xmath976 and @xmath1232 and such that @xmath1233 as well as @xmath1234 .
write @xmath1235 for the law of @xmath28 independent random walks @xmath1236 \to n^{-1 } { \ensuremath{\mathbb{z}}}$ ] , @xmath125 , that satisfy @xmath1237 and @xmath1238 and take steps of size @xmath1239 in discrete time increments of size @xmath1240 .
define @xmath1241 as the law of @xmath1242 } \big)$ ] ( i.e. , the probability distribution of @xmath28 random walk bridges @xmath1243 , @xmath1244\rightarrow n^{-1}{\ensuremath{\mathbb{z}}}$ ] with @xmath1245 conditioned on the event of non - intersection with each other and the barrier @xmath62 ) .
we will now demonstrate that if @xmath1246 then there is a coupling measure @xmath1247 on which @xmath1248 and @xmath1249 are defined with marginal distributions @xmath1241 and @xmath1250 under which almost surely @xmath1251 for all @xmath27 and all @xmath165 $ ] . as @xmath705 , the invariance principle for brownian bridge shows that this coupling measure converges to the desired coupling of brownian bridges necessary to complete our proof .
thus it remains to demonstrate the existence of @xmath1247 . in order to do this ,
we introduce a continuous - time markov chain dynamic on the random walk bridges @xmath1248 and @xmath1249 .
let us write the collection of random walk bridges at time @xmath385 as @xmath1252 and @xmath1253 .
the time @xmath1064 configurations of @xmath1254 and @xmath1255 are chosen to be the lowest possible trajectories of random walk bridges conditioned to touch neither each other nor the barriers @xmath62 and @xmath1256 ( respectively ) .
it is clear that such a lowest collection of paths exists and is unique and that these initial trajectories are ordered @xmath1257 for all @xmath165 $ ] and @xmath173 .
the dynamics of the markov chain are as follows : for each @xmath1258 $ ] , each @xmath173 and each @xmath1259 ( @xmath1260 stands for up flip and @xmath1261 for down flip ) , there are independent exponential clocks which ring at rate one .
when the clock labeled @xmath1262 rings , one attempts to change @xmath1263 to @xmath1264 .
this is the only change at that instant attempted , and it is only successful if the change can be made without disturbing the condition of non - intersection of the random walks with themselves and @xmath62 . likewise , according to the _ same _ clock , one attempts to change @xmath1265 to @xmath1266 , and the same conditions apply .
the first key fact ( which is readily checked ) is that these time dynamics preserve ordering , so that , for all @xmath1267 , @xmath1268 for all @xmath165 $ ] and @xmath173 .
the second key fact is that the marginal distributions of these time dynamics converge to the invariant measure for this markov chain , which is given by the measures we have denoted by @xmath1269 and @xmath1270 .
this fact follows since we have a finite state markov chain which is irreducible with obvious invariant measure . combining these two facts
implies the existence of the coupling measure @xmath1247 as desired .
the proof of lemma [ lemmonotonetwo ] also relies on the same markov chain which preserves ordering .
the ordering of the initial data is a direct consequence of the ordering of the starting and ending points , and hence the same argument carries over .
the right - hand side of equation ( [ strongeqn ] ) is @xmath1271-measurable .
consider functionals @xmath1272 which map line ensembles @xmath50 to @xmath46 . to prove our desired result it suffices to show that for all @xmath1272 which are borel measurable with respect to @xmath1273 , @xmath1274 = { \ensuremath{\mathbb{e}}}\big[h(\mathcal{l}_1,\ldots , \mathcal{l}_n)\cdot { \mathcal{b}_{\bar{x},\bar{y},f}}^{\mathfrak{l},\mathfrak{r}}\big[f(\mathfrak{l},\mathfrak{r},b_1,\ldots , b_k)\big]\big].\ ] ] by the monotone class theorem , it suffices to check equation ( [ stareqn ] ) for functions @xmath156 of the form @xmath1275 where @xmath1276 and @xmath1277 are bounded continuous functions , @xmath1278 , and @xmath1279 $ ] .
we will also make use of the continuity of expectations with respect to boundary conditions .
in particular , consider @xmath127\to { \ensuremath{\mathbb{r}}}$ ] fixed and @xmath1280 , @xmath1281 , @xmath1282 and @xmath1283 converging to @xmath134 , @xmath135 , @xmath469 and @xmath425 .
then for @xmath156 , @xmath1284 = \mathcal{b}_{\bar{x},\bar{y},f}^{\ell , r}\big[f(\ell , r , b_1,\ldots , b_k)\big].\ ] ] we will approximate the stopping domain @xmath143 via a sequence of stopping domains @xmath1285 which are defined by @xmath1286 where @xmath237 is such that @xmath1287 ( if no such @xmath237 exists then @xmath1288 ) ; and @xmath1289 where @xmath1290 ( if no such @xmath237 exists then @xmath1291 ) .
it is clear that @xmath1285 form stopping domains which converge to @xmath143 . on the event @xmath297 , from the convergence of the stopping domains and the continuity of the @xmath1295 , ( [ continuityeqn ] ) holds with @xmath1296 replaced by @xmath1297 , and with @xmath1298 , @xmath1299 and @xmath1300 or @xmath161 if @xmath162 .
@xmath1301\cdot{\bf 1}_{a}\big]\\ & = & { \ensuremath{\mathbb{e}}}\big[h(\mathcal{l}_1,\ldots , \mathcal{l}_n)\cdot \lim_{n\to \infty } \mathcal{b}_{\bar{x}_n,\bar{y}_n , f}^{\mathfrak{l}_n,\mathfrak{r}_n}\big[f(\mathfrak{l}_n,\mathfrak{r}_n , b_1,\ldots , b_k)\big]\cdot { \bf 1}_a\big]\\ & = & \lim_{n\to \infty } { \ensuremath{\mathbb{e}}}\big[h(\mathcal{l}_1,\ldots , \mathcal{l}_n)\cdot \mathcal{b}_{\bar{x}_n,\bar{y}_n , f}^{\mathfrak{l}_n,\mathfrak{r}_n}\big[f(\mathfrak{l}_n,\mathfrak{r}_n , b_1,\ldots , b_k)\big]\cdot{\bf 1}_a\big ] . \ ] ] by the law of total probability and fubini s theorem , @xmath1302\cdot { \bf 1}_a\big]\\ & = & \sum_{-\infty <
i < j<\infty } { \ensuremath{\mathbb{e}}}\big[h(\mathcal{l}_1,\ldots , \mathcal{l}_n)\cdot \mathcal{b}_{\bar{x}_n,\bar{y}_n , f}^{\mathfrak{l}_n,\mathfrak{r}_n}\big[f(\mathfrak{l}_n,\mathfrak{r}_n , b_1,\ldots , b_k)\big ] \cdot { \bf 1 } \{\mathfrak{l}_n=(i+1)2^{-n } , \mathfrak{r}_n= j2^{-n}\}\big].\end{aligned}\ ] ] then using the brownian gibbs property on each term in the summation above , it follows that the above is equivalent to @xmath1303 \\ & = & { \ensuremath{\mathbb{e}}}\big[h(\mathcal{l}_1,\ldots , \mathcal{l}_n ) \cdot f\big(\mathfrak{l}_n,\mathfrak{r}_n,\mathcal{l}_1\big\vert_{(\mathfrak{l}_n,\mathfrak{r}_n)},\cdots , \mathcal{l}_k\big\vert_{(\mathfrak{l}_n,\mathfrak{r}_n)}\big ) { \bf 1}_a\big].\end{aligned}\ ] ] by our choice of @xmath1304 , the mapping @xmath1305 is continuous almost surely with respect to the measure @xmath68 on @xmath50 .
therefore , @xmath1306 converges to @xmath1307 on the event @xmath297 . by the dominated convergence theorem
, this implies that @xmath1308\\ & = & { \ensuremath{\mathbb{e}}}\big[h(\mathcal{l}_1,\ldots , \mathcal{l}_n ) \cdot f\big(\mathfrak{l},\mathfrak{r},\mathcal{l}_1\big\vert_{(\mathfrak{l},\mathfrak{r})},\ldots , \mathcal{l}_k\big\vert_{(\mathfrak{l},\mathfrak{r})}\big ) { \bf 1}_a\big ] .\end{aligned}\ ] ] this completes the proof of the strong brownian gibbs property .
j. pitman , m. yor .
decomposition at the maximum for exclusions and bridges of one - dimensional diffusions . in _
it s stochstic calculus and probability theory .
n. ikeda , s. watanabe , m. fukushima and h. kunita .
springer - verlag 293310 ( 1996 ) .
|
consider a collection of @xmath0 brownian bridges @xmath1 \to { \ensuremath{\mathbb{r}}}$ ] , @xmath2 , @xmath3 , conditioned not to intersect .
the edge - scaling limit of this system is obtained by taking a weak limit as @xmath4 of the collection of curves scaled so that the point @xmath5 is fixed and space is squeezed , horizontally by a factor of @xmath6 and vertically by @xmath7 .
if a parabola is added to each of the curves of this scaling limit , an @xmath8-translation invariant process sometimes called the multi - line airy process is obtained .
we prove the existence of a version of this process ( which we call the airy line ensemble ) in which the curves are almost surely everywhere continuous and non - intersecting .
this process naturally arises in the study of growth processes and random matrix ensembles , as do related processes with `` wanderers '' and `` outliers '' .
we formulate our results to treat these relatives as well .
note that the law of the finite collection of brownian bridges above has the property called the brownian gibbs property of being invariant under the following action .
select an index @xmath9 and erase @xmath10 on a fixed time interval @xmath11 ; then replace this erased curve with a new curve on @xmath12 according to the law of a brownian bridge between the two existing endpoints @xmath13 and @xmath14 , conditioned to intersect neither the curve above nor the one below .
we show that this property is preserved under the edge - scaling limit and thus establish that the airy line ensemble has the brownian gibbs property
. an immediate consequence of the brownian gibbs property is a confirmation of the prediction of m. prhofer and h. spohn that each line of the airy line ensemble is locally absolutely continuous with respect to brownian motion .
we also obtain a proof of the long - standing conjecture of k. johansson that the top line of the airy line ensemble minus a parabola attains its maximum at a unique point .
this establishes the asymptotic law of the transversal fluctuation of last passage percolation with geometric weights .
our probabilistic approach complements the perspective of exactly solvable systems which is often taken in studying the multi - line airy process , and readily yields several other interesting properties of this process . [ section ] [ section ] [ theorem]conjecture [ theorem]lemma [ theorem]proposition [ theorem]corollary [ theorem]claim [ theorem]experimental result # 1 # 1#1 # 1 [ theorem]remark [ theorem]comment [ theorem]definition [ theorem]definitions [ theorem]conjecture
|
we dedicate our contribution to the memory of george stell with whom the loup verlet s orsay group has shared a long - lasting friendship .
this study is devoted to a monte - carlo ( mc ) simulation of a two - dimensional ( 2d ) system made of identical dipolar hard disks ( dhd ) in the euclidian plane @xmath2 in the presence of an uniform external electrostatic field @xmath3 .
the dipoles are assumed to be permanent and the configurational energy of @xmath4 dipolar molecules in @xmath2 reads as @xmath5 \nonumber \\ & - & \mu \sum_{i = 1 } ^{n } \mathbf{s}_i \cdot \boldsymbol{\mathcal{e } } \nonumber \\ \ ; .\end{aligned}\ ] ] in eq . , @xmath6 is the hard disk potential of diameter @xmath7 .
the second term is the contribution from the 2d dipole - dipole interaction where @xmath8 , @xmath9 permanent dipole moment , @xmath10 unit vector in the direction of the dipole moment of particle @xmath11 , @xmath12 , the vector joining the centres of mass of the particles , and @xmath13 .
finally , the last term denotes the interaction energy of the dipoles with the external field @xmath0 which is assumed to be uniform .
we note that the dipole - dipole interaction involved in eq . is derived from the solution of the 2d laplace equation in the plane .
this model was studied recently in the absence of @xmath0 by means of mc simulations performed on a sphere and in a square with periodic boundary conditions ( pbc ) @xcite . in both cases the laws of electrostatics governing the interactions in the considered geometries
were scrupulously adopted yielding identical phase diagrams . at high temperature and moderate density an ordinary polar fluid
is observed characterized by a dielectric constant @xmath14 . in the low temperature , low density part of the phase diagram
a phase of living polymers of dipoles organized into closed rings has been observed . at higher density the structure of this phase
is characterized by an entangled structure of chains and rings and the dielectric constant @xmath14 is no more well - defined .
it was found that the critical dipole moment @xmath15 at the transition from fluid to polymeric phase increases slightly with density . in the presence of an applied field , in the low temperature phase
the dipole moments are expected to align in the direction of the field , the tendency getting more effective as the strength of @xmath0 increases , first yielding roughly linear chains which ultimately will collapse into bundle - like structures at high density . as it is impossible to consider uniform fields on the sphere ( without violating the laws of electrostatics , cf .
@xcite ) we present here only simulations performed of dhd contained in a square with pbc .
let us first recall some elements of electrostatics for a square of side @xmath17 along directions @xmath18 with pbc , which will be referred to as space @xmath19 @xcite . the electric field at point @xmath20 ( @xmath21 $ ] , @xmath22 orthonormal basis of @xmath19 ) created by a point dipole @xmath23 located at point @xmath24 of @xmath19 is given by @xmath25 where @xmath26 denotes the bare dipolar green s function and the dot a tensorial contraction ( see _ e.g. _ refs .
@xcite ) .
the green s function @xmath26 depends on the considered geometry and is given in @xmath19 by @xmath27 where @xmath28 is the periodic ewald potential .
the latter satisfies poisson s equation in @xmath19 , _
i.e. _ @xmath29 \ ; , \ ] ] where @xmath30 is the periodical dirac s comb . in eqs .
@xmath31 where @xmath32 is a @xmath16 vector with integer components @xmath33 in the basis @xmath22 .
one notes that @xmath28 identifies with the potential created by a unit point charge immersed in a uniform neutralizing background of charge density @xmath34 .
expanding @xmath28 and @xmath35 in fourier series one finds @xmath36 where @xmath37 .
some comments on the previous developments are in order : * \(i ) it follows from eq . that @xmath38 and therefore the electric field created by a uniform polarization @xmath39 is zero ( see sec . ) .
* \(ii ) clearly @xmath35 exhibits the same singularity for @xmath40 as the green s function of the infinite euclidean plane @xmath41 , _
_ i.e.__@xcite + [ decompo ] @xmath42 + where @xmath43 is the unit dyadic tensor . in eqs .
@xmath44 is an arbitrary small cutoff ultimately set to zero if point dipoles are to be considered .
it must be understood that any integral involving @xmath45 must be calculated with @xmath46 and then taking the limit @xmath47 . in the presence of hard cores , as in the dhd fluid , the point dipoles can be replaced by a continuous charge distribution of symmetry axis @xmath48 and charge density @xmath49 . in that case
@xmath44 can be chosen to take any value @xmath50 . *
\(iii ) the interaction energy of two dipoles of @xmath19 is given by @xmath51 .
as well - known , expression can not be handled in a straightforward way in mc simulations since it converges very slowly .
this fourier series is then splitted into two series , one in direct space , the other in fourier space , both with good convergence properties . in that way one obtains the dipolar ewald potential
which is detailed in ref .
@xcite and which was used in the simulations reported in this paper .
we consider a line of length @xmath17 , parallel to axis @xmath52 of a square with pbc , which bears a continuous distribution of dipoles @xmath53 where the components @xmath54 are constants . without loss of generality the equation of the line can be chosen as @xmath55 .
the electric field @xmath56 created by this infinitesimal dipole at point @xmath57 is given by @xmath58 the total field of the line is obtained by integration @xmath59 \mathbf{e}_x \ ; .\end{aligned}\ ] ] eq .
shows that the field of a uniform distribution of dipoles aligned along the line @xmath55 ( @xmath60 ) vanishes exactly .
this result should be compared with the expression of the field created by an infinite linear chain of point dipoles , each with dipole moment @xmath61 , all aligned along the @xmath52 axis with spacing @xmath62 given in ref .
the latter is expressed as a series , the dominant term of which behaves as @xmath63 . with the correspondence @xmath64 and in the limit of a continuous distribution , _
i.e. _ @xmath65 with @xmath66 fixed , this dominant term vanishes in agreement with our result .
these results suggest that in the polymeric phase and in the presence of a sufficiently large external field the chains of dipoles aligned along the field do not contribute significantly to the maxwell field .
this point has been checked in our simulations and is discussed in sec . .
the expression of @xmath67 also shows that , apart from the singularity @xmath68 , a uniform electric field parallel to the axis @xmath69 of strength @xmath70 can be generated by a uniform distribution of dipoles perpendicular to the line ( @xmath71 ) .
in passing we note that the delta singularity has no effect in actual simulations since the condition @xmath55 , of measure zero , is never obtained .
we also remark that @xmath66 being fixed , the field vanishes in the limit @xmath72 as it should be , since then the result of the euclidean plane @xmath41 with free boundaries at infinity should be recovered . the electrostatic potential obtained by integration ,
is found , up to an arbitrary additional constant , to be @xmath73 yielding the expected discontinuity of the potential across a dipole layer@xcite @xmath74 these results can be extended to the @xmath75 case ( with @xmath76 replaced by @xmath77 ) where one should consider instead a planar uniform layer of dipoles aligned along the normal to the square @xmath55 of the cube @xmath78 .
we stress that it seems to be the only way to generate a uniform electrostatic field within pbc cubic geometries . at thermal equilibrium
the maxwell field @xmath79 is the sum of the external field @xmath80 and the field created by the dipoles @xmath81 .
it follows from sec .
that , in a given configuration of @xmath4 dipoles in the canonical ensemble , the microscopic field @xmath82 is given by @xmath83 where the microscopic polarization at point @xmath57 reads @xmath84 at equilibrium we thus have @xmath85 where the macroscopic polarization @xmath86 . for a fluid in a uniform external field
@xmath3 the polarization @xmath39 is uniform and , for a sufficiently low field strength , @xmath87 , which defines the dielectric constant @xmath88 .
as well known @xcite the relation between the maxwell field and the external field depends on the geometry .
for instance , in the plane @xmath41 with free boundaries at infininity we have @xmath89 while , in @xmath19 we have @xmath90 \cdot \mathbf{p } \ , , \nonumber \\ & = \mathbf{0 } \ ; , \end{aligned}\ ] ] as a consequence of eq .
from which it follows that @xmath91 and therefore @xmath92 .
for the present planar system the mc simulations were performed in a square of surface @xmath93 with pbc for @xmath94 hard disks carrying a dipole moment , in reduced units , @xmath95 varying between @xmath96 and @xmath97 .
the densities , in reduced units , were taken in the range @xmath98 from 0.1 to 0.3 .
the energy @xmath99 of the system in the presence of the external field @xmath100 is calculated , taking into account the pbc , by the ewald summation technique explicated in refs @xcite .
the expression at point @xmath101 of the microscoipic field @xmath102 due to the @xmath4 dipoles @xmath103 is thus given by @xmath104 where we have reported in the second line of eq .
the ewald expression used in our simulations .
the parameter @xmath105 regulates the rate of convergence of the sums in direct and fourier space . at equilibrium
the maxwell , or macroscopic field is defined as @xmath106 where the brackets denote a canonical average , while the macroscopic polarization is the canonical average of @xmath107 . in the region of the phase diagram where it is a defined quantity and in the limit of small fields ,
the dielectric constant @xmath14 of the system can be obtained , as discussed in sec . , through @xmath108 . in the simulations the external field @xmath109
is chosen parallel to the @xmath52 axis and the microscopic field @xmath102 is calculated for an ensemble of @xmath110 points located at the grid points of a square lattice .
it is calculated even if the point @xmath101 is inside the hard core of the disk .
hence the possibility of a rare but large contribution to @xmath111 , which is avoided by setting to zero the values @xmath112 in the tabulation of the gaussian functions in eq.([eq4 ] ) .
clearly it amounts to compute @xmath113 where the truncated green function @xmath114 is that of eq . .
it follows then from eq . that @xmath115 the maxwell field has been computed from eq .
( [ eq5 ] ) for the thermodynamic states given in table i. for dipole moment @xmath96 , and all densities considered , the system reaches equilibrium after of the order of @xmath116 mc trial moves per particle and about @xmath117 mc trial moves per particle appear sufficient to evaluate the canonical averages with a precision of @xmath118 .
such a sampling is also sufficient to achieve average values of @xmath119 at the different grid points @xmath120 independent of @xmath121 for all considered values of the external field . for dipole moments @xmath122 or @xmath97 , a configuration of the system
is typically characterized by the formation of chains and rings .
such an arrangement of the disks evolves to a different but similar arrangement only within @xmath116 trial moves per particle and therefore of the order of @xmath123 trial moves per particle are necessary for the estimate of the canonical averages and to obtain an average value of @xmath124 quasi - independent of @xmath121 .
figures 1 provides snapshots of such an evolution at @xmath125 for @xmath122 and @xmath126 .
the results of table i show that for all thermodynamic states considered the average value @xmath127 @xmath128 as @xmath129 cancels @xmath130 in the limit of statistical errors .
this result allows to obtain two estimates of the dielectric constant @xmath14 in the limit of small external fields , one , denoted @xmath131 , by @xmath132 the other , denoted @xmath133 , by @xmath134 .maxwell field and dielectric constant for thermodynamics states at low and high reduced dipole moments and for densities @xmath135 and @xmath136 .
@xmath131 and @xmath133 are computed from eqs .
( [ eq6 ] ) and ( [ eq7 ] ) , respectively . at @xmath137 , the evaluation of @xmath14
is precluded by the metastability of the thermodynamic states . [ cols="^,^,^,^,^,^,^,^",options="header " , ] .
for @xmath138 and @xmath96 the two estimates of @xmath14 are in excellent agreement for all values @xmath139 .
this value of @xmath140 is also in agreement with the value @xmath141 determined by the fluctuation formula @xcite at @xmath142 , @xmath143 where @xmath144 . at @xmath126 and @xmath122
the estimated value of @xmath14 is equal to @xmath145 if @xmath146 ; for @xmath147 the estimated values lie between @xmath148 and @xmath149 , a dispersion corresponding to the fact that after more then @xmath123 trial moves per particle the statistical error on @xmath150 remains of the order of @xmath151 .
the estimated value is in agreement with that , @xmath152 , obtained from the fluctuation formula . at @xmath138 and @xmath137 ,
the field @xmath153 is also zero in the limit of statistical errors , demonstrating that within the performed trial moves , the values of @xmath154 and @xmath155 remain coherent such that the maxwell field remains equal to the external field .
however , it is manifest that for @xmath137 , sampling of configurations is affected by strong metastability as shown by comparison of the average polarizations , obtained with @xmath156 and @xmath157 trials per particle , using different initial conditions . in one of the runs , starting from an unpolarized initial configuration , the average final polarization is @xmath158 ( first line of table i ) , whereas in a run starting from an initial configuration with a polarization close to the maximal polarization @xmath159 , the average final polarization is equal to @xmath160 ( second line of table i ) .
snapshots of characteristic configurations for these two states are given in figure 2 .
similarly at @xmath161 , for two runs with different initial configuration , the polarizations seem stable for a sampling of @xmath162 trials per particle , with values @xmath163 and @xmath164 , respectively .
this important metastability affecting the polarization values precludes a reliable calculation of @xmath14 from eqs .
( [ eq6 ] ) or ( [ eq7 ] ) .
a 2d dipolar system with periodic boundary conditions in the presence of a uniform external field behaves according to the laws of electrostatics .
this result allows to show by numerical simulation that the maxwell field inside the system is equal to the applied external field even if the system is highly polarized and also for stable or possibly metastable states where the configurations of the system are dominated by chaines and rings .
this equality between maxwell and external field does , however , not permit to overcome the difficulties inherent in the determination of the dielectric constant of the system at low temperature .
indeed , at these temperatures the simulations are affected by important metastability which precludes a unambiguous determination of the stable thermodynamic state .
99 caillol j - m and weis j - j 2015 _ mol .
phys . _ * 113 * 2487 caillol j - m 2015 _ j. chem
. phys . _ * 142 * 154505 perram j w and de leeuw s w 1981 _ physica _ * 109a * 237 morriss g p and perram j w 1985 _ physica _ * 129a * 395 fulton r l 1978 _ j. chem . phys . _ * 68 * 3089 fulton r l 1978 _ j. chem . phys . _ * 68 * 3095 fulton r l 1983 _ j. chem . phys . _ * 78 * 6865 jackson j d _ classical electrodynamics _
( john wiley & sons , new york , 1962 ) william r. toor w r 1993 _ j. of colloid and interfac .
sci . _ * 156 * , 335 c c c + & dipolar discs at @xmath122 , @xmath126 and @xmath125 .
the blue lines indicate the values and orientations of the average values of @xmath124 at the different points @xmath121 .
left : after @xmath165 trial moves per particle .
right : after @xmath166 trial moves per particle , the computed values @xmath167 are quasi - uniform.,title="fig : " ] & dipolar discs at @xmath122 , @xmath126 and @xmath125 .
the blue lines indicate the values and orientations of the average values of @xmath124 at the different points @xmath121 . left : after @xmath165 trial moves per particle .
right : after @xmath166 trial moves per particle , the computed values @xmath167 are quasi - uniform.,title="fig : " ] c c c + & dipolar discs at @xmath137 , @xmath138 and @xmath168 after @xmath157 moves per particle . left : final configuration for a run starting from an initial unpolarized configuration ; in this state @xmath169 @xmath170 . right : final configuration for a run starting from an initial polarized configuration ; in this state the average polarization is @xmath171 and @xmath172.,title="fig : " ] & dipolar discs at @xmath137 , @xmath138 and @xmath168 after @xmath157 moves per particle . left : final configuration for a run starting from an initial unpolarized configuration ; in this state @xmath169 @xmath170 .
right : final configuration for a run starting from an initial polarized configuration ; in this state the average polarization is @xmath171 and @xmath172.,title="fig : " ]
|
we study a two - dimensional system of dipolar hard disks in the presence of a uniform external electric field @xmath0 by monte carlo simulations in a square with periodic boundary conditions .
the study is performed in both the fluid at high temperature and the phase of living polymers at low temperature . in the considered geometry
the macroscopic maxwell field @xmath1 is computed and found to be equal to the external field @xmath0 in both phases .
the dielectric properties of the system in the liquid phase as well as in the polymeric phase are investigated .
|
turbulent flows are often studied under the assumptions of homogeneity and isotropy ( see , for a review , ref .
such assumptions are convenient for theoretical studies but are not always justified physically .
for example , it is well known that stratification or rotation applied to neutral flows lead to anisotropic turbulence ( see ref .
isotropy is even more difficult to justify in astrophysics where a magnetic field is almost always present at the largest scale of the system .
the magnetohydrodynamics ( mhd ) approximation has proved to be quite successful in the study of a variety of space plasmas . during the last quarter of century
many studies have been devoted to incompressible mhd turbulence in the presence of a uniform external magnetic field @xmath0 ( see ref .
one of the most clearly established results is that the presence of @xmath1 leads to a bi - dimensionalization of an initial isotropic energy spectrum : the turbulent cascade transfers energy preferentially to perpendicular wavenumbers , in the direction transverse to @xmath0 .
constant - flux spectra are known to occur in many instances in turbulent flows , the best example of which being the kolmogorov energy spectrum following a @xmath5 law for three - dimensional navier - stokes turbulence ( see ref .
power law spectra are also measured in turbulent mhd flows but the value of the scaling index is still hardly discussed in the community .
the first prediction in mhd was given independently by iroshnikov and kraichnan in ref .
@xcite ( hereafter ik ) .
they argued that the destruction of phase coherence by alfvn waves traveling in opposite directions along the local large scale magnetic field introduces a new timescale and a slowing down of energy transfer to small scales .
assuming isotropy , the dimensional analysis for three wave interactions leads to a @xmath6 spectrum for the total energy ( see also ref .
many direct numerical simulations of strong turbulence in isotropic ( @xmath7 ) mhd have been made during the last years ( see ref .
@xcite ) but a definitive conclusion about the spectrum index is still not achieved mainly because the kolmogorov and ik predictions are very close ; furthermore , such scaling laws may be slightly altered by intermittency effects , and the numerical resolution is barely sufficient to determine such spectral indices .
goldreich and shridar in ref .
@xcite proposed in 1995 a heuristic model of strong turbulence for anisotropic ( moderate @xmath1 ) mhd where the distinction between the perpendicular ( @xmath8 ) and parallel ( @xmath9 ) wavenumbers is made .
this model is based on a critical balance between linear wave periods @xmath10 and nonlinear turnover timescales @xmath11 for which the equality @xmath12 is assumed to hold at all scales in the inertial range .
( only the symmetric case , for which the alfvn waves traveling in opposite directions carry equal energy fluxes , is considered here and in the remainder of the paper . )
then the one dimensional perpendicular power spectrum for the total energy scales as @xmath13 whereas the parallel and perpendicular spatial sizes of eddies are correlated according to the scaling law @xmath2 .
the latter prediction is rather new ( see however ref .
@xcite ) and seems well observed in recent direct numerical simulations for moderate @xmath1 ( see ref .
@xcite ) . as is well known , in the limit of large @xmath1 mhd turbulence becomes strongly anisotropic the formalism of weak alfvn wave turbulence developed by galtier et al . in ref .
@xcite is well adapted to this situation .
it leads to the so called wave kinetic equations for which the exact power law solution for the total energy is @xmath14 .
the function @xmath15 is undetermined because of the dynamical decoupling of parallel planes in fourier space ( this is the signature of the absence of energy transfer along the @xmath1 direction ; it thus represents a shadow of the initial conditions ) .
numerical simulations of the wave kinetic equations show clearly such a constant - flux spectrum but it also reveals the existence of a transient spectrum during the front propagation towards small scales with a steeper power law in @xmath16 , the dynamics of which is not yet clarified ( see ref .
the discovery of such transient spectra in wave turbulence and the possible existence of a family of solutions that are not caught by the usual technique of conformal transform ( given in ref .
@xcite ) constitute a new exciting topic of research where some progress is currently being made ( see , for example , ref .
when using a shell model of strong turbulence , it has also been found in ref .
@xcite that , when considering the decay of energy in time as a power law , initial transients occur that also follow power laws to leading order and that precede the final power law decay of the energy ; the origin of such transients in time may well be a transient in the fourier energy spectrum preceding the establishment of a kolmogorov like spectrum , although this point has not been documented yet . in this paper
, we propose a heuristic model that describes anisotropic mhd flows for the regimes of strong ( moderate @xmath1 ) and weak wave ( strong @xmath1 ) turbulence as well as the transition between them . as a result of our analysis
, a family of solutions is found for the anisotropic total energy spectrum from which the transient spectrum described above is a particular solution .
we show that the model supports the same anisotropic scaling law between the parallel and perpendicular wavenumbers as the one found in the context of critical balance but it is more general in the sense that here we do not impose equality between linear wave periods and nonlinear turnover timescales which allows us to use the ik phenomenology .
we finally propose to extend the anisotropic model to two other types of fluids .
the presence of the mean magnetic field @xmath1 leads to anisotropy and thus to different variations with wavenumber directions , leading us to distinguish between @xmath17 and @xmath18 .
more precisely we will assume that @xmath19 , _
i.e. _ that under the external agent ( here , @xmath20 ) , the turbulent flow develops principally in the direction perpendicular to this agent . we will consider the symmetric case for which , in particular , the _ r.m.s .
_ value scale in the inertial range of the fluctuating velocity field @xmath21 and magnetic field @xmath22 have the same order of magnitude ; note that @xmath22 is taken in velocity units since we are restricting the analysis to the incompressible mhd case ( the density is constant and can be assumed to be equal to unity ) . in the classical kolmogorov phenomenology ( hereafter , k41 ) ,
the fluctuations are distributed isotropically and there is only one timescale , the nonlinear time or eddy turnover time @xmath11 , which is also the transfer time of the energy to small scales within the system , @xmath23 .
the rate of energy transfer per unit mass writes @xmath24 , where @xmath25 is the total energy . in the very same spirit
, we can develop the ik phenomenology for anisotropic mhd turbulence using explicitly the fact that the eddy turnover time is @xmath26 and the alfvn wave period is @xmath27 .
we assume that these two timescales are _ not _ equal which allows us to use the ik phenomenology .
the rate of energy transfer per unit mass now writes _
ik_a ~ ~ , [ relation1 ] where the transfer time will be assumed to scale as _
tr = _ nl , [ relation2 ] as is known to be the case for three - wave interaction processes ( see ref .
the subscript `` a '' in @xmath28 stands for the anisotropic version of ik . including relation ( [ relation2 ] ) into ( [ relation1 ] ) ,
one obtains _
ik_a ~v^4 ^2 b_0 .
[ epsik ] we now define the anisotropic energy spectrum , e ( , ) ~^- ^- , [ alphabet ] where @xmath29 and @xmath30 are unknown .
it is a 2d energy spectrum . in other words ,
the total energy is recovered by directly integrating the spectrum along @xmath31 and @xmath32 , @xmath33 ( see ref .
@xcite , for a rigorous definition ) .
it is important to note at this stage a difference with the classical phenomenology where the energy spectrum is not introduced with unknown indices but rather deduced by the analysis .
we introduce relation ( [ alphabet ] ) into ( [ epsik ] ) , and by noticing that @xmath34 , we obtain : b_0 e_ik_a ~^4 - 2 ^1 - 2 , which can also be rewritten in terms of a ( @xmath31 , @xmath32 ) relationship as : ~ ( b_0 e_ik_a)^1 1 - 2 ^4 - 22 - 1 . [ ani2
] this is our first anisotropic relation . in order to proceed further
, we are now seeking a second relationship between the two scaling indices @xmath29 and @xmath30 .
one way to obtain a unique relation between them is to use the assumption of constant ratio between @xmath11 and @xmath10 at all scales . in other words
, we will assume that = [ chi ] is a constant at all scales but not necessarily equal to one as it is in the critical balance model in ref .
it is important to note that the precise value of @xmath35 is not important in itself since it is just a numerical factor that does not change the power law scaling .
however if @xmath35 is smaller than one we are allowed to use the ik phenomenology instead of the k41 one for @xmath36 .
a ratio of one seems to be very restrictive and does not correspond to some of the results stemming from direct numerical simulations where @xmath35 can be smaller than unity , as observed for example in ref .
@xcite , a two - dimensional geometrical case where local anisotropy is possible , or in solar wind _ in situ _ measurements ( see ref .
@xcite ) where @xmath35 seems to be smaller than unity .
it is also in apparent contradiction with wave turbulence theory for which we have @xmath37 .
we will see that the assumption of constant ratio at all scales is sufficient to achieve a unified model that describes both strong and weak mhd turbulence . with the definition of the timescales given above ,
we obtain ~ .
[ relationchi ] including expression ( [ relationchi ] ) into ( [ epsik ] ) gives : ~e_ik_a^1/3 ^4/3 b_0 ^2/3 .
[ ani4 ] if we use explicitly the fact that both @xmath35 and the rates of energy transfer per unit mass do not depend on the scale ( see ref .
@xcite ) , we finally obtain ~^2/3 b_0 .
[ ani5 ] this leads us to the first main conclusion of this paper : the same scaling in wavenumbers as the critical balance model is obtained when @xmath35 is _ not _ equal to unity , when the ik phenomenology is used .
it thus seems to be a general law for incompressible mhd turbulence , when either the kolmogorov phenomenology for energy transfer or the @xmath28 three - wave formulation of energy transfer is utilized , as long as a critical balance ( generalized to any ratio differing from unity ) is assumed between stretching by velocity gradients and wave motions in the presence of a uniform field @xmath1 .
another important and somewhat unexpected consequence of this analysis has to deal with the scaling of the energy spectrum . in the context of critical balance
, it is claimed that the energy spectrum derived with an anisotropic kolmogorov phenomenology scales like @xmath38 .
here we show that it is in fact possible to find a multitude of spectral indices @xmath29 and @xmath30 that satisfy the assumption of constant ratio @xmath35 at all scales in the inertial range .
such indices follow in fact a general linear relationship . to find this relation we need to equalize the power - law behaviors found in relations ( [ ani2 ] ) and ( [ ani5 ] ) between the parallel and perpendicular distribution of modal energy in fourier space .
if we use explicitly the fact that the rates of energy transfer per unit mass do not depend on the scale ( see ref .
@xcite ) , we finally obtain 3 + 2= 7 .
[ ab ] relationship ( [ ab ] ) is general and can be used for strong turbulence as well as for weak wave turbulence .
note that the relationship ( [ ab ] ) is compatible with the anisotropic kolmogorov spectrum ( with @xmath39 and @xmath40 ) as advocated by goldreich and sridhar in ref .
@xcite , and with the anisotropic ik - like spectrum corresponding to three - wave interactions for weak mhd turbulence ( @xmath41 and @xmath42 ) ( see ref .
we conjecture that it is compatible with the physics of the transitional regime between weak and strong mhd turbulence as well . in other words
the law ( [ ab ] ) we just derived shows that the @xmath43 scaling between a dependence of @xmath31 and @xmath32 is not a unique signature of the kolmogorov spectrum but rather a signature of the rate at which energy is transfered to small scales where it is dissipated , a trace of the decay of energy , and as such an unavoidable dimensional law
the preceding remark may be linked to the heretofore unexplained following finding : using direct numerical simulations in two space dimensions , it was shown in ref .
@xcite that the structure functions of order @xmath44 based on the energy flux to small scales ( as expressed in terms of the exact laws for mhd turbulence derived in ref .
@xcite ) have a self
similar scaling in the inertial range which is compatible with the scaling for the velocity structure functions in fluid turbulence , whereas the structure functions of the basic fields ( velocity and magnetic fields ) are more intermittent insofar as they depart more significantly from a linear scaling with @xmath44 .
what this paper shows is that both models ( k41 and ik ) can be seen in a unifying way , illustrated by the law ( [ ab ] ) .
note that in the case of the advection of a passive tracer such as temperature , it also seems that the scaling of the structure functions based on the flux of the tracer is close to the fluid scaling laws ( see ref .
@xcite ) , whereas the tracer itself is well known to be strongly intermittent .
more surprisingly perhaps , the choice of @xmath45 and @xmath46 also satisfies the law ( [ ab ] ) derived in this paper .
this @xmath47 spectrum was found in ref .
@xcite as a precursor to the constant flux solution of the alfvn wave kinetic equations which establishes itself later in time ; indeed , in that paper , the @xmath48 spectrum is found numerically for the case @xmath49 , without any dependence in @xmath32 ( corresponding to @xmath46 ) .
the simple model proposed here thus sheds some light , albeit heuristically , on two intriguing facts that have emerged recently concerning weak wave turbulence : ( i ) the fact that , preceding in time the constant flux solution , a power - law spectrum ( called the precursor ) establishes itself the origin of which was unrelated to anything known about turbulence spectra ( see ref .
@xcite ) until now , and ( ii ) the fact that in some cases ( see ref .
@xcite ) in the weak wave turbulence regime , a wealth of power - law solutions in the @xmath50 plane can be found numerically as stationary solutions to the wave kinetic equations , although it is not clear whether such solutions are attractive , nor whether they are stable .
the link between the dimensional argument given here which allows to recover the @xmath45 precursor spectrum , and the self - similar argument given in ref .
@xcite , which is compatible with the @xmath45 spectrum , remains to be clarified .
indeed , in ref .
@xcite one recovers the numerically observed law of self - similar decay of the energy spectrum , assuming that the energy spectrum scales in the precursor phase as @xmath47 , whereas in this paper we give a heuristic justification to that same @xmath51 law in terms of dimensional arguments compatible with three - wave interactions and with the assumption that there is no @xmath32 dependency in the precursor solution . we show furthermore that this law stems from the same type of unified approach that also gives the other known spectra for mhd turbulence . in order to ascertain the validity of the model derived here , numerical computations with the full three dimensional mhd equations
can be envisaged but being able to distinguish between different power laws which are close in their respective spectral indices may prove difficult . this point is currently investigated and will be presented elsewhere .
the argumentation delineated here could be used for other types of wave turbulence , like in the case of whistler waves ( see ref .
@xcite ) , inertial waves ( see ref .
@xcite ) or gravity waves ( see ref .
@xcite ) . for whistler waves that can be encountered in hall mhd or electron mhd
, the characteristic wave period scales as @xmath52^{-1}$ ] ( see ref .
@xcite ) . one finds , following the same analysis as presented before in this paper , that @xmath53 ( see also ref .
@xcite ) ; hence , the prediction for scaling exponents as defined in equation ( [ alphabet ] ) for the whistler turbulence energy spectrum becomes : @xmath54 .
( note that here , the eddy turnover time is based on the magnetic field , @xmath55^{-1 } = [ b \ , \kpe^2]^{-1}$ ] , since @xmath56 . )
the known constant flux solution , @xmath57 and @xmath42 , to the wave kinetic equations derived in ref .
@xcite is recovered again as well as the strong turbulence prediction , @xmath45 and @xmath40 ( see ref .
@xcite ) . when considering inertial waves in rotating turbulence for a navier - stokes fluid , the wave period scales as @xmath58 .
the same analysis as before leads to the following relationship between scaling exponents : @xmath59 .
the known constant
flux solution to the wave kinetic equations written in ref .
@xcite for wave turbulence corresponds to @xmath57 and @xmath42 , which does fulfill the above relationship and we recover the kolmogorov prediction for strong turbulence as well , _ viz .
_ @xmath39 and @xmath40 for a weak rotation rate ( see the comment in ref .
this prediction again could be tested using direct numerical simulations , and the possible existence of precursors could also be studied and checked against the compatibility with the above relationship between the scaling exponents @xmath29 and @xmath30 in the rotating case .
we wish to thank pablo mininni and hlne politano for interesting discussions .
financial support from the research training network tostisp through eec grant hprn - ct-2001 - 00310 , from pnst / insu / cnrs , from nsf - cmg grant 0327888 , and from nasa grant in geospace science nra 03-oss-01 , are all gratefully acknowledged .
writing in general that the characteristic wave period scales as @xmath60 , one finds @xmath61 ; hence , the generalization of equation ( [ ab ] ) for waves with the above dispersion relation with the turnover time @xmath62 is found to be : @xmath63 .
|
a heuristic model is given for anisotropic magnetohydrodynamics ( mhd ) turbulence in the presence of a uniform external magnetic field @xmath0 .
the model is valid for both moderate and strong @xmath1 and is able to describe both the strong and weak wave turbulence regimes as well as the transition between them .
the main ingredient of the model is the assumption of constant ratio at all scales between linear wave period and nonlinear turnover timescale .
contrary to the model of critical balance introduced by goldreich and sridhar [ p. goldreich and s. sridhar , apj * 438 * , 763 ( 1995 ) ] , it is not assumed in addition that this ratio be equal to unity at all scales which allows us to use the iroshnikov - kraichnan phenomenology .
it is then possible to recover the widely observed anisotropic scaling law @xmath2 between parallel and perpendicular wavenumbers ( with reference to @xmath0 ) and to obtain the universal prediction , @xmath3 , for the total energy spectrum @xmath4 . in particular , with such a prediction the weak alfvn wave turbulence constant - flux solution is recovered and , for the first time , a possible explanation to its precursor found numerically by galtier is given .
|
a vast research area , often called discrete nonlinear optics deals with evanescently coupled arrayed waveguides featuring material nonlinearity lederer .
discrete arrays of optical waveguides have drawn a great deal of interest not only because they introduce a vast phenomenology of the nonlinear light propagation , such as e.g. the prediction @xcite and experimental creation @xcite of discrete vortex solitons , but also due to the fact that they offer a unique platform for emulating the transmission of electric signals in solid - state devices , which is obviously interesting for both fundamental studies and applications christodoulides2,lederer .
furthermore , the flexibility of techniques used for the creation of virtual ( photoinduced ) @xcite and permanently written @xcite guiding arrays enables the exploration of effects which can be difficult to directly observe in other physical settings , such as anderson localization @xcite .
another field in which arrays of quasi - discrete waveguides find a natural application is the realization of the optical @xmath0 ( parity - time ) symmetry @xcite . on the one hand , a pair of coupled nonlinear waveguides , which carry mutually balanced gain and loss , make it possible to realize @xmath0-symmetric spatial or temporal solitons ( if the waveguides are planar ones or fibers , respectively ) , which admit an exact analytical solution , including their stability analysis @xcite . on the other hand , a @xmath0-symmetric dimer , i.e.
, the balanced pair of gain and loss nodes , can be embedded , as a defect , into a regular guiding array , with the objective to study the scattering of incident waves on the dimer @xcite .
we note here in passing that some times , also the term `` dipoles '' may be used for describing such dimers , however we will not make use of it here , to avoid an overlap in terminology with classical dipoles in electrodynamics as discussed e.g. in @xcite .
discrete solitons pinned to a nonlinear @xmath0-symmetric defect have been reported too raymond .
such systems , although governed by discrete nonlinear schrdinger ( dnls ) equations corresponding to non - hermitian hamiltonians , may generate real eigenvalue spectra ( at the linear level ) , provided that the gain - loss strength does not exceed a critical value , above which the @xmath1 symmetry is broken @xcite [ self - defocusing nonlinearity with the local strength growing , in a one - dimensional ( 1d ) system , from the center to periphery at any rate faster that the distance from the center , gives rise to stable fundamental and higher - order solitons with _ unbreakable _
@xmath0-symmetry @xcite ] .
one- and two - dimensional ( 1d and 2d ) lattices , built of @xmath0 dimers , were introduced in refs .
@xcite and @xcite , respectively .
discrete solitons , both quiescent and moving ones , were found in these systems @xcite . in the continuum limit ,
those solitons go over into those in the above - mentioned @xmath0-symmetric coupler @xcite . accordingly , a part of the soliton family is stable , and another part is unstable .
pairs of parallel and anti - parallel coupled dimers , in the form of @xmath0-symmetric plaquettes ( which may be further used as building blocks for 2d chains ) were investigated too @xcite .
the objective of the present work is to introduce a _ staggered _ chain of @xmath0-symmetric dimers , with the orientations of the dimers alternating between adjacent sites of the chain .
this can also be thought of as an extension of a plaquette from refs .
@xcite towards a lattice .
while this ladder - structured lattice is not a full 2d one , it belongs to a class of chain systems which may be considered as @xmath2d models @xcite . as shown in section
ii , where the model is introduced , the fundamental difference from the previously studied ones is the fact that such a system , although being nearly one - dimensional , actually realizes the @xmath0 symmetry in the 2d form , with respect to both horizontal and vertical directions . in section iii
we start the analysis from the solvable anti - continuum limit ( acl ) @xcite , in which the rungs of the ladder are uncoupled ( in the opposite continuum limit , the ladder degenerates into a single nls equation ) .
using parametric continuation from this limit makes it possible to construct families of discrete solitons in a numerical form .
such solution branches are initiated , in the acl , by a single excited rung , as well as by the excitation confined to several rungs .
the soliton stability is systematically analyzed in section iii too and , if the modes are identified as unstable , their evolution is examined to observe the instability development .
the paper is concluded by section iv , where also some directions for future study are presented .
we consider the ladder configuration governed by the dnls system with intersite coupling constant @xmath3 , @xmath4where evolution variable @xmath5 is the propagation distance , in terms of the optical realization .
coefficients @xmath6 and @xmath7 with @xmath8 represent @xmath0-symmetric gain - loss dimers , whose orientation is staggered ( alternates ) along the ladder , the sites carrying gain and loss being represented by amplitudes @xmath9 and @xmath10 , respectively .
cubic nonlinearity with coefficient @xmath11 is present at every site , and @xmath12 accounts for the vertical coupling along the ladder s rungs , each representing a @xmath0-symmetric dimer .
the system is displayed in fig .
[ xypic ] .
as seen in the figure , the nearly 1d ladder realizes the @xmath0 symmetry in the 2d form , with respect to the horizontal axis , running between the top and bottom strings , and , simultaneously , with respect to any vertical axis drawn between adjacent rungs . ) and vertical ( @xmath13 ) coupling constants .
red and blue dots designate sites carrying the mutually balanced gain and loss , respectively .
the dashed lines designate the horizontal and vertical axes of the @xmath14 symmetry.,width=302 ] by means of obvious rescaling , we can fix @xmath15 , hence the nonlinearity coefficient takes only two distinct values , which correspond , respectively , to the self - focusing and defocusing onsite nonlinearity , @xmath16 and @xmath17 .
the usual dnls equation admits the sign reversal of @xmath11 by means of the well - known staggering transformation @xcite .
however , once we fix @xmath18 ( and also @xmath12 ) in eq .
( [ tdnls ] ) , this transformation can not be applied , as it would also invert the signs of @xmath19 and @xmath20 . the single self - consistent continuum limit of system ( [ tdnls ] ) ,
corresponding to @xmath21 , is possible for the fields related by @xmath22 , with phase shift @xmath23 .
replacing , in this limit , the finite - difference derivative by the one with respect to the continuous coordinate , @xmath24 , yields the standard nls equation , @xmath25given its `` standard '' nature , leading to a full mutual cancellation of the gain and loss terms , we will not pursue this limit further . instead ,
as shown below , we will use as a natural starting point for examining nontrivial localized modes in the discrete system ( [ tdnls ] ) the opposite acl , which corresponds to @xmath26 , i.e. , the set of uncoupled rungs . stationary solutions to eqs .
( [ tdnls ] ) with real propagation constant @xmath27 are sought in the usual form , @xmath28 and @xmath29 , where functions @xmath30 and @xmath31 obey the stationary equations , @xmath32numerical solutions of these equations for discrete solitons are produced in the next section . to analyze the stability of the solutions , we add perturbations with an infinitesimal amplitude @xmath33 and frequencies @xmath34 , @xmath35the linearization of eq .
( [ tdnls ] ) with respect to the small perturbations leads to the eigenvalue problem , @xmath36 = \omega \left [ \begin{array}{c } a_{n } \\
b_{n}^{\ast } \\
d_{n}^{\ast } % \end{array}% \right ] , \label{lineq}\]]where @xmath37 is a @xmath38 matrix for the ladder of length @xmath39 . using standard indexing ,
@xmath40 submatrices of @xmath37 are defined as @xmath41 @xmath42 where @xmath43 is an @xmath40 matrix of zero elements , except for the super- and sub - diagonals that contain all ones .
for the zero solution of the stationary equation , @xmath44matrix @xmath37 has constant coefficients , hence perturbation eigenmodes can be sought for as @xmath45 . then eq .
becomes a @xmath46 system , whose eigenvalues can be found explicitly : @xmath47so that @xmath34 is real only for @xmath48 . in other words ,
the @xmath0-symmetry is broken , with @xmath49 acquiring a positive real part , which drives the exponential growth of the perturbations , at @xmath50it is interesting to observe here that the coupling between the rungs decreases the size of the interval of the unbroken @xmath0-symmetry of the single dimer @xcite . in the stability region ,
( [ zernu ] ) demonstrates that real perturbation frequencies take values in the following intervals : @xmath51similarly , for the perturbations in the form of @xmath52 the negatives of expressions ( [ zernu ] ) are also eigenvalues of the zero stationary solution , and at @xmath53 , they fall into the negatives of intervals ( [ intervals ] ) . simultaneously , eq . ( [ zernu ] ) and its negative counterpart give the dispersion relation for plane waves ( phonons " ) in the linearized version of eq .
( [ tdnls ] ) .
accordingly , intervals ( [ intervals ] ) , along with their negative counterparts , represent phonon bands of the linearized system .
in section [ sol ] we produce stationary solutions in the form of discrete solitons .
this computation begins by finding exact solutions for the acl , @xmath54 , and then the continuing the solutions numerically to @xmath55 , by means of the newton s method for each @xmath3 ( i.e. , utilizing the converging solution obtained for a previous value of @xmath3 as an initial seed for the newton s algorithm with @xmath56 ) . as suggested by eq .
( [ cr ] ) , we restrict the analysis to @xmath57 , so as to remain within the @xmath0-symmetric region at @xmath58 . subsequently , the stability interval of the so constructed solutions is identified , in a numerical form too .
to construct stationary localized solutions of eqs .
( [ tdnls ] ) at @xmath58 , when individual rungs are decoupled , we substitute @xmath59 with real @xmath60 in eq .
( [ stat ] ) , which yields relations @xmath61 for the uncoupled ladder , one can specify either a single - rung solution , with fields at all sites set equal to zero except for @xmath62 satisfying eq .
( [ rung_rel ] ) , or a double - rung solution with nonzero fields @xmath63 and @xmath64 satisfying the same equations .
we focus on these two possibilities in the acl ( although larger - size solutions are obviously possible too ) .
these are the direct counterparts of the single - node and two - node solutions that have been extensively studied in 1d and 2d dnls models @xcite .
we take parameters satisfying constraints @xmath65to make the second equation ( [ rung_rel ] ) self - consistent .
then , two solution branches for @xmath60 are possible .
the first branch satisfies @xmath66 and @xmath67 . choosing a solution with @xmath68 in the rung carrying nonzero fields
, we name it an _ in - phase rung _ , as the phase shift between the gain and loss poles of the respective dimer is smaller than @xmath69 , namely , @xmath70 $ ] .
the second branch satisfies @xmath71 and @xmath72 .
the rung carrying the solution with @xmath73 is called an _ out - of - phase _ one , as the respective phase shift between the elements exceeds @xmath69 , _
viz_. , @xmath74 $ ] .
the two branches meet and disappear at @xmath75 , when @xmath76 .
recall that @xmath77 [ see eq .
( [ cr ] ) ] is the boundary of the @xmath0-symmetric region for @xmath58 .
these branches can be also be considered as stemming from the hamiltonian limit of @xmath78 , where @xmath79 and @xmath80 correspond , respectively , to the usual definitions of the in- and out - of - phase hamiltonian dimers .
the stability eigenfrequencies for stationary solitons at @xmath58 can be readily calculated analytically in the acl @xcite . in this case
, @xmath37 has the same eigenvalues as submatrices @xmath81which is associated with zero - amplitude ( unexcited ) rungs , and @xmath82associated with the excited ones , which carry nonzero stationary fields , with @xmath31 , @xmath30 taken as per eqs .
( [ rung - ansatz ] ) and ( rung_rel ) . in other words ,
each of the four eigenvalues of @xmath83 , @xmath84is an eigenvalue of @xmath37 with multiplicity equal to the number of zero - amplitude rungs , while each of the four eigenvalues of @xmath85 , @xmath86appears as an eigenvalue of @xmath37 with multiplicity equal to the number of excited rungs .
here @xmath87 , and @xmath88 for an in- and out - of - phase rung , respectively .
equation ( [ exeig ] ) shows that the out - of - phase excited rung is always stable , as it has @xmath89 .
the in - phase excited rung is stable for @xmath90 , and unstable for @xmath91 .
thus , for solutions that contain an excited in - phase rung in the initial configuration at @xmath58 , there are the two critical values , _ viz_. , @xmath92 given by eq .
( [ cr ] ) , and the additional one , which designates the instability area for the uncoupled in - phase rungs : @xmath93 a choice alternative to eq .
( [ par ] ) is @xmath94 in this case , the analysis differs only in that the sign of @xmath95 in eq . is switched .
that is , the in - phase rung is now associated to negative @xmath96 , while the out - of - phase one to positive @xmath97 . in this case
, the in - phase rung is always stable , while its out - of - phase counterpart is unstable at @xmath98 , see eq .
( [ cr2 ] ) . to construct soliton solutions for coupling constant @xmath3 increasing in steps of @xmath99 , we write eq .
( [ stat ] ) as a system of @xmath100 equations for @xmath101 real unknowns @xmath102 , with @xmath103 , @xmath104 .
then we apply the newton s method with the initial guess at each step taken as the soliton solution found at the previous value of @xmath3 , as mentioned above .
thus , the initial guess at @xmath105 is the analytical solution for @xmath58 given by eqs .
( rung - ansatz)-([rung_rel ] ) with parameters taken according to either eq . or eq . .
figure [ s=1abs ] shows @xmath106 for the solutions identified by this process on a ( base @xmath107 ) logarithmic scale as a function of @xmath3 for parameters taken as per eq . .
the logarithmic scale is chosen , as it yields a clearer picture of the variation of the solution s spatial width , as @xmath3 varies .
the different solutions displayed in fig .
[ s=1abs ] include those seeded by the single excited in- and out - of - phase rungs ( the top row ) , and two - rung excitations for which there are three possibilities : in- in both and out - of - phase structures in both rungs ( the second row ) , as well as a mixed structure involving one rung initially excited in - phase , and the other one excited out - of - phase ( the bottom row ) . in figure [ s=1prof ]
we plot @xmath106 for fixed @xmath3 across the various configurations . a point that is clearly illustrated by this figure , which is not evident on the logarithmic scale of fig .
[ s=1abs ] , is the asymmetric spatial structure of the mixed - phase solution of the bottom row ( asymmetric solitary waves have also been proposed in full 2d lattices @xcite ; see also the detailed analysis of @xcite ) .
equations ( [ rung - ansatz])-(rung_rel ) show that , for @xmath58 , since the out - of - phase case corresponds to @xmath108 , the amplitude of the out - of - phase rung , @xmath109 , with @xmath110 , is larger in comparison with its in - phase counterpart , which has @xmath67 .
the asymmetry for the mixed - phase solution persists for @xmath55 , fig .
[ s=1prof ] showing an example of this
. then , fig .
[ s=1w ] shows more explicitly the increasing width of the soliton , using the second moment of the density distribution , as the respective diagnostic , @xmath111versus @xmath3 for the solutions shown in fig .
[ s=1abs ] .
it is relevant to point out that the variation of this width - measuring quantity is fairly weak in the case of the out - of - phase solutions and mixed ones , while it is more significant in the case of the single and double in - phase excited rungs . , where @xmath30 at @xmath58 is given by eqs .
( [ rung - ansatz ] ) and ( [ rung_rel ] ) , and at @xmath55 the soliton solutions @xmath30 are obtained by the continuation in @xmath3 , see the text .
common parameters are @xmath112 , @xmath113 and @xmath114 .
the initial configuration of the excited rungs at @xmath58 and parameters are : a single in - phase rung with @xmath115 , @xmath116 ( top left ) , a single out - of - phase rung with @xmath117 , @xmath116 ( top right ) , two in - phase rungs with @xmath118 , @xmath119 ( middle left ) , two out - of - phase rungs with @xmath120 , @xmath121 ( middle right ) , and , finally , a mixed state carried by two rungs with @xmath122 , @xmath123 , @xmath124 ( bottom center ) .
plots of @xmath125 are identical to those of @xmath126 .
as @xmath3 increases , small amplitudes appear at adjacent rungs , and the soliton gains width .
the corresponding second moment , @xmath127 , defined as per eq .
, is shown in fig .
the stability of the solitons shown here is predicted by eigenvalue plots displayed in fig .
[ s=1mxre ] for @xmath112 . ] .
the configurations of the initial ( @xmath58 ) solution and other parameters follow the same pattern as in fig .
[ s=1abs ] . ] , defined as per eq .
( [ w ] ) , corresponding to each of the plots in fig .
[ s=1abs ] . ] in fig .
[ s=1arg ] the absolute value of the phase difference between fields @xmath30 and @xmath31 at two sides of the ladder is shown .
in other words , this figure shows whether each rung of the ladder belongs to the in - phase or out - of - phase type , as a function of @xmath3 .
this figure reveals that , as @xmath3 increases , there is a progressive spatial expansion ( across @xmath128 ) in the number of sites supporting a phase difference that develops around the initially excited sites .
the individual phases of @xmath129 and @xmath130 are shown in fig .
[ s=1arguv ] . we show in the bottom two plots of fig .
[ s=1arguv ] that two different types of phase profiles can arise ; one type has phase with the same sign on both the left and right sides of the outer portion of the ladder , and the second type has phases that are of opposite sign on the left and right sides of the outer portion of the ladder .
we address this point more in the next section where we discuss stability .
, plotted as a function of @xmath3 and @xmath128 , where @xmath131 is the solution whose absolute value is presented in fig .
[ s=1abs ] .
the value of @xmath132 is set to zero for any @xmath128 at which @xmath133 in fig .
[ s=1abs ] . in other words ,
the phase shift is shown as equal to zero when the amplitude is too small .
for the top left and middle left plots , the soliton s field is different from zero at one or two in - phase rung(s ) when @xmath58 , and as @xmath3 increases the solutions stay in - phase . for the top right and middle right plots , the field at @xmath58 is nonzero and out - of - phase at the one or two central rungs , and , as @xmath3 increases , the fields at these rungs , and at two rungs on either side of the central ones , tend to be out - of - phase , while the field at other rungs , located farther away , tend to be in - phase .
similarly , in the bottom plot , where at @xmath58 the @xmath134 rung is in - phase and the @xmath135 one is out - of - phase , as @xmath3 increases , most rungs tend to be in - phase , except for @xmath136 . ]
( blue ) and @xmath137 ( green ) with values in the interval @xmath138 $ ] , plotted as a function of the spatial variable @xmath128 .
the top four plots and bottom - most left plot have fixed parameter values @xmath139 and @xmath140 .
that is , the @xmath141 solutions for these five plots are the same as those whose modulus is plotted in fig .
[ s=1prof ] . in the bottom right plot
we show @xmath142 for the same fixed parameter values as the bottom left plot except with @xmath143 .
the bottom two plots show the different phase profiles can that arise for different @xmath144 values , with either same ( left ) or opposite ( right ) signs for the phase on the outer portions of the ladder . ]
figures [ s=-1abs ] , [ s=-1prof ] , [ s=-1w ] , [ s=-1arg ] , [ s=-1arguv ] are similar to their counterparts [ s=1abs ] , [ s=1prof ] , [ s=1w ] , s=1arg , [ s=1arguv ] , respectively , but with the parameters taken as per eq . instead of eq . .
comparing figs .
[ s=1arg ] and [ s=-1arg ] shows that @xmath110 favors the solutions with in - phase rungs as @xmath3 increases , while @xmath17 favors the out - of - phase rungs .
in other words , the progressively expanding soliton keeps the in- and out - of - phase structures , in the case of the self - focusing ( @xmath110 ) and defocusing ( @xmath17 ) onsite nonlinearity , respectively , in agreement with the well - known principle that discrete solitons feature a staggered pattern in the case of the self - defocusing @xcite .
also , according to eq .
( [ rung_rel ] ) , for @xmath17 the asymmetry of the mixed - phase solution is switched in comparison to the @xmath110 case , lending the in - phase rung a larger magnitude of the fields than in the out - of - phase one . , but for @xmath145 .
common parameters are @xmath146 and @xmath147 .
the initial values of @xmath148 and @xmath149 at @xmath58 follow the pattern of fig .
[ s=1abs ] .
other parameters are : @xmath150 on the top left , @xmath151 on the top right , @xmath152 on the middle left , @xmath153 on the middle right , and finally @xmath154 on the bottom center .
as @xmath155 increases , small amplitudes appear at adjacent sites , and the soliton gains width , as shown by means of @xmath156 in fig .
the stability of the solitons shown here is predicted by the eigenvalue plots in fig .
[ s=-1mxre ] at @xmath112 . ] , for @xmath157 .
the configurations of the initial @xmath58 solution and parameters follow the same pattern as in fig .
[ s=-1abs ] . ] ) , corresponding to each of the plots in fig .
[ s=-1abs ] . ]
but for @xmath158 , i.e. , for the discrete solitons presented in fig .
[ s=-1abs ] . for the top
left and middle left plots , the soliton s field is nonzero at one or two in - phase rung(s ) when @xmath58 , and as @xmath3 increases these rungs remain in - phase , while all the others are out - of - phase . for the top right and middle right plots , the soliton s field at @xmath58 is nonzero and
out - of - phase at one or two central rungs , and all rungs remain out - of - phase with the increase of @xmath3 . in the bottom plot ,
only the @xmath134 rung remains in - phase , while all others are out - of - phase . ] ( blue ) and @xmath137 ( green ) in the interval @xmath138 $ ] , plotted as a function of the spatial variable @xmath128 for the fixed parameter values @xmath139 and @xmath159 .
that is , the @xmath141 solutions are the same as those whole modulus is plotted in fig .
[ s=-1prof ] . ]
it is relevant to stress that the discrete solitons seeded in the acl by double rungs feature a _ bi - dimer _ structure which does not carry a topological charge @xcite .
i.e. , the solitons can not take the form of vortices , according to our numerical results , ( contrary to what is the case , e.g. , for a ring containing a single @xmath0-symmetric dipole @xcite ) .
figure [ s=1mxre ] shows two - parameter stability diagrams for the solitons by plotting the largest instability growth rate ( if different from zero ) , @xmath160 , as a function of @xmath3 and @xmath19 for parameter values taken per eq . , and fig .
[ s=-1mxre ] shows the same as per eq . .
the respective stability boundaries are shown by green lines ( white , in the - black - and - white version of the figures ) .
some comments are relevant here .
recall that eqs .
( [ cr ] ) and ( [ cr2 ] ) impose stability limitations , respectively , from the point of view of the zero - background solution in the former case , and the single - site excitation in the latter case .
the former background stability condition indicates that the line of @xmath161 ( parallel to the antidiagonal cyan line in fig .
[ s=1mxre ] ) poses an upper bound on the potential stability of any excitation , as it is the condition for the stability of the zero background , on top of which any solitary wave is constructed .
it can be seen in both figs .
[ s=1mxre ] and [ s=-1mxre ] , especially in the right panels of the former and left panels of the latter [ where the instability defined by eq .
( [ cr2 ] ) is less relevant ] , that the background - instability threshold given by eq .
( [ cr ] ) is an essential stability boundary for the family of the discrete solitons . of course
, additional instabilities due to the localized core part of the solution are possible too , and , as observed in these panels , they somewhat deform the resultant stability region .
the additional instabilities stemming from the excited in - phase rungs in fig .
[ s=1mxre ] , and their out - of - phase counterparts in fig .
[ s=-1mxre ] , are separately observed in the left panels of the former figure and right panels of the latter one .
given that this critical point was found in the framework of the acl , it features no @xmath3 dependence , but it clearly contributes to delimiting the stability boundaries of the discrete solitons ; sometimes , this effect is fairly dramatic , as in the middle - row left and right panels of figs .
[ s=1mxre ] and [ s=-1mxre ] , respectively , i.e. , the two - site , same - phase excitations may be susceptible to this instability mechanism . although the precise stability thresholds may be fairly complex , arising from the interplay of localized and extended modes in the nonlinear ladder system , a general conclusion is that the above - mentioned instabilities play a critical role for the stability of the localized states in this system ( see also the discussion below ) .
another essential conclusion is that the higher the coupling ( @xmath3 ) , the less robust the corresponding solutions are likely to be , the destabilization caused by the increase of @xmath3 being sometimes fairly dramatic . , determined by matrix @xmath37 in eq .
( [ lineq ] ) , for parameter values following the same pattern as in fig .
s=1abs , except that here @xmath162 varies along the horizontal axis . in the top two plots in the right column ,
the cyan line represents the analytically predicted critical value , @xmath163 ; this line originates from the cyan dot in the corner at @xmath164 , see eq .
( [ cr ] ) .
if an in - phase excited rung is present , then the second cyan dot is located at @xmath165 , in accordance with eq .
green lines indicate stability boundaries , between the dark region corresponding to stability [ or very weak instability , with @xmath166 , and the bright region corresponding to the instability . ] , but for parameter values from fig .
[ s=-1abs ] . here
the cyan line is drawn on the top two plots in the left column , and the second cyan dot on the @xmath58 axis appears only in the case where the out - of - phase excited rung is present at @xmath58 .
] the values of @xmath49 whose maximum real part is represented in figs .
[ s=1mxre ] and [ s=-1mxre ] were computed with the help of an appropriate numerical eigenvalue solver . at @xmath58 ,
the eigenvalues agree with eqs . and .
as shown in figs .
[ s=1eigs ] and [ s=-1eigs ] , following the variation of @xmath3 and @xmath19 , eigenvalues , associated with the empty ( zero - value ) sites , vary in accordance with the prediction of eq .
( [ zernu ] ) , and eigenvalues , associated with excited rungs , also shift in the complex plane upon variation of @xmath167 . in the case of the mixed - phase solutions with asymmetric amplitude ( seen in the bottom - most plot of fig .
[ s=1prof ] ) , there is a stable region for low values of the parameters @xmath3 and @xmath144 . for larger values of @xmath144
there are parametric intervals ( across @xmath55 for fixed @xmath19 ) in which discrete solitons with phase profiles different from those initialized in the acl of @xmath58 have been identified ; see the bottom two plots in fig .
[ s=1arguv ] .
these distinct branches of the unstable solutions give rise to streaks " observed in the bottom middle panel of fig .
[ s=1mxre ] .
the amplitude profiles of such alternate solutions are similar to those shown in the bottom plot of fig .
[ s=1prof ] , and the gain in width function defined in ( [ w ] ) as a function of @xmath3 is similar to the examples shown in the bottom plots of figs .
[ s=1abs ] , s=1w .
mechanisms by which solutions become unstable for these alternate solutions are outlined below . in the complex plane , for parameters chosen in accordance with the top - most left panel of fig .
[ s=1mxre ] with @xmath168 . for @xmath58 , in the top left plot we show the agreement of the numerically found eigenvalues ( blue circles ) with results produced by eqs .
( green filled circles ) and ( red filled ) . for @xmath169 , in the top right panel we show that the eigenvalues associated with the zero solution indeed lie within the predicted intervals , the boundaries of which are shown by dashed lines .
next , for @xmath170 , in the bottom left plot we observe that values of @xmath171 associated with the excited state have previously ( at smaller @xmath3 ) merged with the dashed intervals , and now an unstable quartet has emerged from the axis . for @xmath172 , in the bottom right panel the critical point corresponding to eq .
( [ cr ] ) is represented , where unstable eigenvalues emerge from the axis at the values of @xmath173 , as the intervals in eq .
comparing plots in the bottom row , we conclude that the critical point of the latter type gives rise , in general , to a stronger instability than the former one . ] , but for parameters chosen in accordance with the top right panel of fig .
s=-1mxre , with @xmath174 . at @xmath58 , in the top left plot we show the agreement of the numerically found eigenvalues ( blue circles ) with eqs .
( green filled circles ) and ( red filled ) . at @xmath175 ,
in the top right we see that eigenvalues associated to the red x s have moved inward towards zero .
next , for @xmath169 , in the bottom left panel we observe that , after merging with zero , the eigenvalues now emerge from zero on the imaginary axis .
finally , at @xmath176 in the bottom right panel , we observe that , after the eigenvalues merge with the dashed - line intervals , an unstable quartet emerges from the axis . ]
the most obvious type of the instability is associated with initializing a solution at @xmath58 from a single unstable rung , i.e. , at @xmath98 in ( [ cr2 ] ) .
eigenvalues for this type of the instability are shown in the top two panels of fig .
[ s=-1eigs ] .
there are three other scenarios of destabilization of the discrete solitons with the increase of @xmath3 , each corresponding to a particular type of a critical point ( transition to the instability ) .
these transitions are demonstrated in figs .
[ s=1eigs]-[s=-1eigs ] .
the first type occurs when eigenvalue @xmath177 associated with an excited rung collides with one of the intervals in eq . .
this weak instability generates an eigenfrequency quartet and is represented in figs .
[ s=1mxre ] and [ s=-1mxre ] , where the green boundary deviates ( as @xmath3 increases from @xmath178 ) from the threshold given by eq .
( [ cr2 ] ) . figures [ s=1eigs ] and [ s=-1eigs ]
illustrate this type of transition in more detail by plotting the eigenvalues directly in the complex plane .
the second type of the transition occurs when the intervals in eq .
come to overlap at @xmath179 , see eq .
( [ cr ] ) .
this is the background instability at empty sites , as shown in figs .
[ s=1mxre ] and [ s=-1mxre ] by bright spots originating from the corners of the diagrams , where @xmath77 . a more detailed plot of these eigenvalues and the corresponding collisions in the complex eigenvalue plane is displayed in fig . [ s=1eigs ] .
the third type of the instability onset occurs for essentially all values of @xmath3 in the case of two in - phase rungs at @xmath180 , or two out - of - phase ones at @xmath181 .
it may be thought of as a localized instability due to the simultaneous presence of two potentially unstable elements , due to the instability determined by eq .
( [ cr2 ] ) . at @xmath55 , it is seen as the bright spots in figs .
[ s=1mxre ] and s=-1mxre originating from @xmath182 .
the eigenvalues emerge from the corresponding zero eigenvalues at @xmath58 .
that is , in the middle - row left plot of fig .
s=1mxre at @xmath58 for @xmath183 there are four zero eigenvalues ; as @xmath3 increases , two of the four eigenvalues move from zero onto the real axis in the complex plane .
a similar effect is observed at @xmath183 in the middle - row right plot of fig .
[ s=-1mxre ] .
finally , it is worth making one more observation in connection , e.g. , to fig .
[ s=-1eigs ] and the associated jagged lines in the top right panel of fig .
[ s=-1mxre ] .
notice that , as @xmath3 increases , initial stabilization of the mode unstable due to the criterion given by eq .
( [ cr2 ] ) takes place , but then a collision with the continuous spectrum on the imaginary axis provides destabilization anew .
it is this cascade of events that accounts for the jaggedness of the curve in the top right of fig .
s=-1mxre and in similar occurrences ( e.g. , in the top left plot of fig .
[ s=1mxre ] ) .
we add this explanation to the set of possible instabilities discussed above , to explain the complex form of the stability boundaries featured by our two - dimensional plots . to verify the above predictions for the stability of the discrete solitons ,
we simulated evolution of the perturbed solutions in the framework of eq . by means of the standard runge - kutta fourth - order integration scheme . in figs .
[ type1prop ] , [ type2prop ] , and [ type3prop ] we display examples of the evolution of each of the three instability types which were identified in section [ cneq0 ] . for the first type , when the instability arises from the collision of eigenvalues associated with the excited and empty rungs , the corresponding unstable eigenmode arises in the form of a quartet of eigenfrequencies . in fig .
[ type1prop ] we demonstrate that this instability leads to the growth of the solution amplitudes and oscillations at the central rung .
the corresponding ( chiefly localized , although with a weakly decaying tail ) instability eigenvectors are shown in the top panels of the figure , while the bottom panels show how the initial conditions evolve in time through the oscillatory growth , in accordance with the presence of the unstable complex eigenfrequencies . for @xmath170 and @xmath168 .
the complex plane of all the eigenvalues for this solution is shown in the bottom left plot of fig .
[ s=1eigs ] .
the top right plot shows the squared absolute values of perturbation amplitudes @xmath184 ( higher amplitudes ) and @xmath185 ( lower amplitudes ) , defined in eq . .
the top left plot shows the solution at @xmath186 , and the bottom left plot shows the solution at @xmath187 with @xmath188 in blue and @xmath189 in green .
in the course of the evolution , the soliton maintains its shape , while the amplitude at the central rung ( @xmath134 ) grows with oscillations ; the growth on the gain side , associated to @xmath190 , is ultimately dominant .
quantities @xmath191 and @xmath192 for the second type of the instability , which arises from the collision of eigenvalues in intervals , which are all associated with empty rungs , the corresponding unstable eigenmode is delocalized .
it is shown in fig .
[ type2prop ] that the corresponding unstable soliton does not preserve its shape .
instead , the instability causes delocalization of the solution , which acquires a tail reminiscent of the spatial profile of the corresponding unstable eigenvector . and @xmath168
. the complex plane of all the eigenvalues for this solution is shown in the bottom right plot of fig .
[ s=1eigs ] .
the top right plot has the same meaning as in fig .
[ type1prop ] , where here @xmath193 are mostly zero while @xmath194 have nonzero amplitudes .
in the course of the evolution the soliton does not maintain its shape .
in particular , the solution profiles at @xmath195 are shown in the left panel at the bottom bearing the apparent signature of the delocalized , unstable eigenmode ; the delocalization is stronger in the @xmath196 solution associated with gain .
similar to fig .
[ type1prop ] we plot @xmath197 in the bottom right plot .
this plot shows that the central node experiences oscillations similar to fig .
[ type1prop ] but the oscillatory effect is dominated by the delocalization seen in the bottom left plot , which grows and exceeds past the shorter peaks in the center . ]
lastly , the third type of the instability is shown in fig .
[ type3prop ] .
it displays the case of two excited in - phase rungs at @xmath198 .
other examples of the same type are similar
e.g. , with two out - of - phase excited rungs at @xmath17 .
the instability has a localized manifestation with the amplitudes growing at the gain nodes of each rung and decaying at the loss ones . for @xmath199 and @xmath200 .
the complex plane including all the eigenvalues for this solution is similar to the top right plot in fig .
[ s=-1eigs ] .
the top right plot here shows the squared absolute values of perturbation amplitudes @xmath201 ( lower amplitudes ) and @xmath202 ( higher amplitudes ) defined in eq . .
the top left plot shows the solution at @xmath186 , and the bottom left plot shows the solution at @xmath203 with @xmath204 in blue and @xmath205 in green .
in the course of the evolution , the soliton maintains its shape , while the amplitude at the rung @xmath135 grows with weak oscillations and with higher growth on the @xmath206 side associated with gain .
quantities @xmath207 ( blue ) , @xmath208 ( green ) , @xmath209 ( cyan ) and @xmath210 ( red ) are shown in the bottom right plot to highlight the growth . ]
we have introduced the lattice of the ladder type with staggered pairs of mutually compensated gain and loss elements at each rung , and the usual onsite cubic nonlinearity , self - focusing or defocusing .
this nearly - one - dimensional system is the simplest one which features__two - dimensional _ _ @xmath0 symmetry . it may be realized in optics as a waveguide array .
we have constructed families of discrete stationary solitons seeded by a single excited rung , or a pair of adjacent ones , in the anti - continuum limit of uncoupled rungs .
the seed excitations may have the in - phase or out - of - phase structure in the vertical direction ( between the gain and loss poles ) .
the double seed with the in- and out - of - phase structures in the two rungs naturally features an asymmetric amplitude profile .
we have identified the stability of the discrete solitons via the calculation of eigenfrequencies for small perturbations , across the system s parameter space .
a part of the soliton families are found to be dynamically stable , while unstable solitons exhibit three distinct scenarios of the evolution .
the different scenarios stem , roughly , from interactions of localized modes with extended ones , from extended modes alone , or from localized modes alone . a natural extension of the work may be the consideration of mobility of kicked discrete solitons in the present ladder system .
it may also be interesting to seek nonstationary solitons with periodic intrinsic switching , cf .
@xcite . a challenging perspective is the development of a 2d extension of the system .
effectively , this would entail adding further alternating ladder pairs along the transverse direction and examining 2d discrete configurations
. it may be relevant in such 2d extensions to consider different lattice settings that support not only solutions in the form of discrete solitary waves but also ones built as discrete vortices , similarly to what has been earlier done in the dnls system @xcite , and recently in another 2d @xmath1-symmetric system @xcite .
d. n. christodoulides , f. lederer and y. silberberg , discretizing light behaviour in linear and nonlinear waveguide lattices , nature * 424 * , 817 - 823 ( 2003 ) ; f. lederer , g. i. stegeman , , d. n. christodoulides , g. assanto , m. segev and y. silberberg , discrete solitons in optics , " phys .
rep . * 463 * , 1 - 126 ( 2008 ) ; i. l. garanovich , s. longhi , a. a. sukhorukov , and y. s. kivshar , light propagation and localization in modulated photonic lattices and waveguides , " phys .
518 * , 1 - 79 ( 2012 ) ; z. chen , m. segev and d. n. christodoulides,optical spatial solitons : historical overview and recent advances , " rep .
75 * , 086401 ( 2012 ) .
b. a. malomed and p. g. kevrekidis , discrete vortex solitons , " phys .
e * 64 * , 026601 ( 2001 ) ; p. g. kevrekidis , b. a. malomed , and yu . b. gaididei , solitons in triangular and honeycomb dynamical lattices with the cubic nonlinearity , " phys .
e * 66 * , 016609 ( 2002 ) ; p. g. kevrekidis , b. a. malomed , z. chen , and d. j. frantzeskakis ,
stable higher - order vortices and quasivortices in the discrete nonlinear schrdinger equation , " phys .
e * 70 * , 056612 ( 2004 ) ; m. ster and m. johansson ,
stable stationary and quasiperiodic discrete vortex breathers with topological charge s=2 , " phys . rev .
e * 73 * , 066608 ( 2006 ) ; meja - corts , j. m. soto - crespo , m. i. molina , and r. vicencio , dissipative vortex solitons in two - dimensional lattices , " phys .
a * 82 * , 063818 ( 2010 ) .
d. neshev , t. j. alexander , e. a. ostrovskaya , y. s. kivshar , h. martin , i. makasyuk , and z. chen , observation of discrete vortex solitons in optically induced photonic lattices , " phys . rev . lett . * 92*,123903 ( 2004 ) ; j. w. fleischer , g. bartal , o. cohen , o. manela , m. segev , j. hudock , and d. n. christodoulides , observation of vortex - ring discrete solitons in 2d photonic lattices , " phys .
rev . lett . * 92 * , 123904 ( 2004 ) ; b. terhalle , t. richter , a. s. desyatnikov , d. n. neshev , w. krlikowski , f. kaiser , c. denz , and y. s. kivshar , observation of multivortex solitons in photonic lattices " , phys .
* 101 * , 013903 ( 2008 ) .
n. k. efremidis , s. sears , d. n. christodoulides , j. w. fleischer , and m. segev , discrete solitons in photorefractive optically induced photonic lattices , " phys .
e * 66 * , 046602 ( 2002 ) ; j. w. fleischer , m. segev , n. k. efremidis , and d. n. christodoulides , observation of two - dimensional discrete solitons in optically induced nonlinear photonic lattices , " nature * 422 * , 147 - 150 ( 2003 ) .
a. a. sukhorukov , y. s. kivshar , h. s. eisenberg , and y. silberberg , spatial optical solitons in waveguide arrays , " ieee j. quantum electron .
* 39 * , 31 - 50 ( 2003 ) ; a. szameit , j. burghoff , t. pertsch , s. nolte , a. tnnermann , and f. lederer , two - dimensional soliton in cubic fs laser written waveguide arrays in fused silica , " opt . exp . * 14 * , 6055 - 6062 ( 2006 ) ; a. szameit and s. nolte , discrete optics in femtosecond - laser - written photonic structures , " j. phys .
b : at . mol .
* 43 * , 163001 ( 2010 ) .
t. schwartz , g. bartal , s. fishman , and m. segev , transport and anderson localization in disordered two - dimensional photonic lattices , " nature * 446 * , 52 - 55 ( 2007 ) ; y. lahini , a. avidan , f. pozzi , m. sorel , r. morandotti , d. n. christodoulides , and y. silberberg , anderson localization and nonlinearity in one - dimensional disordered photonic lattices , " phys . rev . lett . * 100 * , 013906 ( 2008 ) .
a. ruschhaupt , f. delgado , and j. g. muga , physical realization of @xmath0-symmetric potential scattering in a planar slab waveguide , " j. phys .
a * 38 * , l171-l176 ( 2005 ) ; k. g. makris , r. el - ganainy , d. n. christodoulides , and z. h. musslimani , beam dynamics in @xmath1 symmetric optical lattices , `` phys . rev . lett . * 100 * , 103904 ( 2008 ) ; s. longhi , ' ' spectral singularities and bragg scattering in complex crystals,phys
a * 81 * , 022102 ( 2010 ) ; c. e. rter , k. g. makris , r. el - ganainy , d. n. christodoulides , m. segev , and d. kip , observation of parity - time symmetry in optics , " nature phys . * 6 * , 192 - 195 ( 2010 ) ; k. g. makris , r. el - ganainy , d. n. christodoulides , and z. h. musslimani ,
@xmath0 symmetric periodic optical potentials , " int
. phys . * 50 * , 1019 - 1041 ( 2011 ) .
r. driben and b. a. malomed , stability of solitons in parity - time - symmetric couplers , opt .
36 * , 4323 - 4325 ( 2011 ) ; n. v. alexeeva , i. v. barashenkov , a. a. sukhorukov , and y. s. kivshar , optical solitons in @xmath0-symmetric nonlinear couplers with gain and loss phys .
a * 85 * , 063837 ( 2012 ) ; yu .
v. bludov , r. driben , v. v. konotop , and b. a. malomed , instabilities , solitons and rogue waves in @xmath211 -coupled nonlinear waveguides " , j. opt . * 15 * , 064010 ( 2013 ) .
a. e. miroshnichenko , b. a. malomed , and y. s. kivshar ,
nonlinearly @xmath0-symmetric systems : spontaneous symmetry breaking and transmission resonances " , phys . rev .
a * 84 * , 012123 ( 2011 ) ; s. v. dmitriev , s. v. suchkov , a. a. sukhorukov , and y. s. kivshar , scattering of linear and nonlinear waves in a waveguide array with a @xmath1-symmetric defect , " _ ibid_. * 84 * , 013833 ( 2011 ) ; s. v. suchkov , a. a. sukhorukov , s. v. dmitriev , and y. s. kivshar , scattering of the discrete solitons on the @xmath0-symmetric defects , " epl * 100 * , 54003 ( 2012 ) ; a. regensburger , m. a. miri , c. bersch , and j. nger , observation of defect states in @xmath0-symmetric optical lattices , " phys .
lett . * 110 * , 223902 ( 2013 ) .
j. dambroise , p. g. kevrekidis , and s. lepri , asymmetric wave propagation through nonlinear @xmath211-symmetric oligomers `` , j. phys .
a * 45 * , 444012 ( 2012 ) ; j. dambroise , p. g. kevrekidis , and s. lepri , eigenstates and instabilities of chains with embedded defects '' , chaos * 23 * , 023109 , 10 pages ( 2013 ) .
w. h. weber and g. w. ford , propagation of optical excitations by dipolar interactions in metal nanoparticle chains " phys .
b * 70 * , 125429 ( 2004 ) ; p. a. belov and c. r. simovski , homogenization of electromagnetic crystals formed by uniaxial resonant scatterers " phys rev .
e * 72 * , 026615 ( 2005 ) ; i.v .
shadrivov , a.a .
zharov , n.a .
zharova , yu.s .
kivshar , `` nonlinear magnetoinductive waves and domain walls in composite metamaterials '' , photonics and nanostructures fundamentals and applications * 4 * , 69 ( 2006 ) ; r.e .
noskov , p.a . belov and yu.s .
kivshar , subwavelength modulational instability and plasmon oscillons in nanoparticle arrays " phys .
108 * , 093901 ( 2012 ) .
x. zhang , j. chai , j. huang , z. chen , y. li , and b. a. malomed , discrete solitons and scattering of lattice waves in guiding arrays with a nonlinear @xmath0-symmetric defect , opt . exp . * 22 * , 13927 - 13939 ( 2014 ) .
k. li and p. g. kevrekidis @xmath211-symmetric oligomers : analytical solutions , linear stability and nonlinear dynamics `` , phys .
e * 83 * , 066608 ( 2011 ) .
d. a. zezyulin , and v. v. konotop , nonlinear modes in finite - dimensional @xmath1-symmetric systems , '' phys .
lett . * 108 * , 213906 ( 2012 ) .
|
we introduce a ladder - shaped chain with each rung carrying a @xmath0-symmetric gain - loss dimer .
the polarity of the dimers is _ staggered _ along the chain , meaning alternation of gain - loss and loss - gain rungs .
this structure , which can be implemented as an optical waveguide array , is the simplest one which renders the system @xmath0-symmetric in both horizontal and vertical directions .
the system is governed by a pair of linearly coupled discrete nonlinear schrdinger ( dnls ) equations with self - focusing or defocusing cubic onsite nonlinearity .
starting from the analytically tractable anti - continuum limit of uncoupled rungs and using the newton s method for continuation of the solutions with the increase of the inter - rung coupling , we construct families of @xmath0-symmetric discrete solitons and identify their stability regions .
waveforms stemming from a single excited rung and double ones are identified .
dynamics of unstable solitons is investigated too .
|
since the early days of high - energy gamma - ray astronomy it has become clear that the gamma - ray flux received at earth is dominated by emission from the galactic disk @xcite .
this emission can be well understood in terms of cosmic - ray interactions with the interstellar medium @xcite . at energies
@xmath0 mev , the generation of diffuse gamma - ray emission is dominated by the decay of @xmath1 produced in collisions between cosmic - ray nuclei and interstellar medium nuclei . ultimately , the study of this hadronic gamma - ray emission may provide hints on the still mysterious origin of the galactic cosmic - rays .
however , the interpretation of the galactic diffuse gamma - ray emission is complicated by the fact that a large number and variety of individual sources contribute along the line of sight to the observed emission , thus blurring the link between individual cosmic - ray acceleration sites and observed gamma - ray signatures in our galaxy .
gamma rays from cosmic - ray interactions are also expected from nearby galaxies , and indeed , the egret telescope aboard the compton gamma - ray observatory ( _ cgro _ ) has for the first time detected gamma - ray emission from the large magellanic cloud ( lmc ) @xcite .
the lmc is an excellent target for studying the link between cosmic - ray acceleration and gamma - ray emission since this galaxy is nearby ( bringing the sources fluxes in reach of modern gamma - ray telescopes ) and since the system is nearly seen face - on ( avoiding the superposition of sources along the line of sight that hampers studies in our own galaxy ) .
in addition , the lmc is rather active , housing many supernova remnants , bubbles and superbubbles and massive star forming regions that are all potential sites of cosmic - ray acceleration @xcite .
the large area telescope ( lat ) aboard the _ fermi gamma - ray space telescope _ ( fgst ) provides now the capabilities to study diffuse gamma - ray emission from nearby galaxies in depth , and of the lmc in particular @xcite .
we report here on the initial analysis of observations taken in the course of the first year s all - sky survey by the lat .
the lat is the primary instrument on the fgst satellite which has been launched from cape canaveral on june 11th , 2008 .
the lat is an imaging , wide field - of - view , high - energy gamma - ray telescope , covering the energy range from below 20 mev to more than 300 gev .
the lat is a pair - conversion telescope with a precision tracker made of a stack of 18 x , y silicon tracking planes and a calorimeter made of 96 csi(tl ) crystals .
the tracker array is covered by a segmented anticoincidence shield allowing for the rejection of charged particle backgrounds .
the lat has a large @xmath2 sr field of view , and compared to earlier gamma - ray missions , has a large effective area ( @xmath3 @xmath4 on axis at @xmath5 gev for the event selection used in this paper ) , improved angular resolution ( @xmath6 68% containment radius at 1 gev ) and low dead time ( @xmath7 @xmath8s per event ) .
the @xmath9 energy resolution in the 100 mev - 10 gev energy range is better than @xmath10% .
a detailed description of the instrument is given by @xcite .
the on - orbit instrument calibration is presented by @xcite .
the data used in this work covers the period august 8th 2008 april 24th 2009 and amounts to 211.7 days of continuous sky survey observations . during this period
a total exposure of @xmath11 cm s@xmath12 ( at 1 gev ) has been obtained for the lmc region .
the data analysis presented in this paper has been performed using the version v9r11 and the instrument response functions p6_v3 .
we collected all data obtained within the period august 8th 2008 april 24th 2009 and applied the _ diffuse event class _ filter that has been designed to minimize contamination by instrumental background while retaining a substantial fraction of the signal .
as has been pointed out by @xcite , any harsher event cut would not significantly improve the signal - to - noise ratio .
we further excluded from the data all periods where the spacecraft has entered the south atlantic anomaly ( saa ) and for which the spacecraft z - axis points more than @xmath13 away from the zenith direction ( the zenith direction being defined by the vector running from the earth center through the spacecraft ) .
while the saa cut excludes periods of particular large instrumental background from the analysis , the latter cut excludes periods where the earth enters the field of view .
furthermore , to minimize contamination from earth albedo photons we exclude photons with zenith angles above @xmath14 from the analysis .
we further restricted the analysis to photon energies above 200 mev where our current knowledge of the instrument response implies systematic uncertainties that are smaller than @xmath15 and where the redistribution of photons in energy due to incomplete energy measurements becomes negligible . to illustrate the distribution of observed gamma - ray photons in the lmc region we show in fig .
[ fig : image ] a counts - map of the area .
the arrival directions of observed photons in the 200 mev - 100 gev energy range have been binned into @xmath16 large pixels covering a @xmath17 large area around the position @xmath18 .
the binned map has then been smoothed using a 2d adaptive gaussian kernel smoothing technique @xcite to remove poissonian noise that arises from the relatively small number of counts that have been registered .
the signal - to - noise ratio ( s.n.r . ) has been set to 10 to reduce statistical noise variations to below @xmath19 in the image . )
_ fermi_/lat counts map of a @xmath17 large region centered on the lmc for the energy range 200 mev - 100 gev ( greyscale ) .
the contours show the extinction map of @xcite as tracer of the total gas column density in the lmc .
ten linearly spaced contour levels are plotted .
the diamond in the north - east of the image designates the location of the blazar crates j060106 - 703606 @xcite that contributes at a low level to the gamma - ray emission in this area .
[ fig : image ] , width=377 ] we overlay as contours on the _ fermi_/lat counts map the extinction map of @xcite as tracer of the total gas column density in the lmc . to first order
the extinction scales linearly with total gas column density , and we chose 10 linearly spaced contours to allow the reader to visually appreciate the distribution of gas column densities in the lmc .
obviously , a substantial fraction of the gas is found in a small area in the north of the lmc , at roughly @xmath20 , which coincides with the 30 doradus star forming region . in this area of the sky .
[ fig : profile ] , width=377 ] the high - energy gamma - ray photons that are observed from the lmc also peak in this area .
the photon intensity in the 30 dor region exceeds @xmath21 counts / deg@xmath12 while in most of the remaining regions of the lmc is remains below @xmath22 counts / deg@xmath12 ( the background rate around the lmc is around @xmath23 counts / deg@xmath12 ) .
the excess near 30 dor is also clearly seen in the longitude and latitude profiles of the photon intensity observed by lat that is shown in fig .
[ fig : profile ] . within a rectangular box covering galactic longitudes @xmath24 and galactic latitudes
@xmath25 we find a total number of @xmath26 counts within the energy range 200 mev - 100 gev above the expected contributions from galactic diffuse emission , extragalactic diffuse emission , instrumental background , and the blazar crates j060106 - 703606 .
these background contributions have been estimated by fitting spatial and spectral templates of their emission components together with a spatial template for the lmc emission to the data .
galactic diffuse emission has been modelled spatially and spectroscopically using the galprop model @xcite version 54_59xvarh7s while for the combination of extragalactic diffuse emission and residual instrumental background we assume an isotropic emission with a power law spectral distribution .
crates j060106 - 703606 is modeled as a point source at @xmath27 with a power - law spectral distribution . for the lmc we use the extinction map of @xcite as spatial template
from which we subtract a pedestal level of @xmath28 from all pixels and for which we set all pixels outside a radius of @xmath29 around @xmath30 to zero in order to extract the lmc emission .
as spectral model we assume a power law for the lmc . to describe the morphology of the high - energy gamma - ray emission from the lmc we first fit a point source with free position and flux on top of the background model to our data .
this results in a best - fitting point - source position of @xmath31 with a statistical 95% confidence error radius of @xmath32 ( the systematic position uncertainty is estimated to less than @xmath33 ) .
we note that this position is close to that of r 136 , the central star cluster of 30 dor , which is located at @xmath34 , i.e. at an angular distance of @xmath35 from our best - fitted point - source location .
the detection significance of the lmc can be estimated using the so - called _ test statistics _ ( ts ) which is defined as twice the difference between the log - likelihood @xmath36 that is obtained by fitting the lmc model on top of the background model to the data , and the log - likelihood @xmath37 that is obtained by fitting the background model only , i.e. @xmath38 . under the hypothesis that our model satisfactorily explains the _
fermi_/lat data , ts follows a @xmath39 distribution with @xmath40 degrees of freedom , where @xmath40 is the number of free parameters in the lmc model @xcite . in the particular case of a point source with free position , flux and spectral index we have @xmath41 and the measured ts of 869.1 corresponds to a significance of @xmath42 . as next step
we replace the point source model by an extended source model which we implement as axisymmetric 2d gaussian shape with variable angular size @xmath43 .
in addition to the size we again fitted the position , flux and power law spectral index of the source .
this results in a best - fitting source position of @xmath44 ( with a 95% confidence radius of @xmath45 ) , and source extent of @xmath46 .
the ts amounts to 1088.5 which is larger by 219.4 than the value obtained for the point - source model .
since we added one additional parameter ( the source extent @xmath43 ) with respect to the point - source model we obtain the significance of the source extension from the @xmath47 distribution to @xmath48 . alternatively to the geometrical models we also compare the _
fermi_/lat data to various spatial templates that trace the interstellar matter distributions in the lmc .
for neutral hydrogen ( h i ) we use the aperture synthesis and multibeam data that @xcite have combined from atca and parkes observations . for molecular hydrogen
we use co observations of the lmc obtained with the nanten telescope @xcite .
we further used the extinction map of @xcite ( sfd ) as tracer of the total gas column density and compare also our data to the 100 @xmath8 m iris map that has been obtained by reprocessing the iras survey data @xcite .
the results of this comparison are summarized together with that of the geometrical models in table [ tab : models ] .
the best fits are obtained for the sfd extinction map and the iras 100 @xmath8 m infrared map which give ts values of @xmath49 and @xmath50 , respectively .
for 2 free parameters ( the total flux in the map and the spectral index ) this corresponds to a detection significance of @xmath51 .
an almost equally good fit is obtained using the neutral hydrogen map .
fitting instead the co map to the lat data provides a rather poor fit , suggesting that the gamma - ray morphology differs from that of molecular gas in the lmc .
fitting the h i and co maps together to the data confirms this result since the fit attributes @xmath52 of the total flux to the h i component .
correspondingly , the ts increase with respect to fitting the h i gas map alone is also negligible . the h i / sfd / iris 100 @xmath8 m maps fit the data considerably better than a single point source , adding further evidence that the observed high - energy gamma - ray emission is extended in nature .
furthermore , the 2d gaussian source model can not reach the fit that is obtained by those tracer maps , suggesting that the emission morphology is more complex than a single gaussian shape .
lrc lmc model & ts & parameters + point source & 869.1 & 4 + 2d gaussian source & 1088.5 & 5 + h i gas map & 1173.4 & 2 + co gas map & 932.2 & 2 + h i + co gas maps & 1176.1 & 4 + sfd extinction map & 1179.6 & 2 + iris 100 @xmath8 m infrared map & 1179.1 & 2 + , width=377 ] using the extinction map of @xcite ( i.e. our best fitting spatial template of the high - energy emission ) we extract a spectrum of the lmc by fitting the data in 12 logarithmically - spaced energy bins covering the energy range 200 mev - 20 gev . above 20 gev , photons from the lmc
become too sparse in our actual data set to allow for meaningful spectral points to be derived .
figure [ fig : spectrum ] shows the lmc spectrum that has been obtained by this method .
our analysis indicates a spectral steepening of the emission with increasing energy , suggesting that a simple power law is an inadequate description of the data .
we confirm this trend by fitting the data using a broken power law instead of a simple power law .
this results in an improvement of the ts by @xmath53 corresponding to a significance of @xmath54 ( @xmath55 ) of the spectral steepening .
fitting alternatively an exponentially cutoff power law improves the ts by @xmath56 with respect to the simple power law , corresponding to a significance of @xmath57 ( @xmath58 ) of the spectral cutoff . integrating the broken power law or the exponentially cutoff power law model over the energy range 100 mev
500 gev gives identical photon fluxes of @xmath59 and an energy fluxes of @xmath60 for the lmc .
the systematic uncertainty in these flux measurements amounts to @xmath15 .
@xcite reported the first detection of the lmc in @xmath61 mev gamma rays based on 4 weeks of data collected with the egret telescope aboard _
cgro_. due to egret s limited angular resolution and the weak emission detected from the lmc , details of the spatial structure of the galaxy could not have been resolved .
however , it has been obvious from egret data that the lmc was an extended gamma - ray source . _
fermi_/lat allows now for the first time to clearly resolve the gamma - ray emission of the lmc and to attribute the emission maximum to the 30 dor star forming region .
while this coincidence could be taken as a hint for an enhanced cosmic - ray density in 30 dor with respect to the rest of the galaxy , we note that a substantial fraction of the interstellar gas of the lmc is also confined to the 30 dor area .
consequently , the target density for cosmic - ray interactions is greatly enhanced in this region which implies a corresponding enhancement of the gamma - ray luminosity . whether the data do also support an enhanced cosmic - ray density in 30 dor with respect to the rest of the galaxy needs a more detailed analysis of the observations
. the rather poor fit of the co map to the lat data suggests that the overall distribution of gamma - ray emission differs from that of molecular hydrogen .
the distribution of neutral hydrogen fits the data considerably better and the combined fit of h i and co maps indicates that any contribution to the gamma - ray emission that is correlated to the molecular gas is at best marginal .
this agrees well with expectations since the gas budget of the lmc is largely ( 90 - 95% ) dominated by neutral hydrogen @xcite .
consequently we are presently unable to determine the co - to - h@xmath62 conversion factor , x@xmath63 , from our lmc data .
@xcite performed a detailed modelling of the cosmic - ray distribution in the lmc and predicted an integrated @xmath61 mev photon flux of @xmath64 for the galaxy .
@xcite predicted an integrated @xmath61 mev photon flux of @xmath65 based on estimates of the lmc supernova rate and total gas densities .
our observed flux of @xmath59 falls at the hide side of these estimates , yet given the uncertainties in the models the agreement can be judged satisfactorily .
further studies of the lmc with _
fermi_/lat will now concentrate on the spectral analysis of the data , with particular emphasize on variations of the spectral shape throughout the galaxy .
thanks to the excellent sensitivity and angular resolution of the lat , this is the first time that such studies become possible . and other nearby galaxies await their detection , such as the small magellanic cloud or the andromeda galaxy ( m31 ) .
both should be in reach of _ fermi _ and the comparative study of their diffuse gamma - ray emission should help to understand the impact of the environment and metallicity on the physics of cosmic rays .
|
we report on observations of the large magellanic cloud with the _ fermi gamma - ray space telescope_. the lmc is clearly detected with the large area telescope ( lat ) and for the first time the emission is spatially well resolved in gamma - rays .
our observations reveal that the bulk of the gamma - ray emission arises from the 30 doradus region .
we discuss this result in light of the massive star populations that are hosted in this area and address implications for cosmic - ray physics .
we conclude by exploring the scientific potential of the ongoing fermi observations on the study of high - energy phenomena in massive stars .
|
exploration of the properties of quark gluon plasma ( qgp ) produced in relativistic heavy ion collisions represents a major theme of modern nuclear physics research .
the formation of qgp is heralded by numerous signatures , e.g. , jet - quenching @xcite , elliptic flow of hadrons @xcite , quark number scaling of elliptic flow of hadrons@xcite , radiation of thermal photons @xcite , and suppressed production of @xmath6 @xcite , etc .
the celebrated phenomenon of jet - quenching has its origin in energy loss suffered by high energy quarks and gluons as they traverse the qgp , colliding with other partons and radiating gluons .
it leads to a suppressed production of hadrons having large transverse momenta as compared to the case for nn collisions at the corresponding centre of mass energy per nucleon .
it is measured in terms of the nuclear modification factor @xmath7 , given by : @xmath8 where @xmath9 is the impact parameter and @xmath10 is the nuclear overlap function for impact parameter @xmath9 .
the jet - quenching has several other manifestations .
thus , for example , in a non - central collision the jets moving in and out of the reaction plane would cover differing distances inside the plasma , lose differing amounts of energy , and would be quenched to a different extent .
this will lead to an azimuthal anisotropy of momentum distribution of hadrons which does not have its origin in the so - called elliptic flow @xcite .
next consider a high energy quark which is produced in a hard scattering .
a photon fragmented off this quark contributes to the prompt photon production at high transverse momenta
. the quark may lose energy before its fragmentation and thus this photon production may get suppressed @xcite .
following the same arguments as above , this may also lead to an azimuthal anisotropy in photon production for non - central collisions @xcite .
the jet - photon conversion @xcite and jet - induced bremsstrahlung @xcite lead to very small contributions with opposite signs for the azimuthal anisotropy of the photon productions @xcite and are not considered in the present work . a new dimension to the azimuthal asymmetry and
the jet - quenching has been added by the realization that the event by event fluctuations of the initial conditions may affect the elliptic flow of the final state hadrons @xcite .
it is not very clear these seriously affect the azimuthal variation of the transverse momenta , which arises from jet - quenching @xcite as well .
it is expected that the magnitude of all these effects will depend on the dynamics of the collision and the properties of the qgp .
this makes jet - quenching a powerful probe of the qgp . while there are quite a few detailed studies on the jet - quenching at rhic and lhc energies @xcite , in the present work we continue the study reported in ref .
@xcite , where the wang , huang , and sarcevic model @xcite was used to study the evolution of the mechanism of energy loss at rhic and lhc energies in central collision of heavy nuclei .
the model treats the medium produced in the collision as a plasma of length @xmath11 through which the partons move , undergoing multiple collisions , losing energy through radiated gluons , before fragmenting to form hadrons . the only parameters which then enter the calculations are the path - length- @xmath11 , the mean - free path- @xmath12 , and the energy loss per collision- @xmath1 .
the energy loss per collision is then assumed to take a form as proportional to the energy of the parton- @xmath3 , @xmath4 , or a constant ; depending upon the formation time of the radiated gluons , which is less than the mean free path , larger than the mean free path but less than the path length @xmath11 , and greater than the path length @xmath11 , respectively .
the three regimes are called the bethe - heitler regime , the lpm regime , and the complete coherence regime ( as the entire medium radiates as a whole ) @xcite . the formation time
( or length ) of the radiated gluon is @xmath13 , where @xmath14 is the energy of the gluon and @xmath15 is its transverse momentum . in the present work
we explore the limits of the above approach by performing a detailed and systematic study of medium modification of production of neutral pions and prompt photons having large transverse momenta as a function of centrality and system size , by analyzing the corresponding data at 200 gev / nucleon for collisions of gold nuclei and copper nuclei .
we first seek an accurate description of the medium modification parameter @xmath7 for different centralities and @xmath16 , using the three approaches for the energy loss mechanisms , by varying the coefficients of energy loss , while accounting for the variation of the path length @xmath11 .
we find that the hadrons having low @xmath16 ( @xmath17 5 gev/@xmath18 ) fall into the bethe - heitler regime , those having intermediate @xmath16 ( 5 gev/@xmath18 @xmath17 @xmath16 @xmath17 10 gev/@xmath18 ) fall into the lpm regime , while those having large @xmath16 @xmath19 810 gev/@xmath18 are covered by the complete coherence regime .
we find a systematic decrease of the energy loss per collision as we go to more peripheral collisions from central collisions , as one might expect .
we then use the same parameters to determine the azimuthal anisotropy coefficient @xmath20 for neutral pions having large @xmath16 .
we find that in general , the experimentally obtained @xmath20 is lower than those predicted by our calculations , though the trend is generally reproduced .
part of this discrepancy may have its origin in the uniform densities for the nuclei , used in the present work .
next we calculate the prompt photon production as well as @xmath21 .
again we get reasonable description of the experimental data .
further improvements can perhaps be made by relaxing the condition of equality of mean - free path and energy loss per collision for quarks and gluons assumed while obtaining a description for @xmath22 .
this would , however , increase the number of parameters .
we give our formulation for particle production , multiple scattering , and energy loss as applied to pion production at @xmath23 in sec .
the centrality dependence of nuclear modification of neutral pions and parton energy loss ( de / dx ) for au - au and cu - cu collisions at the highest rhic energy are discussed in sec .
the results for the azimuthal asymmetry of high @xmath16 neutral pions are shown in sec .
4 . in sec . 5
, we give the results for the nuclear suppression and azimuthal anisotropy of prompt photons which arise due to the modification of the fragmentation contribution .
finally we summarize our findings in sec . 6 .
in perturbative qcd , the inclusive cross - section for the production of a particle in proton - proton collisions is given by @xcite = _ a , b , c dx_a dx_b & & f_a / a(x_a , q^2 ) f_b / b(x_b , q^2 ) + & & d_c / c(z , q^2_f ) , where @xmath24 is the parton distribution function ( pdf ) for the parton @xmath25 and @xmath26 is the pdf for the parton @xmath9 , for the nucleons a and b respectively .
@xmath27 gives the fragmentation probability of parton @xmath18 into a particle @xmath28 . in case of a hadron ; @xmath29 is the fragmentation function evaluated at @xmath30 , where z is the fraction of the parton s momentum carried by the hadron . in case of photon production , @xmath31 gives the fragmentation probability of a photon fragmented off a quark with the momentum fraction @xmath32 .
in addition , we have an extra term where photon is directly produced in the hard collision ( @xmath33 ) . in that case
the fragmentation function reduces to @xmath34 .
@xmath35 is the parton - parton cross - section , calculated for the leading order processes @xmath36 such as : q+q q+q , + q+|q q+|q , + q+g q+g , + g+g g+g , + ...... at the next - to - leading order , @xmath37 we include subprocesses like : q+q q+q+g , + q+|q q+|q+g , + q+q^ q+q^+g , + q+|q q^+|q^+g , + g+g g + g+ g + ...... the running coupling constant @xmath38 , is calculated at next - to - leading order @xmath39 where @xmath40 is the renormalization scale , @xmath41 is the number of flavours , and @xmath42 is the @xmath43 scale .
we have used the _
cteq4 m _ structure function @xcite and the bkk fragmentation function @xcite .
we set the factorization , renormalization , and fragmentation scales as equal to @xmath16 , though we have checked scale dependence on particle production in pp collisions using @xmath44 and @xmath45 as well .
= 200 gev with nlo pqcd calculations.,width=434 ] we show the results of our calculations for the inclusive neutral pion yield for proton - proton collisions at the centre of mass energy of 200 gev along with the data from the phenix measurements @xcite in fig . [ fig1 ] . a very good agreement with the experimental data
is seen , as in similar calculations by other authors ( see e.g. , ref .
the quantitative description of the data from pp collisions provides us with a reliable base - line to discuss nuclear modification of hadron production in nucleus - nucleus collisions . in order to calculate the inclusive pion spectra from nucleus - nucleus collisions we have used the effects of shadowing and energy loss of partons .
the nuclear shadowing accounts for the modification of the parton distribution function for nuclei compared to that for a free nucleon .
the shadowing function is defined as : @xmath46 where @xmath24 is the parton distribution for the nucleus a and @xmath47 is that for a nucleon .
we have used the eks98 parametrization of nuclear shadowing obtained by eskola , kolhinen , and salgado @xcite .
we use the simple phenomenological model , first used by jeon et al .
@xcite to obtain a description of preliminary @xmath7 at rhic energies , to estimate parton energy loss .
the key features of this treatment are : average energy loss per collision ( @xmath48 ) by the parton , the mean free path of the parton ( @xmath49 ) , and the average path length of the parton in the medium ( l ) .
the probability for a parton to scatter @xmath50 times before leaving the medium can be written as : @xmath51 the multiple scattering and energy loss suffered by the partons then necessitates a modification of the fragmentation function @xmath52 . following the model of wang , huang , and
sarcevic @xcite , the modified fragmentation function can be written as : zd_c / h(z , l , q^2)&=&_n=0^n p_a(n , l ) + & & z_n^a d_c / h^0(z_n^a , q^2 ) + _
m=1^n z_m^a d_g / h^0(z_m^a , q^2 ) [ mod_frag ] where @xmath53 , @xmath54 .
n is the maximum number of collisions for which @xmath55 , @xmath56 is the hadronic fragmentation function which gives the probability that quark or gluon would fragment into a hadron ( in our case @xmath57 ) and @xmath58 .
the first term gives the fragmentation probability of the leading parton @xmath25 with a reduced energy @xmath59 and the second term comes from the emitted gluon having energy @xmath60 .
+ while one may easily ignore the azimuthal ( @xmath61 ) dependence of the average path length @xmath62 for impact parameter @xmath9 for very central collisions , it needs to be considered for non - central collisions .
we use a simple approach , based on glauber model , to evaluate the dependence of the average path - length on the azimuthal angle with respect to the reaction plane . assuming uniform densities for the colliding nuclei , the average path - length for an impact parameter @xmath9 and azimuthal angle @xmath61 can be written as : @xmath63 where @xmath64 and @xmath65 are the transverse co - ordinates for the point where the partons scatter to produce the jet(s ) which
, travelling at an angle @xmath61 with respect to the reaction plane , traverses the path length @xmath66 .
@xmath67 is the nuclear overlap function and @xmath68 and @xmath69 are the transverse density profiles of the two nuclei .
an average of @xmath70 over @xmath61 ( varying from zero to @xmath71 ) gives the average path length @xmath62 ( see fig .
[ fig2 ] ) .
we closely follow the excellent review on jet - quenching by baier et al @xcite , and assume that the collisional energy loss is quite small for light quarks and gluons compared to the energy energy loss due to radiation of gluons .
now one can define the formation time of the radiated gluon as : @xmath72 where @xmath14 is the energy of the radiated gluon and @xmath15 is the transverse momentum .
we would normally have @xmath73 where @xmath74 is a typical momentum transfer in a partonic collision .
the coherence length , @xmath75 , can now be written as : @xmath76 where @xmath77 is the number of coherent scattering centers .
one can then write , @xmath78 where @xmath79 is the energy parameter introduced to separate the incoherent and coherent radiation of gluons .
in the bethe - heitler ( bh ) regime , the energy of the radiated gluon is less than @xmath80 , the coherence length @xmath81 , leading to an incoherent radiation from @xmath82 scattering centers .
thus energy loss per unit length in the soft @xmath14 limit becomes : @xmath83 where @xmath84 and @xmath3 is the energy of the parton .
we shall write @xmath85 for this case and determine @xmath86 from the measurement of @xmath5 . for @xmath87 , we have a regime of coherent radiation , where the coherence length @xmath75 is greater than @xmath49 but less than average path length @xmath11 .
thus , @xmath88 centres of scattering radiate coherently leading to @xmath89 we re - write this as @xmath90 and determine the values of @xmath91 from the measurement of @xmath5 . finally , when the coherence length is larger than @xmath11 ( @xmath92 ) the entire medium acts as one coherent source and energy loss per unit length becomes : @xmath93 this is denoted as constant energy loss ( @xmath94 ) regime .
we use the nlo code of aurenche et al @xcite adopted earlier by jeon et al .
@xcite to account for the energy loss of the partons , by modifying the fragmentation function , for the calculations reported here . a more detailed derivation of the energy loss for this case leads to : @xmath95 where @xmath96 is fourier transform of the normalized differential cross - section for parton - parton collisions for the appropriate momentum transfer scale ( see ref .
the momentum transport coefficient @xmath97 is then defined as : @xmath98 so that we can write : @xmath99 which can be used to deduce the average momentum transport coefficient for any given centrality , for a comparison with other estimates for this .
we have calculated the nuclear modification factor @xmath7 for neutral pions for au - au collisions at the top rhic energy for six centralities using bethe - heitler ( bh ) , lpm , and constant energy loss mechanisms , discussed above .
we have kept the mean free path of gluons as well as quarks fixed at 1 fm and also assumed that their energy loss per collision is identical .
thus once we have accounted for the variation of the average path length @xmath62 with centrality , we are left with the energy loss per collision as the only adjustable parameter .
the results for bh mechanism are shown in fig .
[ fig3 ] for near - central ( 0 - 10% , 10 - 20% , 20 - 30% ) and mid - central ( 30 - 40% , 40 - 50% , 50 - 60% ) collisions .
we see a good description of the data for @xmath16 up to about 6 gev/@xmath18 .
the necessary energy loss per collision is seen to slowly drop from 10% to about 8% of the energy of the partons as we move from near central to mid - central collisions .
= 200 gev , using bh mechanism ( see text ) .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev , using bh mechanism ( see text ) .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev , using bh mechanism ( see text ) .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev , using bh mechanism ( see text ) .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev , using bh mechanism ( see text ) .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev , using bh mechanism ( see text ) .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] we also note that the slope of @xmath7 changes around @xmath16 equal to 5 gev/@xmath18 , which suggests a possible change in the mechanism for energy loss for partons contributing to higher momenta @xcite .
this expectation is confirmed by results shown in fig .
[ fig4 ] , where the so - called lpm mechanism of energy loss is seen to provide an accurate description of the nuclear modification function in the @xmath16 range of 610 gev/@xmath18 .
once again we note that the energy loss coefficient @xmath91 decreases systematically as the collisions become less central .
thus a parton having an energy of 10 gev is likely to lose about 1 gev in the first collision in the most central collision considered here and about 0.7 gev in the collisions having a centrality of about 5060% .
finally , we see ( fig .
[ fig5 ] ) that the medium modification function for hadrons having @xmath100 8 gev/@xmath18 is best described by the mechanism assuming a constant energy loss per collision .
the energy loss per collision is seen to decrease from about 1.4 gev for the most central collisions to about 1 gev for the collisions having a centrality of 5060% .
the transition from the bh regime to the lpm seen above and noted earlier @xcite at @xmath101 5 gev/@xmath18 should not come as a surprise to us , since for the mean free - path @xmath102 1 fm and @xmath103 1 ( gev/@xmath18)@xmath104 per collision , @xmath105 is about 5 gev , which roughly separates the bh and lpm energy loss regimes for all centralities = 200 gev using lpm mechanism .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev using lpm mechanism .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev using lpm mechanism .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev using lpm mechanism .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev using lpm mechanism .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev using lpm mechanism .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] the results for these three figures are summarized in fig .
[ fig5a ] .
several interesting facts emerge .
let us first look at the results for the bh regime operating for @xmath16 up to about 56 gev/@xmath18 .
we note as we go to more peripheral collisions the data , even for lower @xmath16 , are in a better agreement with our calculations .
this , we feel , is due to the reducing importance of radial flow ( which gives a transverse kick to the partons / hadrons ) in peripheral collisions .
the lpm regime operating over the @xmath16 window of 510 gev/@xmath18 and the complete coherence regime operating at higher @xmath16 are seen to correctly describe the details of the changing curvature of the nuclear modification factor .
we add that a look at fig .
[ fig4 ] may prompt one to conclude that the description using the lpm mechanism is also working reasonably well till the largest @xmath16 considered here .
however a closer look at the figs .
[ fig4 ] and [ fig5 ] reveal that the lpm description mostly misses the data for larger @xmath16 values while the description using the constant energy loss per collision correctly reflects the curvature of the data till the largest @xmath16 .
a more accurate data extending up to even larger @xmath16 would be very valuable to settle this question more firmly .
= 200 gev assuming a constant energy loss / collision .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev assuming a constant energy loss / collision .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev assuming a constant energy loss / collision .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev assuming a constant energy loss / collision .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev assuming a constant energy loss / collision .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] = 200 gev assuming a constant energy loss / collision .
the experimental data are taken from ref .
@xcite.,title="fig:",width=283 ] even though we have identified the regions of the applicability of the mechanisms of energy loss , purely on the basis of a good description of the experimental values of the nuclear modification factor , it is heartening to note that the separation of the bh and the lpm regimes is very close to 5 gev/@xmath18 expected by us , as discussed above .
we recall that the lpm and the complete coherence regimes differ by the formation time of the gluons , which is larger than the mean free path but smaller than the path - length for the former , while it is larger than the path - length ( and of - course the mean free path ) for the later .
this is reflected in a slight change in the curvature of the nuclear modification factor , around @xmath101 810 gev/@xmath18 .
= 200 gev for the three energy loss schemes for different @xmath16 regimes.,title="fig:",width=283 ] = 200 gev for the three energy loss schemes for different @xmath16 regimes.,title="fig:",width=283 ] = 200 gev for the three energy loss schemes for different @xmath16 regimes.,title="fig:",width=283 ] = 200 gev for the three energy loss schemes for different @xmath16 regimes.,title="fig:",width=283 ] = 200 gev for the three energy loss schemes for different @xmath16 regimes.,title="fig:",width=283 ] = 200 gev for the three energy loss schemes for different @xmath16 regimes.,title="fig:",width=283 ] in brief , we have noted that the energy loss parameter in the wang , huang , and sarcevic model for energy loss of partons in relativistic heavy ion collisions at rhic can be tuned to obtain an accurate description of the nuclear modification factor @xmath7 for different ranges of the transverse momenta .
this observation is further strengthened by fig .
[ fig6 ] where the variation of @xmath7 integrated over @xmath16 ranging from 6 to 9 gev/@xmath18 is shown as a function of number of participants and compared with the data from phenix collaboration @xcite . in the next step
we apply the above treatment to the collisions of copper nuclei at the same centre of mass energy of 200 gev / nucleon .
these collisions present an interesting system for the study of jet - quenching which complements as well as supplements the results from the collision of gold nuclei discussed earlier .
it is well known that the more central collisions of copper nuclei have number of participants similar to those for mid - central collisions of gold nuclei .
thus a comparison would give us results for two systems , involving similar number of participants , one of which has nearly central collisions and the other one has a large eccentricity .
proceeding as before , we give our results for the nuclear modification factor @xmath7 for cu+cu collisions for centralities varying from 010% to 3040% , in figs .
[ fig7][fig9 ] , using the bh , the lpm , and the constant energy loss mechanisms and compare with the measurements by phenix collaboration @xcite .
( see also ref .
@xcite ) . from fig .
[ fig7 ] , we see that the energy loss mechanism suitable for the bh regime provides an accurate explanation of the data for all the centralities under consideration for @xmath106 6 gev/@xmath18 . the energy loss mechanism denoted as lpm is seen to describe the nuclear modification function over the @xmath16 range of 610 gev/@xmath18 and even beyond for less central collisions fig .
[ fig8 ] . and
finally , the results for the mechanism admitting a constant energy loss per collision is seen to work well for pions having @xmath100 about 68 gev/@xmath18 ( see fig .
[ fig9 ] ) . of neutral pions calculated for lpm mechanism of energy loss .
the data are from the phenix collaboration @xcite for au+au collisions at @xmath107 = 200 gev.,width=434 ] we note that the energy loss co - efficient for the most central collision of copper nuclei is close to that for the collision of gold nuclei for the 5060% centrality , for all the three mechanisms .
we also note the possible separation of the bh and lpm regimes around @xmath16 equal to 56 gev/@xmath18 , as before for gold collisions , though the demarcation of the lpm and the constant energy loss mechanisms is not as sharp for the copper nuclei as for the gold nuclei , especially for the less central collisions .
this may have its origin in smaller @xmath11 for copper nuclei .
a figure similar to fig .
[ fig5a ] can also be prepared for the collisions involving copper nuclei .
we have not given it here , for the want of space .
= 200 gev for bh mechanism of energy loss along with the results from the phenix collaboration @xcite . ,
title="fig:",width=283 ] = 200 gev for bh mechanism of energy loss along with the results from the phenix collaboration @xcite .
, title="fig:",width=283 ] = 200 gev for bh mechanism of energy loss along with the results from the phenix collaboration @xcite .
, title="fig:",width=283 ] = 200 gev for bh mechanism of energy loss along with the results from the phenix collaboration @xcite .
, title="fig:",width=283 ] the results for the largest values of transverse momenta are summarized in fig .
[ fig10 ] . taking the case of constant energy loss per collision ,
@xmath109 , we see that the rate of energy - loss necessary to explain the suppressed production of hadrons at large @xmath16 varies as @xmath110 , the average path length of the partons in the medium , both for cu+cu and au+au collisions at the centre of mass energy of 200 gev / nucleon .
this empirical result confirms the conviction that for the range of transverse momenta covered , @xmath111 810 gev/@xmath18 , the concerned partons interact with the medium as a whole , and the energy lost by the partons , @xmath112 .
this is different from the ads / cft description where the radiated parton cloud is brought on - shell by the drag of the strongly coupled plasma and @xmath113 @xcite , though some moderation of the @xmath11 dependence is expected for an expanding plasma .
we also see that even though the slopes of the results for cu+cu and au+au collisions are similar , the energy loss for a given @xmath110 for au+au collisions is about 4060% larger than that for cu+cu collisions .
we can use the eq .
[ qhat ] to obtain the value of the momentum transport coefficient , @xmath97 from the above .
we find that the @xmath97 varies from 0.25 gev@xmath104/fm for the collisions having 010% centrality for au+au system at 200 agev to 0.39 gev@xmath104/fm for the 5060% centrality .
the decrease for more central collisions can be understood by noting that the systems produced in such cases would be at higher temperatures and thus much more dense than for peripheral collisions .
this would then lead to a smaller suppression of the radiation due to the lpm effect .
the faster rise with the path - length ( @xmath114 ) more than compensates this decrease , finally giving a large energy loss and larger suppression of high momentum particles for more central collisions .
= 200 gev for lpm mechanism of energy loss and compared with phenix results @xcite.,title="fig:",width=283 ] = 200 gev for lpm mechanism of energy loss and compared with phenix results @xcite.,title="fig:",width=283 ] = 200 gev for lpm mechanism of energy loss and compared with phenix results @xcite.,title="fig:",width=283 ] = 200 gev for lpm mechanism of energy loss and compared with phenix results @xcite.,title="fig:",width=283 ] on the other hand , for the case of cu+cu system , we get @xmath115 0.18 gev@xmath104/fm for all the centralities considered by us .
we note that it is about half of what we got for au+au system .
the near identity of @xmath97 for all the centralities for cu+cu system is related to a smaller variation in the path - length for them .
we also note that the au+au system is more efficient in degrading the jet - energy .
we add that various authors have reported widely differing values of @xmath97 .
thus , wiedemann and salgado @xcite have reported a value of 510 gev@xmath104/fm , while gyulassy , levai , and vitev @xcite report a value in the range of 0.350.85 gev@xmath104/fm .
arnold , moore , and yaffe @xcite , on the other hand , suggest a value of about 2 gev@xmath104/fm , in contrast to a value of 34 gev@xmath104/fm reported by majumder @xcite .
it is felt that a part of this difference may be due to different physical attributes of the evolving system over which the average is taken .
= 200 gev for constant energy loss mechanism compared with phenix results @xcite.,title="fig:",width=283 ] = 200 gev for constant energy loss mechanism compared with phenix results @xcite.,title="fig:",width=283 ] = 200 gev for constant energy loss mechanism compared with phenix results @xcite.,title="fig:",width=283 ] = 200 gev for constant energy loss mechanism compared with phenix results @xcite.,title="fig:",width=283 ] with @xmath110 for cu+cu and au+au collisions at 200 agev for the constant energy loss per collision regime obtained in the present work.,width=434 ]
we have seen that the phenomenological model of parton energy loss , which accounts for only the induced gluon radiation of the fast partons , successfully describes the suppression of neutral pion spectra at large @xmath16 at different centralities for au+au and cu+cu collisions at @xmath116 gev . in case of non - central collisions
, the path - length will have an azimuthal variation over the transverse plane ( see fig .
[ fig2 ] ) . in
what follows , we study the azimuthal dependence of the suppression of production of hadrons in terms of the azimuthal anisotropy of their transverse momenta @xcite .
we re - emphasize that this anisotropy is different from the elliptic flow seen for hadrons at lower @xmath16 , which is explained using hydrodynamics which converts the spatial anisotropy of the initial state into the azimuthal anisotropy of the transverse momenta of the hadrons @xcite .
the extent of validity of hydrodynamics for larger @xmath16 is not very clear .
the success of the recombination model @xcite suggests that the fragmentation as a mode of hadronization may be valid for @xmath117 35 gev/@xmath18 .
thus the following results should be taken as indicative of anisotropy which can arise due to medium modification of fragmentation function due to energy loss of partons .
the differential azimuthal anisotropy is measured in terms of the parameter @xmath20 , which is the second fourier coefficient of the azimuthal distribution of hadrons in the reaction plane : @xmath118 we have calculated the azimuthal anisotropy coefficient @xmath20 of neutral pions for @xmath119 2 gev / c for au+au collisions at @xmath107=200 gev for the six centralities mentioned earlier . the energy loss per collision is taken from the earlier analysis and @xmath61 dependent distribution of the pions is calculated by incorporating the @xmath61 dependence of the path - length @xmath11 .
the results of our calculations are displayed in fig .
[ fig11 ] along with the data available from phenix collaboration @xcite .
we show the theoretical results only over the @xmath16 window where the corresponding mechanism was found to describe the nuclear modification data , earlier ( see fig .
[ fig5a ] ) . recalling that the bh mechanism was found to provide a good description of the nuclear suppression , up to @xmath16 equal to 56 gev/@xmath18 ,
while the mechanisms for the lpm and the constant energy loss per collision provided description for higher @xmath16 , the results are quite interesting .
first of all , even for @xmath120 6 gev/@xmath18 while the bh mechanism leads to a @xmath20 increasing with @xmath16 , the experimental results show a contrary behaviour
. it could be that a hydrodynamic expansion of the system under viscosity affects the hadrons having @xmath16 up to 56 gev/@xmath18 @xcite .
the other two mechanisms , though , give a behaviour similar to the experimental data , in the region of transverse momenta where they are applicable .
the theoretical values are larger by up to a factor of two in the worst cases of more central collisions , though they agree reasonably well with the experimental values for the least central collision considered here . of neutral pion calculated using bh , lpm , and constant energy loss mechanisms for au+au collisions at @xmath107=200 gev .
the experimental data are from the phenix collaboration @xcite.,title="fig:",width=283 ] of neutral pion calculated using bh , lpm , and constant energy loss mechanisms for au+au collisions at @xmath107=200 gev .
the experimental data are from the phenix collaboration @xcite.,title="fig:",width=283 ] of neutral pion calculated using bh , lpm , and constant energy loss mechanisms for au+au collisions at @xmath107=200 gev .
the experimental data are from the phenix collaboration @xcite.,title="fig:",width=283 ] of neutral pion calculated using bh , lpm , and constant energy loss mechanisms for au+au collisions at @xmath107=200 gev .
the experimental data are from the phenix collaboration @xcite.,title="fig:",width=283 ] of neutral pion calculated using bh , lpm , and constant energy loss mechanisms for au+au collisions at @xmath107=200 gev .
the experimental data are from the phenix collaboration @xcite.,title="fig:",width=283 ] of neutral pion calculated using bh , lpm , and constant energy loss mechanisms for au+au collisions at @xmath107=200 gev .
the experimental data are from the phenix collaboration @xcite.,title="fig:",width=283 ] of neutral pions calculated for lpm mechanism of energy loss for au+au collisions @xmath107 = 200 gev .
the data points are taken from phenix collaboration @xcite.,width=434 ] it is interesting to see the centrality dependence of the integrated azimuthal anisotropy coefficient in the @xmath16 range of 69 gev / c . for this purpose
, we have used the differential azimuthal anisotropy coefficient @xmath20 for the lpm mechanism obtained earlier ( see fig.[fig12 ] ) .
we see that in general our calculations reproduce the trend of the variation of @xmath121 with the number of participants , though the theoretical values are larger by about a factor of two ( see ref .
@xcite , for similar results ) .
one short - coming of our calculation immediately comes to mind- the description of the nuclei as having a uniform density . a woods - saxon density profile for the colliding nuclei would reduce the difference in the path - lengths for
the partons travelling in the reaction plane and travelling in a plane perpendicular to it , for example , and thus reduce @xmath121 .
we postpone this study to a future work where additionally , the evolution of the medium will be taken into account .
we also give our prediction for @xmath20 for cu+cu collisions at 200a gev for two typical centralities ( fig .
[ fig13 ] ) .
a behaviour similar to that for au+au collisions is seen . of neutral pion
calculated using bh , lpm , and constant energy loss mechanisms for cu+cu collisions at @xmath107=200 gev.,title="fig:",width=283 ] of neutral pion calculated using bh , lpm , and constant energy loss mechanisms for cu+cu collisions at @xmath107=200 gev.,title="fig:",width=283 ]
prompt photons , as indicated earlier , originate from ( a ) quark - gluon compton scattering ( @xmath122 ) , ( b ) quark - anti - quark annihilation ( @xmath123 ) , and ( c ) collinear fragmentation of a high energy quarks ( @xmath124 ) following a hard collision . of these the third process
is affected by the energy loss suffered by the quark before its fragmentation @xcite .
it will also lead to an azimuthal anisotropy of the momentum distribution of hard photons @xcite .
once again , we use the nlo pqcd code of aurenche et al @xcite , suitably modified to account for the energy loss suffered by the quarks before they fragment into photons . as a first step
we show our results for proton - proton collisions at 200 gev ( fig .
[ fig14 ] ) using the fragmentation , factorization , and the renormalization scales as equal , with @xmath125 @xmath15/2 , @xmath15 and 2@xmath15 .
we have used the _
cteq4 m _ parton distribution function and the bkk fragmentation function for these calculations as well .
we see that our results are consistent with the data from phenix @xcite and also with earlier calculations ( see ref .
@xcite ) along similar lines . for the rest of the calculations
we have used a common scale @xmath126 .
= 200 gev with nlo pqcd calculations.,width=434 ] while performing the calculations for the nucleus - nucleus collisions we include nuclear shadowing and energy loss suffered by the quarks as before .
we also properly account for the isospin of the participating nucleons , which is essential while performing calculations for photons .
the nuclear modification factor @xmath127 is then obtained in a manner similar to eq . [ raa ] .
we show our results for the nuclear modification of hard photon production using the energy loss parameters obtained earlier along with the experimental results for collision of gold nuclei at 200a gev in fig .
[ fig15 ] .
we give the theoretical curves over the regions where the nuclear modification for pion production is in agreement with the experimental data ( see fig .
[ fig5a ] ) .
= 200 gev .
the data points are taken from ref .
@xcite , width=434 ] we see that our results are in fair agreement with the data for @xmath15 beyond 10 gev/@xmath18 where the mechanism of prompt photon production included in the present work is expected to dominate @xcite .
we do realize that for lower @xmath15 other mechanisms like jet - conversion @xcite , induced bremsstrahlung @xcite and thermal production @xcite will contribute .
next we calculate the the differential azimuthal anisotropy coefficient @xmath128 for hard photons for au+au collisions at 200a gev .
recall that for lower @xmath15 ( @xmath129 56 gev/@xmath18 ) , the azimuthal anisotropy of the direct photons carries valuable information about the development of the elliptic flow of the plasma and momentum anisotropy of the thermalized partons @xcite . of direct photons calculated using lpm scheme of parton energy loss for au+au collisions at @xmath107=200 gev along with experimental results from the phenix @xcite ( left panel ) and the star @xcite ( right panel ) collaborations.,title="fig:",width=283 ] of direct photons calculated using lpm scheme of parton energy loss for au+au collisions at @xmath107=200 gev along with experimental results from the phenix @xcite ( left panel ) and the star @xcite ( right panel ) collaborations.,title="fig:",width=283 ] in fig . [ fig16 ]
we have shown the results for @xmath130 calculated using lpm and the constant energy loss per collision mechanisms of parton energy loss for two centrality classes ( 20 - 40% and 10 - 40% ) and compared them with the recent results from the phenix @xcite and the star @xcite collaborations .
we make no further adjustments of the energy loss parameters obtained earlier .
we see that the azimuthal anisotropy of direct photons for @xmath131 6 gev/@xmath18 is reproduced reasonably well from our calculations .
the results of figs .
[ fig15 ] and [ fig16 ] come a pleasant surprise as we recall that while calculating the energy loss of the partons and the consequent suppression of hadrons at large @xmath16 , we did not distinguish between quarks and gluons .
the photon production , however , is sensitive only to the energy loss suffered by quarks .
one possible reason for this could be the dominating contribution of quarks to hadrons as well photons having large transverse momenta , which is then rightly sampled by our procedure .
one could use this procedure to first fix the energy loss per collision for quarks and then proceed to determine the energy loss per collision for gluons by fitting the hadronic spectra .
this would , however , require a much more accurate and extensive large transverse momentum data for photons . a further availability of isolated photon spectra along with the ( total ) single photon spectra will help us to subtract out the compton plus annihilation contributions and increase the sensitivity of these calculations to energy loss suffered by quarks .
we have used a simple and transparent model of multiple scattering and energy loss for partons passing through quark gluon plasma produced in relativistic collisions of heavy nuclei , before they fragment into hadrons or photons . taking the centrality dependence of the average path - length into account and taking energy loss per collision as an adjustable parameter
, we have obtained a description of hadronic suppression in au+au and cu+cu collisions at 200a gev .
the energy loss parameters are found to vary systematically with the centrality and the @xmath108 for @xmath100 68 gev/@xmath18 is found to vary linearly with the average path - length @xmath110 , in agreement with the predictions of baier et al .
@xcite .
the same parameters are then used without any further adjustments to obtain the azimuthal anisotropy for high @xmath16 pions .
we find it to be about a factor of two larger than the experimental values .
next , we have calculated the suppressed production of hard photons and their azimuthal anisotropy , again with out changing any of the parameters .
these are in reasonable agreement with the experimental findings .
we have mentioned that in principal , the calculations could be improved by incorporating diffused density distributions for the colliding nuclei , instead of the uniform densities used in the present work . in that case
, it remains to be seen if the @xmath121 could be used to further constrain the energy loss parameters .
the success of the approach , considering that the dynamics of evolution , the temperature and flavour dependence of some of the parameters and a more detailed accounting of various interferences are not considered here , once again confirms the inadequacy of @xmath5 in seriously distinguishing various details of the dynamics of the plasma @xcite .
some improvements are definitely in order .
the use of uniform density for the colliding nuclei can be easily improved upon .
this , as we argued , could improve the description of azimuthal anisotropy of hadron production at higher @xmath16 .
one may also take differing values for the mean free paths for quarks and gluons and the energy loss per collision for them .
note , however , that if we argue that @xmath108 for gluons is twice that for quarks and also that their mean free path is half of that for quarks , then the energy loss for quarks and gluons per collision will be identical .
why does this simple approach work so well ?
we realize that the @xmath16 distribution of jets produced in primary hard scatterings is a rapidly falling function of the transverse momentum .
partons having a transverse momentum @xmath16 after undergoing multiple collisions and losing a momentum @xmath132 will populate the momentum space occupied by partons having transverse momentum @xmath133 while the later will shift to still lower momenta .
thus if we make a good estimate of @xmath132 , we are likely to get a good description of the nuclear modification . but
is this approach really so simple ?
nlo pqcd and accurate structure and fragmentation functions , along with the applicability of pqcd at larger transverse momenta ensure an accurate base - line for p+p collisions .
the influence of the dynamics of evolution is perhaps softened by the short life - time of the plasma at rhic energies and the fact that most of the energy is lost in collisions at early times .
thus we feel that the present study , at the very minimum , gives a reliable average of the rate of energy loss of partons in quark gluon plasma .
we also see that a simultaneous description of nuclear modification of photon and pion production in nuclear collisions , holds out the hope of getting the energy loss suffered by quarks and gluons independently .
it will definitely be of interest to see if the centrality dependence of @xmath108 ( @xmath134 ) for partons having larger transverse momenta at lhc energies ( 2.76 atev and 5.5 atev ) continues to follow the trend seen for the data at rhic energies for cu+cu and au+au collisions , seen in fig .
[ fig10 ] .
if confirmed at lhc energies as well , this will be a very valuable empirical affirmation of the suggestions of baier et al @xcite .
this aspect is under investigation .
sd would like to express his sincere gratitude to department of atomic energy , government of india for financial support during the course of this work .
we also thank christine aidala for drawing our attention to the final data for pp collisions ( fig .
[ fig1 ] ) .
20 r. j. fries , b. mller , c. nonaka and s. a. bass , phys .
lett . * 90 * , 202303 ( 2003 ) ; r. j. fries , b. mller , c. nonaka and s. a. bass , phys .
c * 68 * , 044902 ( 2003 ) ; v. greco , c. m. ko and p. levai , phys .
lett . * 90 * , 202302 ( 2003 ) . t. matsui and h. satz , phys .
b * 178 * , 416 ( 1986 ) ; r. l. thews , m. schroedter and j. rafelski , phys .
c * 63 * , 054905 ( 2001 ) ; x. m. xu , d. kharzeev , h. satz and x. n. wang , phys .
c * 53 * , 3051 ( 1996 ) ; b. k. patra and d. k. srivastava , phys .
b * 505 * , 113 ( 2001 ) ; l. yan , p. zhuang and n. xu , phys .
lett . * 97 * , 232301 ( 2006 ) .
m. gyulassy , p. levai and i. vitev , nucl .
b * 571 * , 197 ( 2000 ) ; m. gyulassy , p. levai and i. vitev , phys .
lett . * 85 * , 5535 ( 2000 ) ; m. gyulassy , p. levai and i. vitev , nucl .
b * 594 * , 371 ( 2001 ) ; c. a. salgado and u. a. wiedemann , phys .
lett . *
89 * , 092303 ( 2002 ) ; c. a. salgado and u. a. wiedemann , nucl .
a * 715 * , 783 ( 2003 ) ; c. a. salgado and u. a. wiedemann , phys .
d * 68 * , 014008 ( 2003 ) .
r. baier , d. schiff and b. g. zakharov , ann .
nucl . part .
sci . * 50 * , 37 ( 2000 ) ; r. baier , y. l. dokshitzer , a. h. mueller , s. peigne and d. schiff , nucl .
b * 484 * , 265 ( 1997 ) ; r. baier , y. l. dokshitzer , a. h. mueller , s. peigne and d. schiff , nucl . phys .
b * 483 * , 291 ( 1997 ) ; r. baier , y. l. dokshitzer , a. h. mueller and d. schiff , phys .
c * 58 * , 1706 ( 1998 ) .
+ see also , d. denterria , arxiv:0902.2011 [ nucl - ex ] ; s. peigne and a. v. smilga , phys .
usp . * 52 * , 659 ( 2009 ) [ usp . fiz .
nauk * 179 * , 697 ( 2009 ) ] [ arxiv:0810.5702 [ hep - ph ] ] .
|
we investigate the system size dependence of jet - quenching by analyzing transverse momentum spectra of neutral pions in au+au and cu+cu collisions at @xmath0 = 200 gev for different centralities .
the fast partons are assumed to lose energy by radiating gluons as they traverse the plasma and undergo multiple collisions . the energy loss per collision , @xmath1 ,
is taken as proportional @xmath2 ( where @xmath3 is the energy of the parton ) , proportional to @xmath4 , or a constant depending on whether the formation time of the gluon is less than the mean path , greater than the mean free path but less than the path length , or greater than the path length of the partons , respectively .
nlo pqcd is used to evaluate pion production by modifying the fragmentation function to account for the energy loss .
we reproduce the nuclear modification factor @xmath5 by treating @xmath1 as the only free parameter , depending on the centrality and the mechanism of energy loss .
these values are seen to explain the nuclear modification of prompt photons , caused by the energy lost by final state quarks before they fragment into photons .
these also reproduce the azimuthal asymmetry of transverse momentum distribution for pions within a factor of two and for prompt photons in a fair agreement with experimental data .
_ keywords _ : relativistic heavy ion collisions , quark gluon plasma , nlo pqcd , jet quenching , pion production , direct photon production , nuclear modification , azimuthal anisotropy .
+ _ pacs no .
_ : 12.38.mh , 12.38.-t , 13.85.qk , 25.75.cj , 25.75.-q
|
with the ever - improving quality of experimental data obtained at the lhc , developing theoretical tools to describe more and more complex physical processes has never been more relevant than it is today .
going to higher - orders in perturbation theory while being able to perform fast numerical computations is requested to achieve accurate theoretical predictions . the conventional way to compute next - to - leading order ( nlo ) or next - to - next - to - leading order ( nnlo ) corrections is dimensional regularization ( dreg ) @xcite . within this formalism
, real and virtual integrals are computed separately in @xmath1 instead of @xmath2 dimensions , and singularities which manifest themselves under the form of @xmath3 poles are subtracted using one of the many existing techniques @xcite , as guaranteed by the kinoshita - lee - nauenberg ( kln ) theorem @xcite . however , this approach may not be the most efficient when dealing with multi - loop multi - leg processes as one has to deal with complex final - state phase - spaces for each of the different contributions . moreover
, building all the necessary counter - terms can be far from being trivial and is in general very time - consuming .
+ we will present in this article an alternative method called four - dimensional unsubtraction ( fdu ) @xcite , which is based on the loop - tree duality ( ltd ) theorem @xcite .
this method allows to combine , at the integrand level , virtual and real contributions using a suitable and physically motivated mapping between momenta , so the cancellation of singularities is achieved locally , meaning that the integration can be performed in @xmath2 dimensions .
the ltd theorem allows to rewrite loop scattering amplitudes as a sum of phase - space integrals , with slightly modified propagators . for a generic @xmath4-particle one - loop integral with massive scalar particles , @xmath5=-\sum\limits_{i=1}^{n}\int_{\ell}\tilde{\delta}(q_i)\prod\limits_{j=1,j\neq i}^{n}g_d(q_i;q_j)\,,\ ] ] where the internal momenta are written @xmath6 , the feynman propagators @xmath7 , and the so - called dual propagators @xmath8 . the delta function @xmath9 sets the internal lines on - shell by selecting the pole of the feynman propagators with positive energy and negative imaginary part .
finally , @xmath10 is an arbitrary future - like vector with positive definite energy , and @xmath11 defines the sign in front of the @xmath12 prescription .
this change of prescription , far from introducing additional difficulties , is essential for the consistency of the method .
more details can be found in @xcite .
in general , the procedure for computing nlo cross - sections is to consider and integrate separately real and virtual corrections .
however , with the fdu technique , it is possible to perform the summation over degenerate soft and collinear states from both contributions before integration , thanks to the use of a suitable mapping of momenta between real and virtual kinematics .
this is achievable because the infrared ( ir ) singularities are located in a compact region of the loop three - momentum , therefore the cancellations occur locally and the introduction of ir subtraction terms is not needed . for a process at one loop with @xmath13 final state particles , we can write @xmath14 where @xmath15 denotes the renormalized virtual correction obtained using ltd and reads @xmath16 where @xmath17 denotes the @xmath4-leg scattering amplitude at leading order ( lo ) , while @xmath18 denotes the one - loop scattering amplitude , including the self - energy corrections of the external legs . the integral is then weighted by the measure function @xmath19 , used to define a given physical observable .
similarly , the real correction reads @xmath20 where the external four - momenta @xmath21 , the phase - space and the tree - level scattering amplitude have been rewritten in terms of the loop three - momentum and the external momenta @xmath22 of the born process .
+ although in dreg both ultraviolet ( uv ) and ir singularities are treated equally , in fdu they are considered separately .
indeed , it is necessary to discriminate them in order to build a complete ltd representation of the virtual contributions , as explained in @xcite .
+ the uv singularities are dealt with by introducing counter - terms for the mass and the wave - function at the integrand level .
those counter - terms are calculated by expanding the unrenormalized amplitude around the uv propagator @xmath23 , then choosing a renormalization scheme and finally adjusting subleading terms in order to subtract the right finite pieces .
for instance , in the @xmath24 scheme , this would be equivalent to make their integrated form in @xmath1 exhibit only an @xmath3 pole , i.e. no finite part .
the counter - terms are process - independent and only depend on the nature of the particles involved , so it is only necessary to compute them once .
more details about the uv renormalization within fdu can be found in @xcite . + ir singularities however , are more tricky to deal with .
first , it is necessary to isolate each of them into different regions of the phase - space so that there is no more than one ir singularity of a given type soft or ( quasi-)collinear in a given region .
then , for each region , one must apply to the real integrand a mapping that will make similar singular behaviors , for the virtual and the real contributions , match and therefore cancel each other locally .
finally , both contributions can be added , as shown in equation ( [ sumrealvirtual ] ) , and the integration can be safely performed in 4 dimensions . + in the next section , we will apply the fdu technique to the three - point scalar function , and illustrate with a few expressions .
let s consider a @xmath0 process at nlo with one massless internal state and the remaining internal and outgoing particles with mass equal to @xmath25 . to simplify the expressions , we define @xmath26 and @xmath27 .
the final - state on - shell momenta are labeled as @xmath28 and @xmath29 , with @xmath30 and with @xmath31 and @xmath32 .
the incoming one is @xmath33 with virtuality @xmath34 .
we write the three internal momenta @xmath35 , and when they are on - shell , @xmath36 , with @xmath37 , @xmath38 and @xmath39 .
+ since the virtual decay - rate is given by @xmath40\,,\ ] ] we apply the ltd to @xmath41 .
its dual representation then consists in the sum of the 3 following integrals : @xmath42 \ , d[v_1]}{1-(1 - 2 v_1)^2 \beta^2}~ , \nonumber \\
i_2 & = & \frac{2}{s_{12 } } \int \frac{\xi_{2}^2 \ , d[\xi_{2 } ] \ , d[v_2]}{\xi_{2,0 } \ ,
\left(1-\xi_{2,0}+\imath 0 \right ) \left(\xi_{2,0}+\beta \ , \xi_{2 } \ , ( 1 - 2 v_2)-m^2\right)}~ , \\ i_3 & = & - \frac{2}{s_{12 } } \int \frac{\xi_{3}^2 \ , d[\xi_{3 } ] \ , d[v_3]}{\xi_{3,0 } \ , \left(1+\xi_{3,0 } \right ) \left(\xi_{3,0}-\beta \ , \xi_{3 } ( 1 - 2 v_3)+m^2\right)}\,,\nonumber \end{aligned}\ ] ] with @xmath43 = \frac{(4\pi)^{\epsilon-2}}{\gamma(1-\epsilon ) } \left(\frac{s_{12}}{\mu^2}\right)^{-\epsilon } \ , \xi_i^{-2\epsilon } \ , d\xi_i~$ ] and @xmath44 = ( v_i(1-v_i))^{-\epsilon } \ , dv_i$ ] . on the other hand ,
the real contributions are given by the interference terms , and are written as @xmath45 with @xmath46 being the normalized scalar product between @xmath47 and the radiated particle four - momentum @xmath48 . in equation ( [ realdecayrate ] ) , we have applied the phase - space partition mentioned at the end of the previous section .
explicitly , we define @xmath49 for @xmath50 . then , for each region , we introduce a suitable mapping between real and virtual kinematics .
for instance , in region 1 we use @xmath51 with @xmath52 and @xmath53 fulfilling @xmath54 .
the parameters @xmath55 and @xmath56 are computed by solving the on - shell conditions @xmath57 , their expressions can be found in @xcite .
a similar transformation is used for region 2 .
it is worth emphasizing that these transformations fulfill two roles : they move the ir singularities to the same integration points in the dual domain , which allows them to cancel each other locally , and they are optimized to deal smoothly with the massless limit in each region . finally , after rewriting the real contributions in terms of the loop variables , we obtain @xmath58 with @xmath59 and @xmath60 being the jacobians of the two transformations , @xmath61 and @xmath62 restricting the integration domain associated to their respective region , and @xmath63
. process with scalar particles , as a function of the dimensionless parameter @xmath13 .
the solid line corresponds to the dreg analytical result , while the dots are the numerical values obtained using fdu .
the horizontal line represents the massless limit.,scaledwidth=50.0% ] the sum of real and virtual contributions , i.e. the sum of equations ( [ virtualdecayrate ] ) , ( [ gammatilde1toy ] ) and ( [ gammatilde2toy ] ) , gives the total decay - rate at nlo , @xmath64 , and is safely integrable in @xmath2 dimensions .
the integration is done numerically , and is shown in figure [ plot ] .
our results are compared with the analytical expression obtained using the dreg technique , and the agreement between the two approaches is very good .
another advantage of the ltd approach is the smooth massless limit , thanks to the mapping being constructed to deal with quasi - collinear configurations .
it is therefore possible to put @xmath13 to 0 before integration and still get the correct result .
throughout this article , we generalized the fdu method to the massive case and applied it on the three - point scalar function . based on the ltd theorem , this technique allows to bypass dreg by expressing real and virtual contributions in terms of phase - space integrals , and adding them after splitting the integration domain and applying a mapping between the two kinematics in each of the different regions .
this allows a safe four - dimensional integration , without the need to deal with @xmath3 poles , which makes the numerical computations easier and faster .
the fdu technique has been applied to physical processes such as @xmath65 @xcite , and can be extended to compute multi - leg multi - loop processes as explained in @xcite .
this research project has been done in collaboration with germn rodrigo and germn sborlini . this work is partially supported by the spanish government and eu erdf funds ( grants fpa2014 - 53631-c2 - 1-p and sev-2014 - 0398 ) and by gv ( prometeu ii/2013/007 ) . 1 c. g. bollini and j. j. giambiagi , nuovo cim .
b * 12 * ( 1972 ) 20 . g. t hooft and m. j. g. veltman , nucl .
b * 44 * ( 1972 ) 189 .
z. kunszt and d. e. soper , phys . rev .
d * 46 * ( 1992 ) 192 .
s. frixione , z. kunszt and a. signer , nucl .
b * 467 * ( 1996 ) 399 .
s. catani and m. h. seymour , phys .
b * 378 * ( 1996 ) 287 .
s. catani and m. h. seymour , nucl .
b * 485 * ( 1997 ) 291 [ nucl .
b * 510 * ( 1998 ) 503 ] .
t. kinoshita , j. math .
* 3 * ( 1962 ) 650 .
t. d. lee and m. nauenberg , phys .
* 133 * ( 1964 ) b1549 .
r. j. hern ' andez - pinto , g. f. r. sborlini and g. rodrigo , jhep * 1602 * ( 2016 ) 044 .
s. catani , t. gleisberg , f. krauss , g. rodrigo and j. c. winter , jhep * 0809 * ( 2008 ) 065 .
i. bierenbaum , s. catani , p. draggiotis and g. rodrigo , jhep * 1010 * ( 2010 ) 073 .
i. bierenbaum , s. buchta , p. draggiotis , i. malamos and g. rodrigo , jhep * 1303 * ( 2013 ) 025 .
|
we present a new method to compute higher - order corrections to physical cross - sections , at next - to - leading order and beyond .
this method , based on the loop tree duality , leads to locally integrable expressions in four dimensions . by introducing a physically motivated momentum mapping between the momenta involved in the real and the virtual contributions , infrared singularities naturally cancel at integrand level , without the need to introduce subtraction counter - terms .
ultraviolet singularities are dealt with by using dual representations of suitable counter - terms , with some subtleties regarding the self - energy contributions . as an example
, we apply this method to compute the @xmath0 decay rate in the context of a scalar toy model with massive particles .
|
modeling the ir radiation of pre - main - sequence stars has traditionally involved `` classic '' geometrically - thin optically - thick disks . this approach fails to produce many features of observed spectral energy distributions ( sed ) in both t tauri stars ( tts ) and herbig ae / be stars ( haebes ) .
one way out of this problem is to supplement the disk emission with a surrounding optically thin halo ( e.g. butner , natta & evans 1994 ; miroshnichenko et al 1999 , mive hereafter ) .
a simpler alternative was proposed by chiang & goldreich ( 1997 , cg hereafter ) : the surface skin of any optically thick object is of course optically thin .
the emission from the disk surface layer can become significant under certain flaring conditions , and cg present sed fits of a number of tts purely in terms of flared disks .
this proposal was extended to sed modeling of haebes by chiang at al ( 2001 ) and natta et al ( 2001 ) .
while the cg model successfully solves the sed problem with a simple physical explanation , recent high - resolution imaging observations of both tts and haebes reveal the presence of halos , ignored in the cg approach .
most striking is the case of gm aur , a classical tts whose sed modeling was presented as evidence of the success of the cg approach ( chiang & goldreich 1999 ) .
indeed , hst / nicmos images by schneider et al ( 2003 ) show a flared disk with radius of 300 au , but they also reveal the presence of a surrounding tenuous envelope extending to a radius of 1000 au . using detailed monte carlo calculations , schneider et al find that simplified models with flared disk without an imbedding halo fail to replicate the scattered light intensity pattern seen in the nicmos images .
the addition of a halo is essential for successful modeling of both the sed and the images .
similarly , hst and ground - based imaging by stapelfeldt et al ( 2003 ) of the tts hv tau c show a nearly edge - on flared disk imbedded in a more extended nebulosity . from detailed model calculations
stapelfeldt et al find that although it is possible to fit the sed purely with a flared disk , the flaring is unreasonably large and such models do not reproduce the image adequately .
they conclude that a flared disk alone is an inadequate model and that an additional , extended component to the circumstellar density distribution is needed to explain the observations .
finally , from detailed modeling of flared disk sed kikuchi , nakamoto & ogochi ( 2002 ) conclude that halos are necessary supplements in explaining flat - spectrum tts .
high resolution observations of haebes give similar evidence for imbedding halos . combining space- and ground - based observations of hd 100546 , the nearest herbig be star , grady et al ( 2001 )
resolve the disk , which extends to 5(500 au ) from the star and displays the elongation of inclined viewing with @xmath4 = 49 .
but they also find that the disk is imbedded in a more extended halo ( 10 radius ) that is roughly circular in shape and optically thin ( background stars are visible through it ) .
polomski et al ( 2002 ) performed high - resolution ground - based observations of a sample of haebes whose mid - ir emission was claimed to be disk - dominated by hillenbrand et al ( 1992 ) .
they find instead that the emission is not confined to an optically thick disk but originates in a more complex environment that includes large , extended dust envelopes . from analysis of iso data ,
creech - eakman et al ( 2002 ) conclude that only a subset of the haebes they observe can be described purely in terms of the cg model ( notably , e.i .
chiang is a coauthor of this study ) . how can the sed of the same t - tauri star , gm aur , be fitted successfully with a flared disk both with a halo ( schneider et al 2003 ) and without one ( chiang & goldreich 1999 ) ?
is it at all possible to distinguish between these two cases ?
what , if any , is the unique radiative signature of each component
? obviously , imbedding a disk in a tenuous halo with a very small optical depth is not going to affect its emission appreciably .
at which stage then does the halo assume a significant role ?
here we address these questions . in appendix
a we derive general results for the emission from optically thin dust in an arbitrary geometry .
we apply these results to the surface layers of flared disks in 2 and to halos in which the disks could be imbedded in 3 .
chiang and goldreich ( 1997 ) noted that the emission from the optically thin surface layer of an optically thick disk , which has been neglected traditionally , can become significant under certain flaring conditions .
the stellar radiation penetrates to an optical distance = 1 along a direction slanted to the surface by angle @xmath5 ( figure [ fig : cg ] ) .
the optical depth of the corresponding skin layer along the normal to the surface @xmath6 is @xmath5 at visual and @xmath7 at wavelength @xmath8 , where @xmath9 and @xmath10 is the dust cross - section at visual . in the case of a flat thin disk
whose inner radius is determined by dust sublimation , the grazing angle is here is the radius of the star , is the dust sublimation radius ( eq.[eq : r1 ] ) and is distance from the axis .
flaring is defined by the radial profile of the disk height @xmath0 ( @xmath11 ) or , equivalently , as is evident from figure [ fig : cg ] , @xmath12 where @xmath13 , therefore the grazing angle of a flared disk is the cg surface layer serves as an effective optically thin disk atop the underlying optically thick disk core , and its flux is obtained from the volume integration listed in equation [ eq : flux ] . in the case of face - on orientation
, the optical depth of the cg layer obeys @xmath14 , therefore its flux at distance @xmath15 is since the temperature profile of optically thin dust depends only on distance from the radiation source , the geometry dependence of this expression enters only from the radial variation of the grazing angle @xmath5 ( reflecting the dust column variation ) .
but other geometries can produce an identical expression .
for example , in the case of spherical geometry the flux is controlled by the dimensionless density profile @xmath16 ( equation [ eq : eta ] ) and the radial optical depth is @xmath17 .
denoting the optical depth at visual wavelengths by , the flux is where @xmath18 . since @xmath19 and @xmath20
enter only as integration variables in the last two integrals , the two expressions are mathematically identical if a minor point is that @xmath21 in eq .
[ eq : fcg ] is strictly a function of @xmath22 rather than @xmath20 ( since temperature is controlled by distance from the star ) ; this slight difference can be ignored because @xmath23 everywhere in the disk .
scattering can be treated similarly and produces the same result since the only difference is that @xmath24 is replaced by @xmath25 , another function of @xmath19 ( see equation [ eq : scat ] ) .
therefore , _ there is a complete equivalence between spherical halos and the cg surface layers of flared disks _ ; either case defines a model with the other geometry and the exact same flux .
specifically , equations [ eq : flat ] and [ eq : equiv1 ] show that a thin flat disk is equivalent to a spherical halo with @xmath26 and = @xmath27 . in the case of a flared disk , equations [ eq : alpha ] and [ eq : equiv1 ]
show that the cg layer will produce precisely the same flux as a spherical halo with in particular , the equivalent halo of a disk with flaring angle @xmath28 has density distribution @xmath29 . and
conversely given a spherical halo , the flared disk with the same outer radius and will produce the exact same flux from its cg - surface layer .
each disk defines a mathematically equivalent halo .
although this equivalence was derived only for face - on viewing of the disk , it carries to most inclination angles since the flux from optically thin dust involves a volume integration ( equation [ eq : flux ] ) and the observed fraction of the disk surface layer remains largely intact as long as internal occultation is not too significant .
similarly , the general analysis presented in appendix a shows that the spherical idealization is not essential for the halo geometry . the dust distribution can be flattened and even distorted into irregular shape before severely affecting the emerging flux .
these results resolve the paradox of successful sed fits for the same star , gm aur , with a flared disk both with a halo ( schneider et al 2003 ) and without one ( chiang & goldreich 1999 ) . the halo contribution to the flux can be absorbed into the disk component by readjusting the flaring law , enabling a successful sed fit without a halo even though it is directly visible in imaging observations .
the problems stapelfeldt et al ( 2003 ) report with disk - only models of hv tau c find a similar explanation .
the halo contribution to the sed can be shifted to the disk , with the halo optical depth added to the flaring angle ( equation [ eq : equiv ] ) .
this leads to unreasonably large flaring , as stapelfeldt et al find .
the same problem was encountered by kikuchi et al ( 2002 ) in modeling flat - spectrum tts .
the halo optical depth in these sources are sufficiently high that they can not be substituted by disks with realistic flaring .
in addition to explaining the shortcomings of disk - only models in these specific cases , the equivalence between halos and cg layers has two important broad consequences : 1 .
_ when the disk is imbedded in a halo that radiates more than its cg layer , the halo becomes the dominant component of the ir flux_. this happens when the halo contains more dust than the halo - equivalent of the disk . from equations [ eq : beta ] and [ eq :
equiv ] , the ir radiation from the system is dominated by the halo contribution whenever where @xmath2 is the optical depth across the halo at visual wavelengths ( 2 for spherical halos ) .
it is impossible to distinguish the cg layer of a flared disk from a halo with flux measurements_. only imaging can produce an unambiguous signature of the cg layer .
the brightness contours of face - on flared disks are concentric circles centered on the star .
inclined viewing changes the contours substantially .
consider the intensity of radiation scattered from the cg surface layer .
it obeys where @xmath30 is the distance to the star and @xmath31 the optical depth toward the observer at the scattering point ( equation [ eq : basic ] ) .
both factors introduce distinct image asymmetries .
the fundamental reason for image distortion by inclination is that the same projected distance from the star corresponds to widely different locations on the surface of the disk . on that surface ,
contours of equal distance from the star are circles of radius .
when viewed face - on from distance @xmath15 , each contour appears as a concentric circle of radius @xmath32 , as seen in the top image in figure [ fig : cgimage ] . at inclination viewing angle @xmath4 to the disk axis , the contour is no longer circular .
absent flaring , the contour becomes an ellipse centered on the star with major axis 2 and minor axis @xmath33 , aligned with the projection of the disk axis on the plane of the sky .
flaring raises the contour to height @xmath34 above the equatorial plane ( figure [ fig : cg ] ) , and the star is shifted toward the observer along the minor axis by @xmath35 .
a point on the contour at position angle @xmath36 from the near side of the minor axis is observed at displacement @xmath37 from the star , where these contours are shown in the bottom image of figure [ fig : cgimage ] .
the off - center position of the star on the minor axis creates an asymmetry such that the far and near portions of this axis obey @xmath38 .
because @xmath39 increases with , this asymmetry increases with distance from the star . at observed displacement @xmath40 from the star , a point on the surface of the disk
is located at @xmath41 . at that point
the optical depth of the cg layer toward the observer is @xmath42 , where therefore the scattering image obeys in this expression , both @xmath5 and @xmath43 are determined at the location @xmath44 on the disk surface .
a power - law grazing angle @xmath45 produces the image @xmath46^{2 + p}/o(\phi)$ ] . this expression and
the resulting brightness contours explain the results of monte carlo scattering calculations for flared disks ( whitney & hartmann 1992 , wood et al 1998 ) .
figure [ fig : asymmetry ] shows the scattering images at three viewing angles of a flared disk with the parameters suggested for ab aur by dullemond et al ( 2001 ) .
the images produced at emission wavelengths are handled in complete analogy .
the only change is the replacement of @xmath47 by the temperature @xmath21 , i.e. , another function of @xmath30 , modifying the dependence of brightness on @xmath48 .
brightness contours not subject to rim occultation are ellipses with eccentricity @xmath49 that directly determines the inclination angle irrespective of the flaring profile .
the images shown in figure [ fig : asymmetry ] possess an additional deviation from circular symmetry , unique to flaring and conveniently measured by the brightness at diametric locations across an axis through the star this asymmetry parameter vanishes for flat disks at all inclination angles , and for pole - on and edge - on viewing irrespective of the flaring .
however , at intermediate inclination angles , flaring introduces substantial asymmetry , as is evident from figure [ fig : asymmetry ] .
non - vanishing @xmath50 is the hallmark of inclined flared disks because it measures the displacement of the isophote centers from the peak brightness position .
its systematic variation with azimuthal angle easily distinguishes it from deviations from the perfect geometry of idealized models or noise in the data .
each flaring profile produces its own characteristic signature @xmath50 .
for example , it is easy to show that the constant grazing angle @xmath5 used in figure [ fig : asymmetry ] gives @xmath51 along the minor axis .
therefore , measuring @xmath50 determines the flaring profile once the inclination is determined from the eccentricity of the brightness contours .
the results of the previous section show that flux calculations of a flared disk in vacuum can be considered a special case of the more general family of models in which the disk is imbedded in an optically thin halo . in the cg case
the `` halo '' is the disk surface layer , fully determined from the flaring geometry .
this layer provides the same ir emission and heating of the underlying optically thick core as its equivalent halo .
therefore , a study of the general halo - imbedded - disk problem contains every possible cg model of flared disks while also covering all cases in which the disk is indeed imbedded in a dusty halo whose optical depth exceeds the bound in equation [ eq : equiv2 ] .
consider a star surrounded by a geometrically thin passive disk and a dusty halo ( figure [ fig : cartoon ] ) .
we study the case of a flat disk and a spherical halo , simplifications that enable us to derive analytical results and broad conclusions that offer important insight .
these simplifications do not cause any serious limitations . as noted previously ,
the halo geometry can be distorted considerably without much impact on the outcome . and
because only the optically thick core of the disk enters into considerations , its only relevant property is its temperature distribution , the surface shape is immaterial . indeed ,
wolf et al ( 2003 ) present detailed model calculation of a flared disk imbedded in an envelope pinched around the equatorial plane , and their results fully conform to our conclusions .
the halo extends from the inner radius to some outer radius @xmath52 .
thanks to scaling , instead of these radii we can specify the dust temperature on each boundary ( see appendix a ) .
the halo is fully characterized by its density profile @xmath53 ( eq . [ eq : eta ] ) and optical depth ; only @xmath54 is relevant in tts and haebes since the star is always visible .
because of its potentially large optical depth , the disk can extend inside the dust - free cavity where its optical depth comes from the gaseous component .
the geometrically - thin disk assumption implies that the disk temperature varies only with radius , vertical temperature structure is ignored .
this temperature is calculated from radiative flux conservation through the gaussian surface in the shape of a small pillbox shown in figure [ fig : cartoon ] .
denote by the radiative flux entering the pillbox from above , including both the stellar and diffuse components .
then where @xmath55 is the 3rd exponential integral , @xmath56 is the disk vertical optical depth and @xmath57 is the cosine of the angle from the disk normal .
the first two term on the left are the disk contribution to the upward flux , the third is the contribution of local intensity transmitted upward through the disk . when the disk is optically thick at frequencies around the peak of @xmath24 ( @xmath58 for @xmath59 ) , the second and third terms can be neglected , leading to the standard expression for disk temperature ( e.g. friedjung 1985 , kenyon & hartmann 1987 ) .
we assume this to be the case everywhere in the disk , an assumption that we check for self - consistency in all our model calculations .
with its temperature derived , the disk emission is calculated from @xmath60 $ ] . because the disk is optically thick around the planckian peak at all radii
, the self - absorption factor can be neglected in the calculation of the disk overall bolometric flux .
then @xmath61 , the disk flux observed at distance @xmath15 and inclination @xmath4 , is proportional to @xmath62 , reflecting the variation of projected area .
this proportionality remains largely unaffected by the envelope attenuation because the short wavelengths , the main contributors to the bolometric flux , emanate from close to the star so that their pathlength is approximately isotropic .
denote by the disk contribution to the overall luminosity @xmath63 and by the contribution of the ( halo + attenuated stellar ) spherical component .
the corresponding flux components are then and the overall flux is @xmath64 where @xmath65 .
the standard `` bare '' disk has @xmath66 ( kenyon & hartmann 1987 ) , therefore in this case @xmath67 .
larger fractions can occur when the disk is imbedded in a halo because of the heating effect of the diffuse radiation , discussed below .
we performed detailed model calculations with the code dusty ( ivezi , nenkova & elitzur 1999 ) which takes into account the energy exchange between the star , halo and disk , including dust scattering , absorption and emission . because its optical depth is typically @xmath68 ,
the halo is transparent to the disk emission in all the models we consider and we neglect the disk effect on the halo . in all the calculations ,
the spectral shapes @xmath69 of the grain absorption and scattering coefficients are those of standard interstellar mix , the sublimation temperature = 1500 k. the spectral shape of the stellar radiation is taken from the appropriate kurucz ( 1994 ) model atmosphere .
the heating rate of a thin flat disk by the stellar radiation at @xmath70 is where is the stellar flux at and @xmath71 , with distance from the axis ( friedjung 1985 ) .
this result reflects the @xmath72 decline of the stellar solid angle and the @xmath73 variation of the grazing angle , yielding disk temperature variation @xmath74 .
natta ( 1993 ) noted that imbedding the disk in a dusty halo can significantly affect its temperature even at small halo optical depths ( see also dalessio , calvet & hartmann 1997 for the effect of optically thick halos ) . with a simple model for scattering at a single wavelength she found that the disk temperature law becomes @xmath75 if the halo density profile is @xmath76 .
our calculations confirm this important point .
figure [ fig : t ] shows the temperature profile for a disk around a = 10,000 k star when `` bare '' and when imbedded in a spherical halo with @xmath77 and = 0.1 and 1 shows that our calculated profile for = 1 dips slightly below that of the bare disk at @xmath78 .
this happens because we neglect the disk emission in the calculation of the halo temperature .
the error introduced by this approximation is of order 2% when = 1 , and even less at smaller . ] .
even though a halo with as small as 0.1 is almost transparent to the stellar radiation , it still causes a large rise in disk temperature .
as is evident from the bottom panel , the halo contribution to the disk heating overtakes the stellar contribution inside the dust - free cavity and dominates completely once the dust is entered .
wolf et al ( 2003 ) present a similar figure for their model .
a dusty envelope with = 0.1 intercepts only about 10% of the stellar luminosity while the disk intercepts 25% of that luminosity .
so how can the halo dominate the disk heating ?
the reason is that direct heating of the disk by the star occurs predominantly at small radii .
the disk absorbs more than 90% of its full stellar allotment within @xmath79 while its entire remaining area , even though so much larger , absorbs only @xmath80 . from equation [ eq : r1 ] ,
the halo starts at @xmath81 , where @xmath82 has already declined to of its value at the stellar surface .
in contrast , the halo emission is isotropic , therefore half of the radiation it intercepts is re - radiated toward the disk , greatly exceeding the direct stellar contribution .
further insight can be gained from the approximate solution presented in ivezi & elitzur ( 1997 ; ie hereafter ) for radiative transfer in spherical symmetry . from equations 20 and b4 of ie
it follows that disk heating by a halo with @xmath68 and @xmath76 is roughly when @xmath83 ; when @xmath84 the factor @xmath85 is replaced by @xmath86 .
this yields @xmath87 , corroborating natta s result and extending its validity beyond the single - wavelength scattering approximation she employed .
the temperature profile is similar to that of a bare disk when @xmath88 but more moderate when @xmath89 . more importantly , and since 100
, the halo dominates the heating at @xmath90 for as small as 0.02 when @xmath91 . as increases , the halo dominance of the heating moves inside the cavity
there the halo heating remains approximately constant while the stellar heating varies as @xmath92 , therefore stellar heating dominates only at @xmath93 , at larger distances the halo takes over .
this explains the results presented in the lower panel of figure [ fig : t ] as well as figure 9 in wolf et al ( 2003 ) .
the figure also shows the temperature profile of the halo .
this profile is largely independent of , varying roughly as @xmath94 when the long - wavelength spectral shape of the dust absorption coefficient is @xmath95 .
the important property evident in the figure is that the disk is much cooler than the envelope at all radii at which both exist and can also contain cooler material in spite of being much smaller , with far reaching consequences for the system radiation . from equation [ eq : fluxes ] ,
the fractional contribution of the disk to the overall bolometric flux is face - on orientation gives the maximal @xmath96 and the standard `` bare '' disk , with @xmath67 , has @xmath97 .
introduce the normalized sed @xmath98 , with similar , separate definitions for the disk and spherical components of the flux .
then since the disk flux obeys @xmath99 for the range of parameters considered here , @xmath100 is independent of the viewing angle @xmath4 , and the entire @xmath4-dependence of the sed comes from the mixing factor @xmath101 .
figure [ fig : seds ] shows sample seds for some representative models .
the halos extend from to 1,000 , with density profiles and overall optical depths as indicated .
the behavior of seds for spherical shells was discussed in ie and since the halo emission is unaffected by the imbedded disk , the seds plotted in dashed lines need no further discussion . the disk , on the other hand , is strongly affected by the halo as is evident from contrasting each disk sed , plotted in long - dashed line , with what it would have been in the absence of a halo ( dot - dashed line ) . the two curves are identical within the first bump around 1 , caused by the stellar heating . in the absence of a halo , the disk sed drops from that peak as @xmath102 . however , halo heating of the outer regions of the disk generates the second broad bump of disk radiation , which is almost two orders of magnitude higher than the `` bare '' disk emission .
the halo heating effect is also evident from other disk properties .
the disks in these models start at the stellar surface and extend to a radius where the temperature is = 25 k. in the absence of a halo , this temperature would be reached at = / = 18 .
as table 1 shows , heating by even a tenuous halo with = 0.1 pushes this radius out by almost factor 5 for the steep density profile @xmath47 and another factor of 2 for the flatter @xmath103 profile which spreads the heating further away from the star .
the impact of halo heating increases with , pushing further out still .
similarly , the disk luminosity , @xmath104 , is only @xmath105 in the absence of a halo but increases as the halo directs more radiation toward it to the point that it becomes @xmath106 when = 1 .
.derived parameters for the models whose seds are presented in figure [ fig : seds ] .
the luminosity ratio of the components is @xmath65 and the disk radius is = , set from the requirement = 25 k. a `` bare '' disk ( = 0 ) has @xmath67 and = 18 . [ cols="^,^,^,^,^ " , ] although it is more compact , the disk can become the stronger emitter at long wavelengths so that the overall sed is dominated by the halo at ir wavelengths and by the disk at sub - mm and mm wavelengths .
this role reversal affects also the wavelength behavior of images .
figure [ fig : images ] shows a series of images at various wavelengths for a sample model . at ir wavelengths
the image is dominated by the halo , displaying the size variation discussed in appendix a. the 0.6 image is dominated by scattering ,
the @xmath107 5images reflect dust emission , leading to size increase with wavelength .
the disk emerges at 17 and dominates the @xmath107 100 images .
the finite beam size and dynamic range of any given telescope could result in an apparent size decrease between 10 and 100 in this case .
such an effect has indeed been discovered in the haebes mwc 137 , whose observed size decreases between 50 and 100 ( difrancesco et al 1994 , 1998 ) .
a switch from envelope to disk domination provides a simple explanation for this puzzling behavior .
no single dust configuration can explain such a decrease , a conclusion reached already in mive and further affirmed by the results of appendix a. difrancesco et al ( 1998 ) suggest that this behavior might reflect multiple , rather than singular , sources of heating but the results of appendix a show the inadequacy of this conjecture .
the region heated by any single source displays an increase of observed size with wavelength , and the superposition of multiple heating sources preserves this behavior .
the opposite trend is possible only when the density distribution contains two distinct components , one optically thick , cool and compact , the other optically thin , warmer and more extended .
a similar effect was detected also in the dust - shrouded main - sequence star vega .
van der bliek , prusti & waters ( 1994 ) find that its 60 size is @xmath108 , yet 850 imaging by holland et al .
( 1998 ) produced a size of only @xmath109 .
so the dust distribution around vega , too , could contain both spherical and disk components with the switch from halo- to disk - dominance occurring somewhere between 60 and 850 .
indeed , imaging at 1.3 mm by willner et al ( 2002 ) reveal the presence of the disk .
the calculation of the disk emission contains two free parameters , involving the temperature and optical depth at the disk edge .
the disk outer radius determines its lowest temperature and a corresponding planck - peak wavelength .
while shorter wavelengths are emitted from a range of disk radii , all longer wavelengths originate from the edge of the disk .
the resulting effect can be seen in figure [ fig : td ] , which shows two representative values of . by the model assumptions
the disk must be optically thick at the local planck - peak wavelength everywhere , 350 at = 25 k. as long as the disk edge remains optically thick also at longer wavelengths , the emission follows the rayleigh - jeans profile @xmath110 .
once the disk edge becomes optically thin , the sed switches to the steeper decline @xmath111 at longer wavelengths .
the break in the disk sed is controlled by the optical depth of the disk edge , which we specify at 350 .
figure [ fig : td ] shows also the effects of varying @xmath112 from its smallest value @xmath113 to a value sufficiently large that the edge is optically thick at all the displayed wavelengths .
the bottom panel of each model in figure [ fig : seds ] shows the variation of the overall sed with viewing angle @xmath4 .
the entire variation comes from the mixing factor @xmath101 ( see equations [ eq : rho ] and [ eq : f ] ) .
since the parameters @xmath114 and @xmath4 enter only in the product @xmath115 but not separately , the sed is subject to a degeneracy : systems viewed at different inclination angles will have the same sed if they have the same @xmath115 in addition to all other properties . because of the rapid decline with distance of the radiation absorbed from the star ( cf eq .
[ eq : hstar ] ) , the disk luminosity , i.e. , @xmath114 , has a steep dependence on its inner radius ; moving the disk inner edge from to only 2 removes 56% of the stellar luminosity intercepted by the disk , 3 results in a 72% removal .
_ such central holes reduce @xmath114 but do not impact any other relevant property _ because they remove only the hottest disk material whose contribution to the overall flux is negligible in comparison with the stellar component .
figure [ fig : inclinations ] plots contours in the @xmath4 plane of constant mixing factor @xmath101 .
it shows , for example , that the seds presented in figure [ fig : seds ] for @xmath4 = 70 would be the same for systems viewed at @xmath4 = 35 if the disk inner radius is increased from 1 to 2.2 in the = 0.1 case and 6 for = 1 .
although the sizes of these holes can not be determined from sed modeling of single stars , from statistical arguments mive conclude that their existence is essential to produce a plausible distribution of inclination angles .
thanks to the cg layer halo equivalence , each model presented here describes also a disk with no halo and with the flaring angle defined by eq.[eq : beta2 ] from @xmath16 and .
in particular , the sed presented for @xmath116 halos describe also flared disks without halo and with @xmath117 , those with @xmath118 cover flared disks with constant grazing angle @xmath5 ( @xmath119 ) .
the equivalence strictly holds only for disks and halos of the same size .
however , the models presented would be little affected if each halo was truncated at because that would only remove halo emission at long wavelengths where the sed is dominated by the disk anyhow .
since the flaring angle of each equivalent disk reaches 2 at its outer edge , halos whose optical depths require excessive flaring can not be realized with disks only .
this was the problem recognized by stapelfeldt et al ( 2003 ) and kikuchi et al ( 2002 ) .
the equivalence holds only for the sed .
high - resolution observations would produce widely different images for each halo and its equivalent flared disk , except when the latter is viewed face - on .
the results of [ sec : cg ] show that every disk imbedded in a halo with @xmath120 can be replaced by a flared disk without a halo and with an identical flux .
this mathematical degeneracy explains the success of sed modeling with disk alone a system like gm aur ( chiang & goldreich 1999 ) even though the halo was subsequently discovered in imaging observations ( schneider et al 2003 )
. it also explains why such modeling runs into difficulties and requires excessive flaring when the halos have larger , as is the case in hv tau c ( stapelfeldt et al 2003 ) and flat - spectrum tts in general ( kikuchi et al 2002 ) . disks generally do not exist in pure vacuum .
equation [ eq : equiv2 ] defines the circumstances under which the surrounding dust becomes the dominant component of the ir flux . even at smaller ,
when not dominating the overall flux , the halo can still dominate the disk heating and make a strong impact on its temperature profile ( [ sec : temp ] ) .
ignoring the surrounding material can produce misleading results regarding the disk properties , such as its flaring profile . in spite of the attractiveness of the flared disk as a simple , physical model without any additional components ,
imaging observations give irrefutable evidence for the existence of extended halos in pre main - sequence stars .
the origin of these halos has not been studied yet .
stapelfeldt et al ( 2003 ) suggest that a replenishment process , either continued infall from the surrounding ism or a dusty outflow from the source itself , is operating .
it is noteworthy in this regard that accretion with the small rates of yr@xmath121 has been deduced from uv spectra of both haebes ( grady et al .
1996 ) and tts ( valenti , basri & johns 1993 ; gullbring et al 1998 ; lamzin , stempels & piskunov 2001 ) , and is consistent with halos that have 0.1 ( mive ) .
these low rates can not correspond to the main accretion buildup of the star but rather a much later phase , involving small , residual accretion from the environment .
the corresponding accretion luminosities are only 0.1 , justifying their neglect in our calculations .
in addition to the cg layer halo equivalence , our results reveal numerous degeneracies that underscore the severe limitations of attempts to determine the dust morphology from sed analysis without imaging observations .
the sed of a halo - imbedded - disk remains the same when the viewing angle and the size of the disk central hole vary together as shown in figure [ fig : inclinations ] . from the results of appendix
a , the sed of a spherical shell with power - law density profile @xmath122 displays a dependence on @xmath123 only when @xmath124 , and then only in the spectral region @xmath125 .
all other regions of @xmath123 and @xmath126 produce the same universal behavior @xmath127 ( [ sec : spherical ] ) .
the dust optical properties introduce additional degeneracies .
the results of [ sec : spherical ] show that the frequency dependence of @xmath128 and the radial dependence of the density profile can be interchanged on occasion without affecting the sed .
the fundamental reason for all these degeneracies is that the flux from an optically thin source involves a volume integration ( eq . [ eq : flux ] ) that tends to remove much of the dependence on the underlying morphology .
although the specific degeneracies we uncovered involve spherical geometry , the general analysis in appendix a shows that the spherical idealization is not essential . the dust distribution can be flattened and even distorted into irregular shape before severely affecting the results .
these degeneracies make it impossible to determine the geometry from a fit to the sed alone without additional input .
only imaging can trace the actual density distribution , and scattering provides a more faithful presentation because , unlike emission which involves also the dust temperature , it involves only the density distribution .
reliance on sed modeling alone can produce misleading results , as was the case for the parameters of the flared disk in gm aur .
the partial support of nsf and nasa is gratefully acknowledged .
.i . acknowledges generous support by princeton university
. = butner h.m . , natta a. & evans n.j .
ii , 1994 , apj , 420 , 326 chiang , e.i . & goldreich , p. , 1997 ,
apj , 490 , 368 ( cg ) chiang , e.i . & goldreich , p. , 1999 ,
apj , 519 , 279 chiang , e.i . , et al 2001 , apj , 547 , 1077 creech - eakman , m.j . , chiang , e.i . , joung , r.m.k . ,
blake , g.a . & van dishoeck , e.f . , 2002 , a&a , 385 , 546 dalessio , p. , calvet , n. & hartmann , l. 1997 , apj , 474 , 397 di francesco j. , et al , 1994 , apj , 432 , 710 di francesco j. , et al , 1998 , apj , 509 , 324 dullemond , c.p . , et al , 2001 , apj , 560 , 957 friedjung , m. , 1985 , a&a , 146 , 366 grady c.a .
, et al . , 1996 , a&as , 120 , 157 grady , c.a . , et al , 2001 , aj , 122 , 3396 gullbring , e. , hartmann , l. , briceo , c. & calvet , n. , 1998 , apj , 492 , 323 harvey , p.m. , lester , d.f . , brock , d. & joy , m. , 1991 , apj , 368 , 558 hillenbrand , l. a. , strom , s. e. , vrba , f. j. , & keene , j. 1992 , apj , 397 , 613 holland , w.s .
1998 , nature , 392 , 788 ivezi . ,
elitzur m. , 1997 , mnras , 287 , 799 ( ie ) ivezi . ,
nenkova m. & elitzur m. , 1999 , user manual for dusty , internal report , university of kentucky , accessible at http://www.pa.uky.edu/@xmath129moshe/dusty/ kenyon s.j . , hartmann l. 1987 , apj , 323 , 714 kikuchi , n. , nakamoto , t. , & ogochi , k. 2002 , pasj , 54 , 589 kurucz r.l .
, 1994 , cd - rom no.19 , smithsonian astrophys .
lamzin , s.a . ,
stempels , h.c . & piskunov , n.e . , 2001 , a&a , 369 , 965 miroshnichenko , a.s . , ivezi , . ,
vinkovi , d. & elitzur , m. , 1999 , apj , 520 , l115 ( mive ) natta , a. 1993 , apj , 412 , 761 natta , a. et al . , 2001 , a&a , 371 , 186 polomski , e.f . ,
telesco , c.m . ,
pia , r. , & schulz , b. , 2002 , aj , 124 , 2207 schneider , g. , et al , 2003 , aj , 125 , 1467 stapelfeldt , k.r . , et al , 2003 , apj , 589 , 410 valenti , j.a . , basri , g. & johns , c.m .
1993 , aj , 106 , 2024 van der bliek , n.s . , prusti , t. & waters , l.b.f.m .
1994 , a&a , 285 , 229 whitney , b.a . &
hartmann , l. , 1992 , apj , 395 , 529 willner , d.j . ,
holman , m.j . ,
kuchner , m.j .
& ho , p.t.p . , 2002 , apj , 569 , l115 wolf , s. , padgett , d.l . , &
stapelfeldt , k.r . , 2003 , apj , 588 , 373 wood , k. , kenyon , s.j . ,
whitney , b. & turnbull , m.w .
, 1998 , apj , 497 , 404
consider a cloud heated from inside by a star of radius and effective temperature ( figure [ fig : imaging ] ) .
the star clears out a dust - free cavity of radius , determined by dust sublimation @xmath130 . when the dust is optically thin , the cavity radius can be found from where @xmath131 is the planck average at temperature @xmath21 of the absorption cross section @xmath128 ( ie ) . with standard interstellar dust and = 1500 k ,
the cavity radius obeys @xmath132 at a typical haebes temperature = 10,000 k. in tts , on the other hand , the dust is much closer to the star : / is only 15 at = 5,000 k and as small as 3 at = 3,000 k. the intensity at frequency @xmath8 and impact parameter @xmath133 ( fig .
[ fig : imaging ] ) is assuming isotropic scattering . here
is the dust density , @xmath134 is distance along the path to the observer , @xmath135 is the albedo at frequency @xmath8 and @xmath136 .
this expression neglects self - absorption by the dust ; the error in this approximation is of order @xmath137 . since the dust temperature can not exceed the sublimation temperature ,
there is no dust emission at @xmath138 4(1000k)/ , only scattering .
diffuse radiation and attenuation between the star and the scattering point can be neglected since our discussion is centered on optically thin dust .
then the only source of scattering is the stellar radiation with energy density @xmath139 , where @xmath140 is the stellar luminosity at frequency @xmath8 . from equation [ eq : basic ]
, the scattered brightness is where @xmath141 and @xmath15 is the distance to the observer . since the frequency- and geometry - dependence separate out , all scattering wavelengths share a common image .
only the brightness level varies with @xmath8 , and because of the wavelength decline of @xmath142 the observed size generally _ decreases with wavelength _ when traced to the same brightness level . in any geometry
the scattering image always traces directly the variation of column density along the line of sight , the dust temperature profile is irrelevant . at wavelengths longer than 3
, @xmath143 and scattering can be neglected .
the planckian enters in equation [ eq : basic ] as a function of @xmath21 at fixed @xmath8 , which can be well approximated by its rayleigh - jeans limit at @xmath144 and a sharp cutoff at @xmath145 . with this approximation
the integration is limited to locations along the path where @xmath146 ; regions with @xmath147 are too cold to emit appreciably at frequency @xmath8 . excluding highly patchy geometries , the highest temperature on the path occurs at @xmath148 ( i.e. , @xmath149 ) , the closet distance to the star , and only paths with @xmath150 contribute to the brightness .
as @xmath134 increases in either direction , @xmath21 decreases .
the integration is truncated either because the temperature becomes too low , in which case the emission is _ temperature bounded _ , or because the edge of the source is reached and the emission is _
matter bounded_. denote the resulting integration limits @xmath151 ( @xmath4 = 1 , 2 ) , then in the matter bounded case @xmath151 is the edge of the source , the integral is independent of @xmath8 and the frequency dependence of the intensity follows @xmath152 . in the case of temperature bounded emission the integration limits introduce additional @xmath8-dependencies that modify this behavior .
however , the integration can be extended to @xmath153 whenever ( 1 ) @xmath154 and ( 2 ) the product @xmath155 of dust density and temperature declines along the path faster than @xmath156 .
therefore , _ when these two conditions are met , the frequency variation of optically thin emission is @xmath157 even when it is temperature bounded_. independent of geometry , all frequencies that obey these conditions produce a common image , similar to the scattering case ; only the scale of brightness varies with @xmath8 .
this result makes it possible to determine the wavelength dependence of the dust cross section directly from imaging observations .
similar to the variation along the line of sight , when @xmath133 increases ( moving away from the star ) the emission again is truncated by either the matter or temperature distribution .
denote by the temperature at the source outer edge .
the corresponding emission cutoff wavelength is when @xmath158 , the dust is sufficiently warm everywhere that the emission is truncated only by the matter distribution .
the observed size is then @xmath159 , the angular displacement of the source edge from the star , the same for all wavelengths .
however , when @xmath160 the brightness is truncated when @xmath161 before the edge of the source is reached , resulting in a wavelength - dependent angular size @xmath162 .
the observed size of optically thin emission _ increases with wavelength _
so long as @xmath160 , the opposite of the trend at scattering wavelengths
. the frequency variation of the dust cross section is well described by @xmath163 with @xmath164 = 12 . then to a good degree of approximation ,
the temperature variation of optically thin dust is @xmath165 , where @xmath166 , producing the wavelength - dependent observed angular size since @xmath167 , @xmath168 increases faster than @xmath169 , a fairly steep rise
. the flux can be obtained from equation [ eq : basic ] by integration over the observed area . at emission wavelengths , the flux at distance @xmath15 is
since the temperature profile of optically thin dust depends only on distance from the star , the dependence on the source geometry enters only from . as before ,
the integration is truncated by either the temperature or the matter distribution . whenever @xmath158 at every point on the surface , the emission is matter bounded everywhere and the integration encompasses the entire source . under this circumstances @xmath170 ,
a universal sed that depends only on the dust properties irrespective of geometry . in particular
, the spectral variation @xmath163 gives @xmath171 , therefore the signature of matter bounded emission is this sed accompanied by wavelength independent images ; this is the expected behavior in any geometry at sufficiently long wavelengths . at @xmath172
the integration volume is truncated by the temperature , and since the emission volume decreases as the frequency increases , the rise of @xmath173 with @xmath8 becomes less steep than in the matter dominated regime .
therefore , the sed changes from @xmath171 at @xmath158 to a flatter slope at @xmath160 .
the break in the slope at can be used to determine the surface temperature .
flux spectral variation shallower than @xmath174 is a clear indication of temperature - bounded emission and should be accompanied by an image size that increases with wavelength .
some explicit results are easily derived in the case of spherical symmetry .
thanks to the scaling properties of dust radiative transfer ( ie ) , only two properties are required to specify the geometry .
the first is the radial optical depth at one wavelength , say @xmath175 where @xmath10 is the cross - section at visual ; at every other wavelength , @xmath176 where @xmath177 .
the second is the dimensionless , normalized profile of the dust density distribution where @xmath18 ; note that @xmath178 .
explicit results follow immediately for all power - law density profiles where the shell extends to the outer radius @xmath179 , subtending the angular region @xmath180 , where @xmath181 and @xmath182 . at scattering wavelengths
@xmath183 whenever @xmath184 the integration can be extended to @xmath153 , yielding @xmath185 ; the brightness decreases as a power law so long as the observation direction is not too close to the halo edge . at emission wavelengths , on the other hand , @xmath186 where the observed size @xmath168 is smaller than @xmath159 when @xmath172 ( equation [ eq : size ] ) .
as long as @xmath187 the integration can be extended to @xmath153 and the brightness then decreases along any radial direction in proportion to @xmath188 . the flux integration in equation [ eq : flux ] is similarly terminated at the observed boundary @xmath168 , producing since @xmath189
, there are two families of sed . in the case of steep density distributions with @xmath190 , the first term inside the brackets
can be omitted because @xmath191 .
such distributions produce @xmath170 irrespective of the actual value of @xmath123 .
since typically @xmath192 , this behavior applies to all cases of @xmath193 . on the other hand , whenever @xmath194 the omitted term dominates and the sed is a broken power law .
the power index switches from the universal @xmath195 at @xmath196 to the geometry - dependent value @xmath197 ( see also harvey et al 1991 ) at @xmath125 .
these results show that the sed displays a dependence on the density profile only when @xmath124 , and then only in the spectral region @xmath198 .
all other regions of @xmath123 and @xmath126 produce the universal behavior @xmath170 .
|
we study the ir emission from flared disks with and without additional optically thin halos .
flux calculations of a flared disk in vacuum can be considered a special case of the more general family of models in which the disk is imbedded in an optically thin halo . in the absence of such halo , flux measurements can never rule out its existence because the disk flaring surface defines a mathematically equivalent halo that produces the exact same flux at all ir wavelengths .
when a flared disk with height @xmath0 at its outer radius @xmath1 is imbedded in a halo whose optical depth at visual wavelengths is @xmath2 , the system ir flux is dominated by the halo whenever @xmath3 . even when its optical depth is much smaller , the halo can still have a significant effect on the disk temperature profile .
imaging is the only way to rule out the existence of a potential halo , and we identify a decisive test that extracts a signature unique to flared disks from imaging observations .
[ firstpage ] circumstellar matter dust infrared : stars radiative transfer stars : imaging stars : formation stars : pre - main - sequence
|
observations in the millimeter continuum provide the best method to identify a galaxy - wide sample of star - forming clumps .
the emission is optically thin , minimally affected by temperature and can be surveyed over large areas . unlike targeted observations ,
blind surveys allow for a complete and systematic study of dense gas clumps . in the past decade
, there have been many galactic plane surveys at millimeter / submillimeter wavelengths , of which the bolocam galactic plane survey @xcite was the first to be completed and publicly released .
atlasgal @xcite surveyed the southern galactic plane at 870 @xmath8 .
the jcmt galactic plane survey , or jps , will survey the northern plane at 850 @xmath8 . in the past 3 years
, the hi - gal galactic plane survey observed the galaxy from 70 to 500 @xmath8with the herschel space observatory , sensing the peak of the dust sed with minimal spatial filtering @xcite .
the survey has provided access to the peak of the dust spectral energy distribution at modest ( @xmath9 ) resolution @xcite .
together , these surveys provide a complete map of long - wavelength dust emission across the galactic plane .
long - wavelength data are essential for constraining the dust emissivity , one of the free parameters in greybody spectral energy distribution ( sed ) fits .
@xcite demonstrated the need for long - wavelength data to accurately determine both @xmath10 , the dust emissivity spectral index , and temperature .
@xcite also showed that adding additional wavelengths to an sed fit , even with lower signal - to - noise , significantly reduces the degeneracy in the fit .
millimeter - wave dust emission also has the advantage of being relatively insensitive to temperature .
when looking at cold gas , @xmath11 k , all of the herschel bands deviate from a rayleigh - jeans temperature approximation .
longer wavelength observations are less affected by temperature assumptions .
the 1.1 mm band is in many cases the longest wavelength unaffected by free - free emission , providing the least environmentally - biased view of optically thin dust emission and therefore total dust mass .
similarly , at millimeter wavelengths , the dust opacity is low enough that all clumps detected in the bgps are expected to be optically thin ( with the possible exception of sgr b2 ; * ? ? ?
* ) . in combination with the weak temperature dependence ,
this feature of 1.1 millimeter emission allows for the most direct and straightforward estimates of total dust mass .
millimeter - bright dust clumps are generally associated with high - density , star - forming gas .
previous surveys have found cold , massive molecular clouds via the @xmath12and @xmath131 - 0 lines @xcite .
however , these clouds are only moderate density , @xmath14 @xmath15 , while dust - detected clumps have typical densities @xmath16 @xmath15@xcite .
the dense gas in these clumps is more directly associated with star and cluster formation @xcite , allowing for systematic studies of pre - star - forming and star - forming gas .
the bgps v1.0data has been public since 2009 , and has been used extensively as both a finder chart and a tool to probe galactic properites .
it was used to examine the properties of maser sources @xcite , outflow sources @xcite , and high - mass star - forming regions @xcite .
it has served as the basis for studies of forming clusters @xcite and intermediate - mass stars @xcite .
the bgps and other surveys have served as finder charts for large - scale millimeter line studies of the galactic plane @xcite .
bgps clumps have been used as the target sample for distance determinations to large cloud populations @xcite .
@xcite used the bgps to measure properties of star forming regions as a function of galactocentric radius .
these and many other ongoing and planned studies demonstrate the need for , and benefits of , publicly available blind legacy surveys .
this paper presents v2.0of the bolocam galactic plane survey ( bgps ) , with a complete data release available at irsa.ipac.caltech.edu / data / bolocam_gps/. in paper i @xcite , the initial processing of the bgps v1.0was described in detail .
it was noted in section 5 of @xcite that there was a discrepancy between our survey and previously published results .
this discrepancy raised the possibility of a flux calibration error in the version 1 ( hereafter v1.0 ) results : we confirm and correct the error in this paper .
in addition , we have made significant improvements to the data pipeline , measured important features of the pipeline including its angular transfer function , improved the pointing accuracy , and added new observations .
the paper is as follows : we resolve the flux calibration discrepancy in . in , we discuss new observations included in the v2.0data
. describes changes to the data reduction process and new data products . and measure the angular transfer function of the bgps v2.0pipeline and properties of extracted sources respectively .
the paper concludes with a brief discussion of the results and a summary .
the original calibration , along with tables of color correction and a detailed treatment of the filter response , are described in @xcite .
we discuss important changes in v2.0 in this section . in @xcite
, we reported that a ` correction factor ' of about 1.5 on average was needed to bring our data into agreement with other 1 mm data sets .
we discovered that the published v1.0bgps images have a different calibration reported in their fits headers than was used in processing the data .
the calibration used in the released data was borrowed from a previous observing run , during which a different bias voltage was used , and differed from the pipeline - derived calibration by a factor @xmath17 , completely explaining the discrepancy .
we checked the data for consistency with the measured calibration offset . in order to compare flux densities in identical sources , we performed aperture photometry on the v2.0data based on the locations of v1.0sources using both the ` source masks ' from bolocat v1.0@xcite and circular apertures centered on the bolocat v1.0peaks .
source masks , also known as label masks , are images in which the value of a pixel is either 0 for no source or the catalog number if there is a source associated with that pixel .
= " 2d we measured the multiplicative offset between v1.0 and v2.0by comparing these aperture - extracted fluxes .
for each aperture size , we measured the best - fit line between the v1.0and v2.0data using a total least squares ( tls ) method weighted by the flux measurement errors as reported in the catalogs . the agreement with @xmath18 ,
as expected based on section [ sec : caloffset ] , is within 10% , although the larger apertures show a slight excess with @xmath19 .
this excess is expected given the improved extended flux recovery in v2.0(see section [ sec : stf ] ) .
the v2.0/v1.0 flux ratio is weakly dependent on the source flux density , with higher v2.0/v1.0 ratios for brighter sources . in section 5.5 of @xcite
, we compared the bgps v1.0 data to other data sets from similar - wavelength observations .
we repeat those comparisons here using the v2.0 data and demonstrate that v2.0achieves better agreement with other data sets than v1.0 .
full details of the comparison were given in @xcite .
we compare to 3 data sets in the same @xmath20 mm atmospheric window .
two data sets from mambo ii , the ( * ? ? ?
* m07 ) cygnus x survey and the ( * ? ? ?
* r06 ) irdc survey , overlap with the bgps .
the simba 1.3 mm survey of the @xmath21 region is the largest survey in the 1 mm band that overlaps with ours ( * ? ? ?
* m09 ) .
the comparison data sets have angular transfer functions that differ from the bgps . in order to account for the difference , we allow for a large angular scale offset between the observations .
we fit a line of the form @xmath22 to the data , where @xmath23 and @xmath24 represent the pixel values gridded to 7.2pixels .
the @xmath25 value allows for a local offset , i.e. a non - zero @xmath25 value indicates a substantial difference in the angular transfer function .
since such an additive offset is unlikely to apply across the entire observed region , we also fit the offset for small sub - regions in the m07 and m09 data , focusing on dr21 and a region centered on g45.5 + 0.1 respectively .
the results of that comparison are displayed in table [ tab : fluxcomparison ] , which includes the original comparison from @xcite .. ] bgps v2.0is in much better agreement with the other data sets than v1.0 , but it retains a significant additive offset , particularly with respect to mambo .
the additive offset is explained by a difference in the angular transfer function ; the mambo observing strategy of fast position switching allows structures on the scale of the array to be preserved , while bolocam s fast - scan strategy does not .
the differing observing strategy explains why there is an additive offset between bolocam and mambo , but no such offset for simba , which was used in a fast - scan mode similar to bolocam .
the varying backgrounds in separate regions account for some of the remaining multiplicative offset .
when individual sub - regions are compared , the additive and multiplicative offsets more clearly separate into independent components , i.e. a line with an additive offset is a better fit to the data . to enable a comparison of flux density between the surveys ,
we must account for the different spectral bandpasses of the instruments .
the relative flux density measured between the instruments depends on the spectral index @xmath26 of the observed source ; @xmath27 corresponds to a perfect black body on the rayleigh - jeans tail . will be unaffected .
the assumption of constant spectral index with flux is reasonable since observed spectral index - flux correlations are shallow @xcite . ] in table [ tab : relflux ] we show the relative flux densities expected for bolocam , mambo and simba ; they differ by at most 19% for spectral indices @xmath28 .
bolocam flux densities are expected to be higher because bolocam has a higher effective central frequency than either of the other instruments . in @xcite , we measured bolocam v1.0/mambo and bolocam v1.0/simba ratios in the range @xmath29 , indicating a clear disagreement between the surveys . with the v2.0 data ,
we measure ratios @xmath30 and @xmath31 .
these numbers still indicate that the bgps is too faint by @xmath32 relative to the expectations laid out in table [ tab : relflux ] , but with a systematic calibration error no better than 20% in each survey , this level of agreement is reasonable .
thirteen nights of additional data were acquired from december 15th , 2009 to january 1st , 2010 .
the target fields and areas covered are listed in table [ tab : observations ] as boxes in galactic latitude and longitude , with position angles to the galactic plane indicated .
the original observations are described in section 2 of @xcite .
the new target fields were selected from visual inspection of fcrao ogs @xmath12 integrated maps , @xcite @xmath12 maps , and iras 100 @xmath8maps .
the fields were selected primarily to provide even spacing in ra in order to maximize observing efficiency , and were therefore not blindly selected . additionally , the orion a and b and mon r2 clouds were observed in observing campaigns by collaborators .
these complexes are not directly part of the bgps , but are included in this data release reduced in the same manner as the galactic plane data .
they are much closer than typical bgps sources and their selection for mapping is very biased , but we include them in the archival data . parts of the orion a nebula remain proprietary as of this release , but are expected to be released upon publication of kauffmann et al ( in prep ) . the california nebula has also been observed and the data published in @xcite . finally , some archival cso data was recovered and added to the bgps .
these data include maps of m16 and m17 .
m17 is an extraordinarily bright 1.1 mm source that was poorly covered in the bgps because it is below @xmath33 .
the bolocat cataloging tool was run on these new fields and they have been included in the v2.0catalog .
some of their properties are displayed in section [ sec : sourceextraction ] .
a total of 591new sources not covered in the v1.0survey were extracted .
and orion b fields were inadvertently excluded ; these are now included in a v2.1 release . ]
observations at wavelengths longer than 2 @xmath8and shorter than 2 cm from the ground are strongly affected by emission and absorption from our own atmosphere .
optical and radio observations from the ground see through a transparent atmosphere , but millimeter observations are dominated by bright foreground emission that dominates the astrophysical signal .
this foreground must be removed in order to create maps of astrophysical emission .
@xcite presented a summary of the techniques used to separate astrophysical and atmospheric signals in ( sub)millimeter bolometric observations .
the bolocam observations reported in this paper were conducted with a fast - scanning strategy that places some of the ` fixed ' astrophysical emission at a different sampling frequency than the varying foreground atmosphere .
this approach is one of the most efficient and flexible and has been used predominantly over alternatives , such as a nodding secondary , in most recent large - scale observing campaigns @xcite .
a variety of different atmospheric removal algorithms have been successfully utilized , but in addition to removing the atmospheric foreground , these approaches remove some of the astrophysical signal . in order to recover signal on angular scales up to the array size ,
the most commonly used approach for bright galactic signals is an iterative reconstruction process .
this process assembles a model of the astrophysical emission , subtracts it from the observed timestream , and repeats , each time reducing the amount of astrophysical signal that is removed by the algorithm .
this general approach was first used on bolocam data by @xcite and refined in @xcite .
we directly examine the effects of the data reduction methods below .
we compared a few different methods for atmospheric subtraction and astrophysical image reconstruction , but settled on an approach very similar to that used in v1.0 .
this subsection recounts the minor changes from v1.0and includes discussion of alternative approaches . the pca method @xcite with iterative flux density restoration was used for v2.0as for v1.0 . in the pca atmosphere removal method ,
the @xmath34 eigenvectors corresponding to the highest values along the diagonal of the covariance matrix ( the most correlated components ) are nulled .
we nulled 13 pca components in both v1.0and v2.0 .
the selection of 13 components produced the best compromise between uniform background noise and fully restored peak signal .
simulations show that the point source recovery is a very weak function of number of pca components nulled ( npca ) , while extended flux recovery is a strong function of npca .
however , residual atmospheric signal was substantially reduced with higher npca . in v2.0 ,
20 iterations were used instead of the 50 used in v1.0 ; in both surveys , convergence was clearly achieved by 20 iterations , and generally individual iterations are indistinguishable by @xmath35 iterations .
the iterative process adopts the non - negative flux density above some cutoff as a model of the astrophysical sky and subtracts that flux density from the timestream before repeating the atmospheric subtraction .
this approach allows large angular scale structures to be recovered by removing them from the timestreams before they can contribute to the correlated signal .
the v2.0pipeline was more succesful than in v1.0at removing negative bowls ( see section [ sec : catalogmatching ] for visual examples ) .
negative bowls are introduced because the atmospheric subtraction process assumes that the mean level of any timestream , and therefore any map , is zero .
the iterative process allows this assumption to be violated , creating maps with net positive signal .
the reduced impact of negative bowls is attributed to a few small changes to the pipeline that each slightly mitigate the bowls . 1 .
the astrophysical model is created by deconvolving the positive emission rather using positive pixels directly .
the deconvolution process , which removes sub - beam - scale noise , was made more stable in v2.0by performing a local signal - to - noise cut using the noise maps described in section [ sec : noisemaps ] ; in v1.0there was no reliable noise map available during the iterative map making process .
better image co - alignment reduced inter - observation spatial offsets .
the astrophysical models therefore better reproduced the timestream data .
improvement in the bolometer gain calibration , which is done on a per - observation basis in v2.0 , improved the convergence of the iterative map maker .
these changes are individually minor , but together resulted in significant improvements to the map quality .
the quadratic planar fit sky subtraction method discussed in @xcite was implemented and tested for 1.1 mm galactic plane data in the v2.0 pipeline , but was not used for the final data products . in principle , this method should do a substantially better job at removing smooth atmospheric signal from timestreams than pca cleaning because it is based on physically expected atmospheric variation .
the spatial recovery was better than the aggressive 13-pca approach , but as with a simpler median subtraction approach ( subtracting the median timestream from all bolometers ) , a great deal of spurious signal from the atmosphere remained in the maps , and the noise properties were highly non - uniform , rendering source extraction difficult .
it was also more computationally expensive and did not remove correlated electronic readout noise , which pca subtraction did .
the @xcite approach is likely more effective at 143 ghz because the atmosphere is better - behaved at lower frequencies .
we speculate that it is also more effective for deep extragalactic fields in which more repeat observations of the same field are able to distinguish atmospheric from real signal on the angular scales of the array .
the bgps data are available from the infrared processing and analysis center ( ipac ) at irsa.ipac.caltech.edu / data / bolocam_gps/. the v2.0data products include the pipeline - processed maps and bolocat label masks as in v1.0 . in the v2.0data release , there are two new map types released : noise maps and median maps
. a variant of the noise maps was produced in v1.0 , while the median maps are an entirely new data product .
residual bolometer timestreams are automatically generated as part of the iterative map - making process .
the residual is the result of subtracting the astrophysical model ( which is smooth , noiseless , and non - negative ) from the atmosphere - subtracted data timestream .
the resulting timestream should only contain the remaining astrophysical noise .
however , maps of the residual timestream contain sharp edge features because the astrophysical model is sharp - edged ( i.e. , transitions from 0 to a non - zero value from one pixel to the next ) .
these sharp transitions are mitigated in the presence of noise .
we therefore created noise maps by taking the local standard deviation of the residual map .
pixels in the original map that were not sampled ( i.e. , represented by ` nan ` in the fits data file ) are ignored when computing this local standard deviation and their values are set to be an arbitrarily high number ( @xmath36 jy / beam ) such that pixels near the map edge are assumed to have extremely high noise ( which is reasonable , since these pixels are affected by a variety of artifacts rendering them unreliable measurements of the true astrophysical flux ) . the local noise is computed within a @xmath37-pixel gaussian , which enforces a high noise level within @xmath38 of the map edge .
this method produces good noise maps ( i.e. , in agreement with the standard deviation calculated from blank regions of the signal map ) and was used both within the iterative process and for cataloging .
we show the noise per pixel for each half square degree in the inner galaxy in figure [ fig : noiseperdeg ] .
the noise level in each outer - galaxy field is summarized in figure [ fig : noiseperfield ] . because the outer galaxy coverage is irregular ,
we show the noise per observed region rather than dividing the regions into degree - scale sub - regions .
some artifacts ( cosmic ray hits , instrumental artifacts ) inevitably remained at the end of the process . in order to mitigate these effects ,
`` median maps '' were created .
the value of each spatial pixel was set to the median value of the timestream points intersecting that pixel ; pixels with fewer than 3 data points were set to ` nan ` .
the noise in the median maps was in some cases lower than that in the weighted mean maps , particularly for fields with fewer total observations .
they uniformly have mitigated instrument - related artifacts such as streaking .
these maps are released in addition to the weighted - mean maps , which often have higher signal - to - noise . in order to get the best possible pointing accuracy in each field , all observations of a given area were median - combined using the montage package , which performs image reprojections , to create a pointing master map @xcite .
each individual observation was then aligned to the master using a cross - correlation technique @xcite : 1 .
the master and target image were projected to the same pixel space 2 .
a cross - correlation image was generated , and the peak pixel in the cross - correlation map was identified 3 .
sub - pixel alignment was measured by performing a 2nd - order taylor expansion around the peak pixel this method is similar to the version 1.0 method , but the new peak - finding method proved more robust than the previous gaussian fitting approach . the v1.0gaussian fitting approach is often used in astronomy ( e.g. , http://www.astro.ucla.edu/~mperrin/idl/sources/subreg.pro ) , but it is biased when images are dominated by extended structure
this bias occurs because the least - squares fitting approach will identify the broader peak that represents auto - correlation of astrophysical structure rather than the cross - correlation between the two images . in v1.0
, we attempted to mitigate this issue by subtracting off a ` background ' component before fitting the gaussian peak , but this method was not robust .
the improved approach to pointing resulted in typical rms offsets between the individual frames and the master @xmath39 .
the improvement in the point spread function is readily observed ( see section [ sec : catalogmatching ] ) .
we carefully re - examined the pointing throughout the bgps using a degree - by - degree cross - correlation analysis between the v1.0 , v2.0 , and herschel hi - gal 350 @xmath8data .
the herschel data were unsharp - masked ( high - pass filtered ) by subtracting a version of the data smoothed with a @xmath40 gaussian .
the result was then convolved with a @xmath41 gaussian to match the herschel to the bolocam beam sizes .
errors on the offsets were measured utilizing the fourier scaling theorem to achieve sub - pixel resolution ( inspired by * ? ? ? * ) .
the errors on the best - fit shift were determined using errors estimated from the bgps data and treating the filtered hi - gal data as an ideal ( noiseless ) model .
the tools for this process , along with a test suite demonstrating their applicability to extended structures in images , are publicly available at http://image - registration.readthedocs.org/. the cross - correlation technique calculated the @xmath42 statistic as a function of the offset . for a reference image @xmath43 and observed image @xmath44 with error per pixel @xmath45 , @xmath46 where @xmath47 and @xmath48 are the pixel shifts . because y is not actually an ideal model but instead is a noisy image , we increase @xmath45 by the rms of the difference between the aligned images , using a corrected @xmath49 , where @xmath50 are the best - fit shifts . for the majority of the examined 1-square - degree fields , the signal dominated the noise and we were able to measure the offsets to sub - pixel accuracy .
a plot of the longitude / latitude offsets between v2.0 and v1.0 and herschel hi - gal is shown in figure [ fig : offsets ] .
table [ tab : ccoffsets ] lists the measured offsets in arcseconds between images for all 1 degree fields from @xmath51 to @xmath52 .
the offsets represent the galactic longitude and latitude shifts in arcseconds from the reference ( left ) to the ` measured ' field ( right ) .
table [ tab : ccoffsetsmeans ] shows the means of the columns in table [ tab : ccoffsets ] , weighted by the error in the measurements and by the number of sources .
weighting by the number of sources is used for comparison with other works that attempt to measure the pointing offset on the basis of catalog source position offsets .
none of the measured offsets are significant ; in all cases the scatter exceeds the measured offset .
@xcite performed a comparison of the bolocam and atlasgal catalogs , identifying a systematic offset between the catalogs of @xmath53 , @xmath54 . because the offset is measured between catalog points , the meaning of this measured offset is not immediately clear . in the bgps maps in the atlasgal - bgps overlap region
, there were 12 individual sub - regions ( @xmath55 , with @xmath56 regions in the cmz ) that could have independendent pointing . because we did not have direct access to the atlasgal maps or catalog at the time of writing , we compared the bolocam v1.0and v2.0catalogs to each other determine whether the pointing changes in v2.0might account for the observed atlasgal offset , assuming the v2.0pointing is more accurate than the v1.0pointing
. we performed an inter - catalog match between v1.0and v2.0 , considering sources between the two catalogs to be a match if the distance between the centroid positions of the two sources is @xmath57 ( this distance is more conservative than that used in section [ sec : catalogmatching ] ) .
we then compared the pointing offset as measured by the mean offset between the catalogs to the offset measured via cross - correlation analysis of the maps on a per - square - degree basis .
the catalog and image offsets agree well , with no clear systematic offsets between the two estimators .
the scatter in the catalog - based measurements is much greater , which is expected since the source positions are subject to spatial scale recovery differences between the versions and because the sources include less signal than the complete maps .
there is no clear net offset between either version of the bgps and the herschel hi - gal survey , or between the two versions of the bgps .
however , the scatter in the pointing offsets between v1.0and herschel is substantially greater than the v2.0-herschel offsets .
the offset measured in @xcite is likely a result of particularly large offsets in a few fields with more identified sources . as shown in table [ tab : ccoffsetsmeans ] , the mean offset , _ weighted by number of sources _ , is greater for the atlasgal overlap region than overall .
we reproduce a number similar to the atlasgal - measured longitude offset of @xmath53 ( our source - count - weighted @xmath58 ) , despite a much larger standard deviation and despite no significant offset being measured directly in the images .
these measurements imply that the pointing offset measured by @xcite was localized to a few fields _ and _ that the offset is corrected in the v2.0data .
in order to determine the angular response of the bolocam array and bgps pipeline in realistic observing conditions , we performed simulations of a plausible synthetic astrophysical sky with synthetic atmospheric signal added to the bolometer timestream .
to generate the simulated atmosphere , we fit a piecewise power law to a power spectrum of a raw observed timestream ( figure [ fig : powerspecfit ] ) .
the power spectrum varies in amplitude depending on weather conditions and observation length , but the shape is generally well - represented by @xmath59 `` pink '' noise ( @xmath60 ) for @xmath61 hz and flat `` white '' noise ( @xmath62 ) for @xmath63 hz , where @xmath64 is the frequency .
we show a fitted timestream power spectrum in figure [ fig : powerspecfit ] .
the deviations from @xmath59 and white noise have little effect on the reduction process .
the fourier transform of the atmosphere timestream is generated by applying noise to the fitted power spectrum .
the power at each frequency is multiplied by a random number sampled from a gaussian distribution with width 1.2 , determined to be a reasonable match to the data , and mean 1.0 .
the resulting fourier - transformed timestream @xmath65 is @xmath66 , where @xmath67 and @xmath68 are the normally distributed random variables and @xmath69 is the fitted power - law power spectrum .
the atmosphere timestream is then created by inverse fourier transforming this signal .
gaussian noise is added to the atmospheric timestream of each bolometer independently , which renders the correlation between timestreams imperfect .
this decorrelation is important for the pca cleaning , which would remove all of the atmosphere with just one nulled component if the correlation was exact .
the noise level set in the individual timestreams determines the noise level in the output map .
we simulated the astrophysical sky by randomly sampling signal from an azimuthally symmetric 2d power - law distribution in fourier space .
the power distribution as a function of angular frequency is given by @xmath70 where @xmath71 is the angular size - scale and @xmath72 is the power - law spectral index for power spectra .
we modeled this signal using power spectrum power - law indices ranging from -3 to + 0.5 ; in the higal @xmath73 science demonstration phase ( sdp ) field , the power - law index measured from the 500 @xmath8 map is @xmath74 over the scales of interest for comparison with bolocam .
the data were smoothed with a model of the instrument psf to simulate the telescope s aperture and illumination pattern . for each power - law index ,
three realizations of the map using different random seeds were created .
the signal map was then sampled into timestreams with the bolocam array using a standard pair of perpendicular boustrophedonic scan patterns .
examples of one of these realizations with identical random seeds and different power laws are shown in figure [ fig : exp10gridin ] .
simulations performed with @xmath75 yielded no recovered astrophysical emission for normalizations in which the astrophysical sky was fainter than the atmosphere .
such a steep power spectrum is inconsistent with both bgps and other observations : as noted above , herschel sees structure with @xmath74 .
the fact that the bgps detected a great deal of astrophysical signal , none of which was brighter than the atmosphere , confirms that @xmath75 is unrealistic .
we used a subset of these power - law simulations to measure the amount of recovered signal at each angular ( spatial ) scale . for each power - law in the range @xmath76 , we used three different realizations of the map to measure the angular transfer function , defined as @xmath77 where @xmath78 is the angular frequency
, @xmath79 is the azimuthally averaged power - spectrum of the pipeline - processed map , and @xmath80 is the azimuthally averaged power - spectrum of the simulated input map .
the angular transfer function shows only weak dependence on the ratio of astrophysical to atmospheric power , and is approximately constant at @xmath81 recovery over the range of angular scales between the beam size and @xmath82 .
the angular transfer function is shown in figure [ fig : stf ] . at larger angular scales , in the range @xmath83 ,
the recovery is generally low ( @xmath84 ) .
our simulations included the full range of observed astrophysical to atmospheric flux density ratios , from @xmath85 for the central molecular zone ( cmz ) down to @xmath86 for sparsely populated regions in the @xmath87 region .
@xcite perform a similar analysis for the scuba-2 pipeline .
our transfer function ( figure [ fig : stf ] ) cuts off at a scale @xmath88 the scuba-2 scale .
while the angular extent of the bolocam footprint is only slightly smaller than scuba-2 s , some feature of the instrument or pipeline allows scuba-2 to recover larger angular scales .
we speculate that the much larger number of bolometers in the scuba-2 array allows the atmosphere to be more reliably separated from astrophysical and internal electrical signals ( bias and readout noise ) , so the scuba-2 pipeline is able to run with an atmosphere subtraction algorithm less aggressive than the 13-pca approach we adopted .
given an understanding of the angular transfer function , it is possible to compare the bgps to other surveys , e.g. hi - gal , atlasgal , and when it is complete , the jcmt galactic plane survey ( jps ) , for temperature and @xmath10 is the dust emissivity index , i.e. , a modification to a blackbody to create a greybody such that @xmath89 , and @xmath90 .
] measurements .
because of the severe degeneracies in temperature/@xmath10 derivation from dust seds ( e.g. * ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ) , we recommend a conservative approach when comparing bgps data with other data sets . for compact sources , aperture extraction with background subtraction in both the bgps and other data set should be effective .
section [ sec : apextract ] discusses aperture extraction in the presence of typical power - law distributed backgrounds . in order to compare extended structures , which includes any sources larger than the beam
, a different approach is required .
the safest approach is to `` unsharp mask '' ( high - pass - filter ) both the bgps and the other data set with a gaussian kernel with fwhm @xmath91 ( @xmath92 ) .
the filtering will limit the angular dynamic range , but will provide accurate results over the angular scales sampled .
direct comparison of power spectra over the reproduced range is also possible .
a demonstration of this approach is given in figure [ fig : higalpowerlaw ] , which shows the structure - rich @xmath93 field .
the bgps power spectrum has a shape very similar to that of higal .
the spectral index is a commonly used measure of the ratio between flux densities at two different wavelengths in the radio , @xmath94 the spectral index between the bgps at 1.1 mm and herschel at 500 @xmath8is @xmath95 over the range @xmath96 , although because the bgps angular transfer function is low at the large end of this range , this is only an ` eyeball ' estimate . on the rayleigh - jeans tail , @xmath97 ,
so this spectral index is consistent with typical dust emissivity index @xmath10 measurements in the range @xmath98 .
@xcite presented the bolocat catalog of sources extracted from the v1.0data with a watershed decomposition algorithm .
we have used the same algorithm to create a catalog from the v2.0catalog .
we have also performed comparisons between the v1.0and v2.0data based on the extracted sources .
the new catalog was derived using the same bolocat parameters as in v1.0 .
this catalog includes regions that were not part of the v1.0survey area , but we restrict our comparison between v1.0and v2.0to the area covered by both surveys .
one major change from the v1.0catalog is that the fluxes in the v2.0catalog are reported with background subtracted .
the backgrounds are calculated from the mode of the pixels in the range @xmath99 $ ] , where r represents the aperture radius ( 20 , 40 , or 60 ) .
the mode is computed using the idl astrolib routine ` skymod.pro ` , which returns the mean of the selected data if the mean @xmath100 is less than the median @xmath101 ( indicating low `` contamination '' from source flux ) or @xmath102 otherwise , then performs iterative rejection of bad pixels @xcite .
we performed aperture extraction on simulated data sets to determine what size apertures are appropriate when comparing to other data sets . in figure
[ fig : bgsub ] , we show the results of aperture extraction with and without background subtraction on a simulated power - law generated image with @xmath103 before and after pipeline processing .
the map has had point sources added to it randomly distributed throughout the field with flux densities randomly sampled from the range @xmath104 $ ] jy , and the power - law extended flux has an amplitude @xmath105 jy .
sources are extracted from the pipeline - processed map using bolocat , then the same source locations and masks are used to extract flux measurements from the input map .
figure [ fig : simptsrc ] shows the input , pipeline - processed , and point - source - only maps along with the bolocat apertures to give the reader a visual reference for an @xmath103 background with point sources .
the scatter between the flux density measurements derived from the input simulated sky map and the iteratively produced map is small when background subtraction is used ( the blue points ) , but large and unpredictable otherwise ( the red points ) .
the agreement between the flux densities extracted from the iterative map and the input synthetic map is excellent for 40 diameter background - subtracted apertures . for these apertures ,
the rms of the difference between the iterative map and the input map fluxes is @xmath106 jy when background subtraction is used , indicating the utility and necessity of this approach .
the agreement is similarly good for 80apertures ( @xmath107 jy ) , but the 120apertures exhibit a source- and background - brightness dependent bias , so we recommend against apertures that large when comparing to other data sets .
there are caveats to this analysis .
if the `` background '' power - law map has a peak flux density @xmath108 the peak point - source flux density , the point sources will not be recovered : the cataloging algorithm will pick out peaks in the power law flux distribution .
these can not be analyzed with simple aperture extraction for an @xmath103 flux density distribution .
however , for shallower power - law distributions , i.e. @xmath109 , aperture extraction effectively recovered accurate flux - densities in the processed maps - shallow power - law distributions more strongly resemble point - source - filled maps .
we matched the v1.0 and v2.0 catalogs based on source proximity .
for each source in v1.0 , we identified the nearest neighbor from v2.0 , and found that 5741 v2.0sources are the nearest neighbor for a v1.0source out of 8004 v2.0sources in the v1.0-v2.0overlap region .
similarly , we identified the nearest neighbor in v1.0for each v2.0source , finding 5745 v1.0sources are the nearest neighbor for a v2.0source out of 8358 v1.0sources .
there are 5538 v1.0-v2.0source pairs for which each member of the pair has the other as its nearest - neighbor .
these sources are clearly reliable and stable source identifications and constitute about 70% of the v2.0sample .
most of the unmatched sources have low flux density ( figure [ fig : v1vsv2matchhist ] ) , but some were significantly higher - these generally represent sources that were split or merged going from v1.0to v2.0 .
a few examples of how mismatches can happen are shown in figures [ fig : contourmatch ] and [ fig : contourmatch2 ] .
the low - flux - density sources were most commonly unmatched in regions where the noise in v1.0and v2.0disagreed significantly .
the high - flux - density mismatches tend to be different decompositions of bright sources and are preferentially found near very bright objects , e.g. in the galactic center region .
we reproduce parts of @xcite figures 17 and 19 as our own figures [ fig : bolocathistograms ] and [ fig : sizehistograms ] .
these figures show the distributions of extracted source properties ( flux density , size , and aspect ratio ) for the v1.0 and v2.0 data .
the source flux density distributions above the completeness cutoff are consistent between v1.0and v2.0 , both exhibiting power - law flux density distributions @xmath110 with values in the range @xmath111 for sources with @xmath112 jy . in the left panel of figure [ fig : bolocathistograms ] , we have included the v2.0aperture - extracted data both with and without annular background subtraction .
the v1.0catalog had no background subtraction performed because the backgrounds were thought to be negligible , but the v2.0catalog has had background subtraction performed so that the flux densities reported more accurately represent the sky .
the v2.0data include more large sources .
the longitude and latitude source flux density distribution plots , figure 15 of @xcite , are reproduced in figure [ fig : latlonhist ] .
the properties are generally well - matched , although even with the 1.5@xmath113 correction factor to the v1.0 data , there is more flux density per square degree in v2.0sources .
the gain in flux density recovery is both because of an increased flux density recovery on large angular scales and because of improved noise estimation , which results in a greater number of pixels being included in sources ( see section [ sec : catalogmatching ] for more details and figures [ fig : contourmatch ] and [ fig : contourmatch2 ] for examples ) .
a two - dimensional histogram providing a broad overview of the survey contents is shown in figure [ fig : hist2d ] .
the ratio of source counts per half square degree is included in panel 3 .
this figure illustrates that the two catalog versions are broadly consistent , and the regions in which they differ significantly tend to have fewer sources
. the most extreme ratios of v2.0to v1.0source counts tend to occur along field edges both because of preferentially low source counts and because the v2.0images have slightly greater extents in latitude than the v1.0 .
we presented version 2 of the bolocam galactic plane survey , which is a significant improvement over v1.0 in pointing and flux calibration accuracy .
the v2.0data show an improvement in large angular scale recovery .
the v2.0release includes new observations of regions in the outer galaxy . *
we have characterized the angular transfer function of the bolocam pipeline .
flux recovery is @xmath114 for scales between @xmath115 .
the angular transfer function shows a sharp drop in recovered power above @xmath116 scales .
* we compared the pointing of the bgps to that in hi - gal , and found that the surveys are consistent to within the measurement error @xmath117 .
* we measured the power spectral density in some regions and compared it to that in hi - gal , concluding that the power spectra are consistent with the normally used dust emissivity values in the range @xmath118 . *
a new version of the catalog has been released .
the improved quality of the v2.0images has some effects on the bgps catalog but the basic statistical properties of the catalog have not significantly changed . because of changing noise properties within the images , only 70% of the individual sources in v2.0have an obvious v1.0counterpart and vice versa .
the remaining 30% of sources do not have obvious counterparts because of two effects : 1 . at low significance ,
changing noise levels recover different features at marginal significance .
it is likely that low significance sources in v1.0and v2.0are _ both _ real features but have been rejected in the other catalog because of the relatively conservative limits placed on catalog membership .
2 . at high significance , the catalog algorithm is dividing up complex structure using the underlying watershed algorithm . in this case , the precise boundaries between objects are sensitive to the shape of the emission .
all of the high significance features appear in both catalogs , but the objects to which a piece of bright emission is assigned can vary . despite these changes in the catalogs ,
the overall statistical properties of the population show little variation except that the largest sources appear brighter and larger owing to better recovery of the large scale flux density .
* acknowledgements : * we thank the referee for a timely and constructive report . this work was supported by the national science foundation through nsf grant ast-1008577 .
the bgps project was supported in part by nsf grant ast-0708403 , and was performed at the caltech submillimeter observatory ( cso ) , supported by nsf grants ast-0540882 and ast-0838261 .
the cso was operated by caltech under contract from the nsf .
support for the development of bolocam was provided by nsf grants ast-9980846 and ast-0206158 .
er is supported by a discovery grant from nserc of canada .
nje and mm were supported by nsf grant ast-1109116 .
|
we present a re - reduction and expansion of the bolocam galactic plane survey , first presented by @xcite and @xcite .
the bgps is a 1.1 mm survey of dust emission in the northern galactic plane , covering longitudes @xmath0 and latitudes @xmath1 with a typical @xmath2 rms sensitivity of 30 - 100 mjy in a @xmath3 beam .
version 2 of the survey includes an additional @xmath4 square degrees of coverage in the 3rd and 4th quadrants and @xmath5 square degrees in the 1st quadrant .
the new data release has improved angular recovery , with complete recovery out to @xmath6 and partial recovery to @xmath7 , and reduced negative bowls around bright sources resulting from the atmospheric subtraction process .
we resolve the factor of 1.5 flux calibration offset between the v1.0 data release and other data sets and determine that there is no offset between v2.0and other data sets .
the v2.0pointing accuracy is tested against other surveys and demonstrated to be accurate and an improvement over v1.0 .
we present simulations and tests of the pipeline and its properties , including measurements of the pipeline s angular transfer function .
the bolocat cataloging tool was used to extract a new catalog , which includes 8594 sources , with 591 in the expanded regions .
we have demonstrated that the bolocat 40and 80apertures are accurate even in the presence of strong extended background emission .
the number of sources is lower than in v1.0 , but the amount of flux and area included in identified sources is larger .
|
it is often assumed that strong correlation is important for an understanding of the high - t@xmath1 problem @xcite . the failure of the local approximations to the exchange - correlation functional , such as the local ( spin ) density approximation , lda or lsda @xcite to produce the anti - ferro magnetic ( afm ) insulating state of many undoped transition metal oxides
is generally quoted as the essential reason for discarding dft calculations for high - t@xmath1 copper oxides @xcite .
dft is essentially exact , but the approximations to make it practical for use in real applications seem inappropriate for oxides .
however , despite this problem there are several lda results that fit to the observed high - t@xmath1 properties , provided that doping and supercells are considered in order to account for imperfect lattice conditions such as stripes , phonons or spin waves @xcite . here
is presented a short review of those dft results together with a discussion of the correlation problem and an attempt to include further corrections to lsda .
estimations of couplings @xmath2 caused by spin - phonon coupling ( spc ) and pure spin fluctuations are also made .
the spin polarized potential from an afm order ( like that of undoped la@xmath3cuo@xmath4 ) can be generated from @xmath5 , where @xmath6 , is at the brillouin zone ( bz ) boundary , so that there is one cu site with spin up and one with spin down within a distance @xmath7 ( @xmath7 is the lattice constant ) .
the potential of the other spin has a phase shift of @xmath8 .
the band dispersion from a 1-dimensional ( 1d ) nearly free - electron model ( nfe ) is obtained from a 2x2 eigenvalue problem , and it has a gap of 2@xmath9 at the zone boundary .
suppose now that an additional potential modulation @xmath10 exists , and that @xmath11 , so that its periodicity or `` wavelength '' is much larger than 2@xmath7 .
the product of these two modulations can be written @xmath12 , where the two amplitudes are combined into one coefficient , @xmath13 .
1 shows the real space configurations along the 100-direction .
( broken line ) , and the unit cell contains 2 cu sites along @xmath14 .
the second row shows the modulation @xmath15 ( here @xmath16 ) and the last row is the product @xmath17 with the new spin configurations on the cu .
the striped unit cell contains 8 cu sites .
, width=302,height=226 ] this potential describes afm modulations in stripes , with magnetic nodes ( cu with zero moment ) separated by @xmath18 .
the values of @xmath13 for phonons or spin waves of different lengths are not known from the nfe model , but some values will be fed in from the ab - initio linear muffin - tin orbital ( lmto , @xcite ) calculations , as will be described below .
the lmto calculations are slow for large supercells .
the cells need to be large to cover the periodicity of realistic phonon and/or spin - waves , and so far it has been possible to extend the cells in one direction only , usually the cuo bond direction .
in contrast , the nfe model is very simple and it can easily be extended to two dimensions ( 2d ) . the band dispersion for k - points @xmath19
is now obtained from the eigenvalues of a 3x3 matrix @xcite .
not only stripe order , but also `` checkerboard '' configuration can be modeled .
( in mry ) for @xmath20
. , width=302,height=226 ] examples of the density - of - states ( dos ) for 3 different @xmath13 are shown in figure 2 .
the potential modulations are parallel to the cuo bond directions along @xmath14 and @xmath21 , so the gaps appear on the @xmath22 and @xmath23 lines , as shown in figure 3 .
however , almost no effect from @xmath13 appears on the band dispersion in the diagonal direction ( along @xmath24 ) .
therefore , the fermi surface ( fs ) is not affected along the diagonal , but fluctuations , described through amplitude variations of @xmath13 for different positions and different time , will make the fs smeared in the two bond directions .
an example of this is shown in figure 4 .
by 1d ( 1-dimensional ) lmto we mean ab - initio lmto calculations for long ( and narrow ) supercells most often oriented along the cuo bond direction @xcite .
these calculations are based on the local version of dft , the local ( spin ) density approximation , lda or lsda @xcite .
cells with phonon distortions and/or spin waves within lengths of 4,8 and 12 lattice constants are typically considered in these calculations .
the wavelengths of spin waves are twice as long as those of phonons .
hole doping , @xmath25 , in la@xmath26ba@xmath27cuo@xmath4 ( lbco ) is generally modeled by the virtual crystal approximation ( vca ) where the nuclear and electronic la - charges ( 57.0 ) are reduced to ( 57-@xmath28 ) to account for a perfectly delocalized doping ( @xmath25 in holes per cu ) .
but some calculations with real la / ba substitutions show that improved superconducting properties can be expected from periodic doping distributions @xcite . the maximal phonon distortion , @xmath29 , for different sites , @xmath30 , and afm magnetic moments on cu , @xmath31 , depend on temperature , @xmath32 , force constants and spin stiffness .
appropriate values of @xmath33 and @xmath34 at @xmath35k are deduced from experiments and calculations @xcite . and
@xmath24 for @xmath36 and @xmath37 20 mry ( full line ) .
the fe band without potential perturbations ( @xmath370 ) is shown by the broken line .
the thin horizontal line is at the dos minimum , which is situated at 0.15 holes / cell .
, width=302,height=226 ] .
the fs is well defined in the diagonal direction , while the fluctuations makes it diffuse in the @xmath38 and @xmath39-directions.,width=313,height=238 ] a striking result of these calculations ( also made for hgba@xmath3cuo@xmath4 ) is that a gap ( or a pseudogap ) will open up in the dos , very similar to what is found in the nfe models .
the gap will appear near to @xmath40 for the undoped material if the wavelengths ( @xmath41 ) are long , while for short phonons or spin waves the gap moves to lower energy @xcite .
in summary , @xmath42 , where @xmath41 is in units of @xmath7 and @xmath25 is the number of holes per cu .
a second important result is that the spin waves will be stronger , and the pseudogaps will be deeper , if the waves coexist with phonon distortions of the correct wave length and phase @xcite .
interactions between phonons and spin waves have also been suggested in order to explain neutron scattering @xcite .
phonons like the `` half - breathing '' o - mode and modes with z - displacements of la and apical oxygens are most effective in this process of spc .
this mechanism offers direct explanations of phonon softening , q - dependent spin excitations and various isotope effects @xcite .
the rather few lmto results permit to establish only a few values of @xmath13 for different @xmath43,@xmath30 of the phonons and spin waves .
a general trend of larger @xmath13 for long wave lengths seems clear , both for phonons and for spin waves . a procedure based on the partial character of the states above and below
the gap in the nfe model confirms this trend for spin waves , and it permits to fill in some of the q - dependent points not calculated by lmto . in all
this gives some confidence in the @xmath44-variations of @xmath13 .
however , one can expect more vivid variations of the spin part near @xmath45 , because lda underestimates the transition towards afm . when these @xmath13-values are fed into the 2d nfe model it gives an approximately linear variation of @xmath43 as function of @xmath25 , up to a saturation near @xmath46 for @xmath25 larger than @xmath47 @xcite .
hence , the linear dependence for low @xmath25 is in qualitative agreement with the result from 1d - lmto , but the pace is different .
this is probably because the gap opens @xmath48 along @xmath38 and @xmath39 in the 2d nfe model , which makes the progression of the gap a bit slower than in 1d .
the saturation is because the @xmath13 s decrease for increasing @xmath43 and it becomes impossible to open a gap at the low energy where @xmath40 should be for large @xmath25 .
however , if @xmath49 and @xmath50 are assumed to be different in the 2d nfe model it is possible to follow one gap towards lower energy . a second weaker gap moves to higher energy .
the emerging picture from these band results is that moderately strong spin fluctuations , which exist for many combinations of @xmath51 and @xmath52 , will be enhanced through spc to some particular phonon distortions .
the selection depends on the atomic character of the phonon mode as well as on doping and wave length ( i.e. on @xmath51 ) .
the latter is because the pseudogap appears at @xmath40 so that a maximum amount of kinetic energy can be gained from the spc mode .
this is for low t when the states below the gap is of one spin and well separated from the ( unoccupied ) state of the other spin above the gap .
but for increasing t there will be mixed occupations of the states around @xmath40 through the fermi - dirac function ( and through thermal disorder ) , which will decrease the spin density . this will decrease the spin polarization of the potential , which in turn will decrease the spin density even more , and at some temperature t@xmath53 the support of the pseudo gap from the spin wave will collapse @xcite . the fact that @xmath13 is largest at low doping favors larger t@xmath53 when @xmath54 .
a high dos at @xmath40 is important for a high superconducting @xmath0 , and therefore is the pseudogap in competition with superconductivity in under - doped cuprates @xcite .
it was suggested that non - adiabatic electron - phonon coupling could be enhanced in the cuprates , because of the low fermi velocity in the z - direction @xcite .
this velocity is comparable to the vibrational velocity and the electronic screening appeared to be insufficient .
however , the screening can be made by other electrons moving within the planes .
moreover , phonon frequencies calculated within lda without assumptions of incomplete screening , agree satisfactory with experimental frequencies @xcite .
instead , excitations of virtual phonons coupled to spin waves can be important for the mechanism of superconductivity , but spc makes the separation between pure electron - phonon coupling , @xmath55 , and @xmath2 caused by spin fluctuations , @xmath56 , less clear .
the important observation is that atomic phonon distortions will trigger enhancement of spin waves .
if so , the possibility for larger @xmath55 is open , because instead for the common approach to ignore spin effects in electron - phonon coupling , there are larger matrix elements when the spin polarized part of the potential is involved .
this is readily imagined for a system which is nonmagnetic when phonon distortions are absent , but magnetic when the distortions are present .
an estimation of pure @xmath55 and the coupling parameter for spc , @xmath57 , in lbco has been presented earlier @xcite . with a total @xmath58 , and distortion amplitudes and potential shifts as in ref .
@xcite this leads to @xmath59 of the order 0.36 and 0.6 for pure phonons and spc , respectively @xcite .
these values appear sufficiently large for a large @xmath0 in simple bcs - type formulations , but the precise relation for @xmath0 depends also on other parameters @xcite .
.[table1 ] total energy ( @xmath60 , mry per cu ) , matrix element for spin fluctuations ( @xmath61 in mry ) , @xmath56 and local stoner enhancement , @xmath62 , calculated for lattices of lbco with two phonon distortions and without distortions . [ cols="<,^,^,^,^,^",options="header " , ] in preliminary calculations for undoped lbco we do a rescaling of the exchange due to kinetic energy and of correlation due to second gradients .
the comparison is made with standard lsda for an afm unit cell where a staggered magnetic field is applied to the cu sites , and it is found that a field of @xmath63 8.8 @xmath64 is the limit for having a zero gap @xmath65 .
the magnetic moment is then @xmath66 per cu .
the amplitude of the required magnetic field to obtain a zero gap is reduced when the calculations include the correction factors , see table ii .
the absolute values can depend on details of the band calculations @xcite , but the trend towards a stability of an afm state is clear .
corrections of this type will be interesting for seeing enhanced spin fluctuations for long wave length spin waves in doped cuprates .
further enhancements of @xmath56 can also be expected @xcite .
potential corrections must ultimately be tested for other types of materials in order to verify that some properties will not deteriorate . here for the cuprates the corrections permit to proceed in the modeling of more properties from the band results .
results of band calculations for supercells with frozen phonons and spin waves suggest that a fermi - liquid state can cause pseudo gaps and dynamic stripes .
together with a nfe model it is possible to describe the doping dependence of many normal state properties of the cuprates .
the coupling between phonon distortions and spin fluctuations seems to be crucial for the mechanism of superconductivity , so that the spin - polarized part of @xmath2 is most enhanced by simultaneous excitations of phonons and spin waves .
two different mechanisms for superconductivity mediated by spin fluctuations are possible .
the largest coupling parameter is when a phonon is excited together with the spin fluctuation .
lower couplings , @xmath56 at larger energies , are independent of the phonon excitation , but these spin fluctuations can nevertheless profit from possible phonon distortions of the lattice . still , absolute numbers are too small when using lda , as is also concluded from the absence of afm stability in lda calculations for undoped systems .
however , it is argued that corrections due to higher order density gradients and kinetic energy are able to bring the band calculations closer to afm .
this is shown in calculations for undoped lbco by yet very approximate corrections due to non locality and kinetic energy .
refinements of such corrections will be of interest for application to supercells , including doping , phonon distortions and spin waves , since it can be expected that realistic @xmath59 will be obtained .
j. kbler , phys .
a*81 * , 81 ( 1981 ) ; c.s .
wang , b.m .
klein , and h. krakauer , phys .
lett . * 54 * , 1852 ( 1985 ) ; b. barbiellini , e.g. moroni and t. jarlborg , j. phys .
matter * 2 * , 7597 , ( 1990 ) .
|
first principle band calculations based on local versions of density functional theory ( dft ) , together with results from nearly free - electron models , can describe many typical but unusual properties of the high-@xmath0 copper oxides .
the methods and a few of the most important results are reviewed .
some additional calculations are presented and the problems with the commonly used approximate versions of dft for oxides are discussed with a few ideas for corrections . it is concluded that rather modest corrections to the approximate dft , without particular assumptions about strong correlation , can push the ground state towards anti - ferro magnetic ( afm ) order .
spin fluctuations interacting with phonons are crucial for the mechanism of superconductivity in this scenario .
|
theoretical investigation of electromagnetic @xmath1 pair production in relativistic heavy - ion collisions goes back to the paper @xcite where the born cross section of the process at high energy was calculated in the leading logarithmic approximation .
racah , in his remarkable paper @xcite , has calculated the high - energy asymptotics of the born cross section up to power - suppressed terms in @xmath2 ( @xmath3 is a lorentz factor of the colliding nuclei ) .
recently there was a certain rise of the interest to this process connected with the functioning of heavy ion colliders , like rhic and lhc , see ref .
in particular , a great attention has been paid to the investigation of the coulomb corrections to the cross section at high energies @xcite .
speaking of the total born cross section , the problem of its calculation is of a three - loop complexity level .
probably , this is the main reason why this quantity was not calculated exactly at arbitrary velocities of the colliding nuclei .
this is in contrast to the born cross section of pair photoproduction in the field of an ion , where the total born cross section is known exactly for any energy of the initial photon since refs .
now that we have an essential progress in the multiloop calculations , we are in position to fill this gap and to calculate the total born cross section of @xmath1 pair production in relativistic ion collisions .
the consideration of the present paper is based on the following approach . using the optical theorem we express the total cross section via the sum of cut three - loop integrals .
then we apply the standard approach to multiloop calculations , based on the ibp reduction and differential equations for master integrals .
the differential equations for the master integrals are first reduced to @xmath4-form @xcite using the algorithm of ref .
@xcite , and then solved recursively up to the required order in @xmath4 .
thus , we obtain the total born cross section exactly in the relative velocity @xmath5 of the colliding nuclei .
our result perfectly agrees with the celebrated result of racah @xcite in the limit of large relativistic factor . at small @xmath5
we compare our result with estimate obtained in the recent paper @xcite and find a complete disagreement . in order to find the origin of the disagreement
, we perform a straightforward calculation of the low - energy asymptotics of the cross section differential with respect to the electron and positron momenta .
the direct integration then reproduces our result obtained with the help of the differential equations .
using optical theorem , the total cross section of the process @xmath7 can be written as @xmath8 where @xmath9 is given by the sum of two cut diagrams depicted in fig .
[ fig : cut_diagrams ] , @xmath5 is the relative velocity of the colliding nuclei , and @xmath10^{-\frac12}$ ] is the lorentz factor .
pair production in the collisions of relativistic nuclei .
cut thin line denotes the cut propagator @xmath11 of the electron , cut double line denotes the cut propagator @xmath12 of a heavy particle , interaction vertex with the heavy particle is @xmath13 ( @xmath14 is a four - velocity of the heavy particle ) . ] contribution of both diagrams can be expresses in terms of the scalar integrals @xmath15 here @xmath16 and @xmath17 are the four - velocities of the nuclei , so that @xmath18 .
we proceed in the following way .
first , we perform the ibp reduction of the cut integrals from the above topologies in @xmath19 . for this step
we use ` litered ` , refs .
we end up with 8 master integrals @xmath20 introducing the column - vector @xmath21 , we obtain the differential system @xmath22 where @xmath23 is a matrix with entries being rational functions of both @xmath3 and @xmath4 . passing to new variable , @xmath24
, we apply the algorithm from ref .
@xcite to reduce the differential system to @xmath4-form @xcite .
the differential system for the new basis @xmath25 has the form @xmath26\widetilde{\mathbf{j}}\,,\ ] ] @xmath27}\,,\\ m_1&={\operatorname{diag}(2 , 0 , 2 , 2 , -6 , 0 , 2 , 0)}\,,\\ m_2&={\operatorname{diag}(0 , 0 , 0 , 0 , 0 , 0 , 0 , -2)}\,.\end{aligned}\ ] ] we obtain @xmath4-expansion of @xmath28 term - by - term using the formula @xmath29\widetilde{\mathbf{j}}^{(n)}+\text{const}\ ] ] and fixing the constant from small-@xmath5 asymptotics . in order to calculate this asymptotics ,
we use the method of expansion by regions @xcite .
the only nontrivial boundary conditions come from @xmath30 term in small-@xmath5 asymptotics of @xmath31 and @xmath32 : @xmath33 as a result , the @xmath4-expansions of both @xmath34 and @xmath35 is expressed in terms of hpls .
the expansions of @xmath35 are presented in ancillary file . plugging the obtained expansions in the cross section expressed via @xmath34 we observe the cancellation of the terms @xmath36 with @xmath37 .
the @xmath38 term gives us the result @xmath39 @xmath40 given the expression , it is easy to calculate both high - energy and low - energy asymptotics of the total cross section .
for @xmath41 we have @xmath42 + \ldots\bigg\}\,,\end{gathered}\ ] ] where @xmath43 .
the first line of eq .
is the celebrated racah s result @xcite , and the second line is the first correction to it .
it is interesting to note that the correction is amplified by the fourth power of @xmath44 . for @xmath45
there is a strong compensation between separate terms in eq . , which leads to @xmath46 suppression of the cross section .
we have @xmath47 recently , the small-@xmath5 asymptotics of the total cross section was discussed in ref .
the estimate @xmath48 given there is in clear contradiction with our result .
in fact , the estimate @xmath49 can be justified in the following way . using the kinematic constraints @xmath50 for momentum transfers @xmath51 ,
it is easy to understand that the main contribution to the cross section is given by the region where @xmath52 the characteristic momenta of the produced particles are of the order of their mass . using these estimates
, it is easy to count powers of @xmath5 in the total cross section .
we have @xmath53 from @xmath54 , @xmath55 from the photon propagators , @xmath56 from the denominator of electron propagator , and @xmath57 from the flux of the colliding particles . as to the numerator of the electron propagator
, one might check that it does not give @xmath57 factor in the sum of two diagrams due to the estimate @xmath58 . using these estimates
, we have derived the differential cross section at @xmath59 and then obtained the leading term of by the direct integration .
to avoid cluttering , we refrain from presenting fully differential cross section here .
we only present the cross section , differential with respect to the energies of the produced particles : @xmath60\,.\end{gathered}\ ] ] here @xmath61 and @xmath59 is the relative velocity of the nuclei . integrating this cross section over @xmath62 , we obtain the leading term of eq . .
it is interesting to compare our result with the leading high - energy ( racah ) and low - energy asymptotics @xmath63 .
these asymptotics are given by the first line of eq . and the leading term of eq .
, respectively .
[ fig : cross_section ] demonstrates this comparison .
one can see that both low- and high - energy asymptotics essentially depart from the exact result in the region @xmath64 .
, eq . , in units of @xmath66 as a function of @xmath67 ( solid curve )
. dashed and dotted curves correspond to high- and low - energy asymptotics , respectively . right : relative error @xmath68 of the high- and low - energy asymptotics . ]
our approach based on the ibp reduction and calculation of the master integrals allows us , without additional efforts , to calculate the total born cross section of the production of a pair of point - like scalar charged particles .
we have @xmath69 the high- and the low - energy asymptotics of this cross section have the form @xmath70 the leading term @xmath71 in the high - energy asymptotics agrees with the result obtained within the equivalent photon approximation .
curiously , the low - energy asymptotics of the cross section for scalar particles scales differently ( @xmath72 ) than that for spinor particles .
inspection of the contributions of separate diagrams shows that in this limit only the contribution of seagull diagram survive .
it is interesting to discuss the applicability region of our results .
of course , by using the born approximation we assume that @xmath73 . in principle , at small @xmath5 one may expect higher - order corrections of the relative order @xmath74 .
however , since at small @xmath5 the velocities of the produced particles are not small ( see , e.g. , their spectrum ) , we would guess that such corrections are forbidden .
one may also expect corrections of the order @xmath75 due to additional coulomb exchanges between the nuclei , but they seem to be accompanied by the factor @xmath76 , where @xmath77 are the masses of the nuclei . in fact a stronger condition @xmath78 is definitely required because otherwise the limit @xmath79 is no longer valid .
nevertheless , we must admit that the determination of the correct magnitude of the higher - order effects at small @xmath5 requires a separate examination .
we are grateful to a.i .
milstein for the interest to our work and useful discussions .
this work has been supported by russian science foundation ( project no .
14 - 50 - 00080 )
. partial support of rfbr through grant no .
15 - 02 - 07893 is also acknowledged .
a. baltz , g. baur , d. denterria , l. frankfurt , f. gelis , v. guzey , k. hencken , y. kharlov , m. klasen , s. klein , v. nikulin , j. nystrand , i. pshenichnov , s. sadovsky , e. scapparone , j. seger , m. strikman , m. tverskoy , r. vogt , s. white , u. wiedemann , p. yepes , m. zhalov , physics reports 458 ( 13 ) ( 2008 ) 1 171 .
|
we apply the differential equation method to the calculation of the total born cross section of the process @xmath0 .
we obtain explicit expression for the cross section exact in the relative velocity of the nuclei .
|
the detectors having low background and low threshold are required for investigation of rare processes involving low energy neutrino and weak interacting particles .
such detectors can be used effectively for search for the dark matter , for measurement of neutrino magnetic moment and coherent neutrino scattering by nuclei and for investigation of solar neutrino problem .
for above investigations one needs low background detector of mass several @xmath0 and with threshold less than 1000 @xmath1 .
the cryogenic and germanium detectors partly correspond to these requirements .
the drawback of the cryogenic detectors is the complexity of their production and use .
the drawback of germanium detectors is a rather high threshold @xmath2 @xmath3 which is due to a leakage current and electronic and microphonic noises .
it would be very attractive to provide effective decreasing of the detector threshold using the internal proportional amplification of signal .
internal proportional amplification in the semiconductor detectors is realized now in the silicon avalanche photodiodes ( apd ) @xmath4 , where a gain of about @xmath5 is achieved by avalanche multiplication of electrons at electric field @xmath6 @xmath7 in narrow @xmath8 junction with sensitive volume several @xmath9 .
below we shall demonstrate the possibility of realizing the germanium detector with internal amplification ( gda ) and present its design .
in semiconductor detectors ( as in a gas proportional counter - pc , or multiwire proportional chamber - mwpc ) the conditions for internal proportional amplification of electrons can be fulfilled .
it is known that in apd critical electric field @xmath10 , that provides multiplication of electrons at room temperature , is equal @xmath11 @xmath7 .
the @xmath10 for germanium at liquid nitrogen temperature can be defined from the dependence of electron drift velocity on electric field and energy of production of electron - hole pairs and photons @xmath12 .
for germanium at 77k the magnitude of @xmath10 derived in this fashion is equal @xmath13 @xmath7 . in apd and gas pc the critical electric field
is produced by a different way . in the first case @xmath10
is achieved by high concentration of impurities in narrow junction . as a result
the sensitive volume of apd is several @xmath14 . in gas pc
the @xmath10 can be achieved by special configuration of the electric field , due to large difference of cathode and anode sizes . in high purity germanium ( hpge ) with a sensitive volume near 100 @xmath15 the @xmath10 can be obtained by the same manner .
electric field in the gas cylindrical pc is of the form : e(r)= where @xmath16 is applied voltage , @xmath17 and @xmath18 are the radii of cathode and anode correspondingly , @xmath19 is a distance from anode .
one can see from ( 1 ) that at @xmath20 @xmath16 , @xmath21 @xmath22 and @xmath23 @xmath22 @xmath24 is about @xmath25 @xmath7 near the anode . unlike gas pc the electric field in the coaxial detector from high purity germanium
is defined not only by @xmath16 , @xmath17 and @xmath18 but the concentration of donor ( n - type ) or acceptor ( p - type ) impurities as well .
the magnitude of the volume charge in the sensitive volume of crystal depends on these impurities .
the electric field in coaxial detector from hpge with regards to the impurities will be @xmath26 : e(r)= r - , where @xmath27 is impurity concentration , @xmath28 is the electron charge,@xmath29 is the dielectric constant of germanium . the electric field ( 2 ) can be expressed with a depletion voltage @xmath30 , which is a minimum applied voltage necessary to neutralize the volume charge and to provide the sensitive region in the whole crystal volume .
the @xmath31 for the coaxial detector , taking into account that @xmath32 , may be given by : v_d - r_2 ^ 2 . and the equation ( 2 )
may be rewritten for n - type germanium in the final form : e(r)= - r - . the dependence @xmath33 on @xmath19 is shown on fig.1 for coaxial detector from hpge .
electric field near the anode reaches @xmath10 - magnitude required for avalanche multiplication of electrons .
the coaxial germanium detector with internal amplification is the more appropriate for the low background spectrometers but the possibility of fabrication of inner electrode of 20 micron radius is highly conjectural presently .
so we shall consider the more realistic problem - fabrication of planar germanium detector with internal amplification , multistrip planar germanium detector , similar in design to mwpc .
the electric field in mwpc is of the form for one dimension case ( the coordinates x and y relate to an centered on the wire , x is parrelel to the wire plane , y is perpendicular ) : e(0,y)= where @xmath16 is applied voltage , @xmath34 is wire spacing , @xmath35 is a diameter of the wire and @xmath36 is the thickness of the planar detector . as in the case of coaxial germanium detector
one must take into account for multistrip germanium detector the depletion voltage @xmath30 . in the case of planar germanium detector @xmath37 .
the electric field for multistrip germanium detector is of the form : e(0,y ) = -y - where @xmath35 is the strip width .
the dependence of @xmath38 on @xmath39 for multistrip germanium detector is shown on fig.2 . in the cases
being considered the electric field near the anode is sufficient for avalanche multiplication of electrons ( @xmath40 @xmath7 ) .
the amplification factor can be estimated as @xmath41 where @xmath42 is a free electron path for inelastic scattering and @xmath43 is a length of avalanche region where @xmath44 .
the @xmath42 in germanium at 77@xmath45 is equal 0.5 micron and for @xmath46 @xmath22 @xmath43 is equal 5 micron ( see fig.2 ) so for this case it is possible to achieve @xmath47 . if one does not need high amplification factor it is possible to decrease @xmath16 or to increase the strip width @xmath34 .
the gda is assumed to use for investigation of rare processes so the spectrometer including gda must have large mass of detector , it can be fabricated from separate modules of mass about 0.7 @xmath48 each .
one module represents the multistrip planar germanium detector from hpge having impurity concentration less than @xmath49 @xmath50 , measures @xmath51 @xmath52 ( see fig.3 ) .
the 12 anode strips of 20 micron width and of 65 @xmath53 length are fabricated by photomask method @xmath54 .
the cathode area is equal @xmath55 @xmath56 and the fiducial volume is equal to 130 @xmath15 .
there are the guard electrodes in the anode and cathode planes .
the anode strips can be connected together however it is more convenient to take signals from separate strips to suppress the background compton gamma - quanta . for the fabrication of gda
it is necessary to use the germanium crystals of uniform distribution of impurities to provide homogenous electric field near the anode .
second in importance it is the providing small depth and width of junction layer under the strips so the electric field near the strips is defined by junction dimensions .
the design of gda must gurantee the reliable cooling of crystal since the critical electric field and amplification factor depend on free path of charge carriers which in its turn depends on temperature .
the energy resolution of semiconductor detector will be given by e= where @xmath57 is the intrinsic energy resolution of detector which is defined by statistical fluctuation in the number of charge carriers created in detector sensitive volume , @xmath58 is the energy resolution which is defined by associated electronics . in the case of gda these two terms will be of form e_int=2.34 and e_el= where @xmath59 is the energy to create one pair of charge carriers , @xmath60 is the energy deposited in the detector , @xmath61 is the fano factor , @xmath62 is the excess noise factor due to the fluctuation of multiplication , @xmath45 is the amplification factor , @xmath28 is the electron charge , @xmath63 is the absolute temperature of the resistors , @xmath64 is the total capacitance presented at the input of preamplifier , @xmath65 is the time constant of the rc circuits of the preamplifier , @xmath66 is the steepness of the field effect transistor characteristic , @xmath67 is the resistance at the input of the preamplifier , @xmath68 is the detector leakage surface current and @xmath69 is the detector leakage balk current due to the thermal generation of charge carriers .
according to the calculation for gda with k@xmath7010 one must take into account in the formula for @xmath58 only the last term due to the bulk leakage current and the formula ( 7 ) for gda can be rewritten as e2.36k where @xmath59 and @xmath60 is in ev , @xmath69 is in na and @xmath65 is in @xmath71 .
the gda energy threshold is defined by @xmath72 or more exactly by the last term in ( 10 ) : e_th 2.36 one can estimate the @xmath73 for microstrip planar detector from hpge of volume 100 @xmath15 with internal amplification : at n=@xmath49@xmath50 , @xmath69=0.01 na per one strip , @xmath65=0.5 @xmath71 and @xmath62=0.5 : @xmath74 12 ev .
the dependance of relative energy resolution @xmath75e / e on energy in the more interesting energy range @xmath76 @xmath1 for gda is shown in table .
the dependance of relative energy resolution @xmath75e / e on energy [ cols="^,^,^,^,^,^,^,^",options="header " , ] + the internal amplification of gda degrades the performance somewhat , but such energy resolution of gda is adequate to the investigation of above problems .
it is interesting to note that common planar detector from hpge of volume 100 @xmath15 produced by `` canberra '' ( type gl3825r ) has relative energy resolution about 8@xmath77 at e=5900 ev .
presently germanium detectors are in considerable use in low background measurements for high purity of germanium crystals : their content of radioactive impurity does not exceed @xmath78 @xmath79 .
if the internal amplification is realized in germanium detectors and their threshold is lowered to several @xmath1 the possibility for their use will considerably increase .
firstly , the lowering of the detector threshold makes it possible to extend the kinematical domain of investigations , secondly , it enables in some cases significantly to decrease the level of background . following are brief discussions of prospects of gda using for neutrino magnetic moment ( @xmath80 ) measurement , for investigation of neutrino coherent scattering by nuclei and for search for `` light '' wimps .
a laboratory measurement of the nmm is based on its contribution to the ( anti)neutrino - electron scattering . for a non - zero nmm
the differential over the kinetic energy @xmath81 of the recoil electron cross section @xmath82 is given by the sum of the weak interaction cross section and the electromagnetic one . at a small recoil energy
the weak part practically constant , while the electromagnetic one grows as @xmath83 towards low energies . for improving the sensitivity to @xmath80
it is necessary to lower the threshold for detecting the recoil electrons as far as the background conditions allow .
now there is number of the projects dedicated to the nmm measurements with detectors of mass @xmath84 kg and with thresholds of @xmath85 kev .
authors of proposal @xmath86 are going to use germanium detector with mass of 2 kg and threshold of 3 kev to achieve for two years reactor measurement the limit on @xmath87 , with @xmath88 being the bhorh magneton .
the use of gda with internal amplification in this experiment would provide possibility to achieve limit on @xmath80 about 2@xmath89 .
however , one can not significantly low the gda threshold in the reactor experiment for two reasons .
firstly , one must take into account atomic binding effect for electrons
. the total @xmath90 cross section will be 30@xmath77 less due to the binding effects ( in the energy range @xmath91 ev ) .
the secondly at t@xmath92500 ev the effect neutrino coherent scattering ( ncs ) by germanium nuclei will surpass the effect of @xmath93 scatttering and ncs will present physical background relative to the @xmath93 scattering . nevertheless the low gda threshold can be used in full measure if one uses artifical neutrino source ( ans ) instead of reactor .
nowadays there are several proposals on development ans with activity @xmath94 mci @xmath95 .
although the development and construction of the ans is rather expensive and challenging problem the use of ans offers some advantages over the reactor based experiments : presently the more appropriate ans for nmm measurement is a tritium source proposed in @xmath98 for measurement of nmm by means of low threshold cryogenic silicon detector .
the end - point energy of tritium beta spectra is 18,6 kev and for maximum recoil electron energy a kinematical limitation gives : t_max = = 1260 ev . if one uses tritium ans with activity 40 mci then the neutrino density flux will be 6@xmath99 1/@xmath97 in the detector position .
the limit on nmm about 3@xmath100 can be achieved during one year measurement time with silicon detector mass of 3 kg in the energy range @xmath101 ev .
the very same limit on nmm can be achieved by use the gda with mass of 3 kg during one year of measurement time in the energy range @xmath102 ev . in the gda case
the count rate due to electromagnetic neutrino interaction will be 0.24 event / d ( at @xmath103 ) , the count rate due to weak interaction will be one order of magnitude less .
now let us consider the background count rate .
if one takes into account that @xmath104 then the value of @xmath105 in the energy range @xmath106 kev(@xmath107 ) and @xmath102 ev(@xmath108 ) will be the very same .
the best level of background for germanium detector was achieved in @xmath109 and it is equal 0.1 event / kev@xmath110kg@xmath110d . if one assume that level of background does not depend on energy in the energy range of interest then the level of background in the second energy interval will be two order of magnitude less since @xmath111 and will be 0.15 event / d .
in addition in the case of tritium ans there is no physical background due to the ncs since in this case the maximum recoil energy of germanium nuclei @xmath112 is equal 0.01 ev . the use of gda can open up the possibility to investigate the coherent neutrino scattering by nuclei .
this interaction is of great importance into interstellar processes and up to now did not observed in laboratory because the very low kinetic energy transferred to nucleus in the process of neutrino - nucleus scattering .
for germanium nucleus and reactor antineutrino spectrum the maximum kinetic recoil energy is about @xmath112=2500 ev and taking into account quenching effect only one third of this energy are imparted on ionization @xmath113 .
so one needs germanium detector with threshold significantly lower 800 ev for investigation of ncs by germanium nuclei .
such low threshold can be provided by gda having internal amplification about 10@xmath114 .
the differential cross section of ncs by nuclei @xmath115 can be expressed by : n^2m(1- ) ( ) ^2 z^2 ( ) where m , n and z are the mass , neutron and charge numbers of nuclei correspondingly and t@xmath116 - nucleus recoil energy .
if one uses gda with mass of 3 kg and threshold 30 ev in the reactor experiment the count rate due ncs in the energy range @xmath117 ev will be near 100 event / d at antineutrino intensity @xmath118 @xmath119 due to the weak interaction .
the count rate due to the electromagnetic interaction ( at @xmath120 ) will be 0.45 event / d . the level of detector background will be only 0.1 event / d if one uses the above estimations of detector background , taking into account that the energy range is equal 0.3 kev in this case .
so one can see that use of gda for measurement ncs makes it possible : * to improve the weak interaction constants , * to investigate the neutrino oscillations by alternative way , * to make more precise the quenching factor for germanium nuclei at low energy transfer which is of interest for dark matter experiments .
presently in dark matter experiments the main efforts are directed at decreasing of the background and lowering of the energy threshold of the detector , since for more appropriate dark matter candidate- wimps or weakly interacting massive particles - the expected count rate for wimps - nuclear scattering is in the range @xmath121 event / kg@xmath110d and expected recoil energy @xmath122 lies in the wide energy range beginning of some tens @xmath1 and above .
significant success was achieved by cdms ( cryogenic dark matter search ) collaboration in radiation background decreasing @xmath123 .
the rejection of 99@xmath77 of photon background was achieved use detectors which simultaneously measure the recoil energy in both- photon and charge mediated signals . however , this method is effective at detector threshold above 15 kev .
so experiments for search for wimps with mass lower than 10 gev/@xmath124 ( at very low t ) are carried out now by cresst ( cryogenic rare event search with superconducting thermometers ) collaboration @xmath125 which plans to use cryogenic detector including four 250 g sapphire detectors with threshold 500 ev .
the using of several gda modules of mass 1 kg each with threshold about 30 ev would be very effective in this investigations . the investigation of solar neutrino problem with use gda can provide possibility to detect simultaneously the whole neutrino spectrum @xmath126 but one needs of course to use large mass detector in this case .
it is beleived that gda being developed will find also use in the applied fields .
99 j. p. pansart , nucl .
instr . and
a387 * ( 1997 ) 186 .
r. farrel , f. olschner , k. shah , m. r. squillante , nucl . instr . and meth .
* a387 * ( 1997 ) 194 .
bertolini and a .
coche , semiconductor detectors .
north - holland publishing company - amsterdam , 1968 , p. 71 ; e. sakrai , h. l. malm and i. l. fovler , paper presented at the gatlinburg conference ( 1967 ) . j. m. marler , p. v. hewka , ieee @xmath127 , n.1 ( 1974 ) 287 .
d.gutknecht , nucl .
instr . and meth .
@xmath128 ( 1990 ) 13 .
a.a.vasenko , i.v.kirpichnikov , a.s.starostin et al , mod .
a*5 * , n.17 ( 1990 ) 1299 - 1306 .
a. g. beda , e. v. demidova , a. s. starostin and m. b. voloshin , yad . fiz .
* 61*,n.1 ( 1998 ) 72 [ phys.at.nucl .
( engl . transl . )
* 61 * , n.1 ( 1998 ) 66 ] .
v. n. karnoukhov , itep report n 90 - 94 , 1994 ; a. v. davydov and yu .
n. isaev , yadern . fiz .
* 59*,n.3 ( 1996 ) 486 [ phys.at.nucl .
transl . ) * 59 * , n.3 ( 1996 ) 459 ] .
l.a.mikaelyan , v.v.sinev , and s.a.fayans , pisma zh .
eksp . teor.fiz.*67
* , no.7,435 [ jetp letters(engl.transl.),*67*,no.7(1998)p.453 .
v. n. trofimov , b. s. neganov and a. a. yukhimchuk , yad.fiz .
* 61 * ( 1998 ) 1373 .
brodzinski , r.l . ,
avignone , f.t . , et al .
methods a*292 * ( 1990 ) 337 .
beck m. et al .
( haidelberg - moscow coll . ) , phys.lett .
* 336b * ( 1994 ) 141 .
r. bernabei , riv .
nuovo cimento @xmath129 , n.5 ( 1995 ) 35 .
b. cabrera , l. m. krauss and f. wilczek , phys .
@xmath130 ( 1985 ) 25 a. c. dodd , e. papageorgiu and s. ranfone , phys.lett .
@xmath131 ( 1991 ) 434 .
d. s. akerib , d. o. caldwell , b. sadoulet et al .
, astro - ph/9712343 ; rep . at taup97 conf . , sept 7 - 11 , 1997 , lngs , italy
m.buhler , l. zerle , f. probst et al .
instr . and
meth . @xmath132 ( 1996 ) 237 .
|
device of new type is suggested - germanium detector with internal amplification .
such detector having effective threshold about 10 ev opens up fresh opportunity for investigation of dark matter , measurement of neutrino magnet moment , of neutrino coherent scattering at nuclei and for study of solar neutrino problem .
construction of germanium detector with internal amplification and perspectives of its use are described .
10000 -27pt 0.00 in 0.00 in 6.5 in 9.2 in 1 . in * germanium detector with internal amplification for investigation of rare processes
* + +
|
optical solitons were first predicted by hasegawa and tappert @xcite and later observed by mollenauer _ et .
the existence of soliton solutons in non - linear schrdinger equation ( nlse ) is related to the complete integrability of this dynamical system @xcite .
it was found that non - linear schrdinger equation aptly describes the pulse propagation in picosecond regime .
generalization of the nlse is neccessitated to take into account higher order dispersion , self - steepening of the pulse due to the dependence of the slowly varying part of the non - linear polarization on time and the delayed effect of raman response for describing optical pulse propagation in the femtosecond domain @xcite .
higher order nlse which captures these effects is given by @xcite , @xmath1 it has both group velocity dispersion ( gvd ) and third order dispersion term , respectively represented by @xmath2 and @xmath3 .
parameter @xmath4 takes into account kerr non - linearity , whereas self steepening effect arising from time variation of the non - linearity @xmath5 has been incorporated through the coefficient @xmath6 .
the terms with coefficient @xmath7 and @xmath8 account for the self frequency shift , originating from delayed raman response .
this model , unlike nlse , is not integrable in general .
a few integrable cases have been identified ; these are known as ( i ) sasa - satsuma case @xcite , ( ii ) hirota case @xcite and ( iii ) derivative nlse of type i and type ii @xcite .
a few restrictive special solutions of bright and dark type have been obtained for this system @xcite .
effects of these higher order terms on pulse propagation have been studied numerically quite extensively @xcite ; some special solutions to this system are also known @xcite .
third order dispersion becomes important for femtosecond pulses , when group velocity dispersion ( gvd ) is close to zero .
it can be neglected for optical pulses , whose width is of the order of @xmath9 femtoseconds or more , having power of the order of 1 watt and gvd far away from zero . for short range propagation , @xmath8 can also be ignored .
the resulting equation can be written in a dimensionless form by appropriately scaling variables : @xmath10 where @xmath11 $ ] of the incident pulse .
the dimensionless equation obtained reads : @xmath12 where , @xmath13 in a previous study by some of the present authors , localized wave packets with nontrivial chirping were emerged as exact solutions of this equation @xcite .
it was found that chirping has a kinematic component , determined through initial conditions and a dynamical component , originating from self - steepening of pulse and its delayed raman response .
these two chirpings have inverse characteristics and can be used for chirp control .
further , periodic solutions , with sinusoidal functions in fractional linear form , have also been identified . in this paper
, we systematically investigate wide class of solutions of this system containing both bright and dark solitons .
interestingly , the equation governing dynamics of intensity is found to be the one that governs dynamics of field amplitude in non - linear schrdinger equation with a source @xcite .
this connection is then profitably used to obtain exact solitonic intensity waves of this higher order nlse , possessing nontrivial chirping behavior .
the nlse type bright solitons are found to propagate at superluminal velocity , with dark solitons complementing them in parameter domains , which are experimentally accessible .
interestingly , both the solutions also complement each other in velocity space .
further , we find that dark soliton which are static in systems like nlse , are propagating in this system solely due to presence of higher order nonlinearities .
these solutions show chirality , which depends on the sign of gvd parameter . depending on the kinematic chirp parameter
two distinct domains are obtained . in the first case , if solitons are chiral for anomalous dispersion ( @xmath14 ) , then they are non - chiral for normal dispersion ( @xmath15 ) . in the second case , solitons are strictly chiral in both the dispersion domains .
we then obtain more general localized solutions , through a conformal transformation , which can travel at both sub and superluminal velocities . in order to classify and gain physical understanding of the obtained solutions ,
we then use a pseudo - potential picture , which shows that the solutions separate into bounded periodic oscillations around the well minima and unbounded motion , separated by the so called separatrix solution .
the separatrix solution is identified and the connection of this system with hydrodynamical equations are pointed out .
interestingly , the localized kink type solitons of this system representing the separatrix can be mapped to the separatix solutions of the elliptic function equation , through the aforementioned conformal transformation .
the paper is organized as follows . in the following section ,
we explore the localized dark and bright soliton solutions after connecting the present dynamical system with the nlse .
chirality of the solutions and the effect of chirping have been systematically investigated . in sec .
iii , we find more general solutions , through a fractional linear transform and show they can be both sub and super luminal depending upon various parameters . section iv deals with conclusion and future directions of work .
we seek a travelling wave packet type solution of the hnlse , which can generally be written as : @xmath16 here the travelling coordinate @xmath17 , and @xmath18 and @xmath19 are real functions of @xmath20 , @xmath21 is the scale parameter and @xmath22 , where @xmath23 is group velocity of the ansatz solution . substituting ( [ ansatz ] ) in ( [ hnlse ] ) , and equating real and imaginary parts yields two coupled equations : @xmath24 equation ( [ c1 ] )
can be exactly integrated to yield : @xmath25 where @xmath26 is a constant of integration , to be determined by initial conditions .
notice that the phase has a nontrivial form and has two intensity dependent chirping terms , apart from the kinematic first term , which is of usual @xmath27 type .
as is evident , the second term is also of kinematic origin and is infact present in the linear schrdinger equation . the last term is due to higher order nonlinearities , which has a dynamical origin , and leads to chirping that is exactly inverse to that of the former .
this is a novel form of self - phase modulation . using the above expression , eq .
( [ c2 ] ) can be written as : @xmath28 with @xmath29 , @xmath30 , @xmath31 and @xmath32 .
note that , in parameter regime , where @xmath33 , the amplitude dynamics will be like the nlse case , albeit with dressed parameters .
phase dynamics , however will not be the same as that of nlse , and will depend on , whether @xmath34 or @xmath35 . in the latter case only , it will be exactly like nlse , devoid of dynamical chirping . as shown in ref .
@xcite , localized solutions of the above equation , in these parameter domains , show directionality and hence are chiral , and the propagation direction for these solitons is decided by sign of @xmath36 . because , these solutions satisfy nlse with dressed parameters , many of the parametric restrictions on solution space are relaxed , for example both dark and bright solitons are present in both anomalous and normal dispersion regimes . defining @xmath37 ,
( [ thetaeqn ] ) can be written conveniently in terms of @xmath38 ( where @xmath39 ) to yield : @xmath40 where , @xmath41 is to determined via initial conditions . motivated by the similarity of the above intensity equation with that of amplitude equation of nlse in the presence of a source @xcite , below we consider different ansatz solutions . * case i * : first we study localized soliton ansatz of the type @xmath42 which allows for both dark and bright solitons , depending on the sign of @xmath43 .
consistency of the above solution yields the following conditions : @xmath44 solving first three of these relations , one finds @xmath45 since , @xmath46 , reality of @xmath47 demands @xmath48 .
similarly , reality of @xmath43 requires , @xmath49 , whereas positivity of @xmath50 leads to @xmath51 .
these conditions , together with the above relations , lead to : @xmath52 where @xmath53 . in case of grey solitons , _
i.e. _ , @xmath54 , for positivity of @xmath38 requires @xmath55 , which gives , @xmath56 we note that @xmath57 can be positive or negative .
depending on the value of @xmath58 , @xmath59 shows possible chirality of the solutions . in case when @xmath60 , intensity vanishes at origin and these solutions are known as dark solitons .
interestingly , derivative of phase ( from eq.([phase ] ) ) becomes singular at that point , provided @xmath61 .
this inturn implies that phase exhibits a discontinuous jumps across origin .
dark solitons of this kind have long been know in case of nlse where they are strictly static @xcite .
velocity space of dark solitons in our case can be investigated using eq .
( [ dark ] ) .
one can easily check that these solitons in general are not static , unlike nlse , and exhibit superluminal propagation with velocity depending on @xmath26 .
however , if @xmath62 , then these solitons can be static .
interestingly , in the limit @xmath63 one can immediately see that eq.([dark ] ) only allows for static solitons which corroborates with known results .
this essentially implies that presence of higher order nonlinearities allows these kind of dark solitons to propagate with non - zero velocity .
the last consistency relation ( eq.([cc4 ] ) ) determines @xmath64 , the soliton velocity in terms of @xmath41 : @xmath65 we note that @xmath30 and @xmath31 contain @xmath64 explicitly . as defined earlier
, @xmath66 is a constant of motion , and hence its value remains unchanged under time evolution . at @xmath67 , it is seen that @xmath68 and @xmath69 , which allows one to write : @xmath70 using this , along with eq .
( [ u ] ) , yields an equation for @xmath64 involving the chirp parameter : @xmath71 when the constant of integration , @xmath72 one finds a quadratic equation for u : @xmath73 reality of @xmath64 can be ensured by demanding positivity of the discriminant in the above equation , which leads to @xmath48 ; this condition is trivially satisfied as it emerges from the reality of @xmath47 .
we now systematically investigate the chirality of the solutions .
for anomalous dispersion regime , @xmath74 , velocity is found to be positive and bounded from above , @xmath75 .
for @xmath76 , @xmath77 or @xmath78 . for @xmath79
, it is straightforward to see that both the roots of eq .
( [ qu ] ) are negative , hence the possibility of @xmath77 is forbidden . hence , the solution is unidirectional and show chirality , with the direction of propagation governed by the sign of @xmath80 .
typical experimental values corresponding to pulse propagation around @xmath9 femtosecond region are @xmath81 . using these values
, one finds @xmath82 , which after appropriate rescaling leads to @xmath83 , yielding superluminal velocity .
we have checked that in the currently experimentally accessible parameter domain , all the solution are superluminal .
* case ii * : next we consider kink type solutions : @xmath84 where @xmath85 due to positivity of intensity .
proceeding like the earlier case , one finds four consistency conditions , three of which yield : @xmath86 reality of @xmath47 requires @xmath87 , which along with positivity of @xmath50 implies @xmath88 .
this clearly implies that , the kink type solution and the earlier found soliton live in mutually disjoint parameter and velocity spaces .
further , as in the earlier case , the reality of @xmath89 requires , @xmath90 to be satisfied .
since @xmath91 , the above inequality holds for all real values of @xmath64 , provided the discriminant is negative .
this leads to restrictions on the chirp parameter : @xmath92 when above relation is not obeyed , the allowed @xmath64 values are forbidden between the two roots . as noted above , positivity of intensity
demands @xmath93 , which translates to : @xmath94 imposing an upper bound on @xmath26 : @xmath95 since , @xmath96 , for existence of these solutions , @xmath26 has to be a non - zero negative quantity . like the earlier case , the consistency condition @xmath97 requires @xmath98 for the normal dispersion region , this condition gives , @xmath99 and for anomalous region ( @xmath100 ) above inequality leads to @xmath101 or @xmath102 , if @xmath103 when @xmath104 , eq .
( [ tanhcon ] ) forbids the negative values of velocity .
hence for this interval , we have strict chirality for the solutions .
the remaining consistency condition yields an equation for @xmath64 : @xmath105 which has the same form as in the previous case , albeit the value of @xmath41 will be different .
constant of motion , @xmath41 , is determined by evaluating it at @xmath106 : @xmath107 hence , equation for @xmath64 reads : @xmath108 which is a quartic equation .
we have not been able to find real roots of this equation for experimental values around 100 femtosecond pulse width
. this can be proved by looking at the sub determinant structure of this quartic equation , _
i.e. _ , @xmath109 @xcite .
the sign of @xmath110 ( where @xmath111 ) , will determine the characteristics of the roots .
lets us take the list of signs of the @xmath110 , _
i.e. _ , @xmath112 $ ] , where @xmath113 is the sign of @xmath110 .
the number of sign change in the sequence reveals the number of real pairs .
if the sequence has @xmath114 number of change in sign , then @xmath114 pair of complex roots exist . here
@xmath115 are always satisfied in the allowed domain @xmath116 , @xmath117 .
hence for a real solution , we must have @xmath118 . in fig .
[ fig ] we have plotted @xmath119 and @xmath120 as a function of @xmath26 and @xmath121 respectively , one can see from the graph that both quantity do not share same sign for any value of @xmath26 and @xmath121 .
apart from ansatz solutions discussed above , below we show that eq.([sigma ] ) can be solved quite generally , using a conformal mapping involving jacobi elliptic functions .
* case i * : when @xmath122 in this case , eq .
( [ sigma ] ) can be written in a simpler form by going to @xmath123 variable , @xmath124 , @xmath125 where @xmath126 , @xmath127 and @xmath128 . in case when @xmath129 , above equation coincides with the equation satisfied by jacobi elliptic functions .
below we show that , for some real constants @xmath130 , @xmath131 and @xmath132 , the conformal transformation : @xmath133 solves equation ( [ nlsesource ] ) , with @xmath134 being one of the twelve jacobi elliptic functions with modulus parameter @xmath135 @xcite .
these elliptic functions satisfy : @xmath136 with appropriate real constants @xmath137 and @xmath138 .
for example , for @xmath139 , @xmath140 and @xmath141 .
the claim that expression ( [ solun ] ) solves eq .
( [ nlsesource ] ) can readily be seen , by defining the first integral @xmath142 , and substituting ( [ solun ] ) in eq .
( [ nlsesource ] ) .
use of eq .
( [ elliptic ] ) leads to the following consistency conditions : @xmath143 notice that these coupled algebraic equations are nonlinear in @xmath130 , @xmath131 and @xmath132 , but are linear in @xmath144 , @xmath145 , @xmath146 and @xmath147 , and hence can be solved exactly to yield : @xmath148 therefore , this shows that , provided above relations hold , eq . ( [ nlsesource ] )
is indeed solved by ( [ solun ] ) .
notice the dependence of @xmath147 on @xmath130 , @xmath131 and @xmath132 , which simply shows that the initial values in eq .
( [ elliptic ] ) do play a role in determining @xmath130 , @xmath131 and @xmath132 , and hence they can not be fixed , given @xmath144 , @xmath145 and @xmath146 . since , original equation parameters @xmath149 ( @xmath150 ) can be expressed in terms of @xmath144 , @xmath145 and @xmath146 , one can determine the parameter regime for a given solution using these relations .
the solutions presented in ref .
@xcite form a subclass of the ones found above . for @xmath151 ,
one gets localized solutions and are often of experimental and technological interest . setting @xmath152 ,
singles out periodic solutions , which were reported in ref.@xcite . for explicitness
, we consider @xmath153 , which gives @xmath154 and @xmath155 is found to be 0 .
the consistency equation becomes @xmath156 after eliminating @xmath131 and @xmath132 one gets , @xmath157 implying that velocity of soliton is governed by background amplitude in this non - linear system .
since we know the values of @xmath144 , @xmath145 and @xmath146 , by fixing the value of @xmath64 , we can get the corresponding values of @xmath131 and @xmath132 in the form , @xmath158 for the experimental values , @xmath159 , @xmath160 , @xmath161 and putting @xmath162 ( sub - luminal ) , we get @xmath163 , @xmath164 and @xmath165 . in the case of super - luminal propagation , @xmath166 , we get @xmath167 and @xmath168 . as is evident , @xmath169 and @xmath170 . for @xmath171 , in eq .
( [ elliptic ] ) one finds @xmath172 and @xmath173 . in this case , the consistency conditions become @xmath174 after simplification , we get a quadratic equation for @xmath64 along with other consistency relations given by , @xmath175 once @xmath64 is known , @xmath130 can be evaluated from eq .
( [ aeq ] ) through @xmath41 dependence on @xmath130 , then @xmath132 and @xmath131 can be obtained using eq . ( [ ceq ] ) and ( [ beq ] ) respectively . for the consistency of eq .
( [ ceq ] ) , @xmath87 must be satisfied .
regularity of solution requires @xmath176 . also , the positivity of @xmath38 demands @xmath177 , which gives restriction on the velocity of the soliton . above obtained solutions
can be better understood using following pseudo particle picture .
( [ nlsesource ] ) can be thought of as equation of motion for a classical particle of unit mass , with displacement given by @xmath123 , moving under influence of a nonlinear force .
the same also holds for eq .
( [ elliptic ] ) , albeit the nonlinear force does not have a constant term .
also notice that both these dynamical systems are of one degree of freedom and their phase spaces are two dimensional , only position @xmath123 and momentum @xmath178 is required to describe the dynamics completely . looked in this setting , mapping eq .
( [ solun ] ) simply says how the two phase spaces are related ; more precisely it shows how phase space for eq .
( [ nlsesource ] ) can be generated by knowledge of phase space for eq .
( [ elliptic ] ) . since , the mapping is conformal , the singularity structure of both the phase spaces are identical , modulo movable poles .
infact , it indicates that eq .
( [ nlsesource ] ) is integrable in the sense of painlev , since eq .
( [ elliptic ] ) is integrable @xcite .
the kink type solutions of the parent eq .
( [ hnlse ] ) , in this setting , can be identified with the separatrix solutions of eq .
( [ elliptic ] ) via the conformal mapping @xcite .
similarly , singular solutions of eq .
( [ elliptic ] ) can be seen to be related to unbounded solutions lying outside separatrix , and periodic solutions can be seen to be related to bounded solutions lying inside separatrix .
it is very interesting to note that solutions to eq .
( [ nlsesource ] ) are actually solutions to modified kdv equation : @xmath179 .
this can be seen using the travelling variable @xmath180 , integrating out the equation once , and identifying @xmath181 , @xmath182 and @xmath146 as the constant of integration .
so , at the traveling variable level , which restricts one to only one soliton solution , one can say that the modified kdv solitons can be simulated by considering these higher nonlinearities in optical fiber .
* case ii * : when @xmath33 in this case , the above analysis is not valid since the degree of the eq .
( [ thetaeqn ] ) changes and can be written as : @xmath183 with @xmath184 , @xmath185 and @xmath186 .
the above equation can be mapped to eq .
( [ elliptic ] ) via a transformation : @xmath187 which is along the same lines as ( [ solun ] ) . following the same procedure as the former case ,
using the first integral @xmath142 , we find the consistency conditions as : @xmath188 again we observe that , the above equations are linear in @xmath189 , @xmath190 , @xmath191 and @xmath147 , and hence can be solved to give : @xmath192 this shows that , the eq .
( [ thetaeqn2 ] ) is an integrable equation , in the sense of painlev , and the explicit solutions can be expressed in terms of jacobi elliptic functions . by choice , of appropriate values of @xmath130 , @xmath131 , @xmath132 and @xmath193 one can find out the parameter regime in which the given solution is valid .
further , solutions to eq .
( [ thetaeqn2 ] ) actually satisfy kdv equation : @xmath194 , in travelling variable @xmath195 , with @xmath196 , @xmath197 and @xmath191 being an integration constant .
it is very interesting to note that the condition @xmath33 , can be fulfilled if @xmath198 or @xmath199 . in the case
when the former is true then , the intensity profile of these solutions is exactly like nlse solutions except with a non - trivial phase chirping . in the latter case ,
the solutions do not have this non - trivial chirping and the solutions are indistinguishable from nlse solutions with appropriate coefficients .
in conclusion , we have shown that nonlinear schrdinger equation in the presence of self - steepening and self - frequency shift , possesses rich travelling wave dynamics , with non - trivial chirping .
self - phase modulation revealed here indicates the possibility of effective control of pulse dynamics through chirp management .
the dynamics of the intensity is governed by the non - linear schrdinger equation with a source .
hence , intensity waves show solitonic behavior , in complete analogy with the integrable nlse .
these solutions exhibit superluminal propagation in experimentally accessible parameter domain .
further , these solutions reveal directional propagation , akin to hydrodynamical equations like kdv , with which the obtained solutions have close connection . we have found that dark solitons in this system can propagate with nonzero velocity , unlike in the case of nlse .
more general solutions obtained through fractional linear transform , showed both sub and superluminal behavior .
we identified the separatrix solution bounding the regular oscillatory motion from the unbounded ones .
this strengthens the hope that this system may be integrable , since separatrix is not known to exist in a system with chaotic dynamics .
b. grammaticos and a. ramani , in _ integrability of nonlinear systems _ edited by y. kosmann - schwarzbach , b. grammaticos , and k.m .
tamizhmani , lecture notes in physics 638 ( springer - verlag , berlin heidelberg , 2004 ) and references therein .
|
we investigate exact travelling wave solutions of higher order nonlinear schrdinger equation in the absence of third order dispersion , which exhibit non - trivial self phase modulation .
it is shown that , the corresponding dynamical equation , governing the evolution of intensity in the femtosecond regime , is that of non - linear schrdinger equation with a source .
the exact localized solutions to this system can have both super and subluminal propagation belonging to two distinct class .
a number of these solitons exhibit chirality , thereby showing preferential propagation behavior determined by group velocity dispersion .
both localized bright and dark solitons are found in complementary velocity and experimental parameter domains , which can exist for anomalous and normal dispersion regimes .
it is found that , dark solitons in this system propagate with non - zero velocity , unlike their counterpart in nanosecond regime .
interestingly , subluminal propagation is observed for solitons having a nontrivial @xmath0 type intensity profile .
|
Subsets and Splits
Data Mining Abstracts
This query searches for and retrieves 2000 documents containing the phrase 'data mining' in the abstract, providing limited but specific document samples.