article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the relevance of bessel functions to the solution of a great variety of problems in physics and engineering is widely known and does not need to be stressed . in particular , zeros of the different kinds of those functions are closely related to energies of bound or resonant physical systems .
for instance , the @xmath2-zeros of the hankel function @xmath5 for positive values of @xmath0 determine the poles of the amplitude of scattering of various kinds of waves by spheres and cylinders @xcite .
almost four decades ago we considered @xcite the zeros of the modified bessel function of the third kind , @xmath1 , also known as modified hankel function or macdonald function .
information about the @xmath0-zeros of @xmath1 for positive order @xmath2 was found in the book by watson @xcite . however , our interest , motivated by a quantum mechanical problem , was on @xmath0-zeros when @xmath2 assumes purely imaginary values .
for that case it was already known @xcite that @xmath1 presents an infinity of zeros , all of them located on the positive real semiaxis of the complex @xmath0-plane .
we showed that these positive zeros , labelled according to decreasing values , form a sequence tending to be a geometric progression with ratio @xmath6 .
we gave , also , asymptotic approximations to the position of the largest zeros for large values of @xmath3 .
further discussions of the properties of the zeros of @xmath1 , @xmath2 pure imaginary , have been carried out by laforgia @xcite and by dunster @xcite .
zeros of @xmath1 with complex @xmath2 have been less studied . in view of their application in the analysis of electromagnetic wave propagation in anisotropic inhomogeneous conducting media ,
nalesso @xcite has considered the @xmath2-zeros of @xmath7 for fixed positive values of @xmath8 .
but , to our knowledge , a discussion of the @xmath0-zeros of @xmath1 for arbitrary complex @xmath2 is still lacking . in recent years , stimulated by review papers by lozier and olver @xcite and by the dlmf project , a considerable number of new algorithms for the computation of special functions have been published . concerning the macdonald function , refs .
@xcite are to be noticed . as a general rule ,
algorithms lose accuracy in the vicinity of the zeros .
hence the interest of having a knowledge , at least approximate , of the location of those zeros . on the other hand
, @xmath1 presents only a finite number of @xmath0-zeros in the riemann sheet @xmath9 if @xmath2 is real , whereas an infinity of zeros occur if @xmath2 is pure imaginary .
it has therefore seemed to us interesting to study the evolution of the @xmath0-zeros of @xmath1 as the modulus and/or argument of @xmath2 are continuously changed . in what follows
, we consider only values of @xmath2 in the quadrant @xmath10 and @xmath0 is assumed to lie in the principal riemann sheet , @xmath9 , for large values of @xmath2 .
the symmetry relations @xmath11 allow to extend our results to values of @xmath2 in all quadrants . on the other hand ,
the relation ( * ? ? ?
9.6.4 ) @xmath12 valid when @xmath13 and @xmath14 are analytically continued to every riemann sheet , proves that the pattern of @xmath0-zeros of the hankel function @xmath15 results from that of @xmath1 by rotation by an angle of @xmath16 around the origin .
we discuss in section 2 the behaviour of the zeros for decreasing values of @xmath3 .
the case of large values of @xmath3 is considered in section 3 , where asymptotic approximations to the zeros are given .
numerical results for the zeros of @xmath1 with moderate @xmath3 are presented graphically in section 4 .
finally , the particular case of real @xmath2 is commented in section 5 .
the particular case of pure imaginary @xmath2 revealed @xcite that , as @xmath3 goes to zero , all zeros of @xmath1 approach the origin .
therefore we start by assuming @xmath18 in the expression of @xmath1 as a sum of two convergent ascending series ( * ? ? ?
9.6.2 and 9.6.10 ) @xmath19 valid for @xmath2 different from an integer .
obviously , the zeros of @xmath1 should satisfy the relation @xmath20 that , in the case of being @xmath18 , can be approximated by @xmath21 where @xmath22 could in principle take the values @xmath23 , although , according to that obtained in the case of @xmath2 being pure imaginary , actually existing zeros correspond to @xmath24 these values of @xmath22 will be used as a label to characterize each zero in the form @xmath25 .
ignoring , in the crudest approximation , the last term in the right hand side of eq .
( [ seis ] ) , one has @xmath26 the behavior of the zeros @xmath25 as @xmath3 decreases can be immediately obtained from these expressions by using the series expansion ( * ? ? ?
6.1.34 ) @xmath27 to obtain @xmath28 where @xmath29 represents the euler constant , @xmath30 . by retaining terms up to @xmath31
, it turns out @xmath32 these two equations allow us to obtain a picture of the behaviour of the zeros of @xmath1 as @xmath3 tends to zero .
for instance , let us assume that @xmath3 decreases with @xmath4 fixed . if @xmath33 , the zeros @xmath25 approach the origin along the positive real semiaxis , as described in ref .
if @xmath34 , @xmath35 tends to zero whereas @xmath36 decreases without limit . in other words , the zeros approach the origin spiraling clockwise infinitely . according to eq .
( [ ddos ] ) , most of zeros ( or even all of them , for sufficiently small @xmath3 ) occur outside the principal riemann sheet . for @xmath37 , eqs .
( [ duno ] ) and ( [ ddos ] ) are no more valid because , as we will see later , the possibility exists of @xmath25 going to infinity for certain values of @xmath2 .
the behaviour of the zeros in this case is much more complicated and will be described in section 5 .
we have already mentioned that eq .
( [ dos ] ) allows one to obtain the zeros of the macdonald function from those of the hankel function , and viceversa .
previous research , by other authors @xcite , on the @xmath2-zeros of the hankel function shows that , for large values of @xmath39 , they occur at @xmath40 .
asymptotic expansions of @xmath14 valid in the transition region should then be used to obtain approximate expressions of its @xmath0-zeros for @xmath38 . from eqs .
9.3.23 and 9.3.24 of ref .
@xcite , by using eqs .
9.1.3 and 10.4.9 , a convenient expansion can be written , namely @xmath41 where @xmath42 , the other functions @xmath43 and @xmath44 being given in eqs .
9.3.25 and 9.3.26 of the same reference . a more powerful asymptotic expansion of @xmath14 exists ( * ?
9.3.37 ) , but its coefficients are much more complicated and , although reversion of that expansion could be done by means of a procedure due to fabijonas and olver @xcite , we find preferable , for the sake of simplicity and transparency , to work with eq .
( [ dtres ] ) .
a simple inspection of that expansion shows that , for large @xmath3 , it vanishes for values of @xmath45 verifying @xmath46 where @xmath47 represents each one of the zeros of the airy function , given in table 10.13 of ref .
@xcite , and @xmath48 is to be determined . trying an expansion of the form @xmath49 and using taylor expansions of the airy function and of its derivative around @xmath47 , it is straightforward , with the aid of a computer algebra software package , to check that the right hand side of eq .
( [ dtres ] ) cancels out for @xmath50 approximate values of the @xmath0-zeros of @xmath15 are then given by @xmath51 reversion of this relation between @xmath0 and @xmath2 would provide an approximation for the @xmath2-zeros of @xmath15 when @xmath39 is large , @xmath52 to be compared with those given in refs .
@xcite . according to eq .
( [ dos ] ) , the @xmath0-zeros of @xmath53 are given approximately by @xmath54 the usefulness of eqs .
( [ dsiete ] ) and ( [ docho ] ) is restricted to the first values of @xmath55 , due to the fact that @xmath56 increases with @xmath57 .
this makes the the omitted terms in the expansion ( [ dsiete ] ) to increase and invalidates their suppression .
besides this , @xmath58 , given by ( [ dcuatro ] ) , increases with @xmath57 and , for a given @xmath2 , may reach a value such that expansion ( [ dtres ] ) ceases from being useful .
if zeros corresponding to a large @xmath57 are to be considered , approximate values can be obtained following a procedure , already used by magnus and kotin @xcite and by cochran @xcite , consisting in taking logarithms in eq .
( [ cuatro ] ) , @xmath59 and approximating the logarithms of the gamma functions by their asymptotic expansions ( * ? ? ? * eq .
one obtains in this way @xmath60 assuming @xmath61 and keeping the dominant terms , it comes out for the @xmath0-zeros of @xmath1 @xmath62 obviously , since we are considering values of @xmath2 in the first quadrant , only large positive values of the integer @xmath22 are compatible with the assumption @xmath61 .
moreover , for a given value of @xmath3 , the approximation ( [ vuno ] ) looses accuracy as @xmath4 goes from @xmath63 towards @xmath64 and becomes completely useless for real @xmath2 . a numerical comparison of the values obtained with eqs .
( [ docho ] ) and ( [ vuno ] ) for moderate values of @xmath57 and @xmath22 shows that both labels coincide .
we have used a numerical procedure to explore the behaviour of the @xmath0-zeros of @xmath1 , as a function of @xmath2 , for moderate values of @xmath3 . in order to simplify the presentation of the results ,
two different situations have been considered : ( i ) variable @xmath3 with fixed @xmath4 , and ( ii ) variable @xmath65 keeping @xmath3 fixed .
the resulting trajectories of the first , second , and third zeros are represented in figures 1 , 2 , and 3 , respectively . for
the evaluation of @xmath1 , expression ( [ tres ] ) has been used .
more sophisticated algorithms can be found in the literature @xcite , but either they have been conceived only for particular ( real or pure imaginary ) values of the variable or the order , or they become unnecessarily complicated for our purpose of giving a merely qualitative description of the zeros . for the location of these , the newton method has been used .
let us mention , however , other existing techniques like , for instance , the auxiliary tables for the evaluation of the @xmath2-zeros of @xmath1 , for @xmath0 real or pure imaginary , proposed by cochran and hoffspiegel @xcite , the finite element approximation method , applied by leung and ghaderpanah @xcite to a very precise determination of the zeros of @xmath66 , and a procedure , applied by segura @xcite to the computation of zeros of bessel and other special functions , that uses fixed point iterations and does not require the evaluation of the functions . a double precision fortran code was used to compute the expression @xmath67 whose zeros coincide with those of the right hand side of eq .
( [ tres ] ) .
the two series were summed term by term until the absolute value of the ratio of the last term to the partial sum turned out to be less than @xmath68 .
a warning message was foreseen for the case of the number of summed terms becoming larger than @xmath69 , but this limit was never reached .
fortunately , zeros at large values of @xmath39 occur for values of @xmath2 either of large modulus or in the neighbourhood of @xmath70 , @xmath71 .
in the first case , the series in ( [ extra ] ) converge reasonably , from the numerical point of view ; in the second case , analytical methods can be used to approximate the zeros .
typical runs of our procedure can be seen in table 1 , which shows intermediate results in the determination of the first , second and third zeros of @xmath1 , with @xmath72 , @xmath73 .
initial approximate values of the zeros , for a given @xmath2 , were obtained by extrapolation of the trajectories followed by the zeros as @xmath2 varies . besides the smallness of the values of the real and imaginary parts of @xmath74 , their changes of sign in the successive steps of the newton method make us to be confident about our results .
.typical output of the newton method applied to the location of the first , second and third zeros of @xmath1 , with @xmath72 , @xmath75 . besides the successive approximations to the zeros , the values of @xmath74 , given by eq .
( [ extra ] ) , are shown . [
cols=">,>,>,>,>,>,>",options="header " , ] our figures show that for large @xmath3 , with @xmath76 , the zeros lie far from the origin in the half plane @xmath77 , according to eq .
( [ docho ] ) .
when @xmath3 decreases , they approach the origin but , before reaching it , they cross the semiaxis @xmath78 , leaving in this way the principal riemann sheet @xmath9 , for a certain @xmath79 whose value depends on @xmath4 and on the order @xmath57 of the zero @xmath80 . the smaller @xmath4 ( between @xmath64 and @xmath63 ) and the higher the label @xmath57 of the zero , the larger the value of @xmath79 .
( in the case of real @xmath2 , it can be shown that @xmath81 . ) therefore , except for the case of pure imaginary @xmath2 , the number of zeros of @xmath1 in the principal riemann sheet is finite : only those zeros with label @xmath82 such that @xmath83 , for the corresponding value of @xmath4 , remain in that principal riemann sheet . if @xmath3 decreases again , being @xmath84 , the zeros approach the origin spiraling clockwise infinitely , according to eqs .
( [ duno ] ) and ( [ ddos ] ) .
the case of real @xmath2 deserves a particular consideration .
let us now concentrate on the particular case of @xmath2 being real . according to the second of eqs .
( [ uno ] ) , the zeros of @xmath1 appear in complex conjugate pairs .
we refer , in what follows , only to those zeros of negative argument .
let us also recall that , for half - integer @xmath2 , @xmath85 , the macdonald function reduces to a factor , different from zero in the finite plane , times a polynomial of degree @xmath22 .
therefore , @xmath1 presents exactly @xmath22 @xmath0-zeros in each one of the riemann sheets .
of course , the locations of these zeros are the same in all sheets . the aproximate expression ( [ docho ] ) for the first zeros of @xmath1 ( @xmath86 ) is valid also in this case . equation ( [ vuno ] ) , instead , is not applicable because the assumption @xmath87 , used in its derivation , is not justified . also , eqs .
( [ duno ] ) and ( [ ddos ] ) are no more valid due to the possibility of @xmath39 becoming infinite for certain values of @xmath2 , a fact that invalidates eq .
( [ cinco ] ) , from which eqs .
( [ duno ] ) and ( [ ddos ] ) stem . according to eq .
( 2 ) , the @xmath0-zeros of @xmath1 are obtained from the @xmath0-zeros of @xmath15 by a rotation of @xmath88 in the complex @xmath0-plane .
it is then immediate to deduce the behaviour of the zeros of @xmath1 for varying real @xmath2 from a previous discussion @xcite of the zeros of @xmath15 . for convenience of the reader ,
we recall here , without demonstration , the main results of that discussion translated to those zeros in the lower half plane @xmath89 . to be specific ,
let us focus on the trajectory followed by a zero , @xmath80 , as @xmath2 decreases from @xmath90 to @xmath64 . for very large @xmath2
, @xmath80 is given approximately by eq .
( [ docho ] ) .
it lies far from the origin in the quadrant @xmath91 .
as @xmath2 decreases , it moves upwards and to the right , reaching the horizontal semiaxis @xmath92 for @xmath93 .
then , it makes a nearly circular quarter of a turn , crosses the vertical semiaxis @xmath94 for @xmath95 , and goes nearly horizontally towards @xmath96 as @xmath97 from above .
the behaviour of @xmath80 for @xmath98 is not uniquely defined . in fact , the implicit function @xmath99 defined by the condition @xmath100 presents a logarithmic branch point at @xmath101 , and the values of @xmath80 for @xmath98 depend on the chosen branch .
one can pass continuously from the values of @xmath80 for @xmath102 to those for @xmath98 if one avoids the branch point by the usual procedure of adding a small imaginary part to @xmath2 as its real part crosses the value @xmath103 .
but the resulting values of @xmath80 for @xmath98 depend on the sign of that imaginary part . since this ambiguity in the values of @xmath80 has been thoroughly discussed in ref .
@xcite , in the context of the @xmath0-zeros of @xmath15 , we do not consider necessary to pursue further .
besides those described above , @xmath1 presents an infinite set of zeros near and at the left of the vertical semiaxis @xmath104 for every integer @xmath105 @xcite . as @xmath2 increases , they approach that semiaxis , cross it when @xmath106 , and move nearly horizontally to the right , going to infinity as @xmath107 , which is a logarithmic branch point for the position of the zeros as a function of @xmath2 , as it was shown in refs . @xcite for the zeros of @xmath15 .
the zeros make a discontinuous jump of @xmath108 , according to how the branch point is circumvented , move nearly horizontally to the left , reach the vertical semiaxis @xmath104 when @xmath109 , and go to the left , to end , for @xmath110 , at positions intermediate between those that they occupied for @xmath105 . a figure illustrating that behaviour can by obtained by rotating fig . 1 in ref .
@xcite by an angle of @xmath88 .
one of the authors ( ef ) acknowledges and thanks support received from conselho nacional de desenvolvimento cientfico e tecnolgico ( cnpq , brazil ) .
the other author ( js ) is grateful to dr .
yik - man chiang for illuminating discussions and acknowledges financial aid of comisin interministerial de ciencia y tecnologa and of diputacin general de aragn .
k. v. leung and s. s. ghaderpanah , an application of the finite element approximation method to find the complex zeros of the modified bessel function @xmath66 , math . comp .
33 ( 1979 ) 12991306 .
notice that a factor @xmath111 is lacking in the values given in this reference for the imaginary parts of the first zero of @xmath112 , the second one of @xmath113 , the third one of @xmath114 , the fourth one of @xmath115 , and the fifth one of @xmath116 .
d. w. lozier and f. w. j. olver , numerical evaluation of special functions , in : w. gautschi , ed . ,
mathematics of computation 19431993 : a half - century of computational mathematics , proc .
symposia in applied mathematics , vol 48 , american mathematical society , providence , ri , 1994 , pp . | the @xmath0-zeros of the modified bessel function of the third kind @xmath1 , also known as modified hankel function or macdonald function , are considered for arbitrary complex values of the order @xmath2 .
approximate expressions for the zeros , applicable in the cases of very small or very large @xmath3 , are given .
the behaviour of the zeros for varying @xmath3 or @xmath4 , obtained numerically , is illustrated by means of some graphics .
macdonald function , modified bessel function of the third kind , hankel function , zeros _ corresponding author : _ = javier sesma + _ address : _ departamento de fisica teorica , + facultad de ciencias , + 50009 zaragoza , spain .
+ _ phone : _ 34 - 976 761 265 + _ fax : _ 34 - 976 761 264 + _ e - mail : _ |
much recent activity in lattice qcd has focused on the determination of low - energy constants ( lecs ) , which parameterise non - perturbative physics in an effective description at low energies .
a promising approach involves computing in the @xmath2regime , the kinematical region of arbitrarily small quark masses @xmath3 in a finite volume @xmath4 such that @xmath5 . here
discretisations that preserve chiral symmetry at non - zero lattice spacing should be used so that small masses and momenta can be reached safely .
however , simulations in the @xmath2-regime are expensive : owing to spontaneous chiral symmetry breaking the spectrum of the dirac operator becomes arbitrarily dense near the origin , and thus the operator is ill - conditioned .
furthermore , one observes large statistical fluctuations ( `` spikes '' ) in correlation functions of local currents @xcite for small masses .
a number of efficient numerical tools have been developed @xcite , in which the low modes of the dirac operator are treated exactly . in ref.@xcite
we proposed a method , dubbed `` low - mode averaging '' ( lma ) , which was shown to eliminate the spikes in two - point correlation functions for quark masses @xmath6 . here
we report on the extension of lma to three - point correlation functions , in order to study non - leptonic kaon decays .
we consider the massive neuberger operator @xcite d_m = ( 1-m)d+m,=a/(1+s ) , where @xmath7 is a free parameter and the massless neuberger operator , @xmath8 , satisfies the ginsparg - wilson relation .
we are interested in correlation functions of the left - handed flavour current [ j_]_=(__p_- _ ) , p_=(1_5 ) , where @xmath9 , and @xmath10 are flavour indices . using
the left - handed current has the advantage that zero modes of the dirac operator are projected out in correlation functions .
however , modes corresponding to small non - zero eigenvalues can still cause large statistical fluctuations .
let @xmath11 denote the quark propagator from @xmath12 to @xmath13 .
lma works by separating @xmath14 low - lying modes and using the spectral representation of the propagator . in this way
the quark propagator is written as s(x , y)= _ k=1^n_low + s^h(x , y ) , [ eq_sxy ] where @xmath15 is the propagator in the orthogonal complement of the subspace spanned by the @xmath16 lowest modes .
the vector @xmath17 is given by e_k = p_u_k+ p_-d p_u_k , where @xmath18 denotes the chirality of the sector containing zero modes ( if any ) , and @xmath19 is an approximate eigenmode of @xmath20 : p_d_m^d_m p_u_k = _ k u_k + r_k , ( u_k , r_i)=0 . after inserting the rhs . of into the expression for the correlator @xmath21 ,
one picks up three contributions : c(t ) = c^ll(t)+c^hl(t)+c^hh(t ) .
as explained in @xcite , the statistical signal for the correlator is enhanced by exploiting translational invariance in @xmath22 and @xmath23 , so that these contributions are sampled over many different source points .
applications of lma to the calculation of @xmath24 in the @xmath2-regime can be found in @xcite .
in a recent paper @xcite we proposed a strategy to investigate the origins of the @xmath1 rule in @xmath25 decays , i.e. the observed enhancement of the transition amplitude for a two - pion final state with isospin 0 , @xmath26 .
the goal of our programme is to quantify separately the contributions to the @xmath1 rule from physics at the charm quark mass scale and from `` intrinsic '' qcd effects due to soft - gluon exchange .
a crucial part of our strategy is to keep an active charm quark , so that the theory has an approximate @xmath27 chiral symmetry .
the @xmath0 effective weak hamiltonian after the operator product expansion contains the four - quark operators @xmath28 and @xmath29 which are given by & & _ 1^= ( _ p_- ) ( _ p_- ) ( _ p_- ) ( _ p_- ) + & & - ( uc ) .
[ eq_q1pm_def ] @xmath30 and @xmath31 are singlets under @xmath32 and transform under irreducible representations of @xmath33 of dimensions 84 and 20 , respectively . in chiral perturbation theory ( chpt ) , the ratio of amplitudes @xmath34 is , at leading order , related to a ratio of lecs via = ( + ) . here
, the lecs @xmath35 multiply the counterparts of the four - quark operators @xmath36 in the low - energy theory .
they can be determined by studying suitable correlation functions of @xmath36 in qcd and matching them to the corresponding expressions in chpt .
the fact that in the @xmath2-regime this matching can be performed at nlo without the presence of additional interaction terms with unknown coefficients @xcite , makes this an attractive setting for carrying out our programme .
the intrinsic qcd contributions to the enhancement can be isolated by determining @xmath37 in the theory with @xmath38 . then , by studying the ratio of lecs as @xmath39 departs from the mass - degenerate limit , the specific contribution of the charm quark to the @xmath1 rule can be investigated in detail . the complicated renormalisation and mixing patterns of four - fermion operators usually encountered in lattice formulations can be avoided through the use of ginsparg - wilson fermions .
indeed , the use of the modified fields @xmath40 in guarantees that the renormalisation and mixing of @xmath36 are like in the continuum theory @xcite . in particular , no mixings with lower - dimensional operators can occur .
for full details we refer the reader to ref .
an investigation of the effects of the charm quark based on chpt has been published in @xcite .
we now report on our attempts to extract the ratio @xmath37 in the su(4)-symmetric theory . in this case
only `` figure-8''-graphs must be considered , since `` eye''-graphs cancel exactly for @xmath41 .
we define the following three - point functions of @xmath36 and the left - handed flavour current : c_1^(x_0,y_0 ) = _ , _ du [ _ 1^(0 ) ] [ j_0(y)]_us .
the relation between @xmath37 and the ratio of correlators is given by h(x_0,y_0 ) = .
|_ren , [ eq_g1pm_r2 ] where the factor @xmath42 has been computed in chpt in the @xmath2-regime at nlo @xcite .
the factors @xmath43 are wilson coefficients , and it is assumed that @xmath44 is renormalised in some scheme
. from one sees that the ratio of correlators is directly proportional to the ratio of lecs .
our main results were obtained from 638 configurations on a lattice of size @xmath45 at @xmath46 , using the 5 lowest modes in the implementation of lma for figure-8 graphs . in fig.[fig_20over84 ] we plot the ratio @xmath44 as a function of the quark mass in units of @xmath47 .
it is seen that @xmath44 increases as the quark mass is tuned towards smaller values , which , at first sight , might be interpreted as an enhancement in the ratio of lecs from qcd effects alone .
however , the quark masses used in the simulations are too heavy to lie inside the @xmath2-regime , so that the computed nlo matching factor @xmath42 can not be applied to yield @xmath37 . for smaller masses
the statistical signal for the ratio of three - point functions was lost even with lma . in order to demonstrate the feasibility of our strategy one has to find ways of obtaining a strong signal in the @xmath2-regime .
we have thus investigated the quality of the signal for three - point functions using lma with a larger number of low modes @xcite .
preliminary results on a lattice of size @xmath48 at @xmath49 are shown in fig .
[ fig_boot ] , where we have plotted the bootstrap distributions for the ratio of three - point functions @xmath44 .
the plot shows that using @xmath50 hardly improves the signal at all , while for @xmath51 the distribution is much narrower , so that the statistical signal is indeed enhanced .
we conclude that lma leads to an improvement also for three - point functions , but only if @xmath16 is increased relative to the previously studied case of two - point correlators .
obviously , using larger values of @xmath16 leads to a computational overhead , which , however , is outweighed by far through the observed enhancement of the signal .
we are currently running on larger spatial volumes in order to corroborate these findings .
* acknowledgements : * simulations were performed on pc clusters at desy hamburg , lrz munich and university of valencia .
l.g . was supported by the eu under contract hprn - ct-2002 - 00311 ( euridice ) .
p.h . was supported by the cicyt ( fpa2002 - 00162 ) and by the generalitat valenciana ( ctidia/2002/5 ) .
w. bietenholz et al . , jhep 0402 ( 2004 ) 023 l. giusti et al . , jhep 0404 ( 2004 ) 013 l. giusti , c. hoelbling , m. lscher & h. wittig , comput .
commun . 153
( 2003 ) 31 t. degrand and s. schaefer , comput .
159 ( 2004 ) 185 h. neuberger , phys .
lett . b417
( 1998 ) 141 ; _ ibid . _
b427 ( 1998 ) 353 l. giusti , p. hernndez , m. laine , p. weisz and h. wittig , hep - lat/0407007 p. hernndez and m. laine , hep - ph/0407086 l. giusti , p. hernndez , c. pena , p. weisz , j. wennekers & h. wittig , in preparation . | we report on recent work on the determination of low - energy constants describing @xmath0 weak transitions , in order to investigate the origins of the @xmath1 rule .
we focus on numerical techniques designed to enhance the statistical signal in three - point correlation functions computed with overlap fermions near the chiral limit . |
is a space telescope mission concept aimed at measuring the time variation in the equation of state of the dark energy responsible for the accelerating expansion of the universe .
initially focused on the use of type ia supernovae as a means of measuring the expansion history of the universe through the luminosity - distance relation .
however , the instrument suite proved to be quite generic , allowing , for example , the inclusion of gravitational weak lensing as another means of constraining the properties of dark energy .
indeed , s very powerful wide - field imager and the interest in using it for complementary science has in many ways motivated and shaped this `` wide - field imaging from space '' conference .
before looking at the exciting discoveries that holds for the future , it is worthwhile to consider recent past and on - going work to constrain the nature of dark energy . in the six years since the discovery of the accelerating expansion of the universe a wealth of new observations have confirmed and further refined the initial discovery . in concert with newer data from type ia supernovae ( including constraints on flatness from the cmb and on from large - scale structure and clusters ) , the statistical significance of the acceleration is now @xmath4 .
just as importantly , the equation of state of the dark energy is now constrained to be @xmath5 ( assuming the equation of state is constant ) .
space - based observations have played an increasing role in cosmological measurements of type ia supernovae .
this is simply due to the much smaller sky noise present in a point - source photometry aperture in space compared with the ground especially at redder optical wavelengths .
the discovery papers @xcite contained just four supernovae measured with . samples totaling 39 type ia supernovae , segregated into early , middle and late host - galaxy morphologies from an snapshot survey , were shown to support an accelerating expansion @xcite .
precise corrections for dust extinction of eleven more supernovae made possible by accurate photometry were shown to support the accelerating expansion and provided tighter constraints on the dark energy equation of state @xcite .
a sample of high - redshift supernovae reaching @xmath6 has shown that the supernovae follow the transition from deceleration to acceleration expected as the universe goes from matter domination to dark energy domination @xcite .
these space - based measurements have significantly bolstered both the case for dark energy and the efficacy of type ia supernovae as distance indicators .
accordingly , today the most precise and highest - redshift supernova measurements several dozen so far are most often obtained with .
an example of the state of the art in space - based supernova cosmology is a joint endeavor involving the supernova cosmology project ( scp ) and the higher-@xmath7 supernova search team ( hzsst ) which around the time of this conference was conducting a search spanning four epochs , each covering fifteen pointings covering the northern goods field .
the searches are separated by 45 days roughly equal to the time from explosion to maximum light in the restframe of @xmath8 type ia supernovae .
the search uses the advanced camera for surveys ( acs ) with the 850 nm long - pass filter in order to reach redward of the restframe uv cut - off exhibited by type ia supernovae . a short exposure in the f775w filter
is added in order to reject some fraction of type ii supernovae , which in the first 10 days have much more uv flux than type ia supernovae .
the depth of the goods field at optical ( ) and infrared ( ground - based ) wavelengths means that photometric redshifts are known for many supernova host galaxies .
in addition , the galaxy colors at the location of the supernovae provide a handle on whether or not star - formation is likely to be ongoing , which can help eliminate likely core - collapse supernovae .
all of this advance information is used to help select one or two likely type ia supernovae suitable for follow - up .
each team has three target - of - opportunity triggers and 60 orbits for near - infrared follow - up observations with nicmos and slitless spectroscopy with acs . for these searches each supernova lightcurve and its associated search costs roughly 30 orbits .
at the time of this conference the scp had netted a probable @xmath9 type ia supernova from the first of the four search epochs .
the discovery image is show in figure 2 .
since then , several more supernova have been found as part of this program .
unfortunately , all of these supernovae have just the bare minimum of data needed to place them on the hubble diagram , leaving little extra information for testing for systematic effects in the measurement . a satellite such as will increase the number of such supernovae by orders of magnitude , but will also have much higher data quality and systematics controls .
while type ia supernovae have shown tremendous value in revealing the presence of dark energy and exploring its nature , we should be concerned that this purely empirical technique will reach a systematics floor .
the single - parameter light - curve width correction method ( plus extinction correction ) reduces the scatter in type ia supernovae at peak to somewhere in the range 0.100.15 mag @xcite .
however , there is nothing special about the physical conditions in a supernova when it appears its brightest . indeed a newly developed technique , cmagic , standardizes type ia supernovae using the color and magnitude roughly 15 days after maximum , when the supernova reaches a color of @xmath10-@xmath11=0.6 @xcite .
this method has been shown to reduce the scatter to below 0.10 mag .
thus , it is possible that type ia supernovae are more homogeneous at a given color rather than at a given lightcurve epoch .
many more observations of nearby supernovae , focusing on accurate colors with well - determined uncertainties will be a great boon in improving the standardization of type ia supernovae . by lowering the remaining scatter ,
stronger limits are also set on any possible remaining systematics in the type ia methodology .
the objective of the mission is to provide the best possible constraint on the time - evolution of the dark energy equation of state .
the only demonstrated experimental techniques are astronomical , and of these , the luminosity distance from type ia supernovae is by far the most mature .
therefore , it was natural to design a mission focused on supernovae , but as we shall see , the resulting design is sufficiently general to perform exceptionally well for gravitational weak lensing .
pushing the type ia supernova method to its limit requires thousands of supernovae spanning a wide range in redshift .
the present design targets 2000 normal type ia supernovae from redshift 0.1 to 1.7 .
it is extremely critical that the photometry be of high quality continuously from low to high redshift , not only to realize the statistical power of the sample but to avoid introduction of systematics errors @xcite . due to the increasing brightness of the terrestrial night sky at redder wavelengths where
higher redshift type ia supernovae must be observed , coupled with increasing atmospheric water absorption at redder wavelengths , the necessary data quality can be obtained only from space .
the broad redshift coverage needed to control systematics further requires detectors at both optical and infrared wavelengths .
telescope optical design and spacecraft size and mass constraints point to a dedicated 2-m telescope with a field of view several hundred times larger than that possessed by , populated with visible and near - infrared detectors .
further , in order to determine the supernova type , and possible sub - class , a sensitive low - resolution spectrograph is required .
systematics control enabled by spectrophotometric spectra , as well as allowance for spacecraft pointing margin , call for an integral field unit spectrograph .
gravitational weak lensing measurements have become increasingly robust over the past decade .
long viewed as a means of constraining , more detailed theoretical studies several developed within the collaboration indicate that precision weak lensing measurements have the capability to place constraints on the dark energy equation of state similar to those possible with type ia supernovae .
in particular , weak lensing cross - correlation cosmography measuring the association between foreground template mass distributions and background weak lensing shear using galaxies located using photometric redshifts appears to be the most powerful weak lensing approach @xcite .
atmospheric turbulence results in poor spatial resolution for high - redshift galaxies observed from the ground .
furthermore , ground - based telescope and camera optics are far from being diffraction - limited .
so great care must be taken in removing the effects of seeing , tracking errors , and optical aberrations .
stars provide the reference for removing these effects , but the number of usable stars is limited , due to both the limited size of the isoplanatic patch within which seeing is correlated and difficulty in distinguishing the few stars from the large number of galaxies at faint magnitudes . errors in these corrections set a limit on the quality that ground - based weak lensing will be able to achieve . in comparison ,
space - based wide - field observations resolve many more sources and the remaining corrections for tracking and telescope aberrations are much smaller relative to the sizes of the sources .
figure 4 illustrates the dramatic gain in the number of resolved galaxies in going from ground- to space - based observations .
moreover , in contrast to , will be placed in a stable thermal environment so that any corrections can be determined using many spatially uncorrelated images . moreover
, the deep optical and infrared photometry from will provide high - quality photometric redshifts as needed to implement the cross - correlation cosmography method .
therefore , can provide excellent weak lensing data for constraining changes in the dark energy equation of state @xcite .
in detail , the telescope design is a three - mirror anastigmat with a 2-m primary mirror @xcite .
the optics are silver - coated to maximize the optical and near - infrared throughput and minimize the thermal emissivity .
the imager consists of 0.34 sq .
degrees paved with red - sensitive ccd s and a scale of 0.10 /pixel , and 0.34 sq .
degrees paved with hgcdte near - infrared detectors with a scale of 0.18 /pixel @xcite .
the sizes of the detectors are identical and those made of hgcdte have a long - wavelength cutoff which reduces the dark current ( and sensitivity to thermal emission ) .
these features allow the ccd and hgcdte detectors to be mounted in the same focal plane , simplifying the imager construction , mounting , and alignment .
each detector has its own filter , so a large filter wheel which would otherwise be needed is avoided .
detailed simulations show that photometry of sufficient quality and redshift range is not possible from the ground even under optimistic assumptions .
the spectrograph has a prism as its dispersing element , providing resolution @xmath12 across optical and near - infrared wavelengths @xcite .
higher resolution is unnecessary due to the velocity broadening intrinsic to the spectra of type ia supernovae .
again , the optical detector is a red - sensitive ccd while the near - infrared detector is hgcdte with a long - wavelength cut - off near 1.7 @xmath13 m .
the spectrograph is fed by an image slicer integral field unit .
the field is 6@xmath143 , wth the latter dimension being divided into 40 slices .
detailed simulations show quite dramatically that spectroscopy of comparable quality is not possible from the ground , even with a thirty meter telescope with adaptive optics .
.nominal parameters of the surveys [ cols="<,^,^,^,^ " , ]
the nature of dark energy is a fundamental question , which astronomical observations can address .
type ia supernova and weak lensing are the most powerful techniques yet developed for studying dark energy .
reaching the requisite statistical and systematics accuracy with these techniques requires a widefield imager in space
. represents a very advanced concept for such a widefield imager , having an instrument suite which is both powerful and versatile .
table 1 details this power for the type i supernova survey , a nominal wide - field weak lensing survey , and a possible panoramic weak lensing survey . the planned dark energy program for the joint dark energy mission ( ) will produce a treasure trove of data for archival study .
guest observer programs that could further exploit such a powerful instrument are envisioned in the concept .
the following conference presentations describe many of the scientific lines of inquiry which would be greatly advanced with a facility such as . | the accelerating expansion of the universe is one of the most surprising and potentially profound discoveries of modern cosmology . measuring the acceleration
well enough to meaningfully constrain interesting physical models requires improvements an order of magnitude beyond on - going and near - term experiments .
the supernova / acceleration probe has been conceived as a powerful yet simple experiment to use type ia supernovae and weak gravitational lensing to reach this level of accuracy . as fundamentally different causes for the acceleration map into very small differences in observational parameters for all relevant cosmological methods , control of systematics is especially important and so has been built into the mission design from the very beginning .
though focused on the study of the accelerating universe , the overall instrument suite is quite general and able to make unique contributions to a wide variety of astronomical studies .
the baseline satellite consists of a 2-m anastigmat telescope , with a 0.7 square degree focal plane paved with optical and nir imaging arrays .
spectroscopy can be obtained using a high - throughput low - resolution optical+nir integral field spectrograph .
the baseline science programs will result in a 15 square degree `` deep field '' having temporal coverage every 4 days and summing to @xmath0 in all colors to be used for discovery and follow - up of some 2000 type ia supernova in the range @xmath1 and a wide area survey spanning 1000 square degrees and reaching @xmath2 in all colors to be used to measure the weak lensing power spectrum well into the non - linear regime .
a panoramic survey covering 10000 square degrees to @xmath3 in all colors is also possible .
this baseline dataset represents a gold mine for archival astronomical research and follow - up with jwst , while guest observer survey programs will substantially broaden the impact that will have . |
a locally self - interacting random walk on the set of integers is a sequence of integer - valued random variables @xmath2 with @xmath3 for all @xmath4 , that can be defined inductively as follows : one initializes a `` local time profile '' by first choosing some function @xmath5 that associates to each edge @xmath6 of the lattice @xmath7 a real number @xmath8 ( the set @xmath9 of unoriented edges can be identified to the set @xmath10 or to the set of couples @xmath11 where @xmath12 ) and one chooses the starting point @xmath13 .
the simplest choice is of course to set this initial function @xmath5 to be equal to zero on all edges , and to start at the origin ( i.e. , @xmath14 ) with probability one .
the law of the walker is then described via some function @xmath15 from @xmath16 to @xmath17 $ ] .
this function describes the probability of jumping to the right in terms of the local time profile `` seen from the walker at its current position '' .
more precisely , one defines inductively , for each @xmath4 , @xmath18 to be equal to either @xmath19 or @xmath20 in such a way that @xmath21 where @xmath22 is the local time profile seen from the particle at time @xmath23 . at each step ,
one updates this local time profile as follows : @xmath24 denotes the edge of the @xmath25-th jump and @xmath26 is the new local time profile .
when @xmath5 is identically zero , then @xmath27 denotes the number of times the walk has jumped on the edge @xmath6 before time @xmath23 .
this procedure clearly describes completely the law of the family @xmath28 .
similar definitions work of course on other graphs than @xmath7 ( for instance on @xmath29 ) . in the present paper
, we will however restrict our discussion to the one - dimensional case .
one can also use local times on sites instead of edges but the latter case turns out to be often easier to handle . depending on the function @xmath15 ,
the walker can be of self - repelling type ( it prefers to go into the direction it has less visited in the past ) , self - attracting , or more complicated ( it can be for instance repelled by the visits to the neighbouring edge , but attracted by its past visits to further edges ) .
when the function @xmath15 is `` cylindric '' i.e. when @xmath30 is a function on the values @xmath31 for a finite set of edges @xmath32 , the interaction can be described as `` local '' . in other words , for some constant @xmath33 , the past trajectory @xmath34 influences the outcome of @xmath35 only via the values of @xmath27 for the edges in the @xmath33-neighbourhood of @xmath36 .
note that in this cylindric case , if there is a large scale limit to the process @xmath2 its evolution will be `` purely '' local as the finite - range becomes scaled down .
such locally self - interacting walks provide a particular class of random evolutions of functions as the sequence of functions @xmath37 form a markov process .
it is in fact very natural to study the evolutions where the function @xmath15 is also height - invariant i.e. the particular case where @xmath38 for all @xmath39 and all real @xmath40 . in other words ,
the walk is sensitive to the local discrete gradient of @xmath41 but not to its actual height .
this excludes for instance the family of `` linear '' reinforcements ( see for instance @xcite for a survey of these other types of reinforcements ) .
another way to view these evolutions is to define , for all @xmath12 and @xmath4 , the local gradient @xmath42 and to say that @xmath43 is in fact a function @xmath44 of @xmath45 . in @xcite , a case where @xmath46 was just a function of @xmath47
was studied . loosely speaking , this
walk ( sometimes referred to as true self - avoiding walk or true self - repelling walk tsrw ) is driven by the negative gradient of the local time at its actual position i.e. , the walk prefers to go in the direction it has visited less often in the past .
it turns out that a non - trivial scaling shows up , as @xmath48 converges in law to some non - trivial random variable .
in fact , the entire limiting process can be described @xcite ( it is called the true self - repelling motion tsrm ) .
it corresponds to a random and local evolution of functions , that is neither a deterministic partial differential equation nor constructed via a white - noise driven stochastic partial differential equation .
this raises the questions whether other such processes exist , whether the previous scaling limit is `` stable '' ( do all similar discrete models have the same limiting behaviour ? what properties of the discrete model is actually captured by this scaling limit ? ) , and provides a motivation to study self - interacting random walks associated to more general functions @xmath15 .
the family of models that we will focus on in the present paper are those where the self - interaction is governed by a linear combination of the values of the local time .
in other words , we let @xmath49 where all but finitely many @xmath50 s are zero , and we choose the local `` drift '' at step @xmath23 according to the value of @xmath51 .
the fact that this evolution in height - invariant corresponds to the condition @xmath52 then , we take @xmath53 for some increasing function @xmath54 such that @xmath55 . in other words , if @xmath56 is ( very ) positive ( respectively , negative ) , then the walk will tend to jump to the right ( respectively , to the left ) , so that @xmath56 indeed plays the role of a local drift .
a particularly natural choice for @xmath54 turns out to be @xmath57 in order to start investigating the variety of possible behaviours , one can for instance start to look at the case where @xmath58 is a function of @xmath59 , @xmath60 , @xmath61 , @xmath62 .
a first natural choice is to pick two real numbers @xmath63 and @xmath64 , and to drive the walker by the linear combination @xmath65 note that these are the choices such that both height - invariance ( because the sum of the terms @xmath66 is zero ) and _ left - right symmetry _ ( the walks @xmath67 and @xmath68 have the same law because the coefficients in front of @xmath69 and @xmath70 are opposite ) hold .
note that when @xmath64 is positive and @xmath71 , this is the previously mentioned tsrw with tsrm as scaling limit .
when @xmath72 and @xmath73 is very small , one can therefore view this process as a perturbation of the tsrw . at the other end of the picture ,
when @xmath64 is negative , it is easy to check that the walk will ( with positive probability ) eventually jump back and forth on one single edge ( the more it jumps on it , the more it is attracted by it ) which is not a particularly interesting case . in section [ s2 ] , we will describe various possible asymptotic behaviours , depending on the values of @xmath63 and @xmath64 .
we shall see that intuition can sometimes be misleading ; repellance by second neighbours does not always have the same effect as repellance by immediate neighbours as it can create traps .
one interesting feature that we will mention is the existence ( when @xmath74 ) of a phase transition between stuck walks ( that can eventually remain in a finite interval ) and the non - stuck ones ( with almost surely an infinite range ) , depending on whether @xmath75 or not .
a second natural possibility is to break the left - right symmetry , and to drive the walker by a linear combination such as @xmath76 which will be among those studied in section [ s3 ] ( with deterministic transient behaviour of the walker ) .
section [ s2 ] is more of survey - type presenting results proved in @xcite , as well as heuristics and conjectures , while section [ s3 ] , will contain original results with actual proofs . throughout the paper
, we will mention results that hold with positive probability , like `` with positive probability , the walk will eventually remain stuck on a single edge '' .
clearly , one then implicitly conjectures that this is the generic behaviour of the walk ( i.e. that the corresponding asymptotic event will hold almost surely ) , but in general , proving such almost sure results for self - interacting walks is not a straightforward task ( and it requires different arguments ) , see for instance @xcite for the linearly reinforced walk . we will not focus on these questions in the present paper .
we would like to emphasize here that ( despite the `` phase diagram '' that we will draw at the end of the next section ) the aim of the present paper is not to give a full classification of the possible asymptotic behaviour of all these cylindric self - interacting walks .
our goal is rather to point out certain phenomenological features , in order to shed some light on how sensitive ( or not ) these walks are with respect to their microscopic definition , and what feature is kept in the scaling limit . * a warm - up .
* as a warm - up to maybe help ( if it does not help the reader , then she / he should not hesitate to skip this paragraph as it will not be essential later on ) the following analysis , let us make some non - rigorous analogy with fibonacci - type sequences and their related polynomials . recall that our models are `` driven '' by a linear combination of the type @xmath77 where @xmath78 .
if a walk is in some local equilibrium state ( it does not `` shoot in one direction '' , it is somewhere in the middle of its past range , and the values of @xmath41 are all very large ) , then one could expect that the value of @xmath56 is rather small at most of the sites ( at least compared to the total number of steps ) .
after renormalization by the number of steps , this would indicate that @xmath79 i.e. that @xmath80 follows some fibonacci - type progression .
as we know , the behaviour of such progressions is controlled by the values ( and position with respect to the unit circle ) of the roots of the corresponding polynomial @xmath81 where @xmath82 the fact that the dynamics is height - invariant corresponds to the case where @xmath83 is a root of @xmath84 .
hence , it is in a way more appropriate to describe the self - interaction by its polynomial ( in the case where the self - interaction is left - right symmetric , it does indeed characterize the self - interaction ) , and we will use this when we will illustrate some of our simulations . note that the function @xmath85 , defined for @xmath86 is actually the fourier transform of the sequence @xmath87 .
thus , the qualitative behaviour of the walk will depend on the type of the fibonacci polynomial @xmath81 , or , what is equivalent of the fourier transform of the sequence of coefficients @xmath87 .
e.g. , positive definiteness of this sequence plays an essential role .
during most of this section , the setup is the one described in the introduction , where the walk is `` driven '' by @xmath88 , i.e. when we are considering the next - to - neighbouring interactions that are both height - invariant and left - right symmetric ( we will comment on interactions with larger range at the end of the section ) .
the meaning of @xmath63 and @xmath64 is as follows : * if @xmath64 is positive ( respectively , negative ) , then the walk is repelled ( respectively , attracted ) by its previous visits to the neighbouring edge . * @xmath63 plays the same role as @xmath64 but for the next - to - neigboring edges . *
@xmath73 and @xmath89 describe the intensity of these self - interactions .
as we shall see , in most of these cases , the qualitative asymptotic behaviour of the walk will depend on the signs of @xmath63 and @xmath64 , and on the ratio between @xmath73 and @xmath89 .
we can divide the parameter space into several classes .
let us first notice that in the case where @xmath90 , then with positive probability , the walk @xmath36 will forever jump back and forth along one singe edge ( between @xmath91 and @xmath83 , say ) .
this is easily checked using the borel - cantelli lemma .
similarly , when @xmath92 and @xmath63 is negative , then with positive probability , the walk will be stuck on a set of two edges ( say , it will visit only the three sites @xmath91 , @xmath83 and @xmath93 ) .
these are the _
purely self - attractive cases_. let us first very briefly recall here the case where @xmath94 and @xmath71 : the walker is then driven by the negative gradient of the local time at its current position . as we have already mentioned , this is the so - called true self - repelling walk tsrw steps is consistent with the distribution of the first @xmath96 steps and this is not always the case for measures on self - avoiding or self - repelling walks ; the term `` myopic '' random walk has also been used . ] .
this scaling limit was clarified in @xcite and @xcite ; it is the true self - repelling motion ( trsm ) . , @xmath97 , @xmath98.,height=192 ]
it is quite natural to expect that this tsrw scaling behaviour is stable under perturbation , i.e. when choosing @xmath99 , but sufficiently small ( in absolute value ) .
there are some good reasons to guess that in the parameter range @xmath100 ( and for @xmath94 ) , the long time asymptotic behaviour and scaling limit of the walk is similar to the tsrw case . indeed , besides numerical evidence , one can note that : 1 . in this parameter range
, we can identify a stationary and ergodic distribution of the `` environment seen by the random walker '' process .
this distribution is qualitatively similar to the tsrw case , the difference being that instead of a product measure we get an exponentially mixing gibbs measure . for details ,
see subsection [ sstationary ] below . 2 . in this stationary and ergodic regime , using the variational method initiated in @xcite and exploited in @xcite for similar but continuous space - time models
, one could very likely obtain superdiffusive lower bounds of the type @xmath101 for the asymptotic variance of the displacement of the random walker . for details of similar computations in continuous space - time , see @xcite . in this regime
, the competition between the nearest edge self - repellence due to @xmath102 and next - to - nearest edge self - attractiveness due to @xmath104 is won by the latter because @xmath73 is large enough , and the random walker is eventually trapped in an interval of finite length .
the length of the trapping interval increases to infinity as the ratio @xmath105 decreases to the critical value @xmath106 .
more precisely , let us define for all positive integer @xmath107 , @xmath108 this is an increasing sequence ( @xmath109 , @xmath110 , @xmath111 and @xmath112 ) . in @xcite ,
the following theorem is proved : [ thm_sticky ] suppose that @xmath113 .
* if @xmath114 for some @xmath115 , then with positive probability , the walk remains stuck on a set of @xmath116 consecutive sites ( and visits all these sites infinitely often ) . * if @xmath117 , then almost surely , the walk does not get stuck on a set with less than @xmath116 sites . for instance , for some values of the parameter @xmath73 larger but very close to @xmath118 , the walk can remain stuck on a set of exactly @xmath119 sites . on the other hand , when @xmath120 , then the walk does almost surely not get stuck on a finite set . , @xmath121 ) stuck on 11 sites : the trajectory and the local time , title="fig:",height=163 ] , @xmath121 ) stuck on 11 sites : the trajectory and the local time , title="fig:",height=163 ] for proof of theorem [ thm_sticky ] and more refined details on the asymptotic behaviour of the walk in this particular trapping regime see @xcite .
in this regime , nearest and next - to - nearest edges are all repelling .
nevertheless these effects do not just always add up to produce a tsrm - type asymptotic behaviour .
the second neighbours can produce traps that are difficult to get out from . in order to illustrate this ,
let us describe a scenario of building up traps , for the extreme case where @xmath92 and @xmath124 : the random walk starts by jumping back and forth on the edge @xmath125 between @xmath91 and @xmath83 a large even number @xmath126 of times .
this happens with probability @xmath127 .
then , in some @xmath128 number of steps it moves to site @xmath129 .
( note , that due to the large number of visits at edge @xmath125 the walker does not have a chance to return from site @xmath130 . )
now , it performs yet again a number of @xmath131 back and forth jumps on the edge @xmath132 ( between sites @xmath129 and @xmath133 ) .
this happens with probability @xmath134 .
after all this , in @xmath128 number of steps the walker lands on edge @xmath135 ( between sites @xmath136 an @xmath130 ) .
now , due to the @xmath131 number of visits at edges @xmath125 and @xmath132 the walker does not have a chance to get away from edge @xmath135 before jumping @xmath137 times back and forth on this edge .
hence , we see that while it is very unlikely to build up a trap ( and it therefore typically takes a long time to build up one ) , once it is built and if @xmath138 is large enough , then it takes even much longer to escape from it .
in fact , the previously described strategy to build such traps is certainly not the optimal one , and the probability to create one is in fact smaller , so that for all positive @xmath63 , it takes much longer to escape from the trap than to create it .
this slowing - down mechanism is reminiscent of that of random walks in random environment , and it indicates that there is no scaling limit process with continuous trajectories . and @xmath92 : the trajectory and the local time , title="fig:",height=163 ] and @xmath92 : the trajectory and the local time , title="fig:",height=163 ]
when @xmath64 is positive , we do not have precise results at this point for the two borderline cases where @xmath139 and @xmath140 .
* @xmath141 is the borderline case between the conjectural tsrm regime ( where the walk should scale like @xmath142 ) and the sticky regime ( when the walk is eventually stuck to a finite interval ) .
we have seen that the walk does almost surely not stay in a finite interval , but its asymptotic behaviour is not really well understood .
naive scaling arguments ( and the stationary measure computations mentioned at the very end of the next section ) would at first suggest that the non - degenerate scaling of the walk should be like @xmath143 .
but it can be rigorously proved that in a stationary regime there exists a diffusive lower bound on the asymptotic variance of the displacement , and intuitively , one might guess diffusive behaviour ( the stationary local time regime see also figure [ figlt3 ] and the walk displacement might have different scaling exponents and not interact in the large - scale limit ) .
+ : the trajectory , height=192 ] * the case where @xmath140 can be interpreted as the tsrw with self - repulsion defined in terms of occupation times on sites rather than edges .
its scaling behaviour is expected to be similar to the tsrw on edges ( see @xcite for references ) .
figure [ phased ] sums up the rough phase diagram of the asymptotic behaviour of the walk depending on the values of @xmath144 on the unit circle . on the unit circle , height=288 ] we now quickly survey some computations related to the existence of stationary measures for the current environment ( i.e. , the local time fluctuations ) as seen from the random walk in the conjectural `` tsrm '' regime .
we will work in the more general context where the dynamics is driven by a height - invariant linear function of the local time i.e. where @xmath146 with @xmath147 where all but finitely many @xmath148 s are zero .
we will write @xmath149 .
it is convenient to work with an initial profile @xmath150 that belongs to the set @xmath151 it is immediate to check that this parity condition will be conserved for all @xmath152 ( the opposite parity convention i.e. @xmath153 would also be preserved ) so that @xmath154 is a markov chain on @xmath155 .
we can also restrict ourselves to the family of @xmath156 s in @xmath155 with finite norm @xmath157 .
we will use the following notation .
if @xmath158 i.e. @xmath159 jumps to the right , we write @xmath160 , and if it jumps to the left , we write @xmath161 .
in other words , for all @xmath162 , @xmath163 the markov chain is described by the fact that ( given @xmath164 ) @xmath165 is equal to @xmath166 with conditional probability @xmath167 , and that @xmath161 otherwise .
the transition operator of the markov process is therefore @xmath168 as often , it is convenient to view this markov chain as the jump chain associated to a continuous - time markov chain on the same state space .
this continuous - time markov chain @xmath169 jumps to the right with intensity @xmath170 and to the left with intensity @xmath171 .
the particular choice for the dynamics leads us naturally to look for a stationary distribution for @xmath172 of `` discrete gaussian type '' .
we are going to suppose in the present section that : * @xmath173 is even ( i.e. , @xmath174 for all @xmath162 ) .
this corresponds to the left - right symmetry mentioned in the introduction . *
the self - interaction has finite spatial range ( i.e. , all but finitely many @xmath148 s are equal to zero ) * it is positive definite in the sense that for all @xmath175 such that all of them but a finite positive number is equal to zero , @xmath176 ( this implies in particular that @xmath177 ) . note that in the case where @xmath178 and @xmath94 , this last condition precisely means that @xmath179
. under these assumptions there exists a unique gibbs measure @xmath180 on @xmath155 formally given by @xmath181 the proper , rigorous definition of the gibbs measure @xmath180 is in terms of conditional specifications on finite intervals , given fixed boundary conditions .
this gibbs measure is well understood , and its correlations decay exponentially fast with distance and the zero - one law holds for tail events . under the previous assumptions ,
the probability measure @xmath180 is stationary and ergodic for the markov process @xmath172 .
the proof is straightforward and based on the following observation : @xmath182 and similar formulae when one replaces @xmath183 by @xmath184 .
it proves that in fact the measure @xmath180 is invariant under the process with `` right - jumps '' only , and under the process with `` left - jumps '' only .
ergodicity also follows from the above observation .
the dirichlet form @xmath185 is easily computed . using the above observation we get @xmath186 hence , since @xmath187 almost surely , it follows that @xmath188 if and only if almost surely @xmath189 but it is standard fact that the above almost sure identities hold if and only if @xmath190 is tail measurable . hence , by the zero - one law mentioned before , @xmath190 is almost surely constant .
this proves ergodicity .
also , one can note that the mean waiting time of the continuous - time markov chain at @xmath156 is given by @xmath191 and that @xmath192 is finite . from this
, it follows that one can define a probability measure @xmath193 on @xmath155 by @xmath194 and that : this probability measure @xmath193 is stationary for the discrete chain @xmath195 .
hence , we see that in this regime , and under these well - chosen starting distributions , the large - scale behaviour of the local - time profile is of brownian - type .
we can note that in the borderline case where @xmath94 and @xmath141 ( i.e. the discrete `` third derivative '' case ) , the previous stationary distribution analysis can be adapted to show that the discrete gradient of @xmath172 ( that is : the discrete second derivative of the local time profile ) has a stationary measure of discrete gaussian type .
more precisely , we see in this interesting case that if we define for all @xmath196 , @xmath197 then the process @xmath198 is a markov process , and that it has a stationary distribution given by i.i.d . discrete gaussians . for details of similar computations of stationary and ergodic measures of the environment seen by the moving particle , in continuous space - time setup , and
their consequences see also @xcite .
[ figlt3 ] to conclude this section , let us very quickly list a few observations : * in the setting of the tsrw , it was natural to describe the dynamics of the walker by saying that it was driven by an approximate negative gradient of the local time at its current position , and the dynamics of its scaling limit could also be described in this way .
our previous analysis however shows that this type of description has to be handled with care .
the case @xmath199 seems at first to also correspond to a negative discrete gradient , but its large scale behaviour is completely different .
* in fact , in the general case , what seems more relevant to classify the asymptotic behaviour of the walk is the quadratic form appearing in ( [ qform ] ) or alternatively the study of the roots of @xmath200 .
this is also apparent in the study of the stuck walks in @xcite .
note also that the borderline cases correspond to ( multiples of ) the polynomials @xmath201 and @xmath202 . in the general case ( with larger interaction - range )
, the phase diagram is seemingly more complicated , with different connected components of the parameter space corresponding to similar asymptotic behaviours . * a particular role
is then played by the `` odd '' derivatives , i.e. more precisely , the polynomials @xmath203 .
then , the discrete @xmath23-th derivative of @xmath204 ( and therefore of @xmath156 ) has a stationary measure of discrete gaussian type .
let us stress that ( as opposed to the cases studied in the previous section ) , the examples studied in the present section break the left - right symmetry i.e. the law of @xmath2 and @xmath205 are different even if one chooses @xmath5 to be symmetric . in this section , we will first focus on the case where @xmath206 , where @xmath207 we then choose @xmath208 which penalizes jumps in the direction of this second derivative .
the precise form of the function @xmath15 is not really important for what follows .
for instance , our results clearly remain true if we only assume in addition of locality and height - invariance , that * @xmath209 . * @xmath43 is a function of @xmath210 goes to @xmath91 very quickly as @xmath211 .
these conditions can in fact also be relaxed .
note that @xmath212 is a discrete and symmetric version of the second derivative of the local time , but this feature is not essential in what follows .
intuitively speaking , the walk will tend to move to the right if the local time profile is locally concave and to the left if it is locally convex .
the main result of the present section is the following : suppose that the initial profile is identically zero ( again , this is not really necessary we only need this initial profile not to be too `` wild '' ) .
then : [ t2 ] with positive probability , @xmath213 as @xmath214 . on large scale and on a positive fraction of the probability space ,
the walk becomes almost deterministic , but non - ballistic , and the scaling is of brownian type .
as we shall see , the proof gives in fact a more precise description of the behaviour of @xmath2 on this event of positive probability .
let us immediately stress that the proof will in fact show that the results remain true if one replaces @xmath210 by any function @xmath215 of the type @xmath216 the facts that @xmath210 is a local version of the second derivative , that it is translation - invariant and that it is `` anti - symmetric '' are therefore not so crucial after all .
in fact , our conclusion holds true if we replace @xmath210 by any local function @xmath217 such that * for some @xmath218 , @xmath219 for all @xmath220 .
* @xmath221 * @xmath222 and @xmath223 are both negative .
this for instance includes discrete @xmath224-th derivatives for `` even '' @xmath224 s such as @xmath225 that corresponds to a discrete forth derivative . in the `` continuous limit ''
, the occupation time density profile of the motion @xmath226 has a discontinuity at its local position ( i.e. at time @xmath227 , it is equal to @xmath228}$ ] ) , and the size of this jump is growing with time .
hence , one could a priori naively expect that the walk would not slow down , but on the contrary speed up when this gap widens .
this can be explained by the fact that the self - interacting random walk therefore captures some feature of the discrete local time profile that is not visible in the continuous limit .
the coming three subsections are devoted to the proof of proposition [ t2 ] .
let us first describe in plain words a possible behaviour for our walk @xmath159 ( we shall then prove that indeed , this behaviour occurs with a positive probability ) : the walk @xmath159 starts at the origin , jumps to its right i.e. @xmath229 , and it forever remains positive i.e. @xmath230 for all positive times .
in fact , when @xmath214 , @xmath36 goes off to @xmath231 . for each @xmath4 , denote @xmath232 the past maximum of the walk . in our scenario , @xmath233 for all @xmath23 . in other words , for each @xmath23 , either @xmath36 is equal to its past maximum @xmath234 , or it is equal to @xmath235 or @xmath236 .
we see that we can therefore decompose the set of times as follows : define for each @xmath237 the hitting time @xmath238 of @xmath162 by @xmath159 , and let @xmath239 .
then , in our scenario , during each of the intervals @xmath240 , the walk jumps back and forth on the edges between @xmath241 and @xmath242 .
when it stops doing so , it first jumps to @xmath243 , and then jumps back and forth on the edges between @xmath162 and @xmath243 and so on .
imagine for a moment that this has happened for a while and that just before @xmath238 ( for a large @xmath162 ) i.e. when @xmath244 , the following loosely defined event @xmath245 is true : the walker has visited each of the edges @xmath246 , @xmath247 and @xmath248 many times , and the number of times it has visited these three edges remain however comparable ( in the sense that the difference between these three numbers of visits is small ) . in particular , immediately before it chooses to jump to @xmath162 for the first time , the value of @xmath249 is rather close to zero ( mind that at that moment @xmath250 , and that @xmath251 because the walk has not visited @xmath162 yet ) so that the probability to indeed jump from @xmath241 to @xmath162 is therefore not too small
. then , the walk arrives at @xmath162 for the first time . at that moment , the two edges to its right have not yet been visited , the edge between @xmath162 and @xmath241 ( to its left ) has been visited only once ( because the walker arrives at @xmath162 for the first time ) . on the other hand ,
the edge between @xmath241 and @xmath252 has been visited a lot of times .
hence , the walk will ( very likely ) jump back to @xmath241 .
once it is back at @xmath241 , the probability to jump to @xmath162 or to @xmath252 at that moment is neither very small nor very close to @xmath83 ( because the situation can not drastically change because of the two previous jumps ) . note that if the walk then jumps to @xmath252 , the value of @xmath210 will then be very small , so that the walk will want to jump immediately back to @xmath241 .
hence , the walk is ( with high probability ) trapped between @xmath252 and @xmath162 for a while .
how long does this happen ?
well , one should note that each time the walk jumps on @xmath246 or on @xmath253 , it will increase the chance to jump to the right next time it is at @xmath241 .
hence , after a short while , the walk will in fact only jump back and forth between @xmath241 and @xmath162 .
this will be the case until the number of times at which it has jumped on @xmath253 starts to be comparable with the number of times at which it has jumped on @xmath254 , because the walk will then have a significant chance to move to @xmath243 .
hence , we end up in a situation where @xmath255 holds .
one important point in this scenario is that in the end , the walk will tend to jump a little more often on @xmath256 than on @xmath257 .
so , by a law of large numbers type argument , the total number of visits of the edge @xmath257 will turn out to grow `` linearly '' in @xmath162 as @xmath258 .
this will a posteriori justify our assumption ( in the definition of @xmath245 ) that at the time @xmath238 , the walk has visited @xmath246 and @xmath259 many times .
we now start to turn the previous ideas into a rigorous proof .
let us define a family of independent random variables @xmath260 , where @xmath261 , @xmath115 , @xmath12 .
for each @xmath262 , @xmath107 and @xmath162 , the law of the random variable is the following : @xmath263 we can then define ( deterministically ) our random walk @xmath159 using this family of random variables as follows .
let @xmath264 and for each @xmath4 , choose inductively @xmath265 where @xmath23 is the @xmath107-th time at which @xmath266 and @xmath267 .
note that when @xmath268 , then @xmath269 is necessarily even ( because @xmath270 and @xmath271 are odd , while @xmath272 and @xmath273 are even ) .
we will for the time being forget about this coupling between @xmath159 and the @xmath274 s , and just list a few simple observations concerning this family @xmath275 of random variables .
a particular role will be played by the random variables @xmath276 when @xmath237 , @xmath277 and @xmath262 is an integer .
we denote them by @xmath278 ( and drop the subscript @xmath107 ) .
let us insist on the fact that @xmath278 are not defined for semi - integer @xmath262 s ( as opposed to the @xmath276 s ) .
let us also keep in mind that for a given @xmath107 and @xmath162 , when @xmath262 is very large , the probability that @xmath279 and @xmath280 is very close to @xmath83 .
we now define certain events : * let @xmath281 denote the event that for all @xmath162 , all @xmath282 and all @xmath283 , @xmath284 * note that almost surely , for each @xmath162 , the number of positive ( respectively , negative ) @xmath262 s for which @xmath278 is positive ( respectively , negative ) is finite ( it follows immediately from the definition that it has a finite expectation ) .
@xmath285 is the event that for all @xmath237 , @xmath286 * for all @xmath237 , let us define @xmath287 there is no problem with the definition of @xmath288 for the same reason as above .
note that because of symmetry , @xmath289 .
the law of large numbers therefore ensures that almost surely @xmath290 we define the event @xmath291 , that for all @xmath292 , @xmath293 $ ] .
a simple application of the borel - cantelli lemma enables to prove that : with positive probability , @xmath294 holds .
the proof is elementary and safely left to the reader . in order to simplify our notations
, we will assume that at time zero , the initial profile is such that @xmath295 and that @xmath296 is zero otherwise .
then , in our `` good scenario '' , the @xmath269 s will always remain even ( because @xmath270 and @xmath271 are both odd , whereas @xmath297 and @xmath273 are both even ) .
this assumption is not a problem . if we start with @xmath296 being identically equal to @xmath91 , we let @xmath159 move twice to the right , and at this time @xmath298 , the situation is exactly the previous one .
since we will prove asymptotic results about @xmath159 with positive probability , our theorem will follow immediately from the results with this particular initial profile .
recall that our walk @xmath2 is defined deterministically from the @xmath276 s as follows : @xmath14 and for each @xmath4 , @xmath299 where @xmath300 , @xmath301 and @xmath107 is the cardinal of the set @xmath302 clearly , this definition ensures that one uses never the same @xmath276 twice , and that indeed @xmath303 we are going to assume that @xmath304 holds , and we will show that our `` good scenario '' holds as well . note that there is nothing probabilistic in the following arguments . for this , we will inductively prove that for each @xmath237 , certain events @xmath305 hold .
define for each @xmath306 , the stopping time @xmath307 as the first time at which the edge @xmath253 has been crossed more than @xmath308 times and the walk is at @xmath241 .
we say that the event @xmath305 holds if the following four conditions hold : 1 .
the walk did not visit @xmath242 before @xmath309 .
2 . on @xmath310 $ ] , the walk did only visit the three sites @xmath311 and @xmath162 .
3 . the last @xmath312 jumps of the walk before @xmath309 were all on the edge between @xmath241 and @xmath162 .
4 . at the time @xmath313 , the two quantities @xmath314 and @xmath315 differ by not more than @xmath316 , and they are both larger than @xmath317 and smaller than @xmath318 . , width=288 ]
assume that @xmath319 , @xmath320 , and that @xmath321 do hold ( and that @xmath322 hold as well ) . for simplicity ,
let us first assume that @xmath323 are all equal to @xmath324 when @xmath325 and to @xmath130 when @xmath326 ( we will then later see what difference it makes if for finitely many values of @xmath262 this is not the case ) .
because @xmath305 holds , it implies that at the ca . @xmath312 first times at which the walk has been at @xmath162 before time @xmath309 , the value of @xmath269 at those times was greater than @xmath327 ( because @xmath328 was small , @xmath329 was large and @xmath330 ) .
furthermore , the value of @xmath269 at the time @xmath331 is very negative ( recall that @xmath332 , that at this time @xmath333 and @xmath334 are comparable , while @xmath335 is greater than @xmath308 and @xmath336 ) . hence , our assumption on the @xmath276 s ensures that @xmath337 , @xmath338 and @xmath339 ( note that the @xmath107 used in those steps can not exceed the number of times one visited the sites ) .
after these two steps , the value of @xmath269 ( when the walk is at @xmath241 ) has decreased by @xmath93 i.e. @xmath340 , whereas @xmath341 if @xmath342 . hence , we see that the walk will have to jump back and forth a number of times on the edge between @xmath241 and @xmath162 ; this will stop being the case because at some point , the walk will be at @xmath162 and will meet a @xmath276 that is equal to one .
because before that , the @xmath210 s ( at the visiting times of @xmath162 ) have been decreasing , we see that this can only happen when one uses a @xmath277 i.e. when one uses @xmath343 ( this is the @xmath278 that is equal to one , for which @xmath262 is the largest ) . from that moment ( call it @xmath344 ) onwards , the walk will start jumping back and forth on the edge between @xmath162 and @xmath242 ( this is because when it is at @xmath162 , it will use @xmath278 s for negative @xmath262 s , and when it is at @xmath242 , it will use @xmath345 for very large values of @xmath262 ) .
this will happen until the time @xmath346 . at that time , @xmath347 i.e. the number of times at which the walk jumped on @xmath253 is equal to the number of times the walk jumped on @xmath348 , which ensures that in this very particular situation , @xmath349 indeed holds . what is the difference due to those @xmath278 s that are equal to @xmath130 if @xmath350 or to @xmath324 if @xmath325 ?
the first remark is that all these particular @xmath262 s will indeed be used in our process ( note that @xmath210 is decreasing two by two at each visit of @xmath162 , and that it starts from a very positive value and become very negative , and therefore has to use all @xmath278 s for @xmath351 $ ] .
each time it meets a positive @xmath262 such that @xmath352 , it jumps to the right instead of jumping to the left . at the end of the day ( i.e. at @xmath353 ) this will mean that @xmath354 will be diminished by two .
conversely , if it meets a non - negative @xmath262 such that @xmath355 , this will add @xmath93 to the number of jumps on @xmath356 .
hence , we see that in our `` real '' case , @xmath357 and that @xmath349 will still hold because of our assumptions on the sum of the @xmath288 s . as a consequence
, we see that on the event @xmath322 of positive probability , all @xmath305 s hold .
this implies in particular that for each @xmath162 , the walk will not come back to the site @xmath241 after @xmath358 i.e. that @xmath359 is in fact the total number of times the walk does jump on the edge @xmath253 .
recall that @xmath360 and that almost surely , @xmath361 as @xmath362 , so that on our event of positive probability , @xmath363 .
it follows that on our event of positive probability , @xmath364 so that @xmath365 ( recall that between @xmath366 and @xmath358 we know that @xmath36 is equal to @xmath241 , @xmath162 or @xmath242 ) and @xmath367 as @xmath214 . this concludes the proof of proposition [ t2 ] .
we now focus on an example that exhibits yet another possible asymptotic behaviour for the walk : @xmath368 where @xmath369 and we then choose @xmath54 as before .
a naive first guess would be that this walk is self - attractive ( it is `` driven '' by the positive gradient of its local time ) and that it should get stuck .
however : [ p4 ] with positive probability , @xmath370 as @xmath214 .
again , the proof shows that this behaviour is valid for a wider class of self - interactions .
in fact , it is quite similar to the case ( with @xmath210 ) studied in the previous subsections .
the main difference is that ( in the `` good '' scenario that we will describe ) the walk will visit approximatively twice more the edge @xmath371 than the edge @xmath6 for large @xmath6 .
since the proof is otherwise almost identical to the previous case , we will only describe in plain words this `` scenario '' , and leave the details of the proof to the interested reader .
suppose that for some particular large time @xmath372 : * the walk is at its past maximum : @xmath373 .
we call this site @xmath374 , and for all @xmath4 , we define @xmath375 , @xmath376 and @xmath377 to be the respective values of @xmath378 at @xmath379 , @xmath380 and @xmath381 . * the values of @xmath382 , @xmath383 and @xmath384 satisfy : @xmath385 \hbox { and } w_0 \in [ v_0 /2 , v_0 ] .\ ] ] we also suppose that the value of @xmath383 is very large . note that under these assumptions , it very likely that @xmath386 and @xmath387 , i.e. that the walk will jump back and forth along the edge @xmath381 between @xmath162 and @xmath241 , because @xmath388 is very large , while @xmath389 is negative and has a large absolute value .
a quick analysis shows that ( with high probability ) the walk will jump back and forth on this edge until the negative drift at @xmath162 stops being huge , and the walk will then jump to @xmath242 for the first time .
when this happens , this means that at this time , @xmath390 and @xmath377 are comparable
. then , for some time , it will jump on the three sites @xmath241 , @xmath162 and @xmath242 , but while doing so , the drift at @xmath162 ( i.e. , that it feels when it is at @xmath162 ) grows fast , so that it will quickly be forced to jump along the edge between @xmath162 and @xmath242 only .
we then let @xmath391 the first time @xmath23 at which @xmath392 is greater than @xmath393 and note that @xmath391 satisfies the same conditions as thos we required for @xmath372 .
note also that once @xmath372 is found , the scenario is very likely to hold until @xmath391 when @xmath383 is large ( its conditional probability goes to @xmath83 very quickly with as @xmath383 gets larger ) . from this , it follows easily that with positive probability , the `` good scenario '' will indefinitely repeated .
then , clearly the total number of visits to the edge @xmath394 grows like @xmath395 as @xmath362 , and the proposition follows . to conclude ,
let us finally mention without proof a much less surprising possible asymptotic behaviour , namely the ballistic one .
the following pictures correspond to the trajectory and local time for the case where @xmath396 where @xmath397 that could seem at first glance to be again one of the walks driven ( like tsrw ) by the negative gradient of @xmath156 .
this example , like many others in the present paper illustrates how sensitive the discrete model is to little shifts in the definition of the driving dynamics .
let us conclude with a list of ( possibly accessible ) open problems related to the questions that we have just discussed : 1 .
prove the @xmath398 scaling behaviour of the walk in the `` tsrm regime '' ( i.e. where @xmath94 and @xmath399 and the stationary distribution makes sense ) , or even the convergence to tsrm . or construct anote related markovian model ( for instance with another function @xmath15 ) where one can prove this ? 2 .
describe in some detail the `` actual '' ( i.e. , the one that actually dominates ) trapping strategy in the `` trapped case '' @xmath92 and @xmath400 .
get some ( even partial ) description of the dynamics in the `` slow phase . ''
does the qualitative behaviour depend only on @xmath401 ? 4 . in the case where the walk goes deterministically to infinity ,
are other scaling behaviours than @xmath402 , @xmath403 and @xmath404 possible ? 5 . improve some of the results to almost sure statements ( instead of `` with positive probability '' ) .
is it possible that for some choice of the parameters in such self - interacting random walks with finite range , the qualitative asymptotic behaviour is actually not almost sure ( i.e. can two different behaviours that exist with positive probability exist ? ) .
do qualitatively really new asymptotic behaviours arise when one considers larger ( but finite ) self - interaction ranges
? * acknowledgements . *
bt thanks the kind hospitality of ecole normale suprieure , paris , where part of this work was done .
the research of bt is partially supported by the hungarian national research fund , grant no .
ww s research was supported in part by anr-06-blan-00058 .
the cooperation of the authors is facilitated by the french - hungarian bilateral mobility grant balaton/02/2008 . | we study certain self - interacting walks on the set of integers , that choose to jump to the right or to the left randomly but influenced by the number of times they have previously jumped along the edges in the finite neighbourhood of their current position ( in the present paper , typically , we will discuss the case where one considers the neighbouring edges and the next - to - neighbouring edges ) .
we survey a variety of possible behaviours , including some where the walk is eventually confined to an interval of large length .
we also focus on certain `` asymmetric '' drifts , where we prove that with positive probability , the walks behave deterministically on large scale and move like @xmath0 or like @xmath1 . = msbm10 _ dedicated to erwin bolthausen on the occasion of his 65th birthday _ |
a wake field analysis for a high - intensity accelerator typically includes wake and/or coupling impedance computations , and following calculations of loss factors and heating due to various elements of the vacuum chamber , as well as a study of possible instabilities .
beam coupling impedances and loss factors can be obtained from wake fields computed by time - domain codes like abci @xcite and mafia @xcite .
however , this approach works only for an ultrarelativistic bunch , with @xmath1 , because of difficulties formulating open boundary conditions for @xmath2 in time domain .
there are two specific features of the wake - field analysis in proton ( or h@xmath3 ) high - intensity linacs .
first , @xmath4 is significantly less than 1 for the most part of the machine .
results at @xmath5 , while provide useful estimates , can be quite different from those at the design @xmath4 values in some cases , see , e.g. , @xcite .
in particular , the resonance impedances and corresponding loss factors can strongly depend on @xmath4 .
frequency - domain calculations provide an accurate answer for a given @xmath2 , but typically they are limited to just a few lowest modes .
second , the beam in high - intensity linacs is either cw , or consists of macropulses containing many regularly spaced bunches . as a result ,
the beam frequency spectrum is concentrated only near the multiples of the bunch repetition frequency @xmath6 .
of course , the spectrum envelope is defined by the bunch shape , but due to short bunch length it rolls off at frequencies many times higher than @xmath6 .
therefore , an important question to consider is whether any higher mode has its frequency close to a multiple of @xmath6 .
the presence of such modes , especially at relatively low frequencies , can lead to undesired coherent effects .
we use time - domain computations with multiple bunches to answer this question .
the idea is to apply a standard time - domain code with a few identical bunches at @xmath5 , but to set the bunch spacing @xmath7 to @xmath8 for having the correct bunch repetition frequency .
since the resonance frequencies are essentially independent of @xmath4 , so is a conclusion from such simulations . in this note
, we concentrate only on this aspect of the wake - field studies for the apt .
specifically , we apply the code abci @xcite to compute longitudinal and transverse wakes in axisymmetric models of the apt 5-cell superconducting ( sc ) cavities using a varying number of bunches and looking for coherent wake - field effects .
wake potentials of a train of a few identical gaussian bunches passing through a 5-cell apt sc cavity have been computed with the code abci @xcite .
geometrical parameters of the apt cavities are given in @xcite .
the bunch rms length was chosen to be 4.5 mm in the @xmath4=0.82 section , and 3.5 mm for @xmath4=0.64 . while these bunches have @xmath4=1 , their separation is set to s=0.85657 m , which gives the proper bunch repetition frequency @xmath6=350 mhz .
we study the loss factor for the 5-cell apt sc cavities as a function of the number of bunches @xmath9 in the bunch train .
we expect that the loss factor per bunch tends to a constant for incoherent wakes , but it should increase linearly when wakes are coherent .
the coherent effects occur if higher - mode resonances are close to multiples of @xmath6 . correspondingly , the loss factor for the total train is proportional to @xmath9 if there are no coherent effects , or increases faster , up to @xmath10 , otherwise .
the results for the transverse loss factor @xmath11 per bunch are shown in fig . 1 , both for @xmath4=0.64 and @xmath4=0.82 . as one can see from fig . 1
, @xmath11 reaches its asymptotic already for @xmath9 between 5 and 10 in the case of @xmath4=0.82 .
this asymptotic value , in fact , is lower than @xmath11 for a single bunch . for @xmath4=0.64 , however , we observe an almost linear growth up to @xmath9 about 20 , and only after that the behavior changes and the transverse loss factor per bunch saturates . therefore ,
in the @xmath4=0.64 cavity higher - order dipole resonances are closer to multiples of @xmath6 than those for @xmath4=0.82 . for comparison ,
the total longitudinal loss factor for both cavities depends quadratically on @xmath9 , while the loss factor per bunch increases linearly as @xmath9 increases .
this is , of course , due to the fundamental accelerating mode of the cavity , whose frequency is 700 mhz .
the wake potentials for a bunch train with a monopole ( on - axis beam ) and dipole ( off - axis beam ) excitation look quite differently . there is a strong coherent build - up of the amplitude of the longitudinal wake as long as bunches travel through the cavity .
the bunches in the train interact with each other through the excitation of the cavity fundamental mode . on the contrary ,
no apparent increase is observed for the transverse wake potential ; wakes left by individual bunches are incoherent in this case .
therefore , one can use a maximal value of the transverse wake from these simulations as a reasonable estimate of that for a very large number of bunches , cf .
fig . 2 .
the maximum wakes from fig . 2
allow to estimate the strength of beam - induced deflecting fields in the cavities for use in beam - dynamics simulations . to identify the frequency range where a higher dipole resonance in the apt sc @xmath4=0.64 5-cell cavity has its frequency close to the multiple of the bunch frequency @xmath6=350 mhz , we plot in fig .
3 the power spectrum of the wake potential produced by a 30-bunch train in the cavity .
one can see in fig .
3 a regular structure of peaks at multiples of @xmath6 , as well as a peak near 950 mhz , which corresponds to the band of the tm110 dipole mode @xcite .
comparison of the wake power spectra for different @xmath9 shows that the magnitude of this last peak decreases quickly as one goes to longer and longer bunch trains , since there is a smaller and smaller excitation at this frequency .
however , it is the strong peak near 1750 mhz the multiple of the bunch frequency that produces a coherent increase of the dipole loss factor .
fortunately , its resonance frequency is close to the cutoff frequency of the pipe , which means this resonance can be effectively damped by hom power couplers .
nevertheless , a more detailed analysis of this frequency range with frequency - domain codes is required to identify the corresponding eigenmode(s ) , and take its ( their ) properties into account in designing hom couplers . as the number of bunches in the train increases , its frequency spectrum is getting more and more concentrated near the multiples of the bunch repetition frequency .
stronger peaks in the wake power spectrum for a relatively long bunch train indicate the frequency regions where the cavity resonances are close to multiples of @xmath6 .
we show in figs . 3 - 6 the power spectra of both the transverse and longitudinal wake potentials for 30-bunch trains .
the wake potentials have been calculated for 30 m after the leading bunch in all cases , they include about 60,000 points , and their fourier transforms have been performed with @xmath12 .
a logarithmic scale is used for the longitudinal spectra , otherwise the pictures would be dominated completely by the cavity fundamental mode at 700 mhz . comparing relative peak heights in the frequency spectra
shows where higher - order modes are close to multiples of the bunch frequency .
clearly , the potentially dangerous regions for the 5-cell @xmath4=0.64 apt sc cavities are : around 1750 mhz and 1050 mhz with respect to the dipole modes ; and near 2100 mhz for the monopole ones ( of course , apart from 700 mhz ) .
since 2100 mhz is above the beam - pipe cutoff , one should expect a trapped monopole mode near this frequency . for 5-cell @xmath4=0.82 apt sc cavities these regions
are : around 1750 , 700 , and 1050 mhz for the transverse modes ( but all those contributions are relatively weak ) and near 1750 and 1050 mhz for the longitudinal ones .
also , some additional attention is probably required to the transverse modes near 950 mhz for @xmath4=0.64 and in the range 900950 mhz for the @xmath4=0.82 cavities . while these frequencies are not close to a multiple of @xmath6 , the corresponding dipole resonances are strong enough that their effects are observed even for rather long bunch trains .
a new approach to study higher - order mode effects in cavities for non - ultrarelativistic ( @xmath2 ) cw or long - pulse beams is proposed .
it utilizes time - domain simulations using bunch trains which have @xmath4=1 , but a correct bunch repetition frequency @xmath6 .
as the number of bunches @xmath9 increases , the details of the beam frequency spectrum , which are dependent both on @xmath4 and @xmath9 , become unessential since the cavity is excited mostly at multiples of @xmath6 .
the approach allows applying standard time - domain codes , for example , @xcite . using this method
we have found a few potentially dangerous frequency ranges of higher - order modes for the apt superconducting cavities .
more details can be found in @xcite . a further analysis with frequency - domain codes
is required to identify the modes in these frequency ranges , and to take their properties into account in designing hom couplers
. 9 g.p .
lawrence and t.p .
wangler , in proceed .
pac97 , vancouver , bc , 1997 ; also in la - ur-97 - 2582 , los alamos , 1997 .
chin , report lbl-35258 , berkeley , 1994 .
mafia release 4.00 , cst , darmstadt , 1997 .
kurennoy , `` cavity loss factors for non - relativistic beams '' , report la - cp-98 - 55 , los alamos , 1998 ; also in these proceedings .
krawczyk , et al .
, in proceed .
pac97 , vancouver , bc , 1997 ; also la - ur-97 - 1700 , los alamos , 1997 .
krawczyk , in proceed .
pac97 , vancouver , bc , 1997 ; also la - ur-97 - 1710 , los alamos , 1997 .
kurennoy , `` multiple - bunch effects in apt sc cavities '' , report la - cp-98 - 151 , los alamos , 1998 . | the 1.7-gev 100-ma cw proton linac is now under design for the accelerator production of tritium ( apt ) project @xcite .
the apt linac comprises both the normal conducting ( below 211 mev ) and superconducting ( sc ) sections .
the high current leads to stringent restrictions on allowable beam losses ( @xmath0 1 na / m ) , that requires analyzing carefully all possible loss sources .
while wake - field effects are usually considered negligible in proton linacs , we study these effects for the apt to exclude potential problems at such a high current . loss factors and resonance frequency spectra of various discontinuities of the vacuum chamber
are investigated , both analytically and using 2-d and 3-d simulation codes with a single bunch as well as with many bunches .
our main conclusion is that the only noticeable effect is the hom heating of the 5-cell sc cavities .
it , however , has an acceptable level and , in addition , will be taken care of by hom couplers . |
sjh is supported by a pparc advanced fellowship .
the work of jds and rjw was supported in part by the u.s .
national science foundation under grants phy-9412556 and phy-9403869 .
jds and rjw also thank the higher education funding council for wales for support in the early stages of this work .
we thank john mehegan for help in analyzing the fermion propagator .
barbour , p. gibbs , j.p .
gilchrist , h. schneider and m. teper , phys .
136b * ( 1984 ) 80 ; i.m .
barbour , p. gibbs , k.c .
bowler and d. roweth , phys .
158b * ( 1985 ) 61 ; s.j .
hands and m. teper , nucl .
b347 * ( 1990 ) 819 ; g. salina and a. vladikas , nucl . phys .
* b348 * ( 1991 ) 210 .
.comparison of our two determinations of the renormalization factor @xmath51 ; the column labeled @xmath86 comes from the photon configurations via eq .
( [ eq : xxx ] ) , and the column labeled @xmath87 comes from the chiral condensates via eq .
( [ eq : psi ] ) [ cols="^,^,<",options="header " , ] | we present results suggesting that magnetic monopoles can account for chiral symmetry breaking in abelian gauge theory .
full @xmath0 configurations from a lattice simulation are factorized into magnetic monopole and photon contributions .
the expectation @xmath1 is computed using the monopole configurations and compared to results for the full @xmath0 configurations .
it is shown that excellent agreement between the two values of @xmath1 is obtained if the effect of photons , which dress " the composite operator @xmath2 , is included .
this can be estimated independently by measurements of the physical fermion mass in the photon background .
epsf @xmath0 lattice gauge theory provides an ideal laboratory for studying the effects of monopoles . in @xmath0 ,
all non - perturbative effects are known to be caused by monopoles @xcite .
further , there is an efficient procedure for identifying the magnetic current of monopoles in the link angle configurations generated in a @xmath0 lattice gauge theory simulation @xcite .
previously , two of the present authors have studied the heavy quark potential in @xmath0 lattice gauge theory , and shown that in the confined phase , the value of the string tension calculated from monopoles agrees quantitatively with the full @xmath0 string tension calculated directly from link angles @xcite . in this paper
, we turn our attention to the quenched chiral condensate in @xmath0 lattice gauge theory , @xmath3 .
we expect ( and find ) that this has a non - zero value in the confined phase , and vanishes in the deconfined phase .
our interest is in how the value of @xmath3 calculated from the link angle configurations compares with the same quantity calculated from monopoles . ascribing non - perturbative effects to monopoles
would say that in the confined phase , a non - vanishing condensate will be induced by monopoles .
nevertheless , since a matrix element of a product of field operators is involved , this value should be renormalized by short distance perturbative effects _
i.e. _ photons , so to make a quantitative comparison , we also need to calculate this renormalization .
our method for doing so , described in more detail below , is to compute the ratio of renormalized to bare masses for a charge moving solely in the field of photons .
our principle result is that applying this renormalization factor to the monopole value of @xmath3 yields the full @xmath0 answer , to within statistical errors .
let us now briefly describe the factorization of the link variables into photon and monopole parts .
we start by resolving the @xmath0 plaquette angles @xmath4 into fluctuating and monopole parts , @xmath5 where @xmath6 $ ] . from the monopole term @xmath7 , we can define two currents , one electric , the other magnetic .
the electric current is defined by @xmath8 .
the magnetic current of monopoles is defined by @xmath9 , where @xmath10 is the dual of @xmath7 .
the currents @xmath11 reside on direct(dual ) lattices .
we will compute the chiral condensate using the banks - casher formula @xcite ( see below ) which involves the density of eigenvalues of the dirac operator .
since the dirac particle is electrically charged , the effect of monopoles on it must be represented by an electric vector potential .
this four - vector potential derives from the electric current @xmath12 , and is analogous to the familiar three - vector potential @xmath13 used for the case of a static monopole in the continuum .
we have @xmath14 where @xmath15 is the magnetic unit of charge , and @xmath16 the lattice coulomb propagator in feynman gauge , satisfying @xmath17 . by defining @xmath18
, we define the photon link by demanding that @xmath19 be factorized on each link : @xmath20 we can now examine the chiral condensate @xmath3 separately on configurations @xmath21 and @xmath22 , as well as the full @xmath19 . in the quenched approximation @xmath3
can be defined in the chiral limit via the eigenvalues of the dirac operator @xcite : @xmath23 where @xmath24 is the lattice volume and @xmath25 is the number of eigenvalues satisfying @xmath26 in the interval @xmath27 , with @xmath28 $ ] the staggered lattice fermion kinetic operator @xmath29={1\over2}\sum_\mu ( -1)^{x_1+\cdots+x_{\mu-1 } } \left(u_\mu(x)\delta_{y , x+\hat\mu}-u_\mu^*(y)\delta_{y , x-\hat\mu}\right).\ ] ] for finite @xmath24 , @xmath30 vanishes ; therefore an extrapolation to the @xmath31 limit must be made from the spectrum density at non - zero @xmath32 . in practice @xmath33 can be estimated by measuring the lowest @xmath34 eigenvalues of @xmath35 per configuration using the lanczos algorithm , and then using a binning procedure @xcite .
calculations of the chiral condensate were performed with @xmath0 configurations generated using the standard wilson action on a @xmath36 lattice .
three different values of the inverse coupling @xmath37 were chosen : @xmath38 and @xmath39 .
for each value of @xmath37 , 5000 lattice updates were performed before beginning measurements to allow for equilibration .
the next 4000 configurations were used for measurements skipping every 20 lattice updates to reduce correlations . for values @xmath40
, our results exhibited the breaking of chiral symmetry ( i. e. @xmath41 ) .
these two values of @xmath37 are known to correspond to the confinement phase as well@xcite . for @xmath42
the configurations were found to be in the chirally symmetric phase , which is coincident with the coulomb phase .
our results for the chiral condensate calculations for the chirally broken phase are presented in figs .
( 1 ) and ( 2 ) .
we show both the full @xmath0 calculations and the results from the monopole gauge field configurations obtained using eq .
( 1 ) . using a linear fit and eq .
( 3 ) , the value of the chiral condensate was extracted from the 10 lowest values of @xmath33 . the values for @xmath43
are shown in table i. in fig .
( 3 ) the results for the chiral condensate from full @xmath0 fields and monopoles are shown for @xmath42 .
for comparison , the results from the @xmath44 monopole configurations are also included .
linear fits to @xmath33 yield a very small , but finite intercept .
however , the values are found to be about a factor of 100 times smaller than the results in the broken phase .
thus , we are confident of being in the chirally symmetric phase for @xmath42 . for completeness , the @xmath42 results are also included in table i. in fig .
( 4 ) we show the eigenvalue spectrum calculated using the @xmath22 background from both broken and symmetric phases .
there is no signal of chiral symmetry breaking ( note that we choose anti - periodic boundary conditions for the fermions in the temporal direction to avoid zero modes from near - plane wave solutions ) .
this supports our conjecture that chiral symmetry breaking in this model can be ascribed entirely to monopoles .
is it possible to account for the mismatch between @xmath45 and @xmath46 ?
since @xmath2 is a composite field operator , we expect it to be modified by quantum corrections independently of whether it acquires a vacuum expectation value . in perturbation
theory , uv fluctuations in general result in the requirement for all field operators to be renormalized , and for composite operators to have an additional renormalization .
this consideration leads us to the following hypothesis : the mismatch between the two condensate measurements is due to the rescaling of the @xmath2 operator by the fluctuations of the gauge fields contained in the @xmath22 configurations ; more concisely @xmath47 where @xmath48 is the dressed operator which takes into account photon - like fluctuations , and the expectation value is taken in the monopole - only backgrounds .
if the hypothesis is correct , then we can write @xmath49 ie .
the tree - level operator is multiplicatively enhanced by the photon fluctuations , the condensate data suggesting that the factor @xmath50 . in support
, we now describe an alternative and independent determination of @xmath51 , obtained by measuring the physical fermion mass @xmath52 in the @xmath22 background . on the assumption that the fluctuations in @xmath22 are approximately gaussian ,
then the dressing of the @xmath2 operator is given by a set of feynman diagrams .
let @xmath53 denote the complete set of 1pi diagrams describing corrections to the fermion two - point function .
then for the dressed fermion propagator we have @xmath54 where @xmath55 , @xmath52 denote bare and physical fermion masses respectively , @xmath56 is a wavefunction rescaling which by analogy with perturbative qed we expect to be both gauge and cutoff - dependent , and the @xmath57 symbol shows that the second equality holds only in the neighborhood of some subtraction point @xmath58 . equation ( [ eq : dressedprop ] ) may be rearranged to read @xmath59 with @xmath60 now , using the identity @xmath61 we see that the operation of differentiating with respect to @xmath55 is equivalent to a zero momentum insertion of a @xmath2 operator in a feynman diagram . therefore @xmath62 is the set of 1pi diagrams , having one external @xmath63 and one external @xmath64 , which describe corrections to @xmath2 , ie . @xmath65 ie .
@xmath66 where @xmath67 is the rescaling factor we seek , and the factor @xmath56 is included because in the diagrammatic approach matrix elements are evaluated with dressed fermion propagators on the external legs . combining ( [ eq : zmdef ] ) with ( [ eq : zdef ] ) , we find @xmath68 hence a measurement of @xmath52 in the @xmath22 background as a function of @xmath55 gives an independent estimate of @xmath51 .
this argument is essentially the same as that used to establish @xmath69 in standard renormalized perturbation theory , where @xmath70 is the renormalized operator whose green functions with other renormalized fields and operators are finite @xcite .
here we make no attempt to define a renormalized operator , since bare quantities are evaluated in a lattice simulation , but rather use the same formalism to quantify the effects of operator enhancement by quantum fluctuations .
note that we have not specified the subtraction point defining @xmath51 very precisely , and have assumed that @xmath51 is independent of @xmath55 . from experience in perturbation theory , we expect that numerically the most significant contribution to @xmath51 , of @xmath71 , comes from short wavelength fluctuations , and that @xmath51 is relatively insensitive to the details of the subtraction . to measure the physical fermion mass in the @xmath22 background , we performed calculations of the fermion propagator starting with the same configurations used to compute @xmath33 . since the fermion propagator is not gauge invariant , it is first necessary to fix a gauge .
although a gauge transformation of @xmath22 is not strictly a symmetry of the full theory , this procedure is justified since @xmath2 , and by hypothesis @xmath52 , are gauge invariant . in this work
we have used landau gauge @xcite and extracted the lattice fermion timeslice propagator @xmath72+m_0)_{0,x}^{-1}\biggr\rangle^{phot}\ ] ] using a conjugate gradient routine , for bare mass values @xmath73 . the restriction to spatial sites an even number of lattice spacings from the origin in each direction improves the signal @xcite . note that we have not attempted to fix the residual gauge freedom as in @xcite , but have instead relied on the fluctuations in @xmath22 being intrinsically small ( see below ) .
the physical mass @xmath52 can now be estimated by fitting @xmath74 to the following functional form : @xmath75 where @xmath76 is the lattice size in the time direction .
we used 100 configurations from each of the three @xmath37 values previously studied . in order to take into account correlations between renormalized mass estimates due to using the same configurations for different bare masses we applied a bootstrap fitting routine . to determine @xmath51 we plotted the mass obtained from @xmath74 against the input bare mass , and fitted the results to a linear form : @xmath77 fig .
( 5 ) shows a graph of @xmath78 from our simulation for @xmath44 .
the graphs for the other two values of @xmath37 are very similar .
the results for @xmath51 and @xmath79 from the linear fits for each value of @xmath37 are shown in table ii . from the table it can be observed that the value of @xmath51 is a smooth function through the phase transition .
unexpected in our results were the small non - zero values of the intercept .
this will be discussed briefly below .
finally , we compare our value of @xmath51 calculated using @xmath22 with the value observed in the chiral condensate calculations .
we define the renormalization factor from the chiral condensate to be @xmath80 and present the comparison in table iii .
the two independent determinations of the renormalization factor show excellent agreement .
this answers the question asked earlier : is it possible to account for the mismatch between @xmath45 and @xmath46 ?
our results demonstrate that the mismatch is simply operator renormalization due to the photons .
as yet we have no satisfactory explanation for the non - zero intercepts in table ii , implying a small breaking of chiral symmetry in the photon - only configurations .
it could be either a finite volume effect , or perhaps a residual gauge freedom associated with gribov copies and the consequent non - uniqueness of the landau gauge , or perhaps due to spatially constant gauge field modes .
the only effect that we have investigated quantitatively is that of constant modes .
we repeated our previous calculations at @xmath81 , this time performing a global gauge transformation to remove constant modes from the gauge fields .
this was done in the following way : the full gauge fields were shifted by a direction dependent constant , @xmath82 where @xmath83 is chosen so @xmath84 .
the new gauge fields @xmath85 were used to recompute @xmath51 and the chiral condensates . in repeating the calculations with this gauge fixing we observed no measurable difference in the results .
although we are not able to explain the small non - zero intercepts , the overall agreement of the slope with the value of @xmath51 required to explain the condensate results is extremely satisfying , and supports our hypothesis on the role of the residual photon fluctuations enhancing the @xmath2 operator . an obvious direction in which to extend this analysis is the exploration of chiral symmetry breaking in non - abelian gauge theories following the identification of monopole networks after abelian projection . studies of this kind have already appeared @xcite ; it is interesting to note that in the maximal abelian gauge used , there is a similar mismatch in the measured values of @xmath3 between full and monopole - only configurations
. it would be interesting to check whether this effect could be accounted for by the operator renormalization described here .
finally note that even though this argument assumed a split of the background gauge field into gaussian fluctuations , described by feynman diagrams , and non - perturbative monopole - only configurations , the idea of classifying effects into those which rescale a local operator and those which lead to a non - vanishing expectation value for the operator may be generalized .
for instance , it may in principle be possible to extend the analysis by factoring configurations in a scale dependent fashion , including only large monopole loops in the monopole part , and seeing if small monopole loops simply renormalize @xmath2 by comparing the @xmath51 factors measured in two different ways .
this may prove to be an effective probe of the scale at which non - perturbative effects manifest themselves .
the scale dependence of monopole contributions to the string tension was studied in @xcite . |
in the earliest stages of star formation , the class 0/i phases @xcite , the newborn protostar is embedded within a dense envelope of gas and dust . in the class 0 phase , little or no emission is generally detected shortward of @xmath110@xmath3 m due to high extinction toward the central object ; however , the bipolar outflows from the central protostar and disk carve cavities in the envelopes creating scattered light nebulae visible in the near to mid - infrared ( 1 - 8@xmath3 m ) .
the morphology of the outflow cavities then result from a ram pressure balance between the infalling envelope and outflow @xcite .
early models of infalling envelopes assumed spherical collapse @xcite along with slow rotation @xcite , which were later modified for infall from initially flattened geometries @xcite .
the rotating collapse model provides a natural route for the formation of circumstellar disks within the protostellar envelope .
however , disks in class 0 objects have been difficult to study primarily because of the surrounding envelope ( e.g. * ? ? ? * ) and insufficient resolution at millimeter wavelengths to isolate the protostellar disk . in addition
, the small - scale structure of outflow cavities in class 0 protostars has also been elusive due to the envelope extincting near - ir scattered light emission , shifting the scattered light emission peak into the mid - ir @xcite .
one of the nearest low - mass protostars in taurus ( d = 140pc ) , l1527 irs ( hereafter l1527 ) , has been a favorite target for studies of outflow and envelope structure ( e.g. * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
it has been classified as class 0 protostar @xcite but could be a class i system given its edge - on orientation @xcite , which enhances the amount of extinction along the line of sight @xcite .
the moderate resolution irac images of l1527 from _ spitzer _ reveal bright bipolar cavities extending @xmath110000 au in radius @xcite .
the outflow cavities are separated by what appears to be a large - scale ( @xmath11000 au ) dark lane ; however , we observed a central point - source between the cavities within the dark lane from 3.6 to 8@xmath3 m . in ( * ? ? ?
* hereafter paper i ) , we were able to explain this by constructing a model assuming that the point - source was a compact inner outflow cavity , unresolved by _ spitzer _ , connecting with the larger outer cavity .
we then proposed high - resolution mid - ir imaging to test this model . in this paper , we present high resolution l@xmath0-band ( 3.8@xmath3 m ) observations of l1527 from the gemini north telescope .
this is the highest resolution mid - ir image of the inner envelope / outflow cavity around a class 0 protostar .
the observations confirm our prediction of a compact , bipolar scattered light structure ; quite similar to the model constructed in paper i but strongly resembling a disk in scattered light .
we observed l1527 with the 8.1 meter gemini north telescope at mauna kea on 2009 october 24 and 2009 december 11 using the near - infrared imager ( niri ) in l@xmath0-band ( 3.8@xmath3 m ) .
the camera was used in f/14 mode which provides a 0.049@xmath2 pix@xmath4 scale ; we only used the central 512x512 area of the detector for faster read - out which enabled integrations short enough to avoid saturation from sky background .
we observed with a standard 5-point `` dice-5 '' dither pattern with 5@xmath2 steps and a nearby star was used for tip - tilt correction and guiding . at each position
150 coadded images of 0.2 seconds were taken ; a total of 35 images were taken yielding 16 minutes of integration during the first observation and 176 images totaling 88 minutes of integration during the second observation . however , the first 38 images in the second observation were unusable due to variable sky background .
observations of the standard star hd 22686 show that the seeing was @xmath10.3@xmath2 on both nights .
the raw data were reduced using the gemini iraf package .
flat field images were generated from sky flats constructed by median combining the dithered images .
each on - source frame was sky subtracted using a median sky generated from the on - source frame and the two images taken immediately before and after .
this method of sky subtraction compensates for the rapid variations in sky brightness at l@xmath0-band .
the sky subtracted images were then combined using the imcombine task of iraf using the offsets given in the header ; l1527 was quite faint in the individual frames and could not be used for further refinement of the offsets .
the datasets from both epochs of observation were combined yielding a final image with 85 minutes of integration .
we also include data taken with the infrared array camera ( irac ) @xcite on the _ spitzer space telescope_. these data were taken in 24 february 2004 as part of the gto taurus survey and were presented in @xcite and @xcite .
the images of l1527 at l@xmath0-band and 3.6@xmath3 m are shown in figure 1 .
the irac image clearly shows the point - like structure appearing in the center of the envelope between the large scattered light cavities .
the gemini l@xmath0 image resolves the point - like structure into a compact , bipolar scattered light nebula at the center of l1527 .
the two lobes are separated by a narrow dark lane @xmath10.45@xmath2 ( 60 au ) wide , consistent with a circumstellar disk shadow .
the l@xmath0 image reveals structure similar to that of the _ hubble _ nicmos images of class i protostars in taurus from @xcite .
however , this is the first time such an image has been captured of a class 0 protostar .
the total extent of the bipolar structure in the direction of the outflow ( east - west ) is @xmath12.5@xmath2 ( 350 au ) and @xmath11.8@xmath2 ( 250 au ) in width .
the extent of the scattered light along the outflow and dark lane is much thicker than a prototypical disk ( i.e. hh30 , * ? ? ?
* ) indicating that the disk in l1527 is more vertically extended or that the scattering is coming from the inner envelope .
the eastern lobe is about twice as bright as the western lobe ; consistent with eastern side being slightly inclined toward us ( @xmath185@xmath5 ) ( paper i ) .
the orientation of the disk dark lane indicates that the rotation axis of the system has a position angle @xmath185@xmath5 east of north .
we also see extended , low surface brightness outflow cavity emission extending away from the inner envelope in the gemini image .
the cavity is quite narrow until @xmath16@xmath2 ( 840 au ) from the center ; this narrowness was not well resolved in the _ spitzer _ image .
past 6@xmath2 , the cavity expands rapidly to very wide angles . the total flux within a 7@xmath614 ( 1000 au ) aperture radius is 6.5 @xmath7 0.6 mjy , consistent with the irac flux from paper i. we use the monte carlo radiative transfer code of @xcite to interpret the observations .
the model components are described in paper i , but we discuss some important aspects here for completeness . the envelope density structure is the standard rotating collapse solution @xcite in which envelope density scales with the infall rate @xmath8 with @xmath9 @xmath10 r@xmath11 outside the centrifugal radius ( r@xmath12 ) and @xmath9 @xmath10 r@xmath13 inside r@xmath12 .
conical outflow cavities are carved out of the envelope , their shape is defined by a polynomial , the degree of the polynomial is the shape parameter .
we use the dust model from paper i for the envelope which has grains up to 1@xmath3 m in radius .
the disk is defined by its initial scale height h@xmath14 at the stellar radius r@xmath15 and flaring power law h @xmath10 r@xmath16 , radial density profile @xmath9 @xmath10 r@xmath17 , mass , and outer radius .
the vertical structure of the disk is assumed to be gaussian with an initial scale height defined to be in hydrostatic equilibrium at the dust destruction radius ( t@xmath18=1600k ) .
values of @xmath19 = 9/8 , 5/4 are commonly used , corresponding to isothermal scale heights with @xmath20 @xmath10 r@xmath21 and r@xmath13 .
the gas and dust of the disk are assumed to be well mixed .
the dust opacities for the disk are taken from @xcite which were used to model the sed of hh30 .
the dust is distributed in a quasi - settled manner with grains up to 1 mm in the disk midplane while smaller grains remain extended .
the transition between these dust populations is defined to be n@xmath22=10@xmath23@xmath24 or the first scale height . in paper
i , we inferred that the apparent point source seen at 3.6@xmath3 m was in fact a bipolar structure seen in reflected light , with the true central source actually hidden by extinction .
to create this structure , we constructed a dual - cavity model having a narrow inner cavity with an outer cavity offset by 100 au ( see figure 10 of paper i for an illustration ) . comparing the top and middle rows of figure 2
shows that our prediction was qualitatively correct .
we next take advantage of the high spatial resolution of the gemini data to improve our understanding of the inner structure .
we required that models reproduce both the l@xmath25 and irac images and the broadband sed and irs spectrum ( * ? ? ? * and paper i ) , as it is important to use as many constraints as possible ( paper i , * ? ? ? * ) .
we first adjusted the parameters of the dual - cavity model to better reproduce the gemini observations by varying the inner and outer cavity shapes and opening angles .
as shown in figures 2 and 3 , our best - fit dual - cavity model reproduces the observations reasonably well ( note that the images for the refined dual - cavity model are not shown as they nearly identical to the disk model as discussed below ) . the envelope infall rate for this model was 10@xmath26m@xmath27 yr@xmath4 ; the outer cavity was offset from the inner cavity by 85 au .
the disk parameters are h(10au)= 1.87 au with @xmath19 = 1.25 , h@xmath14 = 0.0332 r@xmath15 , @xmath28= 0.05m@xmath27 and @xmath29= 25 au ; a model with @xmath19 = 1.125 was also able to be fit , but with double the initial scale height ; the full parameter set is listed in table 1 .
we did not consider a disk model in paper i because the large - scale dark lane seemed to be too thick to be reproduced with a disk .
however , using the detailed gemini image , we can describe disk parameters which yield scattering surfaces comparable to that of the dual - cavity model .
the highly - flared disk in this model had h(10au)= 1.95 au ( h(190au ) = 82 au ) , with @xmath19 = 1.27 , @xmath30 = 3.0 , h@xmath14 = 0.03 r@xmath15 , @xmath28= 0.005m@xmath27 and @xmath29= 190 au .
the envelope infall rate was lowered to 0.8 @xmath31 10@xmath26m@xmath27 yr@xmath4 in order to allow more scattered light from the disk to escape through the envelope .
we also used the envelope dust model for the disk upper layers rather than the default dust model ; this increased the l@xmath0 intensity and reduced emission from 10 to 60 @xmath3 m ( see table 1 for other parameters ) bringing the model into closer agreement with the observations .
as shown in figures 2 and 3 , this model also provides reasonable agreement with the observations .
the dual - cavity model in paper i was justified based on simulations with a wide - angle outflow and a highly flattened envelope density distribution @xcite .
however , it is unclear if such a flattened envelope is present in l1527 , as there is no evidence for this in submillimeter observations @xcite . in addition , the outflow cavity on scales larger that 840 au from the central object does not easily fit within this framework due to the rapid widening of the scattered light nebula beyond 840 au ( 3 ; figure 1 ) this is difficult to explain with a simple conical cavity because there is no direct line - of - sight from the protostar / disk to these parts of the outflow cavity .
therefore we suggest that the morphology of the scattered light nebulosity on large scales is not due directly to the outflow but to the morphology of the envelope and/or ambient medium @xcite . in any case
it appears difficult to explain the complex shape of the scattered light structures with an simple outflow - envelope interaction
. given these difficulties , the disk scenario seems attractive .
images of disks in silhouette have previously been observed around t - tauri stars ( class ii sources ) and class i protostars ( e.g. * ? ? ?
* ; * ? ? ?
the small - scale scattered light morphology of l1527 bears a resemblance to that of iras 04302 + 2247 and hh30 .
the problem for this model is that the disk must be highly flared , with the aspect ratio of the disk extinction lane relative to the extent of the scattered light surfaces being about a factor of two larger than other objects . to reproduce the compact scattered light and large - scale dark lane with a disk , it was necessary to adopt a high degree of flaring in the disk ( @xmath19=1.27 ) , a steep density profile ( @xmath30=3.0 ) , a large disk radius ( r@xmath18=190au ) , and a small disk mass ( m@xmath18=0.005m@xmath27 ) .
( note that the mass is simply a formal parameter of the model for scaling the density . extrapolating disk masses from fitting scattered light in the disk upper layers
is problematic since they contain a small fraction of the total mass . )
these observations and modeling emphasize not only that disks form early during protostellar collapse @xcite but also that disks with large radii are able to form during the class 0 phase ( see also * ? ? ?
* ) with radii comparable to class ii disks in taurus @xcite .
in general , dust growth and/or depletion in upper disk layers relative to ism conditions is needed to reproduce the scattered light images of edge - on disk systems and the seds of t tauri stars @xcite .
thus , it may be that the postulated l1527 disk has had much less dust evolution and thus has more small dust in upper layers , making it easier to explain the wide extinction lane observed ( e.g. * ? ? ?
@xcite directly imaged the disk of l1527 and a binary companion separated from the primary by 0.17@xmath2 ( 24 au ) with vla @xmath32=7 mm observations .
these observations are overplotted on the l@xmath0 image in figure 4 .
the observed disk is quite compact - only @xmath10.3@xmath2 ( 40 au ) in diameter .
though this system is a close binary , we do not expect the companion to affect the modeled parameters significantly , because our results mostly constrain the outer disk .
the compactness of the disk measured @xcite is not inconsistent with our disk model since the @xmath32=7 mm observations will only be sensitive to the densest part of the disk and where the grains are largest .
their measured disk mass of 0.1@xmath33 is substantially larger than our disk model ( which is uncertain as discussed previously ) .
note that the disk mass is strongly dependent on the assumed dust mass opacities ; @xcite used a dust opacity model based on @xcite which had @xmath34=4.9@xmath3110@xmath35@xmath36g@xmath4 ( p. dalessio private communication ) ) .
@xcite used @xmath37=0.035 @xmath36g@xmath4 with @xmath38 @xmath10 @xmath39 which yields @xmath34=0.00425 @xmath36g@xmath4 when extrapolated .
the difference between the @xmath34 values can be attributed to the @xcite dust model only considering grains up to 1@xmath3 m in size while the dust model used by @xcite considered grains up to 1 mm in size . scaling the @xcite @xmath32=7 mm mass to @xmath34 from @xcite gives @xmath28=0.0095@xmath33 , in agreement with the mean disk mass of taurus . from the observations and modeling it is now clear that the small - scale scattered light in l1527 is most likely due to a disk .
however , the assumed parameters needed to fit the disk model ( e.g. steep density profile ) may only have been necessary to enable our simplistic treatment of disk structure to work for l1527 .
models which solve for the vertical structure of the disk self - consistently ( i.e. * ? ? ?
* ) , given the irradiation from the central object have a density structure that is not gaussian ( as assumed in our modeling ) but falls off less rapidly with @xmath40 because the temperature increases with vertical height @xmath40 .
indeed , even in the context of an isothermal structure , the gaussian density distribution becomes a poor approximation as the ratio of the vertical height to radial distance becomes @xmath41 .
these issues deserve further investigation with physically self - consistent radiative transfer modeling of l1527 .
we have presented high - resolution l@xmath0 imaging of the inner envelope of l1527 demonstrating that much can be learned of inner envelope structure and newly forming disks with high resolution scattered light observations of class 0 protostars .
the observations strongly resemble the dual - cavity model constructed in @xcite ; however , the observations can be equally well reproduced by a large , highly flared disk or a refined dual - cavity structure .
however , the global scattered light structure is difficult to interpret as purely an outflow cavity .
we therefore suggest that the bright scattered light structure is most likely be a vertically extended disk - in - formation with material falling onto it .
the authors wish to thank the anonymous referee for a helpful report making the manuscript more clear , gemini staff astronomer r. mason for critical assistance in conducting the observations , b. whitney for providing her radiative transfer code to the community , and p. dalessio for helpful discussions .
j. t. and l. h. acknowledge funding from hst - go-11548.04-a and partial support from the university of michigan .
l. l. acknowledges the support of dgapa ( unam ) , conacyt ( mexico ) , and the guggenheim memorial foundation .
4 ( diamonds ) and 7@xmath614 ( boxes ) ( 10000 au and 1000 au ) are plotted .
the triangle at 3.8@xmath3 m is the gemini l@xmath25 flux within 1000 au .
the model seds are plotted for the disk model ( left ) and the refined dual - cavity model ( right ) with multiple model apertures of 10000 au ( solid line ) , 6000 au ( long - dashed line ) , and 1000 au ( short - dashed line ) .
the model is clearly deficient in flux at long wavelengths ; however , external heating is not taken into account and the dust temperatures fall below 10k in the outer envelope emitting less in the far - ir and submm.,title="fig : " ] 4 ( diamonds ) and 7@xmath614 ( boxes ) ( 10000 au and 1000 au ) are plotted .
the triangle at 3.8@xmath3 m is the gemini l@xmath25 flux within 1000 au .
the model seds are plotted for the disk model ( left ) and the refined dual - cavity model ( right ) with multiple model apertures of 10000 au ( solid line ) , 6000 au ( long - dashed line ) , and 1000 au ( short - dashed line ) .
the model is clearly deficient in flux at long wavelengths ; however , external heating is not taken into account and the dust temperatures fall below 10k in the outer envelope emitting less in the far - ir and submm.,title="fig : " ] image of l1527 with 7 mm dust continuum observations ( contours ) from vla observations by @xcite .
notice that the position angle of the long axis of the disk is in the same direction as the disk extinction lane , @xmath1 - 5@xmath5 east of north .
the second 7 mm point source just east of the disk is the binary companion . ]
llccl r@xmath15(r@xmath27 ) & stellar radius & 2.09 & 2.09 & fixed ( paper i ) + t@xmath15(k ) & stellar temperature & 4000 & 4000 & fixed ( paper i ) + l@xmath15(l@xmath27 ) & system luminosity & 2.75 & 2.75 & fixed ( paper i ) + m@xmath15(m@xmath27 ) & stellar mass & 0.5 & 0.5 & fixed ( paper i ) + m@xmath42(m@xmath27 ) & disk mass & 0.05 & 0.005 & varied ( both models ) + h(10au ) & disk scale height at 10au & 1.87 & 1.95 & varied ( both models ) + h@xmath43 & disk scale height at r@xmath15 & 0.0332 & 0.03 & varied ( both models ) + @xmath30 & disk radial density exponent & 2.25 & 3.0 & varied ( both models ) + @xmath19 & disk scale height exponent & 1.25 & 1.27 & varied ( both models ) + @xmath44(m@xmath27 @xmath45 ) & disk accretion rate & 3.0 @xmath46 & 3.0 @xmath46 & fixed ( paper i ) + r@xmath47(r@xmath15 ) & magnetosphere co - rotation radius & 3.0 & 3.0 & fixed ( paper i ) + f@xmath48 & fractional area of accretion hotspot & 0.01 & 0.01 & fixed ( paper i ) + r@xmath49(r@xmath15 ) & disk inner radius & 14.25 & 14.25 & fixed ( paper i ) + r@xmath50(au ) & disk outer radius & 25 & 190 & varied ( both models ) + r@xmath51(au ) & centrifugal radius & 25 & 190 & varied ( coupled to r@xmath50 ) + r@xmath52(r@xmath15 ) & envelope inner radius & 42.75 & 42.75 & fixed ( paper i ) + r@xmath53(au ) & envelope outer radius & 15000 & 15000 & fixed ( paper i ) + @xmath8(m@xmath27 @xmath45 ) & envelope mass infall rate & 1.00 @xmath54 & 0.8 @xmath54 & varied slightly ( both models ) + b@xmath55 & inner cavity shape exponent & 1.3 & - & varied ( dual cavity model ) + b@xmath56 & outer cavity shape exponent & 1.7 & 1.5 & varied ( both models ) + z@xmath56(au ) & outer cavity offset height & 85 & - & varied ( dual cavity model ) + @xmath57(@xmath5 ) & inner cavity opening angle & 15 & - & varied ( dual cavity model ) + @xmath58(@xmath5 ) & outer cavity opening angle & 15 & 20 & varied ( both models ) + @xmath59(@xmath5 ) & inclination angle & 85 & 85 & fixed ( paper i ) + @xmath60(g @xmath24 ) & cavity density & 0 & 0 & fixed ( paper i ) + @xmath61(g @xmath24 ) & ambient density & 0 & 0 & fixed ( paper i ) | we present high - resolution l@xmath0-band imaging of the inner scattered light structure of class 0 protostar l1527 irs ( iras 04368 + 2557 ) taken with the gemini north telescope .
the central point - source like feature seen in _ spitzer space telescope _
irac images is resolved in the gemini image into a compact bipolar structure with a narrow dark lane in the center .
two scattered light lobes are extended @xmath11.8@xmath2 ( 200 au ) perpendicular to the direction of the outflow and @xmath12.5@xmath2 ( 350 au ) along the outflow axis ; the narrow dark lane between the scattered light lobes is @xmath1 0.45@xmath2 ( 60 au ) thick . the observations are consistent with our initial modeling of a bright inner cavity separated by a dark lane due to extinction along the line of sight of the central protostar by the disk ( tobin et al .
2008 ) .
the bright , compact scattered light might be due to complex inner structure generated by the outflow , as suggested in our first paper , or it may more likely be the upper layers of the disk forming from infalling matter . |
the self - gravitating gases have remarkable physical properties due to the long range nature of the gravitational force .
they are not homogeneous even at thermal equilibrium .
this fundamental inhomogeneity character suggested that fractal structures can arise in a self - interacting gravitational gas@xcite .
self - gravitating gases are used to describe cold clouds in the interstellar medium as well as the large scale structure of galaxies . in both cases ,
self - gravitating gases provide scaling laws for the mass distribution with haussdorf dimensions compatible with the observations@xcite .
all particles have the same mass in the self - gravitating gases in thermal equilibrium considered till now both in the hydrostatic approach @xcite and in the statistical mechanics approach@xcite .
we study in the present paper the statistical mechanics of a non - relativistic gas with two kinds of particles with masses @xmath6 and @xmath7 .
such a system , besides its own physical interest , has obvious astrophysical motivations since cold clouds in the galaxy are formed typically by several kinds of particles .
for example , hydrogen and helium .
let us begin by briefly recalling some results about the statistical mechanics of the self - gravitating gas with one kind of particles @xcite .
it is a gas of @xmath19 non - relativistic particles of mass @xmath20 interacting through newtonian gravity .
the gas is in a volume @xmath21 and in a thermal bath at temperature @xmath8 . in the usual thermodynamic limit (
@xmath22 and @xmath23 is fixed ) the gaseous phase is not stable and the system collapses in a very dense phase . in the dilute limit @xmath22 and
@xmath24 fixed the system can exist in the gaseous phase .
this is a dilute limit since the average density @xmath25 goes as @xmath26 .
the relevant physical parameter of the system is @xmath27 with newton s constant @xmath5 and the length @xmath28 .
@xmath29 is the ratio of the characteristic gravitational energy @xmath30 and the kinetic energy @xmath8 of the gas .
for @xmath31 the ideal gas is recovered . from @xmath31 till a critical value @xmath32
the gaseous phase is stable in the canonical ensemble .
when @xmath29 reaches the value @xmath33 the gas collapses in a very dense phase .
the velocity of sound becomes imaginary at this point triggering unstabilities that lead to the collapse of the gas@xcite .
the saddle point approximation applies between @xmath31 and the point @xmath34 ( associated to the jeans instability ) where the determinant of small fluctuations is positive . at @xmath35 the determinant of small fluctuations vanishes and the saddle point approximation breaks down@xcite . beyond @xmath35
the gaseous phase is stable and the mean field approximation holds in the microcanonical ensemble . solving the saddle point equation in the spherically symmetric case allows to obtain the particle density and the thermodynamic functions as functions of the physical parameter @xmath29 . as shown in ref.@xcite the mean field approach is equivalent to the hydrostatic description@xcite provided the ideal gas equation of state is postulated in the latter approach .
the mass distribution turns to exhibit a scaling behaviour as a function of @xmath12@xcite .
we consider in this paper the self - gravitating gas with two kinds of particles in the canonical ensemble .
we recast the partition function as a functional integral over the densities of particles of each kind when the number of particles is large .
the statistical weight for each configuration of densities turns to be the exponential of an ` effective action ' which is proportional to the number of particles .
therefore , we can use the saddle point approximation in the thermodynamic limit to evaluate the partition function .
the ` effective action ' turns to be the free energy as a functional of the particle densities . when the saddle point provides a minimum of the free energy , the density solution of the saddle point equation is the most probable .
it is certainly exact for an infinite number of particles , since the minimized free energy exponentially dominates the partition function .
that is , the mean field theory defined by the saddle point becomes exact for an infinite number of particles .
the mean field approximation for the canonical ensemble ceases to be valid on a critical line in the @xmath36 plane ( see below ) . beyond this critical line
the saddle point is not a minimum of the free energy and it fails to reproduce the physics of the system .
more precisely , we consider @xmath37 particles of mass @xmath6 and @xmath38 particles of mass @xmath7 interacting through newtonian gravity in a volume @xmath21 and in a thermal equilibrium at temperature @xmath8 . by analogy with the self - gravitating gas with one kind of particles
, we consider the dilute thermodynamic limit : @xmath39 the two relevant physical parameter are here @xmath40 where @xmath41 for a cubic geometry .
notice that eq.([limter ] ) implies that the ratio @xmath42 stays fixed for @xmath43 the self - gravitating gas with two kinds of particle behaves as a perfect gas in the extremely diluted limit @xmath44 and @xmath45 .
when @xmath46 and/or @xmath47 grow , the gas becomes denser till it collapses into a very dense phase when @xmath9 and @xmath10 reach the collapse line for the canonical ensemble in the @xmath48 plane .
by analogy with the gas with one kind of particles @xcite we expect the collapse line to be very close and below the critical line in the @xmath48 plane .
the gaseous phase keeps stable in the microcanonical ensemble beyond this critical line .
we find here that the saddle point equations are two coupled non - linear differential equations for the densities of the two kinds of particles : @xmath49 .
we succeed to reduce these equations to a single non - linear differential equation .
we solve it in the spherically symmetric case .
we express the densities as functions of the physical parameters @xmath9 and @xmath10 .
we thus find the isothermal sphere with two types of particles .
we compute the mass inside a sphere of radius @xmath50 centered at the origin and show that it scales with a haussdorf dimension @xmath14 .
@xmath14 decreases with @xmath51 and @xmath52 from the value @xmath53 for the ideal homogeneous gas till @xmath54 in the canonical critical line .
the haussdorf dimension keeps decreasing beyond the canonical critical line in the stable phase of the microcanonical ensemble .
@xmath14 at the canonical critical line turns to be * independent * of the ratio @xmath55 and coinciding within the numerical precision with @xmath56 for the canonical critical point of the gas with one kind of particles @xcite .
this indicates that the haussdorf dimension at the canonical critical line is an universal value , independent of the gas composition .
the dependence of the critical values of the parameters @xmath9 and @xmath10 with the number of particles ratio is computed . the thermodynamic functions ( free energy , energy , entropy , local pressure and pressure contrast ) are expressed as functions of the physical parameter @xmath9 and @xmath10 .
the pressure contrast [ ratio between the pressure at the origin and the pressure at the boundary ] turns to be * lower * for this mixture of particles than for the gas with one kind of particles .
we compute the pressure at a point @xmath57 of the gas and show that it * locally * obeys the equation of state of an ideal gas , @xmath58 since the gas is inhomogeneous , the pressure acting on any non - infinitesimal volume of the gas does not obey the ideal equation of state .
( this was already the case for a gas with all particles of equal mass@xcite ) .
we plot in figs .
[ pressure3d]-[pressuresec ] the pressure of the gas at the surface .
the mean field equations have a straightforward hydrostatic interpretation .
we show that the mean field equations derived from the partition function are equivalent to the hydrostatic equilibrium equations * provided * that the ideal equation of state is postulated in the latter approach .
we stress that we give here a * microscopic derivation * of the equation of state ( [ ecestl ] ) from the partition function .
we then consider a self - gravitating gas formed by @xmath18 kinds of particles with different masses .
the mean field equations are derived and shown to reduce to a single non - linear differential equation .
this paper is organized as follows , in section ii we present the statistical mechanics of the self - gravitating gas with two kinds of particles in the canonical ensemble , in sec .
iii we present the main thermodynamic magnitudes , the equation of state and the scaling behaviour of the particle distribution for spherical symmetry . in sec .
iv we present the generalization for @xmath18 kinds of particles .
the appendixes contain relevant mathematical developments and the hydrostatic approach to a self - gravitating gas with two kinds of particles .
we present here the partition function for the self - gravitating gas with @xmath37 particles of mass @xmath6 and @xmath38 particles of mass @xmath7 inside a finite volume @xmath21 and derive the mean field approach to it .
we study the statistical mechanics of the self - gravitating gas with two kinds of particles in the canonical ensemble .
the hamiltonian of the system is @xmath59 here , @xmath60 and @xmath61 are the momenta and the coordinates of the particles of mass @xmath6 . @xmath62 and @xmath63 are the momenta and the coordinates of the particles of mass @xmath7 .
therefore , the classical partition function of the gas is @xmath64 it is convenient to introduce the dimensionless coordinates variables @xmath65 and @xmath66 .
the momenta integrals are computed straightforwardly .
hence the partition function becomes the product of the partition function of perfect gases with masses @xmath6 and @xmath7 times the coordinate integral @xmath67 .
@xmath68 where @xmath69 and @xmath70 we approximate the function @xmath67 for a large number of particles ( @xmath71 and @xmath72 ) generalizing the approach of ref.@xcite for two kinds of particles .
the function @xmath67 is expressed as a functional integral over all configurations with particle densities @xmath73 and @xmath74 [ @xmath73 stands for the density of the particles of mass @xmath6 and @xmath74 stands for the density of the particles of mass @xmath7 ] .
@xmath76 with , @xmath77 + i n_2 \ ; b_2 [ 1-\int { \rm d}^3 { \bf x } \ ; \rho_2({\bf x } ) ] \ ; \ ; .\end{aligned}\ ] ] @xmath78 and @xmath79 are lagrange multiplier enforcing the normalization of the densities : @xmath80 @xmath81 stands for the free energy of the gas for the pair of densities @xmath82 , while @xmath83 } -n_2 t \ln{\left[\frac{e v}{n_2 } \left(\frac{m_2 t}{2 \pi}\right)^ { \frac{3}{2}}\right]}\ ] ] is the free energy of the perfect gas with masses @xmath6 and @xmath7 .
one recognizes the gravitational energy in the first two lines of eq.([action ] ) , while the third line contains the entropy . since the free energy becomes large in the thermodynamic limit ( @xmath84 ) , the functional integral @xmath67 ( [ intfonct ] )
is dominated by the minima of @xmath85 .
extremizing the free energy with respect to the pair of densities @xmath82 yields the saddle point equations @xmath86 where we used eq.([etai ] ) .
these equations define the mean field approach .
we set @xmath87 and @xmath88 and we denote by @xmath89 the ratio of masses of the two kinds of particles , @xmath90 we set , @xmath91\quad , \quad \rho_2({\bf x})=\exp[\phi_2({\bf x } ) ] \ ; .\ ] ] eqs.([sadeq ] ) give for the gravitational potential @xmath92 = -\frac{t}{m_2}[\phi_2({\bf x})-a_2]\ ] ] created by the matter densities @xmath93 and @xmath94 at the point @xmath95 . using eqs.([phi])-([potentiel ] ) , we see that these densities obey the boltzmann laws @xmath96 containing the energy of a particle in this mean field gravitational potential , as it must be . @xmath97 and @xmath98 play the role of normalization constants . applying the laplace operator to the saddle point equations ( [ sadeq ] ) we find the differential equations @xmath99 these equations are scale covariant . if @xmath100 is a pair of solutions of eqs.([poisson ] ) , then the pair @xmath101 defined by @xmath102 is also a solution of eq.([poisson ] ) .
this property is due to the scale behaviour of newton s potential . using eq.([potentiel ] ) , we reduce eqs.([poisson ] ) to a single equation @xmath103 using eqs.([potentiel ] ) and ( [ boltzmann ] ) , eq.([poired ] ) becomes in dimensionless coordinates , @xmath104 \ ; .\ ] ] we show in appendix b that this equation is the condition of hydrostatic equilibrium for a two - component fluid once the ideal gas equation of state is postulated . therefore , the mean field approximation is equivalent to the hydrostatic description in the gaseous phase provided the ideal gas equation of state is assumed in the latter approach .
notice that local equations of state other than the ideal gas are often assumed in the context of self - gravitating fluids@xcite .
as stressed in ref.@xcite , one needs long range forces other than gravitational in order to obtain a non - ideal local equation of state in thermal equilibrium .
for spherically symmetric configurations the mean field equations become ordinary non - linear differential equations .
we express here the various thermodynamic quantities in terms of the solution of a single ordinary differential equation .
such equation reduces to the well - known isothermal sphere@xcite if all particles have identical mass .
we consider here the spherically symmetric case where the mean field equations ( [ poisson ] ) take the form @xmath105 where we work in a unit sphere .
therefore the radial variable runs in the interval @xmath106 , @xmath107 . using the scale covariance of the mean field eqs .
( [ poisson ] ) by the transformation ( [ scaletrans ] ) , we can set @xmath108 we use the new parameters @xmath109 and @xmath110 in the spherically symmetric case . hence the mean field eqs.([poissonrad ] ) are transformed into a reduced system of the form , @xmath111 let us find the boundary conditions of these equations . in order to have a regular solution at origin we impose @xmath112 the system ( [ poissonchi ] ) is invariant under the transformation @xmath113 hence , we can choose @xmath114 without loosing generality . as we see below
, the remaining boundary condition on @xmath115 is not independent from @xmath9 and @xmath10 .
the densities of the two kinds of particles @xmath116 and @xmath117 are to be normalized according to eq.([normalisation ] ) .
we obtain using eqs.([phi ] ) and ( [ dilatation ] ) , @xmath118 using the reduced mean field equations ( [ poissonchi ] ) , it is straightforward to show from eq.([etai - norm ] ) that @xmath119 hence @xmath120 recalling that @xmath121 is the gravitational potential ( [ potentiel ] ) we see using eqs.([dilatation ] ) , that @xmath122 is the gravitational field at the boundary of the sphere ( @xmath123 ) consistently with eq.([champ ] ) . from eq.([champ ] ) we introduce a new parameter @xmath124 independent of @xmath125 and function only of the physical parameters @xmath51 and @xmath52 .
notice the boundary condition @xmath126 the reduced mean field equations ( [ poissonchi ] ) become a single equation with the parameter @xmath127 as coefficient @xmath128 and the boundary conditions @xmath129 . using eqs.([phi ] ) and ( [ dilatation ] ) we can express the densities of the two kinds of particle in terms of the solution of eq.([eqchi1 ] ) @xmath130 both @xmath131 and @xmath132 are functions of @xmath127 [ see ( [ constante])-([eqchi1 ] ) ] . inserting @xmath133 and @xmath134 in eq.([etai - norm ] ) , we see that @xmath51 and @xmath52 become functions of @xmath125 and @xmath127 . then expressing @xmath125 and @xmath127 as functions of @xmath51 and @xmath52 , both densities of particles in eq.([dens ] ) become functions of the radial variable @xmath12 and of the physical parameters @xmath51 and @xmath52 .
we find asymptotically from eq.([eqchi1 ] ) @xmath135 for @xmath136 .
notice that this asymptotic behaviour may apply for non - physical values of @xmath125 where the gas is actually collapsed .
we now compute the thermodynamic quantities as functions of the physical parameters @xmath51 and @xmath52 .
let us start by computing the free energy .
using eqs .
( [ action])-([phi ] ) we find @xmath137 we express now the lagrange multipliers @xmath138 and @xmath139 as functions of the physical parameters @xmath51 and @xmath52 . in the spherically symmetric case the integration over the angles in eqs.([sadeq ] ) yields @xmath140 we introduce the densities for the two kinds of particles at the boundary ( @xmath141 ) @xmath142 where we used eqs.([phi ] ) and ( [ dilatation ] ) . notice that @xmath143 and @xmath144 are functions of @xmath145 and @xmath146 as explained by the end of sec .
hence , @xmath147 and @xmath148 are functions of @xmath145 and @xmath149 .
we find for @xmath141 the following expressions for the lagrange multipliers using the normalization of the densities ( [ normalisation ] ) @xmath150 inserting these expressions ( [ lagrange ] ) into the free energy ( [ free1 ] ) , we find @xmath151 + \frac{n_2}{2 } \left[\ln f_2 -\frac{1}{\mu}\ , \eta_1^r-\eta_2^r \right]\\ & & + \frac{n_1}{2 } \int { \rm d}^3{\bf x}\;\phi_1({\bf x})\;e^{\phi_1({\bf x } ) } + \frac{n_2}{2 } \int { \rm d}^3{\bf x}\;\phi_2({\bf x})\;e^{\phi_2({\bf x } ) } \ ; . \nonumber\end{aligned}\ ] ] we compute the integrals in the second line in appendix a and we find @xmath152 + n_2 [ \ln f_2 -\frac{1}{\mu } \ ; \eta_1^r-\eta_2^r + 3(1-f_2)]\;.\ ] ] the free energy as well as the other physical quantities are functions of @xmath51 and @xmath52 .
the parameters @xmath51 and @xmath52 are linked by the relation , @xmath153 hence , for fixed ratios @xmath154 and @xmath155 , the physical quantities depend only on @xmath51 or on @xmath52 . in that case
it is simpler to see the physical quantities as functions of @xmath52 or @xmath51 on a two dimensional plot than to watch the three dimensional surfaces for the physical quantities as functions of @xmath52 * and * @xmath51 .
we plot in fig.[freeen ] the free energy as a function of @xmath52 for fixed @xmath156 . in the limit where the particles of mass @xmath6 dominate ( @xmath157 )
the free energy becomes @xmath158\;.\ ] ] we recognize here the free energy of a self - gravitating gas with @xmath37 particles of mass @xmath6 ( see ref.@xcite ) .
we compute here the gravitational energy of the gas .
the density of gravitational energy is @xmath159 where @xmath121 is the gravitational potential ( [ potentiel ] ) .
hence , @xmath160
-\frac{n_2 t}{2 v } \ ; e^{\phi_2({\bf x } ) } \ ; [ \phi_2({\bf x})-a_2]\;.\ ] ] using the expressions for the lagrange multipliers ( [ lagrange ] ) , we obtain for the gravitational energy density in the spherically symmetric case @xmath161\nonumber\\ & & + \frac{n_2 t}{2 v } \ ; \rho_2(r ) \left[\ln\left(\frac{\rho_2(r_{max})}{\rho_2(r)}\right ) -\frac{1}{\mu}\eta_1^r-\eta_2^r \right ] \ ; .\end{aligned}\ ] ] integrating the energy density ( [ densengrav2 ] ) , with the help of eqs . ( [ normalisation ] ) and ( [ lagrange ] ) we obtain @xmath162 + \frac{n_2 \
, t}{2}\left[\ln{f_2}-\frac{1}{\mu } \ ; \eta_1^r- \eta_2^r\right]\nonumber \\ & -&\frac{n_1 \ , t } { 2 } \int { \rm d}^3{\bf x}\;\phi_1({\bf x})\;e^{\phi_1({\bf x } ) } -\frac{n_2 \ , t}{2 } \int { \rm d}^3{\bf x}\;\phi_2({\bf x})\;e^{\phi_2({\bf x } ) } \ ; .\end{aligned}\ ] ] the integrals in the second line are computed in appendix a yielding for the gravitational energy @xmath163\;.\ ] ] using the gravitational energy ( [ energy ] ) and the free energy ( [ free2 ] ) we obtain for the entropy @xmath164 + n_2 [ 6(f_2 - 1)-\ln{f_2}+ \frac{1}{\mu } \ ;
\eta_1^r+\eta_2^r ] \ ] ] where @xmath165 is the entropy of the perfect gas .
@xmath166}+\frac{5}{2}\right ) + n_2 \left(\ln{\left[\frac{v}{n_2 } \left(\frac{m_2 t}{2 \pi}\right)^ { \frac{3}{2}}\right]}+\frac{5}{2}\right)\;.\ ] ] we plot in fig .
[ entropy ] the entropy @xmath167 as a function of @xmath168 for @xmath156 . since the system is non - homogeneous the local pressure is not uniform
. the density of gravitational force is @xmath169\ ] ] where u(*r * ) is the gravitational potential ( [ potentiel ] ) .
we obtain for the force at the point @xmath170 using eqs.([phi ] ) and ( [ potentiel ] ) , @xmath171 the link between the density of force and the pressure is @xmath172\;.\ ] ] using eq.([phi ] ) , the local pressure is given by , @xmath173 this is the * local * equation of state for the self - gravitating gas with two kinds of particles .
we see that it locally coincides with the equation of state of a perfect gas . since the gas is inhomogeneous , the pressure acting on any finite volume will * not * obey the ideal equation of state . for a point @xmath174 at the boundary , using eqs.([phi ] ) , ( [ densper ] ) and ( [ pessure ] ) yields the external pressure as a function of @xmath51 and @xmath52 as , @xmath175 where @xmath147 and @xmath148 are defined by eq.([densper ] ) .
this is the equation of state of the gas as a whole that we plot in fig.[pressure3d ] . combining eqs.([energy ] ) and
( [ pessureext ] ) yields the virial theorem @xmath176 this self - gravitating system formed by two kinds of particles can be in two phases : gaseous and condensed .
the first one corresponds to @xmath46 and @xmath47 between the origin and their collapse values . in the gaseous phase
the free energy has a minimum for the pair of densities @xmath82 , solutions of the saddle point equations ( [ sadeq ] ) .
this pair of densities is the most probable distribution which become absolutely certain in the thermodynamic limit .
all thermodynamic quantities follows from this pair of densities @xmath82 . in the condensed phase , @xmath82 from mean field
does not describe the particle distribution and the mean field approach fails to describe the condensed phase .
it may be studied by monte - carlo methods as in ref.@xcite . when the physical parameters @xmath177 we retrieve the perfect gas . when @xmath51 and @xmath52 increase , the gas becomes denser at the center of the sphere ( @xmath178 ) and less dense at the boundary @xmath179 because of gravitational attraction [ see figs .
[ densbound]-[densor ] ] .
this effect is more acute for the heavier particles showing that more massive particles diffuse to the denser regions .
the equation of state is depicted in fig.[pressure3d ] . in the ideal gas limit , @xmath180 . in the case
@xmath181 ( gas of particles of mass @xmath7 ) and in the case @xmath182 ( gas of particles of mass @xmath6 ) we recover the equation of state of the self - gravitating gas with one kind of particles @xcite
. we call critical values @xmath183 and @xmath184 the points where @xmath185 $ ] exhibits a vertical slope . @xmath183 and @xmath184 define a critical line in the @xmath186 plane .
namely , for each value of @xmath42 we have a different pair of critical points @xmath183 and @xmath184 .
we plot the critical line in the @xmath186 plane in fig.[critline ] .
the surface pressure has a rim on the canonical critical line [ see fig .
[ pressure3d ] ] .
the projection of this rim on the @xmath187 plane yields the critical line plotted in fig.[critline ] . for a fixed @xmath17
we get a section of the equation of state surface depicted in fig.[pressuresec ] .
this section turns to have a form analogous to the equation of state for the self - gravitating gas with one kind of particles @xcite . by analogy with the gas with one kind of particles
, we expect that the gas collapses in the condensed phase for values of @xmath51 and @xmath52 slightly below @xmath183 and @xmath184 where the saddle point approximation breaks down .
the physical quantities exhibit a square root riemann sheet structure as functions of @xmath51 and @xmath52 .
the branch points are on the critical line for the canonical ensemble .
the lower branch [ see figs . [ pressure3d ] and [ pressuresec ] for the equation of state and figs . [ densbound]-[densor ] for the densities ] describe a phase absent in the canonical ensemble .
such phase is realized and is stable in the microcanonical ensemble as it was the case for the gas with one kind of particles@xcite .
we plot in fig.[etabc ] the value of @xmath188 at the critical line as a function of the number of particles @xmath189 .
this quantity is proportional to the total mass of the gas @xmath190 .
this critical parameter @xmath191 interpolates between the two limiting cases ( @xmath192 and @xmath193 ) where one kind of particles dominate over the others . when @xmath194 the particles of mass @xmath6 dominate and @xmath195 . in this limiting case , @xmath196 which is the critical value for the self gravitating gas with one kind of particles associated to the jeans instability@xcite .
when @xmath197 the particles of mass @xmath7 dominate and @xmath198 . in this limiting case ,
@xmath199 for the mass ratio @xmath200 corresponding to a mixture of hydrogen and helium . from eq.([pessure ] ) we see that the partial pressures of the particles with mass @xmath6 and @xmath7 are given by , @xmath201 the pressure contrast is defined by the ratio of the pressure at the center and the pressure at the boundary : @xmath202 @xcite .
we find from eqs.([limcond2 ] ) , ( [ qui20 ] ) , ( [ dens ] ) and ( [ pessure ] ) @xmath203 we extend this definition to each kind of particles and say that the partial contrasts are given by @xmath204 for the particles of mass @xmath6 and @xmath205 for the particles of mass @xmath7 .
we find from eqs.([limcond2 ] ) , ( [ qui20 ] ) , ( [ dens ] ) and ( [ pres12 ] ) , @xmath206 we plot the contrast and the partial contrasts in fig .
[ contrast ] .
we see that the contrast @xmath207 takes here * lower * values than for a gas with one kind of particles . on the contrary , @xmath208 ,
the partial contrast for the * heavier * particles , takes * higher values * than the contrast for a gas with one kind of particles .
this is due to the fact that the overdensity of particles in the center is more acute for the heavier ones as noticed above .
the particle density and pressure has its maximum at the origin and its minimum at the boundary , as expected .
however , their ratio ( contrast ) is much * larger * for the heavier particles than for the lighter ones .
they are related through [ see eq.([contras ] ) ] , @xmath209^{\mu}\ ] ] where @xmath210 . since @xmath211 , when @xmath212 , we can get @xmath213 [ see fig.[contrast ] ] . in particular
, @xmath214 may be much larger than the contrast in the gas with one kind of particles@xcite .
in summary , this shows that the heavier particles diffuse to the denser regions .
we compute here the mass @xmath11 inside a sphere of radius @xmath12 ( @xmath215 ) .
using gauss s theorem and recalling that @xmath121 is the gravitational potential ( [ potentiel ] ) we find that @xmath216 using eqs.([dilatation ] ) we obtain @xmath217 as for the self - gravitating gas where all particles have the same mass , the mass @xmath11 for the self - gravitating gas with two kinds of particles follows approximately the scaling law @xmath218 this indicates a fractal distribution with haussdorf dimension @xmath14 .
@xmath14 decreases with @xmath51 and @xmath52 from the value @xmath53 for the ideal homogeneous gas ( @xmath177 ) till @xmath219 in the canonical critical line .
the haussdorf dimension keeps decreasing beyond the canonical critical line in the stable phase of the microcanonical ensemble .
we plot in fig.[fractal2 ] the mass @xmath11 for @xmath11 greater than @xmath220 of the total mass of the gas for several values of @xmath51 choosing the ratio @xmath221 to be @xmath222 .
we exclude the region @xmath223 where the mass distribution is almost uniform .
this local uniformity is simply explained by the absence of gravitational forces at the origin @xmath178 due to the spherical symmetry .
[ cols="<,<,<,<",options="header " , ] table 2 .
the fractal dimension @xmath14 and the proportionality coefficient @xmath224 from a fit according to @xmath225 on the critical canonical line for different values of @xmath183 and @xmath184 according to the behaviour of the free energy near the critical line ( see fig . [ freeen ] ) , we find that the first derivatives of the free energy ( energy , pressure ) are continuous , while the second derivatives ( specific heats , compressibility ) are discontinuous . using eq.([energy ] ) and the form of the functions @xmath147 and @xmath148 near the critical line ( see fig .
[ densbound ] ) , we find that the energy has two branches @xmath226 and @xmath227 which behave near the critical point as , @xmath228 where @xmath229 is a positive constant . deriving the energy ( [ enercrit ] ) with respect to @xmath8 and recalling eq.([etai ] ) for @xmath10 , we obtain the two branches of the specific heat at constant volume @xmath230 here , @xmath231 and @xmath232 stand for the gaseous stable phase in the canonical ensemble and @xmath233 and @xmath234 correspond to a phase only realized in the microcanonical ensemble @xcite .
we see that @xmath235 at the critical point and hence @xmath236 is continuous at criticality while @xmath237 exhibits there an infinite discontinuity .
( as is clear , negative values of @xmath237 can not be realized in the canonical ensemble@xcite ) .
the generalization of the treatment given in previous sections to a self - gravitating gas with @xmath18 kinds of particles is straightforward .
we give below the mean field equations and the more relevant results .
the relevant parameters of the gas are now @xmath238 where @xmath239 is the number of particles of mass @xmath240 .
we assume that @xmath241 stay fixed while @xmath242 .
therefore , the ratios @xmath243 also stay fixed for @xmath244 .
the coupled mean - field integral equations for the densities of particles @xmath245 ( @xmath246 ) take the form @xmath247 where lagrange multipliers @xmath248 enforce the normalization of the densities .
eqs.([sadeq - n ] ) give for the gravitational potential @xmath250 = -\frac{t}{m_2}\left [ \phi_2({\bf x})-a_2\right ] = ... = -\frac{t}{m_n}\left[\phi_n({\bf x})-a_n\right]\;.\ ] ] setting , @xmath251 \ ; , \ ; 1 \leq i \leq n\ ] ] and applying to eqs.([sadeq - n ] ) laplace operator , we find the partial differential equations @xmath252 using eq.([potentieln ] ) , we reduce eqs.([poissonn ] ) to a single equation @xmath253 eqs .
( [ poissonn ] ) are scale covariant . if @xmath254 are solutions of eqs.([poissonn ] ) , then @xmath255 defined by @xmath256 are also solutions of eq.([poissonn ] ) .
this property is due to the scale behaviour of newton s potential . in the spherically symmetric case
the mean field equations ( [ poissonn ] ) become ordinaries non - linear differential equations @xmath257 using the scale covariance of the mean field eqs.([poissonn ] ) by the transformation ( [ scaletransn ] ) , we can set @xmath258 with new parameters @xmath259 . in this way , the mean field eqs.([poissonradn ] ) become a reduced system of the form @xmath260 let us find the boundary conditions for these equations . in order to have a regular solution at the origin we impose @xmath261 the system ( [ poissonchin ] ) is invariant under the transformation @xmath262 hence , we can choose @xmath263 without loosing generality .
as in the case of two kinds of particles , the remaining boundary conditions @xmath264 are not independent from @xmath265 . the normalization ( [ normn ] ) of the densities of the @xmath18 kinds of particles @xmath266 has to be imposed .
we obtain from eqs.([phin ] ) and ( [ dilatationn ] ) , @xmath267 using the reduced mean field equations ( [ poissonchin ] ) , it is straightforward to show from eq.([etai - normn ] ) that @xmath268 hence , @xmath269 from eq.([champn ] ) we introduce new @xmath125-independent parameters @xmath270 these new parameters are only function of @xmath271 .
notice that the boundary conditions can be written as , @xmath272 the reduced mean field equations ( [ poissonchin ] ) become a single ordinary differential equation with its coefficients depending on the parameters @xmath273 , @xmath274 and the boundary conditions @xmath275 .
[ here , @xmath276 . using eqs.([phin ] ) and ( [ dilatationn ] )
we can express the densities of the @xmath18 kinds of particles in terms of the solution of eq.([eqchi1n ] ) @xmath277 the thermodynamic functions are expressed as functions of the density of particles at the boundary @xmath278 depending on @xmath271 .
that is , @xmath279 we provide below the expressions for the free energy , the gravitational energy , the entropy and the equation of state .
@xmath280 \cr \cr
\label{energyn } e_p&=&3 t \sum_{i=1}^n n_i \ ; ( f_i-1 ) \\ \cr
s&=&s_0+\sum_{i=1}^n n_i \ ; [ \;6(f_i-1)-\ln{f_i}+\sum_{j=1}^n \frac{m_i}{m_j } \
; \eta_j^r\ ; ] \label{pessureextn}\cr \cr
\frac{pv}{t}&=&\sum_{i=1}^n n_i \ ; f_i\end{aligned}\ ] ] where @xmath281}\ ] ] is the free energy and @xmath282}+\frac{5}{2}\right)\ ] ] is the entropy of the perfect gas with masses @xmath283 .
combining eqs.([energyn ] ) and ( [ pessureextn ] ) yields the virial theorem @xmath284
the self - gravitating gas with two kinds of particle has analogous qualitative properties to the self - gravitating gas with one kind of particles .
physical quantities like energy , free energy and entropy turn to be the sum of a term proportional to @xmath37 plus another term proportional to @xmath38 for large @xmath285 and @xmath21 provided @xmath1 and @xmath2 are kept fixed .
all physical quantities are expressed as functions of @xmath9 and @xmath10 . instead of a critical line as for one kind of particles , we have here a critical line in the @xmath15 plane for the canonical ensemble .
this line is associated to the jeans instability .
the equation of state exhibits a rim on this critical line [ see fig.[pressure3d ] ] .
the thermodynamic functions exhibit a two - sheeted structure as functions of @xmath51 and @xmath52 .
the branch points are on the critical line .
the specific heat is discontinuous and diverges there while the free energy is finite and continuous in the branch line . the local pressure and the local densities of particles are related by the same equation as in a perfect gas [ see sec .
3.5 ] : @xmath286 this can be explained by the dilute character of the self - gravitating gas in thermal equilibrium : @xmath287 for @xmath288 .
this dilution damps the effective interparticle interaction and allows a free particle behaviour .
the particle distribution is inhomogeneous and scales with @xmath50 with a haussdorf dimension @xmath289 .
the haussdorf dimension @xmath289 decreases for increasing @xmath290 and @xmath291 .
its value on the critical line @xmath292 turns to be independent of the ratio @xmath293 implying an universal behaviour .
@xmath289 takes there the same value than for the canonical critical point with one kind of particles @xcite .
the goal of this appendix is to compute the expression @xmath294 using eq.([potentiel ] ) we obtain @xmath295 + \frac{n_2}{2 } \ ; ( a_2-\frac{m_2}{m_1}\ ; a _ 1 ) \;.\end{aligned}\ ] ] using eq.([poisson ] ) in the spherical symmetry the first term of eq.([a1 ] ) becomes @xmath296 using eqs .
( [ dilatation ] ) and ( [ gauss ] ) we obtain @xmath297 integrating by parts ( [ a2 ] ) and using eqs.([dilatation ] ) and ( [ phimax ] ) we obtain @xmath298 = \\ & & \frac{n_1}{2}\left(1+\frac{m_2 n_2}{m_1 n_1 } \right ) \ ; \ln{f_1}+ \frac{n_1}{2 \lambda \eta_1^r}\int_0^{\lambda } { \rm d}x \ ; x^2 \ ; [ \chi_1^{'}(x)]^2 \;.\nonumber\end{aligned}\ ] ] using eq.([lagrange ] ) the second term of ( [ a1 ] ) yields @xmath299 using eqs .
( [ a3 ] ) and ( [ a4 ] ) we express @xmath300 as @xmath301 ^ 2\;.\ ] ] we compute now @xmath302 ^ 2\;.\ ] ] using eqs.([poissonchi ] ) and ( [ champ ] ) , we derive the function @xmath303\ ] ] and find @xmath304 ^ 2 + 3 \ , x^2 [ e^{\chi_1(x)}+\mu^2 e^{\chi_2(x ) } ] \ ; .\ ] ] we integrate @xmath305 between @xmath306 and @xmath125 . integrating by parts the first term and using eq.([gauss ] ) ,
we find @xmath307 the second term of eq.([b ] ) yields @xmath308 using eq.([etai - norm ] ) the third term of ( [ b ] ) yields @xmath309 hence , @xmath310 -\lambda ( \eta_1^r+\mu \ ; \eta_2^r)^2 + 6\lambda(\eta_1^r+\mu^2 \ ; \eta_2^r)\;.\ ] ] therefore , using eqs.([etai ] ) and ( [ densper ] ) we obtain @xmath311 ^ 2&=&n_1 \left [ 3(1-f_1)-\frac{1}{2 } ( \eta_1^r+\mu \ ; \eta_2^r ) \right ] \\ & = & n_2 \left [ 3(1-f_2)-\frac{1}{2 } \left(\frac{1}{\mu } \ ; \eta_1^r+\eta_2^r\right ) \right]\;. \nonumber\end{aligned}\ ] ] hence , the expression of @xmath300 [ using eqs.([a5 ] ) and ( [ search ] ) ] is @xmath312 \nonumber\\ & + & n_2 \left[\frac{1}{2}\ln{f_2}+ 3(1-f_2)-\frac{1}{2 } \left(\frac{1}{\mu } \ ; \eta_1^r+\eta_2^r\right ) \right]\;. \nonumber\end{aligned}\ ] ] the hydrostatic equilibrium condition @xcite @xmath313 \ ; \nabla u({\vec q})\ ; , \ ] ] where @xmath314 stands for the pressure , combined with the equation of state for the ideal gas in local form @xmath315\ ; , \ ] ] yields for the particle densities @xmath316 where @xmath317 and @xmath318 are constants . inserting this relation into the poisson equation @xmath319\ ] ] yields in dimensionless
coordinates eq.([hidro ] ) .
h. j. de vega , n. snchez and f. combes , nature , * 383 * , 56 ( 1996 ) , phys . rev . * d54 * , 6008 ( 1996 ) , ap .
j. * 500 * , 8 ( 1998 ) , in ` current topics in astrofundamental physics : primordial cosmology ' , nato asi at erice , n. snchez and a. zichichi editors , vol 511 , kluwer , 1998 . | we study the statistical mechanics of the self - gravitating gas at thermal equilibrium with two kinds of particles .
we start from the partition function in the canonical ensemble which we express as a functional integral over the densities of the two kinds of particles for a large number of particles .
the system is shown to possess an infinite volume limit when @xmath0 , keeping @xmath1 and @xmath2 fixed .
the saddle point approximation becomes here exact for @xmath0 .
it provides a nonlinear differential equation on the densities of each kind of particles . for the spherically symmetric case
, we compute the densities as functions of two dimensionless physical parameters : @xmath3 and @xmath4 ( where @xmath5 is newton s constant , @xmath6 and @xmath7 the masses of the two kinds of particles and @xmath8 the temperature ) . according to the values of @xmath9 and
@xmath10 the system can be either in a gaseous phase or in a highly condensed phase .
the gaseous phase is stable for @xmath9 and @xmath10 between the origin and their collapse values .
we have thus generalized the well - known isothermal sphere for two kinds of particles .
the gas is inhomogeneous and the mass @xmath11 inside a sphere of radius @xmath12 scales with @xmath12 as @xmath13 suggesting a fractal structure .
the value of @xmath14 depends in general on @xmath9 and @xmath10 except on the critical line for the canonical ensemble in the @xmath15 plane where it takes the universal value @xmath16 for all values of @xmath17 .
the equation of state is computed .
it is found to be * locally * a perfect gas equation of state .
the thermodynamic functions ( free energy , energy , entropy ) are expressed and plotted as functions of @xmath9 and @xmath10 .
they exhibit a square root riemann sheet with the branch points on the critical canonical line .
the behaviour of the energy and the specific heat at the critical line is computed .
this treatment is further generalized to the self - gravitating gas with @xmath18-types of particles . |
muonic atoms are like ordinary atoms except that the electrons are replaced with muons .
these atoms have been studied in many fundamental physics experiments , such as precision measurements of proton size @xcite and nuclear quadrupole moments @xcite .
muonic atoms with pions in the core have been produced from intense @xmath3 beam at brookhaven national lab @xcite and fermilab @xcite .
however hydrogen - like muonic atoms with more exotic particles in the core ( kaons or antiproton ) have never been observed .
heavy - ion experiments , with large amount of thermal muons and hadrons produced , make an ideal environment for the production of such exotic atoms .
this provides us a great opportunity to make these discoveries .
muons that are involved in the atom production make the study particularly interesting in heavy - ion experiments , because thermal leptons are considered to be ideal penetrating probes of hot qcd matter as their production rates rapidly increase with the temperature of the medium . however , one difficulty of measuring the thermal leptons is that they are mixed with a large amount of leptons from weak hadronic decays , which carry little information of the hot and dense matter .
muonic atoms are only produced by particles right after freeze - out , i.e. hadrons and thermal muons or muons from resonance decays like @xmath4 , not by the muons from the weak hadronic decays , which are produced at a relatively late stage .
thus the idea of measuring the distributions of muonic atoms in heavy - ion collisions has been suggested by several theorists , melvin schwartz , gordon baym , gerald friedman , @xcite , joseph kapusta , agnes mocsy , @xcite etc .
the dataset used in this analysis is from au+au collisions at @xmath0 gev collected by the star detector in year 2010 .
central triggered events are selected to maximize the particle multiplicities .
a total of 231 million events passed the event level selections .
particles are identified from the time - of - flight detector and the time projection chamber .
the muon momentum is limited to 0.15 - 0.25 gev / c to ensure the purity of the sample .
the corresponding momenta for kaons and protons / antiprotons are 0.7 - 1.17 gev / c and 1.33 - 2.22 gev / c respectively .
the invariant mass reconstruction is done with the combinatorial method .
the combinatorial signal is constructed by pairing a hadron and a muon with opposite electric charges ( unlike - sign method ) from the same event .
the background is constructed in two ways : a mixed - event method , in which a hadron and a muon with opposite electric charges from two different events are paired ; and a like - sign method , in which a hadron and a muon with the same electric charge from a same event are paired .
note that the coulomb effect becomes stronger when the two charged particles are close in phase space . in the unlike - sign method ,
two particles carry opposite charges , which produce attractive coulomb force and thus enhance the mass distributions , especially at the low mass region .
in contrast , in the like - sign method , the repulsive coulomb force from the same charge suppresses the mass distributions at the low mass region . in the mixed - event method
, there is no coulomb effect for hadron - muon pairs .
therefore , the mixed - event backgrounds are used for acceptance correction of like - sign backgrounds : @xmath5 where @xmath6 and @xmath7 stands for like - sign and mixed - event respectively , and the index stands for the charges for hadrons and leptons .
details of this correction are discussed in @xcite @xcite .
0.5 show peaks at the atom masses.,title="fig : " ] 0.5 show peaks at the atom masses.,title="fig : " ] we adopted the following observable to cancel the trivial coulomb effect and preserve the signal : @xmath8 where @xmath9 stands for unlike - sign @xmath10 like - sign , to cancel the coulomb effect , and me stands for mixed - event for normalization . after the rejection of the coulomb force , we observe the sharp peaks at the expected zero net mass in fig .
[ fig : mass ] .
the error bars show the statistical uncertainties .
the signal is robust in both @xmath11-@xmath2 and @xmath12-@xmath2 systems and their antimatter systems .
femtoscopic correlations between two particles have also been used as a probe of muonic atoms .
the correlation as a function of @xmath13 , which is the magnitude of the momentum of either particle in the pair rest frame , shows how the interactions of the two particles change with respect to their distance in phase space .
star has thoroughly studied the @xmath11-@xmath1 system @xcite , in which only coulomb interation dominates . for non - identified particles ,
a leading particle can be selected , and two cases can be distinguished by @xmath14 and @xmath15 , which stand for the leading particle travels faster and slower , respectively . then the double ratio @xmath16 can be calculated to show the difference of the two cases .
this method was successfully used in previous measurements to probe the space - time asymmetry of the emission of two particles @xcite .
[ sec : figures ] r0.55 the measurement of @xmath11-@xmath1 system is performed from the dataset and kinetic region similar as @xmath11-@xmath2 system which we used for muonic atom signal extraction .
the origin of the non - unity in double ratio comes from the coulomb interactions between the kaons and pions , which are later on enhanced in @xmath17 and suppressed in @xmath18 because of the space - time emission asymmetries of kaons and pions .
the double ratio of @xmath19 system , overlaid on top of @xmath11-@xmath1 system , can be separated in two regions as shown in fig .
[ fig : k4 ] . on the right of the dashed line , where only coulomb interactions are expected in both systems , the double ratios of the two systems are consistent .
this is consistent with the existence of the coulomb force , which is a necessary condition to form muonic atoms . on the left of the dashed line , where the muonic atoms are expected to appear ,
when getting to very low @xmath13 , instead of divergence , the double ratios of @xmath19 system show convergence to unity .
the unity double ratio provides a signature of muonic atoms disassociation at the detector beam pipe , where the hadrons and the muons are separated from the bound state at the same space - time point .
the @xmath1-@xmath2 correlations are also studied .
a large amount of muons from weak decays can pass the track selections , and mix with the primordial muons .
thus these @xmath1-@xmath2 interactions inherit the interactions from @xmath1-@xmath1 interactions , which have two major sources , the electrostatic coulomb interactions and quantum interference from identical pions @xcite .
the later factor generates a strong enhancement on the correlation functions .
we denote the three correlation functions as the follows : * @xmath20 for correlations between pions and muons from simulated weak decays from real pions .
* @xmath21 for correlations between pions and inclusive muons , which is measured from data . * @xmath22 for correlations between pions and primordial muons .
[ sec : figures ] r0.55 the three functions satisfy the linear relationship : @xmath23 , where @xmath24 stands for the fraction of primordial muons from inclusive muons produced from the collisions .
@xmath22 is then estimated by @xmath1-@xmath1 correlations , because of the fact that pion mass and muon mass are fairly close . to avoid quantum statistics enhancement , the correlation function from coulomb between like - sign pairs
is estimated from reversed unlike - sign pairs @xmath25 .
the relation then becomes : @xmath26 .
the minimum @xmath27 fitting is performed in fig .
[ fig : fraction ] .
if two particles have similar trajectories and orientation , implying that they are close in momenta space , the detector will not be able to have enough spacial resolution to distinguish them and will merge the two tracks .
the fitting range is selected between 0.02 - 0.2 gev / c , discarding the very low @xmath13 where the missing track problem is significant .
the fitting results show that the fraction of primordial muons is [email protected]% .
au+au collisions at @xmath0 gev collected by the star experiment are used in this measurement to search for muonic atoms .
the invariant mass distributions show clear signals at the expected mass position for @xmath29-@xmath30 , @xmath31-@xmath32 , @xmath12-@xmath30 , and antiproton-@xmath32 .
the signal is robust after long - range coulomb effect is rejected .
the double ratio of @xmath11-@xmath2 system indicates the kaons and muons that are very close in phase space are emitted at the same space and time , which is consistent with muonic atom ionization .
the fraction of primordial muons is extracted from the correlation method .
99 aldo antognin , _
et al . _ ,
science * 339 * 417 ( 2013 ) .
y. tanaka and b. m. steffen , phys .
* 51 * 18 ( 1983 ) .
r. coombes , _
et al . _ ,
lett . * 37 * 5 ( 1976 ) .
s. h. aronson , _
et al . _ ,
lett . * 48 * 16 ( 1982 ) .
g. baym , g. freidman , r. j. hughes , and b. jack , phys . rev . *
d48 * 9 ( 1993 ) .
j. kapusta and a. mocsy , phys .
c * 59 * 5 ( 1999 ) .
a. adare , _
c * 81 * 5 ( 2010 ) .
l. adamczyk , _
et al . _ ,
c * 86 * 5 ( 2012 ) .
j. adams , _
et al . _ ,
lett . * 91 * 26 ( 2003 ) .
j. adams , _
et al . _ ,
c * 71 * 044906 ( 2005 ) . | we present the search results for muonic atoms on @xmath0 gev au+au collisions collected by the star experiment at rhic . with the muon identification at low momentum ,
the invariant mass spectra were reconstructed .
clear signals are observed at the expected atom masses .
two particle correlations show that the production of the daughter particles happens at the same space - time point , presenting the signature of atom ionization .
the fraction of primordial muons is extracted from @xmath1-@xmath2 correlations . |
the two - dimensional electron gas ( 2deg ) in a perpendicular magnetic field has a very rich phase diagram that includes several phases such as the laughlin liquids that give rise to the integer and quantum hall effectsgirvin1 , the wigner crystal at small filling factor in each landau level@xcite , the bubble crystals and the stripe phase in higher landau levels@xcite and the skyrme crystal@xcite near filling factor @xmath1 in the lowest landau level .
the phase diagram is even more complex when system with extra degrees of freedom such as double quantum wells ( dqws ) are considered@xcite . in dqws
, the orientation of the pseudospin vector associated with the layer degree of freedom can be modified by changing the tunneling and electrical bias between the layers .
another way to modify the properties of the 2deg is by the addition of a lateral two - dimensional superlattice patterned on top of the gaas / algaas heterojunction hosting the 2deg that creates a spatially modulated potential at the position of the 2deg@xcite .
the effect of a one - dimensional periodic potential on the landau levels is particularly interesting@xcite since it leads to commensurability problems due to the presence of different lengths scales : the lattice constant of the external potential @xmath11 the magnetic length @xmath12 ( @xmath13 is the magnetic field ) and the fermi wavelength .
novel magnetoresistance oscillations with period different than that of the well - known shubnikov - de haas oscillations have been detected in such systems .
even more interesting is the effect of a periodic two - dimensional potential on the band structure of the 2deg@xcite .
the intricate pattern of eigenvalues that results from such potential has been studied by many authors and is known as the hofstadter butterfly spectrum@xcite .
its observation in gaas / algaas heterojunction is very difficult due to screening and disorder effects but experimental signature in magnetotransport experiments in 2degs with a lateral surface superlattice potential with period of the order of @xmath14 nm and less have been reported@xcite .
interest in this problem has been revived recently by the experimental observations of the hofstadter s butterfly spectrum that use the moir superlattices that arise from graphene or bilayer graphene placed on top of hexagonal boron nitride@xcite .
another interest of superlattice potentials is their use to create artificial lattices .
for example , a lateral superlattice with a honeycomb crystal structure has recently been proposed to create an artificial graphenelike system@xcite in a gaas / algaas heterojunction . in this work ,
we study theoretically the effect of a square lattice lateral potential with a period @xmath15 on the ground state of the 2deg in gaas / algaas heterojunction at filling factor @xmath1 and in landau level @xmath2 we include the spin degree of freedom and use the hartree - fock approximation to study the combined effects of the external potential and the coulomb interaction .
we assume that the potential is sufficiently weak so that landau level mixing can be neglected .
we also ignore disorder and work at zero temperature .
we vary the potential strength @xmath3 and calculate the ground state for different rational values of @xmath16 $ ] ( where @xmath17 is the flux quantum and @xmath18 are integers with no common factors ) which is the inverse of the number of flux quanta per unit cell of the surface potential .
our formalism allows for the formation of uniform as well as spatially modulated ground states with or without spin texture .
our calculation indicates that , at a critical value , @xmath4 of the external potential , there is a transition from a uniform fully spin polarized state to a charge density wave ( cdw ) with an intricate spin texture .
each unit cell of this cdw contains two positive and two negative amplitude modulations and the vortex spin texture at each maximum(minimum)resembles that of a positively(negatively ) charged meron .
the two positively(negatively ) charged merons in each unit cell have the same vorticity but a global phase that differs by @xmath19 these meronlike textures , however , are not quantized since the amplitude of the cdw varies continuously with @xmath3 . in the vortex - cdw , as we call it , the average spin polarization @xmath5 varies with @xmath3 in a continuous or discontinuous manner depending on whether @xmath6 $ ] or @xmath20 $ ] and saturates at a finite , positive , value of @xmath5 that depends only on @xmath21 in most cases . in the special case @xmath22 the vortex - cdw phase is absent and the transition is directly from a fully spin polarized and uniform 2deg to an unpolarized cdw . the phase diagram for @xmath20 $ ] is richer than that of @xmath6 $ ] as it involves the transition between the vortex - cdw and its conjugate phase , the anti - vortex cdw , obtained by reversing @xmath23 and inverting the vorticity of all merons .
this transition between the two cdws is accompanied by a discontinuous change of @xmath5 that becomes continuous when the zeeman coupling goes to zero .
we study the properties of the vortex - cdw at different values of @xmath21 and with a particular emphasis on its collective excitations which we derive using the generalized random - phase approximation ( grpa ) .
the vortex - cdw has collective modes that have much in common with the collective excitations of the skyrme crystal@xcite that is expected to be the ground state near filling factor @xmath1 in @xmath2 namely , the broken u(1 ) symmetry in the vortex - cdw phase leads to a new gapless mode that can provide a fast channel for the relaxation of nuclear spins@xcite .
this mode and the meronlike spin texture disappear at larger values of the external potential leaving a ground state that is either unpolarized if @xmath24 or partially polarized if @xmath25 our paper is organized as follows . in sec .
ii , we introduce the hamiltonian of the 2deg in the presence of the lateral square lattice potential and briefly review the hartree - fock and generalized random - phase approximation that we use to compute the density of states , the density and spin profiles and the collective excitations of the various phases . in sec .
iii , we present our numerical results for the phase diagram of the 2deg as a function of the potential strength @xmath3 and the inverse magnetic flux per unit cell @xmath21 .
we conclude in sec .
iv with a discussion on the experimental detection of the new vortex - cdw state .
the system we consider is a 2deg in a gaas / algaas heterojunction or quantum well submitted to a perpendicular magnetic field @xmath26 and to a lateral superlattice potential @xmath27 the coupling of the electrons to this external potential is given by @xmath28 , where @xmath29 is the total density operator including both spin states @xmath30 ( we take @xmath31)@xmath32 we assume that only the landau level @xmath33 is occupied but our calculation can easily be generalized to any landau level by changing the effective interactions @xmath34 and @xmath35 and the form factor @xmath36 the hamiltonian of the interacting 2deg is given , in the hartree - fock approximation , by @xmath37where @xmath38 is the 2deg area , @xmath39 is the landau level degeneracy , and the form factor for the @xmath33 landau level is@xmath40where @xmath12 is the magnetic length .
the averages are over the hartree - fock ground state of the 2deg .
the non - interacting single - particle energies , measured with respect to the kinetic energy @xmath41 are given by@xmath42where the zeeman energy @xmath43 with @xmath44 the effective @xmath45factor of bulk gaas and @xmath46 the bohr magneton .
in some experiments on skyrmions , the effective @xmath45factor was tuned in the range @xmath47 to @xmath48 by applying hydrostatic pressure to a sample of gaas / algaas modulation doped quantum wellpotemski . in our study
, we will thus consider that @xmath49 is not determined by the magnetic field , but is instead a parameter than can be adjusted .
the hartree and fock interactions in @xmath33 are given by@xmath50where @xmath51 is the dielectric constant of gaas . finally , the operators @xmath52 are defined by@xmath53where @xmath54 is the operator that creates an electron with guiding - center index @xmath55 ( in the landau gauge ) and spin @xmath56 the four operators @xmath52 are related to the averaged electronic and spin densities in the @xmath57plane by@xmath58 , \\
s_{y}\left ( \mathbf{r}\right ) & = & \frac{1}{2\pi \ell ^{2}}\sum_{\mathbf{q}}{\operatorname{im}}\left [ \left\langle \rho _ { + , -}\left ( \mathbf{q}\right ) \right\rangle e^{-q^{2}\ell ^{2}/4}e^{i\mathbf{q}\cdot \mathbf{r}}\right ] , \\
s_{z}\left ( \mathbf{r}\right ) & = & \frac{\hslash } { 2}\left [ n_{+}\left ( \mathbf{r}\right ) -n_{-}\left ( \mathbf{r}\right ) \right ] .\end{aligned}\]]the @xmath59 can be considered as the order parameters of an ordered phase of the 2deg .
the averaged hartree - fock ground - state energy per electron at @xmath1 is given by@xmath60 the order parameters @xmath61 are computed by solving the hartree - fock equation for the single - particle green s function @xmath62 which is defined by@xmath63where@xmath64they are obtained with the relation @xmath65 the hartree - fock equation of motion for the green s function @xmath66 is given@xcite by @xmath67 g_{\alpha , \beta } \left ( \mathbf{q},i\omega _ { n}\right ) \label{motion } \\ & = & \delta _ { \mathbf{q},0}\delta _ { \alpha , \beta } \notag \\ & & -\frac{e}{\hslash s}\sum_{\mathbf{q}^{\prime } } v_{e}\left ( \mathbf{q}-\mathbf{q}^{\prime } \right ) f\left ( \left\vert \mathbf{q}-\mathbf{q}^{\prime } \right\vert \right ) \notag \\ & & \times \gamma _ { \mathbf{q},\mathbf{q}^{\prime } } g_{\alpha , \beta } \left ( \mathbf{q}^{\prime } , i\omega _ { n}\right ) \notag \\ & & + \frac{1}{\hslash } \sum_{\mathbf{q}^{\prime } \neq \mathbf{q}}u^{h}\left ( \mathbf{q - q}^{\prime } \right ) \gamma _ { \mathbf{q},\mathbf{q}^{\prime } } g_{\alpha , \beta } \left ( \mathbf{q}^{\prime } , i\omega _ { n}\right ) \notag \\ & & -\frac{1}{\hslash } \sum_{\mathbf{q}^{\prime } } \sum_{\gamma } u_{\alpha , \gamma } ^{f}\left ( \mathbf{q - q}^{\prime } \right ) \gamma _ { \mathbf{q},\mathbf{q}^{\prime } } g_{\gamma , \beta } \left ( \mathbf{q}^{\prime } , i\omega
_ { n}\right ) , \notag\end{aligned}\]]where@xmath68and @xmath69 are fermionic matsubara frequencies , @xmath70 is the chemical potential and we have defined the potentials @xmath71 these potentials depend on the order parameters @xmath72 that are unknown .
the equation of motion for @xmath73 must thus be solved numerically@xcite by using a seed for the order parameters and then iterate eq .
( [ motion ] ) until a convergent solution is found . in case
several solutions are found ( corresponding to different choice for the initial seed ) , we choose the one with the lowest energy and compute the dispersion relation of its collective modes to make sure that it is a stable solution .
we remark that this method does not guarantee that the true ground state is the solution that we keep .
the density of states @xmath74 is obtained from the single - particle green s function by using the relation @xmath75 .\ ] ] to find the dispersion relation of the collective modes , we derive the equation of motion in the generalized random - phase approximation for the two - particle green s function@xmath76this equation is@xmath77where @xmath78 is a bosonic matsubara frequency .
equation ( [ grpa ] ) represents the summation of bubble ( polarization effects ) and ladder ( excitonic corrections ) diagrams . the hartree - fock two - particle green s function ( the single - bubble feynman diagram with hartree - fock propagators ) that enters this equation
is obtained from the hartree - fock equation of motion for @xmath79 and is given by @xmath80 \chi _ { \alpha , \beta , \gamma , \delta } ^{\left ( 0\right ) } \left ( \mathbf{q},\mathbf{q}^{\prime } , i\omega _ { n}\right ) \\ & = & \hslash \left [ \gamma _ { \mathbf{q},\mathbf{q}^{\prime } } ^{\ast } \left\langle \rho _ { \alpha , \delta } \left ( \mathbf{q - q}^{\prime } \right ) \right\rangle \delta _ { \beta , \gamma } -\gamma _ { \mathbf{q},\mathbf{q}^{\prime } } \left\langle \rho _ { \gamma , \beta } \left ( \mathbf{q - q}^{\prime } \right ) \right\rangle \delta _ { \alpha , \delta } \right ] \notag \\ & & -\frac{e}{s}\sum_{\mathbf{q}^{\prime \prime } } v_{e}\left ( \mathbf{q}-\mathbf{q}^{\prime \prime } \right ) f\left ( \left\vert \mathbf{q}-\mathbf{q}^{\prime \prime } \right\vert \right ) \left [ \gamma _ { \mathbf{q},\mathbf{q}^{\prime \prime } } ^{\ast } -\gamma _ { \mathbf{q},\mathbf{q}^{\prime \prime } } \right ] \notag \\ & & \times \chi _ { \alpha , \beta , \gamma , \delta } ^{\left ( 0\right ) } \left ( \mathbf{q}^{\prime \prime } , \mathbf{q}^{\prime } , i\omega _ { n}\right )
\notag \\ & & -\sum_{\mathbf{q}^{\prime \prime } \neq 0}u^{h}\left ( \mathbf{q - q}^{\prime \prime } \right ) \left [ \gamma _ { \mathbf{q},\mathbf{q}^{\prime \prime } } ^{\ast } -\gamma _ { \mathbf{q},\mathbf{q}^{\prime \prime } } \right ] \notag \\ & & \times \chi _ { \alpha , \beta , \gamma , \delta } ^{\left ( 0\right ) } \left ( \mathbf{q}^{\prime \prime } , \mathbf{q}^{\prime } , i\omega _ { n}\right )
\notag \\ & & + \sum_{\alpha ^{\prime } } \sum_{\mathbf{q}^{\prime \prime } } u_{\alpha , \alpha ^{\prime } } ^{f}\left ( \mathbf{q - q}^{\prime \prime } \right ) \gamma _ { \mathbf{q},\mathbf{q}^{\prime \prime } } ^{\ast } \chi _ { \alpha ^{\prime } , \beta , \gamma , \delta } ^{\left ( 0\right ) } \left ( \mathbf{q}^{\prime \prime } , \mathbf{q}^{\prime } , i\omega _ { n}\right ) \notag \\ & & -\sum_{\beta ^{\prime } } \sum_{\mathbf{q}^{\prime \prime } } u_{\beta ^{\prime } , \beta } ^{f}\left ( \mathbf{q - q}^{\prime \prime } \right ) \gamma _ { \mathbf{q},\mathbf{q}^{\prime \prime } } \chi _ { \alpha , \beta ^{\prime } , \gamma , \delta } ^{\left ( 0\right ) } \left ( \mathbf{q}^{\prime \prime } , \mathbf{q}^{\prime } , i\omega _ { n}\right ) .
\notag\end{aligned}\ ] ] by defining the super - indices @xmath81 and @xmath82 eq . ( [ grpa ] ) can be rewritten as a @xmath83 matrix equation for the matrix of green s functions @xmath84 this equation has the form @xmath85 \chi = b.$ ] the matrix @xmath86 , which depends only on the @xmath87 is then diagonalized numerically to find @xmath88 .
the retarded response functions are obtained with the analytic continuation @xmath89 we compute the following density and spin responses:@xmath90 \right\rangle \theta \left ( t - t^{\prime } \right ) \right ] _ { \omega } , \label{response}\]]where @xmath91 and the operators @xmath92 , \\
\rho _ { x } & = & \frac{1}{2i}\left [ \rho _ { + , -}-\rho
_ { -,+}\right ] , \\ \rho _ { z } & = & \frac{1}{2}\left [ \rho _ { + , + } -\rho _ { -,-}\right ] , \\
\rho _ { n } & = & \rho _ { + , + } + \rho _ { -,-}.\end{aligned}\ ] ] in a uniform phase , the order parameters @xmath93 are finite only when @xmath94 while in a two - dimensional cdw , they can be non zero each time @xmath95 , where @xmath96 is a reciprocal lattice vector of the cdw .
for the response function , we have to compute @xmath97 in the uniform phase and @xmath98 in the cdw where @xmath99 is , by definition , a vector in the first brillouin zone of the cdw . in the cdw , the grpa matrices @xmath86 and @xmath13 have dimensions @xmath100 , where @xmath101 is the number of reciprocal lattice vectors considered in the numerical calculation .
we typically take @xmath102 the formalism developed in this section can also be applied to the 2deg in graphene if the electrons are assumed to occupy only one of the two valleys . in landau level @xmath33 of graphene ( and only in this level ) , the form factor @xmath103 and the hartree and fock interactions @xmath104 and @xmath35 are the same as those given in eqs .
( [ form],[inter ] ) .
we study the phase diagram of the 2deg at filling factor @xmath1 in landau level @xmath33 and at temperature @xmath106 k. for the external potential , we choose the simple square lattice form@xmath107 , \]]so that @xmath108 in eq .
( [ hamiltonien ] ) with the vectors @xmath109 this external potential tries to impose a two - dimensional density modulation of the 2deg with a square lattice constant @xmath110 we allow the spin texture ( if any ) to have the bigger lattice constant @xmath111by considering the order parameters @xmath112 with reciprocal lattice vectors given by @xmath113 where @xmath114 the _ density _ unit cell has a lattice constant @xmath15 while the _ magnetic _ unit cell has a lattice constant @xmath115 for the potential strength , we use @xmath116 [ see eq . ( motion ) ] , where @xmath117 .
the critical values of @xmath3 ( not @xmath118 ) for the transition between the uniform and the modulated phases at different value of @xmath21 are similar .
hereafter , we give all energies in units of @xmath119 .
we make the important assumption that landau level mixing by both the coulomb interaction and the external potential can be neglected i.e. we work in the limit of a weak superlattice potential .
we also neglect disorder effect and work at zero temperature .
the ratio @xmath120 that enters the hartree - fock energy and equation of motion for the single - particle green s function is given by@xmath121where @xmath17 is the flux quantum .
the important parameter @xmath122 is the number of flux quanta piercing a _ density _ unit cell area . with this definition , the factor
@xmath123 we limit our analysis to @xmath124 where @xmath125 and @xmath126 are integers with no common factors
. in landau level @xmath127 a wigner crystal@xcite with a triangular lattice can form at sufficiently small filling factor @xmath128@xcite . at @xmath129 however , the ground state of the 2deg is a uniform electron liquid with full spin polarization i.e. a quantum hall ferromagnet@xcite ( qhf ) whose energy per electron is given by@xmath130(neglecting the kinetic energy that is a constant in @xmath33 ) .
the qhf remains the ground state even when the zeeman coupling goes to zero because a perfect alignment of the spins minimizes the coulomb exchange energy [ the second term on the right - hand side of eq .
( [ fock ] ) ] . in a uniform state ,
the coulomb hartree energy is cancelled by the neutralizing uniform positive background .
we first consider the case @xmath6 .$ ] figure 1 shows the ground state energy and spin polarization @xmath131 as a function of the potential @xmath3 for @xmath132 and for a zeeman coupling @xmath133 .
the ground state is spatially uniform and has an energy @xmath134 and a spin polarization @xmath5 that remain constant until a critical field @xmath135 this uniform state is described by only one order parameter , i.e. @xmath136 and is fully spin polarized i.e. the spin per electron is @xmath137 the corresponding change in the density of states ( dos ) with @xmath3 is shown in fig .
2 . in the absence of external potential and coulomb interaction ,
the dos has two peaks at energies @xmath138 corresponding to the two spin states .
with coulomb interaction , the zeeman gap @xmath49 is strongly renormalized [ see fig .
2 ( a ) ] as is well known .
when the external potential is present , the dos for each spin orientation , has @xmath126 peaks corresponding to the number of subbands expected when an electron is submitted to both a magnetic field and a weak periodic potentialhofstadter .
this is clearly visible in fig .
2 ( a),(d),(e ) for @xmath139 . the external potential increases the width of the peaks in the dos and decreases the renormalized zeeman gap ( which is also the transport gap ) .
we remark that the rapid oscillations in some of the graphs at @xmath140 are a numerical artefact .
they depend strongly on the number of reciprocal lattice vectors kept in the calculation . as a function of the applied external field @xmath3 for different values of @xmath21 at filling factor @xmath105 in landau level @xmath33 and for the zeeman coupling @xmath141 if we enforce the uniform solution beyond the critical value @xmath142 , we find that the transport gap closes at @xmath143 for @xmath144 where the system becomes metallic .
our code no longer converges in this case .
but , this transition to a metallic state does not occur because the uniform state becomes unstable at @xmath145 the stability of a state is evaluated by computing the dispersion relation of its collective modes .
for the uniform state , the collective excitations reduce to a spin - wave mode@xmath32 when @xmath146 the spin - wave dispersion is given by@xcite in landau level @xmath33 for zeeman coupling @xmath147 ( a)-(c ) @xmath144 and @xmath148 respectively ; ( d)-(e ) @xmath149 and @xmath150 respectively ; ( f ) @xmath151 and @xmath152 .
the rapid oscillations in some of the graphs are a numerical artefact .
all energies are in units of @xmath153 @xmath154 . \label{spinwave}\ ] ] this mode is gapped at the bare zeeman energy and saturates at @xmath155 figure 3 shows its dispersion for @xmath144 and @xmath156 the wave vector @xmath99 runs along the path @xmath157 i.e. along the edges of the irreducible _ density _ brillouin zone ( with @xmath158 in units of @xmath159 ) .
the spin - wave mode softens at a finite wave vector @xmath99 as @xmath160 increases so that the uniform state becomes unstable at @xmath161 which is also the value at which the 2deg is seen to enter a new phase in fig .
when plotted in the reduced zone scheme as in fig .
3 , the spin - wave mode is split into several branches that accumulate into a very dense manifold near @xmath162 . only some of these branches are shown in fig .
3 since we are interested only in the low - energy sector .
the spin - wave dispersion is obtained by following the pole of the response functions @xmath163 with @xmath164 for different values of @xmath165 we remark that the softening of the spin - wave mode by a _
one - dimensional _ external potential was reported previously by bychkov et al.@xcite .
these authors suggested that the resulting condensation of the spin excitons at the softening wave vector would create a new spin density wave ground state .
this is precisely what we find , but this time , for a _ two - dimensional _ surface potential . for @xmath166 @xmath133 and different values of the external potentiel @xmath3 .
the wave vector @xmath99 follows the path : @xmath167 along the irreducible ( density ) brillouin zone of the square lattice .
all energies are in units of @xmath168 the ground state in a small region of @xmath3 after @xmath169 is a charge density wave with a vortex spin texture .
hereafter , we refer to this state as the vortex - cdw .
the range of @xmath3 where the vortex - cdw is the ground state depends on @xmath21 and @xmath170the electronic density and spin texture of the vortex - cdw are shown in fig . 4 for the parameters @xmath144 and @xmath171 the value @xmath172 is close to @xmath173 so that the amplitude of the cdw in this figure is small .
the amplitude increases with @xmath3 however .
minima and maxima of the cdw have the same amplitude and there is no net induced charge in a unit cell as expected .
the spin density @xmath23 ( not shown in the figure ) varies only slightly .
the spin texture of the vortex - cdw is interesting .
there is a @xmath174 spin vortex at each positive and negative modulation of the density . since the @xmath175 component of the spin is everywhere positive and because of the spin - charge coupling inherent to a qhf@xcite , the positive and negative modulations have opposite vorticity
we could , loosely speaking , refer to the positive and negative modulations as merons and antimerons .
a meron is an excitation of a unit vector field @xmath176 that has @xmath177 at its center and @xmath178 far away from the center where the vectors lie in the @xmath57plane and form a vortex configuration with vorticity @xmath179 as @xmath180 increases from the meron core , the spins smoothly rotate up ( if @xmath181 ) or down ( if @xmath182 ) towards the @xmath57plane
. there are four flavors of meron@xcite with a topological charge given by @xmath183 n_{v}.$ ] in a qhf , merons carry half an electron charge . in our vortex - cdw
however , we are dealing with a spin density @xmath184 that does not have @xmath185 everywhere in space ( the vectors do not just rotate ) so that our merons do not have a quantized charge .
but , for a given sign of @xmath186 the meron and antimeron have opposite vorticity and so opposite electrical charge .
moreover , our merons are closely packed in a square lattice and have a large core so that the spin vector tilts towards the @xmath57plane but @xmath186 does not go to zero between two adjacent merons . in each magnetic unit cell of the vortex - cdw , there are two merons and two antimerons with the same vorticity but opposite global phase for two merons or antimerons .
this bipartite meron lattice is similar to the square lattice antiferromagnetic state ( sla ) of the skyrme crystal that was predicted to occur in a 2deg near ( but not at ) filling factor @xmath1 in the absence of an external potential@xcite . in the skyrme crystal
, the electrons ( or holes ) added to the qhf state at @xmath1 crystallize in the form of skyrmions for @xmath187 or antiskyrmions for @xmath188 .
the vortex - cdw that we find here occurs _ at _
precisely @xmath189 figure 1(b ) shows that the average spin @xmath5 decreases with @xmath3 in the vortex - cdw and saturates at the precise value @xmath190 for @xmath191 . in the saturation region for @xmath5 , the spin texture has disappeared and the cdw has very little modulation in @xmath192 we will call this phase , the normal - cdw .
there is no saturation for the two cases @xmath193 we assume that this is due to the fact that there is another phase very close in energy that wins over the vortex - cdw for @xmath194 in these two cases ( and probably at a larger value of @xmath3 for the other cases ) but we have not been able to stabilize this other phase .
we thus limit our analysis to the range @xmath195 $ ] for most values of @xmath21 in this work .
the change in @xmath5 induced by the external potential should be detectable experimentally .
in particular , the vortex - cdw is absent for @xmath24 and thus the 2deg makes a transition from a fully polarized to an unpolarized cdw with period @xmath15 instead of @xmath196 .
in landau level @xmath33 for @xmath172 and zeeman coupling @xmath197 the electronic density @xmath198 is in units of @xmath199 the the vector field shows the vortex structure in the @xmath57plane for the parallel component of the spin vector .
all energies are in units of @xmath168 the density of states for the vortex - cdw is shown in fig .
5 for @xmath149 and @xmath200 the subband structure gets more and more different from that of the uniform phase as @xmath3 increases [ compare with fig .
the electron - hole gap decreases slowly with @xmath3 in the vortex - cdw and normal - cdw phases .
in landau level @xmath33 for @xmath149 , zeeman coupling @xmath133 and different values of @xmath3 ( all energies are in units of @xmath201 ) . ]
figure 6 shows the hartree , fock ( or exchange ) , external potential and zeeman contributions to the total energy of the vortex - cdw and uniform state for @xmath144 and zeeman coupling @xmath197 the hartree energy is zero in the uniform phase and small in the vortex - cdw phase .
the zeeman energy is also very small in both phases .
the competition is between the fock and the external potential energies .
the former is minimal in the uniform state and increases with @xmath3 in the vortex - cdw phase .
the latter is zero in the uniform phase but decreases with @xmath3 in the vortex - cdw state .
figure 6 shows that the increase in exchange energy is more than compensated by the decrease in the external potential energy when the vortex - cdw is formed . of different contributions to the total energy of the uniform and vortex - cdw states for @xmath144 and the zeeman coupling @xmath197 ] the energy of the vortex - cdw state does not depend on the global phase of its vortices .
this @xmath203 symmetry , which is broken in a particular realization of the vortex - cdw state , leads to a gapless phase mode ( a goldstone mode ) .
this is clearly seen in fig .
7(b ) where the two modes for @xmath204 are the spin wave mode which is gapped at @xmath49 and the gapless phase mode . in fig . 7 , @xmath149 and the wave vector @xmath99 now follows the path @xmath205 along the edges of the irreducible _ magnetic _ brillouin zone of the square lattice [ with @xmath206 in units of @xmath207 . to obtain the dispersions in the cdw phases ,
we have computed the response functions @xmath208 with @xmath209 keeping the first @xmath210 reciprocal lattice vectors in the summation .
the summation allows the capture of modes that originate from a folding of the full dispersion into the first brillouin zone .
it also captures the electron - hole continuum@xcite that starts at the hartree - fock gap . in fig .
7 , we have cut the dispersions at a frequency corresponding to the onset of this continuum .
figure 7 shows the dispersion for : ( a ) the uniform phase , ( b ) the vortex - cdw , and ( c ) the normal - cdw . in landau level @xmath33 for @xmath149 and zeeman coupling @xmath197 ( a ) @xmath211
( b ) @xmath212 and ( c ) @xmath213 all energies are in units of @xmath168 the vortex and normal cdws have a phonon mode gapped by the external potential .
the branch we indicate as the gapped phonon mode in fig .
7(b ) has the strongest peak in the response function @xmath214 ( no summation over @xmath96 ) as @xmath215 while the spin wave and phase modes are stronger in @xmath216 @xmath217 and @xmath218 respectively . at @xmath219 for @xmath220 the ground state has transited to the normal - cdw and the phase mode is gapped as shown in fig .
7(c ) .
the phase diagram for @xmath20 $ ] is different from that of @xmath6 .$ ] for @xmath20 , $ ] we find a transition between two types of vortex - cdw phases .
the first vortex - cdw is the one described in the previous section , the second one , the antivortex - cdw has the sign of all vortices and @xmath5 inverted ( but different amplitude for the charge and spin modulations ) .
this antivortex - cdw evolves from a uniform state that has all spin down as shown in fig .
8 . at @xmath221
these two cdw are degenerate in energy . at finite zeeman coupling
, there is a crossing between the energy curves of these two phases .
the ground state thus evolves from the uniform state with all spins up , to the vortex - cdw , then to the antivortex - cdw and finally into the normal cdw .
figure 8 shows these transitions for the special cases of @xmath222 and @xmath223 the corresponding behavior of @xmath5 is also shown .
the region where @xmath5 varies in each graph is where the vortex-(or antivortex)-cdw is the ground state .
as the zeeman coupling gets smaller , this region increases .
the value of @xmath3 for the crossing between the two vortex - cdw states is shown by the dashed vertical line in the @xmath5 vs @xmath3 curves .
the average spin @xmath5 changes discontinuously at this point but this discontinuity goes to zero as @xmath224 the value of @xmath5 is always positive , however . for @xmath6 , $ ] the energy curve for the anti - vortex cdw is above that of the vortex - cdw for all values of @xmath3 .
the two curves merge at @xmath225 but there is no crossing between the two solutions . if we take advantage of the possibility of changing the value of the @xmath45factor independently of the magnetic field in gaas / algaas heterojunctions , then it is possible to reduce the zeeman coupling and , as fig .
9 clearly shows , to increase the transition region where the vortex - cdw is expected . per electron
( right ) of the vortex - cdw ( squares ) and antivortex - cdw ( triangles ) for @xmath222 and zeeman couplings :
( a),(b ) @xmath226(c),(d ) @xmath227(e),(f ) @xmath228 in units of @xmath229 the vertical dashed lines indicate the value of the potential @xmath3 at which the transition between the two vortex cdws takes place . ]
figure 9 shows the behavior of @xmath5 in the ground state for @xmath230 and a very small zeeman coupling @xmath231 the vertical dashed lines indicate where the transition between the vortex and antivortex - cdw phases occurs for each value of @xmath232 the spin starts at @xmath233 in the uniform state then decreases in the vortex - cdw ( open symbols in fig .
9 ) . when the antivortex - cdw replaces the vortex - cdw as the ground state of the system , the value of @xmath5 changes discontinuously .
this jump is more apparent for @xmath234 in fig .
9 . after this discontinuity , @xmath5 increases ( filled symbols in fig .
9 ) until it reaches the finite value @xmath235 at large @xmath8 a value that is independent of the zeeman coupling @xmath170 the behaviour of @xmath5 is not monotonous .
it the limit @xmath236 the @xmath5 curves for the vortex- and antivortex - cdws would cross at @xmath237 and there would be no discontinuity . in the special case @xmath22
the transition is directly from the uniform and fully polarized state with @xmath233 to the normal cdw where @xmath238 there is thus an important discontinuity in @xmath5 in this case . as a function of the applied external potential @xmath3 for several values of @xmath239 and zeeman coupling @xmath240 the vertical dashed lines indicate the potential strength @xmath3 for each value of @xmath21 where the transition from the vortex ( filled symbols ) to the anti - vortex cdw ( open symbols ) takes place . for @xmath22
the transition is from the uniform ( with @xmath241 to the normal cdw with @xmath238 ] as we mentioned above , our formalism can equally well be used to discuss the energy of the electron gas in landau level @xmath33 in graphene if the electrons are assumed to occupy only one valley@xmath32 an exact diagonalization study by ghazaryan and chakraborty@xcite for a 2deg in graphene finds transition between unpolarized and partially polarized ground states induced by the external potential when @xmath144 ( their @xmath243 ) .
the equation ( [ motion ] ) was also used@xcite to study the effect of coulomb interaction on the density of states for graphene in a modulated potential but the vortex - cdw state that we found was not considered in that work .
we have computed the phase diagram of the 2deg at @xmath1 in landau level @xmath33 in the presence of an applied external potential with a square lattice periodicity for several rational values of @xmath244 @xmath245 $ ] .
we restricted our analysis to @xmath195 .$ ] in this range , the 2deg evolves first from a uniform state with full spin polarization then to a vortex - cdw and ( if @xmath246 ) antivortex - cdw state and finally into a normal cdw with no spin texture but with a finite spin polarization @xmath5 if @xmath25 the change in the spin polarization @xmath5 with the applied field ( smooth for @xmath247 and abrupt for @xmath248 ) is one feature of the phase transition described in this work that should be measurable experimentally .
another one is the gapless spin mode due to the broken u(1 ) symmetry in the vortex - cdw phase .
the same mode occurs in a skyrme crystal . in that system
, it was shown that such mode could provide a fast channel for the relaxation of the nuclear spin in nuclear magnetic resonance experiments@xcite . indeed , the bare zeeman gap in the dispersion of the spin - wave mode is orders of magnitude larger than the nuclear spin splitting , impeding the creation of spin waves by nuclear spins .
the softening of the spin wave mode in the uniform phase may also lead to an increase in nuclear spin - lattice relaxation time as suggested by bychkov@xcite .
we have used @xmath3 for the external potential because the transition from the uniform to the vortex - cdw takes place at roughly the same value of @xmath249 when the potential is expressed in terms of @xmath3 .
the actual external potential however is @xmath250 this means that the critical field @xmath251 translates into different real critical fields for different values of @xmath21 i.e. from @xmath252 for @xmath234 to @xmath253 for @xmath254 it is not clear , then if our assumption of neglecting landau - level mixing can be justified for @xmath21 near unity .
we assumed that the external lattice parameter @xmath15 is fixed experimentally .
when @xmath21 is also given , all other parameters are determined : the magnetic field , the electronic density @xmath255 ( at @xmath1 ) and the ratio , @xmath256 of the coulomb interaction to the cyclotron energy : @xmath257where @xmath258 is the effective bohr radius and @xmath259 is the lateral superlattice constant in nm .
we used , for gaas : @xmath51 and @xmath260 ( where @xmath261 is the electron mass ) .
for @xmath262so that with a very small ( but physically feasible@xcite ) superlattice period of @xmath263 nm , we get @xmath264 t , @xmath265 @xmath266 while for @xmath267 we get @xmath268 t , @xmath269@xmath270 the magnetic field and density pose no problem , but @xmath271 is not small , especially when @xmath272 clearly , a more sophisticated calculation including a certain amount of landau - level mixing and screening is required to confirm that the vortex - cdw phase is effectively the ground state in this system .
we leave this to further work . for reviews , see physics of the electron solid , edited by s.t .
chui ( international , boston , 1994 ) ; h. fertig and h. shayegan , in _ perspectives in quantum hall effects _ , edited by s. das sarma and a. pinczuk ( wiley , new york , 1997 ) , chaps . 3 and 9 , respectively . for a review ,
see m. m. fogler , in high magnetic fields : applications in condensed matter physics and spectroscopy , edited by c. berthier , l .- p .
levy , and g. martinez ( springer- verlag , berlin , 2002 ) , chap .
4 , pp . 98138 .
r. r. gerhardts , d. weiss , and k. v. klitzing , phys .
. lett . * 62 * , 1173 ( 1989 ) ; r. w. winkler , j. p .. kotthaus and k. ploog , phys .
lett . * 62 * , 1177 ( 1989 ) ; vidar gudmundsson and rolf r. gerhardts , phys . ref .
b * 52 * , 16744 ( 1995 ) .
d. j. thouless , in _ the quantum hall effect _ , edited by r. e. prange and s. m. girvin , graduate texts in contemporary physics ( springer - verlag , new york , 1987 ) , p. 101 ; m. c. geisler , j. h. smet , v. umansky , k. von klitzing , b. naundorf , r. ketzmerick , and h. schweizer , physica e * 25 * , 227 ( 2004 ) .
s. melinte , mona berciu , chenggang zhou , e. tutuc , s. j. papadakis , c. harrison , e. p.
de poortere , mingshaw wu , p.m .
chaikin , m .
shayegan , r.n .
bhatt , and r.a .
register , phys .
lett . * 92 * , 036802 ( 2004 ) . | the combined effect of a lateral square superlattice potential and the coulomb interaction on the ground state of a two - dimensional electron gas in a perpendicular magnetic field is studied for different rational values of @xmath0 the inverse of the number of flux quanta per unit cell of the external potential , at filling factor @xmath1 in landau level @xmath2 when landau level mixing and disorder effects are neglected , increasing the strength @xmath3 of the potential induces a transition , at a critical strength @xmath4 from a uniform and fully spin polarized state to a two - dimensional charge density wave ( cdw ) with a meronlike spin texture at each maximum and minimum of the cdw .
the collective excitations of this vortex - cdw are similar to those of the skyrme crystal that is expected to be the ground state _ near _ filling factor @xmath1 .
in particular , a broken u(1 ) symmetry in the vortex - cdw results in an extra gapless phase mode that could provide a fast channel for the relaxation of nuclear spins .
the average spin polarization @xmath5 changes in a continuous or discontinuous manner as @xmath3 is increased depending on whether @xmath6 $ ] or @xmath7 .$ ] the phase mode and the meronlike spin texture disappear at large value of @xmath8 leaving as the ground state partially spin - polarized cdw if @xmath9 or a spin - unpolarized cdw if @xmath10 |
ultrarelativistic electrons moving periodically in coherent radiation sources produce fairly monochromatic radiation .
the intensity of this radiation is proportional to the length of the periodic or confining structure , and if it is sufficiently high , the electron is likely to emit several photons during its passage .
the problem of multiple photon emission per electron plays significant role for crystalline radiators with thickness @xmath0 @xcite , compton sources with optical laser pulse energy densities @xmath1 @xcite , synchrotron radiation @xcite , undulators with the number of periods @xmath2 @xcite . at that
, multiple photon emissions from different electrons can be discriminated by means of coincidences between detectors , except in cases when the electron beam density is very high ( e.g. , in sase fels ) , and so only the problem of photon pileup from one electron remains .
physically , the multiple photon emission from an electron in a strong electromagnetic field may be viewed as a kind of electromagnetic showering .
it greatly simplifies if , as is typical for many coherent radiation sources , energies of emitted photons are much smaller than the electron energy : @xmath3 then , in the first approximation , it is permissible to count only photons stemming directly from the _ classical _ current generated by the initial electron , and neglect @xmath4 pairs produced by those photons downstream . for different experiments ,
though , the quantities of interest under conditions of multiple photon emission may differ .
for instance , in synchrotrons one is concerned with lag fluctuations of an electron accelerated by the running wave , in fels electron bunching , in compton sources electron energy straggling before its last interaction with a photon yielding the observed gamma - quantum . for coherent radiation in oriented crystals
, one does not really care about the spent electron , but instead , measurements of the radiation are usually carried out with calorimeters , and are restricted to spectra integrated over ( relativistically small ) emission angles . then , only the aggregate energy of all the nearly forward moving photons from one electron is measured . in the present article , we will discuss the latter case .
as was mentioned above , the possibility of photon pileup in the calorimeter causes significant deviation of the multiphoton probability ( i.e. , the radiative energy loss ) spectrum @xmath5 from the single - photon spectrum @xmath6 predicted by classical electrodynamics , in spite of condition ( [ omega - ll - ee ] ) .
nonetheless , a multiphoton spectrum can be deduced from the corresponding single - photon one , as is described below .
technically , the situation is the simplest when one s knowledge of initial conditions for the electron and the target is sufficient for completely specifying the electron s trajectory .
then , presuming that photons exert negligible back action on the electron , their probability distribution obeys poisson statistics @xcite @xmath7 where @xmath8 is the number of emitted photons , @xmath9 is the photon non - emission probability , and factor @xmath10 accounts for photon equivalence .
it is important to note , though , that the statistical independence of photon emission events does not imply that they occur sequentially in time .
for instance , in a coherent state of the quantized electromagnetic field ( a closest counterpart to a classical electromagnetic field ) @xcite , there is a quantum superposition rather than mere statistical mixture of photon number states , which is indicative of their simultaneous generation . to derive from ( [ poisson ] ) the radiation spectrum measured by a calorimeter
, one must integrate the fully differential spectrum for any definite number of emitted photons over their entire phase space , under the restriction of the sum of photon energies to a certain value @xmath11 , and then sum over photon multiplicities : @xmath12 series ( [ generic - sum - int ] ) can further be resummed to all orders , by laplace transforming it , whereafter it exponentiates , and then can be transformed back to the @xmath11 variable : @xmath13 the integration contour in ( [ generic - eq ] ) is flexible in the complex @xmath14 plane , but the integrand is not the only possible one , too : the upper limit of the integral in the exponent can be actually set equal to any value greater than @xmath11 , which can be expedient when dealing with incoherent radiation component @xcite . for single - photon spectra with a relatively sharp upper end @xmath15 , it is usually sufficient to let the upper integration limit equal to @xmath16 .
the practical benefit of the exponentiation is that it reduces the number of integrals to be handled . if one attempts instead to calculate termwise the probabilities of emission of a certain number of photons via eg .
( [ generic - sum - int ] ) , computation of the multiple integrals even numerically quickly becomes prohibitive .
sometimes , nonetheless , such multiple integrals are computed by monte - carlo see , e.g. , ref . 2 of @xcite . an approach alternative to ( [ generic - eq ] )
is to view @xmath5 as a solution of an integro - differential equation with the single - photon spectrum playing the role of its kernel .
specifically , presuming that @xmath17 is proportional to the target thickness @xmath18 , one can formally derive the standard transport equation @xmath19\ ] ] holding for the electron energy distribution function @xmath20 equation ( [ kinetic - eq1 ] ) might also serve as an alternative justification for the @xmath10 factor in ( [ poisson ] ) , by virtue of the identity @xmath21 should it not be the fact that each photon may actually be coherently generated along a large length , simultaneously with formation of other photons .
another popular approach for computing multiphoton emission probabilities during semi - classical electron passage is to divide the trajectory into several pieces , the interference of radiation from which is expected to be negligible , and define a probabilistic process of photon emission or non - emission on the junctions
. it also permits taking into account incoherent scattering during radiation , and other accompanying processes . that approach led to construction of monte carlo simulators widely used in modern practice @xcite . at the present state of the art ,
such generators usually employ kernels of generation of only one photon , which is subsequently iterated as well as in the resummation approach described above .
simplified reconstruction of the single - photon spectrum from multiphoton one , for radiation at volume reflection .
points , data of @xcite for a 2 mm crystal bent to radius 11 m. blue band , fit of @xmath22 to experimental data , the adjusted parameters being @xmath23 and @xmath24 .
red band , the reconstructed single - photon spectrum @xmath25 .
dashed green curve , prediction for the single - photon spectrum of coherent bremsstrahlung in a bent crystal @xcite.,width=321 ] adopting an analogy with high - energy physics , one can envisage further development of simulators to include non - factorized evaluation of two - photon emission along with the appropriate account of virtual radiative corrections , simulation of a shower including coherent photon emission from @xmath26 pairs , etc . but for qualitative analysis , the resummation method within the region of its validity is advantageous by its simplicity .
below , we will confine ourselves to discussing a few examples of its application .
before we proceed , it is worth pointing out that in case when there are no random parameters on which the spectrum significantly depends and which are to be averaged over , relation ( [ generic - eq ] ) can be _ inverted _ to reconstruct the single - photon spectrum from the observed multiphoton one @xcite . in fig .
[ fig : compar - lietti ] we show an example where a single - photon spectrum of radiation at volume reflection was reconstructed ( in a simplified fashion ) from the experimental data . without taking multiphoton effects into account ,
the agreement between the experiment and the theory leaves much to be desired @xcite , but after the reconstruction it becomes acceptable .
needless to say , the reconstruction procedure is expected to be efficient only if the radiation intensity ( photon multiplicity ) is not too high . in
what follows , we will discuss cases of high intensity and show that the corresponding spectra may have diverse shapes depending on the physical conditions .
for pure coherent radiation , typically characterized by an edgy single - photon spectrum , multiple photon emission imitates the appearance of higher spectral harmonics . with the increase of the radiation intensity ,
such quasi - harmonics proliferate and coalesce into universal forms depending on just few parameters . for a definite and smooth electron trajectory ,
the ( central ) limiting form of the multiphoton spectrum is a gaussian , as one would expect : and @xmath27 for the coherent radiation component @xmath28 were designated as @xmath29 and @xmath30 . here ,
for simplicity we will not be considering the incoherent contribution at all , therefore omitting subscript c. ] @xmath31.\ ] ] the corresponding regime may be physically categorized as a normal random walk . at large @xmath11 beyond the maximum ,
the spectrum decreases closer to an ordinary exponential rather than gaussian law @xcite . that can be interpreted in terms of a maximal - step walk ,
at which the probability is a product of probabilities for each maximal step , the number of steps being @xmath32 .
a correction to ( [ gauss ] ) breaking the scaling was obtained in @xcite .
it engages the third spectral moment related to skewedness . besides rendering the asymmetry to the spectrum
, it slightly redshifts its maximum .
another source of departure from the gaussianity , even in the absence of dependence of the underlying single - photon spectrum on external parameters , is the possible infinity of the spectral moments defining the gaussian .
for instance , in the case of coherent bremsstrahlung ( at a highly over - barrier electron motion ) , the dependence on electron impact parameters is negligible , but yet there is an incoherent radiation component described by bethe - heitler single - photon spectrum @xmath33 , for which the spectral moments diverge in the ultraviolet ( or more precisely , are dominated by the upper endpoint @xmath34 , around which our soft - photon resummation approach breaks down ) .
that leads to a modification of the limiting spectrum shape by a power - law tail persisting at high @xmath11 .
the tail , however , does not override the whole spectrum , whose maximum is still mainly determined by the coherent part of the spectrum and its convergent first moments .
therefore , the diffusion regime may be regarded as _ weakly _ anomalous , and the resulting distribution as intermediate between gaussian and lvy @xcite . within our framework ,
the anomaly must keep weak , i.e. its index @xmath35 must remain small , or else the shower will not be dominated by soft photons .
our formulation so far held for radiation from a completely prescribed electromagnetic current .
next , it is important to ask what happens in case of channeling radiation , when the radiation intensity , and therewith the photon multiplicity , strongly depends on the electron / positron oscillation amplitude in the channel , in turn depending on the electron impact parameter and the angle of entrance to the crystal . clearly ,
in the latter case the multiphoton spectrum must be duly averaged over the initial electron beam .
an important point here is that the averaging must be carried out _
after _ the resummation , and this order of operations is significant under the condition of high photon multiplicity . clearly , in eq .
( [ generic - eq ] ) , the average of an exponential does not reduce to an exponentiated average , @xmath36 similarly , in eq .
( [ kinetic - eq1 ] ) , @xmath37 invalidating the transport equation for the averaged distribution function with the averaged kernel .
in fact , such equations were heuristically used in the literature both for overbarrier motion and for channeling conditions . at that , sometimes they involved adjustable parameters to be determined from fits to observable spectra .
however , since the fraction of particles near the bottom of the channel radiates weakly , but contributes to the total number of passed electrons strongly , estimates even of the mean photon multiplicity per electron based on kinetic equation with an averaged kernel can be deceptive .
a related issue is that the knowledge of an _ averaged _ single - photon spectrum without detailing its dependence on the electron initial conditions is _ insufficient _ for deducing the ( averaged ) _ _ multi__photon spectrum .
indeed , the averaged single - photon spectrum @xmath38 depends on a single variable @xmath39 , whereas the averaged multiphoton spectrum nontrivially depends on two variables : @xmath11 and the target thickness @xmath18 , while the dependence on @xmath18 is intertwined with the dependence on the electron initial conditions .
it should be stressed yet that the dependence of the radiation on initial conditions must include both the incidence angle and the impact parameter . whereas the dependence on the former is measurable in principle ( cf .
, e.g. , @xcite ) , the dependence on the latter is not ( there are no relativistic beams of angstrom transverse size ) , and thus inevitably must be modeled theoretically .
let us now discuss typical high - intensity channeling radiation spectrum shapes , for different types of confining wells , and for simplicity neglecting dechanneling .
the averaged multiphoton spectrum is most easily tractable for channeling of positrons , when the continuous potential of aligned axes or planes may be regarded as approximately parabolic in the entire channel ( as is known to be the case for silicon crystals in orientation @xmath40 , @xmath41 , or @xmath42 ) .
then , the particle trajectories in the channels are given by harmonic curves , and the single - photon radiation spectrum is explicitly expressible , too .
it is proportional to the amplitude squared of the transverse motion , and therefore to the electron transverse energy @xmath43 , whereas the differential of the phase space over which the initial conditions are uniformly distributed can be written as @xmath44 , with @xmath45 the channeling well dimensionality : @xmath46 here value @xmath47 corresponds to capture to axial channeling from a beam with angular divergence @xmath48 , @xmath49 to capture to planar channeling from a beam with angular divergence @xmath48 , or to axial channeling from a beam co - aligned with the axis and having divergence @xmath50 , and @xmath51 corresponds to planar channeling from a perfectly aligned beam with divergence @xmath50 .
the analysis of the contour integral ( [ generic - eq ] ) averaged in this way is still rather involved , but at high intensity , one can view the spectrum just as a superposition of gaussians . inserting ( [ gauss ] ) into eq .
( [ averaging - positron - def ] ) then yields @xmath52 ^ 2}{2\xi\overline{\omega_{1}^2 } ( e_{\perp\max})}\right\}.\ ] ] the latter integral can be expressed in terms of error functions , but we will restrict ourselves to elucidating its qualitative behavior . at @xmath53 , with @xmath54 ,
the maximum of the integrand at @xmath55 falls well within the integration interval , thereby giving the dominant contribution to the integral .
neglecting contributions from the endpoints leads to the result @xmath56^{\mathcal{d}}.\ ] ] it is independent of the variance @xmath57 , and might be obtained by replacing partial gaussians ( including their pre - exponential factor ) in ( [ d - dep - aver - gauss ] ) by the infinitesimally narrow distribution @xmath58.\ ] ] that also demonstrates the existence of an approximate one - to - one correspondence between @xmath11 and the transverse energy of the positron quantified by the ratio @xmath59 .
( a ) high - intensity approximations ( [ d - dep - aver - gauss ] ) for multiphoton channeling radiation from positrons , for @xmath60 . dashed curve , planar channeling radiation from monokinetic beam , @xmath51 . solid curve , planar channeling radiation , @xmath49 .
dot - dashed curve , axial channeling radiation , @xmath47 .
( b ) high - intensity approximations for multiphoton channeling radiation from axially channeled electrons .
solid curve , eq .
( [ e - ax - gauss - integral ] ) , for @xmath61 .
dashed curve , the same with a lower cutoff @xmath62.,width=321 ] on the contrary , beyond the break , the integrand exponentially increases throughout the integration interval .
therefore , the main contribution to the integral comes from a vicinity of the endpoint @xmath63 , around which the @xmath64-dependence of the exponent can be linearized , whereas @xmath65 be put to unity , giving @xmath66 ^ 2}{2\overline{\omega_{1}^2 } ( e_{\perp\max})}}.\ ] ] examples of behavior of function ( [ d - dep - aver - gauss ] ) are shown in fig .
[ fig : aver - gaussians](a ) .
the spectra strongly depend on the value of @xmath45 , especially at low @xmath11 , in accord with asymptotics ( [ dwch - power - asympt ] ) . that illustrates the profound effect of the averaging on multiphoton channeling radiation spectra .
compared to positrons , channeling of electrons , generally , is affected by dechanneling stronger , because electrons are attracted to atomic nuclei , and thus may enter regions of strong scattering . for axial orientation of the crystal ,
though , electrons can orbit around positively charged atomic strings , and thus keep on distance from nuclear vicinities .
the approximate axial symmetry of a string alone , without introducing oversimplified assumptions about the @xmath67 dependence , does not yet permit explicitly expressing electron trajectories and therefrom the radiation spectra .
nonetheless , qualitative analysis is feasible for a potential modeled by a pure logarithmic function with a sharp cutoff at large radial distance @xmath67 : @xmath68 the motion of electrons in potential ( [ v - log ] ) is determined by two integrals of motion : transverse energy @xmath43 and projection @xmath69 of the angular momentum on the string direction .
estimates of first moments of the single - photon spectrum in a crystal of definite thickness give : @xmath70 and @xmath71 where @xmath72 is the period of the radial motion , exponentially depending on @xmath73 , and relatively weakly dependent on @xmath69 . both spectral moments blow up as @xmath74 or @xmath75 , i.e. , when the electron passes close to the singularity of the continuous potential .
the element of the uniformly populated phase space can be expressed in terms of @xmath72 , @xmath69 , as well : @xmath76 and the region accessible to channeling ( subject to condition @xmath77 ) , which exist for axial channeling .
the contribution from such particles is relatively small and will be analyzed elsewhere . ]
is @xmath78 where @xmath79 , @xmath80 .
ultimately , the averaged multiphoton spectrum assumes the form @xmath81
^ 2}{2\lambda\overline{\omega_{1}^2 } ( q_{\max})}\right\},\end{aligned}\ ] ] with @xmath82 .
again , we restrict ourselves to qualitative analysis of integral ( [ e - ax - gauss - integral ] ) . at @xmath83 , @xmath84 ,
the maximum of the integrand at @xmath85 is beyond the integration domain .
therefore , the main contribution comes from the endpoint @xmath86 , and generally , the integral is exponentially small : @xmath87\left[\overline{\omega_{1 } } ( q_{\max})+\omega\right]}\nonumber\\ \times\exp\left\{-\frac{\left[\omega-\overline{\omega_{1 } } ( q_{\max})\right]^2}{2\overline{\omega_{1}^2 } ( q_{\max})}\right\}.\qquad\quad\end{aligned}\ ] ] on the contrary , at @xmath88 , the maximum of the integrand falls well within the integration domain . neglecting the endpoint contributions , or simply inserting @xmath89\ ] ] in place of the partial gaussians ,
the asymptotics of the multiphoton spectrum derives as @xmath90 the corresponding mean energy loss @xmath91 diverges in the ultraviolet as a double logarithm , in accordance with the double logarithmic divergence of the single - photon average mean energy loss , cf .
( [ mean - omega - ax ] ) .
the delta function ( [ del ] ) implies that a correspondence between @xmath11 and the integrals of motion holds for electron axial channeling radiation , as well .
but in contrast to radiation from channeled positrons , the correspondence here takes place at the outer slope of the spectrum [ see fig . [
fig : aver - gaussians](b ) ] . yet , since there are two integrals of motion , the correspondence of @xmath11 with them is no longer one - to - one , unless one refers to the special integral of motion @xmath92 .
experimentally measured multiphoton axial channeling radiation spectra @xcite decrease considerably faster than by power law ( [ omega-2log ] ) .
that must be attributed to the crudeness of our model ( [ v - log ] ) and the physical finiteness of @xmath93 due to the smearing of the string potential at low @xmath67 through thermal and quantum vibrations .
the simplest way to take all that into account is to introduce a lower cutoff @xmath94 into integrals ( [ e - ax - gauss - integral ] ) , which leads to an exponentially decreasing spectrum [ see fig . [
fig : aver - gaussians](b ) , dashed curve ] .
the beginning of the exponential suppression region depends on the value of the cutoff parameter and therethrough on the crystal temperature .
the intrinsic simplicity of the soft photon resummation procedure makes it a useful tool in various physical situations , predicting generation of higher quasi - harmonics at moderate intensity ( as discussed in @xcite ) and the existence of a set of limiting distributions at high intensity , depending on the shape of the confining well , yet yielding in certain spectral regions approximate correspondence between the radiative loss @xmath11 and one of the integrals of motion of the channeled particle . at the same time
, the resummation method has its limitations and is unable to cover , e.g. , the case when a considerable part of the spectrum concentrates near the kinematic edge @xmath95 @xcite .
also , even in the domain of applicability of the soft photon approximation ( [ omega - ll - ee ] ) , one may wish to evaluate higher order corrections in @xmath96 , e.g. , in order to account for degradation of @xmath97 stipulating additional broadening and redshift of the multiphoton spectra @xcite , or to describe the effect of electron beam bunching in fels @xcite . for the specific problem of channeling radiation , there arise additional issues .
first is the need for knowledge of not only the averaged single - photon spectrum , which might be measured at a sufficiently small crystal thickness , but also of the dependence of the radiation spectrum on electron entrance parameters .
that can be illustrated , e.g. , by the variance of the multiphoton spectrum about the mean value , for positron channeling in a harmonic well discussed in sec .
[ subsec : positron - ch ] : @xmath98 due to the second term , it depends on the target thickness _
nonlinearly _ , yet involving a dependence on the value of @xmath45 , which in turn depends on the shape of the well . also , the existence of a correspondence in some spectral regions between @xmath11 and the particle integrals of motion implies that not all the values of integrals of motion contribute at a given @xmath11 , in contrast to the case of the averaged single - photon spectrum .
further complications emerge because in the presence of dechanneling , relative contributions of different integrals of motion can attenuate differently .
the corresponding factors thus must be introduced into integrals ( [ d - dep - aver - gauss ] ) , ( [ e - ax - gauss - integral ] ) .
thirdly , the dechanneled particles contribute to the radiation spectrum , as well , although in a somewhat harder spectral region . that demands inclusion of their contribution to the multiphoton spectrum , too . on the other hand , studying multiphoton radiation spectra under such conditions may offer deeper insight into the complex dynamics of channeling . | recent results in the theory of multiphoton spectra for coherent radiation sources are overviewed , with the emphasis on channeling radiation . for the latter case ,
the importance of the order of resummation and averaging is illustrated .
limiting shapes of multiphoton spectra at high intensity are discussed for different channeling regimes . in some spectral regions , there emerges an approximate correspondence between the radiative energy loss and the electron integrals of motion . intense coherent radiation , weakly anomalous diffusion , averaged gaussian distributions 61.85.+p , 31.30.j- , 52.59.-f |
let @xmath0 \rightarrow { \mathbb{r}}$ ] be a continuous function and @xmath1 \rightarrow { \mathbb{r}}$ ] a function of bounded variation .
it is a classical result that for such @xmath2 and @xmath3 , and for any @xmath4 $ ] , the riemann stieltjes integral @xmath5 exists ( see , e.g. , @xcite ) .
while the basic properties of riemann stieltjes integrals ( and related lebesgue stieltjes integrals ) are covered in classical textbooks on real analysis and integration @xcite , the following simple question is not addressed in them .
[ question ] if @xmath2 is positive , @xmath3 is non - negative , non - vanishing , and satisfies @xmath6 , can we then select the upper limit of integration @xmath7 so that the integral is positive ?
the answer to the question is obviously yes for riemann integrals ( @xmath8 ) , which could help to explain why it has been overlooked so far .
in fact , after an extensive search of the literature , we believe that question [ question ] has not been answered before in the full generality .
( we shall comment later , in remark [ finalcomment ] , why the present generality may be relevant in applications . )
the only reference known to us is a note by satyanarayana @xcite where an affirmative answer is proven under a slightly different set of assumptions , most importantly , assuming that @xmath3 is _ non - decreasing_. let us stress that _ positive _ will here always mean _ strictly positive_. in particular , since @xmath2 is continuous , under our assumptions @xmath9 .
recall that , since @xmath3 is of bounded variation , there exist non - decreasing functions @xmath10 and @xmath11 such that @xmath12 . therefore , finite @xmath13 and @xmath14 exist for any @xmath15 $ ] ( apart from @xmath16 and @xmath17 , obviously ) .
moreover , the integral is equal to @xmath18 since @xmath6 , we may assume that @xmath10 and @xmath11 are non - negative and satisfy @xmath19 .
we may now distinguish two special cases , where it is evident that the answer to question [ question ] is yes . 1 .
if @xmath20 , where @xmath21 , then there exists @xmath22 such that is positive for @xmath23 .
2 . if @xmath24 and @xmath25 , then is at least @xmath26}f(x)>0 $ ] .
the first item follows from the elementary lower bound , valid for @xmath27 , @xmath28 whereas the second item is a straightforward consequence of @xmath19 .
in general , the integral is positive if and only if @xmath29 obviously , the condition @xmath30 is equivalent to @xmath31 , and we have @xmath32 if and only if @xmath33
. it would be tempting to conjecture that , as a continuous function , @xmath2 is nearly constant " in some neighborhood of @xmath34 and , hence , that the inequality ought to hold for @xmath23 with some small " @xmath22 , suggesting an affirmative answer to question [ question ] in general .
the proviso @xmath6 may seem superfluous as the value of the integral does not depend on @xmath35 .
however , together with the condition @xmath36 it constrains the behavior of @xmath3 near @xmath34 , which is a key part of the formulation of question [ question ] .
it should also be stressed that the continuity of @xmath2 is equally important
aside from the possible non - existence of the integral , there is no reason to expect the answer to question [ question ] to be yes when @xmath2 fails to be continuous .
unfortunately , the heuristic above is too simple - minded since mere continuity does not restrict the _ fine properties _ of the integrand @xmath2 and leaves it with enough room to vary too much " for our purposes .
indeed , the general answer to question [ question ] is no .
we show that for any @xmath2 that exhibits enough " variation , there exists a suitable @xmath3 such that the integral is less than zero for all @xmath4 $ ] .
[ negative ] let @xmath37 \rightarrow ( 0,\infty)$ ] be a continuous function .
suppose that there exist two sequences @xmath38 and @xmath39 with @xmath40 such that for some @xmath41 and @xmath42 , @xmath43 then , there exists a function @xmath1 \rightarrow [ 0,\infty)$ ] of bounded variation such that @xmath6 and @xmath44$}\ , .\ ] ] let @xmath45 and consider @xmath46 \rightarrow [ 0,\infty)$ ] defined by @xmath47 where @xmath48 denotes the characteristic function of a set @xmath49 .
( figure [ fig : unbf ] illustrates the definition for example [ th : negexample ] below . )
this is , by construction , a function of bounded variation such that @xmath50 .
we have for any @xmath51 , @xmath52 where @xmath53 since @xmath54 and @xmath55 , there exists @xmath56 such that @xmath57 we may also note that for all @xmath58 , with @xmath59 , and @xmath60 , @xmath61 thus , defining @xmath62 yields a function with all the properties stated in the theorem . ,
as defined in example [ th : negexample ] for @xmath63 , and the corresponding functions @xmath3 and @xmath64 , @xmath65 , as defined in the proof of theorem [ negative ] using @xmath66 and @xmath67 .
the values have been computed numerically by considering only contributions with @xmath68 . for clarity ,
only the region with @xmath69 $ ] is shown .
[ fig : unbf],title="fig : " ] + , as defined in example [ th : negexample ] for @xmath63 , and the corresponding functions @xmath3 and @xmath64 , @xmath65 , as defined in the proof of theorem [ negative ] using @xmath66 and @xmath67 .
the values have been computed numerically by considering only contributions with @xmath68 . for clarity ,
only the region with @xmath69 $ ] is shown .
[ fig : unbf],title="fig : " ] , as defined in example [ th : negexample ] for @xmath63 , and the corresponding functions @xmath3 and @xmath64 , @xmath65 , as defined in the proof of theorem [ negative ] using @xmath66 and @xmath67 .
the values have been computed numerically by considering only contributions with @xmath68 . for clarity ,
only the region with @xmath69 $ ] is shown .
[ fig : unbf],title="fig : " ] [ th : negexample ] function @xmath70 \rightarrow ( 0,\infty)$ ] given by @xmath71 where @xmath42 , satisfies the condition of theorem [ negative ] with @xmath72 , @xmath73 , @xmath74 , using @xmath75 .
figure [ fig : unbf ] illustrates the behavior of @xmath2 and of the corresponding @xmath3 and riemann stieltjes integral , as defined in the proof of theorem [ negative ] .
any integrand @xmath2 that satisfies the condition of theorem [ negative ] is clearly of _ unbounded variation_. this prompts us to ask , could we actually obtain an affirmative answer to question [ question ] if @xmath2 varied less . " in fact , we are able to show that , if @xmath2 is of _ bounded variation _ , the answer to question [ question ] is yes
. bounds for riemann
stieltjes integrals under these assumptions have been derived by beesack @xcite , ganelius @xcite , and knowles @xcite , but instead of building our argument on them , we give a direct proof which relies on some elementary measure theory and the measure - theoretic version of _ grnwall s inequality_. [ bvcase ] let @xmath0 \rightarrow ( 0,\infty)$ ] be a continuous function and @xmath1 \rightarrow [ 0,\infty)$ ] a non - vanishing function of bounded variation such that @xmath6 .
if @xmath2 is of bounded variation , then @xmath76$.}\ ] ] let us denote by @xmath77 the total variation of @xmath2 on @xmath78\subset [ a , b]$ ] .
recall that total variation is additive in the sense that @xmath79 for @xmath80 ( * ? ? ? * theorem 12.1 ) .
moreover , since @xmath2 is continuous , the mapping @xmath81 from @xmath82 $ ] to @xmath83 is continuous ( * ? ? ?
* theorem 12.2 ) .
thus , there exists a finite , positive borel measure @xmath84 on @xmath82 $ ] such that @xmath85 for any @xmath86 and @xmath87 such that @xmath88 .
let us define another finite , positive borel measure by @xmath89 . by construction
, we have then @xmath90 by an approximation with suitable riemann
stieltjes sums , where the values of @xmath3 are chosen to be sufficiently close to their respective infima on all subintervals of the partitions , we can prove that @xmath91 the riemann stieltjes integral on the left hand side is well - defined by the integration by parts formula ( * ? ? ?
* theorem 12.14 ) @xmath92 whenever @xmath93 exists and this is always true under the present assumptions .
now suppose that , contrary to our assertion , we have @xmath94$.}\ ] ] rearranging the integration by parts formula and using the assumption @xmath6 , , and , we obtain for any @xmath4 $ ] @xmath95 but the measure - theoretic version of grnwall s inequality ( * ? ? ?
* theorem a.5.1 ) implies that then @xmath96 for all @xmath97 $ ] , whence @xmath98 , a contradiction .
there are two straightforward refinements to theorem [ bvcase ] .
firstly , it clearly suffices that @xmath2 is of bounded variation on @xmath99 $ ] for some @xmath22 . secondly ,
if @xmath3 is right - continuous , then also the mapping @xmath100 is right - continuous and , under the assumptions of theorem [ bvcase ] , there exists an interval @xmath78 \subset [ a , b]$ ] such that @xmath101$.}\ ] ] [ finalcomment ] theorem [ bvcase ] explains why the answer to question [ question ] is perhaps elusive .
experimentation with nicely behaving integrands will not suffice , since a pathological " @xmath2 is required in order to discover the general answer . however , such integrands need not be mere curiosities .
in fact , our study of question [ question ] was originally motivated by an application in financial mathematics involving as the integrand a _ path _ of a continuous - time _ stochastic process _ , which is typically of unbounded variation .
this version corrects a mistake in the formulation of theorem [ negative ] , pointed out to us by gerald teschl .
j. lukkarinen was supported by the academy of finland .
m. s. pakkanen acknowledges support from the finnish cultural foundation , from creates , funded by the danish national research foundation , and from the aarhus university research foundation regarding the project
stochastic and econometric analysis of commodity markets " . | we study the question , whether a riemann stieltjes integral of a positive continuous function with respect to a non - negative function of bounded variation is positive . |
it is well established that the nuclei of active galaxies exhibit a broad range of luminosities , from the most energetic quasars to more modest seyferts .
active galactic nuclei ( agn ) with nuclear x - ray luminosities below @xmath710@xmath9 erg s@xmath6 are classified as low luminosity agns ( llagns ; koratkar et al .
1995 ) , and are the most common variety of agn observed in the local universe .
about 30% of all nearby bright galaxies exhibit llagn activity , and many additionally host low - ionization nuclear emission - line regions ( liners ; ho et al .
many recent studies have successfully linked liners to agn activity ; for example , the detections of x - ray cores ( dudik et al . 2005 ; flohic et al . 2006 ; gonzlez - martn et al .
2006 ; zhang et al . 2009 ; gonzlez - martn et al .
2009 ) , radio cores ( nagar et al . 2005 ) , and mid - ir coronal lines ( satyapal et al .
2004 ) are all indicative of an accreting black hole ( bh ) energy source . at low nuclear luminosities ,
however , several alternative explanations for the power source have been investigated , such as circumnuclear starbursts ( gonzales delgado et al .
2004 ; colina et al .
2002 ) , shock heating by supernovae ( sne ) in a high - density environment ( alonso - herrero et al . 2000 ; olsson et al . 2007 ) , and photoionization by very hot o stars ( terashima et al .
moreover , a number of studies of the seds of weak agns in liners find that the agn does not produce enough photons to power the emission lines ( e.g. , ho 2008 ; eracleous et al .
2010 , and references therein ) . at a distance of 3.1 mpc ( karachentsev et al .
2004 ) , ngc 404 is the nearest s0 type galaxy to the milky way and the closest galactic nucleus to be classified as a liner by ho
et al . ( 1997 ) . while the observed nuclear x - ray luminosity is low , only a few times 10@xmath10 erg s@xmath6 ( eracleous et al .
2002 ) , the presence of a llagn in ngc 404 remains ambiguous .
the x - ray luminosity is consistent with that of a single high mass x - ray binary ( hmxb ) or a giant star - forming region such as 30 doradus ( wang & helfand 1991 ) , and the soft x - ray emission is consistent with a hot gas origin , potentially blown out by a compact starburst or sne .
no radio core has been observed at 15 ghz ( to a limiting flux of 1.4 mjy ; nagar et al .
2005 ) , however an unnresolved 3 mjy continuum source is detected at 1.4 ghz by del ro et al .
( 2004 ) , comparable in luminosity to the crab nebula .
a compact x - ray source was previously detected in the central region of ngc 404 ( lira et al . 2000 ; eracleous et al . 2002 ) , but its low luminosity and soft thermal spectrum indicate a possible starburst event origin .
mid - ir observations of the ngc 404 nuclear region show high ionization lines consistent with agns ( satyapal et al .
2004 ) ; however , the [ ne v ] lines ( a more reliable indicator of agn activity ) are not detected ( abel & satyapal 2008 ) . additionally , optical _ hubble space telescope _ ( hst ) observations show h@xmath2 emission occurring in both a compact source @xmath1116 north of the nucleus and in structures reminiscient of supernova remnants ( pogge et al . 2000 ) , and [ o iii ] emission originates from a double - lobed structure along the major axis of the galaxy ( plana et al .
1998 ) with a higher velocity dispersion than the central h@xmath2 emission ( bouchard et al . 2010 ) .
while the uv spectrum of the nucleus reveals signatures of o stars , the dilution of the lines suggest that @xmath760% of the uv flux may originate from a nonthermal source ( maoz et al .
1998 ) . the observed level of uv variability ( the uv emission declined by a factor of 3 between 1993 and 2002 ; maoz et al .
2005 ) provides the strongest evidence for the existence of an accreting , massive black hole in the ngc 404 nucleus .
analysis of nicmos data by ravindranath et al .
( 2001 ) reveal a nuclear star cluster ( nsc ) within the central arcsecond of ngc 404 .
nscs are present in @xmath770% of galaxies ( graham & guzman 2003 ; muoz marn et al .
2009 ) , independent of the host galaxy morphology ( bker et al . 2002 ; carollo et al . 1997 ; ct et al .
2006 ) , and are considered to be the foundation of circumnuclear starbursts and supernovae that could drive liner activity ( meurer et al . 1995 ; tremonti et al .
2001 ; chandar et al . 2005 ) . however , if ngc 404 is dominated by star formation , the rate is exceptionally low , with only two to six o stars being sufficient to explain the observed luminosity ( del ro et al . 2004 ) .
dynamical modeling of stellar and gas kinematics in the nucleus by seth et al .
( 2010 ) provide mixed evidence for the presence of a smbh .
they derive a firm upper limit of @xmath12 @xmath1 , and a best fitting gas dynamical mass of @xmath13 @xmath1(3@xmath14 errors ) .
although other low - mass galaxies have been identified as candidate imbh hosts through reverberation mapping ( i.e. , ngc 4395 ; peterson et al . 2005 ) and indirect mass measurements ( i.e. , by narrow optical line measurements ; see greene & ho 2007 ) , ngc 404 is potentially the lowest - mass central bh ever dynamically detected in the center of a galaxy . with both a nsc and possible imbh ,
the nuclear region of ngc 404 is a complicated environment .
however , ngc 404 provides an ideal test case to address several key questions relating to llagn activity : is there an intrinsic lower limit to the luminosity of the agn phenomenon , and what fraction of liners are powered by stellar processes n versus those that host a dwarf version of more powerful seyferts and quasars ?
deep , high spatial resolution x - ray observations can potentially resolve many ambiguities surrounding the ngc 404 nucleus , such as the morphology of the x - ray emission , the shape of the x - ray spectrum , and the variability properties of the source . in this paper
, we present the analysis of a new , 100 ks _ chandra _
observation of the ngc 404 nuclear region . in 2
, we provide a description of the observations and our data analysis procedures , including our imaging analysis , timing analysis , and spectral modeling .
our results are presented in 3 , and we give a discussion of our results in 4 .
a summary of our work is given in 5 .
ngc 404 was observed with _ chandra _
acis - s on 2010 october 21 - 22 for a total useable time of 97 ks . the optical center of the ngc 404 nucleus centered on the s3 chip and detected at @xmath15 = 01@xmath1609@xmath1726.@xmath1899 and @xmath19 = + 35@xmath2043@xmath2105.@xmath81 ( with a positional uncertainty of @xmath70.@xmath802 ) in good agreement ( within 0.@xmath84 ) with the position reported in simbad ( cotton et al . 1999 ) and previous _ chandra _
detections of the ngc 404 nucleus .
the background count rate was estimated using an annular region devoid of any obvious sources , centered at the nucleus with an inner and outer radius of 25@xmath8and 35@xmath8 , respectively .
we find the 0.3 - 8 kev background count rate to be low , @xmath75@xmath22 ct s@xmath6 arcsec@xmath23 , throughout the observation .
we additionally obtained two archival _ chandra _ acis - s observations of ngc 404 from 2000 aug 30 ( 2 ks ) and 1999 dec 19 - 20 ( 24 ks ) .
these observations are summarized in table 1 .
all observations were reduced using the x - ray data analysis package ciao version 4.3 and using standard reduction procedures .
we created exposure maps for the images using the ciao script ` merge_all ` .
point sources , including the ngc 404 nucleus , were identified using the ciao task ` wavdetect ` .
we compared our x - ray point source detections to a 3.6@xmath24 _ spitzer _ irac image of ngc 404 ( aligned with the usno - b1.0 star catalog ) and find one x - ray source other than the ngc 404 nucleus to be coincident with a likely ir counterpart ( with a positional offset of 0.@xmath81 ) , although this source is not a confirmed x - ray emitting object .
we additionally searched for potential optical counterparts using optical observations in the hubble legacy archive , but were unable to unambiguously identify any optical sources within the hst field of view coincident with our x - ray point sources .
we thus conclude that our residual systematic uncertainty in absolute pointing is conservatively 0.@xmath84 .
all spectra were extracted using ` psextract ` , and spectral fitting was performed in ` xspec ` ( arnaud 1996 ) v.12.6.0q , and spectral models were fit to the unbinned spectra ( binned spectra are shown for display purposes only ) .
all models include a column of neutral absorption fixed at the galactic value , @xmath25 = @xmath26 cm@xmath23 ( kalberla et al . 2005 ) , estimated using the heasarc @xmath28 calculator .
we use @xmath29-statistics in lieu of traditional @xmath30 statistics , due to the low number of x - ray counts .
errors correspond to the 90% confidence level .
in addition to reporting the @xmath29-statistic and degrees of freedom ( @xmath31 ) for each model , we used the ` xspec ` task _ goodness _ to perform monte carlo simulations of the spectra using each best - fit model .
the procedure returns the percentage of the simulated spectra that had a fit statistic less than that obtained from the fit to the real data .
a value of 50% indicates the best - fitting model is a good representation of the data ; values much less than this indicate that the data are overparameterized by the model ( i.e. , random statistical fluctuations in the majority of simulated spectra are not able to produce a fit statistic as low as that obtained from the real data ) , and values much higher than this indicate the model is a poor fit to the data ( i.e. , a large majority of simulated spectra have a fit statistic less than that obtained from the real data ) .
we perform @xmath32 realizations for each model , and we denote the resulting percentage as @xmath33 .
we co - added the available _ chandra _ observations to investigate the morphology of the ngc 404 x - ray emission . earlier work by eracleous et al .
( 2002 ) saw evidence for extended soft x - ray emission as far out as 10@xmath8(@xmath70.15 kpc ) from the nucleus , with potential shell - like structures suggestive of a hot gas superbubble ( see the discussion by chu & mac low 1990 ) .
we divided our _ chandra _ image into three energy bands : soft ( 0.3 - 1 kev ) , medium ( 1 - 2 kev ) , and hard ( 2 - 7 kev ) .
each image was adaptively smoothed using ` csmooth ` .
figure 1 shows a smoothed rgb rendering of the ngc 404 nucleus .
the image shows a hard core , coincident with the optical and radio center of the galaxy , and soft surrounding emission .
extended x - ray emission is detected out to @xmath715@xmath8(@xmath70.2 kpc ) .
we additionally see a soft source not detected in optical images , @xmath710@xmath8northeast of the nucleus . for comparison
, the nsc extends out to 0.@xmath87 ( @xmath710 pc ; seth et al . 2010 ) , and the disk scale length of ngc 404 is @xmath7130@xmath8(@xmath72 kpc ; baggett et al .
the adaptively smoothed 0.3 - 1 kev and 2 - 7 kev images , with contours superimposed , are shown in figure 2 to emphasize the compact nature of the hard central source .
we next used the smoothed soft and hard energy images to construct a hardness ratio map of the ngc 404 nuclear region ( figure 3 ) .
we define a hardness ratio hr = ( hard - soft)/(hard+soft ) ; dark areas in the hardness ratio map correspond to softer x - ray emission , while light regions indicate hard emission .
there is clearly a point - like region of hard emission , coincident with the location of the nsc , surrounded by an extended area of predominantly soft x - ray emission . in figure 4 , we show a h@xmath2-@xmath34 color map of the ngc 404 nuclear region ( seth et al .
2010 ) , with contours from our hr map overlaid .
dark regions on the h@xmath2-@xmath34 color map correspond to regions with high h@xmath2 flux , likely associated with young stars .
we find one such region is coincident with soft , extended x - ray emission seen in our rgb rendering .
we find no evidence for long - term ( @xmath710 year ) variability ; our best - fit 0.5 - 2 kev luminosity ( @xmath710@xmath10 erg s@xmath6 , see next section ) agree with that found by eracleous et al .
we generated light curves in four energy ranges ( the total 0.3 - 10 kev band , and the soft , medium , and hard bands described above ) to search for short - term variability over the course of our 97 ks observation .
we generated cumulative arrival time distributions of counts for each light curve , then ran a two - sided ks test against the expected cumulative arrival time distrbution for a constant count rate .
figure 5 shows the cumulative arrival time distributions for our total 0.3 - 8 kev light curve and the hard 2 - 7 kev light curve .
we ran a two - sided ks test for each of our cumulative arrival time distributions against a constant count rate .
as summarized in table 2 , we found no evidence from our ks tests for variability at energies softer than 2 kev ; however , our hard light curve yields a ks chance probability of 5.8@xmath35 . both agns and xrbs exhibit strong and rapid variability ;
the apparent detection of hard x - ray variability on time scales of @xmath71 day therefore lends support to the idea that the ngc 404 nucleus hosts an accreting object .
detailed spectral fitting ( described in the next section ) is needed to constrain the nature of the compact accreting object further .
hard x - ray spectra of llagns in liners are typically well represented by a two - component model : a power - law component plus soft thermal emission ( terashima et al . 2000 ) , and the h@xmath2 luminosoties of liners are positively correlated with the x - ray luminosities in the 2 - 10 kev band ( ho 2001 ) . to test whether the ngc 404 nucleus is consistent with being a llagn
, we use our observation with archival _ chandra _ acis - s observations to perform spectral fitting in the 0.3 - 8 kev energy band .
the total nuclear region ( @xmath3617@xmath8 ) of ngc 404 contained @xmath71200 counts .
our imaging analysis of the ngc 404 nuclear region indicated the presence of two distinct regions within the nucleus : a hard point source coincident with the optical and radio center of the galaxy , and a soft , extended region likely associated with an area of recent star formation . using our hr contours and h@xmath2-@xmath34 color map , we extracted spectra using elliptical regions for each of the two distinct sources , with an elliptical area of 11 square arcseconds for the hard point source and 27 square arcseconds for the soft extended source .
the hard nuclear point source contained @xmath7500 counts , and our soft , extended region contained @xmath790 counts .
all our models fix an absorption component due to the galactic column and an intrinsic absorption inferred from optical extinction from schlegel et al .
we find no evidence in our spectral fitting for additional absorption , and we find no evidence for the presence of a reflection component in any of the 0.3 - 8 kev spectra .
we use the mass - metallicity relation derived in tremonti et al .
( 2004 ) to estimate the ngc 404 metallicity to be 12 + log(o / h ) @xmath78.6 - 9 ( i.e. , near solar abundances ) .
we first attempted to model the soft , diffuse emission as a power law , a disk blackbody , and a thermal plasma ( ` apec ` ; smith et al .
we find the thermal plasma model provides the best fit , with @xmath31= 5.5/4 and @xmath33 = 51% for @xmath4 = 0.67@xmath37 kev and abundances fixed at their solar values .
the 0.3 - 10 kev luminosity of the soft x - ray emission is ( 2.5@xmath38 erg s@xmath6 , with an estimated 2 - 10 kev luminosity of 8.8@xmath3910@xmath40 erg s@xmath6 .
our data are consistent with the idea of gas being ejected from the central region in a superbubble .
we attempted to model the hard , nuclear point - source as a simple power law , but were unable to obtain an acceptable fit ( @xmath31= 33/25 and @xmath33 = 76% ) .
single - temperature thermal plasma models , with abundances either fixed at their solar values or allowed to vary , additionally did not result in acceptable fits ( @xmath31= 68/25 for @xmath41fixed at solar , and @xmath31= 42/24 with @[email protected] ) .
two - temperature thermal plasma models , with abundances fixed at solar values or allowed to vary , severely over - parameterized the data .
we find @xmath33 = 0.4% with @xmath31= 11/23 for our model with @xmath41= 1 , and @xmath33 = 0.3% with @xmath31= 11/22 for @[email protected] .
we next attempted to model the hard , nuclear point source as a power law contaminated by thermal emission from the soft , diffuse x - ray source .
for each fit , we fixed the thermal plasma temperature to be within the 90% confidence interval of our best - fit model to the soft , diffuse source .
abundances were kept fixed at solar values .
we find the hard , nuclear point source to be best described by @xmath4 = 0.78 kev , with a power law photon index @xmath3= 1.88@xmath42 contributing to @xmath772% of the total 0.3 - 8 kev photon flux and @xmath31= 21/24 .
we find a 0.3 - 10 kev luminosity of 2.4@xmath43 erg s@xmath6 and a 2 - 10 kev luminosity of 1.2@xmath43 erg s@xmath6 .
although the fit moderately overparameterizes the data ( @xmath33=21% ) , this model produces our best fitting parameters and is consistent with our imaging analysis .
finally , we applied the results of our spectral fitting to the soft , diffuse emission and the hard , nuclear point source to model the entire nuclear region of ngc 404 as a power law contaminated by thermal plasma emission .
we assume the best - fit thermal plasma temperature ( 0.67 kev ) from the soft , diffuse emission and the best - fit power law photon index ( @xmath3= 1.88 ) from the hard nuclear point source .
we find the power law component contributes @xmath755% of the 0.3 - 8 kev photon flux , with @xmath31= 124/126 , with @xmath4430% . to estimate the errors on the power law photon index , we find the best - fit value of @xmath3when the thermal plasma temperature is set at our 90% confidence interval lower limit ( 0.56 kev ) and upper limit ( 0.78 kev ) .
we find a best - fit photon index @xmath3= 1.85@xmath45 using this approach .
we estimate the 0.3 - 10 kev luminosity to be 3.0@xmath46 erg s@xmath6 and a 2 - 10 kev luminosity of 1.3@xmath5 erg s@xmath6 .
the results of our spectral fitting are summarized in table 3 .
figure 6 shows the 0.3 - 8 kev spectra for the soft , diffuse emission , the hard nuclear point source , and the total nuclear region ( with our best - fit models superimposed ) .
the presence of a power law component in the ngc 404 nucleus , in addition to variability in the 2 - 7 kev emission , lends support to the idea that the ngc 404 nucleus hosts an accreting black hole , but its low luminosity , on the order of a few times 10@xmath10 erg s@xmath6 , is comparable to that of a single xrb .
by modeling the hard nuclear point source and diffuse soft emission found in the ngc 404 nucleus separately , we are able to resolve the ambiguity of the x - ray emission : a hard point source provides a power law component , and extended , diffuse gas supplies the thermal plasma emission .
the eddington luminosity is defined as @xmath47 = 1.3@xmath48 ( @xmath49/@xmath1 ) erg s@xmath6 , and the eddington ratio @xmath50 is commonly defined as @xmath50 = log@xmath51(@xmath52 ) , where the bolometric luminosity @xmath53 is typically estimated as @xmath54 = 16 for agns ( ho 2008 ) , whereas the bolometric correction factor for xrbs is roughly 2 - 5 times lower than for agn ( wu & gu 2008 ; hereafter , wg08 ) .
the relationship between the x - ray power law photon index @xmath3and eddington ratio has been investigated for xrbs ( wg08 ) and low luminosity agns ( llagns ; constantin et al .
2009 ; hereafter c+09 ) .
an anticorrelation is found for llagns , whereas a positive correlation is observed for xrbs and luminous agns ( wang et al .
2004 ; shemmer et al .
2006 ) .
we estimate @xmath50 for ngc 404 assuming the black hole is an xrb , with @xmath5510 @xmath1 , and an imbh agn , with @xmath56 @xmath1 . in figure 7
, we use our best - fit photon index for the hard nuclear point source and estimates of @xmath50 to compare our ngc 404 data to the observed relationships for both xrbs and llagns from wg08 and c+09 , respectively .
the errors in @xmath50 indicate a factor of 3 change in bh mass ( i.e. , an agn ranging from 3.3@xmath57 @xmath1to 3@xmath58 @xmath1and an xrb ranging from 3.3 @xmath1to 30 @xmath1 ) .
we find that while our data fall well below the wg08 relationship for a high / soft state xrb , our data are consistent with the anticorrelation found for the c+09 llagn sample .
however , due to the large scatter observed in the llagn sample and the errors in our observed photon index , we can not decisively rule out the possibility that the ngc 404 central engine is powered by an xrb in the low / hard state .
our data moderately favor the imbh agn interpretation .
although the current work moderately favors the low - mass agn interpretation , the uv spectrum and nuclear star cluster still allow the possibility of an xrb component . in table 4
, we summarize the observed multiwavelength properties of ngc 404 , found in both the literature and presented in this work , and indicate if the origin is likely to be an agn or xrb . additionally , the upper limits on the radio core ( nagar et al .
2005 ) and the detection of an unresolved radio continuum ( del ro et al .
2004 ) can be combined with our deep x - ray observations to place upper limits on the central bh mass of ngc 404 a correlation has been established relating the radio luminosity @xmath59 and x - ray luminosity @xmath60 ( corbel et al .
2003 ; gallo , fender & pooley 2003 ; gallo et al .
2006 ) cover many orders of magnitude in bh mass and luminosity , forming the `` fundamental plane of black hole activity . ''
we use the best - fit bh `` fundamental plane '' recently presented by bell et al .
( 2011 ) and the upper limit on the radio core flux of ngc 404 to estimate @xmath61 @xmath1 .
if the unresolved 3 mjy continuum source is indeed powered by an accreting bh , it would imply @xmath62 @xmath1 .
additionally , we use the following relation between radio luminosity and star formation rate ( condon 1992 ) , @xmath63,\ ] ] where @xmath64 is the nonthermal spectral index and @xmath65 is the star formation rate ( in @xmath1yr@xmath6 ) , to test whether the observed radio luminosity is consistent with the observed low star formation rate . using the @xmath65 upper limit ( @xmath66 @xmath1yr@xmath6 )
estimated in seth et al .
( 2010 ) , we predict an upper limit on the radio luminosity at 1.4 ghz to be @xmath74@xmath67 erg s@xmath6 .
this upper limit is roughly eight orders of magnitude below the observed 1.4 ghz upper limit for the ngc 404 nucleus .
we therefore conclude the observed radio luminosity can not be explained by star formation alone , and is evidence for the presence of an agn in the ngc 404 nucleus .
the x - ray luminosity of the agn in ngc 404 bears directly on the question of whether accretion power can account for the observed luminosities of the optical emission lines , whose relative intensities are the defining characteristic of liners . in a recent study of the energy budgets of three dozen liners , including ngc 404 , eracleous et al .
( 2010 ) found that in the majority of cases the weak agn does not provide enough ionizing photons to account for the observed luminosities of the hydrogen recombination lines .
this conclusion is in general agreement with previous studies , as discussed in eracleous et al .
( 2010 ) . in the particular case of ngc 404
the number of ionizing photons was found to be deficient by a factor of @xmath68 .
the x - ray luminosity of ngc 404 measured here is only @xmath69% higher than that measured by eracleous et al .
( 2002 ) , after accounting for the different distance used in that paper , and is consistent ( within errors ) with the result obtained here .
therefore , the situation regarding the ionizing photon output of the agn remains the same .
however , maoz et al . (
1998 ) have estimated the ionizing luminosity of hot stars in the nucleus of ngc 404 based on their measurements of the uv spectrum with the _ hst _ and found it to be adequate to power the emission lines .
this conclusion was re - iterated by seth et al .
( 2010 ) , who also noted that this ionizing luminosity could be provided by a relatively small number of o stars in the nuclear star cluster .
therefore , the liner in ngc 404 appears to be powered by stellar processes .
we present the results of a 100 ks _
chandra _ observation of the nearby liner and s0 galaxy ngc 404 . the deep exposure has allowed us to critically test several forms for the 0.3 - 10 kev spectrum , and the excellent spatial resolution of _ chandra _ has enabled us to investigate the x - ray morphology of the ngc 404 nuclear region .
we find the 0.3 - 10 kev spectrum to be consistent with emission from hot gas plus a power law continuum , and we are specifically able to separate a point source of high energy photons from a diffuse source of soft x - ray emission
. additionally , find evidence for variability in the hard 2 - 10 kev emission .
the presence of a power law component and a moderate level of variability in the hard emission is indicative of x - ray emission powered by accretion onto a bh .
the estimated 0.3 - 10 kev luminosity ( @xmath72 - 3@xmath70 erg s@xmath6 ) is both comparable to that of a single galactic xrb and consistent with a imbh accreting at extremely low levels , on the order of a few times 10@xmath71 @xmath1yr@xmath6 .
we estimate the eddington ratio for both scenarios ( assuming a 10 @xmath1xrb and a 10@xmath72 @xmath1agn ) and compare our best - fit photon index @xmath3of the hard nuclear point source to the observed trends with @xmath50 for both xrbs and llagn .
we find the ngc 404 x - ray spectral shape and luminosity to be consistent with observed llagns , and inconsistent with observed xrbs .
we therefore favor the scenario in which the ngc 404 nucleus is powered by an imbh , with a mass on the order of @xmath0 @xmath1as dynamically estimated by seth et al .
such a weak agn does not produce a sufficient quantity of ionizing photons necessary to power a liner we therefore conclude that the liner in ngc 404 is powered by stellar processes
. very low accretion rates are common in nearby galaxies with bh masses less than a few times 10@xmath73 @xmath1(e.g .
, baganoff et al . 2001 ; garcia et al . 2000 ; ho et al . 2003 ) .
additional , multiwavelength observations of the ngc 404 nucleus are required to robustly determine the mass of the cental bh for example , resolved stellar populations within the nucleus would enable a robust dynamical mass determination , and a radio detection of the compact source would verify the location of ngc 404 on the fundamental plane of bh activity .
the authors would like to thank the anonymous referee for helpful comments .
b. binder and b. f. williams acknowledge support from _ chandra _ grant go1 - 12118x .
t. gaetz and p. p.
plucinsky acknowledge support of nasa contract nas8 - 03060 .
abel , n. p. & satyapal , s. 2008 , apj , 678 , 686 alonso - herrero , a. , rieke , m. j. , rieke , g. h. , & shields , j. c. 2000 , apj , 530 , 688 arnaud , k. a. 1996 aspc , 101 , 17 baganoff , f. k. et al .
2001 , nature , 413 , 45 baggett , w. e. , baggett , s. m. , & anderson , k. s. j. 1998 , aj , 116 , 1626 bell , m. e. et al .
2011 , mnras , 411 , 402 bker , t. , laine , s. , van der marel , r. p. , sarzi , m. , rix , h .- w . , ho , l. c. , & shields , j. c. 2002 , aj , 123 , 1389 bouchard , a. , prugniel , ph . , koleva , m. , & sharina , m. 2010 , a&a , 513 , 54 carollo , c. m. , stiavelli , m. , de zeeuw , p. t. , & mack , j. 1997 , aj , 114 , 2366 chandar , r. , leitherer , c. , tremonti , c. a. , calzetti , d. , aloisi , a. , meurer , g. r. , & de mello , d. 2005 apj , 628 , 210 chu , y .- h .
& mac low , m .- m .
1990 , apj , 365 , 510 colina , l. , gonzlez delgado , r. , mas - hesse , j. m. , leitherer , c. 2002 , apj , 579 , 545 condon , j. 1992 , araa , 30 , 575 constantin , a. , green , p. , aldcroft , t. , kim , d .- w . ,
haggard , d. , barkhouse , w. , & anderson , s. f. 2009 , apj , 705 , 1336 cotton , w. d. , cotton , j. j. , & arbizzani , e. 1999 , apjs , 125 , 409 corbel , s. , nowak , m. a. , fender , r. p. , tzioumis , a. k. , & markoff , s. 2003 , a&a , 400 , 1007 ct , p. et al .
2006 , apjs , 165 , 57 del ro , m. s. , brinks , e. , & cepa , j. 2004 , aj , 128 , 89 dudik , r. p. , satyapal , s. , gliozzi , m. , & sambruna , r. m. 2005 , apj , 620 , 113 eracleous , m. , shields , j. c. , chartas , g. , & moran , e. c. 2002 , apj , 565 , 108 eracleous , m. , hwang , j. a. & flohic , h. m. l. g. 2010 , apj , 711 , 796 flohic , h. m. l. g. , eracleous , m. , chartas , g. , shields , j. c. , & moran , e. c. 2006 , apj , 647 , 140 gallo , e. , fender , r. p. , & pooley , g. g. 2003 , mnras , 344 , 60 gallo , e. , fender , r. p. , miller - jones , j. c. a. , merloni , a. , jonker , p. g. , heinz , s. , maccarone , t. j. , & van der klis , m. 2006 , mnras , 370 , 1351 garcia , m. r. , murray , s. s. , primini , f. a. , forman , w. r. , mcclintock , j. e. , & jones , c. 2000 , apj , 537 , 23 graham , a. w. & guzmn , r. 2003 , apj , 125 , 2936 gonzlez delgado , r. m. , cid fernandes , r. ; prez , e. , martins , l. p. , storchi - bergmann , t. , schmitt , h. , heckman , t. , & leitherer , c. 2004 , apj , 605 , 127 gonzlez - martn , o. , masegosa , j. , mrquez , i. , guerrero , m. a. , & dultzin - hacyan , d. 2006 , a&a , 460 , 45 gonzlez - martn , o. , masegosa , j. , mrquez , i. , guainazzi , m. , & jimnez - bailn , e. 2009 , a&a , 506 , 1107 greene , j. e. & ho , l. c. 2007 , apj , 670 , 92 ho , l. c. , filippenko , a. v. , & sargent , w. l. w. , 1997 , apjs , 112 , 315 ho , l. c. 2001 , apj , 549 , 5 ho , l.c . , terashima , y. , & ulvestad , j. s. 2003 , 589 , 783 ho , l. c. 2008 , ara&a , 46 , 475 kalberla , p. m. w. et al .
2005 , a&a , 440 , 775 karachentsev , i. d. , karachentseva , v. e. , huchtmeier , w. k. , & makarov , d. i. 2004 , aj , 127 , 2031 koratkar , a. , deustua , s. e. , heckman , t. , filippenko , a. v. , ho , l. c. , & rao , m. 1995 , apj , 440 , 132 lira , p. , lawrence , a. , & johnson , r. a. 2000 , mnras , 319 , 17 maoz , d. , koratkar , a. , shields , j. c. , ho , l. c. , filippenko , a. v. , & sternberg , a. 1998 , aj , 116 , 55 maoz , d. , nagar , n. m. , falcke , h. , & wilson , a. s. 2005 , apj , 625 , 699 meurer , g. r. , heckman , t. m. , leitherer , c. , kinney , a. , robert , c. , & garnett , d. r. 1995 , aj , 110 , 2665 muoz marn , v. m. , gonzlez delgado , r. m. , schmitt , h. r. , cid fernandes , r. , prez , e. 2009 , ap&ss , 324 , 253 nagar , n. m. , falcke , h. , & wilson , a. s. 2005 , a&a , 435 , 521 olsson , e. , aalto , s. , thomasson , m. , beswick , r. , & httemeister , s. 2007 , a&a , 473 , 389 plana , h. , boulesteix , j. , amram , ph . , carignan , c. , & mendes de oliveira , c. 1998 , a&as , 128 , 75 peterson , b. m. et al .
2005 , apj , 632 , 799 pogge , r. w. , maoz , d. , ho , l. c. , & eracleous , m. 2000 , apj , 532 , 323 ravindranath , s. , ho , l. c. , peng , c. y. , filippenko , a. v. , & sargent , w. l. w. 2001 , aj , 122 , 653 satyapal , s. , sambruna , r. m. , & dudik , r. p. 2004 , a&a , 414 , 825 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , apj , 500 , 525 seth , a.c .
2010 , apj , 714 , 713 shemmer , o. , brandt , w. n. , netzer , h. , maiolino , r. , kaspi , s. 2006 , apj , 646 , 29 smith , r. k. , brickhouse , n. s. , liedahl , d. a. , & raymond , j. c. 2001 , apj , 556 , 91 terashima , y. , ho , l. c. , ptak , a. f. , mushotzky , r. f. , serlemitsos , p. j. , yaqoob , t. , kunieda , h. 2000 , apj , 533 , 729 tremonti , c. a. , calzetti , d. , leitherer , c. , & heckman , t. m. 2001 , apj , 555 , 322 tremonti , c. a. et al .
2004 , apj , 613 , 898 van paradijs , j. 1981 , a&a , 103 , 140 wang , q. , & helfand , d. j. 1991 , apj , 370 , 541 wang , j .- m . ,
watarai , k .- y . & mineshige , s. 2004 , apj , 607 , 107 wu , q. & gu , m. 2008 , apj , 682 , 212 zhang , w. m. , soria , r. , zhang , s. n. , swartz , d. a. , & liu , j. f. 2009 , apj , 699 , 281 | we present the results of a 100 ks _
chandra _ observation of the ngc 404 nuclear region . the long exposure and excellent spatial resolution of _ chandra _ has enabled us to critically examine the nuclear environment of ngc 404 , which is known to host a nuclear star cluster and potentially an intermediate - mass black hole ( on the order of a few times @xmath0 @xmath1 ) .
we find two distinct x - ray sources : a hard , central point source coincident with the optical and radio centers of the galaxy , and a soft extended region that is coincident with areas of high h@xmath2 emission and likely recent star formation .
when we fit the 0.3 - 8 kev spectra of each region separately , we find the hard nuclear point source to be dominated by a power law ( @xmath3= 1.88 ) , while the soft off - nuclear region is best fit by a thermal plasma model ( @xmath4 = 0.67 kev ) .
we therefore find evidence for both a power law component and hot gas in the nuclear region of ngc 404 .
we estimate the 2 - 10 kev luminosity to be 1.3@xmath5 erg s@xmath6 .
a low level of diffuse x - ray emission was detected out to @xmath715@xmath8(@xmath70.2 kpc ) from the nucleus .
we compare our results to the observed relationships between power law photon index and eddington ratio for both x - ray binaries and low luminosity active galaxies and find ngc 404 to be consistent with other low luminosity active galaxies .
we therefore favor the conclusion that ngc 404 harbors an intermediate - mass black hole accreting at a very low level . |
spin - orbital physics was initiated by kugel and khomskii @xcite who realized that orbital operators contribute to superexchange in mott insulators in a similar way as spins , so both degrees of freedom contribute jointly to spin - orbital superexchange @xcite . while the spin part has su(2 ) symmetry , the orbital part has remarkable orbital fluctuations @xcite and a much lower cubic symmetry in the perovskite systems where it is intrinsically frustrated @xcite .
this frustration is frequently released by spin order which coexists with orbital order following the goodenough - kanamori rules @xcite , and phases arise with alternating orbital ( ao ) order along ferromagnetic ( fm ) bonds coexisting with ferro - orbital ( fo ) order along antiferromagnetic ( af ) bonds .
well known examples are spin - orbital orders in lamno@xmath9 @xcite , lavo@xmath9 @xcite , and ca@xmath10ruo@xmath11 @xcite where spin - orbit coupling plays a role @xcite the latter two examples ( v , ru ) involve @xmath12 orbitals . indeed , a rather unique example of a spin - orbital system are perovskite vanadates , where a challenging competition between two types of spin - orbital order was observed @xcite . but a different scenario is also possible frustrated spin - orbital interactions may lead to the collapse of long - range order , as for instance in linio@xmath10 @xcite . another possibility is a spin - orbital liquid emerging from frustration @xcite .
doping of mott insulators leads to several remarkable phenomena .
recently short - range charge density wave called stripe phase was reported in doped cuprates @xcite . when coupled to spins it arises as self - organization of charge and spin degrees of freedom @xcite .
it has been suggested that the critical charge , orbital , and spin fluctuations near the quantum critical point provide the pairing interaction @xcite and the spectral properties of stripe phases may be seen as a signature of their stability @xcite . in manganites
hole doping generates orbital polarons that emerge in an af system by double - exchange mechanism @xcite , responsible for a change from af insulator to fm metal with @xmath13 orbital liquid @xcite . as in doped cuprates ,
the holes doped in @xmath13 orbitals are mobile in lamno@xmath9 , or in kugel - khomskii system @xcite , as well as in @xmath12 stripe phases with orbital polarons @xcite .
more complex phenomena are found in doped vanadates where a @xmath12 hole may hop only in two cubic directions @xcite , and robust @xmath14-type af ( @xmath14-af ) coexists with @xmath15-type ao ( @xmath15-ao ) order @xcite .
doping generates finite spectral weight within the mott - hubbard gap already at low @xmath16 ca doping in y@xmath17ca@xmath18vo@xmath9 @xcite , while at higher doping the system remains insulating and develops a kinetic gap @xcite .
the purpose of this paper is to consider the consequences of _ immobile holes _ in a ruthenate due to transition metal ions with a lower valence which are substituted for ru ( @xmath5 ) ions . as an experimental motivation we mention that : ( @xmath19 )
dilute cr doping for ru reduces the temperature of the orthorhombic distortion and induces fm order in ca@xmath10ru@xmath17cr@xmath18o@xmath11 ( with @xmath20 ) @xcite , and ( @xmath21 ) mn - substituted single crystals of sr@xmath9ru@xmath22mn@xmath18o@xmath23 reveal an unusual @xmath24-type af structure at @xmath25 @xcite .
therefore we consider : ( @xmath19 ) _ orbital dilution _ by @xmath26 ( mn@xmath27 , cr@xmath28 ) impurities in ( ca , sr)@xmath10ruo@xmath11 @xcite , and ( @xmath21 ) _ charge dilution _ due to @xmath6 doping which generates hole - doublon interactions .
these findings motivate the theoretical search for the consequences of both orbital and charge dilution .
we show that spin - orbital order may even be globally changed and dictated by defects in case of orbital dilution with @xmath29 @xcite .
note that in contrast to orbital dilution in cuprates @xcite where holes remove both spin and orbital degree of freedom , the present doping ( @xmath19 ) removes the orbital doublon but increases spin from @xmath4 to @xmath30 , while in ( @xmath21 ) spins @xmath4 stay unchanged but a doublon is replaced by a hole in an orbital ( holon ) .
the spin - orbital model for ca@xmath10ruo@xmath11 mott insulator is equivalent by an electron - hole transformation to that introduced some 15 years ago for vanadates @xcite .
thus we consider an interplay between @xmath4 spins and the @xmath12 orbital doublet @xmath31 for a doublon active along a given cubic axis .
for instance , the doublon involves @xmath32 active orbitals along the @xmath33 axis , while @xmath34 orbitals are filled by one electron each .
the model is isomorphic to the vanadate @xmath2 model @xcite , with doublons transforming into empty orbitals ( occupied by two holes ) .
we label @xmath12 orbitals by index @xmath35 when a given orbital is inactive along a cubic axis @xmath36 : @xmath37 we consider a two - dimensional ( 2d ) ruo@xmath10 @xmath38 plane in ca@xmath10ruo@xmath11 ( srruo@xmath9 ) , with ru ions connected via @xmath39 oxygen orbitals . for this plane @xmath40 ( @xmath41 )
orbitals are active along the @xmath42 ( @xmath43 ) axis , while @xmath44 orbitals are active along both @xmath43 and @xmath42 planar axes .
the superexchange for the host bonds @xmath45 along the @xmath46 axis @xcite , @xmath47 is @xmath48 and stabilizes @xmath14-af phase for the realistic parameters of ca@xmath10ruo@xmath11 , see fig .
[ fig : bond](a ) .
the above generic form @xcite follows from charge excitations , @xmath49 the interactions depend on the intraorbital coulomb @xmath50 element and hund s exchange @xmath51 in the host .
the orbital operators , @xmath52 and @xmath53 , are defined by active @xmath12 orbital along the axis @xmath35 @xcite . for a bond @xmath54
it may be rewritten in the orbital - only form by taking average over spin degrees of freedom , i.e. , @xmath55 with orbital operators @xmath56 defined by the axis @xmath35 and with coefficients @xmath57 and @xmath58 depending on spin average @xmath59 on a bond @xmath45 . for
the @xmath1-@xmath5 hybrid bonds charge @xmath60 excitations with the lowest energy do not generate an extra doublon but simply move it from @xmath61 to @xmath0 ion .
we use the convention that @xmath62 stands for the impurity site and @xmath63 for its host neighboring ion .
the energy involved in the charge excitation is , @xmath64 where @xmath65 and @xmath66 are the respective coulomb and hund s elements .
in addition it depends on the ionic energy @xmath67 , i.e. , the mismatch between the energy levels of the two atoms . for mn or cr impurities in ruthenates @xmath68 and plays a role of charge - transfer energy .
the energy @xmath69 ( [ delta ] ) defines two parameters which characterize the interactions along the hybrid @xmath1-@xmath5 bonds @xcite : @xmath70 the spin - orbital interaction on hybrid bond has the generic form similar to eq .
( [ eq : hhost ] ) , but the orbital operators are now defined only by the host site , and the spin operators are for @xmath4 on host and @xmath30 on impurity ion . a more transparent form of the impurity - host bond couples the impurity spin @xmath71 with the neighboring host spin @xmath72
it can be written as follows @xcite , @xmath73 here the spin exchange couplings @xmath74 depend on doublon projection operator @xmath75 at host site @xmath76 , and the doublon energy @xmath77 depends on hund s exchange @xmath78 eq .
( [ eq : imp ] ) .
it can be shown @xcite that the latter is the dominant energy scale , so for a single @xmath1-@xmath5 bond the doublon avoids the inactive ( @xmath35 ) orbital and spins couple with @xmath79 which can be either af if @xmath80 , or fm if @xmath81 .
the sign change at @xmath82 marks a quantum phase transition from af to fm spin correlations , see fig . [
fig : bond](c ) .
similar to the 2d kugel - khomskii model @xcite , this sign change leads here to rather exotic phases with nearly frustrated spins for several doping levels @xmath83 $ ] @xcite .
a single @xmath1 impurity at site @xmath19 modifies the spin - orbital order at its nearest neighbors ( nns ) @xmath84 , while second nns are little affected and thus they typically follow the @xmath14-af/@xmath15-ao order in the host .
in particular , the impurity spin reorients within the orbital polaron at @xmath85 which marks a transition from af to fm regime . both for a single impurity and at low @xmath86 doping , the @xmath1-@xmath5 bonds influence strongly spin - orbital order @xcite . for a higher periodic doping @xmath29 when half of the superexchange bonds are @xmath1-@xmath5 hybrid bonds , the overall spin - orbital order is dictated by them @xcite .
one finds that only every second undoped vertical line @xmath87 is fm , as in the @xmath14-af host phase , and host spins are inverted on any other vertical line and the doublon flips from orbital @xmath43 to @xmath42 .
such a modification of the orbital order stabilizes the fm interactions for @xmath33-@xmath43 doublon pairs on the horizontal bonds by double - exchange , in analogy to a doped @xmath12 system @xcite . in the phase diagram at @xmath29 doping
@xcite one finds a broad region of parameters where impurity spins are also frustrated .
frustration is released by quantum fluctuations which stabilize impurity spin orientation opposite to that expected in the host . at sufficiently large @xmath78 ( [ eq : imp ] ) , fm spin order takes over and @xmath15-ao order is then the same as in the undoped host .
thus , the most interesting novel spin - orbital phases are indeed found at the crossover from af to fm interactions along hybrid bonds .
charge dilution ( with periodic cluster of @xmath88 sites ) in two af phases in an @xmath38 plane : left @xmath14-af phase with fm vertical ( @xmath87 ) lines , and right nel af phase ( shown by arrows for @xmath89 ) .
orbital fluctuations on the bonds around the impurity at site @xmath90 are @xmath91 ( [ eq : h24 ] ) and increase from top to bottom .
bottom right color convention for an orbital occupied by a doublon at host sites and a hole at impurity site ( red circles ) , as well as the cluster of @xmath88 sites ( dotted square ) with the periodic boundary conditions on outer ( red ) bonds . ]
doping a spin - orbital @xmath92 system by @xmath93 ions is even more challenging . in this case
the impurity itself has another orbital degree of freedom , a hole ( holon ) in place of a doublon .
a particular feature of @xmath2 ions is that they promote orbital fluctuations in case when @xmath33 orbital ( @xmath34 ) is filled by one electron at all sites , and the other two @xmath94 ( @xmath32 ) fluctuate along the @xmath33 axis , see eqs .
( [ eq : or_defs ] ) .
this supports fm spin order even in the absence of hund s exchange @xcite .
it is thus challenging to consider such ions within a ruthenate where similar orbital fluctuations are also expected .
the generic form of spin - orbital superexchange ( [ eq : hhost ] ) applies as well to the hybrid @xmath2-@xmath5 bonds where charge @xmath95 excitations may generate an extra doublon in intermediate states at a @xmath0 ion for af bonds ( not shown ) , or instead they create a high - spin state along fm bonds , see fig .
[ fig : bond](d ) . as a result ,
the spin - orbital interactions are here quite complex and include several terms , so we do not reproduce them here .
they are controlled again by the charge - transfer energy and the same effective parameters as for orbital dilution , see eqs .
( [ eq : imp ] ) .
again , the spin exchange depends on whether the doublon at the host ion is in the active or inactive orbital on the considered @xmath2-@xmath5 bond .
it also depends on the configuration at the impurity site , i.e. , whether the holon is in an active or inactive orbital .
the derivation of spin - orbital superexchange for the hybrid @xmath2-@xmath5 bonds demonstrates a remarkable difference to the orbital dilution on @xmath1-@xmath5 bonds discussed above .
consider first the af spin bond shown in fig .
[ fig : bond](d ) .
charge excitations , @xmath95 , create , _ inter alia _ , two high - spin configurations which generate novel orbital quantum fluctuations , @xmath7 , which go beyond those found in the host @xcite .
in addition , a doublon may be created at the @xmath0 impurity ion and such configurations have to be combined with others to obtain eigenstates at both sites at finite @xmath96 . as a result , the fluctuations @xmath7 are amplified .
they are accompanied by ising orbital exchange terms and the usual fluctuations @xmath97 activated by finite hund s exchange ( not shown ) . for a fm spin configuration
the doublon gets a partner empty orbital at the host site and again , orbital quantum fluctuations @xmath98 arise at finite @xmath51 , see fig .
[ fig : bond](e ) . altogether , after averaging spin correlations , the purely orbital superexchange is , @xmath99 with orbital operators @xmath56 defined by the bond direction and the coefficients @xmath100 , @xmath101 , and @xmath102 .
as these parameters for the present hybrid bond are similar to those in the host ( [ eq : h44 ] ) , we use below , @xmath103 i.e. , we characterize the orbital superexchange on the hybrid bonds by the parameter @xmath104 which stands for double excitations in fig .
[ fig : bond](d ) . due to
such terms the orbital quantum number @xmath105 is not conserved .
a more complete analysis of charge dilution and the resulting phase diagrams will be presented elsewhere . here
we focus on the consequences of orbital fluctuations @xmath106 , both in @xmath14-af and nel af spin configurations in the @xmath38 plane , see fig .
[ fig : of ] . quantum fluctuations in the spin subsystem are small for @xmath4 spins and we neglect them . for realistic hund s exchange in the host and at the holon ion one
expects @xmath14-af order .
its orbital pattern at @xmath89 has equal orbital densities in @xmath43 and @xmath33 orbitals , see left column in fig .
[ fig : of ] .
indeed , this configuration guarantees maximal doublon fluctuations along the @xmath42 axis where the order is fm , as in the vanadium perovskites with holon fluctuations involving @xmath43 and @xmath42 orbitals along the @xmath33 axis @xcite . increasing
orbital fluctuations gradually destroy this optimal state and favor ao order along vertical fm columns , with enhanced @xmath43 or @xmath33 holon density at every other site . for small @xmath107
the @xmath108 term is active only along @xmath42-bonds and consequently there is no admixture of @xmath42 orbitals in the ground state . at larger @xmath109
we see the appearance of the @xmath42 orbital polarization at the impurity site and two neighboring host sites along @xmath43 axis which goes together with a severe change of the global orbital order . at @xmath110
a new state appears with almost complete separation into horizontal chains of @xmath43 ( @xmath33 ) orbitals and a small admixture of @xmath42 orbital density .
holon site has a similar orbital density distribution , with somewhat enhanced @xmath42 character .
we suggest that the orbital fluctuations of alternating @xmath43 and @xmath33 orbitals along the vertical bonds replace here double exchange to stabilize fm spin order , while such fluctuations for the same orbitals along the horizontal bonds support af spin order .
actually , fo order in the @xmath38 plane supports af nel state following the goodenough - kanamori rules @xcite , see the right column in fig .
[ fig : of ] .
this state is more robust and almost unchanged by double excitations @xmath7 with a realistic @xmath109 , and only at a high value of @xmath111 one finds increased @xmath43 or @xmath42 orbital densities following the checkerboard pattern around the impurity .
this effect seems to be purely local because already the second neighbors of the impurity in the host are unaffected even for such a large value of @xmath104 . unlike in @xmath14-af spin order
, we see that the defects in the orbital configurations are strongly localized or screened by the host .
this screening has an interesting geometrical mechanism : if we look around the impurity we see that the up and down neighbors get an admixture of the @xmath43 orbitals from the @xmath7 terms induced by the impurity .
these @xmath43-orbital defects can not however delocalize in the @xmath43 direction ( along two such bonds ) .
the defects can only move up , to the site number @xmath112 in the cluster ( see bottom right of fig .
[ fig : of ] ) . on the other hand ,
the horizontal neighbors of the impurity are in analogical situation with @xmath42-orbital defects which for the same reason can only go to the same site number @xmath112 .
finally , at this site one expects a destructive interference of the @xmath43-orbital defects going from the top and bottom bonds and the @xmath42-orbital defects going from the left and right ones as it exhibits only the initial @xmath33-orbital polarization .
it is thus possible that for a different distribution of impurities we would observe a much weaker localization of the defects .
orbital ( charge ) dilution plays a role in several mott insulators with spin - orbital order doped by transition metal ions with either @xmath1 or @xmath2 electron configuration .
we have shown that the orbital order around such impurities changes in general , so even in the dilute limit one may expect observable effects such as islands of reversed spins or doublon fluctuations .
we argue that the general trends reported here are generic and a meaningful insights into phase diagrams of systems with orbital dilution may be gained by investigating classical phase diagrams as quantum fluctuations are small @xcite .
double exchange leads to local or global changes of spin - orbital order , similar to orbital polarons @xcite or dimensional crossover with a change from @xmath15-af to @xmath14-af order in electron doped manganites @xcite .
doping may also generate novel spin - orbital - charge modulated patterns reported recently for @xmath12 systems @xcite .
summarizing , this study highlights the role played by orbital or charge dilution by transition metal ions in cubic spin - orbital systems and provides new insights necessary for theoretical understanding of ( mn , cr)-doped layered ruthenates and related systems .
previous studies @xcite have shown that @xmath1 impurities change radically spin - orbital order , and we expect an even stronger impact of @xmath2 impurities on spin - orbital order in the @xmath5 host the changes may be strongly centered around the impurity sites , as shown in the fo case , or completely smeared out by the host s fluctuations , as in the @xmath14-af case .
the theoretical studies of this doping are under way .
it is challenging to investigate the consequences of enhanced orbital fluctuations in experimental systems .
open access funding provided by max planck society . w.b .
acknowledges support by the european union s horizon 2020 research and innovation programme under the marie sklodowska - curie grant agreement no .
we acknowledge support by narodowe centrum nauki ( ncn , national science center ) , project 2012/04/a / st3/00331 .
ole , a.m. , khaliullin , g. , horsch , p. , feiner , l.f . :
fingerprints of spin - orbital physics in cubic mott insulators : magnetic exchange interactions and optical spectral weights .
b * 72 * , 214431 ( 2005 ) fujioka , j. , yasue , t. , miyasaka , s. , yamasaki , y. , arima , t. , sagayama , h. , inami , t. , ishii , k. , tokura , y. : critical competition between two distinct orbital - spin ordered states in perovskite vanadates
b * 82 * , 144425 ( 2010 ) campi , g. , innocenti , d. , bianconi , a. : cdw and similarity of the mott insulator - to - metal transition in cuprates with the gas - to - liquid - liquid transition in supercooled water
. j. supercond .
* 28 * , 1355 ( 2015 ) tranquada , j.m . ,
axe , j.d . ,
ichikawa , n. , nakamura , y. , uchida , s. , nachumi b. : neutron - scattering study of stripe - phase order of holes and spins in la@xmath113nd@xmath114sr@xmath115cuo@xmath11 .
b * 54 * , 7489 ( 1996 ) fujioka , j. , miyasaka , s. , tokura , y. : doping variation of anisotropic charge and orbital dynamics in y@xmath17ca@xmath18vo@xmath9 : comparison with la@xmath17sr@xmath18vo@xmath9 . phys . rev .
b * 77 * , 144402 ( 2008 ) avella , a. , horsch , p. , ole , a.m. : defect states and excitations in a mott insulator with orbital degrees of freedom : mott - hubbard gap versus optical and transport gaps in doped systems . phys . rev .
b * 87 * , 045132 ( 2013 ) qi , t.f . ,
korneta , o.b . , parkin , s. , de long , l.e . ,
schlottmann , p. , cao , g. : negative volume thermal expansion via orbital and magnetic orders in ca@xmath10ru@xmath17cr@xmath18o@xmath11 ( @xmath20 ) .
lett . * 105 * , 177203 ( 2010 ) mesa , d. , ye , f. , chi , s. , fernandez - baca , j.a . , tian . w. , hu , b. , jin , r. , plummer , e.w . , zhang , j. : single - bilayer @xmath24-type antiferromagnetism in mn - substituted sr@xmath9ru@xmath10o@xmath23 : neutron scattering study .
b * 85 * , 180410(r ) ( 2012 ) brzezicki , w. , ole , a.m. , cuoco , m. : spin - orbital order modified by orbital dilution in transition - metal oxides : from spin defects to frustrated spins polarizing host orbitals . phys .
x * 5 * , 011037 ( 2015 ) | we compare the effective spin - orbital superexchange triggered by magnetic @xmath0 impurities with @xmath1 and @xmath2 configurations and either no orbital degree of freedom ( orbital dilution ) or hole replacing a doublon ( charge dilution ) in a @xmath3 mott insulator with @xmath4 spins .
impurities causing orbital dilution act either as spin defects decoupled from the surrounding ions , or generate orbital polarons along @xmath1-@xmath5 hybrid bonds .
the exchange on these bonds determines which orbital is occupied by a doublon on the host site . in case of charge dilution by @xmath6 impurities additional @xmath7 terms
arise which enhance orbital fluctuations .
we show that such terms may radically change orbital pattern at relatively low doping by @xmath8 hole defects .
our findings provide new perspective for future theoretical and experimental studies of doped transition metal oxides . |
on november 12 , 2014 , the esa / rosetta descent module philae landed at the abydos site on the surface of comet 67p / churyumov - gerasimenko ( hereafter 67p / c - g ) . as part of the scientific payload aboard philae , the ptolemy mass spectrometer ( wright et al .
2007 ) performed the analysis of several samples from the surface at agilkia ( wright et al .
2015 ) and atmosphere at abydos ( morse et al .
the main molecules detected by ptolemy on the abydos site were h@xmath0o , co and co@xmath0 , with a measured co / co@xmath0 molar ratio of 0.07 @xmath1 0.04
. meanwhile , the co / co@xmath0 ratio has also been sampled in 67p / c - g s coma between august and september 2014 by the rosina double focusing mass spectrometer ( dfms ; balsiger et al .
2007 ; hssig et al . 2013 ) aboard the rosetta spacecraft .
strong variations of the co and co@xmath0 production rates , dominated by the diurnal changes on the comet , have been measured by rosina , giving co / co@xmath0 ratios ranging between 0.50 @xmath1 0.18 and 1.62 @xmath1 1.34 over the august - september 2014 time period ( hssig et al .
large fluctuations correlated with the sampled latitudes have also been observed and explained either by seasonal variations or by a compositional heterogeneity in the nucleus ( hssig et al .
2015 ) . further investigation of the coma heterogeneity performed by luspay - kuti et al .
( 2015 ) in the southern hemisphere of 67p / c - g at a later time period led to conclusions in favor of compositional heterogeneity .
this latter hypothesis is also reinforced by the ptolemy measurement of the co / co@xmath0 ratio at the abydos site , which is found outside the range covered by the rosina measurements ( morse et al .
2015 ) . here
, we aim at investigating the physico - chemical properties of the abydos subsurface which can reproduce the co / co@xmath0 ratio observed by ptolemy , assuming that the composition of the solid phase located beneath the landing site initially corresponds to the value in the coma . to investigate the possibility of a heterogeneous nucleus for 67p / c - g
, we have employed a comet nucleus model with i ) an updated set of thermodynamic parameters relevant for this comet and ii ) an appropriate parameterization of the illumination at the abydos site .
this allows us to mimic the thermal evolution of the subsurface of this location . by searching for the matching conditions between the properties of the abydos subsurface and the ptolemy data , we provide several constraints on the structural properties and composition of philae s landing site in different cases
the one - dimensional comet nucleus model used in this work is described in marboeuf et al .
this model considers an initially homogeneous sphere composed of a predefined porous mixture of ices and dust in specified proportions .
it describes heat transmission , gas diffusion , sublimation / recondensation of volatiles within the nucleus , water ice phase transition , dust release , and dust mantle formation .
the model takes into account different phase changes of water ice , within : amorphous ice , crystalline ice and clathrates .
the use of a 1d model is a good approximation for the study of a specific point at the surface of the comet nucleus , here the abydos landing site .
however , since 67p / c - g s shape is far from being a sphere ( sierks et al .
2015 ) , we have parameterized the model in a way that correctly reproduces the illumination conditions at abydos .
this has been made possible via the use of the 3d shape model developed by jorda et al .
( 2014 ) which gives the coordinates of abydos on the surface of 67p / c - g s nucleus , as well as the radius corresponding to the abydos landing site and the normal to the surface at this specific location .
the abydos landing site is located just outside the hatmethit depression , at the very edge of the area illuminated during the mapping and close observation phases of the rosetta mission , roughly until december 2014 .
it is also at the edge of a relatively flat region of the small lobe illuminated throughout the perihelion passage of the comet .
geomorphologically , abydos is interpreted as being a rough deposit region composed of meter - size boulders ( lucchetti et al .
other geometric parameters specific to 67p / c - g , such as the obliquity and the argument of the subsolar meridian at perihelion , are calculated from the orientation of the spin axis computed during the shape reconstruction process ( jorda et al . 2014 ) .
table 1 summarizes the main parameters used in this work .
the porosity and dust / ice ratio of the cometary material are set in the range of measurement of 80 @xmath1 5 % ( kofman et al .
2015 ) and 4 @xmath1 2 ( rotundi et al .
2014 ) , respectively .
these two parameters are linked through the density of the cometary material , and are set to be compatible with the preliminary value determined by jorda et al .
( 2014 ) ( 510 @xmath1 20 kg / m@xmath3 ) .
67p / c - g s thermal inertia is estimated to be in the 10150 w k@xmath4 m@xmath5 s@xmath6 range based on the measurement obtained by the rosetta / virtis instrument ( leyrat et al .
according to the same study , regions surrounding abydos are characterized by a thermal inertia in the lower half of this range .
we have therefore chosen a low thermal inertia close to 50 w k@xmath4 m@xmath5 s@xmath6 .
in addition to water ice and dust , the solid phase of our model includes co and co@xmath0 volatiles .
although coma abundances do not necessarily reflect those in the nucleus , they constitute the most relevant constraint available on its composition .
we thus considered the co / co@xmath0 ratio ( 1.62 @xmath1 1.34 ) measured by rosina on august 7 , 2014 at 18 hours , as representative of the bulk ice composition in the nucleus , and more specifically in the abydos subsurface .
this ratio is derived from the co / h@xmath0o and co@xmath0/h@xmath0o rosina measurements performed at this date , which are equal to 0.13 @xmath1 0.07 and 0.08 @xmath1 0.05 , respectively ( hssig et al .
we selected the date of august 7 because i ) the corresponding rosina measurements were performed at high northern latitudes where the abydos site is located , and ii ) the large co / co@xmath0 range obtained at this moment covers all values measured by rosina at other dates ( including the late value obtained by le roy et al .
( 2015 ) on october 20 for the northern hemisphere , namely co / co@xmath0 = 1.08 ) .
the three main phases of ices identified in the literature , namely crystalline ice , amorphous ice and clathrate phase , are considered in this work .
outgassing of volatiles in 67p / c - g could then result from the sublimation of ices , amorphous - to - crystalline ice phase transition , or destabilization of clathrates in the crystalline ice , amorphous ice and clathrates cases , respectively .
because the properties of volatiles trapping in the nucleus matrix strongly depend on the considered icy phase , the following models have been considered : [ [ crystalline - ice - model ] ] crystalline ice model + + + + + + + + + + + + + + + + + + + + + water ice is fully crystalline , meaning that no volatile species are trapped in the water ice structure . here ,
co and co@xmath0 are condensed in the pores of the matrix made of water ice and dust ; [ [ amorphous - ice - model ] ] amorphous ice model + + + + + + + + + + + + + + + + + + + the matrix itself is made of amorphous water ice with a volatile trapping efficiency not exceeding @xmath210% . in this case , the cumulated mole fraction of volatiles is higher than this value , implying that an extra amount of volatiles is crystallized in the pores . with this in mind ,
we consider different distributions of co and co@xmath0 in the both phases of this model ; [ [ clathrate - model ] ] clathrate model + + + + + + + + + + + + + + + water ice is exclusively used to form clathrates . similarly to amorphous ice , clathrates have a maximum trapping capacity ( @xmath217% ) .
the extra amount of volatiles , if any , also crystallizes in the pores . in our case , however , co is fully trapped in clathrates and escapes only when water ice sublimates . in contrast , we assume that solid co@xmath0 exists in the form of crystalline co@xmath0 in the pores of the nucleus because this molecule is expected to condense in this form in the protosolar nebula ( mousis et al . 2008 ) .
lcl rotation period ( hr ) & 12.4 & mottola et al .
( 2014 ) + obliquity ( @xmath7 ) & 52.25 & + subsolar meridian @xmath8 ( @xmath7 ) & -111 & + co - latitude ( @xmath7 ) & -21 & + initial radius ( km ) & 2.43 & + bolometric albedo ( % ) & 1.5 & fornasier et al .
( 2015 ) + dust / ice mass ratio & 4 @xmath1 2 & rotundi et al .
( 2014 ) + porosity ( % ) & 80 @xmath1 5 & kofman et al . (
2015 ) + density ( kg / m@xmath3 ) & 510 @xmath1 20 & jorda et al . ( 2014 ) + i ( w k@xmath4 m@xmath5 s@xmath6 ) & 50 & leyrat et al .
( 2015 ) + co / co@xmath0 initial ratio & 1.62 @xmath1 1.34 & hssig et al .
our results show that the illumination at the surface of the abydos site is a critical parameter for the evolution of the nucleus , regardless the considered ice structure .
consequently , all three models described in section 2 present the same behavior up to a given point .
we first describe the characteristics displayed in common by the three models by presenting the thermal evolution of the crystalline ice model , before discussing the variations resulting from the different assumptions on the nature of ices .
figure 1 shows the time evolution of the nucleus stratigraphy , which corresponds to the structural differentiation occurring in the subsurface of the abydos site .
this differentiation results from the sublimation of the different ices . after each perihelion passage
, the sublimation interfaces of co and co@xmath0 reach deeper layers beneath the nucleus surface , with a progression of @xmath220 m per orbit .
the co sublimation interface always propagates deeper than its co@xmath0 counterpart because of the higher volatility of the former molecule . on the other hand , because surface ablation is significant , the progression of these interfaces is stopped by the propagation of the water sublimation front after perihelion .
this allows the abydos region to present a `` fresh '' surface after each perihelion .
5 orbits of 67p / c - g ) , showing the interfaces of sublimation of considered volatile species .
the surface ablation occurs at each perihelion ( represented by the vertical lines ) and reaches all interfaces.,width=340 ] at the surface of the abydos site , the outgassing rates of co and co@xmath0 vary with the illumination conditions , reaching maxima at perihelion and minima at aphelion . because the sublimation interface of co@xmath0 is closer to the surface , its production rate is more sensitive to the illumination conditions than that of co. as a result , the outgassing rate of co@xmath0 presents important variations with illumination while that of co is less affected .
this difference strongly impacts the evolution of the co / co@xmath0 outgassing ratio at the surface of abydos ( see figure 2 ) .
close to perihelion , this ratio crosses the range of values measured by ptolemy ( 0.07 @xmath1 0.04 ) and reaches a minimum .
note that the co / co@xmath0 outgassing ratio presents spikes during a certain period after perihelion .
these spikes appear when the co and co@xmath0 interfaces of sublimation are dragged out to the surface by ablation and result from temperature variations induced by diurnal variations of the insolation .
we define @xmath9 as the time difference existing between the ptolemy co / co@xmath0 observations ( here 0.07 measured on november 12 , 2014 ) and the epoch at which our model reproduces these data ( see figure 2 ) . in each case investigated , we vary the input parameters of the model to minimize the value of @xmath9 ( see table 2 for details ) . we have also defined the quantity @xmath10 , namely the fraction of 67p / c - g s orbital period ( @xmath11 yr ) corresponding to @xmath9 .
the results of our simulations are indicated below . outgassing ratio at abydos in the case of the crystalline ice model , during one orbit .
the green line and green area represent the ptolemy central value and its range of uncertainty , respectively .
the blue dots correspond to the measurement epoch ( november 12 , 2014 ) .
vertical lines correspond to passages at perihelion .
see text for a description of @xmath9.,width=340 ] lccc parameters & & & + dust / ice mass ratio & 6 & 4 & 4 + porosity ( % ) & 78 & 76 & 76 + density ( kg / m@xmath3 ) & 516 & 510 & 519 + i ( w k@xmath4 m@xmath5 s@xmath6 ) & @xmath260 & 4060 & @xmath250 + co / co@xmath0 initial ratio & 0.46 & 1.62 & 0.46 + co / h@xmath0o initial abundance & 6% & 13% & 6% + co@xmath0/h@xmath0o initial abundance & 13% & 8% & 13% + results & & & + @xmath9 ( day ) & 52 & 4 & 34 + @xmath12 ( % ) & 2.2 & 0.17 & 1.4 + [ [ crystalline - ice - model-1 ] ] crystalline ice model + + + + + + + + + + + + + + + + + + + + + in this case , we find that the initial co / co@xmath0 ratio and the dust / ice ratio adopted in the nucleus have a strong influence on @xmath9 . this quantity is minimized i ) when the adopted dust / ice ratio becomes higher than those found by rotundi et al .
( 2014 ) and ii ) if the selected initial co / co@xmath0 ratio is lower than the nominal value found by rosina ( hssig et al .
figure 3 shows the evolution of @xmath9 as a function of the dust / ice ratio for two different values of the initial co / co@xmath0 ratio , namely 1.62 ( the central value ) and 0.46 ( close to the lower limit ) .
these results confirm the aforementioned trend : with co / co@xmath0 = 0.46 and a dust / ice ratio of 6 or higher , the ptolemy s measurement epoch is matched with @xmath9 = 52 @xmath1 27 days or lower , which corresponds to @xmath12 lower than @xmath22% .
these results can be explained by the thermal conductivity of crystalline water ice , which is in the 320 w m@xmath4 k@xmath4 range ( klinger 1980 ) , considering the temperatures in the comet . because dust has a conductivity of 4 w m@xmath4 k@xmath4 ( ellsworth & schubert 1983 ) , the global conductivity decreases with the increase of the dust / ice ratio in the nucleus . since heating in the nucleus
is mostly provided by surface illumination , a low conductivity increases the temperature gradient between the upper and deeper layers ( where the co@xmath0 and the co sublimation interfaces are located , respectively ) .
this gradient enhances more the sublimation rate of co@xmath0 ice than that of co ice , leading to smaller co / co@xmath0 outgassing ratios at the surface and to a smaller @xmath9 ( see figure 2 ) . and @xmath12 as a function of the dust / ice mass ratio for two different values of the initial co / co@xmath0 ratio in the case of the crystalline ice model
( see text).,width=340 ] [ [ amorphous - ice - model-1 ] ] amorphous ice model + + + + + + + + + + + + + + + + + + + here , values of @xmath9 are significantly lower than those obtained with the crystalline ice model ; a @xmath9 of 4 @xmath1 9 days is obtained using the central values listed in table 2 for the initial co / co@xmath0 and dust / ice ratios , leading to @xmath13 .
higher values of the co / co@xmath0 ratio never allow our model matching the ptolemy data : the co / co@xmath0 outgassing ratio is increased sufficiently high so that even its minimum becomes higher than the range of measurements performed by ptolemy . on the other hand ,
lower co / co@xmath0 ratios increase @xmath9 .
interestingly , the results of this model are poorly affected by the considered dust / ice ratio . within the 0.121.35 w m@xmath4 k@xmath4 range ( klinger 1980 ) , the conductivity of amorphous water ice lays under those of dust ( 4 w m@xmath4 k@xmath4 ) and crystalline ices ( 320 w m@xmath4 k@xmath4 ) . since amorphous ice dominates the volatile phase , the mean
conductivity is never too far from to that of dust , irrespective of the dust / ice ratio considered within the observed range .
[ [ clathrate - model-1 ] ] clathrate model + + + + + + + + + + + + + + + in this case , low co / co@xmath0 ratios ( still within the range given in table 1 ) are required to get values of @xmath9 below 50 days ( i.e. , @xmath12 below 2% ) .
similarly to the case of the amorphous ice model , @xmath9 is poorly sensitive to the variation of the dust / ice ratio because the conductivity of clathrates ( 0.5 w m@xmath4 k@xmath4 ; krivchikov et al .
2005a , 2005b ) is small compared to that of dust .
our goal was to investigate the possibility of recovering the value of the co / co@xmath0 outgassing ratio measured by the ptolemy instrument at the surface of abydos .
interestingly , all considered models match the ptolemy value with @xmath12 lower than @xmath22% , provided that an optimized set of parameters is adopted for the abydos region . despite the fact it is poorly sensitive to the adopted dust / ice ratio , a nucleus model dominated by amorphous ice ( and possibly including a smaller fraction of crystalline ices ) gives the best results ( @xmath9 @xmath14 4 days , i.e. @xmath12 @xmath14 0.2% ) for a primordial co / co@xmath0 ratio equal to the central value measured by the rosina instrument .
on the other hand , the crystalline ice and clathrate models require a primordial co / co@xmath0 ratio close to the lower limit sampled by rosina to obtain values of @xmath9 under 50 days ( i.e. @xmath12 under 2% ) .
we stress the fact that the co / co@xmath0 range of validity is used under the assumption that the rosina measurements correspond to the bulk nucleus abundances , which is not necessarily true .
a second requirement to minimize the value of @xmath9 in the crystalline ice model is the necessity to adopt a dust / ice ratio at least equal to or higher than the upper limit determined by rotundi et al .
( 2014 ) for 67p / c - g .
this is supported by the pictures taken at abydos by the civa instrument ( bibring et al .
2007 ) aboard the philae module .
the very low reflectance of 35% of the abydos region ( bibring et al .
2015 ) is in agreement with the osiris and virtis reflectance measurements in the visible ( sierks et al .
2015 ) and near - ir ( capaccioni et al .
2015 ) , which are consistent with a low ice content in the upper surface layer ( capaccioni et al .
2015 ) .
surface illumination can also greatly influence the co / co@xmath0 outgassing ratio on 67p / c - g . to quantify this effect , we have simulated a point of 67p / c - g s nucleus which is more illuminated in comparison to abydos .
we have performed a set of simulations at a co - latitude of -52.25@xmath7 , a point that receives permanent sunlight around perihelion . at the date when the co / co@xmath0 outgassing ratio at abydos is equal to 0.07 ( the central value of ptolemy s range of measurement )
, we obtain a different value for the co / co@xmath0 outgassing ratio at this new location , irrespective of the adopted model . for the crystalline ice model , the outgassing ratio at the illuminated site reaches a value of 0.11 , which is still within ptolemy s range of measurement .
this implies that a different illumination can not explain a strong variation of the co / co@xmath0 outgassing ratio if the nucleus presents a homogeneous composition . in this case , the difference between the ptolemy and rosina measurements is clearly due to a heterogeneity in the nucleus composition . on the other hand , the co / co@xmath0 outgassing ratio at the illuminated site is equal to 0.74 and 0.76 in the cases of the amorphous ice and clathrate models , respectively .
these values are within rosina s range of measurements , implying that the difference of illumination is sufficient to explain the difference with the co / co@xmath0 ratio sampled at abydos , assuming a homogeneous nucleus . in summary ,
all possible water ice structures are able to reproduce the observations made by ptolemy , assuming that the primordial co / co@xmath0 ratio is the one inferred by rosina .
each case requires a unique set of input parameters taken from the range of values inferred by rosetta and which describes the structure and composition of the material . according to our simulations , a heterogeneity in the composition of 67p
/ c - g s nucleus is possible only if the nucleus is composed of crystalline ices .
however , if we consider different ice phases like amorphous ice or clathrates , the difference between the ptolemy and rosina measurements could simply originate from the variation of illumination between different regions of the nucleus . in the upcoming months
, the philae module could awaken and allow the ptolemy mass spectrometer to perform additional measurements of the co / co@xmath0 ratio . by comparing these new values with the different co / co@xmath0 outgassing ratios predicted by our three models at the same date
, we would be able to see which model is the most reliable , and thus to determine which water ice structure is dominant at the surface of 67p / c - g s nucleus .
o.m . acknowledges support from cnes .
this work has been partly carried out thanks to the support of the a*midex project ( n^o^ anr-11-idex-0001 - 02 ) funded by the `` investissements davenir '' french government program , managed by the french national research agency ( anr ) .
funding and operation of the ptolemy instrument was provided by the science and technology facilities council ( consolidated grant st / l000776/1 ) and uk space agency ( post - launch support st / k001973/1 ) .
k . acknowledges support from the nasa jet propulsion laboratory ( subcontract no .
1496541 ) . | on november 12 , 2014 , the esa / rosetta descent module philae landed on the abydos site of comet 67p / churyumov - gerasimenko . aboard this module , the ptolemy mass spectrometer measured a co / co@xmath0 ratio of 0.07 @xmath1 0.04 which differs substantially from the value obtained in the coma by the rosetta / rosina instrument , suggesting a heterogeneity in the comet nucleus . to understand this difference , we investigated the physico - chemical properties of the abydos subsurface leading to co / co@xmath0 ratios close to that observed by ptolemy at the surface of this region .
we used a comet nucleus model that takes into account different water ice phase changes ( amorphous ice , crystalline ice and clathrates ) , as well as diffusion of molecules throughout the pores of the matrix .
the input parameters of the model were optimized for the abydos site and the rosina co / co@xmath0 measured ratio is assumed to correspond to the bulk value in the nucleus .
we find that all considered structures of water ice are able to reproduce the ptolemy observation with a time difference not exceeding @xmath250 days , i.e. lower than @xmath22% on 67p / churyumov - gerasimenko s orbital period .
the suspected heterogeneity of 67p / churyumov - gerasimenko s nucleus is also found possible only if it is constituted of crystalline ices .
if the icy phase is made of amorphous ice or clathrates , the difference between ptolemy and rosina s measurements would rather originate from the spatial variations in illumination on the nucleus surface .
an eventual new measurement of the co / co@xmath0 ratio at abydos by ptolemy could be decisive to distinguish between the three water ice structures . |
the quark model @xcite can reproduce the behavior of observables such as the spectrum and the magnetic moments in the baryon and meson sector , but it neglects quark - antiquark pair - creation ( or continuum - coupling ) effects . above threshold ,
these couplings lead to strong decays and below threshold , they lead to virtual @xmath2 ( @xmath3 ) components in the hadron wave function and shifts of the physical mass with respect to the bare mass .
the unquenching of the quark model for hadrons is a way to take these components into account .
pioneering works on the unquenching of quark model were done by trnqvist and collaborators , who used an unitarized qm @xcite , while van beveren and rupp used an t - matrix approach @xcite .
these methods were used ( with a few variations ) by several authors to study the influence of the meson - meson ( meson - baryon ) continuum on meson ( baryon ) observables .
these techniques were applied to study of the scalar meson nonet ( @xmath4 , @xmath5 , etc . ) of ref .
@xcite in which the loop contributions are given by the hadronic intermediate states that each meson can access .
it is via these hadronic loops that the bare states become `` dressed '' and the hadronic loop contributions totally dominate the dynamics of the process . on the other hand , isgur and coworkers in ref .
@xcite demonstrated that the effects of the @xmath0 sea pairs in meson spectroscopy is simply a renormalization of the meson string tension .
also , the strangeness content of the nucleon and electromagnetic form factors were investigated in @xcite , whereas capstick and morel in ref .
@xcite analyzed baryon meson loop effects on the spectrum of nonstrange baryons . in meson sector , eichten _ et al .
_ explored the influence of the open - charm channels on the charmonium properties using the cornell coupled - channel model @xcite to assess departures from the single - channel potential - model expectations . in this contribution
, we discuss some of the latest applications of the uqm ( the approach is a generalization of the unitarized quark model @xcite ) to study the flavor asymmetry and strangeness of the proton , in wich the effects of the quark - antiquark pairs were introduced into the constituent quark model ( cqm ) in a systematic way and the wave fuctions were given explicitly .
finally , the uqm is applied to describe meson observables and the spectroscopy of the charmonium and bottomonium .
in the unquenched quark model for baryons @xcite and mesons @xcite , the hadron wave function is made up of a zeroth order @xmath6 ( @xmath0 ) configuration plus a sum over the possible higher fock components , due to the creation of @xmath1 @xmath0 pairs .
thus , we have @xmath7 ~ , \end{aligned}\ ] ] where @xmath8 stands for the @xmath1 quark - antiquark pair - creation operator @xcite , @xmath9 is the baryon / meson , @xmath10 and @xmath11 represent the intermediate state hadrons , see figures [ figbaryon ] and [ figmeson ] .
@xmath12 , @xmath13 and @xmath14 are the corresponding energies , @xmath15 and @xmath16 the relative radial momentum and orbital angular momentum between @xmath10 and @xmath11 and @xmath17 is the total angular momentum .
it is worthwhile noting that in refs .
@xcite , the constant pair - creation strength in the operator ( [ eqn : psi - a ] ) was substituted with an effective one , to suppress unphysical heavy quark pair - creation .
two diagrams can contribute to the process @xmath18 .
@xmath19 and @xmath19 stand for the various initial ( i = 1 - 4 ) and final ( i = 5 - 8) quarks or antiquarks , respectively . ] two diagrams can contribute to the process @xmath18 .
@xmath19 and @xmath19 stand for the various initial ( i = 1 - 4 ) and final ( i = 5 - 8) quarks or antiquarks , respectively . ] in the uqm the matrix elements of an observable @xmath20 can be calculated as @xmath21 where @xmath22 is the state of eq .
( [ eqn : psi - a ] ) .
the result will receive a contribution from the valence part and one from the continuum component , which is absent in naive qm calculations .
the introduction of continuum effects in the qm can thus be essential to study observables that only depend on @xmath0 sea pairs , like the strangeness content of the nucleon electromagnetic form factors @xcite or the flavor asymmetry of the nucleon sea @xcite .
in other cases , continuum effects can provide important corrections to baryon / meson observables , like the self - energy corrections to meson masses @xcite or the importance of the orbital angular momentum in the spin of the proton @xcite .
the first evidence for the flavor asymmetry of the proton sea was provided by nmc at cern @xcite . the flavor asymmetry in the proton
is related to the gottfried integral for the difference of the proton and neutron electromagnetic structure functions @xmath23 ~.\end{aligned}\ ] ] under the assumption of a flavor symmetric sea , one obtains the gottfried sum rule @xmath24 .
the final nmc value is @xmath25 at @xmath26 ( gev / c)@xmath27 for the gottfried integral over the range @xmath28 @xcite , which implies a flavor asymmetric sea .
the violation of the gottfried sum rule has been confirmed by other experimental collaborations @xcite .
theoretically , it was shown in ref .
@xcite , that the coupling of the nucleon to the pion cloud provides a natural mechanism to produce a flavor asymmetry .
comparison the value of gottfried sum rule calculated within uqm with the experimental data from nmc 1994 , nmc 1997 , hermes , and e866 . figure taken from ref .
@xcite ; aps copyright .
, width=226 ] in the uqm , the flavor asymmetry can be calculated from the difference of the probability to find @xmath29 and @xmath30 sea quarks in the proton @xmath31 ~. \label{asym}\end{aligned}\ ] ] note that , even in absence of explicit information on the ( anti)quark distribution functions , the integrated value can be obtained directly from the left - hand side of eq .
( [ asym ] ) .
our result is shown in fig .
[ protonasy ] .
the results for the two strangeness observables were obtained in a calculation involving a sum over intermediate states up to four oscillator shells for both baryons and mesons @xcite . in the uqm formalism , the strange magnetic moment of the proton is defined as the expectation value of the operator @xmath32\end{aligned}\ ] ] on the proton state of eq .
( [ eqn : psi - a ] ) , which represents the contribution of the strange quarks to the magnetic moment fo the proton ; @xmath33 is the magnetic moment of the quark i times a projector on strangeness and the strange quark magnetic moment is set as in ref .
our result is @xmath34(see fig.[smagnetic ] ) .
similarly , the strange radius of the proton is defined as the expectation value of the operator @xmath35 on the proton state of eq .
( [ eqn : psi - a ] ) , where @xmath36 is the electric charge of the quark @xmath37 times a projector on strangeness , @xmath38 and @xmath39 are the coordinates of the quark @xmath37 and of the intermediate state center of mass , respectively .
the expectation value of @xmath40 on the proton is equal to @xmath41 . in fig .
[ sradio ] our result is compared with the experimental data .
comparison between our resulting value for the strange radius of the proton in the uqm .
figure taken from ref .
@xcite ; aps copyright . ]
comparison between our resulting value for the strange radius of the proton in the uqm .
figure taken from ref .
@xcite ; aps copyright . ]
in refs . @xcite , the method was used by some of us to compute the charmonium ( @xmath42 ) and bottomonium ( @xmath43 ) spectra with self - energy corrections , due to continuum coupling effects . in the uqm , the physical mass of a meson , @xmath44 is given by the sum of two terms : a bare energy , @xmath12 , calculated within a potential model @xcite , and a self energy correction , @xmath45 computed within the uqm formalism . bottomonium spectrum with self energies corrections .
black lines are theoretical predictions and blue lines are experimental data available .
figure taken from ref .
@xcite ; aps copyright . ]
bottomonium spectrum with self energies corrections .
black lines are theoretical predictions and blue lines are experimental data available .
figure taken from ref .
@xcite ; aps copyright . ] our results for the self energies corrections of charmonia @xcite and bottomonia @xcite spectrums , are shown in figures [ charm ] and [ botton ] .
in the baryon sector , our results for asymmetry and `` strangeness '' observables , as shown in figures [ protonasy ] , [ smagnetic ] and [ sradio ] , are in agreement with the experimental data .
these observables can only be understood when continuum components in the wave function are included .
our results in the meson sector for the self energies corrections of charmonium and bottomonium spectra , see figures [ charm ] and [ botton ] , show that the pair - creation effects on the spectrum of heavy mesons are quite small . specifically for charmonium and bottomonium states ,
they are of the order of @xmath46 and @xmath47 , respectively . the relative mass shifts , i.e. the difference between the self energies of two meson states , are in the order of a few tens of mev .
however , as qm s can predict the meson masses with relatively high precision in the heavy quark sector , even these corrections can become significant .
these results are particularly interesting in the case of states close to an open - flavor decay threshold , like the @xmath48 and @xmath49 mesons . in our picture
the @xmath48 can be interpreted as a @xmath42 core [ the @xmath50 , plus higher fock components due to the coupling to the meson - meson continuum . in ref .
@xcite , we showed that the probability to find the @xmath48 in its core or continuum components is approximately @xmath51 and @xmath52 , respectively . in conclusion ,
the flavor asymmetry in the proton can be well described by the uqm .
the effects of the continuum components on the `` strangeness '' observables of the proton are found to be negligible
. nevertheless , our results are compatible with the latest experimental data and recent lattice calculations . in the meson sector
our self energies corrections for charmonia and bottomonia are found to be significant .
this work is supported in part by papiit - dgapa , mexico ( grant in107314 ) and infn sezione di genova .
100 eichten e , gottfried k , kinoshita t , b kogut j , lane d k and -m yan t 1975 _ phys .
lett . _ * 34 * 369 ; eichten e , gottfried k , kinoshita t , d lane k and -m yan t 1978 _ phys . rev .
_ d * 17 * 3090 ; 1980 _ phys .
_ d * 21 * 203 . godfrey s and isgur n 1985 _ phys
_ d * 32 * 189 .
capstick s and isgur n 1986 _ phys .
_ d * 34 * 2809 .
ferraris m , giannini m m , pizzo m , santopinto e and tiator l 1995 _ phys .
. _ b * 364 * 231 ; santopinto e , iachello f and giannini m m 1998 _ eur . phys .
j. _ a * 1 * 307 ; santopinto e and giannini m m 2012 _ phys . rev .
_ c * 86 * 065202 ; giannini m m , santopinto e 2015 _ chin . j. phys . _ * 53 * 020301 ; aiello
m 1996 _ et al .
phys.lett._ b 387 , 215 ; aiello _ et al .
_ 1988 _ j. of phys .
_ g * 24 * 753 ; bijker r , iachello f , santopinto e 1988 _ j. of phys .
_ a * 31 * 9041 ; m.
de sanctis _ et al .
_ 2007 _ phys .
rev . _ c * 6 * 062201 .
santopinto e 2005 _ phys .
_ c * 72 * 022201 ; ferretti j , vassallo a and santopinto e 2011 _ phys .
_ c * 83 * , 065204 ; de sanctis m , ferretti j , santopinto e and vassallo a arxiv:1410.0590 . de sanctis , m , ferretti j 2011 santopinto e _
_ c * 84 * 055201 .
van beveren e , rijken t a , k. metzger , dullemond c , rupp g and ribeiro j e 1986 _ z. phys .
_ c * 30 * 615 .
ono s and trnqvist n a 1984 _ z. phys .
_ c * 23 * 59 ; heikkila k , ono s and trnqvist n a 1984 _ phys .
_ d * 29 * 110 [ 1984 erratum - ibid .
* 29 * 2136 ] ; ono s , sanda a i and trnqvist n a 1986 _ phys .
_ d * 34 * 186 .
ferretti j , galat g and santopinto e 2014 _ phys .
_ d * 90 * 054010 .
kalashnikova y s 2005 _ phys .
_ d * 72 * 034010 .
amaudruz p _
_ 1991 _ phys .
_ * 66 * 2712 ; arneodo m _
_ 1997 _ nucl .
_ b * 487 * 3 . | in this contribution , we briefly analyze the formalism of the unquenched quark model ( uqm ) and its application to the description of several observables of hadrons . in the uqm ,
the effects of @xmath0 sea pairs are introduced explicitly into the quark model through a qcd - inspired @xmath1 pair - creation mechanism .
we present our description of flavour asymmetry and strangeness in the proton when baryon - meson components are included . in the meson sector , the charmonium and bottomonium spectra with self - energy corrections due to the coupling to the meson - meson components . |
study of unstable nuclei with radioactive ion beam ( rib ) facilities has opened an exciting channel to look for some crucial issues in context of both nuclear structure and nuclear astrophysics @xcite .
unstable nuclei play an influential , and in some cases dominant , role in many phenomena in the cosmos such as novae , supernovae , x - ray bursts , and other stellar explosions . at extremely high temperatures ( @xmath0 k ) of these astrophysical environments
, the interaction times between nuclei can be so short ( @xmath1 seconds ) that unstable nuclei formed in a nuclear reaction can undergo subsequent reactions before they decay .
sequences of ( predominantly unmeasured ) nuclear reactions occurring in exploding stars are therefore quite different than sequences occurring at lower temperatures , for example , characteristic of those occurring in sun .
the direct study of stellar properties in ground - based laboratories has become more attractive , due to the availability of ribs , for example the study of @xmath2ne induced neutron pick - up reaction could reveal information about the exotic @xmath3o+@xmath4ne reaction happening in the cno cycle in burning stars .
study of the structure and reactions of unstable nuclei is therefore required to improve our understanding of the astrophysical origin of atomic nuclei and the evolution of stars and their ( sometimes explosive ) deaths
. experimental study of unstable nuclei has considerably advanced through the technique of using secondary radioactive beams .
the quantities measured in the study include various inclusive cross - sections , for example , reaction or interaction cross - sections , nucleon - removal cross - sections , coulomb breakup cross - sections and momentum distributions of a fragment .
these quantities have played an important role to reveal the nuclear structure of unstable nuclei , particularly halo structure near the drip line @xcite .
the total reaction cross - section ( @xmath5 ) is one of the most fundamental quantities characterizing the nuclear reactions and to probe for nuclear structure details .
recent studies using rib have demonstrated a large enhancement of @xmath6 induced by neutron rich nuclei , which has been interpreted as neutron halo @xcite ( such as @xmath7li , @xmath8be , etc . ) and neutron skin structure @xcite ( such as @xmath9he and @xmath10he ) .
the halo structure of @xmath7li seems to be consistent with all the experimental results including the enhancement of interaction cross - section @xmath11 , the enhancement of two - neutron removal cross - section ( @xmath12 ) and the narrow peak in the momentum distribution of fragmentation of @xmath13li . at present
the glauber model is a standard tool to calculate the cross - sections because it can account for a significant part of breakup effects which play an important role in the reaction of a weakly bound nucleus @xcite .
the glauber model , based on the independent individual nucleon - nucleon collisions in the overlap zone of the colliding nuclei @xcite , has been used extensively to explain the observed nuclear reaction cross - section for various systems at high energies @xcite .
this model requires the structure information , namely the density profiles , of the nuclei involved .
this information has to be provided by nuclear structure models , like the relativistic mean field ( rmf ) theory which is effectively used for this purpose recently @xcite .
the rmf formalism is well suited for the studies of exotic nuclei @xcite .
it takes into account the spin orbit interaction automatically , unlike to the non - relativistic case .
the parameters are fitted by taking into account the properties of few spherical nuclei .
the inclusion of rho - meson takes into account the proton - neutron asymmetry and gives an impression that the theory can be used to nuclei far away from the valley of @xmath14-stability . apart from these
, the advantage of the rmf model is the microscopic calculations of nuclear structure starting from the lagrangian with same parameters applicable for the whole nuclear chart and beyond .
the recent extension of this formalism with field theory motivated effective lagrangian approach ( e - rmf ) @xcite could extend the applicability of this model to neutron stars and infinite nuclear matter @xcite . hence this model provides us better predictability to explore the features of exotic nuclei .
the main objective of the present work is to study the nuclear reaction cross - section using rmf and e - rmf nuclear densities in conjunction with the glauber model .
the paper is presented as follows . in section
ii , we discuss in brief the formalism used in the present work . in section iii , we discuss our results for the ground state properties of few selected light mass nuclei and cross - sections of reactions involving them .
the summary and concluding remarks are given in section iv .
the details of the standard rmf formalism for finite nuclei can be found in refs .
@xcite . the recent extension of rmf formalism based upon the field theory motivated effective lagrangian approach , known as e - rmf , can be found in refs .
the energy density functional of the e - rmf model for finite nuclei @xcite is written as @xmath15 & & - \frac{i}{2 m } \beta \mbox{\boldmath$\alpha$}\!\cdot\ !
\left ( f_v % \mbox{\boldmath$\nabla$ } w + \frac{1}{2}f_\rho\tau_3 \mbox{\boldmath$\nabla$ } r + \lambda \mbox{\boldmath$\nabla$ } a \right ) + \frac{1}{2m^2}\left ( \beta_s + \beta_v \tau_3 \right ) \delta a \bigg\ } \varphi_\alpha \nonumber \\[3 mm ] & & \null + \left ( \frac{1}{2 } + \frac{\kappa_3}{3!}\frac{\phi}{m } + \frac{% \kappa_4}{4!}\frac{\phi^2}{m^2}\right ) \frac{m_{s}^2}{g_{s}^2 } \phi^2 - \frac{\zeta_0}{4 ! } \frac{1 } { g_{v}^2 } w^4 \nonumber \\[3 mm ] & & \null + \frac{1}{2g_{s}^2}\left ( 1 + \alpha_1\frac{\phi}{m}\right ) \left ( \mbox{\boldmath $ \nabla$}\phi\right)^2 - \frac{1}{2g_{v}^2}\left ( 1 + \alpha_2\frac{\phi}{m}\right ) \left ( \mbox{\boldmath $ \nabla$ } w \right)^2 \nonumber \\[3 mm ] & & \null - \frac{1}{2}\left(1 + \eta_1 \frac{\phi}{m } + \frac{\eta_2}{2 } \frac{\phi^2 } { m^2 } \right ) \frac{{m_{v}}^2}{{g_{v}}^2 } w^2 - \frac{1}{% 2g_\rho^2 } \left ( \mbox{\boldmath $ \nabla$ } r\right)^2 - \frac{1}{2 } \left ( 1 + \eta_\rho \frac{\phi}{m } \right ) \frac{m_\rho^2}{g_\rho^2 } r^2 \nonumber \\[3 mm ] & & \null - \frac{1}{2e^2}\left ( \mbox{\boldmath $ \nabla$ } a\right)^2 + \frac{1}{3g_\gamma g_{v}}a \delta w + \frac{1}{g_\gamma g_\rho}a \delta r , \label{eqfn1}\end{aligned}\ ] ] where the index @xmath16 runs over all occupied states @xmath17 of the positive energy spectrum , @xmath18 , @xmath19 , @xmath20 and @xmath21 . the terms with @xmath22 , @xmath23 , @xmath24 and @xmath25 take care of effects related with the electromagnetic structure of the pion and the nucleon ( see ref .
@xcite ) . specifically , the constant @xmath22 concerns the coupling of the photon to the pions and the nucleons through the exchange of neutral vector mesons .
the experimental value is @xmath26 .
the constant @xmath23 is needed to reproduce the magnetic moments of the nucleons .
it is defined by @xmath27 with @xmath28 and @xmath29 the anomalous magnetic moments of the proton and the neutron , respectively .
the terms with @xmath24 and @xmath25 contribute to the charge radii of the nucleon @xcite .
the energy density contains tensor couplings and scalar - vector and vector - vector meson interactions in addition to the standard scalar self interactions @xmath30 and @xmath31
. the e - rmf formalism can be interpreted as a covariant formulation of density functional theory as it contains all the higher order terms in the lagrangian by expanding it in powers of the meson fields .
the terms in the lagrangian are kept finite by adjusting the parameters .
further insight into the concepts of the e - rmf model can be obtained from ref .
it is worth mentioning that the standard rmf lagrangian is obtained by ignoring the vector - vector and scalar - vector cross interactions , and does not need any separate discussion .
the field equations and numerical details can be obtained in refs .
the set of coupled equations is solved numerically by a self - consistent iteration method .
the baryon , scalar , isovector , proton and tensor densities are @xmath32 \rho_s(r ) & = & \sum_\alpha \varphi_\alpha^\dagger(r ) \beta \varphi_\alpha(r ) \ , , \label{eqfn7 } \\[3 mm ] \rho_3 ( r ) & = & \sum_\alpha \varphi_\alpha^\dagger(r ) \tau_3 \varphi_\alpha(r ) \ , , \label{eqfn8 } \\[3 mm ] \rho_{\rm p}(r ) & = & \sum_\alpha \varphi_\alpha^\dagger(r ) \left ( \frac{1 + \tau_3}{2 } \right ) \varphi_\alpha(r ) \ , , \label{eqfn9 } \\[3 mm ] \rho_{\rm t}(r ) & = & \sum_\alpha \frac{i}{m } \mbox{\boldmath$\nabla$ } \!\cdot\ ! \left
[ \varphi_\alpha^\dagger(r ) \beta \mbox{\boldmath$\alpha$ } \varphi_\alpha(r ) \right ] \ , , \label{eqfn10 } \\[3 mm ] \rho_{\rm t,3}(r ) & = & \sum_\alpha \frac{i}{m } \mbox{\boldmath$\nabla$ } \!\cdot\ ! \left [ \varphi_\alpha^\dagger(r ) \beta \mbox{\boldmath$\alpha$ } \tau_3 \varphi_\alpha(r ) \right ] \ , .
\label{eqfn11}\end{aligned}\ ] ] for the calculation of ground state properties of finite nuclei , we refer the readers to refs .
@xcite .
the theoretical formalism to calculate the nuclear reaction cross - section using glauber approach has been given by r. j. glauber @xcite .
for sake of completeness , here , we briefly outline the steps of derivations following the notation of ref .
@xcite .
the standard glauber form for the reaction cross - section at high energies , is expressed @xcite as : @xmath33db \;,\ ] ] where @xmath34 , the transparency function , is the probability that at an impact parameter @xmath35 the projectile pass through the target without interacting .
this function @xmath34 is calculated in the overlap region between the projectile and target where the interactions are assumed to result from single nucleon - nucleon collision and is given by @xmath36 \;.\ ] ] here , the summation indices @xmath37 , @xmath38 run over proton and neutron and subscript @xmath39 and @xmath40 refers to projectile and target respectively .
@xmath41 is the experimental nucleon - nucleon reaction cross - section which varies with respect to energy .
the @xmath42-integrated densities @xmath43 are defined as @xmath44 with @xmath45 .
the parameters @xmath46 , @xmath47 , and @xmath14 usually depend on either the proton - proton ( neutron - neutron ) or proton - neutron case , but we have used some appropriate average values in the present calculations @xcite .
the argument of @xmath34 in eq .
( 10 ) is @xmath48 , which stands for the impact parameter between @xmath49 and @xmath50 nucleons .
the glauber model agrees very well with the experimental data at high energies .
however , this model fails to describe , reasonably , the collisions induced at relatively low energies . in such case the present version of glauber model is modified in order to take care of finite range effects in profile function and coulomb modified trajectories .
thus for finite range approximations , the transparency function is given by @xmath51 \overline{\rho}_{pi}\left ( \vec{t}\right ) \overline{\rho } _ { tj}\left ( \vec{s}\right ) d\vec{s}d% \vec{t}\right]\;.\ ] ] here the profile function @xmath52 is given by @xmath53 where @xmath54 , @xmath55 is the impact parameter and @xmath56 and @xmath57 are just the dummy variables for integration over the @xmath42-integrated target and projectile densities .
the differential elastic cross - section by the ratio to the rutherford cross - section is given by , @xmath58 @xmath59 and @xmath60 are the elastic and coulomb ( elastic ) scattering amplitudes , respectively .
the elastic scattering amplitude @xmath59 is written as @xmath61t(b)\right\}\ ] ] with the coulomb elastic scattering amplitude @xmath60 given by @xmath62 here @xmath63/@xmath64 is the sommerfield parameter , @xmath65 is the incident velocity , and @xmath66 with @xmath67 being screening radius @xcite .
the elastic differential cross - section does not depend on the screening radius @xmath67 .
there exist a number of parameter sets for solving the standard rmf as well as e - rmf lagrangians . in our previous paper @xcite
we have calculated reaction cross - sections with densities obtained from various interactions . in the present work ,
we employ the sig - om set for rmf which has been recently used by haidari et al.@xcite and g2 @xcite for e - rmf calculations .
we have presented the calculated binding energy using rmf and e - rmf theories in table i. the experimental data taken from ref.@xcite have also been given for comparison .
it is evident from table ii that both the calculated binding energies are similar and slightly overestimate in comparison with experimental binding energies .
however , these differences with experimental values are small and may be attributed to the fact that for light mass region of the periodic table , mean field is not saturated . to get a qualitative estimation of the binding energy , the rmf as well as e - rmf results are trust worthy and
can be used for further calculations in this region .
.the ground state properties of the nuclei involved in reaction cross - section study .
experimental binding energies have been taken from @xcite .
the sig - om and g2 sets are chosen for rmf and e - rmf parametrizations , respectively . [ cols="^,^,^,^,^,^,^ " , ] the calculation of profile function @xmath68 requires some phenomenological parameters related to nucleon - nucleon cross - section .
these parameters @xmath69 @xmath70 and @xmath71 at various energies are taken from refs . @xcite and tabulated in table iii for sake of completness . here
@xmath46 represents the total cross - section of n - n@xmath72collision , @xmath73 is the ratio of the real to the imaginary part of the forward nucleon - nucleon scattering amplitude and @xmath74 is basically the slope parameter which determines the fall - of the angular distribution of the n - n elastic scattering .
though these parameters in general depend on the isospin of the nucleons ( pp , nn , pn ) , appropriate average values are taken by interpolating a given set . ) for @xmath75c+@xmath75c system .
experimental data are also shown in the figure with error bars @xcite . ]
the total reaction cross - sections at different incident energies have been calculated for various systems and compared with the experimental results @xcite , if available .
the reaction cross - section with stable and unstable beams using stable target such as @xmath75c are within experimental reach and are being studied extensively @xcite . as a first application to nuclear reaction studies
, we have calculated the total reaction cross - section for @xmath75c+@xmath75c system and compared with the experimental results @xcite .
it can be seen from fig .
2 that the agreement of @xmath5 using e - rmf nuclear densities is excellent for almost all incident energies ( energy / nucleon ) , particularly at higher values .
the calculated e - rmf cross - section is slightly off at lower energies which is easily understandable as it is well studied that the glauber model works better at higher incident energies in comparison to the lower incident energies .
this disagreement is due to the significant role played by the repulsive coulomb potential whose effects are obvious in the low - energy range .
such a coulomb effect breaks the characteristic glauber assumption that the projectile travels along straight - line trajectories .
however , our results using e - rmf nuclear densities successfully produce the qualitative trend of experimental results . here
it is interesting to see that the calculations using rmf densities also matches reasonably with the experimental values but the results at lower incident energies are quite off . ) at 800 mev / nucleon for li isotopes as projectile and @xmath75c as target .
experimental data with error bars @xcite are also shown . ] in figs . 3 and 4 , we have shown the comparison of experimental @xcite and calculated @xmath5 for @xmath76li+@xmath75c , @xmath77be+@xmath75c and @xmath78b+@xmath75c systems at fixed incident energy ( 800 mev / nucleon ) .
the trends of the calculations for the different projectiles [ figs
. 3 and 4 ] are basically the same .
li+@xmath7li , @xmath7be+@xmath7be , @xmath79be+@xmath79be and @xmath80b+@xmath80b with rmf and e - rmf densities as input for various incident energies . ]
so far we have discussed the reactions involving stable and unstable beams on stable target .
to measure the nuclear reaction cross - section with unstable beam and unstable target is one of the major challenges for experimental nuclear physicists .
such measurements would be helpful for the better understanding of many astrophysical phenomena as well as in determining the energy and matter evolution at stellar sites .
as more extensive observational data is gathered from earth and space observatories , an ever - greater demand is placed on our knowledge of the basic physical processes that probe astrophysical phenomena .
considerable efforts at institute of modern physics , cas ( china ) and at riken ( japan ) are underway to look for rib+rib cross - section using rib as internal target with rib projectile .
although such measurements are not feasible with presently available experimental techniques , yet the fast advancement in rib techniques may provide us this facility in next few years or so .
such experiments will be decisive in getting precise information about the structure of halo nuclei .
in this view , we have presented the calculated @xmath5 for few rib+rib systems , namely for @xmath7li+@xmath7li , @xmath7be+@xmath7be , @xmath79be+@xmath79be and @xmath80b+@xmath80b in fig . 5 , which may serve as a guiding tool for the experiments under planning .
we see from fig .
5 that rmf and e - rmf predict almost similar trend for the variation of cross - section with respect to energy .
a further inspection of the figure reveals that the e - rmf results are marginally higher than the rmf results .
c+@xmath75c system : ( a ) at 30 mev / nucleon of incident energy and ( b ) at 85 mev / nucleon of incident energy .
the experimental data are taken from @xcite . ]
results for elastic differential cross - section @xmath81 for the @xmath75c+@xmath75c system have been shown in fig . 6 at 30 mev / nucleon as well as 85 mev / nucleon of incident energies .
we see that the elastic scattering angular distributions for @xmath75c+@xmath75c , are better reproduced using e - rmf ( g2 set ) nuclear densities than rmf ( sig - om ) one while demanding conformity with experimental data @xcite .
this example clearly shows the importance of nuclear densities and highlights the sensitivity of the experimental differential cross - section to details of nuclear structure .
results for the elastic scattering angular distributions for rib projectiles have been shown in fig . 7 . from the above study
, it is interesting to observe that the two relativistic approaches give slightly different cross - sections which could be attributed to the different results obtained for ground state properties .
hence the details of structure information have to be considered crucial as they are well reflected in the reaction cross - sections . at low energy region ( 30 mev / nucleon ) , both differential scattering cross - section ( sig - om and g2 ) are similar to each other as shown in fig .
the experimental trend is reproduced well using both the densities as input in the evaluation of differential cross - section . however , if one analyse the data at 85 mev / nucleon as shown in fig .
6(b ) , the values of @xmath82 obtained with both rmf and e - rmf approaches agree well with the experiment , both qualitatively and quantitatively .
similar results of differential cross - section for exotic nuclei which are predicted as likely halo candidates , namely @xmath7li , @xmath7be and @xmath79be @xcite , with @xmath75c as target nucleus is shown in fig . 7 taking incident energies as 30 mev / nucleon and 85 mev / nucleon . in all these systems
i.e. @xmath7li+@xmath75c , @xmath7be+@xmath75c , @xmath79be+@xmath75c , the differential cross - sections obtained with both , the rmf and the e - rmf densities are almost similar .
in summary , we have used the glauber model to calculate nuclear reaction cross - section with densities obtained from rmf and e - rmf calculations using sig - om and g2 set of parameters respectively .
we have seen that the calculation of total reaction cross - section can be performed well with the glauber model using rmf and e - rmf nuclear densities as the main ingredient .
the good quality of results shows that the nuclear reaction cross - section predictions from the glauber model calculations using rmf and e - rmf nuclear densities will be helpful in more stringent analysis of the high energy reactions involving the nuclei on either side of the valley of @xmath83-stability .
the comparison with the results of double folding potential analysis using the same rmf and e - rmf nuclear densities would be enriching our knowledge more in the low energy region in this regard .
while analysing the differential cross - section , we found that the e - rmf density suited well to reproduce the experimental data .
the rmf basically fails for larger scattering angle with higher incident energy .
this study clearly shows the importance of extending rmf to e - rmf formalism to reduce the central density of the target .
thus , from these calculations it is clear that the e - rmf theory not only describe the ground state and nuclear matter properties of the nuclei successfully , but also explains the nuclear reaction data quite well .
overall , these calculations give an excellent account for the existing experimental data for ground state properties namely , nuclear radii and binding energy as well as for nuclear reaction cross - section results .
further , it is hoped that such precise studies for cross - section calculations of exotic nuclei may also be very crucial in view of upcoming radioactive ion beam facilities .
99 a. ozawa , t. suzuki and i. tanihata , nucl .
* a693 * , 32 ( 2001 ) .
d. t. khoa , h. sy than , t. h. nam , m. grasso and n. v. giai phys . rev . *
c69 * , 044605 ( 2004 ) .
f. m. nunes , n. c. summers , a. m. moro and a. m. mukhamedzhanov , _ proceedings of nuclei at the limits _ , anl 26 - 30 july 2004 ; arxiv : nucl - th/0505046 . i. tanihata , j. phys .
* g22 * , 157 ( 1996 ) . i. tanihata , h. hamagaki , o. hashimoto , y. shida , n. yoshikawa , k. sugimoto , o. yamakawa , t. kobayashi and n. takahashi , phys .
. lett . * 55 * , 2676 ( 1985 ) .
t. kobayashi , o. yamakawa , k. omata , k. sugimoto , t. shimoda , n. takahashi and i. tanihata , phys .
lett . * 60 * , 2599 ( 1988 ) .
p. g. hansen and b. jonson , europhys .
lett . * 4 * , 409 ( 1989 ) .
w. mittig et al . , phys .
* 59 * , 1889 ( 1987 ) .
k. varga , s. c. pieper , y. suzuki and r. b. wiringa , phys . rev . *
c66 * , 034611 ( 2002 ) . b. abu - ibrahim , y. ogawa , y. suzuki and i. tanihata , comp .
. comns . *
151 * , 369 ( 2003 ) .
r. j. glauber , _ lectures on theoretical physics , edited by _ w. e. brittin and l. c. dunham ( interscience , new york , 1959 ) , vol.1 , p.315 .
g. f. bertsch , b. a. brown and h. sagawa , phys .
rev . * c39 * , 1154 ( 1989 ) . m. y. h. farag , eur .
j. * c12 * , 1112 ( 2001 ) .
a. bhagwat and y. k. gambhir , phys . rev . * c69 * , 014315 ( 2004 ) ; a. bhagwat , y. k. gambhir and s. h. patil , eur . phys . j. * a8 * , 511 ( 2000 ) ; j. phys .
* g27 * , b1 ( 2001 ) . b. k. sharma , s. k. patra , raj k. gupta , a. shukla , p. arumugam , p. d. stevenson and walter greiner , j. phys .
g * 32 * , 2089 ( 2006 ) .
s. k. patra and c. r. praharaj , phys . rev .
* c44 * ( 1991 ) 2552 ; y. k. gambhir , p. ring and a. thimet , ann .
( n.y . ) * 198 * , 132 ( 1990 ) .
b. d. serot and j. d. walecka , int .
e6 * , 515 ( 1997 ) .
r. j. furnstahl , b. d. serot and h. b. tang , nucl .
* a598 * , 539 ( 1996 ) ; _ ibid _ * a615 * , 441 ( 1997 ) .
p. arumugam , b. k. sharma , p. k. sahu , s. k. patra , t. sil , m. centelles and x. vias , phys .
b601 * , 51 ( 2004 ) .
m. del estal , m. centelles , x. vias and s. k. patra , phys .
rev * c63 * , 024314 ( 2001 ) .
m. del estal , m. centelles , x. vias and s. k. patra , phys . rev . *
c63 * , 044321 ( 2001 ) .
p. j. karol , phys . rev . * c11 * , 1203 ( 1975 ) . m. m. haidari and m. m. sharma , nucl - th/0702001 .
g. audi , a. h. wapstra and c. thibault , nucl .
* a729 * , 337 ( 2003 ) . i. angeli , at .
data and nucl .
data tables * 87 * , 185 ( 2004 ) .
y. ogawa , k. yabana and y. suzuki , nucl .
a543 * , 722 ( 1992 ) .
j. chauvin , d. lebrun and m. buenerd , phys . rev . *
c28 * , 1970 ( 1983 ) ; m. buenerd , a. lounis , j. chauvin , d. lebrun , p. martin , g. duhamel , j.c .
gondrand and p.d .
saintignon , nucl . phys . *
a424 * , 313 ( 1984 ) ; j.y .
hostachy et al , nucl
a490 * , 441 ( 1988 ) .
yong - xu yang , qing - run li and wei - qin zhao , j. phys .
* g28 * , 2561 ( 2002 ) .
c * 18 * , 2273 ( 1978 ) .
b. k. sharma , p. arumugam , a. shukla and s. k. patra _
proceedings of nuclear structure physics at extremes _ , h.p .
university , shimla , india ( 2005 ) . | we study nuclear reaction cross - sections for stable and unstable projectiles and targets within glauber model , using densities obtained from various relativistic mean field formalisms .
the calculated cross - sections are compared with the experimental data in some specific cases .
we also evaluate the differential scattering cross - sections at several incident energies , and observe that the results found from various densities are similar at smaller scattering angles , whereas a systematic deviation is noticed at large angles . in general , these results agree fairly well with the experimental data . |
the detection of cosmic neutrinos poses great challenges , but has attractive features as well .
as opposed to other particles and to electromagnetic radiation , neutrinos can travel through huge amounts of matter without being absorbed .
they can thus lead to the discovery of distant production sites hidden from observation with conventional means .
furthermore , they do not get deflected by magnetic fields during their journey and so their path points straight back to their point of origin .
some known extra - terrestrial sources that are believed to accelerate particles to very high energies are prime candidates for neutrino production . in that scenario , neutrinos are produced as a result of the interaction of accelerated charged particles with surrounding matter or photons , producing pions , which themselves decay into muons and neutrinos . the actual observation of neutrinos would thus shed light on the hadronic nature of the accelerating process .
the energy feeding that process is of gravitational origin and possible candidates are active galactic nuclei ( agns),supernova remnants ( snrs ) or gamma ray bursters ( grbs ) .
models predict a hard energy spectrum @xmath0 , which at high enough energies has a flux detectable above the background of the much steeper atmospheric neutrino flux .
another possible scheme for neutrino production is one where neutrinos come from the decay of massive relic particles @xcite .
high energy neutrinos interact via charged and neutral currents with ordinary matter and produce charged particles in the process .
these particles in turn send out cherenkov light when they travel through a medium with a refractive index greater than one . in order to detect the weakly interacting neutrino , large volumes of matter
must be instrumented with light sensors . in the cases of amanda and icecube
, the ice sheet at the south pole serves as the cherenkov medium .
notice , that the clarity of this ice allows for a sparse instrumentation , reducing the cost of the detector for a given geometry .
muon neutrinos produce long range secondary muons ( 1 km at 200 gev ) sending out light along their path through the detector .
the direction of the parent neutrino can be directly inferred from that of the muon since the mean angle between the two particles is proportional to @xmath1 , with a value of of @xmath2 at 1 tev .
electron and tau neutrinos will produce cascades , reconstructed as point sources using the cherenkov light , but from which some directional information can be extracted as well . the main background to these events comes from atmospheric muons produced in the atmosphere above the detector , so only events reconstructed as up - going are considered as signal , since atmospheric muons from this direction get absorbed by the earth . at pev energies ,
the earth becomes opaque to neutrinos , including the background made by those produced in the atmosphere and one has to look for events coming from close to the horizon and above . at those energies
, the @xmath3 has a characteristic `` double bang '' signature : after the first cascade , produced when the @xmath3 interacts with the ice , the resulting tau particle decays after hundreds of meters , producing a second cascade .
thus the topologies of the three neutrino flavors become quite distinguishable @xcite . _ , height=377 ]
amanda is an array of light sensors installed deep in the 2800 meter thick south pole ice sheet . over the years
, 19 holes have been melted with a hot - water drill and optical modules ( oms ) were frozen - in at depths between 1500 and 2000 m below the surface .
the oms are arranged string - wise , with a distance between them of 10 to 20 meters , depending on their location .
each module consists of an 8-inch hamamatsu photomultiplier , enclosed in a thick glass sphere designed to stand the high pressure of the ice .
there are 677 oms arranged on 19 strings in total , deployed between 1996 and 2000 .
the signals produced by the photomultipliers travel along cables , connected to a counting house at the surface . after triggering , time- and charge - information
is extracted from these pulses and stored for further analysis .
spase - ii is an air - shower array located at the surface of the ice and operated in coincidence with amanda .
muons passing through both detectors have been used to verify the performance of the detector and for position calibration of its oms . _ , width=302
we present here only a selection of results .
several topics are not included here , such as : grb searches , detection and search for gravitational collapse supernovae , search for exotic particle ( q - balls , nuclearites , monopoles etc . ) cosmic rays impinging on the atmosphere produce up - going neutrinos which are indistinguishable from extra - terrestrial neutrinos .
however , the energy dependence of the flux of the first goes as @xmath4 above 1 tev , in contrast to the flux of the latter which is expected to vary as @xmath5 and thus decrease much faster with energy .
the energy spectrum of the year 2000 data with 570 neutrino events was extracted using neural net techniques and then unfolded @xcite . the ( preliminary ) result
is shown in figure [ unfolded ] , together with the frejus measurements @xcite at lower energies .
amanda has performed the first measurements going up to 300 tev .
both the frejus and the amanda results lie within the horizontal and vertical fluxes parametrized by volkova ( @xmath6 gev ) and honda ( @xmath7 gev ) @xcite .
the maximum additional flux with an @xmath5 spectrum ( at a @xmath8 confidence level ) in the range 100 - 300 tev is also marked ( `` limit '' ) on the figure .
3329 @xmath9 events , collected during 2000 - 2003 , were used in a search for point sources @xcite .
the sky map of these events is shown in equatorial coordinates in figure [ skymap ] .
a search for a localized source in that map gives figure [ excessmap ] , representing the significance of the excess in each direction .
the highest value found in any directions in the whole map corresponds to 3.4@xmath10 .
background simulations tell us that the probability of observing an excess @xmath11 , taking into account trial factors , is @xmath12 . a selective search for an excess from a list of 33 pre - selected sources does not lead to the discovery of any significant excess either . _ , width=377 ] _ , width=377 ] _ , width=377 ] even if the fluxes from individual objects are too low to be detected separately from each other , the combined flux of many such sources may still be used to offer evidence of the production of extra - terrestrial neutrinos . if that diffuse flux exists
, it is expected to have a harder spectrum than that of atmospheric neutrinos and will be exceeding it above a high enough energy .
figure [ diffuse ] shows the limits set by amanda and other detectors on diffuse neutrinos : the limit labeled ( 1 ) is the one set by macro @xcite , ( 2 ) is from the 10-string stage of amanda in 1997 using @xmath9 @xcite , ( 3 ) comes from the same data as ( 2 ) but the analysis combines all flavors of uhe @xmath13s@xcite . ( 4 ) is the amanda 2000 all - flavor cascade limit@xcite , ( 5 ) is the baikal limit from their 1998 - 2000 cascade analysis@xcite . lines ( 6),(7 ) and ( 8) are all preliminary results from 2000 data .
the line for the limit ( 6 ) comes the @xmath9 analysis@xcite , ( 7 ) the sensitivity for all the flavors of uhe muons@xcite and ( 8) the 2000 - 2003 @xmath9 sensitivity@xcite .
we assume full mixing , so that in analysis ( 1 ) , ( 2 ) , ( 6 ) and ( 8) , the limits for @xmath9 are shown multiplied by 3 .
the line marked `` agn core ( ss ) '' is the expected flux of diffuse neutrinos according to the model described in @xcite and the one marked `` agn jet ( mpr ) '' refers to the model in @xcite .
both models have been scaled by 1.5 to account for oscillations .
the waxman - bahcall upper limit ( `` wb'')@xcite is also indicated on the figure . c.l . on the muon flux from neutrinos produced in neutralino annihilations in the center of the earth ( left ) and the center of the sun ( right ) .
the dots show models disfavored by direct search results from cdms and the crosses correspond to predictions from different mssm models .
[ wimp ] _ , title="fig:",width=238 ] c.l . on the muon flux from neutrinos produced in neutralino annihilations in the center of the earth ( left ) and the center of the sun ( right ) .
the dots show models disfavored by direct search results from cdms and the crosses correspond to predictions from different mssm models .
[ wimp ] _ , title="fig:",width=226 ] a prime candidate wimp particle to make up for the cold dark matter of the universe is the lightest supersymmetric particle , the neutralino .
an indirect observation of these particles can be made using amanda in the following scenario .
neutralinos lose their energy in collisions with matter and get gravitationally trapped inside massive objects , where they annihilate with each other , emitting neutrinos in that process .
the closest and in that sense most interesting sources of accumulation are the center of the earth and the sun . the limits at @xmath8 c.l .
using 1997 - 1999 data for the earth @xcite and using 2000 data for the sun@xcite , are shown on figure [ wimp ] , together with the limits for other experiments and projected results for icecube .
_ , width=264 ] _ , width=377 ] icecube follows the successful concept demonstrated by amanda , but on a larger scale and with more advanced hardware . with icecube
, we enter the era of cubic kilometer sized neutrino telescopes . at completion , it will consist of 80 strings arranged on a regular hexagon and with 60 digital optical modules ( doms ) each ( see figure [ icecube - detector ] )
. the instrumented region will be between 1450 m and 2450 m below the surface and the hexagonal pattern will span over a @xmath14 area .
the distance between doms within a string is 17 m and the distance between the strings is 125 m , with the existing amanda detector embedded inside the icecube volume .
strings will be added each year until completion but the detector is taking data since the start . with it
s large size , icecube should observe @xmath13 events with energies up to @xmath15 and collect many contained events .
an air shower array , icetop , is deployed at the surface of the ice above icecube .
it consists of 160 cherenkov tanks located close to the 80 holes .
the tanks are grouped in pairs ( stations ) near a hole and are filled with water which is then allowed to freeze around two doms inside .
icetop will serve several purposes : cosmic ray studies up to energies up to @xmath16 ) , search for a possible second `` knee '' for heavy particles like iron , calibration of the icecube doms using coincident atmospheric muons , vetoing of highly energetic background events originating in the atmosphere .
the dom ( figure [ dom ] ) is the main component of the detector .
it contains a 10-inch hamamatsu photomultiplier tube ( pmt ) , enclosed in a glass vessel .
an integrated circuit ( analog transient waveform digitizer , or atwd ) records the pmt pulse with three different gains ( x1/4 , x2 and x16 ) .
the atwd records 128 samples with a rate of 3.3 ns / sample .
the same pulse is also digitized by a 40mhz fadc . the linear dynamic range is 400 photo - electrons ( p.e ) in 15 ns and the integrated dynamic range is more than 5000 p.e . in @xmath17 .
neighboring doms are connected to each other and the atwd is read out after a local triggering condition is met .
each waveform is time - stamped locally .
the main board contains an fpga that can be re - programmed from a computer at the surface .
the doms communicate with the counting house with twisted pair copper cables and receive their power through the same cables .
each dom has its own high - voltage generator .
since the pulses are sent digitally , the pmt amplification can be kept low at @xmath18 .
this benefits both the dynamic range and the aging of the pmts .
+ the doms are deployed in holes melted with a hot - water drill .
a new 5 mw drill , the enhanced hot water drill ( ehwd ) was developed for icecube and put into operation in january 2005 .
this new construction is designed to drill 2500 m deep holes with a 60 cm diameter in less than 40 h , much faster than the amanda drill .
the actual deployment of the doms takes less than 12 hours . with the ehwd
, icecube will be able to deploy 14 strings / year during the november to mid - february period when the south pole station is open for construction work .
simulations of the detector have been made @xcite to predict its capabilities .
the effective area is about @xmath19 for 1 tev muons the angular resolution is expected to be @xmath20 at @xmath21 . .
[ noise]_,width=302 ] _ , width=453 ] a first string , labelled 21 was deployed in february 2005 , together with four icetop stations , totalling 76 doms together@xcite .
since then , all doms of string 21 have been taking data and behaving according to expectation .
in particular , the time calibration of quartz oscillator in each dom has been studied .
these oscillators are synchronized to a master gps clock every 3.5 seconds , by sending down pulses from the surface to the doms .
the calibration procedure yields a time precision of 3 ns .
another measurement is the dark noise rate of the doms , which is around 700 hz . with after - pulse suppression ( implemented with an artificial dead - time of @xmath22 ) , this rate goes down to about 350 hz , see figure [ noise ] . in the same way as amanda
, icecube will be able to detect the burst of low energy neutrinos that accompany a supernova explosion by measuring an excess rate above the dark noise in all its sensors .
a low dark noise improves the sensitivity for such a search .
string 21 has also been used to reconstruct atmospheric muons .
the resulting multiplicities and zenith angles agree with simulations ( see figure [ comparison ] for the comparison ) . in january 2006 , the installation of eight more strings and 12 additional icetop stations was completed .
an event passing through icetop and the nine icecube strings is shown in figure [ event ] .
these new strings total 540 doms ( and icetop 64 doms ) which cover a larger volume than amanda s 677 modules .
@xmath23 of the doms have survived the deployment and re - freezing period and icecube took data for several months at an average rate of 79hz ( with a trigger condition of 8 doms hit in a window of @xmath24 ) .
we have presented a selection of the amanda results from data taken over the past ten years .
the transition from amanda to the much larger icecube detector is under way .
the performance of the first icecube string is as expected .
the experience gained in deploying eight additionalstrings with new equipment bodes well for the higher rate of deployment in coming years . _ , width=113
we acknowledge the support from the following agencies : national science foundation - office of polar program , national science foundation - physics division , university of wisconsin alumni research foundation , department of energy , and national energy research scientific computing center ( supported by the office of energy research of the department of energy ) , the nsf - supported teragrid system at the san diego supercomputer center ( sdsc ) , and the national center for supercomputing applications ( ncsa ) ; swedish research council , swedish polar research secretariat , and knut and alice wallenberg foundation , sweden ; german ministry for education and research , deutsche forschungsgemeinschaft ( dfg ) , germany ; fund for scientific research ( fnrs - fwo ) , flanders institute to encourage scientific and technological research in industry ( iwt ) , belgian federal office for scientific , technical and cultural affairs ( ostc ) ; the netherlands organisation for scientific research ( nwo ) ; m. ribordy acknowledges the support of the snf ( switzerland ) ; j. d. zornoza acknowledges the marie curie oif program ( contract 007921 ) .
99 j.g . learned and k. mannheim , ann .
nucl . part .
2000.50:679 f.w .
stecker and m.h .
salamon , space .
rev.*75 * ( 1996 ) 341 k. mannheim , r. protheroe and j. rachen , phys . rev.*d63 * ( 2001 ) 023003 e. waxman and j. bahcall , phys . rev.*d59 * ( 1999 ) 023002 a. achterberg _
et al . _ , accepted for publication in astropart .
m. ackermann _
et al . _ , astropart
* 24 * ( 2006 ) 459 j. ahrens _ et al .
_ , astropart .
phys . * 20 * ( 2004 ) 507 a. achtenberg et al . , astro - ph/0604450 a. achterberg@xmath26 , m. ackermann@xmath27 , j. adams@xmath28 , j. ahrens@xmath29 , k. andeen@xmath30 , d. w. atlee@xmath31 , j. n. bahcall@xmath32 , x. bai@xmath33 , b. baret@xmath34 , m. bartelt@xmath35 , s. w. barwick@xmath36 , r. bay@xmath37 , k. beattie@xmath38 , t. becka@xmath29 , j. k. becker@xmath35 , k .- h .
becker@xmath39 , p. berghaus@xmath40 , d. berley@xmath41 , e. bernardini@xmath27 , d. bertrand@xmath40 , d. z. besson@xmath42 , e. blaufuss@xmath41 , d. j. boersma@xmath30 , c. bohm@xmath43 , j. bolmont@xmath27 , s. bser@xmath27 , o. botner@xmath44 , a. bouchta@xmath44 , j. braun@xmath30 , c. burgess@xmath43 , t. burgess@xmath43 , t. castermans@xmath45 , d. chirkin@xmath38 , j. clem@xmath33 , d. f. cowen@xmath46 , m. v. dagostino@xmath37 , a. davour@xmath44 , c. t. day@xmath38 , c. de clercq@xmath34 , l. demirrs@xmath33 , f. descamps@xmath47 , p. desiati@xmath30 , t. deyoung@xmath31 , j. c. diaz - velez@xmath30 , j. dreyer@xmath35 , m. r. duvoort@xmath26 , w. r. edwards@xmath38 , r. ehrlich@xmath41 , j. eisch@xmath48 , r. w. ellsworth@xmath41 , p. a. evenson@xmath33 , o. fadiran@xmath49 , a. r. fazely@xmath50 , t. feser@xmath29 , k. filimonov@xmath37 , b. d. fox@xmath31 , t. k. gaisser@xmath33 , j. gallagher@xmath51 , r. ganugapati@xmath30 , h. geenen@xmath39 , l. gerhardt@xmath36 , a. goldschmidt@xmath38 , j. a. goodman@xmath41 , r. gozzini@xmath29 , s. grullon@xmath30 , a. gro@xmath52 , r. m. gunasingha@xmath50 , m. gurtner@xmath39 , a. hallgren@xmath44 , f. halzen@xmath30 , k. han@xmath28 , k. hanson@xmath30 , d. hardtke@xmath37 , r. hardtke@xmath48 , t. harenberg@xmath39 , j. e. hart@xmath31 , t. hauschildt@xmath33 , d. hays@xmath38 , j. heise@xmath26 , k. helbing@xmath39 , m. hellwig@xmath29 , p. herquet@xmath45 , g. c. hill@xmath30 , j. hodges@xmath30 , k. d. hoffman@xmath41 , b. hommez@xmath47 , k. hoshina@xmath30 , d. hubert@xmath34 , b. hughey@xmath30 , p. o. hulth@xmath43 , k. hultqvist@xmath43 , s. hundertmark@xmath43 , j .- p .
hl@xmath39 , a. ishihara@xmath30 , j. jacobsen@xmath38 , g. s. japaridze@xmath49 , a. jones@xmath38 , j. m. joseph@xmath38 , k .- h .
kampert@xmath39 , a. karle@xmath30 , h. kawai@xmath53 , j. l. kelley@xmath30 , m. kestel@xmath31 , n. kitamura@xmath30 , s. r. klein@xmath38 , s. klepser@xmath27 , g. kohnen@xmath45 , h. kolanoski@xmath54 , l. kpke@xmath29 , m. krasberg@xmath30 , k. kuehn@xmath36 , h. landsman@xmath30 , h. leich@xmath27 , i. liubarsky@xmath55 , j. lundberg@xmath44 , j. madsen@xmath48 , k. mase@xmath53 , h. s. matis@xmath38 , t. mccauley@xmath38 , c. p. mcparland@xmath38 , a. meli@xmath35 , t. messarius@xmath35 , p. mszros@xmath46 , h. miyamoto@xmath53 , a. mokhtarani@xmath38 , t. montaruli@xmath56 , a. morey@xmath37 , r. morse@xmath30 , s. m. movit@xmath57 , k. mnich@xmath35 , r. nahnhauer@xmath27 , j. w. nam@xmath36 , p. nieen@xmath33 , d. r. nygren@xmath38 , h. gelman@xmath30 , ph .
olbrechts@xmath34 , a. olivas@xmath41 , s. patton@xmath38 , c. pea - garay@xmath58 , c. prez de los heros@xmath44 , a. piegsa@xmath29 , d. pieloth@xmath27 , a. c. pohl@xmath59 , r. porrata@xmath37 , j. pretz@xmath41 , p. b. price@xmath37 , g. t. przybylski@xmath38 , k. rawlins@xmath60 , s. razzaque@xmath46 , f. refflinghaus@xmath35 , e. resconi@xmath52 , w. rhode@xmath35 , m. ribordy@xmath45 , a. rizzo@xmath34 , s. robbins@xmath39 , c. rott@xmath31 , d. rutledge@xmath31 , d. ryckbosch@xmath47 , h .-
sander@xmath29 , s. sarkar@xmath61 , s. schlenstedt@xmath27 , d. seckel@xmath33 , s. h. seo@xmath31 , s. seunarine@xmath28 , a. silvestri@xmath36 , a. j. smith@xmath41 , m. solarz@xmath37 , c. song@xmath30 , j. e. sopher@xmath38 , g. m. spiczak@xmath48 , c. spiering@xmath27 , m. stamatikos@xmath30 , t. stanev@xmath33 , p. steffen@xmath27 , t. stezelberger@xmath38 , r. g. stokstad@xmath38 , m. c. stoufer@xmath38 , s. stoyanov@xmath33 , e. a. strahler@xmath30 , k .- h .
sulanke@xmath27 , g. w. sullivan@xmath41 , t. j. sumner@xmath55 , i. taboada@xmath37 , o. tarasova@xmath27 , a. tepe@xmath39 , l. thollander@xmath43 , s. tilav@xmath33 , p. a. toale@xmath31 , d. turan@xmath41 , n. van eijndhoven@xmath26 , j. vandenbroucke@xmath37 , a. van overloop@xmath47 , b. voigt@xmath27 , w. wagner@xmath35 , c. walck@xmath43 , h. waldmann@xmath27 , m. walter@xmath27 , y .- r .
wang@xmath30 , c. wendt@xmath30 , c. h. wiebusch@xmath62 , g. wikstrm@xmath43 , d. r. williams@xmath31 , r. wischnewski@xmath27 , h. wissing@xmath27 , k. woschnagg@xmath37 , x. w. xu@xmath30 , g. yodh@xmath36 , s. yoshida@xmath53 , j. d. zornoza@xmath63 _ ( 1 ) dept . of physics and astronomy , university of alaska anchorage , 3211 providence dr . , anchorage , ak 99508 , usa ( 2 )
ctsps , clark - atlanta university , atlanta , ga 30314 , usa ( 3 ) dept . of physics , southern university ,
baton rouge , la 70813 , usa ( 4 ) dept . of physics ,
university of california , berkeley , ca 94720 , usa ( 5 ) institut fr physik , humboldt universitt zu berlin , d-12489 berlin , germany ( 6 ) lawrence berkeley national laboratory , berkeley , ca 94720 , usa ( 7 ) universit libre de bruxelles , science faculty cp230 , b-1050 brussels , belgium ( 8) vrije universiteit brussel , dienst elem , b-1050 brussels , belgium ( 9 ) dept . of physics , chiba university ,
chiba 263 - 8522 japan ( 10 ) dept . of physics and astronomy , university of canterbury , private bag 4800 , christchurch , new zealand ( 11)dept . of physics , university of maryland ,
college park , md 20742 , usa ( 12 ) dept . of physics ,
universitt dortmund , d-44221 dortmund , germany ( 13 ) dept . of subatomic and radiation physics , university of gent , b-9000 gent , belgium
( 14 ) max - planck - institut fr kernphysik , d-69177 heidelberg , germany ( 15 ) dept . of physics and astronomy , university of california , irvine , ca 92697 , usa ( 16 ) dept . of physics and astronomy , university of kansas ,
lawrence , ks 66045 , usa ( 17 ) blackett laboratory , imperial college , london sw7 2bw , uk ( 18 ) dept . of astronomy ,
university of wisconsin , madison , wi 53706 , usa ( 19 ) dept . of physics ,
university of wisconsin , madison , wi 53706 , usa ( 20 ) institute of physics , university of mainz , staudinger weg 7 , d-55099 mainz , germany ( 21 ) university of mons - hainaut , 7000 mons , belgium ( 22 ) bartol research institute , university of delaware , newark , de 19716 , usa ( 23 ) dept . of physics ,
university of oxford , 1 keble road , oxford ox1 3np , uk ( 24 ) institute for advanced study , princeton , nj 08540 , usa ( 25 ) dept . of physics , university of wisconsin
, river falls , wi 54022 , usa ( 26 ) dept . of physics , stockholm university , se-10691 stockholm , sweden ( 27 )
dept . of astronomy and astrophysics , pennsylvania state university , university park ,
pa 16802 , usa ( 28 ) dept . of physics , pennsylvania state university , university park ,
pa 16802 , usa ( 29 ) division of high energy physics , uppsala university , s-75121 uppsala , sweden ( 30 ) dept . of physics and astronomy , utrecht university / sron ,
nl-3584 cc utrecht , the netherlands ( 31 ) dept . of physics ,
university of wuppertal , d-42119 wuppertal , germany ( 32 ) desy , d-15735 zeuthen , germany ( a ) deceased ( b ) on leave of absence from universit di bari , dipartimento di fisica , i-70126 , bari , italy ( c ) affiliated with dept . of chemistry and biomedical sciences , kalmar university , s-39182 kalmar , sweden ( d ) now at rwth aachen university , d-52056 aachen , germany ( e ) affiliated with ific ( csic - universitat de valncia ) , a. c. 22085 , 46071 valencia , spain _ | the amanda neutrino detector has been in operation at the south pole for several years .
a number of searches for extraterrestrial sources of high energy neutrinos have been performed .
a selection of results is presented in this paper .
the much larger icecube detector will extend the instrumented ice volume to a cubic kilometer and 9 out of 80 planned icecube strings have been deployed to date .
we present the status for both detectors .
= 11.6pt |
in the previous study , we introduced a stochastic aggregation process involving both active and passive clusters@xcite .
we generalized smoluchowski s rate equations and obtained exact results for several kernels . in this study
, we apply stochastic aggregation to reaction - diffusion , coarsening , and ballistic agglomeration problems .
our goal is to examine the range of validity the mean - field results , and to determine whether the overall scaling behavior extends to low dimensional systems .
the rate equations approach is mean - field in nature , i.e. , it is valid only when spatial correlations are absent .
formally , it is applicable only in infinite spatial dimension , or in the presence of an effective mixing mechanism .
this mean - field theory should also be _ asymptotically _ exact when the spatial dimension is sufficiently high .
in low spatial dimensions , however , significant spatial correlations eventually develop , and the rate equation approach does not apply in the long time limit .
we therefore focus on one - dimensional systems where spatial correlations are most pronounced .
we performed numerical simulations of stochastic aggregation processes with both diffusive and ballistic particle transport .
the simulations show that the scaling behavior suggested by the mean - field theory is indeed generic , as it extends to one - dimensional systems .
we find that two nontrivial model - dependent exponents characterize the survival probabilities of the particles and monomers , respectively .
smoluchowski s theory provides reasonable estimates for these exponents . additionally , we studied the limiting mass distribution of passive clusters .
surprisingly , over a substantial mass range , this distribution depends only weakly on the underlying transport mechanism .
furthermore , mean - field theory provides an excellent approximation for the limiting mass distribution .
the rest of this paper is organized as follows .
the general scaling behavior is outlined in sec .
predictions of the mean - field theory are summarized in sec .
numerical simulations of stochastic aggregation processes with diffusive and ballistic transport mechanisms are described in secs .
iv and v , respectively .
a discussion of the results is presented in sec .
stochastic aggregation involves two types of clusters : active and passive@xcite . initially , the system consists of active monomers only .
when two active clusters merge , the newly - born aggregate remains active with probability @xmath0 , or becomes passive ( i.e. , it never aggregates again ) with probability @xmath1 . eventually , all active clusters are depleted and the system consists of passive clusters only .
this process can be viewed as an aggregation - annihilation process since it interpolates between aggregation ( @xmath2 ) and annihilation ( @xmath3)@xcite .
quantities of interest include @xmath4 and @xmath5 , the distributions of active and passive clusters at time @xmath6 , as well as the final distribution of passive clusters , @xmath7 .
as shown in @xcite , two conservation laws underly this system .
the first is mass conservation , @xmath8={\rm const}$ ] .
the second conservation law reflects the fact that changes in the overall densities are coupled , @xmath9 , where @xmath10 and @xmath11 are the number densities of active and passive clusters , respectively .
therefore , it is sufficient to study the time evolution of the number density and the mass density of active clusters , @xmath12 and @xmath13 , respectively .
the latter quantity is the survival probability of an active particle , i.e. , the probability that it still belongs to an active cluster at time @xmath6 .
smoluchowski s theory suggests that both quantities decay algebraically in the long time limit @xmath14 as will be shown below , this as well as other scaling properties suggested by this theory hold qualitatively even for low dimensional stochastic aggregation processes . while the decay exponent @xmath15 is typically robust in that it depends only on the major characteristics of the process such as the spatial dimension or the transport mechanism , the exponent @xmath16 is non - universal as it depends on the details of the model , i.e. , on the probability @xmath0 . in turn
, this implies a non - universal growth law for the average mass of an active cluster @xmath17 . for the system to follow a scaling behavior
, the average mass must be the only relevant scale in the long time limit , and conversely , any scale characterizing the initial mass distribution must be `` erased '' eventually .
in other words , the mass distribution is characterized by a single rescaled variable @xmath18 with the time dependent prefactor fixed by the decay laws ( [ rs ] ) .
this scaling behavior is similar to that found for deterministic aggregation - annihilation processes@xcite and for aggregation - annihilation of domains in coarsening processes@xcite .
these studies suggest that another independent exponent describes the decay of small clusters .
specifically , the monomer density decays according to @xmath19 with a model - dependent exponent @xmath20 .
the monomer density decay reflects the small argument behavior of the scaling function @xmath21 with @xmath22 .
one of our main results is that the mass distribution of active clusters is described by a set of nontrivial exponents @xmath23 .
these exponents can be viewed as persistence exponents@xcite as they characterize the survival probability of an active particle , and an active monomer @xcite .
several properties of the scaling exponents are general .
for instance , the inequalities @xmath24 hold since @xmath25 .
the two exponents are equal @xmath26 in the annihilation case ( @xmath3 ) , since @xmath27 . in the aggregation limit ( @xmath2 ) the mass density of active clusters is conserved and therefore @xmath28 .
we now turn to the mass distribution of passive clusters .
the smoluchowski theory suggests that the same scaling form underlies both mass distributions @xmath29 in contrast with the active cluster distribution , the passive cluster distribution approaches a nontrivial final distribution @xmath7 . such a _ time
independent _ final distribution is consistent with the above scaling form only when the scaling function diverges , @xmath30 in the limit @xmath31 , with @xmath32 . as a result , the final mass distribution of passive clusters decays algebraically in the large mass limit @xmath33 at a given time @xmath6 , this decay is realized for clusters whose mass @xmath34 does not exceed the characteristic mass @xmath35 .
note also that @xmath36 implies @xmath37 .
generally , the mass conservation restricts the large mass decay exponent to @xmath38 .
since the @xmath39 exponent varies between @xmath40 and @xmath41 , we see that the entire range of acceptable exponents is realized by tuning the probability @xmath0 .
it is well established that spatial correlations can be safely neglected only in spatial dimensions larger than some upper critical dimension , @xmath42@xcite .
for example , for irreversible aggregation with mass - independent diffusion and reaction rates , one has @xmath43 ; for a general aggregation process , however , the upper critical dimension may be arbitrarily large@xcite .
below the upper critical dimension , substantial spatial correlations develop , and the most important features including the scaling exponents and the scaling functions are changed .
generally , the lower the spatial dimension , the larger the difference with the mean - field predictions .
although the smoluchowski rate equations approach does not apply in low spatial dimensions , it can still serve as a useful approximation after an appropriate modification .
this can be accomplished by replacing the overall reaction rate with an effective density - dependent reaction rate @xmath44 @xmath45 we are primarily interested in situations where aggregation is independent of the mass , and therefore we use a mass independent rate kernel .
the reaction rate @xmath46 is model dependent . in reaction - diffusion processes
, the reaction rate decays algebraically with the density ( see , e.g. , refs.@xcite ) .
assuming @xmath47 yields @xmath48 , and consequently , the density decay exponent is found @xmath49 in general , a reduction in the reaction rate , i.e. , @xmath50 , leads to a slowing down in the density decay rate , @xmath51 .
apart from the change in @xmath15 , all other aspects of this approximation are identical to the smoluchowski theory with a constant rate kernel . indeed , the above rate equations reduce to the smoluchowski s rate equations with a redefined time variable , @xmath52 . in particular , the scaling exponents @xmath39 and @xmath53 are independent of @xmath54 : @xmath55 one can verify the expected limiting behaviors @xmath56 , and @xmath57 .
furthermore , the scaling functions are as in the constant kernel solution @xcite , and for example , @xmath58 is purely exponential .
the corresponding small argument exponents @xmath59 and @xmath60 follow from @xmath39 and @xmath53 using the aforementioned scaling relations .
the final mass distribution of passive clusters is independent of the reaction rate @xmath61 @xcite @xmath62 below , we compare these mean - field predictions with simulation results for one - dimensional stochastic aggregation where spatial correlations are most pronounced .
we also examine the role of the aggregates transport mechanism by considering both diffusive and ballistic transport .
in diffusive stochastic aggregation , identical particles are placed onto a @xmath63-dimensional lattice .
all particles perform independent random walks , i.e. , they hop to a randomly chosen nearest - neighbor site with a constant rate .
if this site is occupied , the two particles coalesce irreversibly , and with probability @xmath0 the resulting aggregate remains active , while with probability @xmath1 it becomes passive . effectively ,
passive particles are removed from the system . in the case of single - species reaction diffusion processes , the effective reaction rate can be obtained from dimensional analysis .
( [ mft ] ) implies @xmath64=[l]^d[t]^{-1}$ ] , and since the reaction rate can only be a function of the diffusion coefficient @xmath65=[l]^2[t]^{-1}$ ] and the density @xmath66=[l]^{-d}$ ] , one finds @xmath67 .
hence , @xmath68 and eq .
( [ nu ] ) yields the correct decay exponents @xmath69@xcite below the upper critical dimension @xmath43 . to examine the above scaling picture we performed numerical simulations of diffusive stochastic aggregation processes in one dimension . unless noted otherwise , the data was obtained from an average over 10 independent realizations in a system of size @xmath70 with periodic boundary conditions .
initially , all sites were occupied .
first , we verified that the number density , the mass density , and the monomer density indeed decay algebraically in the long time limit , in accord with eqs .
( [ rs ] ) and ( [ r1 ] ) . the case @xmath71 is shown in fig . 1 , and the corresponding decay exponents were found : @xmath72 , @xmath73 , and @xmath74 .
mean - field theory correctly predicts @xmath75 .
furthermore , the predictions @xmath76 and @xmath77 provide a reasonable approximation .
one can compare with the case of disordered ( sinai ) diffusion where a real - space decimation procedure@xcite was used to determine _ exact _ values of these exponents@xcite .
remarkably , the disorder case exponent @xmath78 is in excellent agreement with the simulation value .
there is a small discrepancy with the second exponent @xmath79 .
additionally , we verified that the densities of active and passive clusters follow the scaling forms of eqs .
( [ rk ] ) and ( [ pk ] ) , respectively ( see fig . 2 ) . in agreement with the mean - field theory , the scaling functions decay exponentially for large masses . the number density , the mass density , and the monomer density versus time for @xmath71 .
= 7.5 cm scaling of the active and passive mass distributions . shown
are the scaling functions @xmath80 , and @xmath81 , versus the scaling variable @xmath82 at three different times @xmath83 , @xmath84 , and @xmath85 .
different scales correspond to @xmath58 and @xmath86 in the main figure since the latter diverges at the origin .
the data represent an average over @xmath87 independent realizations in a system of size @xmath88 for the case @xmath71 .
the exponent value @xmath89 was used .
the tail of the distribution is shown in the inset .
we also studied how the exponents vary with the probability @xmath0 , as shown in figs . 3 and 4 .
the exact exponents found for the disordered case by le doussal and monthus@xcite provide an excellent approximation ( within 0.1% ) for @xmath39 . in the case of @xmath53 ,
a different behavior emerges in the aggregation limit , @xmath90 , where the exact value is @xmath91 @xcite , and the disagreement with both mean - field theory and the disordered case are most pronounced . the exponent @xmath39 versus @xmath0 .
monte carlo simulation results for the pure diffusion case are compared with the mean - field theory ( [ pd ] ) and the exact value for the disordered case .
the latter is obtained from @xmath92@xcite , where @xmath93 is the confluent hypergeometric function . the exponent @xmath53 versus @xmath0 .
monte carlo simulation results for the pure diffusion case are compared with the mean - field theory ( [ pd ] ) and the exact value for the disordered case obtained from @xmath94@xcite .
the above scaling arguments suggest that the limiting mass distribution of passive clusters decays algebraically with the exponent @xmath32 .
for @xmath71 , one therefore expects @xmath95 ( compare with @xmath96 and @xmath97 , predicted by the disordered case and the mean - field theory ) .
this corresponds to a very strong suppression of large masses , and therefore , it is much more difficult to confirm this behavior numerically . nevertheless , our simulations ( fig . 5 )
are consistent with a power law decay with an exponent @xmath98 . in one dimension , the diffusion - controlled stochastic aggregation is equivalent to the potts model with zero - temperature glauber dynamics@xcite . for the @xmath99-state potts model with
spatially uncorrelated initial conditions , aggregation of domain walls occurs with probability @xmath100 , and annihilation occurs with probability @xmath101 .
therefore , the above can be reformulated in terms of domain walls rather than aggregates . in the coarsening context , the domain wall mass ( or number ) distribution is dual to domain number distribution @xcite . the distribution @xmath5 versus @xmath34 for three different times @xmath83 , @xmath84 and @xmath85 .
the typical mass at these three times is proportional to @xmath102 , @xmath103 , and @xmath104 , respectively .
hence , the distribution is fully developed only over a short range of masses .
the data represents an average over @xmath87 realizations in a system of size @xmath88 with @xmath71 .
the situation when particles move ballistically involves several complications . first ,
while the annihilation limit is uniquely defined @xcite , the aggregation limit has various realizations . in traffic flows ,
the velocity of a newly - born cluster is the smaller of the two velocities @xcite , while in application to astrophysics and granular gases the velocity follows from momentum conservation @xcite .
second , numerical results for the annihilation case @xcite and analytical results for the traffic case @xcite show that the initial conditions are remembered forever , in contrast with the diffusive case . specifically , the small velocity characteristics of the initial velocity distribution influence the long time asymptotic behavior , including the scaling exponents .
we consider the momentum conserving case , also known as `` ballistic aggregation '' or `` sticky gas '' @xcite .
the initial velocities are assigned according to the distribution @xmath105 .
the mass ( momentum ) of a newly - born cluster is equal to the sum of masses ( momenta ) of the two colliding clusters . after an agglomeration event
, the newborn particle remains active with probability @xmath0 , or becomes passive with probability @xmath1 .
to apply the smoluchowski rate equations approach , we again use dimensional analysis to calculate the decay exponent @xmath15 .
the collision rate is @xmath106 , where @xmath107 is the typical velocity and @xmath108 is the typical radius of an aggregate .
a particle of radius @xmath108 contains of the order @xmath109 monomers whose initial momenta are uncorrelated .
momentum conservation therefore implies @xmath110 .
using @xmath111 gives the collision rate @xmath112 . from eq .
( [ nu ] ) one finds @xmath113 with @xmath39 given by eq .
( [ pd ] ) . apart from the exponent @xmath15 , features such as the exponential mass distribution and the exponents @xmath39 and @xmath53 are given by the mean - field theory outlined above . in two dimensions ,
the collision rate does not depend on @xmath39 and hence , the asymptotic behavior @xmath114 agrees with that found for deterministic ballistic agglomeration@xcite . for @xmath115 , stochastic and deterministic asymptotics
differ : stochasticity enhances decay of the number density @xmath116 for @xmath117 and weakens it for @xmath118 .
a more detailed mean - field theory can be carried .
it yields a factorizing joint mass - velocity distribution , with an exponential mass distribution , and a gaussian velocity distribution @xcite . in the aggregation case , @xmath28 and
therefore the correct scaling exponent @xmath119 @xcite is recovered from eq .
( [ nub ] ) . for the annihilation case ,
however , initial conditions are `` remembered '' forever and therefore the above dimensional arguments no longer hold .
the predicted exponent in the annihilation case is always mean - field @xmath120 , while one - dimensional numerical simulations yield an exponent continuously varying from @xmath40 to @xmath41 depending on the initial velocity distribution @xmath105 , e.g. , @xmath121 for uniform initial distributions@xcite .
we have simulated the stochastic aggregation process on a one - dimensional ring with @xmath85 particles .
the initial velocity distribution was uniform in @xmath122 $ ] .
we measured the scaling exponent @xmath39 via the scaling relation @xmath123 , rather than directly versus time , since the exponent @xmath124 is not known analytically .
we have found that the mean - field prediction , @xmath125 , provides a reasonable approximation for the exponent @xmath39 , as shown in fig .
furthermore , this approximation should improve in higher dimensions .
we compared the mean - field prediction for the mass distribution of passive clusters , eq .
( [ pkinf ] ) , with the numerically obtained distributions in both ballistic and diffusive cases .
interestingly , the rate equations provide an excellent approximation for small and moderate masses ( see fig .
given that the discrepancy in @xmath39 is maximal for the case @xmath71 , one may expect an even better approximation for other values of @xmath0 .
noting the strong decay of this distribution , the contribution of very large masses is extremely small ; for example , @xmath126 for @xmath71 . hence , the most pronounced part of the distribution is well approximated by the rate equations theory .
surprisingly , the transport mechanism does not play an important role as far as the final mass distribution of passive clusters is concerned . the scaling exponent @xmath127 versus @xmath0 for ballistic aggregation compared with the mean - field value of eq .
( [ pd ] ) . the final distribution of passive clusters for the @xmath71 stochastic aggregation with diffusive and ballistic transport .
also shown is the mean - field distribution @xmath128 .
we have investigated diffusion- and ballistic - controlled stochastic aggregation in one dimension .
we have seen that the rate equations approach captures the overall scaling behavior and additionally it provides reasonable estimates for the decay exponents . in general ,
the mass distribution is characterized by two nontrivial model - dependent decay exponents . in the diffusion - controlled case ,
the exponent @xmath39 underlying the survival probability of a particle is in excellent agreement with the exact results from the disordered case .
in fact , one can not dismiss the possibility that the disordered and the pure values are identical , based on numerics alone .
however , there is an evident discrepancy in the exponent @xmath53 as the disordered case exponent diverges logarithmically in the aggregation limit .
stochastic aggregation is equivalent to domain coarsening in the zero - temperature potts - glauber model .
the above exponents @xmath23 characterize the domain wall number distribution in analogy with @xmath129 for the domain number distribution @xcite . in the latter case as well , exact values calculated for the disordered case provide an excellent approximation for the domain exponents . in general , the particle survival probability exponent @xmath39 is robust , while the monomer survival probability exponent @xmath53 is very sensitive to the details of the process . in the ballistic - controlled case
, we have shown that even in the absence of a consistent mean - field theory , some characteristics such as the exponent @xmath39 are well approximated by the rate equations .
understanding of reaction processes with an underlying ballistic transport remains largely incomplete .
the asymptotic behavior is highly sensitive to the initial conditions , and the critical dimension is apparently infinite .
in fact , exact analytical results are available mostly in the aggregation limit @xcite .
the most intriguing property of the stochastic aggregation is the profound lack of universality .
indeed , the weak dependence on the transport mechanism is in contrast with the strong dependence on the parameter @xmath0 .
for example , our numerical results show that the final distribution of passive clusters is very close in diffusion- and ballistic - controlled situations .
another very impressive manifestation of this is the excellent agreement between the values of the exponent @xmath127 in the disordered and pure cases .
p. l. krapivsky and e. ben - naim ,
stochastic aggregation : rate equations approach .
somewhat similar _
deterministic _ aggregation - annihilation processes have been investigated in refs.@xcite .
l. krapivsky , physica a * 198 * , 135 ( 1993 ) ; _ ibid _ * 198 * , 150 ( 1993 ) ; _ ibid _ * 198 * , 157 ( 1993 ) .
e. ben - naim and p. l. krapivsky , phys .
e * 52 * , 6006 ( 1995 ) .
l. g. zhang and z. r. yang , phys .
e * 55 * , 1442 ( 1997 ) .
p. l. krapivsky and e. ben - naim , phys .
e * 56 * , 3788 ( 1997 ) .
e. ben - naim and p. l. krapivsky , j. stat .
phys . * 93 * , 583 ( 1998 ) .
p. le doussal and c. monthus , phys .
e * 60 * , 1212 ( 1999 ) .
b. derrida , a. j. bray , and c. godrche , j. phys . a * 27 * , l357 ( 1994 ) .
l. krapivsky , e. ben - naim , and s. redner , phys .
e * 50 * , 2474 ( 1994 ) .
s. n. majumdar and c. cornell , phys .
e * 57 * , 3757 ( 1998 ) .
for a recent review , see s. redner , in _ nonequilibrium statistical mechanics in one dimension _
v. privman ( new york : cambridge university press , 1997 ) , p. 1 .
g. j. van dongen , phys .
lett . * 63 * , 1281 ( 1989 ) .
s. f. burlatsky , o. f. ivanov , and j. m. deutch , j. chem .
phys . * 97 * , 156 ( 1992 ) .
e. ben - naim , phys .
e * 53 * , 1566 ( 1996 ) .
d. s. fisher , phys .
b * 51 * , 6411 ( 1995 ) : p. le doussal , c. monthus , and d. s. fisher , phys .
e * 59 * , 4795 ( 1999 ) .
j. l. spouge , phys .
lett . * 60 * , 871 ( 1988 ) .
r. j. glauber , j. math .
phys . * 4 * , 294 ( 1963 ) .
y. elskens and h. l. frisch , phys .
a * 31 * , 3812 ( 1985 ) .
e. ben - naim , s. redner , and f. leyvraz , phys .
* 70 * , 1890 ( 1993 ) . p. l. krapivsky , s. redner , and f. leyvraz , phys .
e * 51 * , 3977 ( 1995 ) .
m. rey , p. a. rey , l. frachebourg , and j. piasecki , phys .
lett . * 75 * , 138 ( 1995 ) .
p. a. rey , m. droz , and j. piasecki , phys .
e * 57 * , 138 ( 1998 ) .
e. ben - naim , p. l. krapivsky , and s. redner , phys .
e * 50 * , 822 ( 1994 ) .
s. f. shandarin and ya .
b. zeldovich , rev .
phys . *
61 * , 185 ( 1989 ) .
e. ben - naim , s. y. chen , g. d. doolen , and s. redner , phys .
lett . * 83 * , 4069 ( 1999 ) .
g. f. carnevale , y. pomeau , and w .- r .
young , phys .
lett . * 64 * , 2913 ( 1990 ) .
j. piasecki , physica a * 190 * , 95 ( 1992 ) .
y. jiang and f. leyvraz , j. phys .
a * 26 * , l179 ( 1993 ) .
a. martin and j. piasecki , j. stat .
phys . * 76 * , 447 ( 1994 ) .
p. l. krapivsky and e. ben - naim , phys .
e * 53 * , 291 ( 1996 ) .
l. frachebourg , phys .
82 * , 1502 ( 1999 ) .
l. frachebourg , ph .
a. martin , and j. piasecki , _ cond - mat/9911346_. | we study scaling properties of stochastic aggregation processes in one dimension . numerical simulations for both diffusive and ballistic transport show that the mass distribution is characterized by two independent nontrivial exponents corresponding to the survival probability of particles and monomers .
the overall behavior agrees qualitatively with the mean - field theory .
this theory also provides a useful approximation for the decay exponents , as well as the limiting mass distribution .
pacs numbers : 05.40.-a , 05.20.dd , 82.20.mj 2 |
in an imaging - air - cherenkov - telescope ( iact ) experiment like magic @xcite the energy @xmath1 of the cosmic ray particle ( @xmath0-ray ) is not exactly known .
it has to be estimated , the energy resolution being in the order of 20 to 40% . as a consequence ,
the experimentally measured energy spectrum is biased .
the procedure to correct for the effects due to the finite energy resolution is called unfolding . while in high - energy - physics experiments unfolding is a widely used technique , this is not the case in present day s iact experiments
. this paper deals with the unfolding procedure , which is applied as a standard tool in the magic experiment .
the different unfolding methods are explained in detail .
emphasis is put on the discussion of questions and problems related to the application of the unfolding to real data .
it is not the aim of the paper to give a complete derivation of all formulas . for this the reader
is referred to the publications @xcite to @xcite .
an excellent review of unfolding methods is given in @xcite .
the present paper makes use of many ideas discussed in that paper .
although here only differential energy spectra are considered , the procedures are equally well applicable to distributions of other quantities , including distributions in more than one dimension @xcite .
the layout of the paper is as follows . in section [ section : unfoldingaim ]
the notation is defined and the motivation for the unfolding procedure is given .
the different unfolding methods , which means the different ways of regularization , are presented in section [ section : regularization ] .
the so - called forward unfolding , which represents an implicit unfolding under the assumption of a certain parametrization of the unfolding result @xmath2 , is explained in section [ section : forwardunfolding ] . in section [ section : usefulquantities ] two quantities are introduced , which are useful for an optimal choice of the regularization strength .
the criteria for this choice are collected in section [ section : bestweight ] .
section [ section : comments ] discusses various technical aspects which are important in the application of the unfolding procedure to real data .
two particular technical procedures , which ensure an unbiased and robust unfolding , are presented in sections [ section : aeffandm ] and [ section : combine ] .
some unfolding results , obtained by applying the unfolding procedure to data taken in the magic experiment , are discussed in section [ section : application ] . finally the pros and contras of the method of correction factors , which is an alternative way of correcting data for effects due to the finite experimental resolution , are listed in section [ section : correctionfactors ] .
a summary is given in section [ section : summary ] .
in this section the notations are defined and the motivation for the unfolding procedure is explained .
the true and measured ( estimated ) values of the energy of the cosmic ray particle are denoted by @xmath3 and @xmath4 respectively .
the data are assumed to be binned in histograms , and certain binnings are chosen independently for the distributions in @xmath3 and @xmath4 .
furthermore , the following definitions are introduced : + [ cols="^ , < , < " , ] + with @xmath5 in the following it is assumed that the rank @xmath6 of @xmath7 is equal to ( and not less than ) the minimum of @xmath8 and @xmath9 , where @xmath8 and @xmath9 are the number of bins in @xmath4 and @xmath3 respectively , which are used in the unfolding .
this can always be achieved by a proper choice of the binnings in @xmath3 and @xmath4 .
the migration matrix contains the most likely fraction of events moving from a bin @xmath10 in @xmath3 into a bin @xmath11 of @xmath4 , due to the finite experimental energy resolution : @xmath12 the migration matrix @xmath13 is obtained from monte carlo ( mc ) simulations , in which the development of the air shower ( induced by the cosmic ray particle in the atmosphere ) , the emission of cherenkov light in the air shower , the geometrical , optical and electronic properties of the telescope and the experimental procedures in the data analysis ( shower reconstruction , @xmath0/hadron separation , energy estimation , selections , cuts ) are simulated @xcite .
@xmath13 is computed from a 2-dimensional plot of the number of reconstructed mc events in the @xmath4-@xmath3 plane , which was produced under the same conditions ( selections , cuts ) as the experimental distribution @xmath14 .
it is the aim of the unfolding procedure to determine the true distribution @xmath2 , given @xmath14 and @xmath13 .
+ it should be noted that @xmath13 only describes the migration of events .
it does not describe losses of events , which will occur due to the finite acceptance of the detector , due to the trigger conditions and due to additional selections and cuts . in an iact experiment these losses
are also determined by monte carlo simulations , and the corresponding correction factor is the effective collection area @xmath15 .
@xmath15 has to be computed again under the same conditions as the experimental distribution @xmath14 .
apart from minor effects ( see section [ section : aeffandm ] ) , this correction can be performed quite independently of the unfolding .
+ the unfolding can be understood as a reshuffling of events from the bins of @xmath4 into the bins of @xmath3 . in this procedure
the numerical values of the bin edges , both for @xmath3 and @xmath4 , are completely irrelevant . @xmath3 and
@xmath4 may be even two different physical quantities , with completely different ranges of values and different units , like @xmath3 = ( the true energy of the cosmic ray particle ) and @xmath4 = ( the total number of cherenkov photons measured for the shower ) @xcite .
of course , unfolding makes only sense if @xmath3 and @xmath4 are sufficiently strongly correlated , otherwise the distribution of @xmath3 can not be inferred from a distribution of @xmath4 .
this is in contrast to the method of correction factors ( see section [ section : correctionfactors ] ) , where @xmath4 has to be a good estimate of @xmath3 in any case . because the binnings in @xmath3 and @xmath4 can be chosen independently and to a certain degree arbitrarily ( see section [ section : comments ] ) , the migration matrix @xmath13 is in general not square . for this reason eq .
( [ eq : yms ] ) can in general not be inverted to obtain @xmath2 as @xmath16 .
+ as the unfolding is equivalent to a reshuffling of events from the bins of the measured distribution into the bins of the true distribution the unfolding is not restricted to 1-dimensional distributions but can in the same way be applied to multi - dimensional distributions .
the information necessary for the unfolding procedure is completely contained in the corresponding migration matrix , which in the case of multi - dimensional distributions describes the migration of events from the bins of the true multi - dimensional distribution into the bins of the measured multi - dimensional distribution .
the dimensions of the measured and true distributions may also be different .
an example for an unfolding in 2 dimensions is given in @xcite .
very generally , the solution @xmath2 of the system of linear equations ( [ eq : yms ] ) can be obtained by minimizing the least - squares expression @xmath18 where the @xmath9 components of @xmath2 are the free parameters .
minimizing @xmath19 will yield solutions for @xmath2 which , after folding with @xmath13 , are best compatible with the measurement @xmath14 .
two cases have to be distinguished : * * the underconstrained case @xmath20 . * + because @xmath21 the @xmath22 matrix @xmath7 can be inverted and a particular solution @xmath23 can be written as @xmath24 if @xmath25 , the solutions @xmath26 form a space of @xmath27 dimensions with @xmath28 . for @xmath29 eq.([eq : s0 ] ) reduces to @xmath30 and @xmath31 is the only solution . in both cases
@xmath32 , implying @xmath33 .
the solutions are independent of the covariance matrix @xmath34 . moreover , because of ( [ eq : normalizem ] ) the total number of events is not changed : @xmath35 * * the overconstrained case @xmath36 . * + minimizing @xmath19 by varying @xmath2 yields @xmath37 where @xmath38 is the @xmath39 matrix @xmath40 the solution @xmath41 now depends on @xmath34 .
the minimum value of @xmath19 becomes @xmath42 expression ( [ eq : slsq ] ) is also valid for @xmath43 , in which case it reduces to @xmath44 .
it can be shown that the direct solutions ( [ eq : s0 ] ) and ( [ eq : slsq ] ) may lead to large errors of @xmath45 , reflected in large absolute values of the elements of the error matrix @xmath46 of @xmath2 .
this behaviour can be traced back to small eigenvalues of the matrix @xmath7 and @xmath38 respectively @xcite .
in order to reduce the large errors of @xmath2 , a procedure called regularization is applied . by the regularization additional constraints
are imposed on @xmath2 , by which some information in the measurements @xmath47 is discarded .
regularization can be viewed as a smearing of the unfolded distribution with some finite resolution , which reduces the correlations between the @xmath48 of adjacent bins at the expense that @xmath48 is no longer an unbiased estimate of the true distribution @xcite .
+ the bias increases with increasing regularization strength .
nevertheless , it turns out that with properly tuned regularization ( see section [ section : bestweight ] ) solutions can be obtained which are much closer to the true distribution than the direct solutions ( [ eq : s0 ] ) or ( [ eq : slsq ] ) .
+ it is evident that regularization is particularly important in the underconstrained case .
however , also in the overconstrained case regularization makes sense : even if the system of equations ( [ eq : yms ] ) is formally overconstrained ( @xmath49 ) , it may be effectively underconstrained .
this happens for example if some of the measurements @xmath47 have much larger errors than the other @xmath47 .
+ in the following , three different ways of regularization are described @xcite . in some unfolding methods regularization
is performed by adding a regularization term @xmath51 in the expression for @xmath19 ( eq.[eq : chi20 ] ) @xmath52 @xmath53 , also called regularization parameter , is a weight which allows to steer the regularization strength : large values of @xmath53 correspond to weak regularization , small values to strong regularization .
+ * * tikhonov s method * + [ section : tikhonov ] in tikhonov s method @xcite the regularization term is defined as @xmath54 for the second derivative @xmath55 of @xmath2 in bin @xmath10 different approximations may be used .
the expression used in the magic software @xcite is @xmath56 this is actually an approximation for the bin - to - bin variation of @xmath57 . in @xcite @xmath55
is calculated from a spline representation of @xmath2 .
+ for a given value of @xmath53 and after specifying @xmath51 , expression ( [ eq : chi2reg ] ) can be minimized numerically by varying the components of @xmath2 .
the minimization also provides the error matrix @xmath46 of @xmath2 .
the regularization matrix @xmath58 , defined in ( [ eq13a ] ) , can be calculated numerically by performing minimizations with modified values of @xmath47 . + * * schmelling s method * + [ section : schmelling ] in this method , which is discussed in great detail in @xcite , the regularization term is set equal to the `` cross entropy '' @xmath59 @xmath60 is the normalized distribution @xmath2 @xmath61 and @xmath62 is a normalized prior distribution , which describes a prior knowledge about @xmath2 .
the cross entropy @xmath51 quantifies by how much @xmath63 deviates from @xmath62 .
finding @xmath2 by minimizing the cross entropy @xmath51 simultaneously with the least squares expression @xmath19 is called the method of `` reduced cross entropy '' .
+ with @xmath51 from ( [ eq : regschmelling ] ) , expression ( [ eq : chi2reg ] ) can now be minimized to obtain the unfolded distribution @xmath2 .
note that all components of @xmath2 and of the prior distribution are required to be @xmath64 , because otherwise @xmath51 in ( [ eq : regschmelling ] ) can not be defined .
the expressions for the error matrix @xmath46 of @xmath2 and for the regularization matrix @xmath58 are given in @xcite .
in the magic software , the condition @xmath65 is used as an additional constraint , when minimizing @xmath66 . by this one degree of freedom is gained . in some unfolding methods regularization
is performed by suppresing small eigenvalues @xmath67 of @xmath7 by a factor @xmath68 @xcite . by the suppression factor
@xmath68 the matrix @xmath7 and its inverse are modified . in terms of the eigenvectors @xmath69 of @xmath7 they read @xmath70 where the sums extend over all eigenvalues @xmath67 which are different from zero .
like @xmath7 and @xmath71 , @xmath72 and @xmath73 are @xmath22 matrices . without suppression , @xmath74 , @xmath73 is equal to @xmath71 in the underconstrained case @xmath20 . in the overconstrained case ,
@xmath75 , @xmath71 is undefined but @xmath73 can be calculated . a similar factor
@xmath76 can be defined to suppress small eigenvalues @xmath77 of @xmath38 ( eq . [ eq : h ] ) .
there is considerable freedom as to the choice of the values or expressions for @xmath68 and @xmath76 .
one may introduce a parameter @xmath11 such that in the limit @xmath78 the suppression factors @xmath79 and @xmath80 tend to 1 .
@xmath11 has a similar meaning as the weight @xmath53 in eq.([eq : chi2reg ] ) : it determines the regularization strength and for @xmath78 the solutions tend to the direct solutions ( [ eq : s0 ] ) and ( [ eq : slsq ] ) respectively .
+ the expressions for the error matrix @xmath46 of @xmath2 and for the regularization matrix @xmath58 are given in @xcite .
another way of unfolding is to calculate a solution @xmath2 iteratively @xcite .
the regularization is done by stopping the iteration at some point . in this case
the number of iterations @xmath11 plays a similar role as the weight @xmath53 in ( [ eq : chi2reg ] ) . in the limit of an infinite number of iterations , which is equivalent to a very large weight @xmath53
, the solution tends to the direct solution ( [ eq : s0 ] ) or ( [ eq : slsq ] ) respectively .
+ * @xmath81 + a simple iteration scheme @xcite for solving @xmath82 with @xmath83 is + @xmath84 + where @xmath11 is the iteration number and @xmath85 is a relaxation parameter .
the latter should be chosen in the range @xmath86 , where @xmath87 is the largest eigenvalue of @xmath7 .
eq.([eq : c ] ) leads to @xmath88 where @xmath89 is the starting value of @xmath90 .
the unfolded distribution is obtained as @xmath91 .
+ in terms of the suppression factor @xmath92 this regularization can be expressed as @xcite @xmath93 \label{eq23}\end{aligned}\ ] ] if @xmath89 is set to zero , and @xmath94 \label{eq24}\end{aligned}\ ] ] if @xmath89 is chosen to be equal to @xmath14 .
the solutions @xmath95 tends to the direct solution @xmath96 ( eq .
( [ eq : s0 ] ) ) .
+ this procedure can also be applied in the overconstrained case , @xmath75 . in this case @xmath71
is undefined and it has to be replaced by @xmath73 ( eq.[eq : gtilde ] ) .
the solutions @xmath95 tend to @xmath97 , with @xmath74 . + * @xmath98 + in a similar way one may define suppression factors @xmath99 for the eigenvalues of the matrix @xmath38 . in this case
the solution @xmath2 tends to the solution @xmath41 ( eq . ( [ eq : slsq ] ) ) . in the last 2 methods , spectral window method and regularization by iteration ,
@xmath19 does nt appear explicitly .
however , its value can be calculated and it is taken into account when determining the optimum regularization strength ( see section [ section : bestweight ] ) .
once the the unfolded distribution @xmath100 is determined , by any of the methods described in this section , the differential energy spectrum @xmath101 of @xmath0-rays is calculated using eq .
( [ eq : phik ] ) .
@xmath101 has the meaning of the average differential @xmath0-ray flux in the @xmath102-th bin of @xmath3 ( eq.([eq : phie ] ) ) .
an implicit unfolding can be done by representing @xmath2 as a parametric function @xmath103 with parameters @xmath104 and minimizing @xmath19 in ( [ eq : chi20 ] ) @xmath105 with respect to the parameters @xmath106 .
the number of measurements is equal to @xmath8 , the number of unknowns @xmath107 .
thus , the problem is overconstrained if @xmath108 , independent of the value of @xmath9 .
one degree of freedom is gained if the total number of events is required to stay constant : @xmath109 . in many cases
the minimization of @xmath19 can be performed analytically , by solving @xmath110 , similarly to the procedure described in section [ subsection : directsolution ] .
the parametrization of @xmath2 can be written in the form @xmath111 with @xmath112 here @xmath113 is the assumed parametrization of the differential energy spectrum of @xmath0-rays , @xmath114 is the effective observation time , @xmath115 denotes the @xmath102-th bin in @xmath3 or its width , @xmath116 is the effectice collection area , and @xmath117 is the reduction factor due to an additional cut ( for example in @xmath4 , see section [ section : additionalcuts ] ) .
a further correction @xmath118 can be introduced , if @xmath113 is supposed to represent the differential energy spectrum of @xmath0-rays before absorption , either at the @xmath0-source or by interaction with the extragalactic photon background .
@xmath119 can be determined from mc simulations , whereas @xmath120 can be calculated in models about the extragalactic photon background @xcite .
parametrizing @xmath2 as an analytic function of @xmath3 , with some free parameters @xmath106 , can be understood as a kind of regularization , because it forces the solution @xmath2 and its derivatives to be continuous , leading to a suppression of the noise component of @xmath2 .
the forward unfolding does not provide an unfolded distribution @xmath2 .
it provides those parameter values @xmath106 for an assumed parametrization of @xmath121 , which minimize @xmath19 in ( [ eq6a ] ) . of course , @xmath2
can then be calculated from @xmath121 via ( [ eq : phik ] ) . under the assumption of a certain parametrization of @xmath121 ,
the forward unfolding is a very robust method of determining the best parameter values @xmath106 .
moreover , since there is no regularization strength to be adjusted , the uncertainty as to its choice does not exist .
therefore forward unfolding represents a powerful and useful check of the unfolding results obtained by any of the methods described in section [ section : regularization ] . in those methods the parametrization is only introduced after the actual unfolding of the measurements @xmath14 .
in this section two quantities are explained which are useful for judging the quality of the unfolding result : the error matrix ( covariance matrix ) @xmath46 of the unfolded distribution @xmath2 and the regularization matrix @xmath58 . in all unfolding methods
the covariance matrix @xmath46 of @xmath2 can be determined .
the trace of @xmath46 , @xmath122 , measures the noise component of @xmath2 , as @xmath123 measures the noise component of @xmath14 . in the methods where the solution @xmath2 is given as @xmath124 , like in schmelling s method or in the spectral window method , the covariance matrix @xmath46 of @xmath2 is obtained by @xmath125 . + in those methods where the solution is determined by a numerical minimization of @xmath66 , like in tikhonov s method , @xmath46 is obtained from the shape of @xmath66 in the region around the minimum . a quantity which describes how the estimates @xmath126 of @xmath127 couple to the measurements @xmath47 is given by the @xmath22 matrix @xmath128 also called regularization matrix @xcite .
the trace of @xmath58 can be interpreted as the effective number of measurements used in the unfolding procedure .
the number of effectively rejected measurements is then equal to @xmath129 .
+ the maximum value of @xmath130 , which is equal to the rank @xmath6 of @xmath7 , is reached with the direct solutions ( [ eq : s0 ] ) and ( [ eq : slsq ] ) , corresponding to the cases @xmath20 and @xmath75 respectively : for the direct solution @xmath31 the number of rejected measurements @xmath131 is equal to @xmath132 , which means that no information is discarded .
the measurements @xmath14 are completely reproduced by the unfolding : @xmath133 . for the least squares solution @xmath41 the number of effectively rejected measurements is @xmath134 , which means that some information is discarded .
the measurements @xmath14 are not exactly reproduced by the unfolding : @xmath135 .
the fact that the system is overconstrained has a similar effect as regularization . in both cases , with increasing regularization strength @xmath130 is reduced and @xmath131 is increased .
for a given unfolding method the result @xmath2 depends on the regularization strength , which is given by the weight @xmath53 ( or by the number of iterations @xmath11 respectively ) . in the literature various criteria for choosing the `` best '' weight
are proposed @xcite .
unfortunately , none of them provides a choice which is optimal for all cases .
reasons for this are : the optimum regularization strength in general depends on the shape of the unknown distribution @xmath2 .
it also depends on the binnings in @xmath3 and @xmath4 and on the prior distribution ( if applicable ) .
+ the effect of the regularization is illustrated in fig .
[ fig : iterbertero ] , where different quantities are plotted as a function of the iteration number @xmath11 . in this example an experimental energy distribution of @xmath0-rays from the crab nebula @xcite , i.e. the number of excess events in bins of the estimated energy @xmath4 , was unfolded .
the unfolding was performed for 30 different @xmath11 in the range @xmath136 to @xmath137 , using the method of bertero ( eq.([eq24 ] ) ) .
more results from the analysis of these data are given in section [ section : application ] . with decreasing @xmath11 , i.e with increasing regularization strength , one observes an increase of @xmath19 ( eq . [ eq : chi20 ] ) and a decrease in the quantities @xmath138 , @xmath130 , @xmath139 ( eq . [ eq : regtikhonov ] ) and @xmath140 ( eq . [ eq : regschmelling ] ) .
very similar behavior is found for the other unfolding methods , discussed in section [ section : regularization ] .
obviously , an acceptable unfolding result should satisfy the following conditions : * the @xmath66-probability , calculated from the value of @xmath19 and the number of degrees of freedom in the unfolding procedure , should be acceptable , say @xmath141 .
otherwise the unfolding result is incompatible with the measured distribution @xmath14 . +
* the noise term @xmath122 of the unfolded distribution @xmath2 should be comparable to the noise term @xmath123 of the measurements .
the main aim of regularization is a suppression of the large noise term of @xmath2 , which one often obtains if no regularization is applied . a large noise term @xmath122 , as well as large correlation terms of @xmath46 ,
indicate a too fine binning in @xmath3 , leading to small eigenvalues of @xmath7 or @xmath38 ( see section [ sect : optimalbinnings ] ) . +
* @xmath130 should not be much lower than its maximum possible value , which is equal to the rank @xmath6 of the matrix @xmath7 .
otherwise the solution is too strongly dominated and biased by the regularization . for determining the `` best '' regularization strength a compromise has to be found between the above requirements .
it has turned out that the criterion @xmath142 in general leads to solutions which satisfy the above conditions reasonably well , provided the problem is not strongly overconstrained . in the latter case , where the unfolding result is better constrained , a solution with @xmath143 is more apropriate . in magic the standard criterion for determining the optimal regularization strength is @xmath142 .
the full circles in fig . [ fig : iterbertero ] indicate this choice .
however , any other regularization strength can be chosen by hand , if this is suggested by the behaviour of the quantities @xmath144 or @xmath145 .
unfolding with regularization is a procedure which allows freedom in the choice of the regularization method and in the choice of the regularization strength .
the above criteria for an acceptable solution strongly restrict this freedom .
nevertheless a certain degree of arbitrariness remains as to which unfolding result should be considered representative and final . in magic
a selected unfolding result is considered representative if all other unfolding methods yield results , which are also acceptable and statistically consistent with the selected result .
in addition , it is required that also the forward unfolding ( section [ section : forwardunfolding ] ) , using a reasonable parametrization of @xmath121 , gives a consistent result .
an uncertainty due to the unfolding is determined from the spread of the @xmath45 , obtained from the different unfolding methods .
in the actual application of the unfolding procedure to real data some technical details have to be considered , and they are discussed in this section .
the binning of the experimental distribution @xmath14 is often dictated by the available statistics and by the experimental errors .
the binning should not be chosen too fine in order to assure significant measurements in all bins . in the case of an iact experiment ,
a sufficiently large sample is required to determine the number of signal ( excess ) events with sufficient accuracy .
the binning should not be chosen too wide either , because the binning in @xmath14 limits the reconstruction of the fine structure of the unfolded distribution @xmath2 .
another criterion for the binnings is the behavior of gram s matrix @xmath7 .
a too fine binning for @xmath2 leads to strong correlations between neighboring columns of the migration matrix @xmath13 , implying small eigenvalues of @xmath7 , which lead to a large noise component of @xmath2 . given a certain choice of the binning in @xmath4 and thus of @xmath8 , the bin size in @xmath3 or @xmath9
should be set such , that the system of linear equations ( [ eq : yms ] ) is not underconstrained .
this usually leads to wider bins in @xmath3 than in @xmath4 . in magic
a typical value of @xmath146 is 1.4 ( see section [ section : application ] ) . + it should also be noted that the unfolding procedure does nt require equidistant bins , neither in @xmath4 nor in @xmath3 .
if @xmath147 specifies the range of bins of the measured distribution @xmath14 which are to be considered in the unfolding procedure , also the range in @xmath3 to be considered in the unfolding has to be chosen properly : one has to make sure that all bins @xmath10 of @xmath3 are present , for which the column @xmath10 of the migration matrix @xmath148 contributes to the selected bins of @xmath14 , i.e. for which at least one of the elements @xmath149 is different from zero .
there is an exception to this rule , if for some reason certain bins @xmath10 of @xmath3 are not expected to contribute to the selected bins in @xmath4 .
this is for example the case if one of the factors @xmath150 in ( [ eq : phik ] ) is so small that @xmath100 can be neglected .
as explained in section [ section : notation ] , the distribution @xmath14 and the migration matrix @xmath13 have to be produced under identical conditions ( selections , cuts ) .
if an additional cut is imposed when generating @xmath14 , also @xmath13 has to be recalculated before doing the unfolding .
if this additional cut is a cut in @xmath4 one may proceed in the following way : * renormalize the columns @xmath10 of @xmath13 to the selected range in @xmath4 ( see eq.([eq : normalizem ] ) ) .
* perform the unfolding of @xmath14 in the usual way . *
apply a correction @xmath151 to the unfolded distribution @xmath45 , where @xmath151 is the renormalization factor for column @xmath10 of @xmath13 . in the cases where the unfolding procedure involves numerical minimizations , like the ones discussed in section [ subsection : regularizationterm ] ,
the minimization may not converge .
this problem can be often solved by choosing different starting values .
another reason for non - convergence is discussed in section [ section : cannotbedetermined ] . in magic
the standard choice of the starting distribution for @xmath2 , and also of the prior distribution @xmath62 in schmelling s method , is a distribution which is close to the measured distribution @xmath14 . according to eq .
( [ eq : yms ] ) those @xmath45 for which the column @xmath148 @xmath152 is a null vector have no influence on @xmath14 and can therefore not be determined in the unfolding .
these components should not be varied in the minimization because they would unnecessarily complicate the minimization process and may lead to non - convergence . for
the unfolding the migration matrix @xmath148 is the crucial quantity .
obviously , if it does nt describe the real migration of events correctly , the unfolding result will be wrong .
this means that at fixed @xmath10 , i.e. at fixed @xmath3 , the mc simulation has to describe the migration in @xmath1 correctly . this will be the case if at fixed @xmath3 the shower simulation is realistic and if the detector response is simulated correctly .
on the other hand , the distribution of @xmath3 in the mc need not agree with the real distribution of @xmath1 : due to the normalization of @xmath13 ( eq . ( [ eq : normalizem ] ) )
the bin - to - bin - variation of the number of mc events in @xmath3 has no influence on @xmath148 at all .
this is one of the great advantages of unfolding methods like those presented in section [ section : regularization ] as compared to methods based on correction factors ( see section [ section : correctionfactors ] ) .
however , there is a residual dependence of @xmath148 due to the finite binning in @xmath3 : depending on the shape of the @xmath3 distribution within an @xmath3 bin in the mc simulation , the calculated @xmath148 may be more representative for the lower , middle or upper part of the @xmath3 bin .
if the @xmath3 distribution in the real data is different from that in the mc simulation the calculated @xmath148 may not be exactly the right ones .
this residual dependence of @xmath13 on the shape of the @xmath3 distribution in the mc can be nearly completely removed by an iteration procedure in which the @xmath13 for the next iteration step is determined from a mc sample , in which the @xmath3 distribution has been corrected using the unfolding result of the last iteration step ( see section [ section : aeffandm ] ) .
in an iact experiment it is important that the effective collection area @xmath153 , which enters in the flux calculation , is computed taking into account a realistic shape of the differential flux @xmath121 and the actual distribution of effective observation times @xmath155 in the zenith angle @xmath154 .
this is also important for the migration matrix @xmath13 , which enters in the unfolding , because of the residual dependence on the flux spectrum , as discussed in section [ section : mcdependence ] .
recalculating @xmath153 and @xmath13 with the proper @xmath3 and @xmath154 spectra is the more important the bigger the @xmath156 interval and the stronger the variations of @xmath13 and @xmath153 within this bin are .
often , for statistics reasons , large bin sizes in @xmath4 and @xmath154 , and thus also in @xmath3 , have to be chosen . in the following it is assumed that the effective collection area @xmath153 and the migration matrix @xmath13 are known functions of @xmath3 and @xmath154 .
this can be achieved by determining them from a sample of mc @xmath0-ray events in very fine bins of @xmath3 and @xmath154 .
the aim is to calculate an average @xmath157 of the effective collection area and an average @xmath158 of the migration matrix , which are representative for a finite bin @xmath156 in @xmath3 and @xmath154 .
the number of observed events in a @xmath156 bin is given by @xmath159 here @xmath160 is the differential @xmath0-ray flux to be measured , @xmath161 is the distribution of observation times in the experimental data and @xmath162 is the known dependence of the effective collection area on @xmath3 and @xmath154 . with the definitions @xmath163 equation ( [ eq41 ] )
can be rewritten as @xmath164 this is the usual formula for converting numbers of events @xmath2 into differential fluxes @xmath165 . because of the definitions ( [ eq42a ] ) and ( [ eq44x ] ) the differential flux @xmath166 in an @xmath3 bin , as determined in the unfolding , is the average differential flux in this bin .
therefore , when quoting or plotting a result for @xmath166 the bin edges in @xmath3 should also be given or shown .
the number of reconstructed mc events in a bin @xmath11 of @xmath4 can be written as @xmath167 @xmath168 is the element of the normalized migration matrix for the @xmath11-th bin in @xmath4 , at an energy @xmath3 .
the index @xmath10 of @xmath148 is replaced by the variable @xmath3 .
the dependence of @xmath168 on @xmath3 and @xmath154 is assumed to be known from mc simulations .
the average migration matrix @xmath169 for the selected @xmath170 bin , to be used in the unfolding , is then obtained by @xmath171 the averages @xmath169 and @xmath157 are calculated according to ( [ eq60 ] ) and ( [ eq44x ] ) respectively , using an approximation @xmath172 of the function @xmath160 .
the measured distribution @xmath14 is unfolded using @xmath169 , yielding the unfolded distribution @xmath2 . a new approximation @xmath173 is then determined from @xmath2 according to eq .
( [ eq47 ] ) .
the procedure is iterated until @xmath160 has converged . in practice parametric functions
are used as approximations of @xmath160 , and in general the convergence is found to be very fast .
one often has the situation that there exist several measured distributions @xmath174 of the same quantity . if the @xmath174 were obtained under different conditions also the migration matrices @xmath175 and the effective collection areas @xmath176 will be different for the different measurements . in an iact experiment
the different conditions may be * different modes of observation ( on / off mode , wobble mode , observation in the presence of moon light , ... ) .
* different ranges of the zenith angle .
* different detector conditions .
* etc . in order to determine a final unfolded distribution @xmath2 one may proceed in different ways : * * individual unfolding * : + unfold each @xmath174 using @xmath175 to obtain @xmath177 , and combine the @xmath177 to obtain the final solution @xmath2 . +
* * global unfolding * : + combine the @xmath174 and @xmath175 to obtain a global @xmath14 and @xmath13 , do an unfolding of @xmath14 using @xmath13 , which will give the final solution @xmath2 .
there is one important argument in favour of the second option : each of the @xmath174 , or some of them , may have large statistical errors , making the unfolding of the individual @xmath174 unstable .
one common unfolding of the global @xmath14 using the global @xmath13 will be more robust , in general . in the case of an iact experiment ,
the following relations hold for each measurement @xmath178 : @xmath179 @xmath10 denotes the @xmath10-th bin in @xmath3 , @xmath180 is its width , @xmath181 is the average flux and @xmath182 is the effective collection area in this bin , and @xmath46 is the effective observation time . since @xmath165 and @xmath183 are the same for all @xmath178 the relations for the combined data read : @xmath184 inserting @xmath185 and @xmath186 in ( [ eq32 ] ) and using ( [ eq31 ] ) one obtains @xmath187 the relations ( [ eq32a ] ) and ( [ eq33 ] ) give the prescription how to combine the individual @xmath175 to obtain the global @xmath13 , and how to combine the individual @xmath188 to obtain the global @xmath182 . the measured distribution @xmath14 is unfolded using @xmath13 , and the unfolded distribution @xmath2 is converted into a flux @xmath165 according to ( [ eq32 ] ) . as can be seen from ( [ eq32a ] ) and ( [ eq33 ] ) the @xmath182 are weighted averages of the @xmath188 with weights @xmath189 , and the @xmath148 are weighted averages of the @xmath190 with weights @xmath191 .
the weights can be interpreted as fractions because they add up to 1 .
they are correlated and their covariance matrix has to be taken into account when calculating the errors of @xmath182 and @xmath148 . the equations ( [ eq32a ] ) and ( [ eq33 ] ) also show that @xmath182 and @xmath148 can be obtained without knowning which spectra @xmath121 and @xmath155 were used to compute the @xmath192 and @xmath193 .
in this section an experimental energy distribution of @xmath0-rays from the crab nebula is unfolded , which was obtained in an analysis of data taken with the magic telescope @xcite .
the migration matrix @xmath13 , as determined from a sample of @xmath0-mc events , is plotted in fig .
[ fig : crabdata]a ) as a function of @xmath4 and @xmath3 .
the size of the boxes is proportional to the value of @xmath148 , where @xmath11 and @xmath10 are the bin numbers in @xmath4 and @xmath3 respectively .
the experimental distribution @xmath14 of the number of @xmath0-excess events as a function of @xmath4 is displayed in fig .
[ fig : crabdata]b ) .
both distributions are after all cuts and selections , except a cut in @xmath4 . the vertical and horizonthal lines in figs .
[ fig : crabdata]a ) and c ) indicate the ranges in @xmath4 or @xmath3 , which were selected for the unfolding .
the range in @xmath4 is given by those @xmath4 bins for which a significant number of excess events could be determined .
the range in @xmath3 comprises all those @xmath3 bins which are expected to contribute to the selected range in @xmath4 .
[ fig : crabdata]a ) suggests that these are all 15 @xmath3 bins .
however , according to the plot of the effective collection area @xmath194 in fig . [
fig : crabdata]c ) the contribution from @xmath195 gev is expected to be negligible .
the same holds for @xmath196 tev , due to the strongly decreasing @xmath0-ray flux with increasing energy .
this is confirmed by a forward unfolding of @xmath14 in which a differential @xmath0-ray flux of the form @xmath197 , with @xmath198 , was assumed .
this leads to a number of @xmath4 and @xmath3 bins of @xmath199 and @xmath200 respectively , which are used in the unfolding procedure .
the size of the @xmath201 bins was deliberately chosen wider than the @xmath202 bins by a factor of 1.4 , in order to better constrain the unfolding .
the rank of gram s matrix @xmath7 is equal to @xmath203 , as can be seen from fig .
[ fig : crabdata]d ) , which shows the size of the eigenvalues @xmath67 of @xmath7 as a function of @xmath204 .
two to three of these eigenvalues are much smaller than the maximum eigenvalue , and they are the reason for the large values of @xmath138 at large iteration number ( low regularization strength ) in fig .
[ fig : iterbertero]b ) .
the optimum regularization strength and thus the final solution @xmath2 was determined using the criterion @xmath205 .
the estimates @xmath206 ( open circles ) of @xmath47 are compared with @xmath47 ( histogram ) in fig .
[ fig : crabdata]b ) .
these data enter in the calculation of @xmath19 ( eq . [ eq : chi20 ] ) .
the number of degrees of freedom in the unfolding procedure is @xmath207 , because the number of measurements is @xmath8 , the number of unknowns is @xmath9 , and the relation @xmath208 ( eq . [ eq : addconstraint ] ) is used as additional constraint . as can be seen from figs .
[ fig : iterbertero]d ) and e ) the values of @xmath139 and @xmath140 are much lower at the selected regularization strength than without regularization .
this means that the solution @xmath2 is smoothed by the regularization .
the black symbols in fig . [
fig : crabresults]a ) represent the final solution @xmath2 .
the solution before correctiong for the cut in @xmath4 is drawn with grey symbols .
+ the final differential @xmath0-ray flux @xmath165 , as computed from @xmath2 according to eq.([eq : phik ] ) , is drawn in fig .
[ fig : crabresults]b ) .
the solid line represents the result of a fit of the expression @xmath209 , with @xmath198 , to the data points .
the number of degrees of freedom for this fit is @xmath210 , because the number of data points is @xmath200 and the number of free parameters is 3 ( @xmath211 , @xmath212 and @xmath213 ) .
the @xmath66 is 8 for 7 degrees of freedom .
setting @xmath214 in the fit yields a @xmath66 of 24 for 8 degrees of freedom .
this fit is clearly disfavoured as compared to the fit in which the slope @xmath215 is energy dependent . in these fits the full correlation matrix @xmath46 of @xmath2 has been taken into account .
the result of the latter fit for @xmath121 was used to recalculate the averages @xmath169 ( [ eq60 ] ) and @xmath157 ( [ eq44x ] ) for the individual @xmath3 bins . the unfolding was repeated using the recalculated @xmath169 and @xmath157 , yielding new results for @xmath2 , @xmath165 and the fit parameters .
after 1 iteration this procedure converged .
very similar results were obtained with the other unfolding methods .
the spread of the @xmath45 , obtained with the different unfolding methods , can be seen in fig .
[ fig : crabresults]c ) . this spread can be regarded as an estimate of the a systematic error due to the unfolding .
a widely used method of correcting experimental distributions is the application of correction factors : using monte carlo data , both the true distribution @xmath216 and the reconstructed distribution @xmath217 of some quantity are produced under certain conditions ( selections , cuts ) .
correction factors are determined according to @xmath218 an experimental distribution @xmath219 , obtained under the same conditions as @xmath220 , is then corrected by @xmath221 to obtain the corrected distribution @xmath100 .
the following properties of this procedure can be stated @xcite : * @xmath222 is undefined if @xmath223 . *
if @xmath224 also @xmath225 and @xmath226 , which means that @xmath219 is ignored .
* @xmath222 depends on the shape of the mc distribution @xmath227 ; the corrected distribution @xmath100 is always biased towards @xmath227 . *
if @xmath228 also @xmath100 is zero . * the standard linear error propagation often yields too small errors of @xmath100 .
the correction factors are only right if @xmath227 is identical to the true @xmath100 distribution .
if this is not the case one may iterate @xmath227 , setting @xmath227 equal to the last corrected experimental distribution @xmath100 . however
, this often leads to instabilities .
the reason for the instabilities appears to be similar to that causing a large noise component of the direct solution ( [ eq : s0 ] ) .
in contrast to the unfolding methods presented in section [ section : regularization ] , there is very little freedom in choosing the binnings for @xmath2 and @xmath14 . by definition ,
the range of values and the binnings for the true and reconstructed quantity are identical .
advantages of the method of correction factors are that it is simple and stable .
the drawbacks have been listed above , the severest one being the strong dependence of the correction factors on the assumptions made in the mc about @xmath2 .
in this paper the procedures to unfold experimental energy distributions of @xmath0-rays , as applied in the magic experiment , are described .
it is explained , how the uncertainties , which are inherent in any unfolding process , can be handled successfully .
possible problems in the unfolding are discussed and suggestions are given which can help to avoid them .
various techniques are presented , which allow to reconstruct the energy spectrum in a rather unbiased way .
all algorithms are impleneted in the magic software , which is based on the @xmath229 language and root @xcite .
their application to real data has shown to provide robust and reliable results .
the methods and procedures are applied in most of the magic analyses .
we thank michael schmelling for fruitful discussions and critical comments . 10 baixeras , c. et al . , 2004 ,
instrum . meth . , a518 , 188 .
r. gold , anl-6984 ( 1964 ) g.i .
marchuk , `` methods of numerical mathematics '' , springer , berlin ( 1975 ) a.n .
tikhonov and v.ja .
arsenin , methods of solution of ill - posed problems - m ( nauka , 1979 ) s.w .
provencher , computer physics communications 27 ( 1982 ) 213 and 229 v. blobel , `` unfolding methods in high - energy physics experiments '' , desy 84 - 118 ( 1984 )
. v. blobel , 1984 cern school of computing , ajguablava , spain , cern 85 - 09 ( 1984 ) p.88 .
belogorlov et al . , `` interpretation of the solution to the inverse problem for the positive function and the reconstruction of neutron spectra '' , nim a 235 ( 1985 ) 146 . s.f .
giljazov , `` methods of solution of linear ill - posed problems '' , msu , moscow ( 1987 ) v.p .
zhigunov et al . , `` on estimating distributions with the maximum entropy principle '' , nim a 273 ( 1988 ) 362 m. bertero , infn / tc-88/2 ( 1988 )
. m. bertero , advances in electronics and electron physics , vol.75 ( 1989 ) .
anykeyev et al . , nim a 303 ( 1991 ) 350 .
m. schmelling , `` the method of reduced cross - entropy . a general approach to unfold probability distributions '' , nim a 340 ( 1994 ) 400
. v. blobel , `` the run manual '' , opal technical note tn361 ( 1996 ) a. hcker and v. kartvelishvili , nim a 372 ( 1996 ) 469 .
m. schmelling , `` numerische methoden der datenanalyse '' , mpi - k heidelberg ( 1998 ) v. blobel , `` an unfolding method for high energy physics '' , ippp workshop on advanced statistics techniques in particle physics , durham ( 2002 ) w. wittek,``correlations between parameters of extended air showers and their proper use in analyses '' , 26th int .
cosmic ray conference , salt lake city , utah , usa ( 1999 ) majumdar , p. et al .
( magic collab . ) , 2005 , proc . of the 29th icrc , pune , india , 5 - 203 , astro - ph/0508274 .
mizobuchi , s. et al .
( magic collab . ) , proc .
29th icrc , pune , india , 5 - 323 , astro - ph/0508274 .
bretz , t. and wagner , r. ( magic collab . ) , 2003 , proceedings of the 28@xmath230 icrc , tsukuba , japan , 2947 .
kneiske et al . , a&a 413 ( 2004 ) 807 j. albert et al . , submitted for publication in apj ( 2007 ) , arxiv:0705.3244 .
r. brun , f. rademakers , http://root.cern.ch/. | the paper describes the different methods , used in the magic experiment , to unfold experimental energy distributions of cosmic ray particles ( @xmath0-rays ) .
questions and problems related to the unfolding are discussed .
various procedures are proposed which can help to make the unfolding robust and reliable .
the different methods and procedures are implemented in the magic software and are used in most of the analyses . , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
solar and stellar irradiance records are often plagued by data gaps .
the proper interpolation of these missing data is a longstanding and notoriously delicate problem that requires a good understanding of the data @xcite .
considerable attention has been given to this problem in fields such as climate science @xcite but much less so in solar physics and in astrophysics .
often , the limited attention that is paid to data gaps contrasts with the sophistication of the analysis that is subsequently performed on these data .
while short gaps can easily be filled by linear or by nonlinear interpolation , data gaps whose duration exceeds the characteristic time scales are much more difficult to handle .
a notable exception is when multichannel synoptic observations of the same process are available , with gaps in some or in all of them .
spectral irradiance observations , which we shall concentrate on , precisely belong to that category .
our examples will be taken from the sun , but the results can be easily extended to other types of multichannel observations .
our method applies to any set of observations that are recorded simultaneously ( i.e. the time stamps are the same for all records ) , are correlated with each other , and whose time intervals fully or partly overlap .
our main assumption is their linear correlation , in the sense that each record can be approximated by a linear combination of the other ones .
a strong linear correlation is typically observed between spectral irradiance observations made at different wavelengths or between simultaneous measurements of different proxies .
these synoptic records are frequently used to assess subtle changes in the variability of the sun ; they are often remarkably coherent in time @xmath0 and in wavelength @xmath1 . as a consequence
, their variability can be explained in terms of a few contributions only .
this property is well known for the extreme ultraviolet ( euv ) @xcite but also for the visible range @xcite , when measured from space .
the same coherency is observed among different proxies for solar activity @xcite .
this property is rooted in the structuring effect of the solar magnetic field .
the coherency partly breaks down during the impulsive phase of solar flares because the spectrum then considerably depends on the local conditions of the solar atmosphere . here
, however , as in many applications , we consider daily or hourly averages , so that the effect of short transients can be discarded .
this coherency in both time and wavelength is the key to the reconstruction technique we shall introduce below . by interpolating along two dimensions ( in time and along different records ) ,
we not only improve the quality of the reconstruction , but we also can fill arbitrarily large data gaps without having to rely on the tedious bookkeeping that is required by most interpolation schemes . the nonparametric and data - adaptive method we advocate is based on the svd or singular value decomposition @xcite , which is to linear algebra what the fourier transform is to spectral analysis .
the svd allows the extraction of the coherent part of the solar spectral irradiance , which is then used to fill the data gaps iteratively .
the method is described in section 2 , and two applications are detailed .
the first one ( sec .
3 ) deals with solar spectral irradiance data in the euv . in the second application ( sec .
4 ) we consider a set of solar proxies with numerous gaps .
let @xmath2 be a multichannel record that represents either the solar spectral irradiance at different wavelengths ( or in different spectral bands ) or a set of solar proxies , or a combination thereof .
all these quantities must be sampled simultaneously ; the sampling rate , however , does not need to be constant .
these data are conveniently stored in a matrix @xmath3 $ ] , in which columns are time series .
each column may have an arbitrarily large number of data gaps , as long as a reasonable fraction of observations are available , say at least 20% .
the method we propose exploits either the coherency in wavelength , or both the coherency in wavelength and in time .
we start with a description of the first option , because the second one can be readily obtained by data embedding .
let us first assume that there are no gaps .
the svd of the data matrix then yields a separable set of functions ( hereafter called _ modes _ )
@xmath4 which are orthonormal @xmath5 the weights @xmath6 are positive by construction .
the number @xmath7 of modes equals the rank of the matrix , which is usually the smallest of the number @xmath8 of samples or the number @xmath9 of records .
this decomposition is unique .
the svd of the data matrix directly yields a set of three matrices @xmath10 that respectively contain @xmath11 , the weights @xmath12 , and @xmath13 .
a key property of the svd is that modes with heavy weights describe salient features of the data .
that is , the truncated expansion @xmath14 will capture the coherent part of the data while deferring incoherent fluctuations to the remaining modes .
this property has made the svd popular in multichannel and array data processing @xcite .
we shall use it here to reconstruct the missing values .
the performance of the reconstruction can be quantified by the mean square error @xmath15 which shows that by taking the few largest modes , the reconstruction error can be made arbitrarily small .
as it turns out with spectral irradiance data , the first few weights are often orders of magnitude heavier than the subsequent ones , so that excellent reconstructions can be achieved with a few modes only .
we implicitly assume here that features departing from the behaviour observed at other wavelengths are unlikely to have a solar origin ( except during the impulsive phase of flares ) , so that they can be readily discarded .
this will be illustrated below in sec .
3 . let us now assume that some samples are missing .
the data covariance matrix and the svd then can not be computed anymore .
this problem , however , can be circumvented by using the following iterative scheme with two embedded loops : 1 .
fill each gap with some adequate value ( typically the temporal mean of the record ) .
2 . compute the svd .
3 . compute the approximation @xmath16 of the data by retaining the @xmath17 largest mode(s ) of the svd .
initially , @xmath18 .
4 . fill the gaps with @xmath16 , as defined in eq .
3 . as long as these values have not converged , go back to 2 .
( inner loop ) .
increment the number of modes @xmath17 and start again at 2 .
iterate until @xmath19 ( outer loop ) .
this method seems to have emerged independently in different contexts @xcite ; it has mostly been used for spatio - temporal data sets , with some subtle differences @xcite .
we refer to @xcite for discussions on optimality , convergence , etc .
three additional adaptations , however , need to be considered before the method can be applied to irradiance data . the relative variability and the average value of the solar spectral irradiance vary by orders of magnitude between the soft x - ray and the visible range .
the svd , however , is scaling - dependent and so a renormalisation is required .
we do so by standardising each record : first , the time average is subtracted and then a normalisation with respect to the standard deviation @xmath20 or the noise level ( if known ) is performed .
both operations are affected by the value of the missing samples , so they must be repeated at each iteration .
this is particularly important for the offset subtraction .
the renormalisation may be done only once .
the solar spectral variability contains a mix of scales that are driven by different processes : 27-day variations are due to solar rotation , the 11-year periodicity is caused by the solar cycle , etc .
each of these processes leads to a specific spectral dependence ; different scales should therefore be processed separately when filling gaps .
this feature considerably improves the reconstruction skill and to the best of our knowledge has not yet been used .
the two ranges of scales that are most frequently encountered in solar studies are : below 81 days ( which captures solar rotation and the evolution of active regions ) and above 81 days .
we apply the iterative svd procedure described in sec .
2.1 separately to both scales .
the _ trous _ wavelet transform @xcite is used to decompose the data into two records at each iteration : one with short time - scales and one with long time - scales .
classical bandpass filters may also be used because this has no significant impact on the results .
the wavelet transform , however , is better suited for non - stationary data .
one may also want to extract additional scales , such as the 13-day periodicity associated with centre - to - limb effects of hot coronal lines .
this indeed results in a small but discernible improvement in the reconstruction of the euv , at the expense of a longer computation time .
the methodology so far only exploits the coherency between different wavelengths ( or proxies ) , which is the key property .
one may also want , however , to make use of the temporal coherency .
this is useful when there are specific times at which there is no single observation , or if the number of records is small ( typically @xmath21 ) , or if each record can be considered as a smoothly varying waveform with incoherent noise superimposed on it .
the main asset of the iterative svd reconstruction method is its straightforward extension to such a filtering in time , using the concept of embedding , which has been pioneered in the study of chaotic systems by @xcite .
let us expand the data matrix by appending replicates that are shifted in time , i.e. @xmath22 .\ ] ] by applying the svd to this embedded matrix we exploit both the coherency in wavelength and in time .
it is important ( but not mandatory ) that the data be regularly sampled since the method essentially computes a weighted average of each sample with its nearest neighbours . the higher the value of the embedding dimension @xmath23 , the more adjacent time steps are used in the reconstruction , thus leading to a stronger smoothing in time .
this is equivalent to using a symmetric finite - impulse filter whose coefficients are obtained data - adaptively .
the particular case wherein one single record is embedded and decomposed by svd is called singular spectrum analysis ( ssa ) . in the ssa ,
only temporal information is used and so it is important for the embedding dimension to exceed the value of the dominant period in the data @xcite .
our reconstruction , however , mostly relies on the strong coherency across wavelengths or proxies to fill the gaps and so the conditions on the value of the embedding dimension are much less stringent . in practice ,
low values ( @xmath24 ) already bring a significant improvement .
the main reason for keeping this dimension as low as possible is to reduce the computational load .
the three tuneable parameters of the method are : a ) the number @xmath25 of significant modes , b ) the number of scales into which the data are decomposed , and c ) the embedding dimension @xmath23 .
only the first one really affects the outcome .
a separation into two scales only ( with a threshold between 50 - 100 days ) is enough to properly capture both short- and long - term evolutions , and embedding dimensions of @xmath24 are usually adequate for reconstructing daily averages .
the determination of the optimum parameters and the validation of the results is made by cross - validation and will be illustrated below .
the only critical question is memory and computational load .
for an irradiance data set with five years of daily values at 100 wavelengths , and an embedding dimension of @xmath26 , the size of the embedded matrix is @xmath27 $ ] .
the computation of the svd at each iteration typically takes several seconds .
for that reason , it may be desirable to process separately those spectral bands that evolve differently , such as the soft x - ray , the euv and the muv bands .
the routine in matlab@xmath28 is available from the author .
the solar euv monitor ( sem ) is a solar extreme ultraviolet ( euv ) spectrometer that has been operating continuously on the soho satellite since january 1996 @xcite . in its first - order mode , sem measures the irradiance within an 8 nm bandpass centred about the bright 30.38 nm he ii line . on june 25 , 1998 , soho suffered a mission interruption , leading to the loss of several months of data .
this long data gap considerably complicates the use of sem data for upper atmosphere model validation .
the sem , however , mostly captures chromospheric emissions , which are highly correlated with other gauges of solar activity .
foremost among these are : * the @xmath29 or decimetric index , which is the solar radio flux at 10.7 cm .
this index , which is measured from the ground , captures a mix of thermal and electron gyro - resonance emissions , and has been shown to be highly correlated with the euv flux @xcite . * the mg ii index , which is the core - to - wing ratio of the mg ii line at 280 nm .
this index is widely used as a proxy for chromospheric activity @xcite . *
the intensity of the h i lyman @xmath30 line at 121.57 nm , which is the brightest spectral line below 200 nm @xcite . together with the flux from the sem , we have four quantities that have different physical origins and yet are highly correlated , thereby opening the prospect of filling the large gaps in the sem data .
we consider daily averages made from january 1 , 1996 until april 29 , 2011 .
the linear correlation between the @xmath29 index and the other proxies improves when taking its square root , which we shall systematically do from now on .
the correlation between these four proxies on both long and short time - scales is illustrated in fig .
[ fig_sem_excerpt ] .
our working hypothesis is that each of the missing samples from the sem can be reconstructed from a linear combination of ( possibly non - simultaneous ) observations of the other proxies . as we shall see shortly ,
the best value of the embedding dimension is 4 ; let us therefore select @xmath31 and first determine the optimum number of modes . with four variables and an embedding dimension of 4 , the total number of svd modes is 16 ; their weights are displayed in fig .
[ fig_sem_k ] .
the first weight surpasses all the others because the first mode is an average of all four proxies , which is by far the most conspicuous coherent feature .
the inflexion point between the few heaviest weights and the flat tail provides a convenient but visual criterion for determining the number of significant modes @xcite . according to this criterion , the best interpolation skill is for @xmath32 modes out of 16 . a better validation test consists in generating a small number of synthetic gaps , reconstructing them , and then checking how the residual error varies with the model parameters .
to do so , we remove 5 - 10 % of the samples from each record and then compute the normalised error @xmath33 where the average is computed for synthetic gaps only .
this procedure is repeated ten times to obtain an estimate of the average value of the normalised error .
a value of 100 % can be interpreted as an error whose standard deviation equals the solar cycle variability of the original data .
this value truly reflects the error made by filling short data gaps .
note that it tends to underestimate the error for larger gaps , unless the length distribution of the synthetic gap matches that of the original data .
the evolution of the normalised error with @xmath25 is illustrated in fig .
[ fig_sem_k ] , which shows a broad minimum around @xmath34 , in agreement with the estimate obtained by visualisation .
note that the four minima occur at different values of @xmath25 .
the normalised error is on average larger for the @xmath35 index , which suggests that this quantity is relatively more difficult to reconstruct than the others .
this is not so surprising , because it is the only emission from the radio band .
the smallest normalised error is obtained for the sem , with @xmath36 .
this value is about half that of the estimated normalised uncertainty @xcite , which shows the excellent quality of the reconstruction . in practice , the optimum value of @xmath25 is frequently found to be one or two units higher than the value obtained by visual inspection . as fig .
[ fig_sem_k ] suggests , an overestimation of @xmath25 is preferable to an underestimation .
the choice of the embedding dimension @xmath23 is mostly based on physical insight . with @xmath37 ( i.e. no embedding ) we assume that the missing samples are reconstructed from simultaneous observations only , whereas @xmath38 implies that the information contained in past and future observations is also used . setting @xmath38
therefore involves a weighted averaging over time , which is appropriate for records whose samples are highly correlated in time . in fig .
[ fig_sem_d ] we estimate the normalised error for different embedding dimensions , using the optimum number of modes for each of them .
the smallest error is obtained for an embedding dimension of @xmath31 .
larger dimensions hardly reduce the error but do increase the computational load substantially
. as expected , the higher the value of @xmath23 , the smoother the reconstruction and the more likely that fine features may be missed .
this is particularly evident in august 1998 , when a group of rapidly evolving active regions were moving across the solar disc .
an embedding dimension of 4 properly captures their evolution , whereas a dimension of 15 smears out all but the most pronounced peaks .
this example illustrates a relatively simple case because only one record has gaps in it .
let us now , however , consider a more frequent case in which several of the records have large gaps .
filling these gaps by standard interpolation schemes can become very time - consuming because of the amount of bookkeeping that is required to test whether gaps occur simultaneously in several records , etc .
the svd - based interpolation does not require any of these tests .
the ca k index is the normalised intensity of the ca ii k - line at 393.37 nm and has been advocated as a proxy for magnetic activity , including plages , faculae , and the network .
this line is measured from the ground , so it can not be observed continuously .
here we consider a record of daily observations made at the national solar observatory at sacramento peak @xcite , in which about 66% of the samples are missing .
this index is known to be highly correlated with other solar indices , in particular with the mg ii index @xcite , so that the svd method is ideally suited for filling its gaps . to reconstruct the missing values
, we consider the following set of proxies that are highly correlated with the ca k index : the square root of the @xmath29 index , the intensity of the h i lyman @xmath39 line , the mg ii index and the magnetic plage strength index ( mpsi ) @xcite .
the time interval ranges from nov . 1
, 1980 to july 1 , 2010 ; all proxies have data gaps except for the first two .
these gaps occur erratically and 6% of them exceed 10 days . in this particular example
, the coherency between proxies is crucial and indeed the choice of the embedding dimension @xmath23 does not significantly affect the results .
let us take @xmath40 , which is the value that is recommended by the reconstruction error .
the maximum number of svd modes is 10 because we have five records . out of these ,
three only are found to be significant .
the result of the reconstruction is illustrated in fig .
[ fig_cak1 ] for periods of high and low solar activity .
note that the results obtained with different number of modes lead to similar temporal evolutions .
the reconstruction at solar maximum looks reasonable because it passes through the observations while staying highly correlated with the mg ii index . during solar minimum , however , the observed values of the ca k index continue to fluctuate whereas the reconstructed values and the other proxies stay almost constant .
the difference between the observed ca k index and the smoothly varying reconstruction varies randomly in time , which questions its solar origin . to further investigate the origin of this difference between the observed and reconstructed index , we filtered the reconstructed data with the _ trous _ wavelet transform , which allows the separation of the sharp peaks from the more regular reconstruction .
the residuals , i.e. the difference between the filtered reconstruction and the original observations , are shown in fig .
[ fig_cak2 ] : they are found to be independent and their gaussian distribution only weakly varies with the solar cycle .
this is a strong indication that the residuals are measurement errors rather than solar fluctuations .
their standard deviation is 0.0008 , which represents 20 % of the solar cycle variability of the ca k index .
our reconstruction thereby provides a means for fitting the numerous data gaps in the ca k index while also evaluating the confidence interval of the observations .
this study shows that svd - interpolation is a powerful technique for filling arbitrarily large gaps in multi - wavelength , multichannel or in synoptic records .
we focused here on solar spectral irradiance observations , which are frequently plagued by missing data .
these gaps may be distributed at random in time or in wavelength .
the main tuneable parameter is the number of svd modes that is needed to reconstruct the data ; this value may be estimated either by visualisation or by cross - validation .
the method works best when each record can be approximated by a linear combination of the others .
since it relies on linear combinations only , it may be desirable to apply a nonlinear static transform beforehand to increase the linear correlation between the records . for the method to work ,
the observations must be sampled simultaneously but not necessarily evenly .
non - simultaneous observations can be handled by resampling all variables to a common grid , for example by fourier decomposition ( e.g. * ? ? ?
* ) , and then filling the gaps by svd . by alternating between the two ,
both the gaps and the interpolated values can be progressively refined .
this method has several applications in addition to mere interpolation .
the first one is the cross - calibration of measurements of the same quantity by different instruments .
the mg ii index , for example , is at present measured by different instruments that give different amplitudes .
these data sets are incomplete and only partly overlap , which considerably impairs their inter - comparison . the iterative svd method is ideally suited for filling these gaps because the records are by definition strongly correlated .
a second potential application is the stitching together of total solar irradiance ( tsi ) observations . merging tsi records from several instruments
is a delicate and controversial task @xcite because instruments disagree on the absolute value of the tsi and often do not operate simultaneously .
the iterative svd provides a means for estimating the different offsets in a self - consistent way because it allows us to extrapolate each tsi record by assuming that its statistical properties with respect to the other records do not change in time .
this property is particularly useful for checking composites that are built from different records , such as the tsi , the h i lyman @xmath30 intensity , the mg ii index and the sunspot index @xcite .
this will be detailed in a forthcoming publication .
i thank the following institutes for providing the data : the laboratory for atmospheric and space physics ( university of colorado ) for the mg ii and h i lyman @xmath30 composites , the national solar observatory at sacramento peak ( data produced cooperatively by nsf / noao , nasa / gsfc and noaa / sec ) for the ca k index , the mount wilson observatory ( operated by ucla , with funding from nasa , onr and nsf , under agreement with the mt .
wilson institute ) for the mpsi index and the space sciences center ( university of southern california ) for the sem data .
this study received funding from the european community s seventh framework programme ( fp7/2007 - 2013 ) under the grant agreement nr .
218816 ( soteria project , http://www.soteria-space.eu ) .
, s. l. , henry , t. w. , and fleck , b. ( 1998 ) . .
in balasubramaniam , k. s. , harvey , j. , and rabin , d. , editors , _ synoptic solar physics _ , volume 140 of _ astronomical society of the pacific conference series _ , pages 301308 . | data gaps are ubiquitous in spectral irradiance data , and yet , little effort has been put into finding robust methods for filling them .
we introduce a data - adaptive and nonparametric method that allows us to fill data gaps in multi - wavelength or in multichannel records .
this method , which is based on the iterative singular value decomposition , uses the coherency between simultaneous measurements at different wavelengths ( or between different proxies ) to fill the missing data in a self - consistent way .
the interpolation is improved by handling different time scales separately .
two major assets of this method are its simplicity , with few tuneable parameters , and its robustness .
two examples of missing data are given : one from solar euv observations , and one from solar proxy data .
the method is also appropriate for building a composite out of partly overlapping records . |
gamma - ray bursts ( grbs ) are rapid , powerful flashes of gamma - ray radiation occuring at an average rate of one event per day over the whole sky at cosmological distances .
the high energy prompt emission is followed by a broadband ( x - rays to radio ranges ) fading emission , named afterglow , ( costa et al .
1997 ; van paradijs et al . 1997 ; frail et al . 1997 ; bremer et al . 1998 ; heng et al .
2008 ) that can be observed up to weeks and months after the onset of the event . the distribution of grb durations observed by the batse instrument ( fishman et al .
1989 ) is bimodal , with peaks at @xmath2 and @xmath3s and a boundary at @xmath4 s ( kouveliotou et al .
1993 ) is defined as the time during which the cumulative counts increase from 5% to 95% above background , adding up to 90% of the total grb counts . ] .
the two classes of long ( lgrbs , @xmath5 s ) and short grbs ( sgrbs , @xmath6 s ) , show substantial evidences for different origins . while it has been firmly established that lgrbs , or at least a significant fraction of the nearby events ( with redshift @xmath7 ) for which it has been possible to search for the presence of a supernova ( sn ) ,
are associated with core - collapse explosions of massive stars ( see hjorth & bloom 2011 , for a recent review ) , the nature of sgrb progenitors is still under debate .
current models suggest that they are associated with the merging of compact objects in binary systems , such as a double neutron star ( ns ) , or a ns and a black hole ( bh ) system ( eichler et al .
1989 ; narayan et al .
1992 ; nakar 2007 ; berger 2014 ) .
such systems can originate from the evolution of massive stars in a primordial binary ( narayan et al .
1992 ) or by dynamical interactions in globular clusters during their core collapse ( grindlay et al . 2006 ; salvaterra et al .
2008 ) . very recently
, a direct evidence supporting the merger scenario has been claimed by tanvir et al .
( 2013 ) and berger , fong & chornock ( 2013 ) who reported the possible detection of a kilonova associated to the sgrb130603b ( but see jin et al .
2013 for further discussion ) .
sgrbs are not distinguished from lgrbs only by their duration . if we consider the observed prompt emission , negligible spectral lag ( norris , marani & bonnell 2000 ; norris , scargle & bonnell 2001 ) and harder spectra ( kouveliotou et al . 1993 ) are common for sgrbs . on the other hand ,
the prompt emission properties of short grbs are similar to the first 12 s of long events ( ghirlanda , ghisellini & celotti 2004 ) and both classes of objects show a similar spectral evolution ( ghirlanda , ghisellini & nava 2011 ) .
this might suggest a common emission mechanism for both long and short grbs .
since 2005 , with the advent of the fast - repointing _ swift _ satellite ( gehrels et al .
2004 ) , the discovery of sgrbs afterglows and the identification of their host galaxies made possible to study their distances , energy scales and environments ( gehrels et al . 2005 ) .
sgrbs are found to be typically less energetic ( their isotropic equivalent energy , @xmath8 , is of the order of @xmath9 ) than lgrbs and to occur at a lower redshift ( nakar 2007 ; berger 2011 ; fong et al .
their afterglows tend to be significantly fainter on average than those of lgrbs ( kann et al .
2011 ; nicuesa guelbenzu et al .
2012 ; margutti et al . 2013 ) , but scaled to their @xmath8 similar to faint lgrbs ( nysewander et al .
2009 ) . concerning the host galaxies , sgrbs occur in both early and late type galaxies with low star formation rate and are associated with an old stellar population ( berger 2009 ; leibler & berger 2010 ; fong et al .
a different origin for sgrbs with respect to the lgrb class is also supported by the lack of detection of the underlying sn in the light curves of their optical afterglows down to very stringent magnitude limits ( hjorth et al .
2005a ; hjorth et al .
2005b ; fox et al . 2005
; covino et al . 2006 ; kann et al .
2011 ; davanzo et al . 2009 )
and by their inconsistency with the correlation , valid for lgrbs , between the rest frame spectral peak energy and @xmath10 ( @xmath11 correlation ; amati et al .
2008 and references therein ) . on the other hand , ghirlanda et al .
( 2009 ) showed that sgrbs are consistent with the same @xmath12 correlation ( where @xmath13 is the prompt emission isotropic peak luminosity ) defined by lgrbs ( yonetoku et al .
sgrbs are also consistent with the three parameter @xmath14 correlation ( with @xmath15 being the afterglow energy emitted in the soft x - ray band ; bernardini et al .
2012 ; margutti et al . 2013 ) .
furthermore , the distributions of the intrinsic x - ray absorbing column densities of long and short grbs do not show significant differences when compared in the same redshift range ( @xmath7 ; kopac et al .
2012 ; margutti et al . 2013 ) .
finally , given that the measured duration of the grb prompt emission can vary for different instrument ( e.g. due to the different energy band used ) , it has been recently proposed that the value of @xmath16 used to divide the long and short grbs should be reduced to about 0.8 s for the _ swift _ bursts ( bromberg et al .
a recent review of the properties of sgrbs has been presented by berger ( 2014 ) .
the majority of the studies reported above is based over the entire sample of sgrbs with measured redshifts .
although this approach has the clear advantage of describing the intrinsic physical properties of these objects , it can be severely affected by observational biases , given that almost 3/4 of the _ swift _ sgrbs are lacking a secure redshift measurement ( berger 2014 ) .
in this paper , we present a carefully selected sub - sample of the _ swift _ sgrbs .
inspired by the criteria we followed to build a complete sample of lgrbs ( the bat6 sample ; salvaterra et al .
2012 and references therein ) , we selected sgrbs with favorable observing conditions for redshift determination from the ground , ending up with a sample of 36 events , almost half of which with a redshift measure .
the redshift completeness increases up to about 70% by restricting to those events that are bright in the 15 - 150 kev _ swift _
burst alert telescope ( bat ; barthelmy et al .
2005 ) energy band , obtaining a sample of 16 sgrbs , which is complete in flux and has the highest completeness in redshift with respect to the sgrb samples presented in the literature to date .
such flux - limited sample provides a robust base for the study of the energetics , redshift distribution and environment of the _ swift _ population of sgrbs .
a statistical study of these properties is a useful tool to indirectly investigate their elusive progenitors , find additional parameters for the grb classification ( that can go beyond the duration of the prompt emission ) and check for the existence of possible sub - classes . the paper is organized as follows . in section 2
we describe the sample and the selection criteria and in section 3 the data analysis methods .
the observed and rest - frame properties of the grbs of the sample are discussed in section 4 .
our conclusions are presented in section 5 . throughout the paper
we assume a standard cosmology with @xmath17 and @xmath18 .
errors are given at the 68% confidence level unless stated differently .
we selected our sample among all the _ swift _ grbs with @xmath19 s and adding those classified as sgrbs in light of their prompt emission hardness ratio and absence of a spectral lag . with these criteria
we could include in our sample also sgrbs whose _
swift_-bat light curve shows a short - duration peak followed by a softer , long - lasting tail ( the so - called `` extended emission '' ) . for these events , @xmath16 can be longer than 2 s. we then selected : 1 ) all the events promptly re - pointed by the _ swift _ x - ray telescope ( xrt ; burrows et al .
2005 ) and 2 ) with favorable observing conditions for ground - based optical follow - up aimed at redshift determination , i.e. events with low galactic extinction in the direction of the burst ( @xmath20 mag ) .
such criteria are inspired , although less tight , to the ones used to built the tough and bat6 sample for lgrbs ( jakobsson et al .
2006 ; hjorth et al . 2012 ; salvaterra et al .
in particular we note that , in order to avoid an x - ray selected sample , we do not require a detection of the x - ray afterglow ( but just a prompt observation with the _ swift_-xrt ) .
the sample built this way ( the _ total _ sample from now on ) consists of 37 sgrbs , with a completeness in redshift of @xmath21 ( table 1 ) .
this _ total _
sample includes 5 sgrbs with extended emission ( ee ) and 19 with short - lived x - ray afterglow ( sl ) after the trigger is less than @xmath22 . ] .
we restricted this sample to those events with a peak photon flux @xmath23 ph s@xmath24 @xmath25 , computed using the 15150 kev _
swift-_bat light curves binned with @xmath26 ms .
this further criterium selects 17 sgrbs , 12 with a measured redshift , providing a sample that has , at the same time , a size large enough to perform statistical studies and a high level of redshift completeness ( 71% ) .
an analogous , although less tight , cut was used in salvaterra et al .
( 2012 ) to built the bat6 sample of lgrbs . being free of selection effects ( except for the flux limit ) , this sample ( the _ complete _ sample from now on ) , although relatively small , provides a useful benchmark to compare the rest - frame physical properties of sgrb prompt and afterglow emission .
[ cols="^,^,^,^,^,^,^,^,^ " , ] [ tab : log_corr_rf ] as a general trend , we note that the afterglow x - ray luminosity at early times ( @xmath27 min ) correlates with the prompt emission quantities @xmath8 , @xmath28 and @xmath29 with null probabilities of @xmath30 and dispersion @xmath31 dex . at later times ( @xmath32 , 11 and 24 hr )
the scatter increases and these correlations become much less significant . the early time
prompt - afterglow correlation we find ( table [ tab : log_corr_rf ] ) are rather similar to the same correlations found for the bat6 sample of lgrbs ( davanzo et al .
in particular , the early time @xmath33 correlations for short and long grbs from the two samples have the same slope , with the sgrbs lying on the faint end of the correlation ( in agreement with what found by nysewander et al . 2009 ) . to compare them qualitatively ,
we show in fig .
[ fig : correlations ] the time resolved prompt - afterglow correlations for the sgrbs of our _ complete _ sample and for the lgrbs of the bat6 sample . concerning the @xmath34 plane we note that at all times , assuming the same @xmath28 , sgrbs have on average lower x - ray luminosity with respect to lgrbs .
however , we note that the systematics in the procedure of @xmath28 estimate for sgrbs discussed in sect .
4.2.1 for the @xmath35 relation might amplify this effect .
similarly , compared to lgrbs , for a given @xmath29 , sgrbs are found at an average lower x - ray luminosity .
considering the correlation found between @xmath36 and @xmath8 , such a result can be considered a natural consequence of the inconsistency of sgrbs with the @xmath37 correlation ( fig .
3 , upper panel ) .
another feature arising from the comparison of the short and long grb samples is that the sgrbs prompt - afterglow correlations show a scatter much larger than what observed for lgrbs , especially at late times ( @xmath38 hr ) .
although this effect can be related to the lower number of data points in the sgrb sample , it may suggest that the sgrbs x - ray luminosity is not a proxy for @xmath8 and @xmath28 as good as it is found to be for the lgrbs ( davanzo et al .
2012 ) . such phenomenon might be indicative of a lower radiative efficiency of the central engine related to the different progenitor or be the consequence of a different circumburst medium .
finally , we note from fig .
[ fig : correlations ] that the two debated short grb090426 and grb100816a always fall in the lgrbs region of the different correlation planes .
while this can be considered as a further evidence against the classification of grb100816a as a sgrb , we stress that the same reasoning can not be applied safely to grb090426 given the poor constraints on its prompt emission spectrum ( see sect .
4.2.1 ) .
the redshift distribution of sgrbs can provide an indirect tool to constrain the nature of their progenitors and discriminate among the evolutionary channels .
the redshift distribution of merger events of dynamically formed double compact object systems is expected to be different from that of primordial binaries . in particular , given the relatively short delay between formation and merging ( @xmath39 gyr ) , sgrbs originated by the `` fast '' primordial channel ( sect .
4.1 ) should have a redshift distribution which broadly follow that of the star formation , especially at low redshift .
the average ( median ) redshift for the sgrbs of our _ complete _ sample is @xmath40 ( 0.72 ) . this value is higher than the one obtained by fong et al .
( 2013 ) by considering the whole _
sgrb sample ( @xmath41 ) while it is in agreement with the mean redshift ( @xmath42 ) reported by rowlinson et al ( 2013 ) for their sgrb sample limited to the events with @xmath43s ( which is thus excluding all sgrb with extended emission ) .
indeed , the average redshift of our _ complete _ sample is consistent with the expected peak for the redshift distribution of sgrbs originated by the primordial formation channel ( salvaterra et al .
2008 ) . in the primordial binary scenario , the intrinsic formation rate , defined as the number of bursts per unit time and unit comoving volume at redshift @xmath44 ,
@xmath45 is given by the convolution of the massive star binary formation rate and a delay time distribution function @xmath46 .
the delay is the time interval between the formation of the massive star binary and the merging of the compact objects .
the formation rate is assumed to be proportional to the cosmic star formation rate , @xmath47 , parametrized as in hopkins & beacom ( 2006 ) .
@xmath48 is then given by : @xmath49,\ ] ] where @xmath50 is the age of the universe at redshift @xmath44 .
for the delay time distribution function @xmath46 , we adopt the simple expression : @xmath51 with @xmath52 , as suggested from an updated analysis of double compact object mergers performed using population synthesis methods ( portegies zwart & yungeson 1998 ; schneider et al . 2001 ; belczynski et al .
2006 ; oshaughnessy et al .
2008 ) . however , some uncertainty remains on the exact value of the @xmath53 exponent ( berger 2014 , and references therein ) . for a comparison with the data , we compute the observed distribution of sgrbs for three different values of @xmath53 , namely @xmath54 , @xmath52 and @xmath55 ( see also behroozi , ramirez - ruiz & fryer 2014 ) , with characteristic delay times varying from @xmath56 myr to @xmath57 gyr .
given that a @xmath58 distribution would diverge as @xmath59 get close to zero , we will consider an initial time cutoff of 10 myr .
as noted by behroozi , ramirez - ruiz & fryer ( 2014 ) , the choice of the time cutoff has little effect on the mean time delay value . , -1 , and -0.5
are shown with the long - dashed , short - dashed and dotted line , respectively . in computing the expected redshift distribution for the different model
we apply the same photon flux cut , @xmath60 ph s@xmath24 @xmath25 in the _ swift-_bat 15150 kev band , used in the definition of our _ complete _ sample .
, width=317 ] the observed photon flux , @xmath61 , in the energy band @xmath62 , emitted by an isotropically radiating source at redshift @xmath44 is : @xmath63 where @xmath64 is the differential rest frame photon luminosity of the source , and @xmath65 is the luminosity distance .
we define the isotropic equivalent intrinsic burst luminosity as @xmath66 and , to describe the typical burst spectrum , we adopt a band function with photon indexes fixed to @xmath67 and @xmath68 and rest - frame peak energy @xmath69 , assuming that the @xmath70 is valid for sgrbs ( see sect .
4.2.2 ) .
given a normalized sgrb luminosity function , @xmath71 , the observed rate of bursts with peak flux between @xmath72 and @xmath73 is : @xmath74 where @xmath75 $ ] is the comoving volume element , and @xmath76^{1/2}$ ] .
@xmath77 is the solid angle covered on the sky by the survey , and the factor @xmath78 accounts for cosmological time dilation .
@xmath48 is the comoving burst formation rate normalized to unity at @xmath79 as computed in eq . 3 and @xmath80 is the sgrb formation rate at @xmath79 .
finally , we assume that the luminosity function of short grbs is well described by a single power - law of power - index @xmath81 for luminosity in the range @xmath82 erg s@xmath24 .
following salvaterra et al .
( 2008 ) , we obtain the value of @xmath81 by fitting the differential photon flux distribution of sgrbs with @xmath83 s detected by batse in the 50 - 300 kev band ( paciesas et al .
we obtain a reasonable fit to batse data in all formation scenarios with @xmath84 , @xmath85 , and @xmath86 ( errors at 1@xmath87 confidence level ) for @xmath54 , @xmath88 and @xmath89 , respectively .
we then compute the expected redshift distribution of short grbs matching the selection criteria of our _ complete _ sample , i.e. those with photon fluxes @xmath60 ph s@xmath24 @xmath25 in the 15150 kev band of _ swift-_bat ( see sect . 2 ) .
we note that having a flux - limited sample , with a clear photon flux cut , is an asset for the comparison with the model predictions .
model results computed for the primordial formation channel with different time delay distributions are shown in fig .
[ fig : z_distrib ] .
the shaded area accounts for the observed cumulative redshift distribution including the uncertainties due the lack of a secure redshift determination for @xmath90 of the sgrbs of the _ complete _ sample . from the plot
it is clear that the model with @xmath55 can be firmly discarted .
a model with @xmath52 is still acceptable ( although at the lower limit of the observed redshift distribution ) , while the model with a time delay distribution @xmath91 looks to be favoured in accounting for the observed redshift distribution of the sgrbs of our sample , suggesting that they are mainly originated by primordial double compact object systems merging in a relatively short time .
this differs from the results of a similar analysis carried on for the bat6 sample of long grbs : to account for their observed redshift distribution , the rate of long grb formation has to increase with redshift on the top of the known evolution of the sfr density as @xmath92 ( salvaterra et al .
2012 ) .
finally , we note that a significant contribution of sgrbs with dynamical origin would require a lower mean redshift ( salvaterra et al . 2008 ; guetta et al . 2008 ) , suggesting that the contribution of this formation channel to the sgrbs should be negligible and/or limited to the faintest events ( which are not included in our flux - limited sample ) . on the other hand one can conclude that dynamically formed sgrbs should be intrinsically sub - energetic given that , occurring at an average lower redshift , they should show a relatively high observed flux ( and , in principle , fall in a flux - limited sample like the one presented in this paper ) .
a tentative estimate of the fraction of sgrbs with dynamical origin in our sample is given in the next subsection .
the distribution of the intrinsic x - ray absorbing column densities of the sgrb of our _ complete _ sample has a mean value @xmath93 and a standard deviation @xmath94 ( fig .
[ fig : nh ] ) .
we compared our values with the distribution of the intrinsic x - ray absorbing column densities of the _ complete _ sample of lgrbs presented in campana et al .
( 2012 ) . in order to make a proper comparison , we considered only those lgrbs
whose redshift is lower or equal to @xmath95 ( which is the highest redshift recorded in our sgrb _ complete _ sample ) obtaining a mean @xmath96 and a standard deviation @xmath97 ( fig .
[ fig : nh ] ) . when compared in the same redshift bin ,
the distribution of the intrinsic x - ray absorbing column densities of the two sample are fully consistent . a two - sample kolmogorov - smirnov ( k - s ) test gives a probability of 34% , likely indicating that the two distributions are drawn from the same population .
such result is in agreement with the findings ( obtained following a similar procedure ) by kopac et al .
( 2012 ) and margutti et al .
( 2013 ) .
although this result can be intepreted as the evidence of a common environment for long and short grbs , we caution that the intrinsc x - ray @xmath98 might be a good proxy of the grb host galaxy global properties but not for the specific properties of the circumburst medium .
furthermore , the possibility that gas along the line of sight in the diffuse intergalactic medium or intervening absorbing systems can contribute to the absorption observed in the x - ray emission of grbs has to be taken into account ( behar et al .
2011 ; campana et al .
2012 ; starling et al .
however , such effect is expected to dominate at @xmath99 , while at lower redshifts , comparable to the values found for our _ complete _ sample , the absorption within the grb host galaxy is expected to dominate ( starling et al .
2013 ) . for lgrbs ,
the massive star progenitor is expected to significantly enrich the surrounding environment with metals ( whose x - ray @xmath98 is a proxy ) before the collapse with its stellar wind .
alternatively , it has been recently proposed that the helium in the h ii regions where the burst may occur is responsible for the observed x - ray absorption in lgrbs ( watson et al .
2013 ) . under these hypothesis , a high intrinsic x - ray @xmath98 , can be interpreted as the evidence of a dense circumburst medium .
something similar can happen for sgrbs , under the condition that a short time ( of the order of myrs ) separates the supernova explosions which gave origin to the compact objects in the primordial binary system progenitor and its coalescence , with the result that the burst would occur inside its host galaxy and near its star forming birthplace ( perna & belczynski 2002 ) .
such formation channel of `` fast merging '' primordial binaries is in agreement with the observed redshift distribution of our _ complete _ sample discussed above .
indeed , the only case for which combined x - ray and optical afterglow spectroscopy could be performed for a genuine sgrbs ( grb130603b , which is included in our sample ) , provided evidence for a progenitor with short delay time or a low natal kick ( de ugarte postigo et al .
2013 ) .
sgrbs originated by double compact object systems which experienced a large natal kick or which are dynamically formed in globular clusters are expected to be associated with a low - density environments . as shown in table 4
, four sgrbs of the _ complete _ sample have only upper limits on the intrinsic x - ray @xmath98 . among these , grb100625a is the only event whose upper limit is significantly below the average @xmath98 of the distribution . assuming that such limit is indicative of a low - density circumburst medium
, we can estimate that at least 10% of the events of our sample are originated by coalescing binaries formed via the dynamical channel ( or having experienced a large natal kick ) .
furthermore , three ( out of five ) events of the _ complete _ sample missing a robust redshift measure ( grb061201 , grb080503 , grb090515 ) have no detected host galaxy coincident with the afterglow position in spite of the precise ( sub - arcsecond ) location and the deep magnitude limits ( berger 2010 and references therein ) .
as discussed in berger ( 2010 ) `` hostless '' sgrbs may lie at moderately high redshifts @xmath100 , and have faint hosts , or represent a population where the progenitor has been kicked out from its host or is sited in an outlying globular cluster . a statistical study carried out recently by tunnicliffe et al .
( 2014 ) pointed out that the proximity of these events to nearby galaxies is higher than is seen for random positions on the sky , in contrast with the high redshift scenario . following this interpretation ,
up to 4 events ( 25% of the sgrbs of the _ complete _ sample ) might have occurred in low - density environments .
finally , we remark that the _ complete _ sample presented in this paper is built by selecting the events with a bright prompt emission that , according to the standard grb model , is independent on the type of circumburst environment . of the bat6 sample ( data taken from campana et al .
the statistical study of the rest - frame properties of sgrbs gives the best opportunity to characterize the physics of these events , although such studies are often biased by the fact that almost 3/4 of grbs are lacking a secure redshift measurement . in this paper
, we overcome this problem working with a carefully selected sample of _ swift _ sgrbs having a completeness in redshift of @xmath101 which increases up to @xmath102 by considering the events with the brightest @xmath103ray prompt emission . from the study of the observer and rest - frame properties of this sample we obtained the following main results :
- the sgrbs x - ray afterglow fluence correlates linearly with the prompt emission fluence at all times , with no difference for the sgrbs with extended emission or with a short / long - lived x - ray afterglow .
- the percentage of _ swift _ sgrbs lacking an x - ray afterglow detection at early times is more than 6 times higher than for lgrbs .
these events lay in the faint end of the @xmath103ray fluences distribution suggesting that their lack of an x - ray afterglow is likely due to the intrinsic faintness of the prompt emission , although a high redshift origin for these events can not be excluded .
- compared to bright lgrbs ( bat6 sample ) the spectral hardness of the sgrbs of the _ complete _ sample seems to be mainly due to a harder low energy spectral component present in short bursts , rather than to a higher peak energy . -
all the sgrbs of our _ complete _ sample are consistent with the @xmath35 and the @xmath104 correlations , with the significant exception of grb080905a .
we note that such event has the lowest values for @xmath8 and @xmath28 among the burst of our sample .
we speculate that the redshift of this grb might be higher than the value inferred from its association with a nearby ( @xmath105 ) spiral galaxy observed edge - on . apart from this exception
, we found evidence for a @xmath35 correlation followed by sgrbs being sistematically fainter than the correlation defined by lgrbs .
although such finding is intriguing , we caution that it can be affected by the choice of the temporal bin in the estimate of the isotropic peak luminosity for both long and short grbs . - on the @xmath37 plane ,
sgrbs define a region with the same slope measured for the correlation holding for lgrbs but with a different normalization .
two exceptions are grb090426 and grb100816a , both consistent within @xmath106 with the @xmath37 lgrb correlation .
- the rest - frame afterglow x - ray luminosity at early times ( @xmath107 min ) correlates well with the prompt emission @xmath8 and @xmath28 . at later times
( @xmath108 hr ) these correlations become much less significant .
this is at variance with respect to what observed for lgrbs , where , although with increasing scatter from early to late times , the @xmath109 and @xmath110 correlations hold at all times .
such effect might be indicative of a lower radiative efficiency of the sgrb central engine related to the different progenitor or be the consequence of a different circumburst medium for long and short grbs . - in light of its duration , prompt emission and x - ray afterglow properties , grb100816a ( initially classified as a sgrb ) is likely a long - duration event .
no firm conclusion can be derived for the classification of grb090426 , mainly due to the lack of strong constraints on the properties of its prompt emission spectrum .
- the redshift distribution of the sgrbs of our sample has a mean value of @xmath111 .
no evidence for a different environment for long and short grbs can be derived from the rest - frame x - ray column densities . when compared on the same redhift bin ,
the distributions for long and short grbs are fully comparable .
both these results are consistent with the scenario of `` primordial binary '' progenitors , with short coalescence times .
however , a minor contribution ( 10%25% ) of dynamically formed ( or with large natal kicks ) compact binaries progenitors can not be excluded .
the _ complete _ sample of sgrbs presented in this paper consists of 16 events observed by the _ swift _ satellite over 8.5 years .
as can be evinced from table 1 , the efficiency of redshift measurements almost doubled ( from @xmath90 to @xmath112 ) in the last 5.5 years ( since 2008 ) .
this is likely due to the improving procedures of automatic follow - up from the ground and to the advent of effective facilities like grond ( greiner et al .
2008 ; nicuesa guelbenzu et al .
2012 ) . as a consequence of this , within a few years of further _ swift _ operations we will handle a well - selectd sample of @xmath113 sgrbs with a redshift completeness higher than 70% . with such numbers , the statistics of this sgrb sample will become comparable to , e.g. , the lgrb bat6 sample , opening the possibility to study , e.g. , the sgrbs luminosity function and measuring their intrinsic extinction through their optical and near - infrared spectral energy distributions ( salvaterra et al . 2012 ; covino et al .
this work made use of data supplied by the uk swift science data centre at the university of leicester .
we acknowledge the italian space agency ( asi ) for financial support through the asi - inaf contract i/004/11/1 .
99 amati , l. et al .
, 2002 , a&a 390 , 81 amati , l. et al . , 2008 , mnras , 391 , 577 amati , l. , 2008 , aipc , 966 , 3 antonelli , l. a. et al . , 2009 ,
a&a , 507 , l45 barthelmy , s. d. , barbier , l. m. , cummings , j. r. et al .
2005 , ssr , 120 , 143 barthelmy , s. d. et al .
2009 , gcn circ .
9494 barthelmy , s. d. et al .
2013 , gcn circ .
14741 band , d. et al .
, 1993 , apj , 413 , 281 behar , e. et al .
, 2011 , apj , 734 , 26 behroozi , p. s. , ramirez - ruiz , e. , fryer , c. l. , 2014 , arxiv:1401.7986 belczynski , k. & kalogera , v. , 2001 , apj , 550 , l183 belczynski , k. , bulik , t. & kalogera , v. , 2002 , apj , 571 , l147 belczynski , k. , perna , r. , bulik , t. et al .
2006 , apj , 648 , 1110 berger , e. et al .
2007 , apj , 664 , 1000 berger , e. , 2009 , apj , 690 , 231 berger , e. et al . , 2009 ,
apj , 704 , 877 berger , e. , 2010 , apj , 722 , 1946 berger , e. , 2011 , newar , 55 , 1 berger , e. , fong , w. & chornock , r. , 2013 , apj , 774 , l23 berger , e. 2014 , ara&a , 51 in press bernardini , m. g. et al . , 2012 , mnras , 425 , 1199 bhat , p. n. , 2010 , gcn circ .
10912 bloom , j. s. et al .
, 2006 , apj , 638 , 354 bloom , j. s. et al . , 2007 , apj , 654 , 878 blundell , k. m. , rawlings , s. & willott , c. j. , 1999 , aj , 117 , 677 bremer , m. et al .
1998 , a&a , 332 , l13 bromberg , o. et al . , 2013 , apj , 764 , 179 burrows , d. n. , hill , j. e. , nousek , j. a. et al .
2005 , ssr , 120 , 165 campana , s. et al .
2012 , mnras , 421 , 1697 chornock , r. & berger , e. , 2011 , gcn circ 11518 costa , e. et al .
1997 , nature , 387 , 783 covino , s. et al .
2006 , a&a , 447 , l5 covino , s. et al .
2013 , mnras , 432 , 1231 cucchiara , a. et al .
2013 , apj , 777 , 94 davanzo , p. et al .
2009 , a&a , 498 , 711 davanzo , p. et al .
2012 , mnras , 425 , 506 de ugarte postigo , a. et al .
2013 , a&a , in press , arxiv:1308.2984 eichler , d. , livio , m. , piran , t. , schramm , d.n .
, 1989 , nature , 340 , 126 evans , p. et al .
2009 , mnras , 397 , 1177 fishman , g. j. , et al .
1989 , _ gro _ science workshop proc . ,
w. n. johnson , greenbelt : nasa , 2 - 39 fitzpatrick , g.2010 , gcn circ .
11124 fong , w. f. et al .
2011 , apj , 730 , 26 fong , w. f. et al .
2013 , apj , 769 , 56 fox , d. b. , frail , d. a. , price , p. a. , et al .
2005 , nature , 437 , 845 frail , d. a. et al .
1997 , nature 389 , 261 gehrels , n. et al . , 2004 ,
apj , 611 , 1005 gehrels , n. et al . , 2005 ,
nature , 437 , 851 gehrels , n. et al .
, 2008 , apj , 689 , 1161 ghirlanda , g. , ghisellini , g. & celotti , a. , 2004 , a&a , 422 , l55 ghirlanda , g. , nava , l. , ghisellini , g. , celotti , a. , firmani , c. , 2009 , a&a , 496 , 585 ghirlanda , g. , ghisellini , g. & nava , l. , 2011 , mnras , 418 l109 gorosabel , j. et al . , 2010 ,
11125 golenetskii , s. et al .
, 2005 , gcn circ .
4394 golenetskii , s. et al . , 2010 ,
11127 golenetskii , s. et al . , 2013 ,
14771 graham , j. f. et al . , 2009 ,
apj , 698 , 1620 greiner , j. et al . , 2008 , pasp , 120 , 405 grindlay , j. , portegies zwart , s. & mcmillan , s. , 2006 , natph , 2 , 116 heng , k. et al . , 2008 ,
apj , 681 , 1116 hjorth , j. , watson , d. , fynbo , j. p. u. , et al . , 2005a , apjl , 630 , l117 hjorth , j. , watson , d. , fynbo , j. p. u. , et al . ,
2005b , nture , 437 , 859 hjorth , j. , bloom , j. s. , 2012 , gamma - ray bursts , cambridge astrophysics series 51 , pp .
169 - 190 hjorth , j. et al . , 2012 ,
apj , 756 , 187 hopkins , a. m. , beacom , j. f. , 2006 , apj , 651 , 142 hopman , c. et al .
, 2006 , apj , 643 , l91 isobe , t. et al . , 1990 ,
apj , 364 , 104 jakobsson , p. et al .
, 2006 , a&a , 447 , 897 jin , z. p. et al .
, 2013 , apj , 775 , l19 kalberla , p. m. w. et al .
2005 , a&a , 440 , 775 kann , d. a. , et al .
2011 , apj , 734 , 96 kendall , m. , stuart , a. , 1979 , the advanced theory of statistics .
macmillian , new york kopac , d. et al . , 2012 , mnras , 424 , 2392 kouveliotou , c. et al . , 1993 , apj , 413 , l101 kumar ,
p. & pamaitescu , a. , apj , 541 , l51 leibler , c. n. , berger , e. 2010 , apj , 725 , 1202 levesque , e. et al . , 2010 , mnras , 401 , 963 margutti , r. et al .
, 2012 , apj , 756 , 63 margutti , r. et al
. , 2013 , mnras , 428 , 729 mcbreen , s. et al . , 2010 ,
a&a , 516 , 71 nakar , e. , 2007 , phys .
, 442 , 166 nava , l. et al .
2011 , a&a , 530 , a21 nava , l. et al .
2012 , mnras , 421 , 1256 narayan , r. , paczynski , b. , piran , t. , 1992 , apj , 395 , l83 nicuesa guelbenzu , a. et al .
2012 , a&a , 548 , a101 norris , j.p . ,
marani , g.f .
& bonnell , j.t . , 2000 , apj , 534 , 248 norris , j.p . ,
scargle , j.d .
& bonnell , j.t . , 2001 , in gamma - ray bursts in the afterglow era , ed .
e. costa , f. frontera & j. hjorth ( springer - verlag ) , 40 norris , j.p .
et al . , 2010 ,
11113 norris , j.p .
et al . , 2013 ,
14746 nysewander , m. et al .
2009 , apj , 701 , 824 ohno , m. et al .
, 2007 , gcn circ .
6638 oshaughnessy , r. , belczynski , k. , kalogera , v. , 2008 , apj , 395 , l8 paciesas , w. s. et al . , 1999 ,
apjs , 122 , 465 padovani , p. 1992 , a&a , 256 , 399 panaitescu , a. , kumar , p. , & narayan , r. 2001 , apjl , 561 , l171 perna , r. & belczynski , k. , 2002 , apj , 570 , 252 piranomonte , s. et al . , 2008 , a&a , 491 , 183 portegies zwart , s. f. , yungelson , l. r. , 1998 , a&a , 332 , 173 press , w. h. , flannery , b. p. , teukolsky , s. a. , & vetterling , w. t. , 1986 , numerical recipes ( new york : cambridge univ . press ) rowlinson , a. et al . , 2010 , mnras , 408 , 383 rowlinson , a. et al . , 2013 , mnras , 430 , 1061 sakamoto , t. & gehrels , n. , aipc , 1133 , 112 sakamoto , t. et al . , 2011 ,
apjss , 195 , 2 sakamoto , t. et al .
, 2013 , apj , 766 , 41 salvaterra , r. et al . , 2008 ,
mnras , 388 , l6 salvaterra , r. , devecchi , b. , colpi , m. , davanzo , p. , 2010 ,
mnras , 406 , 1248 salvaterra , r. , campana , s. vergani , s. d. et al .
2012 , apj , 749 , 68 schneider , r. et al . , 2001 ,
mnras , 324 , 797 soderberg , a. m. et al . , 2006 , apj , 650 , 261 spearman , c. 1904 , am .
j. psycol . , 15 , 72 starling , r. l. c. et al .
, 2013 , mnras , 431 , 3159 tagliaferri , g. et al . , 2005 , nature , 436 , 985 tanvir , n. et al .
2010 , gcn circ .
11123 tanvir , n. et al .
2013 , nature , 500 , 547 tsutsui , r. et al .
, 2013 , mnras , 431 , 1398 uehara , t. et al . , 2008 ,
gcn circ 7223 van paradijs , j. et al .
1997 , nature 386 , 686 vietri , m. 2000 , astroparticle physics , 14 , 211 yonetoku , d. et al . , 2004 , apj , 609 , 935 watson , d. et al . , 2013 , apj , 768 , 23 zhang , f. et al . , 2012 , apj , 750 , 88 | we present a carefully selected sample of short gamma - ray bursts ( sgrbs ) observed by the _ swift _
satellite up to june 2013 . inspired by the criteria we used to build a similar sample of bright long grbs ( the bat6 sample ) , we selected sgrbs with favorable observing conditions for the redshift determination on ground , ending up with a sample of 36 events , almost half of which with a redshift measure .
the redshift completeness increases up to about 70% ( with an average redshift value of @xmath0 ) by restricting to those events that are bright in the 15 - 150 kev _ swift
_ burst alert telescope energy band .
such flux - limited sample minimizes any redshift - related selection effects , and can provide a robust base for the study of the energetics , redshift distribution and environment of the _ swift _ bright population of sgrbs .
for all the events of the sample we derived the prompt and afterglow emission in both the observer and ( when possible ) rest frame and tested the consistency with the correlations valid for long grbs . the redshift and intrinsic x - ray absorbing column density distributions we obtain are consistent with the scenario of sgrbs originated by the coalescence of compact objects in primordial binaries , with a possible minor contribution ( @xmath110%25% ) of binaries formed by dynamical capture ( or experiencing large natal kicks ) .
this sample is expected to significantly increase with further years of _ swift _ activity .
1_e,-1 53e_53 0n_*,0 02_0,2 28d_28
^ -2 [ firstpage ] gamma - rays : bursts x - rays : general . |
the discovery of the higgs boson completes the list of particles of the standard model . however , there are many open questions regarding the dynamics of electroweak symmetry breaking . addressing these questions
has been a driving force in both theoretical and experimental explorations of the energy frontier .
one of the most outstanding problem is the nature of the electroweak phase transition .
currently , we have measured with precision the size of the higgs vacuum expectation value ( vev ) and the mass of the higgs boson . however , we know very little about the shape of the higgs potential beyond that .
the standard model defines its higgs potential with only renormalizable terms the so - called mexican " hat form . in this case the electroweak symmetry is smoothly restored via a continuous cross - over as the temperature is raised above the electroweak scale @xcite .
however , new physics can modify the nature of the phase transition , possibly turning it into an abrupt first - order phase transition . on the one hand ,
such new physics can be searched for directly at current and future colliders . on the other hand ,
a first order electroweak phase transition in the early universe would generate a stochastic background of gravitational waves that can be searched for with interferometers @xcite .
discovering such a gravitational wave signal would establish the electroweak phase transition as a new milestone in our understanding of the early universe .
it will advance our knowledge into an epoch significantly earlier than nucleosynthesis .
a first order electroweak phase transition requires a significant deviation away from the renormalizable higgs potential , which implies the presence of new physics close to the weak scale .
we can look for such new particles directly , such as at the lhc .
even though it can be powerful in certain cases , the reach of this approach is limited .
the new physics can be weakly coupled , and the searches at the lhc suffers from large background . at the same time , the most model independent effect of such new physics is the induced deviation in the higgs couplings .
measuring such couplings precisely , and uncovering potential new physics , is a major physics goal of proposed higgs factories . in this paper
, we will focus on the potential of probing new physics associated with electroweak symmetry breaking at these facilities .
if a first order electroweak phase transition occurred in the early universe , the collision of bubbles and damping of plasma inhomogeneities would have generated a stochastic background of gravitational waves .
the frequency of these waves is relatively model independent , being related to the scale of the cosmological horizon at the time of the phase transition .
therefore today we expect the waves to have redshifted into the milli - hertz range .
this potential signal is impossible to probe with ground - based gravitational wave interferometers like advligo due to seismic noise .
however , the signal is ideal for a space - based interferometer like elisa @xcite with arm lengths of order millions of kilometers . in this paper
, we assess the possibility of using gravitational waves as probes of a first order electroweak phase transition and the complementarity of this technique with the collider searches .
there are a number of ways in which new particles may cause the electroweak phase transition to become first order . in general a first order phase transition
can occur if the higgs effective potential is modified from its standard model form so as to develop a potential energy barrier separating the phases of broken and unbroken electroweak symmetry . as discussed in , there are three general model classes in which the barrier can arise .
first , if the new degrees of freedom are scalar fields that participate in the electroweak phase transition ( their vev changes at the same time as the higgs ) then _ tree - level interactions _ with the higgs field can make some regions of field space energetically dis - favorable and lead to a barrier .
second , the presence of new particles coupled to the higgs boson affects the running of the higgs mass parameter and self - coupling .
then the barrier can arise by virtue of _ quantum effects_. third , if the new particles are present in the early universe plasma and acquire their mass ( at least partially ) from the higgs field , then a barrier can arise via _
thermal effects_. this can be understood as a tradeoff between minimizing the energy , represented by the tree - level higgs potential , and maximizing the entropy , which prefers the higgs field to take on values where particles in the plasma are light .
then , this third case can be further divided into two categories : a barrier arising from light scalars via the thermal cubic term @xmath6 and a barrier arising from heavy particles that get their mass predominantly from a large coupling with the higgs field . in this paper
, we survey a number of simplified models that demonstrate the basic ingredients necessary for a first order electroweak phase transition .
we focus on four models in which the sm is extended , respectively , to include a real scalar singlet , a scalar doublet , heavy chiral fermions , and varying yukawa couplings .
this set of models exemplifies all of the different phase transition model classes , enumerated above .
in each of the models discussed here , the higgs field is represented by @xmath7 , and the standard model lagrangian contains @xmath8 in calculating the scalar effective potential we write @xmath9 with @xmath10 real .
the vacuum spontaneously breaks the electroweak symmetry , @xmath11 with @xmath12 .
the higgs mass is denoted as @xmath13 , and it takes the value @xmath14 .
first , we add to the sm a real scalar field @xmath15 , which is a singlet under the sm gauge group .
this is probably the simplest extension of the higgs sector of the sm . at the same time , due the lack of other interactions , it gives rise to the most independent signal .
the most general renormalizable lagrangian is written as @xmath16 without loss of generality , we can set @xmath17 .
since the new scalar is a singlet , it only interacts with the standard model via the higgs portal , @xmath18 and @xmath19 .
the electroweak phase transition and collider phenomenology in this model , sometimes called the xsm , have been studied extensively ; see _ e.g. _ and references therein . the gravitational wave signal in this model has been studied recently by refs .
@xcite .
there is no single reason why this model admits a first order electroweak phase transition .
in fact different limits of this simple model exhibit each of the phase transition model classes that were identified in .
most notably , the tree - level interactions play a significant role in most of the parameter space . during the electroweak phase transition , the singlet @xmath20 need
not remain fixed .
if @xmath20 changes along with @xmath21 , then the higgs portal terms , @xmath22 and @xmath23 , can give rise to a barrier in the effective potential , and the phase transition is first order .
after electroweak symmetry breaking @xmath24 , and generically we expect the singlet field to acquire a vacuum expectation value as well , @xmath25 . then , the higgs portal operators allow the higgs and singlet fields to mix
. the mixing angle @xmath26 satisfies @xmath27 where @xmath14 is the physical higgs boson mass , and @xmath28 is the physical mass of the singlet .
interactions between the higgs boson and the singlet scalar affect the coupling of the higgs to the z - boson . writing the effective @xmath29 coupling as @xmath30 , we define the fractional deviation from the sm value as @xmath31 where the couplings are evaluated at a center of mass energy @xmath32 .
we calculate @xmath33 as @xmath34 the first term arises from the tree - level higgs - singlet mixing ( [ eq : singlet_mixing ] ) .
this is typically the dominant contribution to @xmath33 . at the one - loop order
, the singlet contributes to the wavefunction renormalization of the higgs .
this gives rise to the second and third terms in .
we have generalized the calculation in refs .
@xcite , to allow for the cases without a @xmath35 symmetry .
the bosonic loop function is given by @xcite @xmath36 the wavefunction renormalization terms are typically sub - dominant , except for the @xmath35 limit ( discussed in ) where @xmath37 and the @xmath38 term is dominant .
the fourth term in also arises at the one - loop order . as recognized in this term
appears when the higgs trilinear coupling @xmath39 deviates from its sm value @xmath40 .
the effect on @xmath33 depends on the center of mass energy , and for @xmath41 the prefactor evaluates to @xmath42 @xcite
. the cubic self - coupling of the mass eigenstate higgs ( @xmath43 ) is calculated as @xmath44 in the standard model we have @xmath45 .
the last term in @xmath39 arises from a three - vertex , one - loop graph .
as we will see , models exhibiting a strongly first order phase transition , typically have an @xmath46 deviation in @xmath47 , and therefore , this effect on @xmath33 can be sizable .
if the singlet is sufficiently light , @xmath48 , the higgs decay channel @xmath49 opens .
this decay contributes to the higgs invisible width .
the invisible width is calculated as @xcite @xmath50 where @xmath51 is the effective tri - linear coupling of the mass eigenstates . if the invisible channel is open , the branching fraction is typically so large as to be excluded already by lhc limits , @xmath52 @xcite
. therefore we require @xmath53 . in general the model has @xmath54 parameters corresponding to the potential terms in with @xmath17 .
two parameters can be exchanged for the higgs mass and vev , leaving @xmath55 free parameters . in
we present the main result for this model , which entails a scan over the @xmath55-dimensional parameter space . in the following sub - sections we discuss a couple of special limiting cases of this model .
even though they do not represent generic models with first order electroweak phase transition , they give rise to different predictions .
we include them in our discussion for completeness .
we impose the @xmath35 discrete symmetry under which the singlet is odd , @xmath56 , and the standard model fields are even . in terms of the singlet lagrangian ( [ eq : singlet_l ] ) the symmetry enforces @xmath57
we also require that the @xmath35 symmetry is not broken spontaneously , and thus @xmath58 .
the only interaction between the standard model and the singlet is through the higgs portal , @xmath59 .
the @xmath35 symmetry forbids a mixing between the higgs and singlet fields . in the absence of mixing ,
modifications to the @xmath29 coupling ( [ eq : singlet_dzh ] ) first arise at the one - loop level .
the tree level modifications to the trilinear coupling ( [ eq : singlet_lam3 ] ) are also suppressed in this limit , and therefore , only the @xmath60 term contributes to @xmath33 .
thus we expect that this corner of parameter space can evade constraints on @xmath33 from future colliders as discussed later . as such
, this model has been identified as a `` worst case scenario '' for finding evidence of a first order ew phase transition at colliders @xcite . despite the vanishing mixing ,
the electroweak phase transition may still be first order .
morally speaking , the higgs - singlet mixing is a `` local '' property of the theory , related to the behavior of small fluctuations about the vacuum , but the nature of the phase transition depends also up `` global '' properties of the theory , _
e.g. _ whether the theory admits other metastable vacua .
the presence or absence of such metastable vacua is not _
directly _ related to the mixing at the true vacuum .
two specific scenarios have been studied .
if the higgs portal coupling is sufficiently large , the singlet can affect the running of the higgs self - coupling , which may induce a barrier in the effective potential @xcite .
alternatively , the higgs portal interaction , @xmath61 with @xmath62 , may give rise to a barrier in the effective potential at tree - level when the phase transition passes from a vacuum with @xmath63 and @xmath64 to a vacuum with @xmath65 and @xmath66 . the tree - level higgs - singlet mixing ( [ eq : singlet_mixing ] ) vanishes when we take @xmath67 unlike the @xmath35 symmetric limit of , the choice of parameters in is not associated with any enhanced symmetry .
significant fine - tuning among tree - level parameters is necessary to reach this limit .
furthermore , such a tuning is not technically natural ; the mixing is induced radiatively . at the one - loop order ,
the induced mixing is @xmath68 .
otherwise , the first order electroweak phase transition and collider phenomenology is similar to the @xmath35 case . in the unmixed limit
, the singlet can be pair produced through an off - shell higgs via the @xmath69 coupling @xmath70 , and can decay to @xmath71 , @xmath72 , and @xmath73 final states via the radiatively generated mixing
. then by the goldstone equivalence theorem , the singlet decay branching ratios for the @xmath71 , @xmath72 , and @xmath73 channels are 25@xmath74 , 50@xmath74 , and 25@xmath74 respectively .
those final states contribute to a multi - lepton , multi - jet signature , which can be probed at the lhc . with a large @xmath72 branching ratio
, the 4@xmath75 channel leads to a same - sign dilepton with multiple jets ( zero b - jet ) final state .
the background processes for this channel include @xmath76 , @xmath77 , @xmath78 , @xmath79 , and same sign @xmath72 plus jets .
all backgrounds except same sign @xmath72 plus jets , can be estimated from the @xmath80 searches in the same sign dilepton with at least two b - jets channel , by replacing the b - tagging with a b - jet veto @xcite .
we assume the b - tagging efficiency is 70@xmath74 .
for the same sign @xmath72 plus jets , we include both single parton scattering and double parton scattering @xcite , and assume a 90@xmath74 acceptance to account for the lepton efficiency , and kinematics .
we assume the same 90@xmath74 acceptance for signal as well .
then at the hl - lhc , we expect a 2@xmath81 significance for @xmath82 , which corresponds to @xmath83 , and @xmath84 @xcite . in this section
, we go beyond the singlet to consider new particles in non - trivial representations of the sm gauge group .
some of the simplest cases are obtained by introducing @xmath85 scalar doublets and singlets with @xmath86 charge .
perhaps the most well - known example is the mssm stop .
however , the light stop scenario is very restricted and , at least in simple cases , it can not give rise to a first order electroweak phase transition without running afoul of collider constraints @xcite , see also @xcite .
many of these constraints are a consequence of the supersymmetry .
for example , the scalar top partner must to be colored and hence the stop is subject to stringent limits from collider searches . to avoid the collider constraints , models like folded susy
have been proposed @xcite , in which the stops can still solve the hierarchy problem , but are not colored . in the following ,
we consider a similar stop - like model .
the new particles are taken to have the same electroweak gauge quantum numbers as the stop , but they are not colored .
in addition , their couplings are not subject to the constraints of supersymmetry .
we extend the sm to include @xmath87 copies ( flavors ) of scalar doublets and complex scalar singlets .
we will denote the doublets and singlets as @xmath88 and @xmath89 where the index @xmath90 runs from @xmath91 to @xmath92 . in order to mimic the interactions of colored squarks
, we require the lagrangian to respect the global @xmath93 symmetry , under which the @xmath94 and @xmath89 transform in the fundamental representation , and the sm fields are invariant .
notice that we have used a susy - like notation to indicate the electroweak gauge quantum numbers , but no susy relations are implied . with the new stop - like particle content
, the scalar potential can be written as @xmath95 \nonumber \per\end{aligned}\ ] ] the sum over @xmath96 flavors has been suppressed . in general
the model has @xmath97 parameters , but @xmath98 of these can be exchanged for the higgs mass and vev , leaving @xmath99 free parameters .
additionally , we will later assume a universal dimensionless coupling , @xmath100 , which reduces the free parameters to four : @xmath101 .
we present the results of a parameter - space scan in . in the well - known light
stop scenario of the mssm @xcite , the electroweak phase transition can become first order due to the presence of these scalar particles in the plasma .
their contribution to the higgs thermal effective potential ( background - dependent free energy density ) goes as @xmath102^{3/2 } t$ ] where @xmath103 is the number of colors , and the effective stop mass @xmath104 depends on the background higgs field @xmath10 and the plasma temperature @xmath105 ( daisy correction ) .
this non - analytic term in @xmath106 arises only for relativistic bosonic fields , due to the non - analyticity of the bose - einstein distribution function at @xmath107 . in a regime where the stop mass can be approximated as @xmath108 ,
the effective potential acquires a cubic term , @xmath109 , which can provide the requisite barrier for a first order phase transition . in our stop - like model , the same thermal effects can give rise to a first order electroweak phase transition .
however , since we do not impose the susy relations , @xmath110 is replaced by some combination of the quartic couplings @xmath111 in . in principle
, @xmath111 can be larger than @xmath110 , which increases the height of the potential energy barrier and strengthens the first order the phase transition .
for simplicity , we neglect the daisy resummation , _
i.e. _ the one - loop temperature - dependent mass correction , which plays an important role in the mssm s light stop scenario @xcite .
we expect that this effect will weaken the pt strength on a parameter point - by - point basis without affecting our broader conclusions regarding testability of models with a first order electroweak phase transition .
we assume that the new scalar fields do not acquire vevs , @xmath112 .
thus , the the flavor symmetry prevents the higgs from mixing with the squark - like fields .
nevertheless , the trilinear interactions allows the stop - like fields to mix after electroweak symmetry breaking .
the mixing angle satisfies @xmath113 the spectrum consists of two `` stops '' and one `` sbottom , '' and we denote their mass eigenvalues by @xmath114 , @xmath115 , and @xmath116 .
their couplings to the higgs are given by [ eq : g_stop ] @xmath117 these dimensionful couplings can be read off of upon diagonalizing the stop mass matrix . since the stop- and sbottom - like particles are not actually colored , they do not affect the higgs coupling to gluons . however , the new charged scalar particles do increase the strength of the higgs - photon coupling radiatively .
this increases the higgs diphoton decay rate @xcite @xmath118 where @xmath119 is the electromagnetic fine structure constant and [ eq : stoplike_a ] @xmath120 with @xmath121 and @xmath122 .
we take the electromagnetic charges of the stop- and sbottom - like particles to be @xmath123 and @xmath124 .
the functions @xmath125 are defined in . to compare with the sm contribution , @xmath126
, we drop the @xmath127 and @xmath128 terms from .
the new charged scalars also affect the higgs coupling to z - bosons .
at one - loop order the dominant effect typically comes from a wavefunction renormalization on the higgs leg .
this graphs brings two factors of the quartic couplings @xmath111 , which appear in .
the other one - loop graphs , _
i.e. _ the vertex renormalization and the z - boson wavefunction renormalization , are suppressed compared to the higgs wavefunction renormalization by factors of @xmath129 where @xmath130 is the electromagnetic coupling . as long as @xmath131
the higgs wavefunction renormalization is the dominant effect on the @xmath29 coupling .
consequently the @xmath29 coupling deviates from its sm value by @xcite @xmath132 where the loop function is defined in . in the previous sections we have discussed how new scalar states can lead to a first order electroweak phase transition . here
we investigate models in which the first order transition derives from new fermions .
the key ingredient is that the fermions have a large coupling to the higgs , which gives them a large mass during the electroweak phase transition .
prior to the phase transition the fermions were lighter , since the higgs vev does nt contribute to their mass .
since it is energetically unfavorable for the fermion masses to grow , the effective potential develops a barrier , which leads to a first order phase transition .
we discuss two implementations of this idea .
extend the sm to include the left - chiral fermions @xmath133 , @xmath134 , @xmath135 , and @xmath136 .
they have the same gauge charge assignments as higgsinos and gauginos in the mssm .
two doublets @xmath137 are required for anomaly cancellation , and either the singlet @xmath135 or the triplet @xmath138 is required to allow a yukawa coupling with the standard model higgs .
working in the two - component weyl spinor notation , the lagrangian is @xmath139 - \frac{1}{2 } m_2 \bigl [ \tilde{w}^a \tilde{w}^a + \hc \bigr ] - \mu \bigl [ \tilde{h}_2 \cdot \tilde{h}_1 + \hc \bigr ] \nn & \quad - \bigl [ \phi^{\dagger } \bigl ( h_2 \sigma_a \tilde{w}^a + h_2^{\prime } \tilde{b } \bigr ) \tilde{h}_2 + \hc \bigr ] - \bigl [ \phi \cdot \bigl ( -h_1 \sigma_a \tilde{w}^a + h_1^{\prime } \tilde{b } \bigr ) \tilde{h}_1 + \hc \bigr ] \per\end{aligned}\ ] ] we have again used a susy - like notation for the fields so as to easily identify their gauge quantum numbers .
reference @xcite studied the phenomenology in this model and identified a region of parameter space in which the electroweak phase transition can be first order .
in general , the model has @xmath54 free parameters : @xmath140 mass parameters @xmath141 and @xmath142 couplings @xmath143 . suggests to focus on the restricted parameter space @xmath144 this reduces the free parameters to @xmath140 : one mass parameter @xmath145 and two couplings @xmath146 and @xmath147 . in this restricted parameter space ,
the spectrum consists of two degenerate `` charginos , '' two degenerate lighter `` neutralinos , '' and two degenerate heavier `` neutralinos '' : [ eq : fermion_spectrum ] @xmath148 the mass eigenstates have couplings with the higgs given by [ eq : fermion_g ] @xmath149 where the neutralino @xmath150 can be decomposed into @xmath151 .
we only consider @xmath152 to avoid the higgs and z invisible decays
. identified that the phase transition can be strongly first order in the limit @xmath153 with @xmath146 or @xmath154 . in this case , the charginos and neutralinos are light near in the symmetric phase ( @xmath155 in ) but heavy in the broken phase .
this makes it energetically preferable for the system to remain in the symmetric phase , and the phase transition is delayed until a lower temperature thereby becoming more strongly first order . since we will be interested in a region of parameter space with large couplings , there is a threat that the new fermions will exacerbate the electroweak vacuum instability .
following we counter this problem by introducing new scalar particles that have field - dependent masses @xmath156 where @xmath157 .
we do not expect these particles to play any significant role in the phenomenology or phase transition dynamics .
the presence of new charged fermions tends to suppress the higgs diphoton decay rate .
we calculate @xmath158 by generalizing the mssm chargino calculation in : @xmath159 where @xmath160 and @xmath161 appear in and the chargino contribution is given by @xmath162 we will see that if @xmath163 the model is already strongly constrained by lhc limits on the higgs diphoton decay width .
therefore we also consider the case @xmath164 where the charginos do not contribute to the higgs diphoton decay rate at one - loop order .
both the charginos and neutralinos affect the higgs coupling to z - bosons .
as argued above , the dominant effect comes from the higgs wavefunction renormalization provided that @xmath165 .
thus we calculate the deviation to the @xmath29 coupling as @xmath166 where the fermion loop integral is given by @xmath167 where @xmath168 appears as the @xmath169 renormalization scale .
since the @xmath29 coupling arises already at tree - level , @xmath170 , there can be another one - loop correction when the @xmath169 parameters @xmath171 and @xmath172 are matched onto physical quantities such as the @xmath173 boson mass and @xmath174 @xcite . in this model
, we expect a large contribution from the matching because the custodial symmetry is broken . using ref .
@xcite we estimate that this matching effect can contribute to @xmath33 with a magnitude that is comparable to , but a more precise calculation of @xmath33 is beyond the scope of our work .
since we are primarily interested in establishing that models with a first order phase transition are testable at future colliders , we are satisfied that the predicted value of @xmath33 in this model is comparable to , or larger than , . ] the value given by . as discussed previously , the custodial symmetry is broken in this model , and therefore , we expect a large contribution to the electroweak precision parameter , t @xcite . to avoid the constraints , we can introduce new particles to the theory to cancel the contribution .
a heavy higgs can generate a negative contribution to the t parameter , and compensate the contribution to the t parameters from the neutralinos and the charginos .
references @xcite recently studied the electroweak phase transition in an electroweak - scale implementation of the froggatt - nielsen mechanism .
we investigate their model in this section .
extend the standard model to include a real scalar field @xmath175 , which is a singlet under the standard model gauge group .
we call @xmath175 the flavon field . by enforcing an appropriately chosen flavor symmetry ,
the standard model yukawa interactions can be forbidden . instead , the yukawa interactions are generated from dimension - five operators after the flavon gets a vev .
these dimension - five operators are written schematically as @xmath176 where the @xmath177 s are flavor charges of the fermions and the higgs .
the associated yukawa matrix is @xmath178 . in it
is assumed that the expectation value of @xmath175 changes during the electroweak phase transition .
( it need not be zero before the transition . ) across the two - dimensional field space , the fermion masses vary due to the explicit dependence on the higgs field and the implicit dependence on the flavon field , via the yukawa coupling . for fermion species @xmath179 we can write the field - dependent mass as @xmath180 the electroweak phase transition can be strongly first order if @xmath181 in the symmetric phase and @xmath182 in the broken phase @xcite . in this case
the fermion is light in both the symmetric and broken phases , but heavy at intermediate field values . as a result
, intermediate field values are energetically disfavored , and the corresponding potential energy barrier induces the phase transition to become first order .
one can view this scenario as a different implementation of the heavy fermion model of .
one could study the electroweak phase transition in this model by specifying a potential for @xmath175 and tracking the evolution of both @xmath175 and the higgs @xmath10 through the two - dimensional field space @xcite . for simplicity ,
we follow in assuming that @xmath175 can be parametrized in terms of @xmath10 along the phase transition trajectory .
then the model is specified by writing @xmath183 . taking a phenomenological perspective
, we consider three models for the yukawa couplings @xmath184 \times \sqrt{2 } m_f / v & \qquad \text { , for $ f$ a lepton } \end{cases } \nonumber\end{aligned}\ ] ] where @xmath185 is the mass of fermion @xmath186 . in each case , the yukawa coupling is fixed to @xmath187 for @xmath188 .
the first two models are chosen to match the parametrization studied in . here
, the varying flavon field leads to a universal ( flavor - independent ) _ shift _ in the yukawa couplings for either the quarks or the leptons .
the third model is chosen to avoid possible constraints from flavor - changing neutral currents . here
, the flavon leads to a flavor - independent _ rescaling _ of the lepton yukawa couplings while leaving the quark yukawa couplings unaffected .
the factor of @xmath189 is introduced to balance the @xmath190 yukawa coupling @xmath191 ; then , @xmath192 and @xmath193 .
whereas the first two models can be thought to arise from the operator in , the third model would arise from the operator @xmath194 where vev of the flavon field sets the scale of the yukawa matrix but not the flavor - dependent mixings and splittings .
since we have not specified an explicit interaction between the flavon and standard model fields , it is difficult to assess collider constraints on this model . in general
, we expect that the model can be constrained via higgs - flavon mixing and flavor - changing neutral currents ( fcnc ) . generally , the physical higgs boson and the singlet flavon will mix , unless a symmetry forbids it .
as we discussed in , this mixing leads to a number of potential constraints .
second , if the flavon s interactions are not flavor - diagonal , then it may contribute to fcnc , which are strongly constrained .
for instance , as discussed in @xcite , the scenario with one froggatt - nielsen flavon is ruled out , and the scenrio with two froggatt - nielsen flavons are constrained from higgs and top exotic decays , and other flavor constraints .
in the models we consider , the higgs boson couples to new states that are typically below the tev scale , and therefore , the models are subject to collider tests .
the new states we consider can contribute to the wavefunction renormalization of the higgs .
this affects the higgs couplings universally . among all the higgs couplings ,
a lepton collider can measure the higgs coupling to z - bosons very well .
this is accomplished in `` higgs factory '' mode by running the lepton collider at @xmath195 to @xmath196 where the higgs - strahlung cross section is maximized and the threshold for @xmath197 just opens up .
deviations in the @xmath29 coupling from the sm expectation are parametrized by @xmath33 , which was defined in .
the current lhc limit @xcite and improved sensitivities with future colliders are summarized below : @xmath198 here we show sensitivities for the high - luminosity large hadron collider ( hl - lhc ) @xcite , circular electron positron collider ( cepc ) @xcite , international linear collider at @xmath199 ( ilc-500 ) @xcite , and future circular collider study with both @xmath200 and @xmath201 measurements ( fcc - ee ) @xcite .
new charged states coupled to the higgs will radiatively affect the higgs decay rate into a pair of photons .
both current and future colliders can constrain a deviation from the sm prediction in this channel .
we summarize below the current sensitivity of the atlas experiment @xcite , the projected sensitivity of the cms experiment at the hl - lhc @xcite , and the expected sensitivity of future lepton colliders @xcite : @xmath202 the cubic self - coupling of the higgs boson @xmath47 is currently unconstrained by the lhc .
the @xmath203 and the @xmath204 options for ilc can also measure @xmath47 from the higgs - strahlung , ww fusion , and the zz fusion processes @xcite .
future hadron colliders with higher center of mass energy , for example , sppc and fcc - hh , are expected to increase the sensitivity in @xmath47 significantly @xcite , to the ball park of @xmath205 . at the same time ,
much more detailed studies are needed to produce more solid estimate by carefully taking into all possible channels and systematics .
these sensitivities are summarized as follows : @xmath206 the matter / anti - matter asymmetry of the universe may have been generated at the electroweak phase transition ; this scenario is known as electroweak baryogenesis .
( for a review , see . ) in this framework , the electroweak phase transition needs to be first order , such that phase coexistence occurs during the transition . at the boundary between phases , _
i.e. _ the bubble wall , the properties of particles and their interactions can vary rapidly .
specifically , thermal diffusion of @xmath85 chern - simons number leads to efficient violation of baryon number outside of the bubbles ( symmetric phase ) , but inside of the bubbles the weak gauge fields become massive , @xmath207 , and these processes ( electroweak sphalerons ) acquire a boltzmann suppression .
thus , a baryon asymmetry can be generated outside of the bubble and diffuse into the bubble , where it is approximately conserved .
to avoid washout of the baryon asymmetry in the higgs phase , the electroweak order parameter must satisfy @xcite @xmath208 at the temperature of the phase transition , @xmath209 . if is satisfied , the phase transition is said to be `` strongly '' first order . for our numerical results ,
we evaluate the washout criterion at the bubble nucleation temperature @xmath210 ; see .
although is a necessary condition for electroweak baryogenesis , it is not a sufficient condition .
successful generation of the baryon asymmetry also requires a source of @xmath211-violation , which is model - dependent .
additionally it requires that the bubble walls are not expanding too quickly @xcite .
these details are beyond the scope of our work .
therefore , we do not consider electroweak baryogenesis explicitly , and we do not calculate the relic baryon asymmetry .
rather , we assess whether the phase transition is strongly first order depending on whether it satisfies ; if the phase transition is not strongly first order , electroweak baryogenesis is not viable . a first order cosmological phase transition is expected to generate a stochastic background of gravitational waves @xcite . in general , the gravitational waves arise from several sources .
when bubbles of the higgs phase meet one another and collide , the localized energy density generates a quadrupole contribution to the stress - energy tensor , which sources gravitational waves .
additionally , the passage of bubbles through the plasma creates magnetohydrodynamic turbulence and sound waves .
decay of the turbulence and damping of the sound waves can continue for multiple hubble times , even after the phase transition is completed , providing additional sources of gravitational waves .
the spectrum of stochastic gravitational waves is very model - dependent .
( we provide additional details in . )
a key parameter is the ratio @xmath212 , which compares the vacuum energy density liberated in the phase transition with the energy density of the relativistic plasma .
the gravitational wave spectrum is proportional to @xmath213 , up to some efficiency factor .
writing the phase transition temperature as @xmath209 we have @xmath214 . in the standard model @xmath215 , but a first order phase transition ( due to bsm physics ) can have appreciable supercooling , typically @xmath216 .
thus the amplitude of the stochastic gravitational wave spectrum varies by many orders of magnitude across the parameter space .
conversely , the peak frequency of the gravitational wave spectrum is less model - sensitive .
the frequency of these gravitational waves today is related to the size of the higgs phase bubbles at the time of collision .
typically , this is some ( model - dependent ) fraction of the cosmological horizon at the time of the phase transition .
( the complete formulas appear in . ) for an electroweak - scale phase transition , the spectrum typically peaks in the range @xmath217 .
such a signal is potentially within reach of future space - based gravitational wave interferometers , such as elisa .
the geometry of elisa has not been finalized , and a number of designs are under investigation .
configurations differ in regard to number of links ( 4 or 6 ) , arm length ( 1 , 2 , or 5 million kilometers ) , duration ( 2 or 6 years ) , and noise level ( comparable to pathfinder or greater ) . the corresponding sensitivities to a stochastic gravitational wave background are calculated in for four configurations . in its more optimistic configurations ,
the sensitivity of elisa could reach @xmath218 in the next section , we compare the predicted gravitational wave signals in our models against these projected sensitivities .
for each of the models we scan approximately @xmath219 points in the parameter space .
the parameter ranges vary from model to model , but unless otherwise specified we generally allow the dimensionless couplings to vary between @xmath220 and @xmath98 . at each parameter point
, we calculate the relevant collider observables ( @xmath33 , @xmath221 , @xmath47 ) , the phase transition order parameter ( @xmath222 ) , and the gravitational wave spectrum . throughout this section ,
the scatter plots are color - coded : an orange point indicates a first order phase transition , a blue point indicates a strongly first order phase transition ( [ eq : washout_crit ] ) , and a green point indicates a very - strong first order phase transition with potentially detectable gravitational wave signal at elisa . in we extended the standard model by a real scalar singlet , which is able to mix with the standard model higgs .
this mixing ( in addition to radiative effects ) leads to a modification of the @xmath29 and @xmath43 couplings , parametrized by @xmath33 and @xmath39 . in
we show the distribution of models over this parameter space .
models with a first order phase transition ( orange , blue , or green points ) tend to predict a large suppression of the @xmath29 coupling , ranging from @xmath223 to as much as @xmath224 .
for the vast majority of the models in the scan , this suppression is detectable at higgs factories such as the cepc , which has an expected sensitivity of @xmath225 ( [ eq : dzh_sensitivity ] ) .
focusing on the models with a gravitational wave signal that is potentially detectable by elisa ( green points ) , the suppression of @xmath29 is larger still , typically @xmath226 .
therefore if the higgs factory measures a significant suppression of the @xmath29 coupling , it would motivate a search for relic gravitational waves with space based interferometers .
parameter space scan for the singlet model of .
an orange point indicates a first order phase transition , a blue point indicates a strongly first order phase transition ( [ eq : washout_crit ] ) , and a green point indicates a very - strong first order phase transition with potentially detectable gravitational wave signal at elisa .
the right panels shows the predicted gravitational wave spectrum today along with the projected sensitivity of elisa @xcite .
, title="fig:",height=226 ] parameter space scan for the singlet model of .
an orange point indicates a first order phase transition , a blue point indicates a strongly first order phase transition ( [ eq : washout_crit ] ) , and a green point indicates a very - strong first order phase transition with potentially detectable gravitational wave signal at elisa .
the right panels shows the predicted gravitational wave spectrum today along with the projected sensitivity of elisa @xcite .
, title="fig:",height=226 ] in the @xmath227-symmetric limit of , the only free parameters are @xmath28 , @xmath228 , and @xmath229 .
the discrete symmetry forbids a mixing between the higgs and singlet fields , and deviations in the @xmath29 and @xmath43 couplings arise first at one - loop order .
the loop - induced contribution to @xmath33 typically falls below the projected sensitivity of future higgs factories .
the region with a viable first order phase transition are shown in .
this limit of the singlet model admits strongly first order , two - step phase transitions in which the singlet field acquires a vev prior to electroweak symmetry breaking .
the density of very strong phase transitions ( green points ) is higher , in part , because of a sampling effect ; here we scan @xmath140 model parameters whereas we scan @xmath55 in .
our results broadly agree with more detailed analyses in the literature , _
e.g. _ .
parameter space scan for the singlet model of where a discrete @xmath35 symmetry forbids the higgs - singlet mixing and suppresses the bsm modification to the @xmath29 coupling .
, title="fig:",height=226 ] parameter space scan for the singlet model of where a discrete @xmath35 symmetry forbids the higgs - singlet mixing and suppresses the bsm modification to the @xmath29 coupling .
, title="fig:",height=226 ] in we discussed a limit of the scalar singlet model in which the mixing is tuned to zero .
then the @xmath29 coupling is only modified at one - loop order , as seen in .
this region of parameter space is identified on as the `` funnel '' region where @xmath230 and @xmath231 . we have performed a parameter scan focusing on this region , and we show the results in .
most of the parameter space corresponds to a weakly first order phase transition ( orange points ) . in this limit
, the phase transition occurs in two steps with the singlet first acquiring a negative vev at @xmath232 , and the electroweak symmetry is broken later @xmath233 when the singlet vev becomes positive . due to the large field excursion and the barrier provided by tree - level potential terms ( @xmath234 and @xmath61 )
there is a significant amount of supercooling , and the phase transition is very strongly first order .
however , at the zero - temperature vacuum , the model is very sm - like , and the deviation in the @xmath29 coupling is too small to probe with future higgs factories . since the model admits strongly first order phase transition , but is inaccessible to collider probes , this limit can be viewed as a new class of `` nightmare scenario . ''
parameter space scan for the singlet model of where the higgs - singlet mixing it tuned to zero .
, title="fig:",height=226 ] parameter space scan for the singlet model of where the higgs - singlet mixing it tuned to zero . ,
title="fig:",height=226 ] in we extend the sm by three scalar doublets and complex scalar singlets , which can be viewed as colorless stops and sbottoms . as the text discusses below
, we restrict to a 4-dimensional parameter space by assuming a common quartic coupling @xmath111 .
the new charged scalars contribute to the higgs diphoton decay width @xmath235 and lead to a deviation in the @xmath29 coupling , parametrized by @xmath33 .
figure [ fig : stoplike_observables ] shows the result of a scan over the 4-dimensional parameter space . in the region of parameter space with a first order phase transition ( orange , blue , and green points ) , the higgs diphoton decay width is enhanced by more than @xmath236 , and it is enhanced by more than @xmath237 in the region with a potentially detectable gravitational wave signal ( green ) . given current lhc limits ( [ eq : higgs_diphoton_sensitivity ] ) some of this parameter space is already at tension with the data .
more importantly , the projected sensitivity of figure higgs factories ( cepc , ilc-500 , fcc - ee ) is sufficient to test the entire region of parameter space where the phase transition is first order . parameter space scan for the stop - like model of .
the projected sensitivity of figure higgs factories ( cepc , ilc-500 , fcc - ee ) is sufficient to test the entire region of parameter space where the phase transition is first order ( orange , blue , & green points ) .
, title="fig:",height=226 ] parameter space scan for the stop - like model of .
the projected sensitivity of figure higgs factories ( cepc , ilc-500 , fcc - ee ) is sufficient to test the entire region of parameter space where the phase transition is first order ( orange , blue , & green points ) .
, title="fig:",height=226 ] in we extend the sm by four chiral fermions : a pair of doublets ( `` higgsinos '' ) , a triplet ( `` wino '' ) , and a singlet ( `` bino '' ) .
we vary the mass parameter @xmath145 , the wino yukawa coupling @xmath146 , and the bino yukawa coupling @xmath147 . for simplicity
we focus on the two cases , @xmath238 , @xmath239 and @xmath240 , @xmath241 .
nonzero @xmath146 allows the charginos to suppress the higgs decay rate into photons ( [ eq : fermion_ghgg ] ) .
as seen in the left panel of for @xmath163 and @xmath238 this suppression is already at tension with current limits on @xmath158 from the lhc ( [ eq : higgs_diphoton_sensitivity ] ) .
a large yukawa coupling @xmath242 is required to achieve a first order phase transition ( indicated by the dashed curves ) , and an even larger coupling @xmath243 is required for a strongly first order phase transition ( thick solid curve ) . although the phase transition is strongly first order in this region , the electroweak order parameter only gets as large as @xmath244 , and the corresponding gravitational wave signal is not within reach of elisa .
taking instead the wino yukawa coupling to vanish ( @xmath240 ) removes the tree - level interaction between higgs and charginos , and the constraint from higgs diphoton decay is avoided .
we find that a first order phase transition is possible provided that the coupling is sufficiently large , @xmath245 .
however , this model is ruled out by a large deviation in the peskin - takeuchi t - parameter . as we discuss at the end of , it may be possible to avoid constraints on @xmath105 by adding new particles to restore the custodial @xmath246 symmetry .
then the model can still be tested through precision measurements of the @xmath29 coupling .
using we estimate the modification to @xmath29 to be at the level of @xmath247 if @xmath240 and @xmath248 .
such a large deviation can potentially be probed by the high - luminosity lhc or certainly by a future higgs factory ( [ eq : dzh_sensitivity ] ) .
finally let us comment on the size of the yukawa couplings considered here .
if @xmath146 and @xmath147 are as large as @xmath98 or @xmath140 , as shown in , then the model has a landau pole not far above the electroweak scale . in principle
new particles can be added to cancel the radiative corrections from the heavy chiral fermions and raise the cutoff .
the new particles may enter higgs physics , possibly making them testable at the lhc , but certainly testable at a future @xmath249 collider .
results for the heavy chiral fermion model of .
the higgs - to - diphoton decay rate is shown as a function of the chargino mass .
we fix the higgs - bino - higgsino yukawa coupling @xmath238 and we show various values of the higgs - wino - higgsino yukawa coupling @xmath146 .
thick lines indicates parameters with a strongly first order phase transition ( [ eq : washout_crit ] ) , dashed lines indicate a weakly first order transition , and thin lines indicate a cross - over or second order transition .
, height=207 ] in we allowed the yukawa couplings of the standard model fermions to depend on the vacuum expectation value of a scalar flavon field .
provided that the flavon vev changes during the electroweak phase transition , the yukawa couplings acquire an implicit dependence on the vev of the higgs field .
in we discussed three phenomenological models : ( a ) quark yukawa couplings receive a universal ( flavor - independent ) _ shift _ and lepton yukawa couplings are fixed , ( b ) lepton couplings are shifted and quark couplings are fixed , and ( c ) lepton couplings receive a universal _ rescaling _ and quark couplings are fixed .
results of the phase transition analysis appear in where we show the electroweak order parameter @xmath222 in terms of the universal yukawa parameter @xmath250 . for cases @xmath251 and @xmath252 above , the electroweak phase transition is strongly first order for @xmath253 for varying quark yukawas and @xmath254 for varying lepton yukawas . for the same value of @xmath250
the phase transition is stronger in case @xmath251 , because more degrees of freedom have the anomalous field dependence . for case @xmath255 ,
the phase transition only becomes strongly first order for @xmath256 . in this model , all the lepton yukawa couplings are enhanced by the same factor , and consequently the electron and muon remain negligible compared to the tau .
effectively , only the one degree of freedom ( @xmath190 ) is playing any role in making the phase transition first order .
we focus on @xmath257 to avoid issues associated with loss of perturbatively . in this parameter regime ,
the predicted stochastic gravitational wave background is not within the reach of elisa s most optimistic design sensitivity .
results for the varying yukawas model of . in the top row ,
the quark yukawa couplings are allowed to vary with lepton couplings fixed to their sm values . in the middle and bottom rows ,
the lepton couplings vary , and the quark couplings are fixed . in the first two rows , the yukawa coupling changes by a flavor - independent shift , and in the bottom row it is a flavor - independent rescaling instead . , title="fig:",height=226 ] results for the varying yukawas model of . in the top row ,
the quark yukawa couplings are allowed to vary with lepton couplings fixed to their sm values . in the middle and bottom rows ,
the lepton couplings vary , and the quark couplings are fixed . in the first two rows ,
the yukawa coupling changes by a flavor - independent shift , and in the bottom row it is a flavor - independent rescaling instead . , title="fig:",height=226 ] + results for the varying yukawas model of . in the top row ,
the quark yukawa couplings are allowed to vary with lepton couplings fixed to their sm values . in the middle and bottom rows ,
the lepton couplings vary , and the quark couplings are fixed . in the first two rows , the yukawa coupling changes by a flavor - independent shift , and in the bottom row it is a flavor - independent rescaling instead . , title="fig:",height=226 ] results for the varying yukawas model of . in the top row ,
the quark yukawa couplings are allowed to vary with lepton couplings fixed to their sm values . in the middle and bottom rows ,
the lepton couplings vary , and the quark couplings are fixed . in the first two rows , the yukawa coupling changes by a flavor - independent shift , and in the bottom row it is a flavor - independent rescaling instead . , title="fig:",height=226 ] + results for the varying yukawas model of . in the top row ,
the quark yukawa couplings are allowed to vary with lepton couplings fixed to their sm values . in the middle and bottom rows ,
the lepton couplings vary , and the quark couplings are fixed . in the first two rows , the yukawa coupling changes by a flavor - independent shift , and in the bottom row it is a flavor - independent rescaling instead . , title="fig:",height=226 ] results for the varying yukawas model of . in the top row ,
the quark yukawa couplings are allowed to vary with lepton couplings fixed to their sm values . in the middle and bottom rows ,
the lepton couplings vary , and the quark couplings are fixed . in the first two rows , the yukawa coupling changes by a flavor - independent shift , and in the bottom row it is a flavor - independent rescaling instead . , title="fig:",height=226 ]
in this paper we have explored a few minimal extensions of the standard model in which new particles below the tev scale cause the electroweak phase transition to become first order . although the new particle content is motivated from a bottom - up and minimalist perspective , this new physics
can easily be embedded in a broader uv theory such as supersymmetry . despite their simplicity ,
our models exhibit various different mechanisms giving rise to a first order phase transition @xcite . for instance , the value of the scalar singlet field can change along with the higgs field thereby inducing a first order phase transition through tree - level interactions , or the presence of stop - like scalar particles in the electroweak plasma can lead to a first order transition via thermal effects . despite this diversity of particle content , phenomenology , and phase transition dynamics
, we find a generic relationship between models with a first order electroweak phase transition and those that are testable at colliders .
namely , in the region of parameter space where the electroweak phase transition becomes first order , we typically find such a large deviation in the higgs coupling with z - bosons ( @xmath29 ) that it should be discovered by a future higgs factory .
it generically predicts a large , @xmath46 , deviation in the triple higgs coupling , which can be measured well by next generation hadron collider and high energy lepton collider , such as the @xmath204 version of the ilc .
the models can also be probed by the higgs diphoton decay rate , and other methods .
while future colliders may shed light on the dynamics behind electroweak symmetry breaking , the most direct probe of a cosmological first order phase transition is the associated stochastic background of gravitational waves .
we have calculated the spectrum of gravitational waves that would have been produced during a first order electroweak phase transition in the early universe .
we find that only models with especially strong first order phase transitions , typically @xmath258 , are within reach of a future space - based gravitational wave interferometer experiment like elisa .
nevertheless , models with weaker first order transitions can still be probed by future colliders .
this provides the exciting opportunity to make complimentary measurements of the electroweak phase transition using a combination of next - generation colliders and interferometers .
detecting signals of first order electroweak phase transition in both cosmological observation and collider experiments would be an achievement , which can rival that of the bbn , and open a new page in our understanding of early universe .
+ * acknowledgments : * we are grateful to michael fedderke for various discussions of the heavy fermion model .
we thank jing shu and carlos wagner for interesting discussions .
work at anl is supported in part by the u.s .
department of energy under contract no .
de - ac02 - 06ch11357 .
is partially supported by u.s .
department of energy grant de - fg02 - 04er41286 .
ajl is supported at the university of chicago by the kavli institute for cosmological physics through grant nsf phy-1125897 and an endowment from the kavli foundation and its founder fred kavli .
ltw is supported by doe grant de - sc0013642 .
for each of the models , we use the thermal effective potential to calculate the parameters of the phase transition . for a pedagogical approach to this calculation , see the lectures in . for additional details see appendices f and g of . here
we enumerate the various assumptions and approximations .
we calculate the thermal effective potential in the one - loop approximation .
the zero - temperature quantum contribution takes the form of the well - known coleman - weinberg potential .
the coleman - weinberg potential is dominated by the particles with largest coupling to the higgs .
thus we include the top and anti - top quarks , the weak gauge bosons , the higgs boson , and model - dependent new physics .
we neglect the contributions from leptons and lighter quarks , while photons and gluons do not contribute at one - loop order .
we work in the landau gauge where the ghosts are massless and do not contribute to the effective potential .
we regulate the one - loop ultraviolet divergences in the dimensional regularization scheme .
we renormalize at the scale @xmath259 by requiring all possible derivatives of the one - loop correction to vanish at the vacuum .
that is , we determine the counterterms by imposing @xmath260 , @xmath261 , and so on .
( see also appendix f of . ) in this way , the zero - temperature correction to the one - loop thermal effective potential is determined .
the finite - temperature correction is evaluated from the bosonic and fermionic thermal functions using their representation as a sum over bessel functions .
we truncate the sums at @xmath262 , which is a good approximation to the full result .
we neglect the daisy resummation . for the models in which the effective potential has a barrier already at @xmath263 , such as the singlet models
, we expect this to be a good approximation .
the phase transition is studied by varying the thermal effective potential with respect to temperature . for some range of temperatures ( possibly including @xmath263 ) the potential displays a barrier separating a local minimum with @xmath264 ( symmetric phase ) from a local minimum with @xmath265 ( broken or higgs phase ) .
we construct an interpolating trajectory between these local minima ( see appendix g of ) and calculate the @xmath266-symmetric bounce solution and its action @xmath267 along the one - dimensional trajectory . the phase transition is said to begin when the bubble nucleation rate @xmath268 exceeds the hubble expansion rate @xmath269 . the condition @xmath270 evaluates to @xmath271 and we define the bubble nucleation temperature @xmath272 such that @xmath273 .
the phase transition continues until an @xmath46 fraction of the universe is in the higgs phase , and we identify this as the percolation temperature @xmath274 .
after percolation completes , the energy stored in the bubble walls is transferred back to the plasma as heat , and the temperature increases to @xmath275 . for simplicity
, we will evaluate @xmath272 and assume @xmath276 .
for a more sophisticated approach , see .
the phase transition is characterized by two parameters : @xmath277 and @xmath278 . in the broken phase at temperature @xmath209 ,
the vacuum energy @xmath279 is calculated as the value of the zero - temperature effective potential at the location of the minimum of the thermal effective potential .
the vacuum energy in the unbroken phase @xmath280 follows from a similar calculation .
then the dimensionless parameter @xmath281 quantifies the energy liberated ( latent heat ) . here
, @xmath282 is the radiation energy density in the broken phase at the temperature of the phase transition .
if reheating is negligible , as we have assumed , then @xmath283 , and larger @xmath277 corresponds to a stronger phase transition with greater supercooling .
the second parameter @xmath278 quantifies the duration of the phase transition . during the phase transition
, the bubble nucleation rate has an exponential time dependence @xmath284 where @xmath285 is the phase transition duration .
the exponent is calculated as @xmath286 this affects the size of bubbles at the time of collision , and therefore it factors into the gravitational wave spectrum .
we must have @xmath287 otherwise the phase transition would not occur .
we calculate the spectrum of gravitational waves generated at the electroweak phase transition by following , which summarizes the results of various original sources .
gravitational waves at a first order phase transition are produced in three ways : from the collision of bubbles , from the decay of magnetohydrodynamic turbulence , and from the propagation of sound waves .
the spectrum of gravitational waves produced by bubble collisions is given by the envelope approximation to be @xmath288 where @xmath289 is the fraction of the liberated vacuum energy transferred into motion of the bubble wall and @xmath290 is the speed of the wall .
the spectrum peaks at a frequency @xmath291 given by @xmath292 the decay of magnetohydrodynamic turbulence contributes to the gravitational wave spectrum as @xmath293 where @xmath294 is the fraction of energy transferred to turbulent motions of the plasma , and @xmath295 this spectrum peaks at a frequency @xmath296 given by @xmath297 finally , sound waves contribute to the gravitational wave spectrum as @xmath298^{7/2 } } \com\end{aligned}\ ] ] where @xmath299 is the fraction of energy transferred to bulk motions of the fluid .
the spectrum peaks at @xmath300 note that @xmath301 and @xmath302 are larger than @xmath303 by a factor of @xmath304 .
this is because gravitational wave production from plasma effects can continue even after bubble collisions have completed
. the efficiency factors ( @xmath289 s ) are model - dependent and vary with the strength of the phase transition .
if the phase transition is weakly first order , the energy in the scalar field is depleted by interactions with the plasma , and it reaches a terminal velocity ( possibly relativistic ) . then the energy transferred to the plasma ( @xmath294 and @xmath299 ) is limited by the vacuum energy that was initially available in the field , which is parametrized by @xmath277 . on other hand , for a strongly first order transition the pressure gradient drives the bubble wall to expand and `` run away '' with @xmath305 . in this regime , the amount of energy transferred to the plasma saturates , and the surplus energy causes the bubble wall to accelerate .
consequently , for large @xmath277 we have @xmath306 and @xmath307 .
since only the strongest phase transitions will be detectable by elisa , we can simplify by assuming that all models lead to runaway phase transitions . in this way
, we accurately assess the gravitational wave signal for the strongest phase transitions .
following , the following expressions parametrize the fraction of energy that is carried by the bubble wall , that is carried by the bulk motion of the plasma , and that is lost as heat : @xmath308 a fraction @xmath309 of the energy carried by bulk motions is transferred to mhd turbulence , @xmath310 .
simulations motivate @xmath311 and @xmath312 .
we also take @xmath313 .
these expressions are valid for sufficiently strong phase transitions , @xmath314 where @xmath315 we calculate @xmath209 , @xmath316 , @xmath277 , @xmath317 , and @xmath318 from the one - loop thermal effective potential , as described in .
m. donofrio and k. rummukainen , phys . rev . *
d93 * , 025003 ( 2016 ) , arxiv:1508.07161 . c. caprini _ et al .
_ , jcap * 1604 * , 001 ( 2016 ) , arxiv:1512.06239 .
d. j. h. chung , a. j. long , and l .- t .
wang , phys . rev . *
d87 * , 023509 ( 2013 ) , arxiv:1209.1819 .
a. v. kotwal , m. j. ramsey - musolf , j. m. no , and p. winslow , ( 2016 ) , arxiv:1605.06123 .
l. leitao and a. megevand , ( 2015 ) , arxiv:1512.08962 .
f. p. huang , p .- h .
gu , p .- f .
yin , z .- h .
yu , and x. zhang , ( 2015 ) , arxiv:1511.03969 .
f. p. huang , y. wan , d .-
wang , y .- f .
cai , and x. zhang , ( 2016 ) , arxiv:1601.01640 .
n. craig , c. englert , and m. mccullough , phys .
. lett . * 111 * , 121803 ( 2013 ) , arxiv:1305.5251 .
d. curtin , p. meade , and c .- t .
yu , jhep * 11 * , 127 ( 2014 ) , arxiv:1409.0005 . j. fan , m. reece , and l .- t .
wang , jhep * 08 * , 152 ( 2015 ) , arxiv:1412.3107 .
m. mccullough , phys . rev . *
d90 * , 015001 ( 2014 ) , arxiv:1312.3322 , .
chen , s. dawson , and i. m. lewis , phys .
rev . * d91 * , 035015 ( 2015 ) , arxiv:1410.5488 .
atlas , g. aad _
_ , jhep * 11 * , 206 ( 2015 ) , arxiv:1509.00672 .
cms , c. collaboration , ( 2016 ) .
a. ashoorioon and t. konstandin , jhep * 07 * , 086 ( 2009 ) , arxiv:0904.0353 .
j. r. espinosa and m. quiros , phys
. rev . * d76 * , 076004 ( 2007 ) , arxiv : hep - ph/0701145 . atlas , g. aad _
et al . _ ,
b749 * , 519 ( 2015 ) , arxiv:1506.05988 .
t. melia , k. melnikov , r. rontsch , and g. zanderighi , jhep * 12 * , 053 ( 2010 ) , arxiv:1007.5313 .
s. group , ( 2015 ) .
d. curtin , p. jaiswal , and p. meade , jhep * 08 * , 005 ( 2012 ) , arxiv:1203.2932 .
t. cohen , d. e. morrissey , and a. pierce , phys . rev . *
d86 * , 013009 ( 2012 ) , arxiv:1203.2924 .
m. carena , g. nardini , m. quiros , and c. e. m. wagner , jhep * 02 * , 001 ( 2013 ) , arxiv:1207.6330 .
w. huang , j. shu , and y. zhang , jhep * 03 * , 164 ( 2013 ) , arxiv:1210.0906 .
g. burdman , z. chacko , h .- s .
goh , and r. harnik , jhep * 02 * , 009 ( 2007 ) , arxiv : hep - ph/0609152 .
m. carena , m. quiros , and c. e. m. wagner , phys .
b380 * , 81 ( 1996 ) , arxiv : hep - ph/9603420 .
a. djouadi , v. driesen , w. hollik , and j. i. illana , eur .
j. * c1 * , 149 ( 1998 ) , arxiv : hep - ph/9612362 . m. carena , s. gori , n. r. shah , and c. e. m. wagner , jhep * 03 * , 014 ( 2012 ) , arxiv:1112.3336 .
m. carena , a. megevand , m. quiros , and c. e. m. wagner , nucl .
b716 * , 319 ( 2005 ) , arxiv : hep - ph/0410352 .
n. craig , m. farina , m. mccullough , and m. perelstein , jhep * 03 * , 146 ( 2015 ) , arxiv:1411.0676 . m. a. fedderke , t. lin , and l .- t .
wang , jhep * 04 * , 160 ( 2016 ) , arxiv:1506.05465 .
i. baldes , t. konstandin , and g. servant , ( 2016 ) , arxiv:1604.04526 .
i. baldes , t. konstandin , and g. servant , ( 2016 ) , arxiv:1608.03254 . atlas , cms , g. aad _
et al . _ , ( 2016 ) ,
cms , , in _ community summer study 2013 : snowmass on the mississippi ( css2013 ) minneapolis , mn , usa , july 29-august 6 , 2013 _ , 2013 , arxiv:1307.7135
. k. fujii _
et al . _ , ( 2015 ) , arxiv:1506.05992 .
d. denterria , frascati phys .
ser . * 61 * , 17 ( 2016 ) , arxiv:1601.06640 .
w. yao , , in _ proceedings , community summer study 2013 : snowmass on the mississippi ( css2013 ) : minneapolis , mn , usa , july 29-august 6 , 2013 _ , 2013 , arxiv:1308.6302 .
a. j. barr , m. j. dolan , c. englert , d. e. ferreira de lima , and m. spannowsky , jhep * 02 * , 016 ( 2015 ) , arxiv:1412.7154 .
a. azatov , r. contino , g. panico , and m. son , phys . rev . *
d92 * , 035001 ( 2015 ) , arxiv:1502.00539 .
he , j. ren , and w. yao , phys . rev .
* d93 * , 015003 ( 2016 ) , arxiv:1506.03302 .
p. huang , a. joglekar , b. li , and c. e. m. wagner , phys .
rev . * d93 * , 055049 ( 2016 ) , arxiv:1512.00068 .
a. riotto and m. trodden , ann.rev.nucl.part.sci . *
49 * , 35 ( 1999 ) , arxiv : hep - ph/9901362 .
j. kozaczuk , jhep * 10 * , 135 ( 2015 ) , arxiv:1506.04741 . c. j. hogan , mon . not .
* 218 * , 629 ( 1986 ) .
m. kamionkowski , a. kosowsky , and m. s. turner , phys . rev . *
d49 * , 2837 ( 1994 ) , arxiv : astro - ph/9310044 .
m. quiros , , in _ high energy physics and cosmology .
proceedings , summer school , trieste , italy , june 29-july 17 , 1998 _ , pp . 187259 , 1999 , arxiv : hep - ph/9901312 .
d. j. h. chung and a. j. long , phys . rev .
* d84 * , 103513 ( 2011 ) , arxiv:1108.5193 .
l. dolan and r. jackiw , phys . rev . * d9 * , 3320 ( 1974 ) .
n. k. nielsen , nucl .
b101 * , 173 ( 1975 ) .
h. h. patel and m. j. ramsey - musolf , jhep * 07 * , 029 ( 2011 ) , arxiv:1101.4665 .
m. garny and t. konstandin , jhep * 07 * , 189 ( 2012 ) , arxiv:1205.3392 . c. wainwright , s. profumo , and m. j. ramsey - musolf , phys
. rev . * d84 * , 023521 ( 2011 ) , arxiv:1104.5487 . c. l. wainwright , s. profumo , and m. j. ramsey - musolf , phys
. rev . * d86 * , 083537 ( 2012 ) , arxiv:1204.5464 . | after the discovery of the higgs boson , understanding the nature of electroweak symmetry breaking and the associated electroweak phase transition has become the most pressing question in particle physics . answering this question is a priority for experimental studies .
data from the lhc and future lepton collider - based higgs factories may uncover new physics coupled to the higgs boson , which can induce the electroweak phase transition to become first order .
such a phase transition generates a stochastic background of gravitational waves , which could potentially be detected by a space - based gravitational wave interferometer . in this paper
, we survey a few classes of models in which the electroweak phase transition is strongly first order .
we identify the observables that would provide evidence of these models at the lhc and next - generation lepton colliders , and we assess whether the corresponding gravitational wave signal could be detected by elisa .
we find that most of the models with first order electroweak phase transition can be covered by the precise measurements of higgs couplings at the proposed higgs factories .
we also map out the model space that can be probed with gravitational wave detection by elisa .
* probing the electroweak phase transition + with higgs factories and gravitational waves * peisi huang,@xmath0 andrew j. long,@xmath1 and lian - tao wang@xmath2 _ @xmath3enrico fermi institute , university of chicago , chicago , illinois 60637 , usa .
_ _ @xmath4high energy physics division , argonne national laboratory , argonne , illinois 60439 , usa _ _ @xmath5kavli institute for cosmological physics , university of chicago , chicago , illinois 60637 , usa _ |
the determination of helium abundances in stars has been a long - standing problem in astronomy . for the old stars in globular clusters and in the galactic halo ,
it is only possible to spectroscopically measure the helium abundance in a direct way for hot horizontal branch ( hb ) stars .
studies of this kind ( e.g. moehler , heber , & durell 1997 ; moehler , heber , & rupprecht 1997 ) have found abundances significantly higher _ and _ lower than the primordial helium abundance determined from observations of low - metallicity extragalactic h ii regions ( e.g. @xmath5 from olive , skillman , & steigman 1997 ) .
the interpretation typically given for the low helium abundances is that gravitational settling has acted preferentially on helium atoms in the atmospheres of these stars , rendering the measurements unusable for determining the initial helium abundance of the stars .
high helium abundances may be the result of mass loss before the hb phase , which could remove most of the hydrogen - rich envelope of the stars .
indirect methods must be used to measure the helium abundance in earlier phases of stellar evolution .
the most useful methods are applied to globular clusters because individual clusters provide us with large samples of stars with seemingly identical compositions and ages .
globular cluster stars also seem to be among the oldest that can be found in the galaxy , so that if the helium abundance could be determined , it would give us a hint of the primordial helium abundance .
there are two other reasons for examining the helium abundances of globular clusters .
first , it is important to look at the overall trend in the helium abundance as a function of the cluster metal content .
this provides a check of our understanding of nucleosynthesis and galactic evolution .
in addition , previous studies of rr lyrae stars have indicated that an anticorrelation of helium abundance with metal content can explain the sandage period - shift effect ( sandage 1982 ) .
second , it is important to see if it is possible to detect helium variations among clusters of similar metallicity , since helium remains a plausible candidate for the `` second parameter '' determining the morphology of the hb . while age seems to play a role in some clusters , it appears to be inadequate to explain all of the variations seen . in the past
, clusters have been marked as possibly having high helium abundances ( e.g. , dickens 1972 ) , but to date there has not been definitive evidence of a relative helium abundance difference .
we will examine three different helium abundance indicators that can be measured from cluster photometry .
we do this in the hopes that anomalous clusters will appear with unusual values in two or more of the parameters , so as to provide better evidence for variations .
a cursory examination of older tabulations of data reveals an additional reason for recomputing values for these indicators : the trend of helium abundance with metallicity appears to depend on the method used ( caputo & castellani 1983 ) .
the three indicators have been discussed to varying degrees in earlier studies : @xmath6 ( iben 1968 ; buzzoni et al 1983 ; caputo , martinez roger , & paez 1987 ) , @xmath3 ( caputo , cayrel , & cayrel de strobel 1983 ; hereafter , ccc ) , and @xmath2 ( carney 1980 ; ccc ) .
photometric studies of globular clusters have become rather numerous in the past decade , so that the size and quality of the dataset for each parameter can be considerably increased . where it has been possible , we have also analyzed data for old globular clusters in the magellanic clouds , and for stars in local group dwarf spheroidals with old populations . in the following sections , we discuss each parameter in turn their definitions , the datasets , potential errors in measurement , and calibration . in the final two sections , we compare helium abundance values derived from the different indicators , and
examine the evidence for trends as a function of [ fe / h ] .
for all three of the indicators discussed below , higher values imply higher helium abundances .
this indicator is simply defined as the ratio @xmath7 , where @xmath8 is the number of horizontal branch stars , and @xmath9 is the number of rgb stars brighter than the luminosity level of the hb .
this primarily reflects the dependence of the hydrogen - burning shell s progress on the hydrogen content of the envelope material being fed into it ( and to a lesser extent , the change of helium core mass at helium flash , which affects the hb luminosity , and hence , the hb lifetime ) .
because the ratio is computed using the brightest stars in each cluster , it can be calculated for any cluster having deep enough photometry . however , because these stars are relatively scarce in globular clusters , it usually requires wide - field data , or accurate photometry of the cluster core .
we have restricted our sample to those clusters with photometric samples of at least a total of 100 hb and rgb stars .
in addition , we require photometry of sufficient quality to separate the hb , rgb , and agb populations readily , and to eliminate field stars from the sample when proper motion data is not available . we have chosen to use the ratio @xmath6 rather than the similar ratio @xmath10 , which would be more easily measured in most cases , because @xmath6 is more straightforward to calculate theoretically .
for this reason , the stellar sample was sometimes restricted in radius in order to remove portions of the cluster center where image blending made identifications of hb , rgb , and hb stars difficult . in
re - examining the determinations of @xmath6 , we found that the vast majority of earlier calculations used the `` average '' differential bolometric correction @xmath11 of 0.15 mag to determine the faint end of the rgb sample .
this value was used in the original buzzoni et al .
( 1983 ) paper , but only because uncertainties of various sorts made the correction unimportant .
however , the correction is actually a function of metallicity , and it is applied to the faint end of the rgb sample , which means it can be the source of significant systematic error . to update @xmath12 , we have examined the hb models of dorman ( 1992 ) in conjunction with the isochrones of bergbusch & vandenberg ( 1992 ; hereafter bv92 ) .
these stellar models were computed with a consistent set of physics and compositions .
although the composition is somewhat out of date ( it does not include full @xmath13-element enhancement ) , the differential bolometric corrections should be satisfactory because they are differential in nature , and the oxygen abundance does not affect the color of the giant branch relative to the instability strip . we have determined the corrections as a function of [ fe / h ] , and have derived the following fitting formula : @xmath14 } + 0.229 \mbox{[m / h]}^{2 } + 0.034 \mbox{[m / h]}^{3 } .\ ] ] because @xmath13-element enhancements influence the position of the hb and rgb in the cmd like a change in [ fe / h ] ( salaris , chieffi , & straniero 1993 ) , they must be taken into account when computing [ m / h ] .
to do so , we assumed a constant @xmath13-element enhancement of 0.3 dex for all of the globular clusters examined ( carney 1996 ) .
[ the halo field population is believed to show a different abundance pattern as a function of metallicity ( wheeler , sneden , & truran 1989 ) . ] the @xmath13-element abundance was taken to contribute to the effective metal content of the cluster as 0.7 [ @xmath13/fe ] ( salaris , chieffi , & straniero 1993 ) .
the contribution does not go directly as [ @xmath13/fe ] because oxygen , the most abundant @xmath13 element , has a relatively high ionization potential . as a result
, it does not contribute as significantly to the opacity in the envelope of rgb stars as it does for higher temperature hb stars .
error in observed @xmath6 values comes from two sources : error in the numbers of hb and rgb stars due to misidentification or poisson fluctuations , and error in the determination of the faint limit for the rgb sample ( resulting from the determination of the magnitude level of the hb , or the metallicity uncertainty that affects the size of the differential bolometric correction ) .
the magnitude level of the hb was taken to be the average magnitude of stars in the instability strip . for clusters with many well - studied rr lyrae stars , this
could be taken from the average magnitude of the variables . often though the rr lyrae stars were not observed frequently enough to derive good average magnitudes .
if the cluster had populated red and blue edges of the intability strip , a linear interpolation between the edges was made . in cases where the cluster had only a populated red or blue hb ,
a theoretical correction was made using the models of dorman ( 1992 ) to @xmath15 using the cluster reddening and a well - determined point on the populated portion of the hb .
dorman s scaled - solar abundance hb models were used with a correction for enhanced @xmath13-element abundances as described above .
an examination of the table indicates that the quoted errors for the hb magnitudes for the clusters are in general less than 0.05 mag .
this is because we are not interested in the _ absolute _ value of @xmath16 for any of the clusters only the value for the sample used .
it is clearly unnecessary for the photometric calibration of the sample to be perfect , since we only require that the relative photometry of the hb and rgb be good enough that the faint limit of the rgb sample can be determined well . for this reason ,
the tabulated values of @xmath16 should not be taken as being good absolute values .
this column is provided to make verification of the results easier .
the error due to the differential bolometric correction is illustrative , so we present a semi - analytic model for its contribution to the error . the important quantity to compute is @xmath17 .
this can be derived assuming that i ) the differential luminosity function ( lf ) of the rgb has a constant slope @xmath18 , and ii ) the number of red giant branch stars goes to zero at the tip of the rgb ( trgb ) .
so , we use @xmath19 , and find @xmath20 where @xmath21 is the normalization of the luminosity function . from this
we find @xmath22 n_{rgb } .\ ] ] the fraction in brackets is close to one , and from bergbusch & vandenberg ( 1992 ) lfs we find @xmath23 nearly independent of composition . from this
we find that for each tenth of a magnitude added to the faint limit of the rgb sample , the number of rgb stars increases by about 9% .
( alternately , making the faint limit fainter by one magnitude for a given cluster area increases the total number of red giants by approximately 85% . )
this demonstrates the importance of accurate values for the differential bolometric correction @xmath12 .
the above derivative affects the error in @xmath9 through errors in metallicity and in the determination of the hb magnitude . because of the metallicity dependence of the differential bolometric correction , helium determinations for high metallicity clusters will be relatively more uncertain
a polynomial fit to the derivative gives @xmath24 } + 0.0257\mbox{[m / h]}^{2 } .\ ] ] we note that the error from this source goes as @xmath9 , whereas the poisson counting errors go as @xmath25 .
we have used [ fe / h ] values from the tabulation of djorgovski ( 1993 ) in computing the values of @xmath12 for the clusters .
djorgovski s values fairly closely follow the values of zinn & west ( 1984 ; hereafter , zw ) .
we have also considered an alternative metallicity scale since this affects the differential bolometric corrections .
using metallicities from carretta & gratton ( 1996 ) , we recomputed the @xmath6 values for 16 clusters from their sample .
the results are shown in fig .
[ ratcomp ] . for the most part ,
the values were little affected by the changes in the scale .
computed @xmath6 values are insensitive to factors that merely shift the rgb in temperature or color ( age , for example ) .
factors that cause changes in the absolute brightness of the hb ( such as mean mass of the helium cores of stars , and cno abundances ) will cause systematic errors in measured values for individual clusters .
our sample is composed of 42 galactic globular clusters ( ggcs ) and 5 globulars in the large magellanic cloud ( lmc ) , and is presented in table [ tratio ] .
the listings are grouped in bins of approximately 0.2 in [ fe / h ] .
references carrying a `` ( pm ) '' designation were used to remove field stars from the sample using proper motion information .
there tend to be few clusters at the very metal - rich end due to a combination of substantial field star contamination and confusion between rgb and red hb cluster stars .
@lclcccll@ i d ( ngc / ic ) & @xmath8 & @xmath16 & @xmath12 & @xmath9 & @xmath26 & @xmath6 & references + + 5927 & 134 & @xmath27 & 0.66 & 168 & 20 & @xmath28 & samus et al 1996 , + & & & & & & & sarajedini & norris 1994 + & & & & & & & + 6496 & 69 & @xmath29 & 0.62 & 53 & 5 & @xmath30 & richtler et al 1994 + 6624 & 126 & @xmath31 & 0.63 & 177 & 30 & @xmath32 & richtler et al 1994 + 6637 ( m69 ) & 127 & @xmath33 & 0.53 & 101 & 21 & @xmath34 & sarajedini & norris 1994 + & & & & & & & + 104 ( 47 tuc ) & 368 & @xmath35 & 0.49 & 307 & 38 & @xmath36 & lee 1977a + & & & & & & & + 6171 ( m107 ) & 117 & @xmath37 & 0.41 & 135 & 29 & @xmath38 & ferraro et al 1991 , + & & & & & & & dickens & rolland 1972 , + & & & & & & & cudworth et al . 1992 ( pm ) + 6652 & 75 & @xmath39 & 0.41 & 61 & 20 & @xmath40 & ortolani et al 1994 + & & & & & & & + 6712 & 122 & @xmath41 & 0.41 & 102 & 19 & @xmath42 & cudworth 1988 ( pm ) + 6723 & 101 & @xmath43 & 0.38 & 79 & 15 & @xmath44 & menzies 1974 + & & & & & & & + 362 & 94 & @xmath45 & 0.35 & 83 & 14 & @xmath46 & harris 1982 + 1261 & 148 & @xmath47 & 0.34 & 115 & 26 & @xmath48 & ferraro et al 1993 + 1851 & 209 & @xmath49 & 0.33 & 174 & 24 & @xmath50 & walker 1992a + 2808 & 247 & @xmath51 & 0.33 & 249 & 22 & @xmath52 & ferraro et al 1990 + 6121 ( m4 ) & 114 & @xmath53 & 0.34 & 112 & 24 & @xmath54 & cudworth & rees 1990 ( pm ) + 6266 ( m62 ) & 114 & @xmath55 & 0.34 & 80 & 18 & @xmath56 & caloi et al 1987 + & & & & & & & + 288 & 100 & @xmath57 & 0.33 & 85 & 20 & @xmath58 & buonanno et al 1984b + 4499 & 145 & @xmath59 & 0.31 & 112 & 23 & @xmath48 & ferraro et al 1995 + 5904 ( m5 ) & 555 & @xmath60 & 0.33 & 515 & 94 & @xmath61 & sandquist et al 1996 + & & & & & & & cudworth 1979 ( pm ) + & & & & & & & rees 1993 ( pm ) + 6229 & 92 & @xmath62 & 0.31 & 100 & 19 & @xmath63 & carney et al 1991 + 6681 ( m70 ) & 82 & @xmath64 & 0.31 & 50 & 8 & @xmath65 & brocato et al 1996b + 6752 & 225 & @xmath66 & 0.31 & 144 & 13 & @xmath67 & buonanno et al 1986 + 6934 & 55 & @xmath68 & 0.31 & 62 & 12 & @xmath69 & brocato et al 1996b + 6981 ( m72 ) & 45 & @xmath70 & 0.31 & 56 & 10 & @xmath71 & brocato et al 1996b + 7006 & 96 & @xmath72 & 0.30 & 62 & 24 & @xmath73 & buonanno et al 1991 + & & & & & & & + 1904 ( m79 ) & 122 & @xmath74 & 0.29 & 114 & 16 & @xmath75 & ferraro et al 1992b + 3201 & 179 & @xmath76 & 0.30 & 150 & 14 & @xmath77 & lee 1977b + 5272 ( m3 ) & 562 & @xmath78 & 0.29 & 473 & 65 & @xmath79 & ferraro et al 1997 + 5897 & 68 & @xmath80 & 0.29 & 50 & 8 & @xmath81 & ferraro et al 1992a + 6093 ( m80 ) & 170 & @xmath82 & 0.30 & 174 & 39 & @xmath83 & alcaino et al 1998 + 6205 ( m13 ) & 90 & @xmath84 & 0.29 & 88 & 12 & @xmath85 & cohen et al 1997 + 6218 ( m12 ) & 91 & @xmath86 & 0.30 & 56 & 12 & @xmath87 & brocato et al 1996b + 6254 ( m10 ) & 69 & @xmath88 & 0.30 & 71 & 13 & @xmath89 & harris et al 1976 + rup 106 & 65 & @xmath90 & 0.29 & 46 & 12 & @xmath91 & buonanno et al 1993 + & & & & & & & + 5694 & 56 & @xmath92 & 0.27 & 43 & 14 & @xmath93 & ortolani & gratton 1990 + 6397 & 105 & @xmath94 & 0.27 & 91 & 10 & @xmath95 & kaluzny 1997 + 6809 ( m55 ) & 203 & @xmath96 & 0.28 & 197 & 37 & @xmath97 & lee 1977c + & & & & & & & + 4590 ( m68 ) & 108 & @xmath98 & 0.26 & 119 & 34 & @xmath99 & walker 1994 + 5024 ( m53 ) & 302 & @xmath100 & 0.27 & 278 & 39 & @xmath101 & rey et al .
1998 + 5053 & 53 & @xmath102 & 0.25 & 58 & 10 & @xmath103 & sarajedini & milone 1995 + 5466 & 86 & @xmath104 & 0.26 & 71 & 12 & @xmath105 & buonanno et al 1984a + 6341 ( m92 ) & 140 & @xmath106 & 0.26 & 92 & 20 & @xmath107 & buonanno et al 1983b + 7078 ( m15 ) & 153 & @xmath108 & 0.26 & 124 & 23 & @xmath109 & buonanno et al 1983a + 7099 ( m30 ) & 202 & @xmath110 & 0.27 & 124 & 11 & @xmath111 & sandquist et al 1998 , + & & & & & & & yanny et al 1994 + @lclcccll@ i d ( ngc / ic ) & @xmath8 & @xmath16 & @xmath12 & @xmath9 & @xmath26 & @xmath6 & references + + 1466 & 91 & @xmath112 & 0.28 & 96 & 36@xmath113 & @xmath114 & walker 1992c + 1841 & 165 & @xmath115 & 0.26 & 158 & 32 & @xmath116 & brocato et al 1996c + 2210 & 67 & @xmath117 & 0.27 & 70 & 15 & @xmath118 & brocato et al 1996c + 2257 & 164 & @xmath119 & 0.28 & 138 & 24 & @xmath77 & testa et al 1995 + reticulum & 46 & @xmath120 & 0.29 & 54 & 12 & @xmath121 & walker 1992b + notes : @xmath122 a theoretical correction has been applied to compute the level of the instability strip .
@xmath123 the possibility of incompleteness exists at the faint end of the sample .
@xmath113 the number may be affected field star contamination .
the variation of the @xmath6 values as a function of hb type is shown in fig .
. there may be a slight decrease in the average @xmath6 for clusters with bluer hb type .
several clusters with the bluest morphologies have unusually high @xmath6 values : ngc 6752 , ngc 7099 ( m30 ) , ngc 6341 ( m92 ) , ngc 6218 ( m12 ) , and ngc 6681 ( m70 ) .
however , this is not universal there are more clusters in the same range of @xmath124 having @xmath6 values that are closer to the galactic average .
there is not a clear reason for this difference : well - measured clusters are found in both groups , and a metallicity difference does not appear to be present . an important effect of the inclusion of the metallicity dependence of @xmath12 is the improved agreement between the @xmath6 values of several of the most metal - rich clusters and the metal - poor clusters . in a similar fashion , the increased @xmath12 values will also significantly reduce the estimates of the helium abundances of galactic bulge fields ( @xmath125 according to minniti 1995 ) because of the high mean metallicity of those stars . the indicator @xmath2 ( ccc ) is defined simply as the magnitude difference between the ms at @xmath126 and the hb at the instability strip .
ccc originally defined the hb point to be at the blue edge of the instability strip , but we have chosen to revise this definition to make it more easily calculable theoretically and observationally .
this also reduces the color difference between the hb and ms points , reducing possible systematic effects from photometric calibration .
the various sensitivities of the indicator are not significantly changed by the revision to the definition .
increases in envelope helium abundance influence this indicator by increasing the luminosity of the hb ( via the strength of hydrogen shell - burning during that phase ) and by increasing the effective temperature and the luminosity of the ms through a decrease in the envelope opacity .
the net effect of the luminosity and temperature changes on the ms is to make it fainter at a _ given _ color . as an indicator
, @xmath2 has the advantage of being sensitive to the helium abundance ( @xmath127 mag ; ccc ) , and the disadvantages of having a definite metallicity dependence ( @xmath128 } \approx -0.5 $ ] mag / dex ) and of requiring photometry from the hb to well below the ms turnoff .
there is an additional disadvantage in requiring the knowledge of the cluster s reddening to determine @xmath129 .
however , @xmath2 has no dependence on age , since the chosen ms point reaches unevolved stars .
we can calculate theoretical values of @xmath2 using a self - consistent set of isochrones ( bv92 ) and hb models ( dorman 1992 ) having @xmath130 .
we have fitted the following polynomial to the data after making small corrections for different initial helium content in the theoretical models : @xmath131 } - 0.7938 ( \mbox{[m / h]})^{2}\ ] ] @xmath132})^3 .\ ] ] as can be seen from table [ tdelta ] , there are only 20 clusters with enough information to compute a value of @xmath2 .
we included clusters for which @xmath16 and @xmath129 were derived from different studies , in spite of the possibility of zero point differences .
the error budget for @xmath2 was computed using : @xmath133 @xmath134 where the last two terms account for reddening and zero point uncertainties respectively . as was discussed in
[ secr ] , we do not need to worry about how well calibrated the data are in an absolute sense , as long as there are not significant systematic errors in the relative photometry for the ms and hb ( for instance , a nonlinearity in one of the ccds ) . [ cols="<,<,^,^,^,<,<",options="header " , ] the derived mean abundance is considerably lower than the expected primordial value @xmath135 . because we believe we have removed the most important ( known ) systematic errors in the @xmath6 measurements , we should ask what might cause the @xmath136 values to be low .
brocato , castellani , & villante ( 1998 ) suggest that the uncertainties in the rate of the @xmath137c(@xmath138)@xmath139o reaction lead to an uncertainty in the derived helium abundance of about 0.02 .
another possibility is that there is a systematic effect throwing the @xmath12 values off . in lieu of systematic errors in @xmath12
, we can ask what physical processes can slow rgb evolution .
langer , bolte , & sandquist ( 1999 ) have suggested that deep mixing processes on the upper rgb may result in lengthened or reduced evolutionary times for those stars , which could modulate the resulting @xmath6 values .
this kind of scenario would mean that the helium abundances measured by this method ( as well as those using hb stars in some way ) would be affected since some of the helium produced by the hydrogen - burning shell would be mixed into the envelope of the stars that undergo the process .
we can also attempt to measure the trend in @xmath140 as a function of [ fe / h ] .
the primary difficulty is the clusters at [ fe / h]@xmath141 ( essentially on the zw scale ) with abnormally low values
. there are good reasons for removing m4 and m12 from consideration due to differential reddening and large reddening uncertainty . along with m4 ( which has a relatively low @xmath2 value also ) , m12 , m107 , and ngc 1851
were the clusters for which we had to use metallicities from the rhs study since they were not observed by cg .
we find from clusters with @xmath2 measurements ( excepting m4 , m12 , m107 , and ngc 1851 ) has a linear trend : @xmath142 } + ( 0.049\pm0.011 ) .\ ] ] the slope has marginal significance , but systematic error in the zeropoint of the metallicity scale affects the slope of the best - fit line since the @xmath140 values of high metallicity clusters are modified to a larger degree by metallicity shifts than low metallicity clusters .
if we reduce the metallicity of all clusters by 0.21 dex ( equivalent to ignoring @xmath13 elements ) , we derive a best - fit slope of @xmath143 a change in sign .
thus , the dependence of helium abundance on [ fe / h ] ( in addition to the absolute helium abundance ) as derived from @xmath2 will not be certain until the absolute metallicity scale is improved .
the data from the @xmath3 indicator for the two oosterhoff groups are individually consistent with constant helium abundance as a function of [ fe / h ] .
it is unlikely that the difference in between the two groups is due to a difference in helium abundance since a similar jump does not appear in the other indicators . to explain the apparent difference in @xmath3 values between oosterhoff groups
, there must either be some combination of a mean luminosity or a mean mass difference between the rr lyrae stars in the two groups .
if the difference were entirely due to a difference in mean mass , there would be no observable effect in either the @xmath2 or @xmath6 indicators . if there was only a difference in the mean luminosity ( the oo ii clusters rr lyraes would have to be about 15% more luminous than
oo i variables ) , oo ii clusters should have @xmath6 values that are 13% smaller ( since @xmath6 requires the hb magnitude to compute the faint limit of the rgb sample ) , and @xmath2 values that are higher by about 0.15 mag . from 15 oo i clusters having more than 5 variables as well as @xmath6 values , we find @xmath144 . for the 7 oo ii clusters meeting the same criteria , we find @xmath145
. the values for the two groups are consistent to within the errors . for @xmath2
, we find @xmath146 and @xmath147 for 7 oo i and 3 oo ii clusters respectively . so again
, we have no evidence for helium abundance differences between the groups , but the small numbers of clusters with @xmath2 values makes this a weak comparison . the difference in mean @xmath3 value between the groups corresponds to a difference in the mean rr lyrae mass of approximately 20% .
this is not so large as to make it unreasonable that there might be differences in the amount of mass loss at the tip of the rgb between the two groups .
the observations require the oosterhoff ii clusters to have variables of lower mass .
this goes in a direction opposite what is needed to explain the hb morphologies of at least some of the clusters .
clearly this does not speak to the exact cause of the oosterhoff dichotomy , but it can give a little guidance on the details of how the dichotomy is brought about .
we begin this section by summarizing what we consider to be the most important results of the surveys tabulated here . with the corrected differential bolometric corrections ( larger for more metal - rich populations ) ,
we now see no significant evidence for variation in the indicator @xmath6 as a function of metallicity or horizontal branch type .
in particular , metal - rich ( [ fe / h ] @xmath148 ) galactic globular clusters now are more consistent with the mean , and galactic bulge fields ( minniti 1995 ) are also likely to show lower helium abundances .
however , only a handful of clusters have helium values @xmath136 consistent with the favored primordial value @xmath149 .
there is evidence that there is real scatter in the @xmath6 values , over and above what can be chalked up to measurement errors .
clusters with anomalously high ( m30 , m62 , and ngc 6752 ) and low ( m68 , m107 , ngc 5927 , ngc
6229 , and ngc 6624 ) values appear over a range of metallicity . for @xmath3 , we find that virtually all of the oosterhoff i clusters ( and the lmc cluster reticulum ) have values consistent with a constant helium abundance .
the oosterhoff ii clusters with the largest number of rr lyrae variables also appear to have a constant value ( as do the ursa minor and carina dwarf spheroidals , and the lmc globular cluster ngc 1841 ) .
the remaining lmc globular clusters , and the dwarf spheroidals draco and sextans have values slightly above the average for the oo i group . for the richest oo ii systems , evolution is unable to explain the numbers of rr lyrae variables , and a systematic offset in helium abundance also seems unlikely given the evidence from the other two indicators .
the lack of any correlation between hb type and is indication that the second parameter problem is completely independent from the oosterhoff dichotomy .
the @xmath3 measurements give us the best constraint on the chemical evolution parameter @xmath150 . using the @xmath151 error bar on the slope of the @xmath3 [ fe / h ] relation ,
we get a limit @xmath152 , which is consistent with measurements from extragalactic hii regions ( pagel et al .
1992 ) . for @xmath2 ,
our results are consistent with constant helium abundance across the range of metallicity sampled .
the absolute value of the helium abundance as well as the exact value of @xmath153})$ ] depend on the metallicity scale used .
the most discrepant clusters can potentially be explained by errors in the cluster reddenings . comparing data from different indicators
, we find that the mean trends are all consistent with constant helium abundance for metallicities [ fe / h ] @xmath154 . we have examined the three indicators so as to use the information to either bolster or dispute claims of unusual helium abundance from just one indicator . in examining the clusters with helium abundances from more than one indicator
, we have not found convincing evidence that any have abnormal helium abundances .
systematic effects clearly appear to varying degrees in the data for all three indicators , but we have not been able to determine the cause in all cases .
we must note that none of the three indicators we have tabulated has much data covering the most metal - rich clusters where evidence of helium enrichment may still reside .
in general , the photometry for these clusters is most subject to field star contamination and heavy reddening , making interpretation difficult . for @xmath6 , there is the additional problem that the red hb begins to overlap the rgb in the cmd , making clean values impossible .
the redness of the hb also tends to preclude the possibility of rr lyrae stars , and hence the possibility of computing @xmath3 .
reddening and difficulties in finding the true hb level makes @xmath2 values unlikely without considerable work .
other means must be devised to determine good helium abundances for these clusters .
i would like to thank many for supplying electronic copies of their datasets : e. brocato , f. ferraro , p. guhathakurta , d. martins , s .- c .
rey , t. richtler , n. samus , a. sarajedini , v. testa , and a. walker .
i would also like to thank m. catelan , g. e. langer , r. taam , and the anonymous referee for very helpful conversations , and m. bolte and r. taam for their support ( under nsf grants ast-9420204 and ast-9415423 , respectively ) while this work was in progress .
catelan m. , sweigart a. v. , borissova j. , 1998a , in bradley p. a. , guzik j. a. , eds , asp conf .
135 , a half century of stellar pulsation interpretations : a tribute to arthur n. cox .
asp , san francisco , p. 41 | we present a survey of helium abundance indicators derived from a comprehensive study of globular cluster photometry in the literature .
for each of the three indicators used , we have conducted a thorough error analysis , and identified systematic errors in the computational procedures . for the population ratio @xmath0
, we find that there is no evidence of a trend with metallicity , although there appears to be real scatter in the values derived .
although this indicator is the one best able to provide useful absolute helium abundances , the mean value is @xmath1 , indicating the probable presence of additional systematic error . for the magnitude difference from the horizontal branch to the main sequence @xmath2 and the rr lyrae mass - luminosity exponent @xmath3 , it is only possible to reliably determine relative helium abundances .
this is due to continuing uncertainties in the absolute metallicity scale for @xmath2 , and uncertainty in the rr lyrae temperature scale for @xmath3 .
both indicators imply that the helium abundance is approximately constant as a function of [ fe / h ] . according to the @xmath3 indicator , both oosterhoff i and ii group clusters have constant values independent of [ fe / h ] and horizontal branch type .
in addition , the two groups have slopes @xmath4}$ ] that are consistent with each other , but significantly smaller than the slope for the combined sample .
globular clusters : general stars : abundances stars : variables : other ( rr lyrae ) |
the magellanic clouds represent a unique laboratory for studying star formation in an environment which differs from that of our galaxy .
they are nearby ( in comparison to most other galaxies ) , at distances of approximately 50 and 60 kpc respectively for the large and small magellanic clouds @xcite .
this enables them to be studied with intrinsically high sensitivity and resolution .
the relative star formation rates in both the large magellanic cloud ( lmc ) and the small magellanic cloud ( smc ) are greater than in the milky way , but the metallicity is significantly lower @xcite .
these conditions are thought to be similar to those in larger galaxies earlier in the history of the universe .
hence studies of star formation in the magellanic clouds can yield insights into star formation at redshifts of @xmath0 .
interstellar masers are excellent signposts of high - mass star formation in the milky way .
they are intense and occur at wavelengths that are not significantly absorbed by intervening clouds of gas and dust @xcite .
the first complete survey of the lmc and smc was recently undertaken in the 6.7-ghz transition of methanol and the 6.035-ghz transition of oh @xcite .
the survey , made with the methanol multibeam receiver on the parkes telescope , had a 1-@xmath1 sensitivity of 0.13 jy and 0.09 jy over the entire smc and lmc respectively ( the extent of both clouds , as defined by the co and/or hi emission ) .
the region of the lmc containing the bulk of the co emission and all the previously known masers were surveyed to a better 1-@xmath1 sensitivity of approximately 60 mjy . despite this comprehensive and sensitive search
only four 6.7-ghz methanol masers have been discovered in the lmc @xcite .
taking into account the relative rates of high - mass star formation in the two systems , this is a factor of 45 lower than expected on the basis of the observed luminosity distribution of galactic 6.7-ghz methanol masers @xcite .
in contrast , the number of detected ground - state oh and water masers is roughly consistent with the difference in the star formation rates between the milky way and lmc .
this discrepancy has been attributed to the lower metallicity of the lmc .
the current rate of high - mass star formation within the magellanic clouds , and its relationship to the dynamical interaction of the clouds with the milky way , are a topic of active investigation .
the _ spitzer space telescope _ legacy program sage ( surveying the agents of a galaxy s evolution ) has made observations of a number of nearby galaxies ( including the lmc ) to examine the current star formation rate @xcite .
the unique association of 6.7- and 12.2-ghz methanol masers with young , high - mass star forming regions has made them very useful signposts of such regions within our galaxy .
a recent comparative study of 6.7- and 12.2-ghz methanol masers in the milky way has shown that 12.2-ghz masers are preferentially associated with more luminous 6.7-ghz sources @xcite .
@xcite also used their observations and comparisons with published 1.2-mm dust continuum and mid - infrared data to start to quantify the maser evolutionary scheme proposed by @xcite .
they found that the 12.2-ghz methanol masers are associated with a later evolutionary phase than those sources which show emission from only the 6.7-ghz transition .
the median ratio of the peak flux density of the 12.2- to the 6.7-ghz transitions was found to be 1:5.9 , similar to previous studies .
it was also shown that the ratio depends on luminosity ( and hence age ) , increasing as the luminosity of the 6.7-ghz emission increases ( see fig . 4 of @xcite ) in order to investigate whether the properties of 12.2-ghz methanol masers in the lmc differ from their galactic counterparts we have undertaken new , sensitive observations towards known 6.7-ghz methanol and oh maser sources in the lmc .
@xcite searched for 12.2-ghz methanol masers towards the two sources known at the time n11 ( mc18 ) and n105a ( mc23 ) , with 3-@xmath1 detection upper limits of 0.1 and 0.3 jy respectively .
there are no published observations at 12.2-ghz towards the two more recently discovered 6.7-ghz methanol masers in the lmc _ iras_05011 - 6815 and n160a ( mc76 ) , nor at the locations of any of the known oh or water masers .
in contrast to methanol masers , the abundance of 22-ghz water masers in the magellanic clouds appears to be consistent with that observed in the milky way .
this suggests that water masers may be more useful signposts of young high - mass star formation regions within the lmc . in order to better characterize the properties of methanol and water masers in the lmc we have also made observations of the 6.7-ghz methanol and 22-ghz water masers in the lmc which were nearly contemporaneous with our 12.2-ghz search
we have searched a sample of eight known maser regions in the lmc for the presence of 12.2-ghz methanol masers .
the sample included all known ground and excited - state oh masers , 6.7-ghz methanol masers and 22-ghz masers in the magellanic clouds .
observations were carried out with the australia telescope national facility ( atnf ) parkes 64-m radio telescope during 2008 june 2025 .
the observations were made with the ku - band receiver which detected two orthogonal linear polarizations .
it had typical system temperatures of 205 and 225 jy for the respective polarizations throughout the observations .
the parkes multibeam correlator was configured to record 8192 spectral channels across a 16-mhz bandwidth for each polarization .
this corresponds to a velocity coverage of @xmath2290 @xmath3 and after hanning smoothing , a spectral resolution of 0.08 @xmath3 .
the parkes radio telescope has rms pointing errors of @xmath210 arcsec and at 12.2-ghz the telescope has a half - power beam - width of 1.9 arcmin .
all sources were observed at a fixed frequency of 12167-mhz ( i.e. with no doppler tracking ) , which alleviated the requirement for a unique reference observation to be made for each of the source positions .
the data were processed using the same method as described in @xcite , except that a running median of 50 channels ( as opposed to 100 channels ) was used to correct the spectrum baseline .
the data were reduced using the asap ( atnf spectral analysis package ) .
alignment of velocity channels was carried out during processing .
absolute flux density calibration was achieved by observing pksb1934 - 638 each day which has an assumed flux density of 1.825 jy at 12167-mhz @xcite .
each of the fields was observed for 160170 minutes , resulting in a typical rms in the final spectrum ( after hanning smoothing and averaging over polarizations and time ) of 27 mjy and a corresponding 5-@xmath1 limit of 0.135 jy .
the adopted rest frequency was 12.178597 ghz @xcite .
the 6.7-ghz observations reported by @xcite were made in 2006 june and september , approximately 2 years prior to the 12.2-ghz observations reported here .
as masers are known to be variable we undertook additional 6.7-ghz observations using the parkes 64-m radio telescope in 2008 august ( @xmath4 2 months after the 12.2-ghz observations ) using the methanol multibeam receiver .
the observations were made in mx mode , with 10 minutes per - beam on source toward each of the four known 6.7-ghz methanol maser sites ( see @xcite for a detailed description of the system parameters ) .
we recorded a 4 mhz bandpass for each beam , for each of two circular polarizations for both the 6.7-ghz methanol and the 6.035 mhz oh transitions .
the spectra were produced by an autocorrelation spectrometer with 2048 spectral channels .
the data were not hanning smoothed and the resulting velocity resolution for the 6.7-ghz observations was 0.11 @xmath3 .
this is very similar to the velocity resolution of the 12.2-ghz observations . a fault with one of the beams ( beam 3 ) meant that its data had to be excluded , resulting in an on - source integration time of 60 minutes per source .
the 22-ghz water maser observations were made with the australia telescope compact array ( atca ) during 2008 august 18,19 , with the array in the 6b configuration .
the observations were made in director s time allocations and spanned approximately 7 hours on the first day and 5 hours on the second day .
a total of eight different pointing centres were observed .
these covered the locations of all the 11 previously known star - formation maser sites in the lmc .
the observing band was centred at a fixed frequency of 22213 mhz .
the correlator was configured with 1024 channels across a 16 mhz bandwidth recording a single linear polarization and also allowing measurement of continuum emission .
the data were not smoothed , giving a velocity resolution of approximately 0.25 @xmath3 and a usable velocity range of approximately 195 @xmath3 .
pksb1934 - 638 was used as the primary flux calibrator , which at this frequency is assumed to have a flux density of 0.83 jy @xcite .
observations of pksb0537 - 441 were used to calibrate the bandpass and pksb0637 - 752 was used as the phase calibrator .
the observations were made as a series of 90 second integrations .
these were regularly interleaved with observations of the phase calibrator . in total , each source was observed 9 or 10 times over the two observing sessions resulting in an onsource time of 13.5 or 15 minutes per source .
we also used the line - free spectral channels from the water maser observations to search for continuum emission associated with the star formation regions .
the observations had only a 16-mhz bandwidth and thus were not very sensitive ; however they do have higher spatial resolution than previous studies of these regions .
the typical rms for the continuum images was 0.20.3 mjy beam@xmath5 .
the known lmc masers are contained in eight distinct fields which were targets for a 12.2-ghz methanol maser search .
table [ tab : results ] lists either the detected peak flux density and velocity , or the 5-@xmath1 noise level in the final spectrum .
a heliocentric velocity range of 120 450 @xmath3 was observed for each source . the only source detected was n105a ( mc23 ) , which coincidentally was also the first 6.7-ghz methanol maser detected in the lmc .
the 12.2 ghz spectrum of n105a is shown in the upper panel of fig . [
fig : n105a ] .
the peak of the 12.2-ghz emission is 0.12 jy ( 4.6-@xmath1 ) .
as it is within the velocity range of the 6.7-ghz emission , and is also apparent when the total 12.2-ghz dataset is split in half , or when the two polarizations are examined independently , we are confident that it is a genuine detection . comparing our 12.2-ghz spectrum of n105a with the 6.7-ghz emission observed in 2006 june @xcite shows that the 12.2-ghz peak coincides with the second strongest spectral peak in this source .
@xcite found that for approximately 75 per cent of the sources where they detected both 12.2- and 6.7-ghz emission the peak of the two spectra occurred at the same velocity .
for the remaining 25 per cent the 12.2-ghz peak coincided with a secondary 6.7-ghz feature .
although the association of the 12.2-ghz methanol maser with a secondary 6.7-ghz spectral feature is not unusual , it is also possible that the 6.7-ghz maser may have varied in the period between the observations reported by @xcite ( june 2006 ) and these observations .
comparison of the august 2008 6.7-ghz methanol maser spectrum , with that shown in fig .
3 of @xcite shows no detectable variation over the two - year period between the observations .
the observations of the other known 6.7-ghz methanol masers in the lmc and the 6.035-ghz oh maser in n160a also show no significant changes in the absolute , or relative intensity of the maser features compared to those published by @xcite ( see figure [ fig : spectra ] ) . [ cols="<,<,<,<,<,<,<,<,<,^ " , ] [ tab : ir ] @xcite have independently searched for yso candidates in the lmc .
their analysis , likewise , used the sage data , but also included a large amount of published and archival data from other wavelengths .
we have compared the positions of the interstellar masers in the lmc with their catalogue and find that 11 of the 16 locations are within 2 arcseconds of an object they have classified as a `` definite '' yso .
the sources without an associated yso candidate all lie in regions where there is a complex background , or a bright source nearby ( robert gruendl , private communication ) .
it is striking that the methods used by @xcite and @xcite to identify ysos in the lmc ( with the bulk of the observational data in common ) produce such very different results .
clearly @xcite have been much more successful in identifying the young , high - mass objects with which masers are associated .
we have compared the infrared properties of the maser - associated sources to those of the whole sample of definite ysos identified by @xcite ( their table 9 ) .
our initial examination of infrared colour - colour and colour - magnitude diagrams suggested that the maser - associated ysos tended to be brighter in all near and mid - infrared wavelength bands and generally have redder colours than the bulk of yso candidates ( figure [ fig : colmag ] ) .
this is consistent with the maser - associated sources being more luminous and more deeply embedded than the bulk of yso candidates , and may be useful for identifying regions of parameter space associated with the rare , and highly sought youngest high - mass stars .
the discrimination of the maser - associated sources is most clearly seen in the colour - magnitude diagrams , which explains why similar investigations undertaken with samples of galactic masers typically produce less conclusive results . for galactic sources
the broad range of distances to the sources adds additional scatter to the observed magnitudes , an effect which is difficult to correct for when the distances to most sources are poorly known . for the [ 8.0 ] versus [ 4.5]-[8.0 ] colour - magnitude diagram
the masers - associated sources all satisfy the criteria [ 4.5]-[8.0 ] @xmath6 2.5 and [ 8.0 ] @xmath7 12.5-([4.5]-[8.0 ] ) .
a total of 352 of the 855 ysos meet these criteria .
for the [ 3.6 ] versus [ 3.6]-[4.5 ] colour - magnitude diagram the maser - associated sources ( with one exception ) satisfy the criteria [ 3.6]-[4.5 ] @xmath6 0.5 and [ 3.6 ] @xmath6 12.2+([3.6]-[4.5 ] ) and a total of 251 of the 855 ysos in the lmc meet these criteria .
the dashed lines in fig .
[ fig : colmag ] indicate these criteria for each colour - magnitude diagram . taking the union of the two sets results in a total of 195 ysos satisfying both sets of criteria ( 22 percent of the sample as a whole ) .
recent statistical investigations of millimetre dust clumps with and without associated masers has enabled the development of models which predict the presence of a maser on the basis of the dust - clump properties @xcite .
these models also allowed some inferences to be made about the critical physical properties which determine the presence of a maser .
motivated by the apparent differences between those ysos with , and those without masers in the colour - magnitude diagrams we decided to undertake a similar investigation using the infrared photometric data compiled by @xcite for the sample of ysos in the lmc .
we constructed boxplots for each of the nine wavelength bands available ( j , h , k from the two micron all sky survey , 3.6 , 4.5 , 5.8 and 8.0 @xmath8 m from irac and 24 and 70 @xmath8 m from mips ) which show that in each of these wavelength bands the ysos associated with masers tend to be more luminous than the sample as a whole ( figure [ fig : boxplot ] ) .
the photometric data in @xcite is only complete for the 4.5- and 8.0-@xmath8 m bands , as these were used in the initial selection of yso candidates .
while most sources have measurements for all of the irac wavelength bands , this is not the case for the 2mass and mips data , especially for the maser - associated ysos . in order to objectively test whether the maser - associated ysos are more luminous than the sample as a whole we performed a @xmath9-test for each wavelength band to determine if the two groups have a different mean .
the @xmath9-test returned a statistically significant result ( @xmath10 95 per cent probability that the means of the two samples are different ) for each of the k- , 3.6- , 4.5- , 5.8- , 8.0- and 24-@xmath8 m bands which gave @xmath11-values of 0.03 , 0.0003 , 0.0004 , 0.0001 , 0.0004 and 0.02 respectively . for the other wavelength bands the number of maser - associated yso with a measurement was in all but one case , less than half the sample ( of 11 sources ) , which prevented meaningful statistical comparisons from being undertaken .
we also investigated the infrared colours formed from the irac data . from
the four wavelength bands six independent colours can be formed .
the three colours which include data from the 3.6-@xmath8 m band all have a statistically different mean ( as measured by a @xmath9-test ) for the maser - associated sources compared with the rest of the sample .
the @xmath11-values for the difference in the mean for the [ 3.6]-[8.0 ] , [ 3.6]-[5.8 ] and [ 3.6]-[4.5 ] colours are 0.03 , 0.03 and 0.01 respectively . in each case
the maser - associated sources have a larger mean colour ( they are `` redder '' ) , than the rest of the sample
. it would be desirable to fit a binomial generalised linear model ( glm ) to the infrared data to predict the properties of sources likely to be associated with a maser .
however , in order to undertake such modeling the only sources which can be used are those where the presence or absence of a maser is well established , and which have data for all the parameters . for this reason , it was only practical to consider using the irac data for the glm fitting as the data for the other wavelength bands are much less complete .
of the 855 definite yso sources identified by @xcite 814 have observations for all four irac wavelength bands .
unfortunately , the small number of detected methanol masers and the targeted / incomplete nature of the water maser searches meant that for neither transition was glm modeling practical due to the small sample sizes .
[ fig : colmag ] uses blue squares to represent those yso within a maser field which do not have an associated maser .
it is clear that the majority of these sources have similar infrared properties to the maser associated sources .
this is in contrast to our findings when we compared the maser - associated ysos to the rest of the sample . when we perform @xmath9-tests comparing those ysos which fall within a maser field with the rest of the sample we find a statistically significant difference between the means of the two samples for all the infrared bands with the exception of the mips 70 @xmath8 m band ( where there is insufficient data in the yso near a maser sample ) .
in addition to this , the @xmath11-values from the @xmath9-tests are in all cases lower than when comparing the masers associated ysos with the rest of the maser sample .
this is likely to be because the presence of a strong water maser signposts a young high - mass star formation region and since they form in clusters there are likely to be other similar sources nearby .
this is in contrast with most of the yso sources , which differ significantly in their infrared properties from those associated with masers .
they are likely to either be from more - evolved regions or associated with low- or intermediate - mass star formation . in summary ,
while we have established that there is a statistically significant difference between the infrared properties of maser - associated ysos compared to the whole sample , the information we have on the presence or absence of masers is insufficient to enable us to undertake predictive statistical modeling at the present time .
colour - colour and colour - magnitude diagrams are useful for identifying regions of parameter space which favour certain types of objects .
they are always subject to a degree of confusion however , due to overlaps in the colour ranges of different types of objects ( as nicely demonstrated by * ? ? ?
in addition , they are often incomplete ( sources need measurements in each of three or four wavelength bands to be included ) , or they do not utilize all the available information ( for sources where there are measurements in other wavebands ) . fitting the spectral energy distribution ( sed ) of a source has the advantage that it can utilize all the available information for a particular source . we have used the online fitting sed tool of @xcite to investigate the sed s of those ysos associated with masers and also the properties of the other lmc ysos .
the sed s are fitted to a grid of two - dimensional radiative transfer models of ysos consisting of a central star surrounded by an accreting disk and envelope @xcite .
the limitations inherent in the assumptions , implementation and results of the model have been extensively discussed in the literature ( see for example * ? ? ?
* ; * ? ? ?
* ) and we will not repeat them here . in order to fit each of the 855 definite yso sources of @xcite , we used an automated script to populate the online fitting form with the infrared data and download the best - fit model parameters for each source .
we have used the photometric measurements of @xcite for each of the bands where they are available .
when there was no reported measurement for a particular waveband we have assumed upper limits of 17.6 , 16.9 and 16.1 magnitudes for the j- , h- and ks - band deep 2mass observations towards the lmc @xcite and upper limits of 19.3 and 16.1 magnitudes for the irac 3.6- and 5.8-@xmath8 m bands ( the observations are complete in the other irac bands due to the selection criteria ) @xcite . where no measurement was available for the mips 24- and 70-@xmath8 m bands we chose not to use the upper limit values in the fit , as in many cases the sources are in regions of extended or saturated emission at these wavelengths , rather than regions with no emission . following @xcite we have assumed aperture sizes of 3 arcseconds for the 2mass observations , 5 arcseconds for irac observations , 10 arcseconds for mips 24 @xmath8 m and 20 arcseconds for mips 70 @xmath8 m .
we allowed the distance to vary between 48 and 52 kpc and the visual extinction ( along the line of sight ) to vary between 0 and 2 magnitudes ( the mean value for the interstellar extinction towards the lmc measured by @xcite is 0.3 magnitudes ) .
[ fig : sedfit ] shows some examples of the resulting sed fits for maser and non - maser associated ysos .
there are between seven and nine infrared intensities or upper limits available for each source and in most cases the chi - squared for the resulting sed fits are reasonable .
however , because of the large number of sources modeled we have made no attempt to remove sources where this was not the case .
figure [ fig : sed_box ] shows the distribution of some of the sed model fit parameters for the ysos associated with masers compared to the other ysos in the @xcite sample . for most of the parameters there is a large range present in both categories of yso and in many cases the distributions are highly skewed ( e.g. the envelope mass shown in fig .
[ fig : sed_box ] ) .
for skewed distributions it is more appropriate to use the mann - whitney test to determine if the two samples have a different median , rather than the @xmath9-test ( which compares the means ) .
the mann - whitney test returns statistically significant results ( @xmath11-value @xmath4 0.05 ) for five of the model parameters , the mass of the central star ( 0.001 ) , the outer radius of the envelope ( 0.05 ) , the ambient density ( 0.01 ) , the inclination to the line of sight ( 0.01 ) and the total luminosity ( 0.0008 ) .
the inferences which can be drawn from these results is clear .
the masers are associated with high - mass , luminous ysos where the sed is dominated by a large envelope and are located in regions where the ambient density is higher than in the vicinity of most ysos . the outer radius of the envelope and the inclination to the line of sight the results appear to be influenced by details of the modeling process .
the maximum outer radius of the envelope used by the model is @xmath12 au , and that is the fitted value for 9 of the 11 maser - associated ysos ( compared to 360 of the other 844 sources ) .
essentially this result is consistent with the masers being associated with higher mass objects , which would be expected to arise from the largest clumps / envelopes of gas and dust .
similarly for the inclination angle the maximum value available in the models is 87.13@xmath13 , and that is the fitted value for 7 of the 11 maser - associated ysos ( compared to 252 of the other 844 sources ) .
naively , it seems unlikely that geometry of the star forming region should play a significant role in whether we see a luminous water maser associated with a particular source and such a result is likely to be an artifact of the modeling process .
however , water masers are typically associated with outflows and an outflow - axis near perpendicular to the line of sight will produce the maximum line - of - sight velocity coherent path length and hence the most luminous water masers .
@xcite suggest that the inclination angle is one of the most important parameters in determining the sed for young objects , as it controls the contribution from the inner , hotter regions of the envelope due to orientation of the bipolar cavity .
the preference for the maser - associated ysos to best fit models with high inclination angles then means that there is little or no contribution from the inner regions in these seds .
an alternative interpretation of these results would be that additional obscuring material is present for the highest mass ysos in comparison to lower - mass objects .
we have compared the sed model parameters of the maser - associated ysos with other ysos in the maser fields .
we applied the mann - whitney test for all of the model parameters to compare these two samples and the only statistically significant result ( @xmath11-value 0.03 ) was for the inclination to the line of sight .
the fact that the inclination angle again shows a statistically significant difference warrants further future investigation to determine whether it is an artifact of the modeling process , or if it has a physical basis . since the sed model fits are based on the same infrared data for which we found no significant difference between the maser - associated ysos and those in the maser fields , it it not surprising that we see a similar lack of difference in the sed models .
this means that , as for the infrared properties it is not possible to undertake glm modeling to try and statistically predict which of the other ysos in the lmc may have an associated luminous water maser .
the best we can do is to look for other ysos for which the fitted models are very similar to those associated with masers .
of the list of 844 ysos in the lmc which are not known to have an associated water maser there are a total of 17 for which each of the mass of the central star , total luminosity , ambient density , outer radius of the envelope and inclination to the line of sight are all greater than or equal to the median of the maser - associated yso sample .
these ysos are 044927.29 - 691206.3 , 045358.57 - 691106.6 , 045426.06 - 691102.3 , 051004.00 - 685438.3 , 051328.26 - 692241.8 , 051922.98 - 693940.5 , 052208.52 - 675922.1 , 052257.55 - 680414.1 , 053205.89 - 662530.9 , 053603.03 - 673234.2 , 53831.62 - 690214.6 , 053912.67 - 692941.4 , 053937.04 - 694536.7 , 053937.53 - 694609.8 , 053945.94 - 693839.2 , 054309.58 - 675006.6 & 054630.37 - 693458.6 .
three of these ( 051004.00 - 685438.3 , 051328.26 - 692241.8 , 053945.94 - 693839.2 ) are within one of the maser fields and so are known not to have an associated maser , however , the remaining 14 should be primary targets in any future searches for water masers in the lmc .
a sensitive search for 12.2-ghz methanol masers towards a sample of eight fields containing all currently known methanol , oh and water masers in the large magellanic cloud has resulted in the detection of one source .
this is the first detection of a 12.2-ghz methanol maser outside our galaxy .
our sample size is too small and our observations have insufficient sensitivity to determine if there is any difference in the properties of the different maser transitions in the lmc compared to the galaxy ; however they are consistent with the previously observed underabundance of methanol masers in the lmc .
we have also observed the water masers associated with known high - mass star formation regions in the lmc and discovered six new sites .
five of these are in regions where emission had previously been detected , while the sixth is a weak new water maser detected towards the strongest 6.7-ghz methanol maser in the lmc _ iras_05011 - 6815 .
we find most of the star - formation masers in the lmc for which accurate positions are available are associated with sources identified as yso on the basis of their infrared properties .
the ysos associated with masers tend to be more luminous and redder than the sample as a whole .
we suggest that this is likely due to the masers being associated with young high - mass star formation regions .
sed modeling of the ysos shows that the masers are preferentially associated with sources with a high central mass , total luminosity and ambient density .
this is consistent with expectations from galactic maser studies .
we would like to thank the referee michele pestalozzi for useful suggestions which have improved the paper .
we thank john whiteoak and jasmina lazendic - galloway for useful discussions on the nature of the water maser emission in n113/mc24 .
we would like to thank robert gruendl for making available information on yso candidates in the lmc prior to publication .
we would like to thank karl menten for useful discussions during the writing of this paper .
spe would like to thank the alexander - von - humboldt - stiftung for an experienced researcher fellowship which has helped support this research .
this research was supported by the eu framework 6 marie curie early stage training programme under contract number mest - ct-2005 - 19669 `` estrela '' .
the parkes telescope and australia telescope compact array are part of the australia telescope which is funded by the commonwealth of australia for operation as a national facility managed by csiro .
this research has made use of nasa s astrophysics data system abstract service .
this research has made use of the nasa / ipac infrared science archive , which is operated by the jet propulsion laboratory , under contract with the national aeronautics and space administration .
this work is based in part on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , under contract with the national aeronautics and space administration . | we report the results of a sensitive search for 12.2-ghz methanol maser emission towards a sample of eight high - mass star formation regions in the large magellanic clouds which have been detected in other maser transitions .
we detected one source towards the star formation region n105a .
this is the first detection of a 12.2-ghz methanol maser outside our galaxy .
we also made near - contemporaneous observations of the 6.7-ghz methanol and 22-ghz water masers towards these sources , resulting in the detection of water maser emission in six new sources , including one associated with the strongest 6.7-ghz maser in the magellanic clouds _
iras_05011 - 6815 .
the majority of the maser sources are closely associated with objects identified as likely young stellar objects ( yso ) on the basis of _ spitzer space telescope _ observations .
we find that the ysos associated with masers tend to be more luminous and have redder infrared colours than the sample as a whole .
sed modeling of the ysos shows that the masers are associated with sources of higher central mass , total luminosity and ambient density than the majority of ysos in the lmc .
this is consistent with the well - established relationship between luminous methanol and water masers and young , high - mass objects observed in the galaxy .
masers stars : formation galaxies : magellanic clouds
ism : molecules |
it must have been a subconscious path , that guided schrdinger to the derivation of his famous equation in the early days of 1926 @xcite .
schrdinger started with the classical hamilton - jacobi equation , transformed the classical action function @xmath0 by @xmath1 $ ] , performed an adventurous variational calculation on the hamilton - integral and arrived at his celebrated result @xmath2 several attempts have since been made to relate the quantum mechanics perfectely described by ( [ schrodingerequation ] ) to quantities , that are known from classical mechanics , action and momentum ; these attempts are twofold .
the first one was taken by madelung @xcite already in 1926 .
madelung introduced the polar wave - function , @xmath3 $ ] , where @xmath4 is the madelung action function and both @xmath5 are real .
madelungs way was further taken by bohm @xcite in his attempt to build a causal quantum theory .
but the initial separation of real and imaginary parts of the wave - function leads to very complicated equations for amplitude and phase and will not be pursued here .
for exaustive information on the madelung - de - broglie - bohm theory see @xcite and for modern application @xcite .
the work of floyd @xcite pursued especially the theory of the stationary hamilton - jacobi theory with quantum potentials , while faraggi and matone @xcite probably worked out some fundamental ideas on the geometrical meaning of the quantum potential . in the appendix
, we will comment especially on @xcite in the light of our results .
the second way is to initially keep the real and imaginary parts of the quantum action function together , @xmath6 $ ] , leading to the hamilton - jacobi equation of quantum mechanics , @xmath7 also in 1926 , wentzel , kramers and brillouin @xcite , starting from ( [ hjs ] ) , proposed an expansion of the full quantum action function @xmath8 by small values of @xmath9 , which finally leads to the classical action , pertubed by small corrections of order @xmath9 .
the wkb - theory works surprisingly well , but it breaks down in the vincity of _ quantal _ regions , where the wkb - wave function is plagued by singularities . for an exhaustive review on this topic see @xcite and references therein .
various methods , most of them are reviewed in @xcite , have been developed to improve the wkb - theory .
the most elaborated among these are the frman - frman series @xcite and the use of the quasi - linearization method by raghunathan and vasudevan @xcite .
the method of quasi - linearization has the advantage , that it initially does not rely on the existence of a small expansion parameter , @xmath9 in our case .
quasi - linearization , as applied to quantum mechanics in @xcite is the starting point of our article .
we first develop the relations between the schrdinger theory and the hamilton - jacobi theory of quantum mechanics , which , uninterpreted , can already be found in @xcite .
then , we are going to develop the use of quasi - linearization techniques from a systematic and physically reasonable point of view , which was neglected in @xcite .
the systematic use of this technique will lead us to the definition of a quantum correction function .
this is a function , that describes all quantum effects of a physical system .
further , the properties of the quantum correction function will be analyzed in detail , especially their relation to the wkb - series expansion .
after the properties of the quantum correction functions are clarified , we will apply them to the eigenvalue - problem and the question of quantum trajectories . in our article
, we will develop a quantum theory from the hamilton - jacobi point of view .
in contrast to the de broglie - bohm ansatz , we will not initially separate the amplitude and phase of the wave - function . in this subsection
, we will examine the interrelation between the schrdinger and hamilton - jacobi equation from a structural point of view .
this will clarify , that the solutions of both equations indeed are related to each other by the well known relations @xmath10,\quad
s(x ) = -i\hbar\ln\left[\psi(x)\right]\quad.\ ] ] by introducing a local gradient field , which can easily be identified with the local quantum momentum , @xmath11 the hamilton - jacobi equation ( [ hjs ] ) can be cast into @xmath12 the full solution of the schrdinger equation ( [ schrodingerequation ] ) is a superposition of two linearly independent solutions @xmath13 in the language of action , this reads @xmath14 + c_{2}\exp\left[\frac{i}{\hbar}s_{2}(x)\right]\quad . \label{hjwavefunction}\ ] ] we want to emphasize here , that ( [ hjwavefunction ] ) allows us to decompose a schrdinger wave - function into running waves in _ any _ point in space , not only asymptotically .
the statistical properties of running waves are given by @xmath15,\quad j_{i}(x ) = \pm\im{\psi_{i}(x)^{*}\psi_{i}'(x ) } = \pm\re{p_{i}(x)}\rho_{i}(x)\quad.\ ] ] any of the two linearly independent wave functions solve the schrdinger equation seperately , and any of the two action functions solve the hamilton - jacobi equation seperately .
but when we take the logarithm , we find @xmath16 = s(x ) & = & -i\hbar\ln\left[c_{1}\exp\left[\frac{i}{\hbar}s_{1}(x)\right ] + c_{2}\exp\left[\frac{i}{\hbar}s_{2}(x)\right]\right]\nonumber\\ & \sim & \frac{s_{1}(x)+s_{2}(x)}{2 } -i\hbar\ln\left[2\cos\left[\frac{s_{1}(x)-s_{2}(x)}{2\hbar}-\frac{\phi}{2}\right]\right]\quad,\end{aligned}\ ] ] where the @xmath17 s have been absorbed in the phase @xmath18 .
the function @xmath19 is again a solution of the hamilton - jacobi equation , but @xmath20 is not .
this fact is due to the nonlinearity of the hamilton - jacobi equation and its physical interpretation is interference .
usually , the two linearly independent solutions @xmath21 will be complex conjugates , such that @xmath22 . the superponed wave function , up to an amplitude factor , then reads @xmath23 \cos\left[\re{\{s_{1}(x)\}}/\hbar - \frac{\phi}{2}\right]\quad.\ ] ]
to be able to distinguish between the influence of classical mechanics and quantum mechanics on the behaviour of a physical system , we propose , that the full quantum momentum field @xmath24 can be decomposed into its classical part and a contribution arising purely by quantum effects .
we set @xmath25 the function @xmath26 is called a _ quantum _ _ correction _ _ function _ , because it decribes purely quantum effects and thus corrects the behaviour of the classical movement , governed by the classical momentum @xmath27 .
more precisely , the function @xmath26 describes the difference between the full quantum momentum @xmath24 and its classical contribution @xmath27 @xmath28 it can be concluded easily , that @xmath26 tends to zero in regions , where the motion is known to be purely classical . + to derive an analytic expression for the quantum correction function @xmath26 , we insert ( [ quasilintransform ] ) into the hamiltion - jacobi equation ( [ hjm ] ) and find an differential equation for @xmath26 : @xmath29 this nonlinear first order equation can be formally be integrated to yield a selfconsistent formula for the quantum correction function @xmath26 : @xmath30 \int^{x}p'(y)\exp\left[\frac{i}{\hbar}\int^{y}\{2p(z)+q(z)\}dz\right]dy\quad.\end{aligned}\ ] ] we emphasize , that this expression for @xmath26 is exact ; so far , no approximation has been made . in the following , we will examine the structure of this expression and formulate a systematic perturbational approach .
this perturbational approach does not , as it is usually done for semiclassical questions , start initially with the expansion in terms of @xmath9 , it will consist of quantum correction functions , that contain every order in @xmath9 .
this will open a more general sight onto the behaviour of physical systems near their semiclassical limit .
the main reason to examine quantum mechanical systems in hamilton - jacobi formulation is to clearly distinguish between classical and quantum behaviour .
as we will see below , the semiclassical limit of hamilton - jacobi quantum mechanics is open to perturbation theory , while its full quantum description belongs to the non - perturbative limit of this theory .
the main result of the following subsection is , that in the one - dimensional case , we are able to introduce an operator formalism , which is , in its structure , closely related to the wellknown dirac- or interaction representation . the central quantity of this operator formalism is given by @xmath31{\bf{\hat{q}}}(x)\exp\left[\frac{i}{\hbar } w(x)\right]\quad , \label{operatorequation}\ ] ] where the quantum action function is given by @xmath32 @xmath33 the classical action and @xmath34 the quantum action correction function .
the structure of ( [ operatorequation ] ) enables us to make use of the baker - hausdorff identity , which allows us a formal expansion : @xmath35{\bf{\hat{q}}}(x)\exp\left[\frac{i}{\hbar } w(x)\right]\nonumber\\ & = & q_{0}(x)-\frac{i}{\hbar}\left[w(x),{\bf{\hat{q}}}(x)\right ] + \frac{1}{2}\left(\frac{i}{\hbar}\right)^{2}\left[w(x),\left[w(x),{\bf{\hat{q}}}(x)\right]\right ] + ... \nonumber\\ & = & q_{0}(x ) + q_{1}(x ) + q_{2}(x ) + ... \quad .
\label{perturbationseries}\end{aligned}\ ] ] the zeroth order correction function @xmath36 comes from @xmath37\int^{x}p'(y)\exp\left[\frac{i}{\hbar } 2 s(y)\right]dy\quad.\end{aligned}\ ] ] this corresponds to the zeroth iteration of equation ( [ operatorequation ] ) , where the action correction @xmath34 is chosen zero .
+ the convergence behaviour of our perturbation series depends crucially on its key quantity , the function @xmath36 .
the whole perturbation series is built on this zeroth order quantum correction function , and thus the behaviour of this function will influence the behaviour of the perturbation series .
the question we have to ask must be , if there are singularities around classical turning points , which plague the wkb - theory . anticipating the answer
, we can state : the function @xmath36 is regular in the vincity of classical turning points @xmath38 , if , and only if the potential function @xmath39 is regular there
. the examination now follows .
+ we want to examine , if the integral , which determines the function @xmath36 posseses upper or lower bounds .
we have @xmath40\int^{x}p'(y)\exp\left[i 2 s(y)\right]dy\quad.\ ] ] since we are interested in the behaviour of @xmath36 in the vincity of a classical turning point , we expand the integrand around the classical turning point : @xmath41 this approximation is suitable , because the function @xmath42 , which is the origin of the turning point singularity is approximately zero almost everywhere , except the turning point region .
thus , the main contribution to the integral comes directely from the turning point region . inserting the approximate quantities and performing the integral
, we get the following estimate @xmath43\left|\frac{u'(x_{t})}{u^{2/3}(x_{t})}\right|\quad.\ ] ] we clearly see , that near the turning point , the behaviour of @xmath36 depends directly on the behaviour of the potential function @xmath39 . when @xmath39 is regular at @xmath38 , which is the case for all potentials , that locally can be represented by powers @xmath44 , @xmath45 , the quantum correction function itself is regular and finite .
when @xmath39 is unbounded near @xmath38 , which is the case for potentials , that locally can be represented by inverse powers @xmath46 , @xmath47 , then the quantum correction function will diverge , too . from a more mahematical point of view ,
mandelzveig @xcite has shown , that expressions like ( [ operatorequation ] ) converge _ quadratically_. physically spoken , quadratical convergence means , that the zeroth iterate @xmath36 is already very close to the exact result @xmath26 , such that qualitatively the whole quantum mechanical behaviour of a system is to be decribed by the zeroth iterate @xmath36 . from the perturbation series ( [ perturbationseries ] ) ,
the order in @xmath9 is not easily visible .
we have seen in ( [ perturbationseries ] ) , that the full quantum momentum can be expressed as a series @xmath48 on the other hand , it is known , that the full quantum momentum can be expanded in power of @xmath9 , when @xmath9 is small compared to the classical action function : @xmath49 the crucial question is now , how these two series are related to each other . to get more insight into this question , we start with the selfconsistent expression ( [ operatorequation ] ) for @xmath26 . due to selfconsistency
, the correction function @xmath26 is a function of _ all _ orders of @xmath9 . to see this more clearly
, we formally can put the indices @xmath50 together and obtain @xmath51 this means , that every order of quantum correction functions carries all orders of @xmath9 , as stated above .
this astounding fact will be demonstrated in the next subsection , where the power series in @xmath9 is calculated directely from the quantum correction functions .
after these formal preparations , we will calculate the power series in @xmath9 .
one of the most interesting results of this calculation will be a reason for the bad convergence of the wkb - series .
+ the wkb - theory follows from the assumption , that the solutions of the hamilton - jacobi equation can be represented as power series in @xmath9 : @xmath52 when these anstze are inserted into the hamilton - jacobi equation ( [ hjs ] , [ hjm ] ) , there arises an infinite hierarchy of equations , which can only be solved by successive approximation . breaking this infinite hierarchy after the first order yields the familiar wkb result for the wave - function .
but the expansion in powers of @xmath9 is plagued by singularities around classical turning points , and the strength of these singularities increases by every order in @xmath9 .
this is not the case , when the series expansion is derived from the quantum correction functions , because their expansion in powers of @xmath9 will create residual terms of higher order in @xmath9 , which will catch the divergencies and result finite expressions as a whole .
+ technically , the expansion can be set up as follows . by use of the auxiliary identity @xmath53
, we get for the zeroth order @xmath54 \int^{x}p'(y)\exp\left[\frac{i}{\hbar } 2 s(y)\right]dy\nonumber\\ & = & -\exp\left[-\frac{i}{\hbar } 2 s(x)\right ] \int^{x}p'(y)\left(\frac{p(y)}{p(y)}\right)\exp\left[\frac{i}{\hbar } 2 s(y)\right]dy\quad.\end{aligned}\ ] ] performing an integration by parts , one gets @xmath55 \int^{x}dz\left(\frac{p'}{p}\right)'(z)\exp\left[\frac{i}{\hbar}2s(z)\right]\quad.\ ] ] using the auxiliary identity once again and integrating by parts yields @xmath56 we observe the following facts : the term proportional to @xmath9 is precisely the first order series expansion following from the wkb - theory ; the following term is of order @xmath57 , and does not contribute to the first order in @xmath9 .
the residual term , denoted by @xmath58 carries only powers of @xmath9 equal or larger @xmath59 .
such a residual term is not known in wkb - theory , it is lost by initially assuming the power expansion in terms of @xmath9 .
+ up to now , we have shown , that the zeroth order quantum correction leads to the correct wkb - term first order in @xmath9 , @xmath60 the second order expansion in wkb - theory gives @xmath61 the first term has already been obtained by the second integration by parts of @xmath36 .
further integrations by parts of @xmath62 in the same manner would not lead to the required second term , it can only be calculated by expanding @xmath63 into powers of @xmath9 . the perturbational expression for @xmath63 as calculated above ( [ perturbationseries ] ) is @xmath64 = -\frac{i}{\hbar}[\hbar\int^{x}q_{0,1}(z)dz,{\bf{\hat{q}}}(x)]\quad.\ ] ] the first integration by parts cancels , such that only term of order @xmath65 remain , @xmath66\nonumber\\ & = & -\hbar^{2}\frac{i}{4}\frac{p'(x)}{p^{2}(x)}q_{0,1}(x ) + r_{1,3}(x ) - [ \int^{x}q_{0,1}(z)dz,{\bf{\hat{r}}}_{0,3}(x)]\quad.\end{aligned}\ ] ] the first order in @xmath9 is already known from above , @xmath67 , leading to the first term of the desired wkb expression : @xmath68 with that , we obtain @xmath69 the residual terms @xmath70 und @xmath71 carry only higher orders in @xmath9 .
+ with this , rather cumbersome , procedure we could show , how all orders of the wkb - series can be obtained , at least in principle .
it further clarifies the fidelity of the quantum correction functions : @xmath36 is exact to the first order in @xmath9 , @xmath72 is exact to second order in @xmath9 and so on . keeping the wkb - form , we can write the quantum mechanical momentum as @xmath73 where in @xmath74 we have subsummized all residual terms . the number @xmath75 gives the order in @xmath9 .
the residual terms can not be obtained when the wkb - series expansion is set initially .
the residual terms result from partially integrating the nonsingular expression for the quantum correction function @xmath26 , but a nonsingular expression will remain nonsingular . the procedure of partial integration splits the nonsingular quantum correction function into a singular wkb - term and a singular residual term .
both singularities kept together will give a nonsingular function , but omitting the residual terms , as wkb - theory does , will give a badly converging , singular power series in @xmath9 .
in the introduction section we have seen , that in our theory the wave - function of a system can be written in terms of superponed exponentials , or , if one prefers , in terms of trigonometric functions .
depending on the problem , both representations have advantages and disadvantages .
the key to the construction of the wave - function is the action integral , that depends on the direction of integration . in a two turning - point system
, we have the choice to integrate from the left ( @xmath76 ) to the right ( @xmath77 ) or vice versa .
we have reserved a whole section to this problem , because its solution is not as easy as it might look at first .
we begin with the wave - function , that is obtained by integrating from the left to the right , @xmath78 .
we find , in trigonometric fashion , @xmath79 \cos\left[\int_{x_{1}}^{x}\left(p(x')+\re{q(x')}\right)dx'-\frac{\phi_{\rightarrow}}{2}\right]\quad .
\label{wavefunctionright}\ ] ] the arrow - index of the reflection phase indicates , that the direction of integration here was chosen as @xmath78 . to construct this wave - function ,
we have made the following choice for the quantum momentum function @xmath80 this choice is of somewhat arbitrary character , we could also have choosen the complex conjugate of @xmath24 or @xmath81 or any other possible combination of signs and complex conjugates . but the choice , once made , gives the system the property of a direction of integration , all other directions of integration have to be consistent with .
this becomes clear , when we now construct the wave - function for the direction of integration @xmath82 . the easiest way to do
this is to introduce a parity transformation by the operator @xmath83 , that performs an axialsymmetric transformation .
the position of this axis will be the minimum of the potential - the potential itself may be of unsymmetric shape , the minimum is the only condition - that any two turning - point system will possess .
we want to know what a wave , that starts at @xmath76 would look like , when it starts at @xmath77 : @xmath84 = \exp\left[i\int_{x_{2}}^{x}\left(p(x')+\re{q(x)}-i\im{q(x')}\right)dx'\right]\quad.\ ] ] the sign of the imaginary momentum function has changed .
this is due to the fact , that the real parts of the quantum momentum are even , the imaginary part is an odd function with respect to the minimum of the potential .
this can be deduced from ( [ operatorequation ] ) by inspection .
thus , when a parity transformation with respect to this minimum is applied , the sign of the imaginary part must be changed . without this fact ,
a consistent construction of a wave - function for the direction of integration @xmath82 is not possible .
we finally obtain @xmath85 \cos\left[\int_{x_{2}}^{x}\left(p(x')+\re{q(x')}\right)dx'-\frac{\phi_{\leftarrow}}{2}\right]\nonumber\\ & = & \exp\left[-\int_{x}^{x_{2}}\im{q(x')}dx'\right ] \cos\left[-\int_{x}^{x_{2}}\left(p(x')+\re{q(x')}\right)dx'-\frac{\phi_{\leftarrow}}{2}\right]\quad .
\label{wavefunctionleft}\end{aligned}\ ] ]
in the framework of hamilton - jacobi theory , soft boundary problems arise , because the space is separated into classically allowed and classically forbidden regions . by solving the appropriate hamilton - jacobi equations , a wave - function @xmath86 for each of these regions
the boundary value problem is to match all these wave - functions , to obtain a wave - function @xmath87 defined on the whole space .
this boundary value problem is also known as the connection problem , see e. g. @xcite . in the schrdinger point of view
the connection problem @xmath88 is a formidable task and will , except some few solvable examples , remain unsovled in general . in the hamilton - jacobi point of view , where we can express the wave - functions as exponentials ,
the connection problem becomes easily treatable on the level of trigonometric and exponential functions : @xmath89 \leftrightarrow \exp\left[-\im{s_{2}(x)}\right]\cos\left[\re{s_{2}(x)}-\frac{\phi}{2}\right ] \leftrightarrow \exp\left[-s_{3}(x)\right]\quad .
\label{spiritofhj}\ ] ] the reflection phase @xmath18 solves the connection problem and with @xmath18 the wave - function can be defined on the whole space .
the points where the wave - function of different spatial regions have to be matched are the classical turning points . the boundary value problem , e.g. @xmath90{\bigg{|}}_{x = x_{2 } } = \frac{\partial}{\partial x}\ln\left[\psi_{3}(x)\right]{\bigg{|}}_{x = x_{2}}\quad , \label{boundary1}\ ] ]
is called a soft boundary problem , because it is neither a dirichlet nor a neumann boundary problem , but a mixture of both . writing ( [ boundary1 ] ) in the spirit of ( [ spiritofhj ] ) , we obtain @xmath91 = -q_{3}(x_{2})\quad.\ ] ] from this , we obtain the reflection phase @xmath92 + 2\int_{x_{1}}^{x_{2}}\left(p_{2}(x')+\re{q_{2}(x')}\right)dx'\quad .
\label{phaseltor}\ ] ] for the reverse direction , we find similary @xmath93 - 2\int_{x_{1}}^{x_{2}}\left(p_{2}(x')+\re{q_{2}(x')}\right)dx'\quad . \label{phasertol}\ ] ] the reflection phases , by their dependence on the turning points , are functions of energy . knowing the reflection phases , we can easily derive the quantization law by the argument , that both directions of integration must give the same wave - function .
the phases of this wave - functions may only differ by @xmath94 . from this , we get @xmath95 note that by ( [ phaseltor ] , [ phasertol ] ) the action - integrals cancel each other in the sum @xmath96 . to make contact with the usual bohr - sommerfeld rule , we rewrite ( [ quantization1 ] ) and introduce the maslov - index @xmath97 @xmath98 with ( [ quantization2 ] ) the maslov - index of a physical system is directely calculable by knowledge of the quantum correction function .
this is a very valuable result , because quantization in the hamilton - jacobi point of view means just to calculate ( [ quantization2 ] ) to obtain the energy - levels .
there is no need to orthonormalize the wave - functions to find the energy - levels , when calculated by ( [ quantization2 ] ) .
the theory of hilbert - spaces ensures , that wave - functions corresponding to eigenvalues of the underlying problem are already orthonormal . by ( [ phaseltor ] , [ phasertol ] )
, we can examine another important property of the reflection - phases .
they are divided into two parts , one gauge - dependent , one gauge - invariant , @xmath99 inspection of ( [ phaseltor ] , [ phasertol ] ) , in combination with ( [ wavefunctionright ] , [ wavefunctionleft ] ) shows , that the lower bound of integration does not necessarily have to be a turning point . dependent on the direction of integration , which has to be defined at first , the lower bound serves as a point of reference and can be chosen arbitrarily as any point lying between the two turning points .
the choice of the lower bound must leave the wave - function invariant .
this is ensured by the gauge - dependent part of the reflection - phase , e.g. @xmath100 the gauge - invariant part is thus given by , e.g. @xmath101\quad.\ ] ]
for some reasons , this paper would be the place to set a profound and philosophical discussion on the sense and usefulness of quantum trajectories into work , but we think that should be done elsewhere . instead , we want to demonstrate how easily quantum trajectories are available within the framework of the stationary hamilton - jacobi theory . the topic of quantum trajectories is often discussed by introducing a quantum potential and a quantum force .
we refuse the introduction of those quantities , because in our opinion they are somewhat misleading .
the so called quantum potential is not a potential function as classical potentials are , but depends on the real part of the reduced quantum mechanical action function , so does the quantum force . from this point of view ,
one anyway has to solve the hamilton - jacobi equation at first and calculate the quantum potential afterwards , which makes the introduction of a quantum potential gratuitous ( for the here discussed stationary theory ) .
there is another crucial point , that leads us to reject the concept of a quantum potential , because , as we will show below , the classical relation between momentum and velocity does not hold within the framework of quantum mechanical hamilton - jacobi theory .
the trajectory has thus to be derived from the quantum mechanical action function directely . from this trajectory , by time derivatives , it would in principal
be possible to construct a lagrangian or newton s equation of motion , but such attempts are , again , gratuitous to our regards .
the use of the stationary hamilton - jacobi equation describes isoenergetic dynamics , meaning that motion takes place on a energy - shell in phase - space .
the presence of quantum mechanical behaviour has no effect on the energy - shell itself .
this can be demonstrated by writing down the the classical and quantum mechanical hamiltonians @xmath102 and comparing them @xmath103 if in ( [ energyconstrained ] ) @xmath104 is zero , the energy of quantum and classical motion is constrained to the same energy - shell , quantum and classical motion are isoenergetic to each other .
the employment of quasilinearization @xmath105 naturally fullfills @xmath106 because @xmath107 the presence of the quantum momentum corrections thus have no effect on the energy of the system , they only change the dynamical flow on the energy - shell . for the dynamics that means ,
that the time the motion between two points takes is changed . in the light of ( [ isoenergetics ] )
it becomes clear how misleading the introduction of concepts like quantum potentials are , because a genuine potential function should of course change the total energy of the system .
the presence of a quantum momentum correction leads to a different trajectory , @xmath108 where the total motion is divised into a purely classical and a purely quantum mechanical part .
naively one would think , that from ( [ trajectory ] ) it immedeately follows that @xmath109 which , as we will show now , is not true . in classical mechanics , the trajectory equation ( [ trajectoryeq ] ) and
the velocity - momentum relation are equal .
we have @xmath110 in quantum mechanics , the quantum momentum correction breaks this equality , leading to @xmath111 it can easily be seen , that the naive assumption ( [ wrongconnection ] ) leads to an obviously wrong result .
see @xcite for statements , that dig in the same hole .
+ the necessary theoretical tool to calculate a trajectory is already known from classical mechanics .
as there , we start with the action function @xmath112 employing the jacobi - connection , which is the same as demanding the phase of the wave - function to be stationary , it follows @xmath113 which we will call the trajectory equation .
the use of the jacobi - connection is a natural consequence of the hamilton - jacobi theory .
theoretically , it is very well founded and will give the right trajectory for any given action - function .
thus , we dwell on a very safe terrain , whereas ( [ wrongconnection ] ) is not more than an assumption .
it is common knowledge , that a standing wave , e.g. a bound state , has a current density of zero , and thus contains no observable motion .
as already stated above , this is true in cases where only the wave - function @xmath87 is known .
the use of the quantum action function @xmath114 instead allows a decomposition into running waves , which certainly posses a dynamical interpretation .
this dynamical interpretation allows us to connect the motion of a quantum particle with a causal picture . to do so
, we have to work out the differences between quantum and classical motion at first . from ( [ trajectoryeq ] )
we get the running time differences between classical and quantum motion moving from an arbitrarily chosen point to a turning point and back : @xmath115 from ( [ runningtime ] ) , we obtain the total time - shift between quantum and classical motion @xmath116 the total time - shift is a positive function of energy and a consequence of the isoenergeticity .
its meaning is the following : the quantum particle moves with a larger local momentum as the classical particle does and reaches the turning point earlier than the classical particle .
but for the backward motion the quantum particle is delayed with respect to the classical particle , resulting a positive , total time - shift .
the delay of the quantum particle at the turning point stems from the finite quantum momentum there and reflects the finite probability for entering the potential barrier .
we emphasize , that a negative sign of the total time - shift ( [ totaltimeshift ] ) would result in acausal behaviour , because then the quantum particle would leave before it would have arrived , which is not possible .
the result ( [ totaltimeshift ] ) allows us to derive a remarkable property of quantum trajectories , that is , they leave the phase - space integral invariant . comparing the classical and the quantum mechanical phase - space integral ,
this can easily be seen .
if we define @xmath117 as the classical period and @xmath118 , we have @xmath119 from the last line of equation ( [ invariance ] ) equation ( [ totaltimeshift ] ) readily follows .
thus , if isoenergeticity is fulfilled , classical and quantum mechanical motion are described by the same action hypersurface . with the time - shift ( [ totaltimeshift ] )
, we now can construct the trajectories @xmath120 by this procedure one obtains quantum trajectories in the classically allowed region , forming a sum of disconnected branches @xmath121 the trajectories form disconnected branches , because their shape in the classically forbidden region is not available by the trajectory equation ( [ trajectoryeq ] ) , but below , when the quantum trajectories of the harmonic oscillator will be discussed , we will see how this problem can be solved in a physically reasonable way .
to demonstrate how quantum mechanics is done from the hamilton - jacobi point of view , we will discuss two important systems , that are widely used as models for various physical systems : the linear potential and the harmonic oscillator potential .
especially , we will work out the differences between the schrdinger point of view and the hamilton - jacobi point of view . by the ladder
, we will obtain an insight into the quantum mechanical behaviour , which can not be gained by the wave - function alone .
we consider a linear potential @xmath122 . by this example , we will focus on the improvement , that is gained by the use of the zeroth order quantum correction function @xmath36 in contrast to the usual wkb - approximation . from the potential ,
we obtain local the classical momentum and the local classical action as @xmath123 the zeroth order quantum corrections function in this case can be calculated analytically , giving @xmath124\gamma_{1/3}\left[-i 2 s(x)\right]\quad .
\end{aligned}\ ] ] during the following , we set @xmath125 and @xmath126 , because the system is translationally invariant on the energy - axis . in figure ( [ linpot1 ] ) , we have plotted the phase - space functions of the linear potential . the quantum correction function @xmath127 , far away from the classical turning point , here @xmath128 , is negligible and the classical momentum @xmath27 is in charge .
approaching the turning point region , the quantum correction function @xmath127 starts to grow , describing the importance of quantum mechanical behaviour in the vincity of the turning point . at the turning point itself , the real - part of the corrected momentum @xmath129 is finite , while the classical momentum @xmath27 is zero as usual .
this allows only one interpretation : for a classical particle , the potential represents an inpenetrable barrier , and reaching the classical turning point leads undoubtedly to a change of the direction of motion . for a semiclassical particle , described by @xmath130 , the classical picture is not true anymore .
the corrected momentum @xmath129 is finite at the classical turning point , meaning that the potential barrier becomes penetrable with some probability . for a classical particle ,
the potential barrier represents a hard wall , but for a semiclassical particle , the potential barrier becomes soft .
the softness of the potential barrier in the semiclassical case is decribed by the finite momentum functions and by the finite probability to find a particle inside the barrier .
we want to emphasize here , that in the wkb - approach the potential barrier remains hard , because of the turning point singularity . the proper connection between classically allowed and classically forbidden region in the wkb - case can only be established by the knowledge of the exact solution of the schrdinger equation , leading to the reflection - phase @xmath131 . by the use of quantum correction functions ,
the turning point singularity is overcome .
the linear potential serves as a model for the one turning point case .
taking the turning point @xmath76 as the turning point of the system , the wave - function can be constructed by ( [ wavefunctionleft ] ) @xmath132 \cos\left[-\int_{x}^{\infty}\left(p(x')+\re{q_{0}(x')}\right)dx'-\frac{\phi_{\leftarrow}}{2}\right]\quad.\ ] ] the second turning point @xmath77 must be set to infinity .
this becomes clear when we remember that ( [ wavefunctionleft ] ) was derived from the symmetry properties of a two turning point system : the particle comes in from infinity until @xmath76 is reached , where it is reflected back to infinity - and at infinity it is reflected again into direction of @xmath76 .
thus , a one turning point system follows from a two turning point system by setting one of the two turning points to infinity .
this procedure is necessary , because it is the only way to take into account the symmetry properties of the quantum correction function as discussed above .
together with ( [ phasertol ] ) and the symmetry property of @xmath133 we obtain the wave - function @xmath134 \cos\left[\int_{x_{1}}^{x}\left(p(x')+\re{q_{0}(x')}\right)dx ' - \frac{\phi_{\leftarrow}^{{\rm{inv}}}}{2}\right]\quad .
\label{wavefunctionlinpot}\ ] ] expression ( [ wavefunctionlinpot ] ) perfectely demonstrates how the gauge - dependent part of the reflection - phase cancels , while the gauge - invariant part @xmath135 remains present . as can be seen in fig .
( [ linpot3 ] ) the invariant phase @xmath136 is not the wkb reflection - phase @xmath137 obtained by the asymptotical expansion of the airy - function .
the asymptotical reflection - phase follows from the limit @xmath138 the harmonic oscillator is surely the most considered and best understood quantum system .
this fact makes it a very interesting test object for the viewpoint on quantum mechanics we haves developed here , because by the harmonic oscillator , we can demonstrate at best the differences between schrdinger and hamilton - jacobi quantum mechanics .
we will work in suitably scaled quantities @xmath139 for simplicity , but without loss of generality , we have set @xmath140 in all numerical calculations . in figure ( [ ho1 ] )
we have plotted the quantum mechanical and the classical phase - space of the harmonic oscillator for the ground state energy ( left picture ) .
again we see the typical effect of the quantum correction functions , breaking up the closed classical phase - space , thus making the potential barrier soft and penetrable .
the effect of the quantum correction functions is largest for the ground state , as one would naturally expect , because the ground state is close to the anticlassical limit of the system .
for the tenth excited state , the right picture in ( [ ho1 ] ) shows that the influence of @xmath26 already has declined , the system has already entered its semiclassical limit and quantum corrections become negligible .
we will not present pictures of oscillator wave - functions here , because we think that up to now the reader might be convinced , that the wave - functions determined by ( [ wavefunctionright ] ) or ( [ wavefunctionleft ] ) are , up to the normalization amplitude , identical with the oscillator wave - functions determined by the schrdinger equation .
instead we will focus on the calculation of quantities , that are not available within the schrdinger equation . in figure ( [ ho2 ] )
we have calculated the quantization curve and the maslov - index according the formulae ( [ quantization1 ] , [ quantization2 ] ) .
the quantization curve determines the oscillator eigenstates by knowledge of the classical momentum and its quantum corrections .
although the wave - function is implicitly known when the reflection phases ( [ phaseltor ] , [ phasertol ] ) are known , there is no need to employ them for determining the spectrum .
the knowledge of the quantum correction functions is sufficient for a calculation from first principles .
the maslov - index imposes a tiny oscillating structure onto the quantization curve .
the nodes of these oscillations lie perfectely at @xmath141 , coincidence with the bound states and recover the bohr - sommerfeld rule .
this behaviour of the maslov - index was first calculated in @xcite for the half - sided harmonic oscillator and by a different method . for the harmonic oscillator , for further details
see part i of this paper , we now will discuss the topic of quantum trajectories in detail .
we will calculate and discuss trajectories for the ground state and the tenth excited state to illustrate the influence of the quantum corrections functions near the anticlassical and in the semiclassical regime . by ( [ totaltimeshift ] ) the time - shift between quantum and classical motion , see figure ( [ ho4 ] ) is given as @xmath142 the lower bound in ( [ tratimeshift ] ) is zero , because we want to calculate the trajectory of a particle starting its journey from the minimum of the oscillator potential @xmath143 .
we could have chosen any other point , the choice of the point of reference has no influence on the shape of the quantum trajectory . from the trajectory equation ( [ trajectoryeq ] )
, we get @xmath144 where the index @xmath145 denotes the multiple branches obtained by shifting the time by @xmath146 in the classically allowed region for the forward and backward motion , respectively . by the trajctory equation ( [ trajectoryeq ] ) only the quantum motion between the two turning points can be determined .
the cutoff at the turning points is the cause the multiple branches . if we compare ( [ trasubspace ] ) with the classical trajectory @xmath147\quad . \label{traclassical}\ ] ] and unse the fact that quantum and classical motion are described by the same frequency , see ( [ invariance ] ) , we are lead to make the following ansatz for a quantum trajectory @xmath148,\quad \lim_{k\rightarrow\infty}a_{{\rm{q}}}(k,\omega , t ) = \sqrt{\frac{k^{2}}{\omega^{2}}}\quad . \label{traansatz}\ ] ] the limit in ( [ traansatz ] ) requires , that the quantum amplitude approaches its classical value in the classical limit .
the fact that all quantum effects are now located in an amplitude - function reminds us of the path - integral , where the same phenomenon occurs : phase - quantities are given by the classical action function , but the time - evolution amplitude is quantum mechanical .
however , by comparison between ( [ trasubspace ] , [ traansatz ] ) , we can establish a continous quantum trajectory without cutoffs at classical turning points by analytical continuation beyond the classical turning points , illustrated in figure ( [ ho3 ] ) .
the reader might object this procedure by the argument , that by analytical continuation in this case a kind of information is gained , that does not exist .
we agree with this argument to some extent and will therefor call the trajectories in the classically forbidden region apparent trajectories .
the present situation is similar to the quantum time - shifts in @xcite , where it was found that apparent trajectories enter the classically forbidden region .
but the difference between there to present is , that in @xcite the potential was a singular one - and no particle , classical or quantum can cross a singularity .
but the harmonic oscillator is regular and by the quantum correction functions , the turning point region has become regular , too .
further know we from above , that in the quantum case the potential becomes soft and therefor penetrable .
all this put together means the following : for the harmonic oscillator , the reflection probability at a turning point is @xmath149 for all energies , thus the particle comes back in any case .
the particle reaches the classical turning point with finite momentum , which imposes a time - shift in comparison to the classical particle and a phase - shift on the wave - function as a result of the soft boundary condition to be fullfilled .
the phase - shift of the wave - function and the time - shift of the motion are two manifestation of the same phenomenon , the relative softness of the potential barrier in the quantum mechanical limit .
all these facts allow only one interpretation , the particle as it has arrived at the turning point with finite momentum , goes beyond the turning point and enters the classically forbidden region , appears to be decelerated by the soft barrier , comes to rest and finally moves backward to be expelled again into the classically allowed region .
the only explanation thus is the softness - effect of the potential barrier .
we emphasize , that this softness has nothing to do with sticking or absorptive reactions , it is a pure quantum mechanical effect . for energies approaching the classical limit
, the barrier hardens as is described by the limiting process in ( [ traansatz ] ) .
we have shown how quantum mechanical problems can be solved within the framework of the hamilton - jacobi equation of quantum mechanics , where the central quantity of interest is the quantum mechanical action function instead of the wave - function . to our regards ,
hamilton - jacobi theory and schrdinger theory are fully equivalent to each other , they describe the same physics and differ only in their descriptive manner .
in contrast to the well - known madelung - de - broglie - bohm theory , we do not separate real and imaginary part of the action function initially . a very valuable feature of hamilton - jacobi theory is , that by use of the quantum action function , the wave - function can uniquely be decomposed into running waves in every point in space , regardless whether the wave - functions describes bound or scattering states . in the schrdinger point of view , this is possible only asymptotically .
the use of quantum correction functions enopens a deeper insight into the quantum mechanical behaviour of systems , especially in the vincity of turning points , where the local quantum mechanical momentum is finite .
this property overcomes the wkb - singulariy at classical turning points and furthermore clearifies the special behaviour of a potential barrier in the quantum mechanical case : the barrier becomes soft and therefor allows the quantum particle to enter it .
this softness - property of the potential barrier in the quantum mechanical case is modeled by the soft boundary conditions the wave - function has to obey - and that can be solved exactly by help of the quantum action function - and also explain the finite probability of finding a particle there .
the soft boundary - value problem is solved by a reflection phase .
the reflection phase is the key - quantity to obtain the spectrum of a quantum system with help of a quantization curve and allows a direct calculation of maslov - indices . to obtain the spectrum of a quantum system by this method , only action and momentum functions have to be known , orthonormalized wave - functions are not necessary . especially for potentials where exact wave - functions are not analytically known
, it may be much more simple to calculate the quantization curve instead of searching orthonormalized eigenstates with the schrdinger equation . besides the hamilton - jacobi theory allows us to calculate quantities that are not available within the schrdinger point of view
, the calculation of quantum trajectories is possible .
the quantum trajectories obtained by the trajectory equation ( [ trajectoryeq ] ) are only defined in the classically allowed region , but by analytical continuation , they can be established also for the classically forbidden region in a physically reasonable manner .
the behaviour of the trajectories again mirror the softness - property of the potential barrier in the quantum mechanical case and the finite probability to enter the classically forbidden region is all described by a total time - shift @xmath146 ( [ totaltimeshift ] ) .
in contrast to de - broglie - bohm theories , no quantum potential is introduced , and therefor no equation of motion .
moreover , the equation of motion in de - broglie - bohm theories @xcite and our trajectory equation ( [ trajectoryeq ] ) contradict each other .
this result was also found by @xcite , but was interpreted differentely . although the trajectories obtained from the quantum action integral by use of the hamilton - jacobi formalism show a somewhat quirky behaviour , they are consistent with our results about the behaviour of the quantum correction function in the vincity of classical turning points and the behaviour of the wave - function .
both , quantum correction functions and trajectories illustrate the softness of a potential function near the anticlassical limit
. moreover , hamilton - jacobi formalism is theoretically very well founded and thus we expect ourselves so travel on safe terrain , whereas , as stated in @xcite the equivalence between velocity and quantum mechanical momentum in de - broglie - bohm theories is up to now an unproved assumption .
the work of floyd @xcite interprets stationary hamilton - jacobi theory in a very different way as we have done here .
although starting from the same key equation ( [ hjs ] ) , floyd developed a completely different theory by separating real and imaginary parts of the quantum action function . by the help of a constrained condition similar to ( [ isoenergetics ] ) he arrived at a nonlinear , second - order equation for the real part of the quantum momentum function @xmath150 \quad , \label{floydhj}\end{aligned}\ ] ] where @xmath151 denotes a quantum potential .
if @xmath24 is the classical momentum @xmath27 , we obtain from ( [ floydhj ] ) the well - known quantality function , that describes the break - down of wkb - theory in the vincity of classical turning points , see @xcite .
but if we are interested in in the quantum momentum , we are not allowed to simply write @xmath152 because @xmath151 depends on @xmath24 . moreover ,
if @xmath24 has been obtained by ( [ floydhj ] ) , equation ( [ floydmomentum ] ) reduces to an identity @xmath153 which is of no use either . because of this , we have argued above , that the concept of a quantum potential becomes misleading in stationary hamilton - jacobi theory .
our second objection is , that by transforming ( [ hjs ] ) into a real but second order equation , there occurs a loss of information , because for ( [ floydhj ] ) neither a closed solution formula , nor any initial conditions are known . the apparent arbitrariness for the choice of initial conditions in ( [ floydhj ] ) lead floyd to interpret , that there are infinitely many quantum microsates , associated with infinitely many possible initial conditions , that map onto the same energy - eigenstate . for nonlinear , higher order differential equations , its the crux of matter that no unique solution may exist , instead of a large number of different solutions , depending sensitively on initial conditions .
but from such a multitude of solutions , the physically relevant one must be extracted . in our first order approach , this is naturally done by ( [ quasilintransform ] , [ operatorequation ] ) .
moreover , the action deduced from ( [ floydhj ] ) , the real part of our action function , is assumed to be real on the whole space . comparing this with our findings above we can state that this is impossible , because a necessary condition for the wave - function to fall off in the classically forbidden region is a purely imaginary action function there .
if the concept of the quantum potential would be dropped , this could be deduced from ( [ floydhj ] ) , together with the equations that relate real and imaginary parts of the action function .
we close this comment and emphasize again , that to our regards , the hamilton - jacobi theory of quantum mechanics does not offer a kind of hidden information that is lost in the schrdinger point of view .
the information provided by both theories is the same , only their descriptive manner is different .
99 e. schrdinger , ann . phys . * 79 * , 489 , 1926
. e. madelung , z. phys . *
40 * , 322 , 1926 .
d. bohm , phys .
85 * , 166 , 1952 ; phys . rev . * 85 * , 180 , 1952 ; phys .
rev . , * 87 * , 389 , 1952 ; d. drr : _ bohmsche mechanik als grundlage der quantenmechanik _ , springer verlag berlin , 2001 .
r. holland : _ the quantum theory of motion _ , cambridge university press , 2004 .
r. e. wyatt : _ quantum dynamics with trajectories _ - introduction to quantum hydrodynamics , springer verlag new york , 2005 .
e. r. floyd , j. math .
* 17 * , 881 , 1976 ; j. math . phys . * 20 * , 83 , 1979 ; phys . rev .
d * 25 * , 1547 , 1982 ; phys .
d * 26 * , 1339 , 1982 ; phys .
d * 29 * , 1842 , 1984 ; phys . rev .
d * 34 * , 3246 , 1986 .
a. e. faraggi , m. matone , phys .
lett . a * 249 * , 180 , 1998 ; phys .
b * 450 * , 34 , 1999 .
g. wentzl , z. phys . * 38 * , 518 , 1926 ; h. a. kramers , z. phys .
* 39 * , 828 , 1926 ; l. brillouin , c. r. acad .
paris * 183 * , 24 , 1926 .
h. friedrich , j. trost : _ working with wkb waves far from the semiclassical limit _
, phys . rep . *
397 * , 6 , 2004 .
m. v. berry , k. e. mount : _ semiclassical approximation in wave mechanics _ , rep .
phys . * 35 * , 315 , 1972 .
n. frman , p. o. frman : _ phase integral method _ in : springer tracts in modern philosophy , vol .
40 , springer verlag new york , 1996 .
k. raghunathan , r. vasudevan , j. phys .
a * 20 * , 839 , 1987 . v. b. mandelzveig , j. math
* 40 * , 6266 , 1999 .
h. friedrich , j. trost , phys .
a * 54 * , 1136 , 1996 .
j. butterfield , _ quo vadis quantum mechanics _ , eds : a. elitzur , s. dolev , n. kolenda , springer verlag heidelberg , 2005 .
p. holland , nuovo cimento * 116 * b , p. 1043 and
a. jurisch , h. friedrich , phys .
a * 70 * , 032711 , 2004 . | in this article , we develop quantum mechanics upon the framework of the quantum mechanical hamilton - jacobi theory .
we will show , that the schrdinger point of view and the hamilton - jacobi point of view are fully equivalent in their description of physical systems , but differ in their descriptive manner . as a main result of this ,
a wave function in hamilton - jacobi theory can be decomposed into travelling waves in any point in space , not only asymptotically
. the well known wkb - theory will be show up as a special case of the more general theory , we develop below . by the example of the linear potential and the harmonic oscillator
we will discuss quantum mechanics from the hamilton - jacobi point of view .
soft boundary value problems as the connection problem can be solved exactly . quantizised energies and maslov - indices can be calculated directely without orthonormalizing wave - functions .
also , we will focus on trajectory - themes , which , in contrast to the schrdinger point of view , follow naturally from the quantum mechanical action function . |
i would like to thank members of the zeus collaboration who have worked on the inclusion of heavy flavour data in the pdf fits , particularly : c gwenlan , e tassi , j terron , m wing , m botje .
i thank r thorne and p thompson for useful discussions . | at hera heavy quarks may contribute up to @xmath0 of the structure function @xmath1 .
the introduction of heavy quarks requires an extension of the dglap formalism .
the effect of using different heavy flavour number schemes , and different approaches to the running of @xmath2 , are compared using the zeus pdf fit formalism .
the potential of including charm data in the fit is explored , using @xmath3 double differential cross - sections rather than the inclusive quantity @xmath4 .
parton density function ( pdf ) determinations are usually global fits @xcite , which use inclusive cross - section data and structure function measurements from deep inelastic lepton hadron scattering ( dis ) data as well as some other exclusive cross - ections .
the kinematics of lepton hadron scattering is described in terms of the variables @xmath5 , the invariant mass of the exchanged vector boson , bjorken @xmath6 , the fraction of the momentum of the incoming nucleon taken by the struck quark ( in the quark - parton model ) , and @xmath7 which measures the energy transfer between the lepton and hadron systems .
the differential cross - section for the neutral current ( nc ) process is given in terms of the structure functions by @xmath8,\ ] ] where @xmath9 .
in the hera kinematic range there is a sizeable contribution to the @xmath1 structure function from heavy quarks , particularly charm .
thus heavy quarks must be properly treated in the fomalism .
furthermore fitting data on charm production may help to give constraints on the gluon pdf at low-@xmath6 .
the most frequent approaches to the inclusion of heavy quarks within the conventional framework of qcd evolution using the dglap equations @xcite are : * zm - vfn ( zero - mass variable flavour number schemes ) in which the charm parton density @xmath10 satisfies @xmath11 for @xmath12 and @xmath13 in the splitting functions and @xmath14 function .
the threshold @xmath15 , which is in the range @xmath16 , is chosen so that @xmath17 gives a satisfactory description of the data .
the advantage of this approach is that the simplicity of the massless dglap equations is retained .
the disadvantage is that the physical threshold @xmath18 is not treated correctly ( @xmath19 is the @xmath20 cm energy ) .
* ffn ( fixed flavour number schemes ) in which there is no charm parton density and all charmed quarks are generated by the bgf process .
the advantage of the ffns scheme is that the threshold region is correctly handled , but the disadvantge is that large @xmath21 terms appear and charm has to be treated ab initio in each hard process . *
gm - vfn ( general mass variable flavour number schemes ) , which aim to treat the threshold correctly and absorb @xmath21 terms into a charm parton density at large @xmath5 .
there are differing versions of such schemes @xcite for the main zeus - s analysis @xcite , the heavy quark production scheme used was the general mass variable flavour number scheme of roberts and thorne ( tr - vfn ) @xcite .
however we also investigated the use of the ffn for 3-flavours and the zm - vfn . in fig .
[ fig : schemes ] we compare the fit prediction for @xmath4 using each of these schemes to data on @xmath4 from the zeus collaboration @xcite .
one can see the differences between the ffn and the zm - vfn at threshold where the zm - vfn is clearly inadequate . in this kinematic region
the tr - vfn is more like the ffn .
however , the tr - vfn scheme becomes more like the zm - vfn scheme for @xmath22 .
this comparison illustrates the effect of change in scheme when keeping the pdf parameters fixed . in practice
one should refit the pdf parameters using the alternative schemes .
the result of this is shown in fig [ fig : fitschemes ] .
the difference between the ffn and tr - vfn is not so marked .
it is well known that these choices have some effect on the steepness of the gluon at very small-@xmath6 , such that the zero - mass choice produces a slightly less steep gluon . in fig
[ fig : glusea_schemes ] the differing shapes of the sea and the gluon pdfs for these different heavy quark schemes are illustrated .
figure [ fig : recentcharm ] shows the zeus - s fit predictions for more recent @xmath4 data from zeus and h1 @xcite .
the scheme chosen was ffn for 3 flavours with the renormalisation and factorisation scale for light quarks both set to @xmath5 but the factorisation scale for heavy quarks set to @xmath23 .
the reason for these choices of scheme and scale is that these are the choices made in the programme hvqdis @xcite which was used to extract @xmath4 from data on @xmath3 production .
the scale choice does not make any signficant difference to the predictions ( see later ) .
note that in fig [ fig : recentcharm ] the charm data are shown compared to the zeus - s pdf fit predictions but these data were not input to the fit .
including the zeus charm data @xcite in the zeus - s pdf fit gives no visible improvement to pdf uncertainties . to investigate the potential of charm data to constrain the gluon pdf , we modified the zeus - s pdf fit as follows : all zeus inclusive neutral current and charged current cross - section data from hera - i was included but no fixed target data ; the parametrisation was modified to free the mid-@xmath6 gluon parameter @xmath24 and the low-@xmath6 valence parameter @xmath25 , however the @xmath26 normalisation had to be fixed since there is no information on this without fixed taregt data
this fit is called the zeus - o fit . fig .
[ fig : f2charm ] compares the gluon pdf and its uncertainties as extracted from this zeus - o pdf fit with the those extracted from a similar fit including the @xmath4 data .
this illustrates that the charm data has the potential to constrain the gluon pdf uncertainties .
its lack of impact on the global fit may be because we are not using the charm data optimally .
@xmath4 is a quantity extracted from @xmath3 cross - sections by quite a large extrapolation
. it would be better to fit to those cross - sections directly .
the evaluation of the theoretical predictions involves running the nlo programme hvqdis for each iteration of the fit .
however , one can shorten this process by using the same method as was used for the zeus - jets fit @xcite .
the pdf independent subprocess cross - sections are output onto a grid , such that they can simply be multiplied by the pdfs at each iteration .
the data used are the nine double differential cross - section measurements of @xmath27 @xcite , see fig .
[ fig:9xsecns ] there are further theoretical considerations to be accounted for when unputting @xmath3 cross - sections , as opposed to and inclusive quantity like @xmath4 , to a pdf fit . since the grids are calculated using hvqdis the fit must use the ffn scheme to be compatible .
this means that we can not use zeus high-@xmath5 data , since this scheme is not suitable at high-@xmath5 .
hence we chose to use the zeus - s global fit , which incuded fixed target data , with a cut - off @xmath28gev@xmath29 .
furthermore , it has only recently become evident that since we are using the ffn scheme we must also treat the running of @xmath2 differently than in the vfn schemes . in these vfn schemes
@xmath2 is matched at flavour threholds @xcite , but the slope of @xmath2 is discontinuous at the flavour thresholds . for consistency with hvqdis we must use a 3-flavour @xmath30 which is continuous in @xmath5 .
this requires an equivalent value of @xmath31 in order to be consistent , at low @xmath5 , with the results of using a value of @xmath32 in the usual vfn schemes .
such a 3-flavour @xmath30 has also been used in specialised pdf fits of mrst ( mrst2004f3 ) @xcite , which are used to make predictions for charm production . in fig
[ fig : f2calphas ] we compare different heavy quark factorisation scales and different treatments of the running of @xmath2 for predictions of @xmath4 .
fig .
[ fig : f2balphas ] makes the same comparision for @xmath33 .
we see that within the ffn scheme the choice of the heavy quark factorisation scale makes only a small difference at low @xmath5 .
the treatment of @xmath30 gives larger differences .
the ffn scheme and tr - vfn scheme differ for almost all @xmath5 if @xmath30 runs as for the vfn schemes .
however if a 3-flavour @xmath30 is applied in the ffn schemes there is much better agreement of all schemes at higher @xmath5 .
we now return to consider inputting the @xmath3 cross - sections to the pdf fit .
the zeus - s global fit formalism is used including all zeus inclusive neutral and charged current cross - section data from hera - i and the fixed target data .
the parametrisation was also modified to free the mid-@xmath6 gluon parameter @xmath24 and the low-@xmath6 valence parameter @xmath25 .
this fit is called zeus - s-13 .
figure [ fig : gluonffnzs13 ] shows the difference in the gluon pdf uncertainties , before and afer the @xmath3 cross - sections were input to the zeus - s-13 global fit .
disappointingly the uncertainty on the gluon is not much improved .
should we have expected much improvement ?
there are two aspects of the fit which could be improved .
the predictiond for the @xmath3 cross - sections have more uncertainties than just the pdf parametrization .
a further uncertainty is introduced in the choice of the @xmath34 fragmentation the petersen fragmentation function was used for the fit predictions .
however , looking back at fig [ fig:9xsecns ] we can see that the lund fragmentation function seems to describe the data better .
to best exploit the charm data in future we need to address such aspects of our model uncertainty .
secondly , fig [ fig : d*errors ] compares the fractional errors on the @xmath3 cross - sections with the uncertainty on the prediction for these quantities derived from the uncertainty on the gluon pdf in the zeus - s-13 pdf fit , before inputting the @xmath3 cross - sections .
the data errors are larger than the present level of uncertainty . thus we eagerly await the 5-fold increase in statistics expected from hera - ii charm and beauty data . |
one of the crucial issues in astronomy and astrophysics is to understand formation processes of stars , planetary systems , and their clusters .
they are the most basic constituents in the galaxy and hence , play essential roles in its formation and evolution . in order to understand physics and dynamics in the star - formation processes ,
it is necessary to obtain accurate physical properties of target sources such as size , mass , and luminosity . because all of these values strongly depend on the adopted distance , the distance is the most fundamental parameter for quantitative discussion . in spite of its importance
, however , it has long been difficult to measure accurate distances to star - forming regions without assumptions . the most reliable and direct way to determine
the distance is an annual trigonometric parallax method , based on precise measurements of the position and motion of the object . in 1990
s , the hipparcos satellite measured annual parallaxes for more than 100 000 stars with a typical precision of 1 mas level @xcite .
nevertheless , accurate distance measurements with hipparcos for star - forming regions were limited to nearby sources within a few hundred pc from the sun ( e.g. lombardi et al . 2008 ) . although distances to several ob clusters associated with nearby star - forming regions up to @xmath4600 pc have been reported @xcite
, hipparcos could not directly determine the distances to newly born stars deeply embedded in dusty molecular cloud cores , because hipparcos was capable of observing only in the optical wavelengths .
in the last few years , phase - referencing vlbi astrometry has been developed drastically , yielding the annual parallaxes of nearby star - forming regions .
for instance , radio continuum sources in the taurus @xcite , ophiuchus @xcite , and orion @xcite regions were observed with the nrao very long baseline array ( vlba ) to derive their annual parallaxes with typical uncertainty of better than a few percent . in addition , we have started the initial scientific project `` measurements of annual parallaxes of nearby molecular clouds '' with vera ( vlbi exploration of radio astrometry ) .
vera is a japanese vlbi network dedicated to astrometric observations aimed at revealing 3-dimensional structure of the galaxy @xcite . in our project for nearby molecular clouds , we specially focus on the accurate distance measurements of star - forming regions within 1 kpc from the sun @xcite .
our results will contribute to refine all of the observational studies on nearby star - forming regions by providing their most accurate and directly measured distances .
we have already conducted astrometric observations of the h@xmath0o maser sources in the orion @xcite , ophiuchus @xcite , and perseus @xcite regions .
all of the results provide direct measurements of their distances with much higher accuracy than the previous photometric methods and the parallax measurements with hipparcos . as a part of this project , we present results of astrometric observations of the h@xmath0o maser source iras 22198 + 6336 in a nearby dark cloud core l1204 g .
l1204 is located in the cepheus - cassiopeia molecular cloud complex @xcite and is associated with a well known high - mass star - forming region s140 .
it is also identified to be a molecular cloud wb176 @xcite . according to the molecular line observations of l1204
, it consists of several high density molecular cores labeled by a - g @xcite .
a deeply embedded yso iras 22198 + 6336 is associated with one of the dense cores , l1204 g , and it is identified to be a powering source of molecular outflow @xcite . the centimeter and millimeter counterpart of iras 22198 + 6336 is detected with the vla @xcite .
the h@xmath0o masers were detected @xcite and their distributions were investigated by the vla @xcite and vlba @xcite . however , they did not measure the proper motions of the h@xmath0o masers because observations by @xcite could not achieve sufficient high resolution and @xcite observed only at 1 epoch .
therefore , the kinematics of the circumstellar materials around iras 22198 + 6336 remains still unclear .
the distance to l1204 is estimated by a photometric method to be 910 pc @xcite assuming that a b0v star hd211880 is the ionization source of the h region s140 , which is also associated with l1204 .
on the other hand , @xcite identified hd211880 ( hip110125 ) as a member of the cepheus ob2 cluster . using the hipparcos data , @xcite determined the distance to the cepheus ob2 cluster to be closer value of 615@xmath135 pc from the sun .
the kinematic distance of 1.3 kpc is sometimes adopted in several literatures @xcite according to the systemic velocity of @xmath5 km s@xmath6 @xcite .
depending on the adopted distance , the luminosity of iras 22198 + 6336 could differ by a factor of 4 , and hence , its physical properties sometimes contain large uncertainties . based on our accurate astrometric observations with vera , we will provide the parallactic distance to iras 22198 + 6336 in l1204 g , together with the kinematics of the circumstellar gas traced by the h@xmath0o masers .
we will also discuss the basic properties of the powering source of the h@xmath0o masers by adopting our new distance to l1204 g .
the vera observations of the h@xmath0o masers ( @xmath7-@xmath8 , 22235.080 mhz ) associated with iras 22198 + 6336 have been carried out since nov .
2006 at intervals of about 1 month and are still ongoing .
we present the results from a total of 9 observing sessions from nov .
2006 to dec . 2007
( 2006/312 , 2006/361 , 2007/047 , 2007/096 , 2007/130 , 2007/215 , 2007/282 , 2007/322 , and 2007/353 ; hereafter an observing session is denoted by year / day of the year ) .
all 4 vera stations participated in all sessions , providing a maximum baseline length of 2270 km . observations were made in the dual beam mode ; the h@xmath0o masers associated with iras 22198 + 6336 and a reference source j2223 + 6249 ( @xmath922h23m18.096661s , @xmath10d4933.80393 " ; @xcite ) were observed simultaneously @xcite .
the reference source j2223 + 6249 is separated south from iras 22198 + 6336 by 1.1 degrees .
j2223 + 6249 was detected with the peak flux density of about 200 mjy in all the epochs as shown in figure [ fig - j2223 ] .
left - handed circular polarization was received and sampled with 2-bit quantization , and filtered using the vera digital filter unit @xcite .
the data were recorded onto magnetic tapes at a rate of 1024 mbps , providing a total bandwidth of 256 mhz in which one if channel and the rest of 15 if channels with 16 mhz bandwidth each were assigned to iras 22198 + 6336 and j2223 + 6249 , respectively . a bright continuum source , 3c454.3 , was observed every 80 minutes for bandpass and delay calibration .
system temperatures including atmospheric attenuation were measured with the chopper - wheel method @xcite to be 100 - 500 k , depending on weather conditions and elevation angle of the observed sources .
the aperture efficiencies of the antennas ranged from 45 to 52% depending on the stations .
a variation of the aperture efficiency of each antenna as a function of elevation angle was confirmed to be less than 10% even at the lowest elevation in the observations ( @xmath420 degrees ) .
correlation processing was carried out on the mitaka fx correlator @xcite located at the naoj mitaka campus . for h@xmath0o maser lines , the spectral resolution was set to be 15.625 khz , corresponding to the velocity resolution of 0.21 km s@xmath6 .
data reduction was performed using the nrao astronomical image processing system ( aips ) .
amplitude and bandpass calibrations were done for the h@xmath0o maser source ( iras 22198 + 6336 ) and reference source ( j2223 + 6249 ) independently . for phase - referencing ,
fringe fitting was done with the aips task fring on the phase reference source ( j2223 + 6249 ) , and the solutions were applied to the target source ( iras 22198 + 6336 ) . because the reference source j2223 + 6249 showed the double - peaked structure in the synthesized images as shown in figure [ fig - j2223 ] , we carried out self - calibration to correct for the structure effect on the phase calibration .
in addition , we applied the results of the dual - beam phase calibration , in which difference between instrumentally added phases in the two beams were measured by injecting artificial noise sources into both beams at each station during the observations @xcite .
we also corrected for the approximate delay model adopted in the correlation processing .
the difference in the optical path length error between our newly applied model and the calc9 model developed by the nasa / gsfc vlbi group are confirmed to be less than @xmath42 mm . in this correction ,
drifts of the visibility phase caused by the earth s troposphere were estimated based on the gps measurements @xcite . as discussed in @xcite , the tropospheric zenith delay consists of `` dry '' and `` wet '' parts , and only latter one is variable and
hence significantly cause the phase errors in the calibrations .
the typical tropospheric `` wet '' zenith delay ( path length ) was 5 - 10 cm while it was sometimes as large as 40 cm , in particular at the ogasawara and ishigaki stations in summer . on the other hand ,
the `` dry '' component of the tropospheric zenith delay was stable about 2.3 m. ( 80mm,80mm)fig1.eps synthesis imaging and deconvolution ( clean ) were performed using the aips task imagr with the uniform weighting to achieve the highest spatial resolution , allowing us to resolve complex spatial structures of the masers .
the resultant synthesized beam size ( fwmh ) was typically 1.2 [email protected] mas with a position angle of -40 degrees .
the peak positions and peak flux densities of masers were derived by fitting elliptical gaussian brightness distributions to each image using the aips task sad .
the formal uncertainties in the maser positions given by sad were mostly better than 0.05 mas .
( 80mm,80mm)fig2.eps ( 140mm,140mm)fig3.eps figures [ fig - total ] and [ fig - spectra ] show an example of a total - power spectrum and scalar - averaged cross - power spectra of the h@xmath0o masers associated with iras 22198 + 6336 , respectively .
the h@xmath0o maser lines were detected within the lsr ( local standard of rest ) velocity range from @xmath12 km s@xmath6 to @xmath13 km s@xmath6 .
the intense spectra at the lsr velocity of about @xmath14 km s@xmath6 were detected in all the observed epochs , whereas other velocity components were detected at rarer occurrences . in this paper
, we refer to the brightest components at @xmath14 km s@xmath6 as the main components .
our result is consistent with those of long - term monitoring observations made by @xcite and @xcite that the most prominent maser feature in iras 22198 + 6336 was at the lsr velocity of about @xmath15 km s@xmath6 , while those around the systemic velocity of @xmath5 km s@xmath6 @xcite rarely appeared . according to their long - term monitoring observations ,
the radial velocities of the h@xmath0o masers in iras 22198 + 6336 changed with time over the 11-year monitoring period at a rate of 0.2 km s@xmath6 yr@xmath6 .
the lsr velocity of the @xmath15 km s@xmath6 component has been gradually red - shifted to be about @xmath16 km s@xmath6 @xcite .
this trend predicts @xmath17 km s@xmath6 of the velocity of the main component at around 2007 , as observed in our observations .
however , we can not find any signature of the velocity drift within our observing period probably due to our relatively shorter time span .
since the spectrum taken at the first epoch , 2006/312 , showed most of the velocity components , we first searched for the h@xmath0o masers at the first epoch by producing the synthesized images for all the individual velocity channels with the field of view of @xmath44@xmath114 . as a result , we found roughly 4 clusters of maser spots , or features , spreading over the 300 mas@xmath11300 mas region .
hereafter we define a spot " as emission occurring in a single velocity channel and a feature " as a group of spots which are spatially coincident within the beam size and are detected for at least two consecutive velocity channels . based on this coarse map
, we again carried out synthesis imaging for all the epochs with the field of view of 25.6 [email protected] mas centered at the positions for each of the identified features .
the results of the mapping are shown in figure [ fig - allspot ] , and examples of phase - referenced channel maps are shown in figure [ fig - chmap ] .
we labeled the groups of features a - m in the order of their lsr velocities .
some of the weak components were not detected in the phase - referenced maps but only detected in the self - calibrated images , in which we carried out phase - calibration using the brightest maser spot rather than the reference source j2223 + 6249 , due to the lower dynamic range of the phase - referenced images .
for such spots , we derived their absolute positions by referencing to the absolute position of the brightest spot in the main component that is significantly detected in the phase - referenced images at all the epochs .
the maser distribution shows roughly a circular structure with the radius of about 150 mas as seen in figure [ fig - allspot](a ) .
overall spatial and radial velocity distribution of h@xmath0o maser features are in good agreement with those by the vla @xcite and the vlba @xcite , but we resolve the previously detected components into several spatially distinct maser spots due to our higher spatial resolution .
the main components which are most blue - shifted and brightest , j - m , coincide with both the position and velocity of the c2 component labeled by @xcite , while the most red - shifted components , a - b , coincide in the position and velocity with the c4 component of @xcite .
the northern components near the systemic velocity , f - i , are also coincident with both the position and velocity of the c1 component of @xcite . on the other hand , another group of components c
- e is spatially coincident with c3 of @xcite , whereas their lsr velocities , from -7 to -10 km s@xmath6 , are significantly red - shifted by more than 5 km s@xmath6 .
here we emphasize that the accuracy of our absolute positions with an uncertainty of only 1 mas , which is limited mainly by the position uncertainty of the reference source j2223 + 6249 , 0.90 mas and 1.31 mas in right ascension and declination , respectively @xcite , are about two orders of magnitude higher than those by the previous vla observations @xcite and by the fringe - rate maps obtained with the vlba @xcite .
the properties of the features a - i are summarized in table [ tab - features ] .
these parameters are derived from the phase - referenced images except for the two features , d and e , as noted in table [ tab - features ] .
they are compact and almost no significant spatial structures , and hence , we only summarize the parameters for the peak spot in each feature .
all of the features listed in table [ tab - features ] were detected each only at a single one different epoch probably due to their high variability .
on the other hand , the brightest main features , j - m , show complex spatial structures . we summarize the parameters for the selected 26 spots in table [ tab - spots ] , which are detected in at least three epochs , and
hence , are usable to obtain reliable proper motions as discussed later .
the parameters for the spots l1 - 3 are derived from the self - calibrated images while others are from the phase - referenced images .
the main features are resolved into four subgroups ; j @xmath18 , k @xmath19 , l @xmath20 , and m @xmath21 in right ascension and declination , respectively , and all of them consist of multiple spots . in particular , the component m shows the elongated structure along the north - south direction from 0 mas to @xmath22 mas in declination , as shown in figures [ fig - central ] and [ fig - chmap ] .
the relatively large scale structures found in these features have been inferred from the fact that the h@xmath0o masers are highly resolved - out with the longer baselines in the vlba @xcite .
we also found that the flux densities of the cross - power spectra of the masers are typically only about 30% of those of the total - power spectra as shown in figures [ fig - total ] and [ fig - spectra ] , which are suggestive of the missing flux .
although the peak velocities of the main features , @xmath14 km s@xmath6 were stable , their internal structures were found to be highly variable .
the features k and l appeared later than 2007/215 , while m was detected only at the earlier epochs than 2007/215 . in contrast , the feature j was highly variable but detected throughout whole of the observing period .
( 160mm,160mm)fig4.eps ( 80mm,80mm)fig5.eps as depicted in figure [ fig - allspot](a ) , we found a radial velocity gradient along the east - west direction when we consider the clusters of features a - b and j - m ; blue - shifted ( j - m ) and red - shifted(a - b ) components tend to be distributed at the western and eastern sides of the map , respectively .
the similar radial velocity gradient has been reported by @xcite and @xcite which is attributed to the bipolar outflow . on the other hand , the larger - scale ( @xmath231 ) velocity gradient along the northwest - southeast direction has been found in the spectral line observations of nh@xmath24 @xcite and c@xmath25o @xcite .
although they attributed these velocity gradients to the molecular outflow , their velocity structures showed an opposite trend to our h@xmath0o maser distribution ; the blue - shifted and red - shifted components traced by the nh@xmath24 and c@xmath25o lines tend to be located at southeast and northwest , respectively .
it would be suggested that the large - scale velocity gradient does not directly related to that traced by the h@xmath0o masers .
detailed discussion about the spatial and velocity structures of the h@xmath0o masers will be presented later .
lccrrr & @xmath26 & detected & @xmath27 & @xmath28 & @xmath29 + no & [ km s@xmath6 ] & epoch & [ mas ] & [ mas ] & [ jy beam@xmath6 ] + a & -3.15 & 2006/312 & 293.43 & 60.22 & 1.9 + b & -3.57 & 2006/312 & 302.30 & 47.19 & 0.74 + c & -8.20 & 2006/312 & 171.42 & -128.00 & 3.9 + d@xmath31 & -8.41 & 2007/353 & 173.85 & -141.05 & 2.2 + e@xmath31 & -8.83 & 2007/353 & 174.73 & -136.19 & 2.7 + f &
-9.68 & 2006/312 & 36.88 & 145.81 & 11.5 + g & -13.05 & 2007/215 & -24.59 & 207.43 & 1.9 + h & -14.31 & 2006/312 & -5.22 & 199.21 & 15.7 + i & -14.31 & 2007/353 & -39.40 & 218.13 & 7.2 + + + + + + lccrrrrr & @xmath26 & detected & @xmath27 & @xmath28 & @xmath29 & @xmath32 & @xmath33 + no & [ km s@xmath6 ] & epochs & [ mas ] & [ mas ] & [ jy beam@xmath6 ] & [ mas yr@xmath6 ] & [ mas yr@xmath6 ] + j1 & -16.42 & 1 * 3**67 * 9 & 4.15 & -23.17 & 3.7 & -2.33(25 ) & -2.32(32 ) + j2 & -16.63 & 1 * 3**67 * 9 & 4.14 & -23.31 & 6.4 & -2.48(25 ) & -2.20(32 ) + j3 & -16.84 & 1 * 3**67 * 9 & 4.03 & -23.33 & 3.0 & -2.27(25 ) & -2.10(32 ) + k1 & -17.05 & * * * * * * 789 & 14.21 & 20.60 & 2.7 & -2.02(168 ) & 3.00(213 ) + k2 & -17.26 & * * * * * 6789 & 15.53 & 20.48 & 4.7 & -0.99(83 ) & 3.18(105 ) + k3 & -17.47 & * * * * * 6789 & 15.51 & 20.44 & 8.2 & -0.58(83 ) & 3.71(105 ) + k4 & -17.68 & * * * * * 6789 & 15.48 & 20.46 & 8.5 & -0.55(83 ) & 3.71(105 ) + k5 & -17.89 & * * * * * 6789 & 15.51 & 20.46 & 6.8 & -0.40(83 ) & 3.79(105 ) + k6 & -18.10 & * * * * * 67
* 9 & 15.62 & 20.36 & 2.7 & -1.29(88 ) & 3.80(110 ) + l1@xmath31 & -17.68 & * * * * * 6789 & -0.99 & -11.88 & 12.8 & -1.45(83 ) & -2.63(105 ) + l2@xmath31 & -17.89 & * * * * * 6789 & -0.95 & -11.94 & 10.1 & -2.53(83 ) & -1.85(105 ) + l3@xmath31 & -18.10 & * * * * * 6789 & -1.02 & -11.77 & 2.5 & -2.00(83 ) & -2.44(105 ) + m1 & -17.05 & 123456 * * * & -0.90 & -3.40 & 19.6 & -3.44(41 ) & -2.00(50 ) + m2 & -17.26 & 123456 * * * & -0.88 & -3.37 & 60.2 & -3.45(41 ) & -1.23(50 ) + m3 & -17.47 & 123456 * * * & -0.87 & -3.53 & 48.8 & -3.43(41 ) & -1.35(50 ) + m4 & -17.68 & * 2 * 456 * * * & -1.52 & -5.48 & 10.4 & -2.98(55 ) & -0.75(69 ) + m5 & -17.89 & * 2 * 456 * * * & -1.51 & -5.47 & 8.9 & -2.96(55 ) & -0.26(69 ) + m6 & -18.10 & * * * 456 * * * & -0.12 & -5.75 & 0.6 & -3.73(98 ) & 0.70(125 ) + m7 & -17.47 & 1**456 * * * & -0.68 & -1.82 & 42.6 & -3.49(45 ) & 0.26(55 ) + m8 & -17.68 & 1**456 * * * & -0.72 & -2.28 & 20.4 & -3.44(45 ) & 1.11(55 ) + m9 & -17.89 & * * * 456 * * * & -0.08 & -2.55 & 1.2 & -3.28(98 ) & 0.83(125 ) + m10 & -18.10 & * 2 * 456 * * * & -2.48 & -2.67 & 3.3 & -0.90(55 ) & 0.30(69 ) + m11 & -17.05 & 1 * 3**6 * * * & -0.52 & -0.35 & 2.7 & -3.89(45 ) & -0.01(56 ) + m12 & -17.68 & 1234 * * * * * & -0.50 & -0.39 & 175.7 & -3.39(78 ) & 0.80(96 ) + m13 & -17.89 & 12345 * * * * & -0.47 & -0.22 & 146.1 & -3.62(60 ) & -0.24(72 ) + m14 & -18.10 & 1234 * * * * * & -0.45 & -0.21 & 59.5 & -3.24(78 ) & 0.99(96 ) + + + + + + + + + ( 160mm,160mm)fig6.eps we conducted monitoring observations of the h@xmath0o masers for about one year , and the absolute motions of the maser spots were successfully obtained by referencing to the extragalactic position reference source j2223 + 6249 .
the movement of each maser spot can be expressed by the sum of a linear motion and the annual parallax .
it can be assumed that all the maser features or spots in l1204 g have the same annual parallax within the astrometric accuracy of vera .
therefore , we can simultaneously determine the proper motions ( @xmath34 , @xmath35 ) and the initial positions ( @xmath36 , @xmath37 ) in right ascension and declination for the @xmath38-th feature or spot , along with a common annual parallax of l1204 g , @xmath39 , by the least - squares analysis of all the detected features or spots .
because only the features in the main components , j - m , were detected at more than two epochs , we excluded the features a - i in the following astrometric analysis . to obtain the accurate annual parallax
, we attempted to analyze the data by adopting four different datasets ; ( 1 ) both the right ascension and declination data for all the spots , ( 2 ) only the right ascension data for all the spot , ( 3 ) same as ( 1 ) but for all the features , and ( 4 ) same as ( 2 ) but for all the features .
as already defined , the feature is the sum of the spatially coincident maser spots integrated over at least two consecutive velocity channels .
thus , we determined the peak position of each feature by the gaussian fitting of the velocity - integrated phase - referenced intensity map . on the other hand ,
the peak position of each spot is derived from the individual channel map obtained from the phase - referenced analysis by assuming that the velocity of the maser spot is constant , i.e. the maser spots are regarded to be identical if their velocities are the same
. it would not always be valid if the masers show significant velocity drift @xcite .
however , we applied this assumption to the identification of the detected spots between different epochs because we could not find significant velocity drift in our observations as mentioned above .
comparison of former and latter datasets shows the effect of the structures in the maser features on our astrometric accuracy .
as mentioned above , we found significant spatial structures in the features j - m .
in particular , the elongated structure in the feature m can be clearly seen in figure [ fig - central ] , in which the peak positions of the maser spots do not always coincide with those of the features .
these complex spatial and velocity structures within the feature m would degrade the accuracy of astrometry as reported in the previous results of vera for nearby sources @xcite .
instead , we resolved the internal structure of the feature m by producing the channel maps of all the velocity components .
they could better trace the motion of the masing gas in the multi - epoch astrometric analysis . according to their positions and velocity channels , we identified in total 5 features ( 1 in feature j - l for each and 2 in feature m ) and 26 spots in the main components j - m as listed in table [ tab - spots ] , which were detected in at least three observing epochs .
some of the spots , for instance , the spots j1 - 3 , k6 , m4 - 5 , m7 - 8 , and m10 - 11 , were not always detected probably due to the variability and the insufficient sensitivity .
however , our identification of the maser spots is proved to be appropriate because their motions are well fitted to the linear proper motions and annual parallax which is common for other spots , as shown in figure [ fig - parallax ] .
thus , all of the proper motions identified in our analysis really trace the bulk motion of the masing gas . in addition , we evaluate the effect of the atmospheric zenith delay residual by comparing the datasets 1 and 3 with those of datasets 2 and 4 .
according to the previous results with vera @xcite , the accuracy of the derived annual parallax is significantly improved when only the data for right ascension are used , as this data is less affected by atmospheric modeling errors .
therefore , we ran the fitting by making use of right ascension data alone and both right ascension and declination data for comparison .
all the results for different four datasets can be compared in table [ tab - results ] .
the results of the fitting are also displayed in figure [ fig - parallax ] for the dataset 1 .
the standard deviations of the post - fit residuals are @xmath40.2 mas and @xmath40.3 mas in right ascension and declination , respectively .
if we employ the formal uncertainties in the gaussian fitting of the maser spots , @xmath40.05 , as the positional errors in our observations , a reduced-@xmath40 is as large as 20 .
this result would suggest that we underestimate the positional errors and that these post - fit residuals represent the actual positional accuracy of the astrometric observations with vera . thus , we took into account @xmath41 and @xmath42 listed in table [ tab - results ] in the least - squares analysis to give a weight , which is inversely proportional to the square of the error , for the right ascension and declination data , respectively . as can be seen in table [ tab - results ] ,
all of the 4 datasets yield the consistent parallax values within the mutual errors .
although none of the spots were persistent during whole of the observing period due to their variability , we could determine the annual parallax with an uncertainty of smaller than 10% by successfully identifying multiple spots during almost 1 year .
the highest accuracy is achieved by adopting both right ascension and declination data for all of the 26 detected spots , dataset 1 .
the resultant value of the annual parallax for the dataset 1 is [email protected] mas , corresponding the distance to iras 22198 + 6336 of 764@xmath127 pc .
the associated uncertainties of only @xmath44% for the dataset 1 is smaller than the others , and this dataset provides the most reliable value .
detailed discussions about the possible astrometric error sources will be presented in the section [ sec - error ] .
( 140mm,140mm)fig7.eps lcccc parameter & dataset 1 & dataset 2 & dataset 3 & dataset 4 + dataset & ra and dec & ra only & ra and dec & ra only + & 26 spots & 26 spots & 5 features & 5 features + total number of data & 218 & 109 & 52 & 26 + numbers of fitted parameters & 105 & 53 & 21 & 11 + degrees of freedom & 113 & 56 & 31 & 15 + @xmath39 ( mas ) & 1.309(47 ) & 1.289(66 ) & 1.243(80 ) & 1.288(107 ) + @xmath43 ( pc ) & 764(27 ) & 776(40 ) & 805(52 ) & 777(65 ) + @xmath41 ( mas ) & 0.23 & 0.23 & 0.22 & 0.23 + @xmath42 ( mas ) & 0.29 & & 0.32 & + + + + + along with the annual parallax , proper motions of the maser spots are derived , as summarized in table [ tab - spots ] and illustrated in figure [ fig - allspot](b ) .
the proper motion of 1 mas yr@xmath6 corresponds to 3.6 km s@xmath6 at the distance of 764 pc .
although the proper motions for some of the features contain large uncertainties , we can see a systematic trend in the direction and magnitude of the proper motion vector in figure [ fig - allspot](b ) .
the kinematics traced by the proper motions as well as the radial velocities of the masers will be discussed in the section [ sec - model ] .
as discussed in the recent works on astrometry with vlba and vera @xcite , it is difficult to estimate individual error sources induced in the vlbi astrometry quantitatively . in this paper
, we regard the post - fit residuals of the least - squares analysis listed in table [ tab - results ] as the typical positional errors , which results in a reduced-@xmath40 of unity . because they are significantly larger than the formal uncertainties in the gaussian fitting of the maser spots , @xmath40.05 mas , some systematic errors would have significant effects on the astrometry . here
we discuss about the possible error sources in the present results .
one of the serious error sources is the atmospheric zenith delay residuals mainly due to the tropospheric water vapor .
this effect is caused by the difference in the optical path lengths through the atmosphere between the target and reference sources because their elevation angles are usually different .
@xcite presented results of detailed simulations to calculate the positional errors caused by the atmospheric zenith delay residuals in astrometry with vera .
they demonstrated that the positional error due to the zenith delay residual of 2 cm , which is a typical value of vera observations , is better than 0.030 mas and 0.018 mas for right ascension and declination , respectively , in the case of the declination of 60 degrees , separation angle ( sa ) between the two sources of 1 degree , and the position angle ( pa ) on the sky from the target source to the reference source of 180 degrees , which are valid for the iras 22198 + 6336 and j2223 + 6249 pair aligned along the north - south direction .
it turns out that the simulated positional error for this pair can not fully explain the observed positional uncertainties , unless we introduce an unrealistically large scatter of the zenith delay residual of @xmath430 cm for each of the observation .
therefore , it is unlikely that the atmospheric zenith delay residuals is the most dominant source of our astrometric errors .
alternatively , we propose that the principal error source in the present results would be the temporal variation of the internal structure of the masers , as proposed by some previous results from the vera astrometry @xcite . this effect would be more serious for nearby sources than the distant ones .
in fact , we can see spatial structure of the features in figure [ fig - central ] , and some of the spots in figure [ fig - parallax ] show significant deviations from the best fit model . even in the case of the fitting of the spots rather than the features
, the internal structures in the individual spots might be still significant .
although this effect would limit the positional accuracy of each observation , the accuracy of the annual parallax measurements can be improved with increasing number of spots used in the analysis if the variation of the structure behaves in the random way .
indeed , the post - fit residuals are reduced almost inversely proportional to the square root of the degrees of freedom in the fitting as listed in table [ tab - results ] . on the other hand
, we confirmed that the uncertainties in the station positions , path length errors due to ionosphere , and the delay model employed in our data analysis , which is consistent with the calc9 model , would have negligible effects on our astrometry with vera @xcite .
the uncertainties in the absolute position of the reference source j2223 + 6249 , 0.90 mas and 1.31 mas in right ascension and declination , respectively @xcite , do not affect the derived annual parallax and proper motion because they would add only a constant offset to the position of the maser features . the spatial structure and its temporal variation of the reference source j2223 + 6249
possibly contribute to the positional errors of the maser features .
however the structure effect of the reference source would be much less significant than those of the masers because we performed self - calibration of j2223 + 6249 to solve its structure . according to the dynamic range of the self - calibrated image of j2223 + 6249 ( @xmath420 ; figure [ fig - j2223 ] ) ,
the positional uncertainty caused by the structure of the reference source would be less than 0.025 mas ( hpbw/2snr where hpbw and snr represent the synthesized beam width and signal - to - noise ratio , respectively ) . in summary
, we conclude that the astrometric accuracy in our observations is mainly limited by the structure of the maser features , and that the atmospheric zenith delay residuals also contribute to some extent to the error sources .
as suggested by @xcite , tropospheric zenith delay residual also plays as a random error because they are correlated only over a few days .
thus , one can reduce the post - fit residuals by employing longer observing epochs as well as using larger number of masers .
thus , the longer monitoring observatios will be helpful to achieve higher precision and accuracy of the parallax measurements , which is still underway .
we derived the distance to l1204 g to be 764@xmath127 pc from the sun .
this value is smaller than the earlier photometric result , 910 pc @xcite , and the kinematic distance , 1.3 kpc @xcite .
while they did not report the errors in the derived distances , photometric and kinematic distances would generally contain larger uncertainties by a factor of a few tens of percent than the parallactic distances ( e.g. hirota et al .
2008 , sato et al .
2008 , and references therein ) .
the closer distance , 615@xmath135 pc , is reported by @xcite based on the hipparcos measurements of the annual parallaxes of the cepheus ob2 cluster members , in which one of the members hd211880 is the ionization source of s140 that is possibly associated with l1204 .
this value is offset from our measured distance of iras 22198 + 6336 by 149@xmath144 pc and is slightly larger than the apparent diameter of the cepheus ob2 cluster of about 130 pc ( @xmath44 degrees in the plane of the sky ; de zeeuw et al .
if l1204 g is physically associated with the cepheus ob2 cluster , our result suggests that l1204 g would be located at the farthest side of the cepheus ob2 cluster .
our parallactic distance to iras 22198 + 6336 is found to be very close to the well studied h@xmath0o maser sources in the cepheus - cassiopeia region such as ic1396n @xcite and cepheus a @xcite , although they are not based on the parallax measurements .
together with parallactic distances of the maser sources in this region including ngc281 reported by @xcite and further measurements such as for ic1396n and cepheus a , we will be able to reveal the 3-dimensional structure of this molecular cloud complex and refine the statistical studies of star - formation in this region @xcite .
although the adopted fir luminosity of iras 22198 + 6336 is different from literature to literature , it has been included in the sample of massive ysos judging from the relatively large far - infrared or bolometric luminosity @xcite .
for instance , @xcite adopted @xmath45 to be 250@xmath2 while @xcite adopted @xmath46 to be 580@xmath2 by assuming the same distance @xmath4900 pc .
in contrast , @xcite assumed the kinematic distance of 1.3 kpc and derived the larger fir luminosity of 1240@xmath2 .
very recently , @xcite reported results of a survey of millimeter and centimeter continuum emission from high - mass to intermediate - mass ysos including iras 22198 + 6336 , of which they adopted the luminosity of 1300@xmath2 .
based on the radio continuum observations with the vla of iras 22198 + 6336 along with the observed flux densities reported by previous literatures , @xcite presented basic physical properties of the radio counterpart of the iras source , vla 2 , such as mass and luminosity .
however , all of these results employed the larger distances than our present result . here
we refine the physical properties of the possible powering source of the h@xmath0o masers , iras 22198 + 6336 vla 2 , based on our newly derived distance and previously reported flux values by multiplying a scaling factor proportional to an appropriate power of the distance @xmath43 .
for instance , a factor of ( 764 pc/1300 pc)@xmath47=0.35 is adopted for correcting the luminosity reported by @xcite . first of all , we stress that the bolometric luminosity is reduced to be 450@xmath2 , which is only 35% of previously adopted value , 1300@xmath2 .
such a low luminosity would argue against a classification of iras 22198 + 6336 as a massive yso .
the derived luminosity set limits on the stellar mass of iras 22198 + 6336 .
if it is a zero - age main sequence ( zams ) star , the spectral type of late - b ( between b6 and b7 ) with the stellar mass of about 5@xmath3 @xcite well accounts for the derived luminosity . on the other hand , if it is in a pure accretion phase , the stellar mass of 7@xmath3 is derived from the accretion luminosity with the typical mass accretion rate of 10@xmath48 yr@xmath6 and the stellar radius of 5@xmath49 @xcite . the newly obtained total ( gas+dust ) mass of iras 22198 + 6336
, ( 7 - 10)@xmath3 , which is only 35% of that derived from the sed fitting @xcite , is in good agreement with the above values .
thus , our result implies that iras 22198 + 6336 is not a massive but an intermediate - mass yso with the mass of no more than 7@xmath3 .
the lack of the iras 12 @xmath50 m emission as well as no near - infrared counter part in the 2mass images support the hypothesis that iras 22198 + 6336 is deeply embedded intermediate - mass yso , which is analogous to low - mass class 0 source @xcite . according to @xcite , iras 22198 + 6336
is associated with centimeter radio continuum emission , which could arise either in an ultra - compact h region or a shock - ionized region in the outflow .
they concluded that the contribution of shock - ionized gas around iras 22198 + 6336 should be only a small fraction of the observed centimeter emission because of an extremely high ratio of the centimeter continuum luminosity @xmath51 to the outflow momentum rate , @xmath52 , with an efficiency factor of 90% .
however , if we assume the distance of 764 pc , the ratio is reduced by a factor of 0.59 because the luminosity is proportional to @xmath53 while the outflow momentum rate , which is usually derived from the linear size , radial velocity , and inclination angle of the outflow , is proportional to @xmath43 . even if this is the case ,
the efficiency of 50% is still higher than the typical low - mass ysos , @xmath410% @xcite , suggesting that it would be significantly higher in the very early - stage of intermediate - mass ysos .
the mass and luminosity of iras 22198 + 6336 refined by our newly derived distance is consistent with that of the intermediate - mass deeply embedded yso with the spectral type of late - b .
nondetection of the 6.7 ghz ch@xmath24oh masers , which is thought to be a signpost of massive star - formation site , toward iras 22198 + 6336 @xcite also supports our conclusion that iras 22198 + 6336 is not a massive yso .
although the oh masers , which is another signature of massive star - formation associated with an ultra - compact h region , are detected toward iras 22198 + 6336 @xcite , relatively lower resolution of the oh maser observations , @xmath43 , might give argument against their association with the same powering source of the h@xmath0o masers .
further observational studies with the higher - resolution and higher - sensitivity centimeter , millimeter , and submillimeter interferometers would shed light on the properties of the central star in l1204 g more clearly
. such detailed observational studies on iras 22198 + 6336 would be quite important because it is one of the rare sources in the very early stage of intermediate - mass yso associated with the h@xmath0o masers ( e.g. patel et al .
2000 ) and hence , good target sources for the complete understanding of the very early stage of intermediate - mass star - formation processes .
together with the annual parallax , we successfully measured the proper motions of the maser spots around iras 22198 + 6336 .
these results enable us to discuss in detail about the kinematics of the circumstellar materials . prior to discussions
, we should note that the proper motions obtained with vera do not represent the `` intrinsic '' proper motions of the masers because they are measured with respect to the sun and hence , include the contribution of the solar motion and the galactic rotation . assuming the solar motion relative to the lsr based on the hipparcos satellite data , ( @xmath54)=(10.00 , 5.25 , 7.17 ) km s@xmath6 @xcite
, we estimated the contribution of the solar motion to the observed absolute proper motion to be 3.612 mas yr@xmath6 and @xmath55 mas yr@xmath6 in right ascension and declination , respectively .
in addition , contributions of the galactic rotation is calculated to be ( @xmath56 mas yr@xmath6 , @xmath57 mas yr@xmath6 ) on the assumption of the flat rotation with @xmath58 of 8.0 kpc @xcite and @xmath59 of 236 km s@xmath6 @xcite for l1204 g at the distance of 764 pc .
subtracting these values from the observed proper motions listed in table [ tab - results ] and shown in figure [ fig - allspot](b ) , we obtained the intrinsic proper motions of the h@xmath0o masers corrected for the solar motion and galactic rotation .
the origin of these proper motions is attributed to either due to the systemic motion of the host cloud l1204 g itself , the internal motions of the masers with respect to the powering source iras 22198 + 6336 , such as jets and rotating disk associated with it , or both .
regarding to the systemic motion , observed proper motions might trace the peculiar motion of l1204 g with respect to the galactic rotation .
the discrepancy between the kinematic distance of l1204 g , which is obtained by assuming that l1204 g exactly follows the galactic rotation curve , and our parallactic distance would be suggestive of the effect of the peculiar motion of l1204 g rather than the galactic rotation . indeed , l1204
g ( @xmath60 degrees ) is situated at 75 pc above the galactic plane with a half - thickness of 87 pc @xcite , which seems too large to apply the flat rotation model .
however , the systemic motion of l1204 g is quite uncertain because of a lack of the proper motion measurements of iras 22198 + 6336 , as suggested in the previous result with vera of ngc 1333 @xcite .
therefore , here we only consider several possibilities for the origin of internal proper motions of the masers and the kinematics in the circumstellar material .
as shown in figure [ fig - allspot](a ) , the proper motions of the features j - m provide a hint of systematic motion directed outward from the center of the circular structure with the approximate diameter of @xmath4300 mas or 200 au .
it is naturally interpreted that the most blue - shifted and red - shifted components trace the interacting region with the ambient gas and the collimated jets aligned in the east - west direction , and that the powering source is located at the center of the maser distribution ( see figure [ fig - allspot](a ) ) .
this is roughly consistent with the position and elongation of the vla source observed at the 7 mm band @xcite , although their spatial resolution is much larger than our vlbi observations .
the possible trend of the northward proper motion vectors in figure [ fig - allspot](b ) could be due to the unknown systemic motion of the powering source .
the close - up view of the main features shows an arc - like structure indicative of a bow shock , as seen in figure [ fig - allspot](b ) .
such bow shock structures at the tips of the collimated jets have been found in low - mass class 0 protostars iras 05413 - 0104 @xcite and s106 fir @xcite . in this case ,
the origin of the other maser spots close to the systemic velocity , which seem to be distributed perpendicular to the possible jets , would be accounted for by the rotating disk .
the radial velocity difference along the north - south direction , @xmath48 km s@xmath6 and their separation , @xmath61 au ( @xmath4300 mas ) , implies the keplerian rotation with the centrally enclosed mass of 1.6@xmath3/(cos @xmath38)@xmath47 where @xmath38 is the inclination angle between the rotation axis and the line - of - sight .
the inclination angle of the disk @xmath38 can be estimated from that of the east - west jets traced by the most blue - shifted and red - shifted masers as follows .
the mean proper motions of the most blue - shifted masers are derived to be @xmath62 mas yr@xmath6 and @xmath63 mas yr@xmath6 in right ascension and declination , respectively .
if we adopt the proper motions only in the right ascension direction as the outward velocity of the jets , the mean transverse velocity of 5.5 km s@xmath6 toward west and the mean radial velocity of -6.6 km s@xmath6 with respect to the systemic velocity for the maser spots j - m give the inclination angle @xmath38 of 50 degrees .
similarly , if we adopt the proper motions of both the right ascension and declination direction , the mean transverse velocity of 12.6 km s@xmath6 yields the inclination angle @xmath38 of 66 degrees .
the enclosed mass is estimated to be 3.9@xmath3 and 9.7@xmath3 for former and latter case , respectively , and these values are almost consistent with the result of the sed fitting @xmath64 .
on the other hand , the radio continuum emission in the 1.3 cm and 7 mm bands showed elongated structure along the northwest - southeast direction @xcite . according to their sed analysis , the 1.3 cm emission
is dominated by the free - free emission from the shock - ionized jets while the 44% of the 7 mm flux comes from the thermal dust emission in the circumstellar disk , and hence , the 7 mm emission would be due to a superposition of the ionized jet and the dust disk @xcite .
the direction of the radio continuum emissions are consistent with that of the large - scale velocity gradient @xcite , and hence , @xcite attributed them to the free - free emission from the ionized radio jets .
however , the large - scale velocity gradient showed opposite trend to our h@xmath0o maser distribution as mentioned in section [ sec - distribution ] .
the maser distributions might trace an interface between the ambient material and the outflow lobe which is parallel to the radio continuum source ( e.g. moscadelli et a. 2006 ) , whereas the asymmetric distribution of the masers with respect to the outflow axis can not be accounted for by the simple model only .
it is worth considering another possibility of the rotating disk around the yso with a rotation axis along the north - south direction .
if we assume the radial velocity gradient along the east - west direction , @xmath416 km s@xmath6 per @xmath61 au ( @xmath4300 mas ) , as a cause of rotation , the enclosed mass is estimated to be at least 6.4@xmath3 , which is comparable to the total mass or stellar mass of iras 22198 + 6336 as discussed in the previous section . in this case , the northern and southern components c - i would trace the jets from the central yso . their radial velocity close to the systemic velocity would indicate that the jets are aligned almost within the plane of the sky and hence , the disk is almost edge - on view .
the directions of the radio continuum emissions in the 1.3 cm and 7 mm bands @xcite are in good agreement with that of the h@xmath0o maser features c - i .
nevertheless , this model can not explain the observed velocity drift of 0.2 km s@xmath6 yr@xmath6 for the @xmath15 km s@xmath6 feature @xcite because the velocity drift at the edge of the disk ( at the radius of 100 au ) is estimated to be only 0.01 km s@xmath6 yr@xmath6 in the case of the yso with the mass of 5@xmath3 .
thus , the rotating disk with the axis of north - south direction is inadequate to explain the kinematics in the circumstellar materials around iras 22198 + 6336 probed by the blue - shifted maser features j - m .
alternatively , distributions and proper motions of the features j - m could also be interpreted as a part of the smaller - scale expanding shell as shown in figure [ fig - allspot](b ) , with the radius of [email protected] mas ( [email protected] au ) and the expanding velocity of [email protected] km s@xmath6 in the plane of the sky . in this case , the dynamical time of the expanding shell is calculated to be only 8 years .
basic characteristics of this expanding shell is quite similar to the case of cepheus a @xcite and w75n @xcite .
meanwhile , the nature of the possible powering source of the shell is not consistent with the position and elongation of the radio continuum emission @xcite . in addition , it is unlikely that such a shell is traced only by the most blue - shifted components .
thus , we rule out the possibility of the small - scale expanding shell as a kinematical model of the blue - shifted maser features j - m .
@xcite proposed three possible mechanisms to account for the velocity drift of the maser emission ; a rotating circumstelar disk , acceleration / deceleration of the jets , or a precession of the jets . according to our discussion ,
most plausible explanation for the proper motions , radial velocity distribution , and velocity drifts of the maser features is the collimated bipolar jets along the east - west direction .
thus , the origin of the velocity shift would be the deceleration ( because the radial velocity of the blue - shifted features became closer to the systemic velocity ) or precession of the jets . to distinguish between possible kinematic models , proper motion measurements of red - shifted components located at the eastern end of the maser feature distributions
would be crucial .
in addition , further high resolution observations of centimeter to submillimeter continuum emission as well as near- and mid - infrared observations of the central yso and collimated jets , if they are associated , would help to reveal the overall circumstellar structure of this source with accuracies of much better than 0.1 .
in this paper , we presented the results of multi - epoch vlbi astrometry with vera of the 22 ghz h@xmath0o masers associated with iras 22198 + 6336 in l1204 g . using the absolute positions of total 26 maser spots that were detected in at least three observing sessions
, we derived the annual parallax of iras 22198 + 6336 to be [email protected] mas , corresponding to the distance of 764@xmath127 pc from the sun .
this is the most accurate distance to l1204 g with the uncertainty of only 4% . because the newly derived distance is closer than those well adopted in the previous literatures ( 910 pc or 1.3 kpc ) ,
the bolometric luminosity of iras 22198 + 6336 should be significantly reduced . according to the sed of iras 22198 + 6336
, the bolometric luminosity and total mass of iras 22198 + 6336 were refined to be 450@xmath2 and 7@xmath3 , respectively .
these values are not consistent with a massive yso but are in agreement with an intermediate - mass yso .
we confirmed that iras 22198 + 6336 would be a deeply embedded intermediate - mass yso analogous to a low - mass class 0 source @xcite .
together with the annual parallax , we measured the absolute proper motions of the h@xmath0o masers .
we discussed several kinematical models of the circumstellar gas such as a small scale expanding shell , collimated jets , or rotating disk .
we proposed that the distributions , proper motions , and the previously reported velocity drifts @xcite of the maser features associated with iras 22198 + 6336 are consistent with the collimated bipolar jets and possibly rotating disk perpendicular to them .
iras 22198 + 6336 is one of the rare sources in the very early stage of intermediate - mass ysos .
therefore , further high - resolution and high - sensitivity observations of the continuum emission from centimeter to submillimeter wavelength , in particular with the future interferometers such as evla and alma , and the proper motion measurements of the h@xmath0o masers and continuum sources will be essential to reveal the kinematics and more detailed properties of iras 22198 + 6336 , which will contribute to the understanding of very early stage of intermediate - mass star - formation processes .
we thank the anonymous referee for valuable comments and suggestions .
we also thank guillem anglada and carlos carrasco - gonzalez for providing us with information on the vla observations of iras 22198 + 6336 .
we are grateful to the staff of all the vera stations for their assistance in observations .
th is financially supported by grant - in - aids for scientific research from the ministry of education , culture , sports , science and technology ( 13640242 , 16540224 , and 20740112 ) .
anglada , g. 1995 , rev .
mexicana astron .
conf . , 1 , 67 brand , j. , cesaroni , r. , comoretto , g. , felli , m. , palagi , f. , palla , f. , & valdettaro , r. 2003 , a&a , 407 , 573 chikada , y. , et al .
1991 , in frontiers of vlbi , ed . h.
hirabayashi , m. inoue , & h. kobayashi ( tokyo : universal academy press ) , 79 claussen , m. j. , marvel , k. b. , wootten , a. , & wilking , b. a. 1998 , apj , 507 , l79 crampton , d. & fisher , w. a. 1974 , pub .
, 14 , 283 dame , t. m. , et al .
1987 , apj , 322 , 706 dehnen , w. & binney , j. j. 1998 , mnras , 298 , 387 de zeeuw , p. t. , hoogerwerf , r. , de bruijne , j. h. j. , brown , a. g. a. , blaauw , a. 1999 , aj , 117 , 354 edris , k. a. , fuller , g. a. , & cohen , r. j. 2007 , a&a , 465 , 865 fukui , y. 1989 , in low - mass star formation and pre - main - sequence objects , ed . b. reipurth ( garching : eso ) , 95 furuya , r. s. , kitamura , y. , wootten , h. a. , claussen , m. j. , saito , m. , marvel , k. b. , & kawabe , r. 2000 , apj , 542 , l135 hachisuka , k. et al .
2006 , apj , 645 , 337 hirota , t. et al .
2007 , pasj , 59 , 897 hirota , t. et al .
2008 , pasj , 60 , 37 honma , m. et al .
2003 , pasj , 55 , l57 honma , m. et al .
2007 , pasj , 59 , 889 honma , m. et al . 2008a , submitted to pasj honma , m. , tamura , y. , & reid , m. j. 2008b , pasj , 60 , in press iguchi , s. , kurayama , t. , kawaguchi , n. , kawakami , k. 2005 , pasj , 57 , 259 imai , h. et al .
2007 , pasj , 59 , 1107 jenness , t. , scott , p. f. , & padman , r. 1995 , mnras , 276 , 1024 kawaguchi , n. , sasao , t. , manabe , s. 2000 , in proc .
spie , 4015 , radio telescopes , ed .
h. r. butcher ( washington : spie ) , 544 loinard , l. , mioduszewski , a. j. , rodrguez , l. f. , gonzlez , r. a. , rodrguez , m. i. , torres , r. m. 2005 , apj , 619 , l179 loinard , l. torres , r. m. , mioduszewski , a. j. , rodrguez , l. f. , gonzlez - lpezlira , r. a. , lachaume , r. , vzquez , v. , & gonzlez , e. 2007 , apj , 671 , 546 loinard , l. , torres , r. m. , ; mioduszewski , a. j. , & rodrguez , l. f. 2008 , apj , 675 , l29 lombardi , m. , lada , c. j. , & alves , j. 2008 , a&a , 480 , 785 menten , k. m. , reid , m. j. , forbrich , j. , & brunthaler , a. 2007 , a&a , 474 , 515 migenes , v. et al .
1999 , apjs , 123 , 487 molinari , s. , brand , j. , cesaroni , r. , & palla , f. 1996 , a&a , 308 , 573 moscadelli , l. , testi , l. , furuya , r. s. , goddi , c. , claussen , m. , kitamura , y. , & wootten , a. 2006 , a&a , 446 , 985 palla , f. , brand , j. , cesaroni , r. , comoretto , g. , & felli , m. 1991 , a&a , 246 , 249 patel , n. a. , greenhill , l. j. , herrnstein , j. , zhang , q. , moran , j. m. , ho , p. t. p. , & goldsmith , p. f. 2000 , apj , 538 , 268 perryman , m. a. c. , et al .
1997 , a&a , 323 , l49 petrov , l. , kovalev , y. y. , fomalont , e. , & gordon , d. 2005 , aj , 129 , 1163 reid , m. j. 1993 , ara&a , 31 , 345 reid , m. j. & brunthaler , a. 2004 , apj , 616 , 872 snchez - monge , . ,
palau , a. , estalella , r. , beltrn , m. , & girart , j. m. 2008 , a&a , 485 , 497 sandstrom , k. m. , peek , j. e. g. , bower , g. c. , bolatto , a. d. , & plambeck , r. l. 2007 , apj , 667 , 1161 sato , m. et al .
2007 , pasj , 59 , 743 sato , m. et al .
2008 , pasj , 60 , in press stahler , s. w. & palla , f. 2004 , the formation of stars ( weinheim : wiley - vch ) szymczak , m. , hrynek , g. , & kus , a. j. 2000 , a&as , 143 , 269 tafalla , m. , bachiller , r. , & martn - pintado , j. 1993 , apj , 403 , 175 tofani , g . ,
felli , m . , taylor , g. b. , & hunter , t. r. 1995 , a&as , 112 , 299 torrelles , j. m. et al .
2003 , apj , 598 , l115 torrelles , j. m. et al .
2001 , nature , 411 , 277 torres , r. m. , loinard , l. , mioduszewski , a. j. , & rodrguez , l. f. 2007 , apj , 671 , 1813 ulich , b. l. , haas , r. w. 1976 , apjs , 30 , 247 valdettaro , r. , palla , f. , brand , j. , cesaroni , r. , comoretto , g. , felli , m. , & palagi , f. 2002 , a&a , 383 , 244 wilking , b. a. , blackwell , j. h. , mundy , l. g. , & howe , j. e. 1989 , apj , 345 , 257 wouterloot , j. g. a. & brand , j. 1989 , a&as , 80 , 149 yonekura , y. , dobashi , k. , mizuno , a. , ogawa , h. , & fukui , y. 1997 , apjs , 110 , 21 | we present results of multi - epoch vlbi observations with vera ( vlbi exploration of radio astrometry ) of the 22 ghz h@xmath0o masers associated with a young stellar object ( yso ) iras 22198 + 6336 in a dark cloud l1204 g .
based on the phase - referencing vlbi astrometry , we derive an annual parallax of iras 22198 + 6336 to be [email protected] mas , corresponding to the distance of 764@xmath127 pc from the sun .
although the most principal error source of our astrometry is attributed to the internal structure of the maser spots , we successfully reduce the errors in the derived annual parallax by employing the position measurements for all of the 26 detected maser spots .
based on this result , we reanalyze the spectral energy distribution ( sed ) of iras 22198 + 6336 and find that the bolometric luminosity and total mass of iras 22198 + 6336 are 450@xmath2 and 7@xmath3 , respectively .
these values are consistent with an intermediate - mass yso deeply embedded in the dense dust core , which has been proposed to be an intermediate - mass counterpart of a low - mass class 0 source . in addition , we obtain absolute proper motions of the h@xmath0o masers for the most blue - shifted components .
we propose that the collimated jets aligned along the east - west direction are the most plausible explanation for the origin of the detected maser features . |
two widely studied geometric invariants of a closed connected riemannian manifold @xmath1 are the laplace spectrum and the length spectrum . the _ laplace spectrum _ ( or _ spectrum _ ) is the non - decreasing sequence of eigenvalues , considered with multiplicities , of the laplace operator acting on the space of smooth functions of @xmath2 .
while the spectrum is known to encode some geometric information such as the dimension , volume and total scalar curvature , it is by now well established that the spectrum does not uniquely determine the geometry of a riemannian manifold ( e.g. , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) . in particular , in 1985 sunada devised a method that allows one to construct an abundance of isospectral manifolds by exploiting certain finite group actions @xcite , which were originally studied by gassmann @xcite .
the _ length spectrum _ is the collection of lengths of the smoothly closed geodesics in @xmath1 , where the multiplicity of a length is counted according to the number of free homotopy classes containing a geodesic of that length .
if we ignore multiplicities , the resulting set of non - negative numbers is known as the _ weak _ ( or _ absolute _ ) _ length spectrum_. as with the laplace spectrum , it is known that the length spectrum does not uniquely characterize the geometry of a manifold .
a classical pursuit in geometry , dynamics and mathematical physics is to understand the mutual influences of the laplace and length spectra of a riemannian manifold . by using the poisson summation formula
one can show that any two flat tori are isospectral if and only if they share the same length spectrum .
the work of colin de verdire shows that for a generic riemannian manifold i.e .
, a manifold equipped with a `` bumpy '' metric @xcite the weak length spectrum is determined by its laplace spectrum @xcite .
furthermore , chazarain @xcite and duistermaat and guillemin ( * ? ? ?
1.2 ) demonstrated that for an arbitrary riemannian manifold , the _ weak _ length spectrum contains the singular support of the trace of its wave group ( a spectrally determined tempered distribution ) .
hence , the laplace spectrum always determines a non - trivial subset of the weak length spectrum of a riemannian manifold .
however , the interesting and important question of whether the weak length spectrum is actually a spectral invariant remains open
. in @xcite sormani and wei introduced the covering spectrum : a geometric invariant that is related to the length spectrum of a riemannian manifold @xmath1 , and that `` roughly measures the size of the one dimensional holes in the space '' .
in @xcite the covering spectrum is computed by considering a certain family @xmath3 of regular coverings of @xmath2 , where @xmath4 covers @xmath5 for @xmath6 , and selecting the values of @xmath7 where the isomorphism type of the cover changes .
that is , we look for `` jumps '' in the `` step function '' @xmath8 .
when viewed within this framework the definition actually applies to all complete length spaces , and sormani and wei demonstrated that the covering spectrum is well - behaved under gromov - hausdorff convergence . in this paper
we present a slightly different , yet compatible , definition of the covering spectrum that is applicable to any metric space ( see section [ sec : covspec ] ) . in short
, we form the covering spectrum of @xmath2 by assigning a non - negative real number @xmath9 to _ every _ non - trivial covering @xmath10 of @xmath2 . in the case
that @xmath1 is a compact riemannian manifold this number is half the length of the shortest closed geodesic that has a lift to @xmath10 that is not a closed loop ; see corollary [ cor : geolength ] ( cf .
* lemma 4.9 ) ) .
for example , if @xmath11 is the universal cover of @xmath2 , then @xmath12 is half of the systole of @xmath2 .
the covering spectrum of @xmath1 , denoted @xmath13 , is then defined to be the collection of all @xmath9 as @xmath10 ranges over all non - trivial covers of @xmath2 .
thus , @xmath14 is the portion of the weak length spectrum consisting of those lengths that are `` seen '' by some covering space as the length of the shortest closed geodesic having a non - closed lift . as an example , consider the flat @xmath15 torus @xmath16 , where @xmath17 denotes the circle of circumference @xmath18 .
then one can easily verify that @xmath19 ( * ? ? ?
2.5 ) . in this case
we see that the covering spectrum consists entirely of half the successive minima of the corresponding lattice @xmath20 .
however , as we show in example [ exa : flattori ] this is not the case for all flat tori .
having identified the covering spectrum as a geometrically determined finite part of the weak length spectrum , one may wonder about the mutual influences between the covering spectrum and laplace spectrum . along these lines , sormani and
wei found that so - called komatsu pairs of sunada isospectral manifolds share the same covering spectrum ( * ? ? ?
however , as we will show in section [ sec : isospectral ] , their claim ( * ? ? ?
* example 10.3 ) that certain pairs of isospectral heisenberg manifolds due to gordon have distinct covering spectra is false @xcite , thus keeping alive the question of whether the covering spectrum is a spectral invariant . by engaging in both a geometric and a group theoretic analysis of the covering spectrum and its relation to the sunada condition
, we provide a negative answer to this question .
specifically , we show that in dimension @xmath21 and higher there are sunada isospectral manifolds with distinct covering spectra . in dimension 4
these include certain isospectral flat tori due to conway and sloane . in closing
, we briefly return to the relationship between the weak length spectrum and the laplace spectrum . as we noted above
, the laplace spectrum of a manifold always determines a particular subset of its weak length spectrum , and under certain genericity conditions it is known that the laplace spectrum completely determines the weak length spectrum . however , in general , this relationship is not fully understood . in corollary 8.8 of @xcite ,
sormani and wei make the interesting observation that a continuous family of manifolds sharing the same _ discrete _ weak length spectrum will necessarily share the same covering spectrum .
that is , the covering spectrum is an invariant of a particular flavor of _ iso - length - spectral _ deformation . in light of the fact that we have shown that the covering spectrum is not a spectral invariant it would appear to be of interest to explore the covering spectrum of certain continuous families of _ isospectral _ manifolds in the literature .
in this section , we state our main results and give an overview of the material covered in later sections .
we first recall sunada s construction .
let @xmath22 be a finite group , and let @xmath23 and @xmath24 be subgroups .
suppose that @xmath22 acts on a closed connected manifold @xmath2 , and suppose that the action is free in the sense that every non - trivial element of @xmath22 acts without fixed points .
we then consider the quotient manifolds @xmath25 and @xmath26 , which both cover @xmath27 , as indicated in figure [ fig : sunadadiagram ] , where the labels at regular coverings are their deck transformation groups .
\(n ) at ( 0,4 ) @xmath2 ; ( m ) at ( -2,1.3 ) @xmath25 ; ( mm ) at ( 2,1.3 ) @xmath26 ; ( gm ) at ( 0,0 ) @xmath28 ; ( n )
node[left ] @xmath23 ( m ) ; ( n ) node[right ] @xmath24(mm ) ; ( n ) node[right ] @xmath22(gm ) ; ( m )
( gm ) ; ( mm ) ( gm ) ; [ thm : sunada ] the following are equivalent : 1 . for every conjugacy class @xmath29 of @xmath22 we have @xmath30 ; 2 . for
every @xmath22-invariant riemannian metric on @xmath2 , the quotient riemannian manifolds @xmath25 and @xmath26 have the same laplace spectrum .
the fact that @xmath31 implies @xmath32 is sunada s theorem , and it is the only implication that will be used in this paper ( see @xcite , ( * ? ? ?
3 ) for alternative proofs and comments ) .
many of the examples of isospectral manifolds in the literature can be understood within the framework of sunada s theorem and its generalizations ( e.g. , @xcite , @xcite , @xcite , @xcite , @xcite and @xcite ) .
the group theoretic condition @xmath31 is known by several names in the literature : perlis @xcite says that @xmath23 and @xmath24 are gassmann equivalent after gassmann @xcite , who first used this condition in 1926 , and spectral geometers ( e.g. , @xcite ) frequently say that @xmath23 and @xmath24 satisfy the sunada condition .
others say that @xmath23 and @xmath24 are almost conjugate @xcite , or linearly equivalent @xcite or that @xmath23 and @xmath24 induce the same permutation representation @xcite .
in this paper we will express that condition ( 1 ) holds by saying that @xmath23 and @xmath24 are _ gassmann - sunada _
equivalent , or that @xmath33 is a _ gassmann - sunada triple_. the statement that @xmath32 implies @xmath31 in theorem [ thm : sunada ] follows from the work of pesce @xcite , and it shows that the gassmann - sunada condition is not only sufficient , but also necessary if we want it to ensure that @xmath25 and @xmath26 are isospectral for all possible choices of compatible riemannian metrics in the diagram above . in section [ sec : proof ]
we establish the following analogue of sunada s method for the covering spectrum .
[ thm : covsunada ] let @xmath34 and @xmath2 be as above . if @xmath2 is simply connected and of dimension at least @xmath21 then , the following are equivalent : 1 . for all subsets @xmath35 of @xmath22 that are stable under conjugation we have @xmath36 2 . for every @xmath22-invariant riemannian
metric on @xmath2 , the quotient riemannian manifolds @xmath25 and @xmath26 have the same covering spectrum .
more generally , if @xmath2 is simply connected , then @xmath37 implies @xmath38 , and if @xmath2 has dimension at least @xmath21 , then @xmath38 implies @xmath37 . in the last statement of theorem [ thm : covsunada ]
, the condition that @xmath2 is simply connected can not be omitted . as will become evident after the discussion in section [ sec : proof][explanation ] , this follows from the fact that if @xmath10 is a normal subgroup of @xmath22 that is contained in @xmath39 , then the triple @xmath40 can satisfy condition @xmath37 while the triple @xmath41 need not satisfy this condition ( see remark [ rem : reducedtriples ] for an example ) .
we do not know if @xmath38 implies @xmath37 for all @xmath42-dimensional manifolds . in order to understand the relevance of condition @xmath37 in this result , and to explain the ingredients of the proof
, we introduce the purely algebraic concept of a _ length map _ on a group .
[ dfn : lengthmap ] a _ length map _ on a group @xmath22 is a map @xmath43 from @xmath22 to the set of non - negative real numbers that satisfies 1 .
@xmath44 for all @xmath45 ; 2 .
@xmath46 for all @xmath47 ; 3 .
@xmath48 for all @xmath49 and @xmath50 .
taking @xmath51 and @xmath52 we see that @xmath53 implies that @xmath54 and @xmath55 for all @xmath56 .
the motivation for the preceding definition is found in the following example .
[ exa : minmark ] let @xmath1 be a _ compact _ riemannian manifold , and let @xmath57 be its fundamental group with respect to some base point .
the _ minimum marked length map _
@xmath58 assigns to each @xmath59 the length of the shortest closed geodesic in the free homotopy class determined by @xmath60 .
equivalently , we can set @xmath61 , where @xmath62 is the universal riemannian cover of @xmath1 , @xmath63 is the metric structure induced by @xmath64 and ( after a choice of base point ) the group @xmath65 acts on @xmath66 via deck transformations ( see ( * ? ? ?
we refer the reader to ( * ? ? ?
4.4 ) for the reason why each free homotopy class in the definition above has a shortest closed geodesic . since the conjugacy classes of @xmath57 naturally correspond to the free homotopy classes of loops in @xmath2
, we see that @xmath67 is a length map .
since @xmath2 is compact , the image of @xmath67 is closed and discrete in @xmath68 ( * ? ? ?
[ def : introjumpset ] given a length map @xmath69 on a group @xmath22 , we define for each @xmath70 the subgroup @xmath71 of @xmath22 . when @xmath72 , we have @xmath73 , so this defines an increasing filtration , denoted @xmath74 , of @xmath22 by normal subgroups indexed by the positive real numbers .
we define its _ jump set _
@xmath75 to be the set of all @xmath70 such that for all @xmath76 we have @xmath77 .
we will see that this notion of the jump set of the filtration defined by a length map provides the connection between conditions @xmath37 and @xmath38 of theorem [ thm : covsunada ] . more specifically , in section [ sec : covfund ] we will show the following .
[ prop : covjump ] the covering spectrum of a compact riemannian manifold @xmath1 is given by @xmath78 in lemma [ lemma : jumptriple ] we will see that condition @xmath37 of theorem [ thm : covsunada ] holds if and only if for every length map @xmath69 on @xmath22 the restrictions to @xmath23 and to @xmath24 have identical jump sets .
when @xmath37 holds , we will say that @xmath23 and @xmath24 are _ jump equivalent _ or that @xmath33 is a _
jump triple_. combining lemma [ lemma : jumptriple ] with proposition [ prop : covjump ] it is not hard to show that @xmath37 implies @xmath38 for simply connected @xmath2 . in order to show the other , more difficult , implication of theorem [ thm : covsunada ] , we will assume that @xmath37 does not hold and then proceed to construct a riemannian metric on @xmath2 so that the resulting minimum marked length map has a special property .
in order to construct this metric it will be necessary to understand the extent to which length maps actually arise from minimum marked length maps of riemannian manifolds .
that is , we would like to know which length maps on the fundamental group of a given manifold @xmath2 arise as the minimum marked length map associated to some riemannian metric on @xmath2 . in section [ sec : prescribing ]
we take up this line of inquiry and we establish theorem [ thm : systolelengthmap ] , which can be summarized as follows . [ thm : introsystole ] let @xmath2 be a closed manifold of dimension at least three and let @xmath79 be a finite set . if @xmath80 is the restriction to @xmath81 of a length map on @xmath65 , then there exists a riemannian metric @xmath82 on @xmath2 such that @xmath83 for any @xmath84 .
that is , @xmath69 can be extended to the minimum marked length map associated to some metric @xmath82 on @xmath2 .
hence , in the case where @xmath2 is a closed manifold with finite fundamental group and dimension at least @xmath21 , the above demonstrates that there is no difference between length maps on @xmath57 and the minimum marked length maps associated to @xmath2 . in theorem [ thm : systolelengthmap]the detailed statement of the above
we actually prove that we have some control over the extension of @xmath69 .
more specifically , we demonstrate that we are able to prescribe which free homotopy classes contain the shortest closed geodesics ( see remark [ rem : systolelengthmap ] ) . in section [ sec :
proof ] , we will use this to prove theorem [ thm : covsunada ] .
theorem [ thm : systolelengthmap ] also provides a way to characterize the initial segments of the so - called minimum marked length spectrum of a riemannian metric on a closed manifold of dimension at least three .
since this is perhaps of independent interest we will formulate this after the following definition .
[ def : minlspec ] let @xmath1 be a riemannian manifold with associated minimum marked length map @xmath85 .
the value @xmath86)$ ] depends only on the class of the loop @xmath87 in the set of unoriented free homotopy classes @xmath88 of loops in @xmath2 ; whereby the _ unoriented free homotopy class _ of @xmath89
we mean the collection of all loops freely homotopic to @xmath89 or its inverse @xmath90 .
hence , we obtain an induced map @xmath91 , which is again denoted by @xmath67 , and in this incarnation it is known as the _ minimum marked length spectrum_. alternatively , one can define the minimum marked length spectrum as the set of ordered pairs @xmath92 where each length @xmath93 is `` marked '' by the unoriented free homotopy class @xmath94 .
this set of pairs is a subset of the marked length spectrum as defined in @xcite , for example .
[ thm : varyg ] suppose that @xmath2 is a closed connected manifold of dimension at least three .
let @xmath95 be a sequence of distinct elements of @xmath88 where the first element @xmath96 is trivial .
then for every sequence @xmath97 of real numbers the following are equivalent : 1 .
the sequence @xmath98 is @xmath99-admissible ( see definition [ dfn : admissible ] and example [ exa : admissible ] ) ; 2 .
there is a riemannian metric @xmath82 on @xmath2 such that the minimum marked length map @xmath67 satisfies @xmath100 for all @xmath101 and @xmath102 for all @xmath103 .
in particular , there is a metric @xmath82 on @xmath2 such that the systole is achieved in the unoriented free homotopy class @xmath104 .
the statement that @xmath105 implies @xmath106 in the above is a minimum - marked - length - spectrum analog of a classical theorem due to colin de verdire @xcite , which states that given a connected closed manifold @xmath2 of dimension at least @xmath21 and a finite sequence @xmath107 there is a riemannian metric @xmath82 on @xmath2 such that the sequence gives exactly the first @xmath0 eigenvalues , counting multiplicities , of the associated laplacian .
however , unlike colin de verdire s result , the sequence @xmath108 in the above can not be chosen arbitrarily : it will depend on our choice of the sequence @xmath109 in @xmath88 .
the above theorem then tells us that given such a choice for @xmath99 , the @xmath99-admissibility of @xmath108 is a necessary and sufficient condition for the existence of a metric @xmath82 such that the @xmath101th smallest value of the minimum marked length spectrum is @xmath110 for @xmath111 . with theorem [ thm : systolelengthmap ] in place , we will complete the proof of theorem [ thm : covsunada ] in section [ sec : proof ] . in section [ sec : groups ] we present some group theoretic context and results concerning gassmann - sunada triples ( condition @xmath31 of theorem [ thm : sunada ] ) and jump triples ( condition @xmath37 of theorem [ thm : covsunada ] ) .
there are some key differences in the behavior of jump triples and gassmann - sunada triples .
for instance , we have @xmath112=[g : h']$ ] for every gassmann - sunada triple @xmath33 , but not for every jump triple . the behavior with respect to dividing out normal subgroups
is also different : if @xmath10 is a normal subgroup of @xmath22 that is contained in @xmath113 , then @xmath23 and @xmath24 are gassmann - sunada equivalent in @xmath22 if and only if @xmath114 and @xmath115 are gassmann - sunada equivalent in @xmath116 . however , as we noted earlier , @xmath114 and @xmath115 can be jump equivalent in @xmath116 even when @xmath23 and @xmath24 are not jump equivalent in @xmath22 ( see remark [ rem : reducedtriples ] ) .
while most small examples of gassmann - sunada triples turn out to be jump triples ( see example [ exa : small ] ) we will show how to adapt some well - known constructions of gassmann - sunada triples to obtain gassmann - sunada triples that are not jump triples .
finally , our analysis leads us to the following conclusion in section [ sec : isospectral ] .
[ thm : distinct ] for each @xmath117 there are closed riemannian manifolds of dimension @xmath118 with identical laplace spectra and distinct covering spectra .
as we will see in example [ exa : tori ] , these will include certain isospectral flat tori of dimension @xmath119 due to conway and sloane @xcite . in a forthcoming paper
we employ different techniques to construct isospectral surfaces with distinct covering spectra @xcite .
we conclude the paper with a study of the covering spectrum of heisenberg manifolds that refutes ( * ? ? ?
* example 10.3 ) .
the notion of the covering spectrum was defined by sormani - wei @xcite for complete length spaces . in this section
we provide an alternate definition that works for any metric space . for compact length spaces
the main focus of this paper our definition coincides with the definition given by sormani and wei , and for non - compact complete length spaces the only difference is that @xmath120 is sometimes an element of our covering spectrum . in order to present this definition ,
we first review some terminology from covering space theory . let @xmath121 be a topological space . by a _ space over _ @xmath121 we mean a topological space @xmath122 along with a continuous map @xmath123 .
for such a space @xmath122 over @xmath121 and @xmath124 we say that @xmath122 is _ trivial over _ @xmath125 , or it _ evenly covers _
@xmath125 , if @xmath126 is a disjoint union of open subspaces of @xmath126 that are each mapped homeomorphically onto @xmath125 via the map @xmath127 .
a _ covering space _ of @xmath121 is a space @xmath122 over @xmath121 that is locally trivial ; that is , each point @xmath128 has an open neighborhood @xmath129 so that @xmath122 is trivial over @xmath129 .
we say that a covering space @xmath122 of @xmath121 is _ trivial _ if it is trivial over @xmath121 .
[ dfn : covrad ] let @xmath130 be a metric space and let @xmath131 be a covering space of @xmath121 .
for @xmath132 we say that @xmath122 is _ @xmath7-trivial _ over @xmath121 if for each @xmath128 the cover @xmath122 is trivial over the open ball of radius @xmath7 centered at @xmath133 .
the _ covering radius _
@xmath134 of @xmath122 over @xmath121 is given by @xmath135 clearly , we have @xmath136 for any trivial cover @xmath137 , and @xmath138 for any non - trivial cover @xmath137 .
if @xmath121 is a compact metric space , then the lebesgue number lemma tells us that the covering radius of any cover will be non - zero .
however , in the next example we see that non - compact metric spaces can admit covers that have zero covering radius .
[ exa : zeroradius ] let @xmath121 be the metric _ subspace _ of the euclidean plane equipped with the standard metric that one gets by connecting a sequence of circles with radius going to zero as in figure [ fig : zerocovrad ] . then for any non - trivial covering space @xmath122 of @xmath121 the covering radius @xmath139 is the infimum of the radii of those circles that are not evenly covered .
this space @xmath121 has a simply connected universal cover , which therefore has covering radius zero over @xmath121 .
( 0,0 ) grid ( 4.5,0.999 ) ; ( 0,1 ) circle ( 0.5 cm ) ; ( 1,1 ) circle ( 0.3 cm ) ; ( 2,1 ) circle ( 0.18 cm ) ; ( 3,1 ) circle ( 0.108 cm ) ; ( 4,1 ) circle ( 0.0648 cm ) ; ( 4.7,0.2 ) circle ( 0.02 cm ) ; ( 4.9,0.2 ) circle ( 0.02 cm ) ; ( 5.1,0.2 ) circle ( 0.02 cm ) ; in lemma [ lem : lpc ] below we show that if @xmath130 is a locally path connected space , then @xmath140 is actually an @xmath139-trivial cover .
in general this may fail ; see example [ exa : infrad ] below .
in particular , we can have @xmath141 without @xmath142 being @xmath143-trivial .
[ exa : infrad ] let us define a metric subspace @xmath121 of the euclidean plane that is locally path connected at all but a single point , together with a non - trivial double cover that is trivial over every bounded subset of @xmath121 .
as is shown in figure [ fig : infinitecovrad ] , we take a union of infinitely many circles , all tangent to each other at the same point @xmath133 , with the radius of the circles going to infinity , where in each circle we omit an open interval lying around the point @xmath133 whose size is shrinking to zero as we go through the infinite sequence of circles .
thus , what remains of each circle is a closed interval , and the end points of these intervals give a sequence converging to @xmath133 .
then we also add the single point @xmath133 to our space .
informally one might say that @xmath121 has a loop that only closes on an infinite scale .
now take any point @xmath144 inside the circles that is not in @xmath121 .
then the plane with @xmath144 removed has a connected double cover , and we restrict this cover to @xmath121 .
we leave it to the reader to check that this covering has the stated properties .
( -6.1,-0.3 ) rectangle ( 6.1,3 ) ; ( 0,2 cm ) arc ( 90:0:1 cm ) ; ( 0,2 cm ) arc ( 90:180:1 cm ) ; ( 0,2.8 cm ) arc ( 90:-40:1.4 cm ) ; ( 0,2.8 cm ) arc ( 90:220:1.4 cm ) ; ( 0,4 cm ) arc ( 90:-62:2 cm ) ; ( 0,4 cm ) arc ( 90:242:2 cm ) ; ( 0,5.6 cm ) arc ( 90:-74:2.8 cm ) ; ( 0,5.6 cm ) arc ( 90:254:2.8 cm ) ; ( 0,8 cm ) arc ( 90:-82:4 cm ) ; ( 0,8 cm ) arc ( 90:262:4 cm ) ; ( 0,11.2 cm ) arc ( 90:-86:5.6 cm ) ; ( 0,11.2 cm ) arc ( 90:266:5.6 cm ) ; ( 0:1 cm ) circle ( 0.8pt ) ; ( 180:1 cm ) circle ( 0.8pt ) ; ( -40:1.4 cm ) circle ( 0.8pt ) ; ( 220:1.4 cm ) circle ( 0.8pt ) ; ( -62:2 cm ) circle ( 0.8pt ) ; ( 242:2 cm ) circle ( 0.8pt ) ; ( -74:2.8 cm ) circle ( 0.8pt ) ; ( 254:2.8 cm ) circle ( 0.8pt ) ; ( -82:4 cm ) circle ( 0.8pt ) ; ( 262:4 cm ) circle ( 0.8pt ) ; ( -86:5.6 cm ) circle ( 0.8pt ) ; ( 266:5.6 cm ) circle ( 0.8pt ) ; ( 0,1 ) circle ( 0.8pt ) ; ( 0,0 ) circle ( 0.8pt ) ; at ( 0,1 ) @xmath144 ; at ( 0,0 ) @xmath133 ; one can make this space into a path connected example by connecting a single end point of each component to the point @xmath133 .
one can also make a similar bounded example : a space within the open disk of radius 1 in the plane , with a double cover that is @xmath7-trivial for all @xmath145 but not @xmath146-trivial .
[ rem : nonconstant ] given a covering @xmath123 the degree of a point @xmath128 is the cardinality of the fiber @xmath147 over @xmath133 , which may be infinite .
if this cardinality is the same for all @xmath128 , then we say that the covering has constant degree .
since @xmath122 is locally trivial over @xmath121 this degree is locally constant on @xmath121 , so it is constant on the connected components of @xmath121 .
when considering the covering radius we will often restrict to the case that the covering has constant degree . to see how to express the covering radius in the general case , note that for any covering @xmath148 that is not of constant degree the fibers of the degree map partitions @xmath121 into at least two open subsets @xmath149 , where @xmath101 ranges over a suitable index set @xmath150 , and for each @xmath101 the space @xmath151 is a cover of @xmath149 of constant degree .
then the covering radius of @xmath122 over @xmath121 is the infimum of the following set : @xmath152 for example , if @xmath121 consists of two points and @xmath122 is a non - trival covering space of @xmath121 then @xmath139 is the distance between the two points of @xmath121 .
the _ covering spectrum _ , denoted @xmath153 , of a metric space @xmath121 is the set of all @xmath139 as @xmath122 varies over all non - trivial coverings of @xmath121 of constant degree .
we will show in the next section that for connected locally path connected spaces the covering spectrum is the jump set of a length map on the fundamental group .
we will also see that for compact riemannian manifolds , the prime object of interest in this paper , the definition above is equivalent to the definition given by sormani - wei @xcite ; see remark [ rem : swdef ] .
we conclude this section with some general remarks .
the covering spectrum of @xmath121 is a subset of @xmath154 .
it is empty if and only if all constant degree covers of @xmath121 are trivial . in example
[ exa : zeroradius ] the covering spectrum is the set of diameters of the circles together with the element @xmath120 . if we do not use the euclidean distance , but instead view @xmath121 as a length space by taking the path - length metric , then the covering spectrum consists of half the circumferences of the circles together with the element @xmath120 .
in example [ exa : infrad ] the covering spectrum is @xmath155 and that of its bounded analog is @xmath156 .
the same is true for their path connected versions .
these topological spaces can not be made into length spaces , so this is a situation in which the definition of the covering spectrum as given in @xcite does not apply .
if @xmath121 is connected then one may drop the condition `` of constant degree '' from the definition .
discrete metric spaces have an empty covering spectrum , and more generally , the covering spectrum of a union of disjoint open subspaces of a metric space is the union of their covering spectra .
in this section we see how to express the covering spectrum in terms of additional structure on the fundamental group .
the main result of this section is a proof of proposition [ prop : covjump ] .
first , we recall some terminology .
a _ path _ in a topological space @xmath121 is a continuous map @xmath157\to x$ ] .
we say that @xmath89 is a _ loop _
if @xmath158 , and we then say that the loop is _ based at _
@xmath159 . if @xmath123 is a covering map and @xmath89 is a path in @xmath121 , then a _ lift _ of @xmath89 to @xmath122 is a path @xmath160 in @xmath122 such that @xmath161 .
the lemma of _ unique path lifting _ states that the map @xmath162 gives a bijection between the set of lifts of @xmath89 to @xmath122 and the fiber @xmath163 over the starting point of @xmath89 . given a path @xmath89 its _ inverse _ @xmath164 is the path given by @xmath165 , and if @xmath166 is also a path in @xmath121 and @xmath167 , then @xmath168 is the _ composed path _ sending @xmath169 to @xmath170 if @xmath171 $ ] and to @xmath172 when @xmath173 $ ] ; that is , we travel along @xmath89 followed by @xmath166 . when @xmath174 is a path in @xmath121 with @xmath175 , we recall that @xmath176 is path homotopic to @xmath177 ; hence , there is no ambiguity in writing @xmath178 $ ] for the path homotopy class of both paths .
if @xmath89 is a loop in @xmath121 based at @xmath133 , then we denote its homotopy class in the fundamental group @xmath179 by @xmath180 $ ] .
the group operation in @xmath179 is given by @xmath180[\tau]=[\sigma*\tau]$ ] and the inverse by @xmath180^{-1}=[{\overline \sigma}]$ ] . by a _
filtration _ @xmath181 of a group @xmath22 we mean a family of subgroups @xmath182 of @xmath22 , with @xmath101 ranging over an ordered index set @xmath150 , such that @xmath183 when @xmath184 .
the _ jump set _ of the filtration is the subset of @xmath150 given by @xmath185 that is , @xmath186 if and only if @xmath187 is a _ proper subset _ of @xmath188 for all @xmath189 ( cf . definition [ def : introjumpset ] and example [ exa : flattori ] ) . in this paper
our index set @xmath150 will always be the set of positive real numbers ( with the usual ordering ) and the subgroups of the filtration will always be normal subgroups of @xmath22 . for a metric space @xmath130 with base point @xmath190 , we now define a filtration @xmath191 on the fundamental group @xmath192 of @xmath121 .
this filtration will determine the covering radius of any covering of @xmath121 and hence encodes the covering spectrum .
for @xmath70 we let @xmath193 be the subgroup of @xmath192 generated by the elements of the form @xmath194 $ ] , where @xmath87 is a loop contained inside some open @xmath7-ball and @xmath195 is a path from @xmath190 to @xmath196 .
these generators are precisely the homotopy classes of loops based at @xmath190 that are _ freely _ homotopic to a loop completely contained in some open @xmath7-ball .
we note that in ( * ? ? ?
* section 2 ) the same filtration is considered for complete length spaces , where @xmath193 is denoted by @xmath197 .
the next lemma expresses the covering radius ( see definition [ dfn : covrad ] ) in terms of this filtration .
[ lem : lpc ] let @xmath130 be a connected and locally path connected metric space , with base point @xmath190 , and let @xmath122 be a non - trivial covering space of @xmath121
. then @xmath122 is @xmath139-trivial over @xmath121 .
moreover , @xmath139 is the maximum of all @xmath70 such that for every loop @xmath89 in @xmath121 based at @xmath190 with @xmath180\in { \operatorname{fil}}^\delta\pi_1(x , x_0)$ ] , every lift to @xmath122 of @xmath89 is a closed loop .
recall first that for each open subset @xmath198 the cover @xmath122 is trivial over @xmath129 if and only if every lift to @xmath122 of a loop in @xmath129 is a loop in @xmath122 ; ( * ? ? ?
* lemma 2.4.9 ) . to show the first statement , let @xmath199 be an open ball of radius @xmath139 in @xmath121 centered at some point @xmath200 , and
let @xmath89 be a loop whose image is completely contained inside @xmath199 .
then since the image of @xmath89 is compact we see that there is a @xmath201 such that the image of @xmath89 is completely contained inside the open ball of radius @xmath7 centered at @xmath133 . by the definition of the covering radius @xmath139 , the covering @xmath142 is @xmath7-trivial , hence we see that @xmath89 lifts to a closed loop .
it now follows from our initial remark that @xmath122 is an @xmath139-trivial cover of @xmath121 .
next , we note that the connected components of @xmath122 are open , and they are themselves coverings of @xmath121 . for
each component of @xmath122 one may choose a point @xmath202 in the fiber of this component over @xmath190 . with ( * ? ? ?
* lemma 2.5.11 , 2.4.9 ) one then shows that this component is @xmath7-trivial over @xmath121 if and only if @xmath203 is contained in the image of @xmath204 under the map induced by the covering map .
thus , the second statement follows .
it follows from the lemma that the covering spectrum of a locally path connected space does not contain the element @xmath205 ( cf .
example [ exa : infrad ] ) .
[ prop : covspecjset ] if @xmath130 is a connected and locally path connected metric space , then @xmath206 suppose that @xmath207
. then @xmath208 by lemma [ lem : lpc ] , and by definition there is a covering @xmath209 with @xmath210 .
again by the lemma , the collection of homotopy classes of loops based at @xmath190 all of whose lifts to @xmath122 are themselves loops contains @xmath193 , but it does not contain the set @xmath211 for any @xmath212 .
it follows that @xmath7 is a jump for the filtration , so the inclusion `` @xmath213 '' holds .
now , let @xmath214 .
following ( * ? ? ?
2.3 ) we see with ( * ? ? ?
* theorem 2.5.13 ) that there is a connected and locally path connected covering space @xmath215 such that @xmath216 .
hence , @xmath193 consists of exactly those classes of loops whose lifts to @xmath217 are all loops . with the lemma , or by using (
* lemma 2.5.11 , 2.4.9 ) , we see that for all @xmath218 , we have @xmath219 and `` @xmath220 '' follows .
[ rem : swdef ] the notion of the covering spectrum was first defined by sormani and wei in @xcite for so - called complete length spaces , and the proposition above shows that their definition exactly gives the set @xmath221 .
the only difference with our notion of covering spectrum for complete length spaces is that for non - compact @xmath121 we sometimes have @xmath120 in the covering spectrum such as for the space in example [ exa : zeroradius ] ( endowed with shortest path length metric ) .
we note that under our definition , @xmath222 if and only if there is a covering space @xmath142 such that @xmath121 has open subsets of arbitrary small diameter that are not evenly covered . to conclude this section
we will make the filtration of the fundamental group more explicit for the case of compact riemannian manifolds .
we will identify generators for the normal subgroups in the filtration : the classes of short closed geodesics .
in later sections , this will allow us to compute the covering spectra of some familiar classes of manifolds .
in particular , in example [ exa : heis ] we will see how to compute the covering spectrum of a heisenberg manifold .
let @xmath2 be a manifold and let @xmath82 be a riemannian metric on @xmath2 .
choose a base point @xmath223 .
in section [ sec : results ] we defined the map @xmath224 by sending a class to the infimum of all lengths of loops that are freely homotopic to it .
[ prop : lengthmap ] for every connected manifold @xmath2 with riemannian metric @xmath82 and base point @xmath223 and every @xmath214 we have @xmath225 suppose that @xmath87 is a loop based at @xmath190 in @xmath2 so that @xmath86)<2\delta$ ] .
then @xmath87 is freely homotopic to a loop @xmath166 of length below @xmath226 .
it is clear that @xmath227 is contained in the open ball @xmath228 of radius @xmath7 centered at @xmath229 . if @xmath230 is the path from @xmath190 to @xmath229 , given by the moving base point during such a homotopy , then @xmath231=[{\alpha } * \tau * \overline { \alpha}]\in{\operatorname{fil}}^\delta \pi_1(m , x_0)$ ] .
this shows `` @xmath220 '' .
for the other inclusion , recall that @xmath232 is generated by elements of the form @xmath233 $ ] where @xmath230 is a path from @xmath190 to a point @xmath234 and @xmath89 is a loop based at @xmath235 that is contained in @xmath236 for some @xmath237 . since @xmath236 is locally simply connected we can assume that @xmath89 has finite length .
the image of @xmath89 is compact , so it is in fact contained in @xmath238 for some @xmath239 .
choose any path @xmath240 from @xmath235 to @xmath241 .
we can now break @xmath89 into finitely many sections of length less than @xmath242 , and by connecting the division points with @xmath241 by paths of length smaller than @xmath243 we may write @xmath244\in\pi_1(m , x_2)$ ] as a finite product @xmath245 \cdots [ { \sigma}_n]$ ] with each @xmath246 a loop based at @xmath241 of length less than @xmath247 . writing @xmath248 we now have @xmath249 = [ \beta * { \sigma}_1 * \overline\beta ] [ \beta * { \sigma}_2 * \overline\beta ] \cdots [ \beta * { \sigma}_n * \overline\beta]\ ] ] in @xmath250 , and since we have @xmath251)<2\delta$ ] for each @xmath101 , our other inclusion follows .
[ cor : geolength ] let @xmath1 be a compact riemannian manifold and let @xmath252 be a non - trivial covering space .
then the covering radius @xmath9 is half the length of the shortest closed geodesic @xmath89 in @xmath2 that has a lift to @xmath10 that is not a closed loop . by ( * ? ? ?
* lemma 2.4.9 ) and the fact that the image of @xmath67 is discrete ( see example [ exa : minmark ] ) , there is a shortest closed geodesic in @xmath2 that has a lift to @xmath122 that is not a closed loop .
let @xmath89 be such a geodesic and let @xmath253)$ ] be its length . for every @xmath212 the image of @xmath89 lies within an open ball of radius @xmath254 ,
so the covering @xmath127 is not @xmath254-trivial , and @xmath255 . on the other hand , for any loop in @xmath2 shorter than @xmath7
all lifts to @xmath10 are closed loops .
proposition [ prop : lengthmap ] then tells us that all loops in @xmath2 whose homotopy class in @xmath256 have only closed loops as lifts to @xmath10 , and with lemma [ lem : lpc ] this gives @xmath257 .
this is an immediate consequence of propositions [ prop : covspecjset ] and [ prop : lengthmap ] and the definitions of @xmath258 and @xmath259 .
the moral of proposition [ prop : covjump ] is that it allows us to compute the covering spectrum of a riemannian manifold in an easy way as the jump set of the filtration associated to a length map . under the hypotheses of the proposition ,
the length map has a closed discrete image in @xmath68 , and the group it is defined on , the fundamental group , is finitely generated .
we now describe how to compute the jump set in this scenario .
[ jsa ] let @xmath22 be a group that is _ finitely generated _ and let @xmath260 be a length map with _ closed discrete _ image . then @xmath261 can be computed as follows .
let @xmath262 where we continue until @xmath263 .
the process terminates because @xmath22 is finitely generated : if @xmath2 is the largest value attained by @xmath69 on some finite set of generators of @xmath22 , then @xmath69 assumes only finitely many values up to @xmath2 ( since the image of @xmath69 is closed and discrete ) , and the @xmath264 are among them .
then we have @xmath265 in particular we recover the result that the covering spectrum of a compact manifold is always finite ( * ? ? ? * theorem 3.4 ) . as an application
we show in the following example how to compute the covering spectrum of a flat torus .
[ exa : flattori ] let @xmath266 be a euclidean space , i.e. , a vector space over @xmath267 of finite dimension @xmath118 with an inner product @xmath268 .
let @xmath269 be a lattice in @xmath266 of full rank , and let @xmath270 be the flat torus @xmath271 , where the riemannian metric is induced by @xmath268 .
then we can identify the fundamental group of @xmath270 with @xmath269 , and the minimum marked length map @xmath69 on @xmath269 is given by @xmath272 .
we then have @xmath273 , where for @xmath214 we have @xmath274 . those jumps of the filtration where the rank of the sublattice increases , are the so - called _ successive minima _ of the lattice .
counting these jumps with a _ multiplicity _ , which is the increase in rank of the sublattice at the jump , we see that there are @xmath118 successive minima .
however , the covering spectrum of @xmath270 can have more than @xmath118 elements : there can also be jumps where the rank does not increase .
for instance , when @xmath266 is @xmath275-dimensional consider a lattice @xmath269 in @xmath266 spanned by orthogonal vectors @xmath276 where @xmath277 . with the procedure
above we see that the covering spectrum of the flat torus @xmath278 is given by @xmath279 .
now let @xmath280 and consider the lattice @xmath281 .
the lattice @xmath269 is a sublattice of @xmath282 of index @xmath42 and any vector in @xmath282 that is not in @xmath269 is of the form @xmath283 , where @xmath284 , and its length is at least @xmath285 .
it follows that @xmath286 .
we recall from the introduction that the covering spectrum of @xmath287 is also measuring when a particular family @xmath288 of regular covers of @xmath2 changes isomorphism type .
these coverings , as defined in ( * ? ? ?
3.1 ) , can be given by @xmath289 .
the above example highlights the difference between isomorphism type as coverings and as topological spaces . indeed , if @xmath290 , then @xmath291 , and if @xmath292 , then @xmath293 . these are both @xmath275-tori , but they are distinct regular covers of @xmath2 .
we revisit the covering spectra of flat tori in section [ sec : isospectral ] .
the purpose of this section is to give a proof of theorems [ thm : introsystole ] and [ thm : varyg ] .
as we noted in example [ exa : minmark ] , every riemannian metric @xmath82 on a closed manifold @xmath2 gives rise to a length map @xmath294 , known as the minimum marked length map , sending the homotopy class of a loop to the infimum of the lengths of the loops that are freely homotopic to it .
the general problem underlying this section is to understand the nature of those length maps that arise as the minimum marked length map associated to some riemannian metric .
do these length maps satisfy additional properties not yet implied by the length map axioms of definition [ dfn : lengthmap ] ?
the next result , whose proof will take up most of this section , implies that this is not the case if we restrict our attention to properties involving only finitely many elements .
in addition to helping us establish theorems [ thm : introsystole ] and [ thm : varyg ] , the following result is also instrumental in the proof of theorem [ thm : covsunada ] as we will see in section [ sec : proof ] .
[ thm : systolelengthmap ] let @xmath2 be a closed and connected manifold of dimension at least @xmath21 .
let @xmath81 be a finite subset of @xmath65 and suppose that @xmath295 is a map satisfying : 1 .
@xmath296 for all non - trivial @xmath297 ; 2 .
@xmath298 for all @xmath299 and @xmath300 such that @xmath133 is conjugate to @xmath301 .
then there is a riemannian metric @xmath82 on @xmath2 such that @xmath302 for all @xmath303 .
in addition , for every @xmath304 the metric @xmath82 may be chosen so that @xmath67 satisfies : 1 . for all @xmath305
for which the set @xmath306 is non - empty with infimum @xmath307 we have @xmath308 ; 2 . for all other @xmath305
we have @xmath309 .
[ rem : systolelengthmap ] by definition [ dfn : lengthmap ] conditions @xmath310 and @xmath311 are clearly necessary for the existence of an extension of @xmath69 to a length map on @xmath65 .
the theorem above implies that they are also sufficient , and that this extension can be taken to be @xmath67 for a suitable metric @xmath82 .
conditions @xmath31 and @xmath32 describe the degree of control we have over the behavior of @xmath67 outside @xmath81 .
in particular , this control allows us to ensure that in @xmath1 the unoriented free homotopy classes ( see definition [ def : minlspec ] ) containing geodesics of length less than @xmath199 can only be found _ among _ those classes corresponding to elements of the form @xmath301 for some @xmath312 and @xmath313 : condition @xmath31 determines the collection of such `` short '' classes precisely .
we will use this in the proof of theorem [ thm : varyg ] .
let @xmath2 be a manifold and let @xmath314 denote the space of all riemmannian metrics on @xmath2 .
given two metrics @xmath315 we will say that @xmath82 is _ larger than _ @xmath316 , denoted @xmath317 , if for any tangent vector @xmath318 we have @xmath319 .
clearly , @xmath320 defines a partial order on @xmath314 .
the set @xmath314 is closed under addition , and under multiplication by everywhere positive smooth functions on @xmath2 . first , we note that by increasing @xmath199 we can reduce to the case that @xmath321 for all @xmath322 .
let the manifold @xmath323 be the standard circle . since the dimension of @xmath2 is at least @xmath21 , we can represent the free homotopy classes of the elements @xmath324 by smooth embeddings @xmath325 with pairwise disjoint images . the tubular neighborhood theorem ( * ? ? ?
5.2 , ch . 4 )
says that for each @xmath322 there is a smooth vector bundle @xmath326 over @xmath327 together with a diffeomorphism @xmath328 onto an open subset @xmath329 of @xmath2 , whose composition with the zero section @xmath330 is the map @xmath331 .
the loops @xmath331 have positive distance between each other , so by shrinking the neighborhoods , if necessary , we can also assume that the tubular neighborhoods @xmath332 are disjoint .
let us consider the structure of such a tubular neighborhood @xmath329 .
if @xmath118 is the dimension of @xmath2 , then the bundle @xmath326 on @xmath327 has rank @xmath333 .
there are up to isomorphism exactly two vector bundles of rank @xmath334 over @xmath327 : one orientable and one non - orientable ( * ? ? ?
* ch . 4 , sec . 4 , ex .
2 ) , ( * ? ? ?
5 ) . in view of choosing metrics
later , let @xmath335 be the standard open unit ball around @xmath120 in @xmath336 , which is diffeomorphic to each fiber of the vector bundle @xmath326 over @xmath327 .
now consider the quotient @xmath337 , where @xmath338 acts by sending @xmath339 to @xmath340 if @xmath326 is orientable , and to @xmath341 if @xmath326 is non - orientable .
then it follows that there is a commutative diagram : ; ( t ) edge [ ->,bend right ] node [ above left]@xmath342(b1 ) edge [ ->,bend left ] node [ above right]@xmath331(b3 ) edge [ - > ] node [ left]@xmath120-section(b2 ) ( b1 ) edge [ - > ] node [ above ] @xmath343 ( b2 ) ( b2 ) edge [ - > ] node [ above ] @xmath343 ( b3 ) edge [ - > ] node [ below ] @xmath344 ( b3 ) ; where the horizontal maps are diffeomorphisms . in both the orientable and non - orientable case , the diffeomorphisms give rise to the following additional structure on @xmath329 .
first , the standard riemannian metric on @xmath337 gives a riemannian metric @xmath345 on @xmath329 that has the property that for every @xmath49 each loop within @xmath329 homotopic to @xmath346 has length at least @xmath347 , with equality for the loop @xmath346 itself .
second , we obtain a smooth map @xmath348 by sending the point associated to @xmath349 to the length of the vector @xmath350 .
note that @xmath351 is @xmath120 on @xmath352 .
for each @xmath322 we will consider five open neighborhoods @xmath353 on @xmath352 for @xmath354 , where @xmath355 .
we write @xmath356 and @xmath357 , and we let @xmath358 be the smooth map given by @xmath351 on each @xmath329 .
let @xmath359 be the metric on @xmath270 whose restriction to the tubular neighborhood @xmath329 is given by @xmath360 . then with respect to the metric @xmath359 , the loop @xmath361 has length @xmath362 , for each @xmath84 .
now , let @xmath363 be a constant such that with respect to @xmath364 the distance between @xmath365 and @xmath366 is at least @xmath199 .
since @xmath2 is compact , we may fix a riemannian metric @xmath367 on @xmath2 such that every non - contractible loop in @xmath2 has length at least @xmath199 ; that is , we may choose @xmath367 so that its systole is at least @xmath199 . with the notation and assumptions as above , there is a riemannian metric @xmath82 on @xmath2 that satisfies the following properties : ( 0,0 ) circle ( 5 cm ) ; ( 0,0 ) circle ( 4 cm ) ; ( 0,0 ) circle ( 3 cm ) ; ( 0,0 ) grid ( 10,10 ) ; ( 0,0 ) circle ( 3 cm ) ; ( 0,0 ) circle ( 2 cm ) ; ( 0,0 ) circle ( 1 cm ) ; ( 0,0 ) grid ( 10,10 ) ; ( 0,0 ) circle ( 1 cm ) ; ( 70:2 cm ) ( 70:5.5 cm ) ; ( a ) at ( 70:4.5 cm -| 6.6,0 ) @xmath31 @xmath368 on @xmath369 ; ; ( 70:4.5 cm ) ( a ) ; ( 41:4 cm ) ( 41:5.5 cm ) ; ( a ) at ( 41:4.5 cm -| 6.6,0 ) @xmath32 @xmath370 on @xmath371 ; ; ( 41:4.5 cm )
( a ) ; ( 41:2 cm )
( 41:3 cm ) ; ( b ) at ( 41:2.5 cm -| 6.6,0 ) @xmath37 @xmath372 on @xmath373 ; ; ( 41:2.5 cm ) ( b ) ; ( 41:0 cm ) ( 41:1 cm ) ; ( d ) at ( 41:0.5 cm -| 6.6,0 ) @xmath38 @xmath374 on @xmath375 ; ; ( 41:0.5 cm ) ( d ) ; ( -40:3 cm ) ( 0,0 ) ; ( c ) at ( -40:1.5 cm -| 6.6,0 ) @xmath105 @xmath376 on @xmath377 . ; ( -40:1.5 cm )
( c ) ; first consider the metric @xmath378 on @xmath270 : it clearly satisfies @xmath379 on @xmath270 .
now , let @xmath380 \to [ 0,1]$ ] be a smooth function with @xmath381 for @xmath382 and @xmath383 for @xmath384 .
we define the metric @xmath385 by gluing two metrics on open subsets of @xmath2 that coincide on their intersection @xmath386 : @xmath387 on @xmath2 we have @xmath388 , and on @xmath377 we have @xmath389 .
now let @xmath390 \to [ 0,1]$ ] be a smooth function such that @xmath391 for @xmath392 and @xmath393 for @xmath394 and set @xmath395 then @xmath82 satisfies properties ( 1)(5 ) . we claim that the metric @xmath82 produced above has the desired properties .
indeed , let @xmath396 , and suppose @xmath18 is a loop of shortest length in the free homotopy class associated to @xmath133 : @xmath18 will necessarily have length @xmath397 .
then we are in at least one of the following three cases .
* case a. * if @xmath398 , then there is a unique @xmath322 such that @xmath399 .
then @xmath18 is freely homotopic within @xmath329 to @xmath400 for some @xmath401 .
we have @xmath376 on @xmath402 , so @xmath18 has length at least @xmath403 with respect to @xmath82 .
* case b. * if @xmath404 , then we use that @xmath368 on @xmath405 to see that @xmath18 has length at least @xmath199 . * case c. * if @xmath406 contains points in @xmath407 and in @xmath408 , then the length of @xmath18 is at least the distance measured with metric @xmath82 between @xmath407 and @xmath408 .
but since @xmath409 on @xmath410 the definition of @xmath411 shows that this distance is at least @xmath199 . to finish the proof , we first notice that @xmath346 has length @xmath403 by property ( 4 ) of the lemma . now , suppose @xmath412 , with associated loop @xmath18 ( as above ) , is such that the set @xmath413 is nonempty with infimum @xmath414 .
if this infimum is below @xmath199 , then we must be in case a , and @xmath415 .
in particular , for @xmath303 we know that @xmath414 is less than @xmath199 by our assumption at the start of the proof , and @xmath416 by the assumption on @xmath69 in the theorem , and it follows that @xmath417 .
if the set @xmath418 is non - empty and its infimum is at least @xmath199 , then @xmath309 in all three cases .
if @xmath418 is empty , then we can not be in case @xmath419 and we also deduce that @xmath309 .
this completes the proof of theorem [ thm : systolelengthmap ] . in the first part of remark [ rem : systolelengthmap ] we established that conditions @xmath310 and @xmath311 in the hypotheses of theorem [ thm : systolelengthmap ] are equivalent to @xmath69 being the restriction of a length map on @xmath65 .
hence , we see that theorem [ thm : introsystole ] is just the first statement of theorem [ thm : systolelengthmap ] . before proving theorem [ thm : varyg ] we recall that its moral is that we can find a metric @xmath82 with a prescribed value and location for the first @xmath0 elements of the minimum marked length spectrum ( see definition [ def : minlspec ] ) .
now , given distinct unoriented free homotopy classes @xmath420 and a minimum marked length map @xmath67 , the length map axioms imply certain relations between the values @xmath421 . examining conditions @xmath311 and @xmath31 in the statement of theorem [ thm : systolelengthmap ] and keeping the moral of theorem [ thm : varyg ] in mind we are led to make the following definition .
[ dfn : admissible ] let @xmath2 be a closed manifold and let @xmath88 denote its collection of unoriented free homotopy classes . given a finite sequence @xmath422 of distinct elements of @xmath88 with @xmath423 the trivial class , we will say that a sequence @xmath424 is _
@xmath99-admissible _ if it satisfies the following conditions : 1 .
for @xmath425 we have @xmath426 whenever @xmath427 for some @xmath428 ; 2 . for @xmath429
we have @xmath430 whenever @xmath431 for some non - zero @xmath338 .
[ exa : admissible ] one can quickly verify that a sequence @xmath432 such that @xmath433 is @xmath99-admissible for any choice of @xmath434 in the definition above .
also , we note that if @xmath435 is a subset that is closed under taking powers , the definition of @xmath99-admissibility reduces to condition @xmath31 ( cf .
remark [ rem : systolelengthmap ] ) . to prove that @xmath105 implies @xmath106 first recall from homotopy theory that two homotopy classes @xmath436 give rise to the same unoriented free homotopy class in @xmath88 if and only if @xmath133 is conjugate to @xmath437 or to @xmath438 in @xmath65 . as a consequence , any minimum marked length map @xmath85 induces a map @xmath439 called the minimum marked length spectrum ( cf
. definition [ def : minlspec ] ) .
now , as in the statement of theorem [ thm : varyg ] , let @xmath440 be a sequence of @xmath0 distinct unoriented free homotopy classes , where @xmath423 is trivial and let @xmath441 be a @xmath99-admissible sequence .
for each @xmath442 choose @xmath443 within the class @xmath444 .
let @xmath445 and define @xmath295 by @xmath446 .
then it follows from condition @xmath31 of @xmath99-admissibility that @xmath69 satisfies the hypotheses of theorem [ thm : systolelengthmap ] .
hence , choosing @xmath447 in theorem [ thm : systolelengthmap ] it follows from condition @xmath32 of @xmath99-admissibility that we obtain a metric @xmath82 with the desired properties . to prove that @xmath106 implies @xmath105 , let @xmath82 be a metric on @xmath2 such that @xmath448 and @xmath449 for all @xmath450 .
then an examination of the length map axioms allows us to conclude that the sequence @xmath451 is @xmath99-admissible .
the statement and proofs of theorems [ thm : systolelengthmap ] and [ thm : varyg ] also hold when @xmath2 is a closed manifold of dimension @xmath42 and the sequence of free homotopy classes associated to the elements of @xmath81 can be represented by disjoint simple closed loops in @xmath2 : a condition that is always met in dimension @xmath21 and higher .
it is worth noting that colin de verdire s result @xcite on prescribing the laplace spectrum fails in dimension @xmath42 precisely because not every finite collection of free homotopy classes can be represented by a collection of pairwise disjoint simple closed curves @xcite .
we close this section by noting that in the case where @xmath57 is finite , we may take @xmath81 to be all of @xmath57 in theorem [ thm : systolelengthmap ] and @xmath452 to be any length map in the sense of definition [ dfn : lengthmap ] .
then since every finitely generated group is the fundamental group of some closed @xmath119-manifold ( * ? ?
* exercise 4.6.4(b ) ) we obtain the following result , which says that every length map on a finite group is the minimum marked length map of some riemannian manifold . [ cor : systolelengthmap ] let @xmath43 be a length map on a finite group @xmath22 .
then there exists a riemannian manifold @xmath1 and an isomorphism @xmath453 so that @xmath454 .
the aim of this section is to establish theorem [ thm : covsunada ] .
first , we identify how the covering spectrum of the total space of a riemannian covering can be computed in terms of the base space .
let @xmath455 be a riemannian covering ; that is , @xmath127 is a covering map that is also a local isometry . fixing a base point @xmath456 of @xmath2 and @xmath457 of @xmath10
, we have an injective group homomorphism @xmath458 given by @xmath180 \mapsto [ p \circ \sigma]$ ] . since @xmath127 is a local isometry ,
the lengths of @xmath87 and @xmath459 are the same , so we have the following commutative diagram : \(n ) at ( -1.7,2 ) @xmath57 ; ( nn ) at ( 1.7,2 ) @xmath460 ; ( r ) at ( 0,0.3 ) @xmath68 ; ( n ) node[above ] @xmath461 ( nn ) ; ( nn )
node[below right ] @xmath462(r ) ; ( n ) node[below left ] @xmath463(r ) ; from this the following observation is immediate . [
lem : restrict ] @xmath464 .
hence , if @xmath465 and @xmath466 are riemannian coverings of a common base space @xmath467 , then @xmath468 if and only if @xmath469 , where @xmath470 and @xmath471 are subgroups of @xmath472 .
a word of warning is in order here .
suppose that @xmath69 is a length map on a group @xmath22 and @xmath23 is a subgroup of @xmath22 , then the restriction of @xmath69 to @xmath23 is a length map on @xmath23 that gives rise to the filtration @xmath473 .
note however that @xmath474 may not be the same as @xmath475 .
the first group consists of finite products of elements of length below @xmath7 where all elements are in @xmath23 , and the second group consists of such products where the elements are in @xmath22 and only the product is required to be in @xmath23 .
thus , to obtain the correct covering spectrum of @xmath2 out of the length map @xmath476 on @xmath10 one has to first restrict @xmath476 to @xmath477 , rather than restricting the filtration of @xmath460 to @xmath477 . in order to prepare for the proof of theorem [ thm : covsunada ] , we first give an interpretation of condition @xmath37 of the theorem in terms of length maps .
[ lemma : jumptriple ] let @xmath22 be a finite group , and let @xmath23 and @xmath24 be subgroups
. then the following are equivalent : 1 . for all subsets @xmath478
that are stable under conjugation by elements of @xmath22 we have @xmath479 2 . for every length map @xmath69 on @xmath22 we have @xmath480 [ rem : range ] there is some freedom in choosing the range of conjugacy stable subsets of @xmath22 that @xmath81 and @xmath270 vary over in condition @xmath31 .
for instance , one obtains an equivalent condition when one lets @xmath81 and @xmath270 vary over all conjugacy stable subsets of @xmath22 , so that condition @xmath31 in the lemma is equivalent to condition @xmath37 of theorem [ thm : covsunada ] .
[ dfn : jump ] we will say that @xmath23 and @xmath24 are _ jump equivalent _ subgroups of @xmath22 or that @xmath33 is a _ jump triple _ , if the conditions in the lemma above are met . in the next section we will see how the notion of jump equivalence is related to other equivalence relations of subgroups .
suppose that @xmath31 holds and that @xmath69 is a length map on @xmath22 .
for every @xmath214 the subset @xmath481 is stable under conjugation .
we have @xmath482 , so @xmath483 by condition @xmath31 we see that this is equivalent to the same statement for @xmath24 , and @xmath32 follows . to show that @xmath32 implies @xmath31 suppose that @xmath484 are conjugacy stable subsets of @xmath22 , and that @xmath485 .
we will construct a length map @xmath486 that depends on @xmath81 and @xmath270 so that condition @xmath32 for this length map implies that the equivalence in @xmath31 holds .
we may assume that @xmath81 and @xmath270 are closed under taking inverses .
define @xmath69 by @xmath487 then @xmath69 satisfies the length map axioms of definition [ dfn : lengthmap ] , and we have @xmath488 , and @xmath489 .
it follows that @xmath490 the same holds when we replace @xmath23 by @xmath24 , so indeed condition @xmath32 for @xmath69 implies the equivalence in @xmath31 . with these preliminaries out of the way we now turn to the proof of theorem [ thm : covsunada ] .
first , we notice that since @xmath22 acts freely on @xmath2 we see that the natural projection @xmath491 is a regular covering with @xmath22 serving as the group of deck transformations .
let us choose a base point in @xmath2 for the fundamental group @xmath65 , which also determines a base point in the quotient of @xmath2 by any subgroup of @xmath22 . by ( * ?
2.6.3 ) there is a surjective homomorphism @xmath492 with @xmath493 .
hence , @xmath494 is an isomorphism if and only if @xmath2 is simply - connected .
we also note that , letting @xmath495 and @xmath496 denote the covering maps , we have @xmath497 and @xmath498 . we now prove that if @xmath2 is simply - connected , then @xmath37 implies @xmath38 .
assume that the triple @xmath41 satisfies condition @xmath37 .
then , since @xmath499 is an isomorphism , we see that condition @xmath37 holds for the triple @xmath500 .
the result now follows by applying lemma [ lemma : jumptriple ] and lemma [ lem : restrict ] to the coverings @xmath501 and @xmath502 .
we now assume that @xmath2 is a manifold of dimension at least @xmath21 and prove that in this case condition @xmath38 implies condition @xmath37 .
assuming condition @xmath37 does not hold , our goal will be to construct a riemannian metric on @xmath28 such that the length map attains certain prescribed values on a suitably chosen finite set of conjugacy classes of @xmath503 . pulling this metric back to @xmath2
will then give a @xmath22-invariant riemannian metric on @xmath2 so that the quotient manifolds @xmath25 and @xmath26 have distinct covering spectra .
note first that @xmath503 is finitely generated , and that @xmath22 is finite , so @xmath504 is generated as a group by a finite subset @xmath505 of @xmath506 ; see e.g. ( * ? ? ?
by the assumption that condition @xmath37 does not hold , one sees that there are subsets @xmath81 and @xmath270 of @xmath507 both stable under conjugation by elements of @xmath22 , and taking inverses , so that we have @xmath485 , and @xmath508 , but @xmath509 ( switch @xmath23 and @xmath24 if necessary ) .
now choose any lift @xmath510 of @xmath270 to @xmath503 , i.e. , any subset @xmath511 so that @xmath494 maps @xmath510 bijectively to @xmath270 .
let @xmath512 , which is a lift of @xmath81
. we will prescribe the length map on the finite subset @xmath513 of @xmath503 . note that this union is a disjoint union . for any subset @xmath121 of @xmath503 let us write @xmath514 for its closure under conjugation . since we assumed that the dimension of @xmath2 is at least @xmath21 ,
theorem [ thm : systolelengthmap ] now implies that there is a riemannian metric @xmath82 on @xmath28 such that the associated minimum marked length map @xmath515 satisfies @xmath516 for any conjugacy stable @xmath517 that contains @xmath505 the group @xmath518 contains @xmath519 , and its image under @xmath494 is @xmath520 , so we have @xmath521 since @xmath522 , we see that for every @xmath7 with @xmath523 we have @xmath524 and we conclude that @xmath119 is not in @xmath525 . on the other hand , since @xmath526 , we have for each @xmath527 @xmath528 therefore , @xmath119 is an element of @xmath529 .
now , since we have @xmath497 and @xmath498 , it follows from lemma [ lem : restrict ] that @xmath42 lies in the covering spectrum of @xmath26 but not in the covering spectrum of @xmath25 .
this completes the proof of theorem [ thm : covsunada ] .
suppose that @xmath22 is a finite group , and let @xmath23 and @xmath24 be subgroups . in this entirely group theoretic
section we study the notion of jump equivalence of @xmath23 and @xmath24 , as defined in definition [ dfn : jump ] . like gassmann - sunada equivalence ,
jump equivalence is a strong form of the group theoretic notion corresponding to kronecker equivalence of number fields @xcite .
many instances of gassmann - sunada equivalence , such as the komatsu triples considered in @xcite , satisfy a stronger condition , which we call _ order equivalence_. the definitions and relations of these four equivalences are given in figure [ fig : subgroupequiv ] , where the stated conditions are to hold for all subsets @xmath81 , @xmath270 of @xmath22 that are closed under conjugation by elements of @xmath22 .
it is not hard to show the implications in the diagram we leave this to the reader .
the main purpose of this section is to provide examples of gassmann - sunada equivalent subgroups that are not jump equivalent , which in the next section give rise to isospectral manifolds with distinct covering spectra .
in fact , we will see that there are no relations between the four equivalences that hold for all @xmath33 other than the ones implied by the diagram .
we start with a few basic remarks . in the last row the covering spectra are the same , but the multiplicities in the sense of example [ exa : flattori ] or ( * ? ? ?
* definition 6.1 ) are not : they are @xmath530 ( respectively ) for @xmath270 and @xmath531 ( respectively ) for @xmath510 .
this table was obtained using magma @xcite and can be verified ( tediously ) by hand .
[ rem : tori ] in fact , @xmath532 and @xmath533 are jump equivalent in @xmath534 , but since @xmath535/3a$ ] is not simply connected , this does not contradict theorem [ thm : covsunada ] .
furthermore , @xmath532 and @xmath533 are order equivalent in @xmath534 .
however , one can check that @xmath536 and @xmath537 are _ not _ jump equivalent in @xmath538 .
[ exa : heis ] let @xmath539 be a real inner product space of dimension @xmath540 endowed with a non - degenerate symplectic form @xmath541 ; that is , @xmath542 is a non - degenerate skew - symmetric form .
the heisenberg group @xmath543 associated to @xmath544 is the manifold @xmath545 , together with the group operation @xmath546 with this operation @xmath23 becomes a non - commutative real lie group .
the identity element @xmath547 is the pair @xmath548 .
the center @xmath549 of @xmath23 is equal to the commutator subgroup @xmath550 $ ] of @xmath23 , so @xmath23 is @xmath42-step nilpotent , and the commutator pairing @xmath551 $ ] is given by the symplectic form : @xmath552 the tangent space of @xmath23 at @xmath553 is @xmath554 , which we view as an inner product space by putting the standard inner product on @xmath555 and taking the orthogonal direct sum with @xmath556 .
we extend this inner product uniquely to a left - invariant riemannian metric @xmath82 on @xmath23 .
note that the group isomorphism @xmath557 is an isometry between the euclidean space @xmath544 with its flat metric and the quotient manifold @xmath558 together with its induced structure of a riemannian manifold .
now suppose we have a lattice @xmath269 of full rank in @xmath544 , and a positive real number @xmath18 such that @xmath559 .
then @xmath560 is a discrete subgroup of @xmath23 that acts freely and by isometries on @xmath23 by left multiplication .
the quotient space @xmath561 is then a riemannian manifold , called a _
riemannian heisenberg manifold_. note that @xmath562 acts freely by isometries on @xmath121 , and that the quotient @xmath563 , considered as a riemannian manifold , is just the flat torus @xmath564 .
the covering spectrum of @xmath270 is the jump set of the length map @xmath565 on @xmath269 as in example [ exa : flattori ] .
we will now show how to obtain the covering spectrum of @xmath121 from that of @xmath270 .
the covering spectrum of @xmath121 is the jump set of the length map @xmath566 on @xmath567 , where @xmath568 is the geodesic distance on @xmath23 ( see example [ exa : minmark ] ) . 1 .
@xmath572 for all non - zero @xmath573 , and 2 . among all free homotopy classes represented by central elements @xmath574 of @xmath567 ,
the one that can be represented by a closed geodesic of smallest length is @xmath575 to deduce the proposition , suppose first that @xmath577 . then by property ( 2 ) the element @xmath578 has the smallest non - zero length in @xmath567 .
the intermediate groups of @xmath579 now correspond exactly to the subgroups of @xmath269 , and by property ( 1 ) the filtrations induced by @xmath580 and @xmath581 ( respectively ) correspond as well , so this gives us the first case .
for the second case , suppose that @xmath573 with @xmath582 and @xmath583 .
then we have @xmath584 , and by property ( 1 ) both @xmath585 and @xmath586 have length @xmath587 .
thus , @xmath588 is contained in the smallest non - trivial group in the filtration of @xmath567 , and the second case follows .
[ rem : heisequiv ] by ( * ? ? ?
* prop 2.16 ) any quotient of a @xmath589-dimensional simply connected heisenberg lie group endowed with a left - invariant metric by a cocompact , discrete subgroup is isometric to a riemannian heisenberg manifold as described above , i.e. , to one of the form @xmath590 , where @xmath591 and @xmath560 are as above .
moreover , if two such riemannian heisenberg manifolds are isospectral , then they are isometric to manifolds @xmath592 and @xmath593 given with the same @xmath544 , @xmath594 , @xmath541 , @xmath18 , and two cocompact lattices @xmath269 and @xmath282 in @xmath544 such that the flat tori @xmath564 and @xmath595 are isospectral .
in fact , two heisenberg manifolds @xmath121 and @xmath596 as above are isospectral if and only if @xmath564 and @xmath595 are isospectral . to see these comments
, the reader may consult ( * ? ? ?
* prop . 2.14 , 2.16 ) and ( * ? ? ?
iii.5 ) .
[ cor : heiscovspec ] suppose that @xmath121 and @xmath596 are isospectral heisenberg manifolds as in remark [ rem : heisequiv ] .
then @xmath121 and @xmath596 have the same covering spectra if and only if @xmath270 and @xmath510 have the same covering spectra . by statement
( 2 ) in the proof of proposition [ prop : heiscovspec ] we see that @xmath597 since both are equal to the distance in @xmath543 from @xmath548 to @xmath578 . as the flat tori @xmath564 and @xmath595 are isospectral , they have the same length spectrum , so there is a one to one correspondence between the lengths of the vectors in @xmath269 and the lengths of the vectors in @xmath282 . in particular , we must have @xmath598 .
the result now follows from proposition [ prop : heiscovspec ] .
let @xmath599 $ ] be a pair of lattices of conway and sloane as in example [ exa : tori ] , that give rise to isospectral 4-dimensional flat tori with distinct covering spectra .
one easily checks that @xmath600 and @xmath601 are discrete subgroups of @xmath602 where @xmath603={\mathbb r}^{4}$ ] with the standard symplectic form . by remark
[ rem : heisequiv ] , @xmath604 and @xmath605 are isospectral , since @xmath564 and @xmath595 are isospectral . by proposition [ prop : heiscovspec ]
, they have distinct covering spectra , as @xmath270 and @xmath510 have distinct covering spectra .
it follows from corollary [ cor : heiscovspec ] that the example of isospectral heisenberg manifolds with different covering spectra given by sormani and wei ( see ( * ? ? ?
* example 10.3 ) , @xcite ) is incorrect , as their tori @xmath270 and @xmath510 have the same covering spectrum . in ( * ? ? ?
* definition 6.1 ) , sormani and wei assign a _
basis multiplicity _ to each element @xmath7 of the covering spectrum of a compact riemannian manifold @xmath2 .
it is the minimal number of generators of length @xmath606 that one needs to add to @xmath607 to generate the next bigger group in the filtration .
one can show that if @xmath121 is a riemannian heisenberg manifold with @xmath608 , then the basis multiplicity of @xmath609 in @xmath153 is either equal to the basis multiplicity of @xmath609 in @xmath610 or equal to one more than the basis multiplicity of @xmath609 in @xmath611 all other basis multiplicities of @xmath153 are unaffected ; i.e. , their basis multiplicity in @xmath153 is equal to their basis multiplicity in @xmath611 the isospectral heisenberg manifolds of ( * ? ? ?
* example 2.4a ) have the same covering spectra by corollary [ cor : heiscovspec ] , but @xmath576 has a different basis multiplicity in the two examples . | in 2004 , sormani and wei introduced the covering spectrum : a geometric invariant that isolates part of the length spectrum of a riemannian manifold . in their paper
they observed that certain sunada isospectral manifolds share the same covering spectrum , thus raising the question of whether the covering spectrum is a spectral invariant . in the present paper
we describe a group theoretic condition under which sunada s method gives manifolds with identical covering spectra .
when the group theoretic condition of our method is not met , we are able to construct sunada isospectral manifolds with distinct covering spectra in dimension 3 and higher .
hence , the covering spectrum is not a spectral invariant .
the main geometric ingredient of the proof has an interpretation as the minimum - marked - length - spectrum analogue of colin de verdire s classical result on constructing metrics where the first @xmath0 eigenvalues of the laplace spectrum have been prescribed .
funded in part by the european commission under contract mrtn - ct-2006 - 035495 ] research partially supported by nsf grant dms-0204648 ] research partially supported by an nsf postdoctoral research fellowship , nsf grant dms-0605247 and a woodrow wilson foundation career enhancement fellowship ] |
the leading structures of the nucleon in deeply inelastic scattering processes are described in terms of three twist-2 parton distribution functions the unpolarized @xmath3 , helicity @xmath4 , and transversity @xmath5 distribution .
owing to its chirally odd nature @xmath5 escapes measurement in deeply inelastic scattering experiments which are the main source of information on the chirally even @xmath3 and @xmath4 .
the transversity distribution function was originally introduced in the description of the process of dimuon production in high energy collisions of transversely polarized protons @xcite .
alternative processes have been discussed .
let us mention here the collins effect @xcite which , in principle , allows to access @xmath6 in connection with a fragmentation function describing a possible spin dependence of the fragmentation process , see also mulders:1995dh and references therein .
recent and/or future data from semi - inclusive deeply inelastic scattering ( sidis ) experiments at hermes @xcite , clas @xcite and compass legoff : qn could be ( partly ) understood in terms of this effect desanctis:2000fh , ma:2002ns , efremov:2001cz . other processes to access @xmath6 have been suggested as well , see the review barone:2001sp .
however , in all these processes @xmath6 enters in connection with some unknown fragmentation function .
moreover these processes involve the introduction of transverse parton momenta , and for none of them a strict factorization theorem could be formulated so far .
the drell - yan process remains up to now the theoretically cleanest and safest way to access @xmath6 . the first attempt to study @xmath5 by means of the drell - yan process
is planned at rhic @xcite .
dedicated estimates , however , indicate that at rhic the access of @xmath5 by means of the drell - yan process is very difficult @xcite .
this is partly due to the kinematics of the experiment .
the main reason , however , is that the observable double spin asymmetry @xmath7 is proportional to a product of transversity quark and antiquark distributions .
the latter are small , even if they were as large as to saturate the soffer inequality soffer:1994ww which puts a bound on @xmath5 in terms of the better known @xmath3 and @xmath4 .
this problem can be circumvented by using an antiproton beam instead of a proton beam .
then @xmath7 is proportional to a product of transversity quark distributions from the proton and transversity antiquark distributions from the antiproton ( which are connected by charge conjugation ) .
thus in this case @xmath7 is due to valence quark distributions , and one can expect sizeable counting rates .
the challenging program how to polarize an antiproton beam has been recently suggested in the * p*olarized * a*ntiproton e*x*periment ( pax ) at gsi @xcite .
the technically realizable polarization of the antiproton beam of about @xmath8 and the large counting rates due to the use of antiprotons make the program promising . in this note
we shall make quantitative estimates for the drell - yan double spin asymmetry @xmath7 in the kinematics of the pax experiment . for that we shall stick to the description of the process at lo qcd .
nlo corrections for @xmath7 have been shown to be small @xcite .
similar estimations were done earlier @xcite using different models for the transversity distribution . here for transversity distribution we shall use the result of the covariant probabilistic model developed earlier @xcite . in this model
the quarks are represented by quasifree fermions on mass shell and their intrinsic motion , which has spherical symmetry and is related to the orbital momentum , is consistently taken into account .
it was shown , that the model nicely reproduces some well - known sum rules .
the calculation was done from the input on unpolarized valence quark distributions @xmath9 and it was shown , that assuming @xmath10 symmetry , a very good agreement with experimental data on the proton spin structure functions @xmath11 and @xmath12 can be obtained .
in the paper @xcite we discussed the transversity distribution in the mentioned quark - parton ( qpm ) model .
this model , in the limit of massless quarks , implies the relation between the transversity and the corresponding valence quark distribution : @xmath13the factors @xmath14 represent relative contributions to the proton spin from different quark flavors , which for the assumed @xmath10 symmetry means , that @xmath15 and @xmath16 .
the factor @xmath17 depends on the way , in which the transversity is calculated : _ i ) _ interference effects are attributed to the quark level only , then @xmath18 . in this approach
the relation between the transversity and the usual polarized distribution is obtained@xmath19which means , that the resulting transversity distribution is roughly twice as large as the usual @xmath20 .
the signs of both the distributions are simply correlated .
soffer inequality in this approach is violated for the case of large negative quark polarization , when @xmath21 , which means that the proton @xmath22quarks in the @xmath10 scheme are just on the threshold of violation .
_ ii ) _ interference effects at parton - hadron transition stage are included in addition , but the result represents only upper bound for the transversity . this bound is more strict than the soffer one and roughly speaking , our bound is more restrictive for quarks with a greater proportion of intrinsic motion and/or smaller ( or negative ) portion in the resulting polarization . no simple correspondence between the signs of actual transversity and @xmath20 follows from this approach . in this scenario : @xmath23 . following the papers @xcite ,
the transversity can be measured from the drell - yan process @xmath24 @xmath25 in the transversely polarized @xmath26 collisions in the proposed pax experiment .
the transversity can be extracted from the double transverse spin asymmetry@xmath27where , using momenta @xmath28 of the incoming proton@xmath29antiproton pair and the momenta @xmath30 of the outgoing lepton pair , one defines the physical observables @xmath31the variable @xmath32 can be interpreted as the rapidity of lepton pair .
the asymmetry @xmath7 is obtained from the cross sections corresponding to the different combinations of transverse polarizations in the incoming @xmath26 pair@xmath33 rt!.5 where the last expression corresponds to the double spin asymmetry in the qed elementary process , @xmath34 . so using the above formulas
, one can calculate the double spin asymmetry ( [ dy3 ] ) from the valence quark distribution according to the relation ( [ dy1 ] ) . in fig .
[ fi1 ] the result of the calculation is shown .
the normalized input on the proton valence quark distribution was taken from ref .
@xcite , which corresponds to @xmath35 and the energy squared of @xmath26 system is taken @xmath36 in an accordance with the assumed pax kinematics . in the same figure the curve obtained at @xmath37 from the calculation @xcite based on the chiral quark - soliton model @xcite is shown for a comparison .
all curves in this figure are based on the same parameterization @xcite of the distribution functions @xmath38 appearing in the denominator in eq .
obviously , our calculation gives a lower estimate of the @xmath7 and one of possible reasons can be the effect of quark intrinsic motion , which , as we have shown , can play role not only for the spin function @xmath11@xcite , but also for the transversity @xmath39 @xcite . in an accordance with @xcite , the motion of the lepton pair can be described alternatively with the using the variable @xmath40 rh.5 in the fig .
[ fi2 ] the estimation of asymmetry obtained in the cited paper for @xmath35 and at @xmath41 is compared with our curves from fig .
[ fi1 ] , in which the variable @xmath42 is replaced by the @xmath43 , whereas both the variables are related by the transformation ( [ dy6 ] ) . apparently , for @xmath44
the curve from @xcite is quite compatible with our results .
so , in both the figures we have the set of curves resulting from different assumptions and the experiment should decide , which one gives the best fit to the data .
how many events is necessary for discriminating among the displayed curves ?
after integrating over angular variables one gets@xmath45then@xmath46which implies@xmath47so for approximate estimate of the statistical error we obtain@xmath48where @xmath49 is number of events related to the bin or interval of @xmath32 or @xmath43 in which the curves are compared .
for example , if one requires @xmath50 , which is error allowing to separate the curves in presented figures , then roughly @xmath51 should be the number of events in the considered bin or interval .
of course , this estimation assumes full polarization of the colliding proton and antiproton .
since the expected polarization of antiprotons at the pax will hardly be better than @xmath8 , the minimum number of events will be correspondingly higher .
the covariant probabilistic qpm , which takes into account intrinsic quark motion , was applied to the calculation of transverse spin asymmetry of dileptons produced in the @xmath26 collisions in the conditions , which are expected for the recently proposed experiment pax .
this asymmetry is directly related to the transversity distributions of quarks inside the proton . in our asymmetry calculation
the two approaches for the transversity , which differ in accounting for the interference effects , were applied .
our obtained results are compared with the prediction based on the quark - soliton model .
one can observe , that quite different approaches give the similar results , but both our curves are lower than that obtained from the quark - soliton model .
our results for @xmath44 are also well compatible with the recent estimate @xcite .
this work is supported by votruba - blokhintsev programm of jinr , a.e . and o.t .
are partially supported by grants rfbr 03 - 02 - 16816 .
further , this work has been supported in part by the academy of sciences of the czech republic under the project av0-z10100502 .
99 -1 mm j. p.
ralston and d. e. soper , nucl .
b * 152 * ( 1979 ) 109 .
r. l. jaffe and x. d. ji , nucl .
b * 375 * ( 1992 ) 527 .
j. l. cortes , b. pire and j. p. ralston , z. phys .
c * 55 * ( 1992 ) 409 .
j. c. collins , nucl .
b * 396 * ( 1993 ) 161 .
p. j. mulders and r. d. tangerman , nucl .
b * 461 * ( 1996 ) 197 [ erratum - ibid .
b * 484 * ( 1997 ) 538 ] .
d. boer and p. j. mulders , phys .
d * 57 * ( 1998 ) 5780 .
a. airapetian _ et al . _
[ hermes collaboration ] , phys .
* 84 * ( 2000 ) 4047 ; phys .
d * 64 * ( 2001 ) 097101 ; phys .
b * 562 * ( 2003 ) 182 .
h. avakian _ et al . _
[ clas collaboration ] , arxiv : hep - ex/0301005 .
j. m. le goff [ compass collaboration ] , nucl .
a * 711 * ( 2002 ) 56 .
e. de sanctis , w. d. nowak and k. a. oganesian , phys .
b * 483 * ( 2000 ) 69 .
v. a. korotkov , w. d. nowak and k. a. oganesian , eur .
j. c * 18 * ( 2001 ) 639 .
k. a. oganessian , n. bianchi , e. de sanctis and w. d. nowak , nucl .
a * 689 * ( 2001 ) 784 .
b. q. ma , i. schmidt and j. j. yang , phys .
d * 66 * ( 2002 ) 094001 , d * 65 * ( 2002 ) 034010 , d * 63 * ( 2001 ) 037501 . m. anselmino and f. murgia , phys . lett .
b * 483 * ( 2000 ) 74 .
a. v. efremov , k. goeke and p. schweitzer , phys .
b * 522 * ( 2001 ) 37 , b * 544 * ( 2002 ) 389e and b * 568 * ( 2003 ) 63 ; eur .
j. c * 24 * ( 2002 ) 407 and c * 32 * ( 2004 ) 337 ; phys .
d * 67 * ( 2003 ) 114014 ; acta phys .
b * 33 * ( 2002 ) 3755 ; aip conf .
* 675 * ( 2003 ) 459.p .
schweitzer and a. bacchetta , nucl .
phys . a * 732 * ( 2004 ) 106 .
v. barone , a. drago and p. g. ratcliffe , phys .
* 359 * ( 2002 ) 1 .
l. c. bland [ star collaboration ] , aip conf .
* 675 * ( 2003 ) 98 ; arxiv : hep - ex/0403012 . g. bunce , n. saito , j. soffer and w. vogelsang , ann . rev .
nucl . part .
* 50 * ( 2000 ) 525 .
o. martin , a. schfer , m. stratmann and w. vogelsang , phys .
d * 60 * ( 1999 ) 117502 . c. bourrely and j. soffer , nucl .
b * 423 * ( 1994 ) 329 .
v. barone , t. calarco and a. drago , phys .
d * 56 * ( 1997 ) 527 .
j. soffer , phys .
* 74 * ( 1995 ) 1292 .
g. r. goldstein , r. l. jaffe and x. d. ji , phys .
d * 52 * ( 1995 ) 5006 .
pax collaboration , antiproton - proton scattering experiments with polarization , letter of intent ( darmstadt , 2004 ) , available at http://www.fz-juelich.de/ikp/pax .
w. vogelsang and a. weber , phys .
d * 48 * ( 1993 ) 2073 .
a. p. contogouris , b. kamal and z. merebashvili , phys .
b * 337 * ( 1994 ) 169 .
h. shimizu , g. sterman , w. vogelsang , h. yokoya , arxiv : hep - ph/0503270 . | estimates are given for the double spin asymmetry in lepton - pair production from collisions of transversely polarized protons and antiprotons for the kinematics of the recently proposed pax experiment at gsi on the basis of predictions for the transversity distribution from the probabilistic quark - parton model developed earlier .
* transversity and the polarized + drell - yan process in @xmath0 * a.v .
efremov@xmath1 , o.v . teryaev@xmath1 and @xmath2 |
with a galactic core - collapse supernova rate of 13 per century ( e.g. , * ? ? ?
* ) , young neutron stars ( with age @xmath5kyr ) are rare .
nevertheless , even by those standards the known sample is woefully incomplete .
only 12 rotation - powered pulsars with characteristic age @xmath6kyr are known in the galaxy , and less than 20 such pulsar
supernova remnant ( snr ) associations are firmly established . developing a more complete picture of the young pulsar population contributes to an understanding of the birthrate of neutron stars and of the physics of their creation in stellar core collapses .
while most snrs are in principle good locations to search for young pulsars , wholesale searches require multiple telescope pointings each with potentially inadequate sensitivity and have had limited success ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a more recent and successful approach has been to target , with single deep observations , compact pulsar wind nebulae ( pwne ) , identified via x - ray or radio imaging and spectroscopy , which indicate the presence of a young neutron star even in the absence of the detection of pulsations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
detection of the period @xmath7 , @xmath8 , and derived quantities , underlies significant further understanding of the pulsar , its relativistic wind , pwn , and environment ( see , e.g. , * ? ? ?
* ) . a composite radio and x - ray image of snr and its central pulsar wind nebula ( pwn ) . in red
is shown a 1.5ghz image , using vla data taken in 1984 ( reproduced from * ? ? ?
* ) ; in green is displayed a 0.510kev x - ray image using the epic pn detector on _ xmm - newton _ , produced from an archival observation taken in 2003 ( see * ? ? ?
yellow denotes the radio- and x - ray - bright core of the pwn , within which is located the hard x - ray point source @xcite that likely is emission from .
the vla data are at an angular resolution of @xmath9 while the _ xmm - newton _ data have been smoothed with a @xmath10 gaussian .
the white circle shows the half - power point of the gbt beam for the pulsar search at 2ghz .
the x - ray source inside the snr shell and to the south - west of the x - ray / radio core is the x - ray and @xmath11-ray transient xmmu j174716.1281048 = igr j174642811 @xcite . the x - ray source outside and to the west of the snr shell is coincident with the _ chandra _ source cxo j174705.4280859 @xcite . ]
lcccccc 2002 may 16 & parkes & 1.4 & @xmath12 & 2.0 & 9.3 & + 2002 nov 5 & parkes & 1.4 & @xmath12 & 1.6 & 7.5 & + 2003 nov 5 & parkes & 2.9 & @xmath13 & 0.25 & 5.6 & + 2005 oct 20 & parkes & 1.4 & @xmath14 & 0.25 & 9.4 & + 2006 jan 8 & gbt & 1.9 & @xmath15 & 0.08 & 6.0 & 52.137293(2 ) + 2009 mar 11 & gbt & 2.0 & @xmath16 & 0.16 & 5.8 & 52.152855(2 ) the composite snr consists of a radio shell @xmath17 in diameter surrounding a @xmath18 pwn ( helfand & becker 1987 ; see figure [ fig : snr ] ) . based on the very high interstellar absorption ( @xmath19@xmath20 ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , its distance is large , here parameterized by @xmath21 .
the pwn is luminous in radio , with @xmath22 @xcite , and filled with x - ray synchrotron emission with @xmath23 @xcite .
it is also a very - high - energy @xmath11-ray source , with @xmath24 @xcite
. based on these energetics and a variety of empirical relations , it has been predicted that the pulsar powering this pwn has @xmath250.2s and spin - down luminosity @xmath26@xcite , while the snr shell size implies an age of about 17kyr @xcite .
the hard x - ray point source , with 1% the luminosity of the pwn and surrounded by small - scale ordered structure , is likely emission from the pulsar @xcite .
we have searched this location for a radio pulsar , and in this letter report the discovery of , the central pulsar in snr .
we began our search for the pulsar in snr at the atnf parkes telescope , where during 20022005 we used three combinations of search frequencies and filterbank data acquisition systems to do four very long observations ( see table [ tab : obs ] ) .
the @xcite ne2001 free electron distribution model predicts that in this direction , for @xmath27 , the expected dispersion measure would be @xmath28pc@xmath2 , and that the pulse broadening due to interstellar scattering at the standard search frequency of 1.4ghz would be @xmath29ms .
given the possibility of larger actual dm and @xmath30 , one of our observations was at 3ghz .
we analyzed all search data with standard pulsar search techniques implemented in presto @xcite , including the excision of radio frequency interference ( rfi ) and a nearly optimal set of trial dms ( for more details see , e.g. , * ? ? ? * ) .
we dedispersed at up to twice the maximum galactic dm of 1580pc@xmath2 predicted by ne2001 in this direction .
no new pulsars were detected in any of the parkes data sets .
the parkes sensitivity limits at 1.4ghz correspond to luminosity @xmath31mjykpc@xmath32 , provided that scattering was not the limiting factor . at 3ghz , although the sky background temperature was reduced , any ordinary pulsar would have had an even more significantly reduced flux , and the equivalent luminosity limit was worse .
since young pulsars can have luminosities at least as small as @xmath33mjykpc@xmath32 @xcite , we did a deeper search at the nrao green bank telescope ( gbt ) . in 2006 january we observed at the gbt at a central frequency of 1.95ghz using the spigot autocorrelation spectrometer @xcite .
for a typical spectral index of @xmath34 @xcite , the pulsar would be fainter by a factor of about 1.7 by comparison with 1.4ghz . however , this was more than made up by the larger gain and bandwidth at the gbt , such that our observation was a factor of about 2 more sensitive than the best parkes search , both compared at the same fiducial frequency of 1.4ghz .
in addition , by using a higher search frequency , the scattering timescale was reduced by a factor of about 4 , which proved crucial in light of the actual pulsar period and the amount of observed scattering .
we identified in these data a highly dispersed ( @xmath35pc@xmath2 ) and scattered pulsar candidate with period @xmath36ms ( figure [ fig : psr ] ) .
it is a faint signal , with flux density @xmath37jy , compared to the detection threshold in this observation for a pulsar with these characteristics of approximately @xmath38jy .
we confirmed the pulsar in an equivalent observation using the green bank ultimate pulsar processing instrument ( guppi ) on 2009 march 11 ( figure [ fig : psr ] ) , for which @xmath39jy . in the presence of rfi , guppi , an fpga - based digital spectrometer ,
has vastly improved performance compared to spigot .
this proved vital , since the rfi environment has degraded enormously in the intervening 3 years .
detections of at gbt .
each 2ghz profile ( @xmath36ms ) is repeated in phase , and is shown as a function of time ( _ bottom _ ) and as the sum of more than 400,000 pulses ( _ top _ ) .
left : discovery observation from 2006 january 8 with spigot .
right : confirmation observation from 2009 march 11 with guppi .
the observed pulse fwhm increases from @xmath40 at 2.35ghz to @xmath41 at 1.75ghz , due to interstellar scattering . ] to measure the amount of scattering that clearly affects the pulse profile of ( figure [ fig : psr ] ) , we did a fit to the guppi profile simultaneously in seven 100mhz - wide sub - bands , assuming that the scattering timescale @xmath42 scales with observing frequency as @xmath43 with @xmath44 ( see , e.g. , * ? ? ?
we obtain @xmath45s , scaled to the usually reported frequency of 1ghz ( for @xmath46 in the range 3.64.4 , @xmath47 varies over 0.160.28s ) .
this compares to @xmath48s predicted by @xcite for this dm and direction .
the observed dm is biased by the large degree of scattering , and in our fitting process we obtain the corrected @xmath49 .
a comparison of the barycentric periods for the discovery and confirmation observations ( table [ tab : obs ] ) shows that the pulsar period increased by @xmath50s . the time interval between these observations was @xmath51 days ( mjd @xmath52 ) .
the average period derivative is thus @xmath53 .
the corresponding derived pulsar parameters are @xmath54 , @xmath55kyr , and surface magnetic dipole field strength @xmath56 g . for the observed dm , the ne2001 model yields @xmath57kpc , though with substantial uncertainty .
has a magnetic field strength at its light cylinder of @xmath58 g , which is the seventh highest among pulsars known in the galactic disk .
it has been suggested that @xmath59 is a controlling parameter for the emission of very narrow `` giant '' radio pulses @xcite , and among the four pulsars known to emit giant pulses , psr b054069 , in the lmc , has the smallest value , @xmath60 g .
therefore , in addition to the periodicity searches of the data listed in table [ tab : obs ] ( we also did a more sensitive folding analysis of all the parkes data sets , using the subsequently determined values of @xmath7 and @xmath8 , but did not detect the pulsar ) , we carried out a search for dispersed individual pulses in the two gbt data sets .
we used standard matched - filter techniques described in detail by @xcite and implemented in the sigproc pulsar search package .
no single pulses were detected . however , our sensitivity was much degraded as a consequence of the large single - channel dispersion smearing ( 12ms ) and the huge scattering ( about 13ms at 2ghz ) , but for which we could easily have detected giant pulses like those observed from psr b054069 by @xcite .
is a rare young neutron star in that its spin - down luminosity is surpassed among known galactic pulsars only by the crab .
we have estimated the chance probability of detecting such a pulsar in our search by simulating an inner - galactic gbt survey with individual 6hr spigot observations at 2ghz .
considering the best population model of @xcite , we find that such a survey would detect one ordinary ( non - millisecond ) pulsar every 0.07deg@xmath32 .
our beam area is 11% of this , which is also the probability of finding one such pulsar blindly . until now , only the crab was known with @xmath61among 1600 ordinary galactic pulsars .
the probability of finding any such high @xmath62 pulsar by chance is thus @xmath63 .
is also very distant , based on its large dm , as expected for the pulsar in snr from the high @xmath64 to its pwn .
the current positional uncertainty of is @xmath65 , which is sufficient to locate it in projection within the shell ( figure [ fig : snr ] ) . within this area ,
there are no radio or x - ray sources other than the pwn that could plausibly account for a young pulsar with @xmath66 , despite the great sensitivity of existing vla , _ chandra x - ray observatory _ , and _ xmm - newton _ observations .
there is therefore no doubt that and are associated .
the electron density model is not well constrained toward the distant inner galactic regions ( see * ? ? ?
prior to the discovery of , only nine pulsars were known with @xmath67 , of which only two are within 20 degrees of the galactic center ( gc ) , none with independently estimated distances @xcite .
therefore , the ne2001 distance of 13kpc for could be substantially in error .
it is not excluded that the pulsar could be located at 89kpc , physically near the gc .
however , given both the very large @xmath64 and dm , it is unlikely that the pulsar is located substantially closer than the gc .
also , the ratio @xmath68 for is higher than seen toward all but about four other pulsars ( see * ? ? ?
* ; * ? ? ?
gaensler et al . argue that this indicates a location behind substantial intervening molecular material , which would not be surprising for a location near or beyond the gc .
it is also intriguing that two pulsars located only 0.3 degrees in projection from the gc ( at a lateral distance of about 40pc if near it ) have @xmath69 @xcite , like , which among known pulsars is the third nearest to the gc in projection .
the scattering timescale for those two pulsars is a factor of a few larger than for , but still very small compared to the levels expected if they were located within the scattering screen thought to surround the gc at a distance estimated as 50330pc by @xcite .
@xcite therefore argue that those two pulsars are located somewhat in front of the gc .
the same may apply to , with a larger lateral separation of about 130pc , but we also can not exclude a substantially larger distance .
we consider it likely that @xmath70 .
the spin parameters of the new pulsar are similar to those of in snr , which has @xmath71ms , @xmath72 , @xmath73kyr , and @xmath74kpc @xcite .
it is therefore of interest to compare the properties of both systems .
the radius of the central pwn in is @xmath75pc , nominally 50% larger than the pwn in .
the radio luminosities may be similar , although this comparison is uncertain because of the much better frequency coverage for ( see * ? ? ?
* ; * ? ? ?
* and references therein ) . in x - rays ( 210kev ) , @xmath76 @xcite . at the highest energies , both pwne are unresolved tev sources , and colocated with their lower - energy counterparts , but is @xmath77 times more luminous than in a comparable band @xcite .
although the characteristic ages of both psrs j17472809 and j18331034 are @xmath78kyr , the actual age of and its snr is only @xmath79kyr @xcite . could therefore be the older of the two , and one might be tempted to appeal to this possible difference to explain the larger size of the pwn ( and of its shell , whose radius @xmath80 is also nominally 50% larger than that of ) and ( along with a smaller nebular magnetic field ) its possibly lower x - ray efficiency .
however , the nominal difference in the two efficiencies is small compared to the scatter observed among young pulsars .
dissimilar sn explosion kinetic energies and ejected masses could instead explain the different radii ( see , e.g. , * ? ? ?
some observables point to the possibility that also for , @xmath81 .
the shell is relatively bright in radio , unlike .
this could be due to a different circumstellar environment , or may suggest that the snr has swept enough mass and is transitioning to the adiabatic phase .
nevertheless , the pwn still retains approximate circular symmetry and central location within the snr shell , which likely indicates that a strong reverse shock has not yet formed . in that case , for the observed ratio @xmath82 ( similar to that of ) , the pwn evolutionary models of @xcite indicate a small age .
the pwn energetics also indicate a small age : for , the pwn magnetic and particle energies add to @xmath83erg @xcite . this is smaller than @xmath84erg , suggesting that @xmath85 ( see * ? ? ?
all these ideas are outlined more fully for / by @xcite , including a discussion of the very significant differences with psr j0205 + 6449 and its pwn 3c 58 , despite comparable spin parameters .
all these considerations point in the case of / to a system that , while not necessarily quite as young as / , may have an age of no more than about 23kyr . in turn , for spin evolution under constant magnetic moment with braking index in the observed range 23 @xcite , this would imply a birth period of @xmath86ms .
the tev @xmath11-ray emission observed from both and is most likely due to inverse compton scattering of relativistic pulsar wind electrons .
the seed photons for such scattering in general arise from dust , the cosmic microwave background , and star light .
if is located close to the gc , the main contribution to the photon field in likely originates in star light @xcite , and this could account for much of the greater tev luminosity of compared to .
if on the other hand is located at a substantially greater distance , we would infer an even larger tev luminosity for while the stellar photon field would have lower energy density .
thus , careful modeling of the tev emission from the pwn may help to constrain the distance to the system . in any case
, it may be that much of the observed difference between and arises not so much from intrinsic differences in the particle spectrum injected , respectively , by psrs j17472809 and j18331034 , as from their different environments .
the pulsars themselves have comparable luminosities and are very faint . for , @xmath87mjykpc@xmath32 and @xmath88 @xcite . for , @xmath89mjykpc@xmath32 , and @xmath90 , assuming that is the counterpart ( another potential candidate lies @xmath91 to its north ; * ? ? ?
* ) , which may be tested with the determination of a pulsar position via timing measurements . at gev energies ,
pulsations have already been detected from with the large area telescope on the _ fermi gamma - ray space telescope_. is , at least , at twice the distance , and near the bright gc , and no detection by _ fermi _ has yet been reported .
because of interstellar scattering , is not detectable at 1.4ghz without a much more sensitive observation than is possible in practice ( @xmath92ms , estimated by scaling from @xmath47 ) . until recently ,
there were only four galactic young pulsars known with @xmath93ms ( the crab , j1913 + 1011 , b1951 + 32 in snr ctb 80 , and j2229 + 6114 associated with 3eg j2227 + 6122 ) .
the bright psr j14106132 , with @xmath94ms and @xmath95pc@xmath2 , but very scattered at 1.4ghz and discovered instead at 6ghz @xcite , and the similarly scattered but @xmath96 times fainter , show the importance of doing the utmost in dedicated searches of interesting objects to plumb the depths of the pulsar luminosity distribution and to minimize other biases against detecting short period neutron stars ( see also * ? ? ?
* ; * ? ? ?
* for detection of other highly scattered pulsars ) .
lies within the group of low luminosity young pulsars ( @xmath97mjykpc@xmath32 ) , nearly all of which were discovered in deep directed searches .
these now make up one - third of all known young pulsars ( see discussion in * ? ? ?
* ) , and there is little reason to suppose that such pulsars are intrinsically rare .
the future , if not bright , appears promising .
the gbt is operated by the national radio astronomy observatory , a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
we thank the staff at nrao for quickly approving and scheduling the rapid response science request to confirm the pulsar .
the parkes observatory is part of the australia telescope , which is funded by the commonwealth of australia for operation as a national facility managed by csiro .
this work was supported in part by the nsf through grant ast-0908386 .
b.m.g . acknowledges the support of a federation fellowship from the australian research council through grant ff0561298 .
is partially supported by west virginia epscor through a research challenge grant .
djannati - atai , a. , de jager , o. c. , terrier , r. , gallant , y. a. , & hoppe , s. 2008 , in proc .
30th icrc , vol . 2 , ed .
r. caballero , j. c. dolivo , g. medina - tanco , l. nellen , f. a. snchez , & j. f. valds - galicia ( mexico city : unam ) , 823 | the supernova remnant has long been inferred to contain a central energetic pulsar . in observations with the nrao green bank telescope at 2ghz ,
we have detected radio pulsations from .
the pulsar has a rotation period of 52ms , and a spin - down luminosity of @xmath0 , the second largest among known galactic pulsars . with a dispersion measure of @xmath1pc@xmath2 , is distant , at @xmath3kpc according to the ne2001 electron density model , although it could be located as close as the galactic center .
the pulse profile is greatly scatter - broadened at a frequency of 2ghz , so that it is effectively undetectable at 1.4ghz , and is very faint , with period - averaged flux density of @xmath4jy at 2ghz . |
the most basic classification of active galactic nuclei ( agn ) consists in dividing them in two classes : radio - quiet and radio - loud .
this represents not only an observational distinction but a basic physical difference whose basis is still not understood .
radio morphologies ( lobes , jets etc . ) are obviously the product of some physical mechanisms at work in the nuclei of radio - loud objects .
rees et al .
( 1982 ) suggested , for example , that the primary engine in radio galaxies was a spinning black hole fed by thick , and hot , accretion flow .
later , blandford ( 1990 ) and meier ( 1999 ) have explicitly identified the black hole spin as a possible physical parameter responsible for the radio - loud and radio - quiet dichotomy .
if the hole rotates faster , it is more efficient in producing the jets observed in radio - loud objects .
x - ray photons seem to be the best probe to investigate the radio loudness issue as they are produced ( and reprocessed ) in the inner regions of agns where accretion occurs .
x - ray observations of radio - loud objects therefore provide unique data to test the model hypotheses .
in particular , two questions can be addressed : 1 ) does the jet contribute to the x - ray emission in non - blazar radio - loud agns ? 2 ) are the same accretion processes at work in radio galaxies ( and in general in radio - loud agns ) and seyfert galaxies ?
bepposax is a broad band ( 0.1 - 200 kev ) x - ray satellite ( scarsi 1993 ) which has ideal characteristics for this purpose . here
we present a bepposax analysis of 6 broad line radio galaxies and compare our sources with a sample of 13 seyfert 1 galaxies also observed by bepposax ( matt these proceedings , hereafter m00 ) . based on the properties of optical - uv spectra , blrgs are considered the radio - loud counterpart of seyfert 1s ( urry and padovani 1995 ) .
the comparison of the x - ray ( nuclear ) properties of these two classes of objects is then particularly appropriate to investigate the radio - loud and radio - quiet dichotomy .
the radio galaxies presented here are part of a larger sample of radio - loud emission - line agns observed by bepposax , whose analysis is still on - going . in this paper , we restrict the discussion to a subsample of 6 blrgs , which can be directly compared with the sample of seyfert 1s discussed by m00 .
the sample is shown in table 1 .
all the 3c sources have been observed for about 100 ksec .
pks2152 - 69 and pictor a were observed for @xmath0 and @xmath1 ksec , respectively .
lcccc & & & & + source & z & n@xmath2 & radio@xmath3 & l@xmath4 + & & 10@xmath5 @xmath6 & morphology & @xmath7 + & & & & + pks2152 - 69 & 0.027 & 0.25 & fri / frii & 0.2 + pictor a & 0.035 & 0.42 & frii & 1.0 + 3c120 & 0.033 & 1.1 & fri / s & 2.4 + 3c111 & 0.048 & 3.2 & frii / s & 2.9 + 3c390.3 & 0.057 & 0.42 & frii / s & 3.3 + 3c382 & 0.059 & 0.88 & frii & 9.4 + + + + the spectral analysis results , obtained by simultaneously fitting the lecs ( 0.1 - 4 kev ) mecs ( 1.5 - 10 kev ) and pds ( 15 - 100 kev ) instruments are reported in table 2 .
a simple power law plus galactic absorption gave good fits to the data of pks2152 - 69 and pictor a. the iron line is absent in pictor a in agreement with the asca measurement ( eracleous et al .
1998 ) . for pks2152 - 69 ,
the exposure time was too short to allow the detection of even a strong feature . for both , we could obtain only an upper limit for the iron line equivalent width . in 3c111
the continuum was also a simple power law and the iron line only marginally detected ( in spite of the long exposure time ) .
the soft photons appeared strongly absorbed .
it is however possible that the cold absorber is not intrinsic to the source but associated to our galaxy .
3c111 is in fact behind the galactic dark cloud , taurus b ( see reynolds et al .
1998 for more details ) .
more complex models were necessary to fit the other radio galaxies .
absorption in excess of the galactic column density , iron lines , reflection ( ref . ) humps , soft excesses and bendings of the high energy spectrum ( cutoff ) are common spectral features , although not always simultaneously present in each source .
lccccccc & & & & & & & + source & @xmath8 & n@xmath9 & refl .
& cutoff & e@xmath10 & @xmath11 & ew + & & & & kev & kev & kev & ev + & & & & & & & + pks2152 - 69@xmath12 & 1.79@xmath13 & 0.25 ( f ) & ... & ... & 6.4 ( f ) & 0.0 ( f ) & @xmath14 + pictor a & 1.63@xmath15 & 0.42 ( f ) & ... & ... & 6.4 ( f ) & 0 ( f ) & @xmath16 + 3c120@xmath17 & 1.80@xmath18 & 2.4@xmath19 & 0.7@xmath20 & 109@xmath21 & 6.2@xmath22 & 0.27@xmath23 & 59@xmath24 + 3c111@xmath25 & 1.65@xmath26 & 7.1@xmath27 & @xmath28 & @xmath29 & 6.6@xmath30 & 0 ( fixed ) & 58@xmath31 + 3c390.3 & 1.80@xmath32 & 1.3@xmath33 & 1.2@xmath34 & @xmath35&6.4@xmath13 & 0.07@xmath36 & 136@xmath37 + 3c382@xmath17 & 1.79@xmath26 & 0.88 ( f ) & 0.4@xmath38 & 155@xmath39 & 6.5@xmath40 & 0.00 ( f ) & 31@xmath41 + + & & & & & & & + + + the soft excess was detected in two sources , namely 3c120 and 3c382 .
it was parameterized with a steep power law ( @xmath42 ) and contributed to the total emission in the 0.1 - 2 kev band more than 60@xmath43 in both the sources .
the origin of this excess is not clear .
it could be related to thermal emission from a cold thin disk , to radiation coming by extended thin plasma as well as to a very soft jet emission .
all the blrgs with detected iron line show reflection components . although the strength of the reflection ( ref . )
is not very well constrained , data seem to suggest that radio - galaxies with weak iron lines have also weak reflections ( figure 1 ) .
if confirmed , this would imply that the line is generated in the same material which reflects the x - ray primary photons .
we compared the spectral properties of our blrgs with a sample of 13 seyfert 1s observed by bepposax ( m00 ) .
the 2 - 10 kev luminosities of seyferts ( ranging from @xmath44 to @xmath45 erg sec@xmath46 ) and blrgs ( see table 1 ) partially overlap .
the x - ray primary power law of blrgs is generally flat ( @xmath47 , rms dispersion @xmath48 ) .
the seyfert 1 sample is characterized by a steeper average spectral slope .
however , the larger spread of the spectral indices in radio - quiet agns ( @xmath49=1.87 , @xmath50 ) , does not allow to statistically confirm any difference between seyfert 1s and blrgs .
half the 3c sources in our sample shows a bending of the x - ray spectrum at high energies .
when it is modeled with an exponential cutoff , the cutoff energies are similar to those observed in seyferts . while in radio - quiets the iron lines and the reflection components are always present , in blrgs
the reprocessed features are detected in only 3 sources .
it should be noted the absence of the lines can not be only attributed to poor statistics .
in addition in radio - louds , the iron line equivalent widths are significantly smaller than in radio - quiets ( @xmath51 ev , @xmath52 ev ; @xmath53 ev , @xmath54 ev ) . a general weakness of the reprocessed features in blrgs
is also confirmed by the xte results presented by sambruna and eracleous ( these proceedings ) .
note that wozniak et al .
( 1998 ) have already discussed this possibility analyzing asca , ginga and osse data . a cold absorbing column in excess of galactic is observed in 3 ( including 3c111 ) out of 5 blrgs with lecs data .
no source shows absorption edges typical of warm absorber , which , on the contrary , is rather common in seyfert 1s .
it is possible that the absorbing material is different in blrgs and seyfert 1s , being warm in radio - quiets and cold in radio - louds ( see also sambruna eracleous and mushotzky 1999 ) .
this is also supported by a historical study of the column density changes in 3c390.3 .
the long - time ( years ) variability of the intrinsic n@xmath55 does not appear to be correlated to the flux intensity at 1 kev ( see fig . 2 in grandi et al .
it is possible that variations in the geometry of the absorber rather than changes of its ionization state ( as expected in the case of a warm absorber ) are responsible for the n@xmath55 long term variability .
two important points arise from the bepposax study of blrgs : 1 ) there is not a unique type of blrg x - ray spectrum , but a variety of cases ; 2 ) there are several differences between blrgs and seyfert 1 galaxies .
the most impressive difference concerns the reprocessed features that are weaker in radio - louds than in radio - quiets ( fig .
1 ) . how can these results be explained ?
blrgs are complex objects , in which at least three x - ray components , a jet , an accretion flow and a molecular torus , can contribute to the formation of the spectrum .
* jet * in some blrgs the radio jet shows superluminal motion .
it is then reasonable to suppose that doppler - enhanced radiation also contaminates the x - ray spectra and dilutes the reprocessed spectral features when the doppler factor of the jet ( @xmath56=[@xmath57(1-@xmath58 cos)]@xmath46 ) is sufficiently large . if , in agreement with the agn unified schemes , blrgs and intrinsically powerful blazars are the same objects seen at different angles of view , also the jet output of radio galaxies should be inverse compton in the pds band ( fossati et al .
1998 , ghisellini 1998 ) . if this is the case , it is rather improbable that the observed high energy steepening in blrgs is related to the jet . in powerful blazars , the compton break ( i.e. the compton peak in the spectral energy distribution , sed )
is usually revealed at mev - gev energies . in order to occur at 100 - 200 kev
as observed in blrgs , the doppler factor ( @xmath56=[@xmath57(1-@xmath58 cos)]@xmath46 ) should be smaller by about a factor 100 or more .
this would imply a strong de - amplification of the non - thermal radiation ( @xmath59(@xmath60)=@xmath61i@xmath62(@xmath63 ) and , in turn , a difficult detection of the jet . * accretion flow * it is not really clear how gas accretion occurs in radio - loud objects .
the simplest approach is to extend the physical models developed for seyferts to broad line radio galaxies . as in the case of radio - quiet agns ,
the accretion gas flow could be a cold geometrically thin disk with a hot corona above it ( haardt and maraschi 1991 , 1993 ; petrucci et al .
the role of the corona is to transform into x - ray photons , via inverse compton scattering , the uv photons generated by the disk .
down - scattered x - rays , in turn , hit the disk and are reprocessed , generating an iron line and a reflection hump above 10 kev .
a seyfert - like accretion disk should then produce the following signatures in x - ray blrg spectra : a soft excess ( related to the thermal disk emission ) , a high energy thermal cutoff ( related to the corona temperature ) , a broad red - shifted iron line ( which suffers the vicinity of the strong black hole gravitational field ) and a strong reflection component ( produced by a disk subtending a @xmath64 solid angle to the x - ray primary source , i.e. refl@xmath65 ) .
alternatively , one can assume that different physical / geometrical accretion configurations occur in agns which discharge large amount of energy in jets .
rees et al .
( 1982 ) speculated on the possibility that in radio - galaxies the hot accreting gas is in the shape of an ion - supported torus characterized by low radiative efficiency .
the advection dominated accretion flow ( adaf ) models ( narayan e al . 1998 ) , recently proposed , follow similar lines of though .
shapiro lightman and eardley ( sle , 1976 ) found another solution for the hot flow , that resembles the ion - supported torus .
however , in the sle model the energy produced in the accreting gas by viscosity is locally radiated and not advected radially .
then the accretion flow in blrgs could be hot and geometrically thick in the inner region and become cold and geometrically thin only at larger radii ( chen and halpern 1989 ) . given the smaller covering factor of the cold ( and reprocessing ) matter to the x - ray primary source , this accretion configuration predicts less prominent soft excesses , narrower weaker iron lines and smaller reflection components than the seyfert case .
in addition , a lack of correlation between the variations of the primary x - ray source and the reprocessed features is expected ; the entity of the temporal delay depending on the inner radius of the cold disk .
* molecular torus * it is probable that a thick wall of absorbing material ( perhaps in a toroidal shape ) surrounds the accretion flow .
the torus can produce the iron line and reflect the x - ray primary photons ( ghisellini haardt and matt 1994 ) further complicating the x - ray spectrum .
in addition , if the opening angle of the torus is small , its ( less thick ) upper layers could be intercepted by the observer and cause the soft x - ray depletion sometimes observed in blrgs . since
its distance from the x - ray source is large ( @xmath66 pc ) , iron line and reflection components should respond with a considerable delay ( light - years ) to the continuum variations .
it is possible and also probable that all these components are present in blrgs . if they are mixed in different ways in different objects , the variety of x - ray spectra observed with bepposax would be easily explained . as shown in table 3 the main bepposax spectral features of the blrgs with high signal - to noise spectra
( i.e. pointed for about 100 ksec ) can be opportunely reproduced choosing a plausible combination of the nuclear components .
.blrg nuclear components [ cols="^,^,^ " , ] this tentative table shows that a seyfert - like disk alone can not reproduce the observations .
the weakness of the reprocessed features requires the presence either of a non - thermal radiation from a jet or of a cold disk sub - tending a small solid angle to the x - ray primary source ( see figure 2 ) .
the sources in table 3 which seem better to fit the seyfert - like+jet model are 3c111 and 3c120 .
if the model is correct , the weak ( but detected ) iron line in 3c120 assures that seyfert and blazar components have comparable intensity around @xmath67 kev , i.e. the non - thermal power law is important and can effect the continuum shape .
we then tested whether the 3c120 spectrum can be fitted by a mix of beamed and un - beamed radiation on the entire bepposax band ( i.e. on about 3 decades in energies ) .
a simple power law was assumed to represent the jet ( blazar ) spectrum and a power law with cutoff plus iron line and reflection component were utilized to mimic the seyfert emission .
the seyfert equivalent width was assumed ew@xmath68 ev , the average value from the matt sample .
different combinations of spectral slopes for the seyfert and blazar models were tested . in all the cases , the relative normalization at 6.4 kev of the two power laws
was fixed in order to reproduce the observed fe equivalent width reported in table 2 .
as shown in figure 3 ( _ left panel _ ) , 3c120 data can be fitted by a blazar and a seyfert - like power law ( in figure @xmath69 , @xmath70 ) .
however the model requires very low energy cutoff ( @xmath71 kev ) and a strong reflection ( ref@xmath72 ) .
these best fit values are rather extreme if compared to those of seyferts .
this slightly disfavors the jet model , although it can not be completely rejected given the large uncertainties associated to the parameters . a geometry that assumes a hot inner flow and cold disk surrounding it seems to be particularly appropriate for 3c390.3 .
in this source the uv bump is weak , there is not soft excess and the iron line is narrow . in this case a molecular torus ( or a warped disk ) also contributes to increase the strength of the iron line and the reflection component ( grandi et al .
the idea that the iron line emitting regions can not be located near the x - ray is also supported by recent bepposax observations of the very bright radio - galaxy centaurus a ( cen a ) ( @xmath73 erg @xmath6 sec@xmath46 ) .
it is a fr i optically classified as narrow line radio galaxy and shows a radio - optically - x - ray jet pointed far away from us ( @xmath74 ) .
it is characterized by a x - ray spectrum rather complex . however above 3 kev
it is dominated by the nuclear point - like source ( turner et al .
1998 ) .
cen a was observed on 1997 february 20 - 21 ( 30 ksec ) , on 1998 january 6 - 7 ( 50 ksec ) and twice for 40 ksec during the 1999 summer on july 10 - 11 and august 2 - 3 .
the repeated observations of this radio - galaxy , thanks to their high statistics , have allowed a detailed study of the nuclear flux variations versus the iron line flux changes . as clearly shown in figure 3 (
_ right panel _ ) , line and continuum do not vary together , in particular the iron line was more intense when the source was weaker ( compare the 1997 and 1998 observations ) .
since the inclination of the jet is large in cen a , the contribution of non - thermal radiation to the total x - ray nuclear continuum is expected to be negligible .
then figure 3 ( _ right panel _ ) simply indicates that the line emitting region responds with a significant delay to the continuum variations , as it is expected if the reprocessing gas is located far from the primary x - ray source .
we conclude that , although , in principle , the two proposed accretion scenarios are both viable , the idea that a jet reduces the reprocessed features is lightly disfavored .
bepposax analysis of 6 blrgs has shown that a variety of x - ray spectra exits . a mix of different components , a jet , an accretion flow and a molecular torus , could explain the observations .
the comparison between our 6 blrg and a sample of 13 seyfert 1s has pointed out important differences between radio - loud and radio - quiet agns . in particular
, it has been shown that blrg reprocessed features are often absent or weaker than in seyfert 1s .
this result has important implications .
if the accretion flow in blrgs and seyferts is identical ( i.e a cold thin optically thin disk with a hot above it ) a strong contribution by doppler - enhanced radiation is necessary to explain the weakness of the reprocessed features .
alternatively , if the jet emission is not important , the cold re - processing gas has to subtend a smaller solid angle to the x - ray primary source .
a possibility is that the x - ray continuum is produced by a hot geometrically thick ion - supported torus that illuminates a cold thin disk at larger radii .
choosing between the two scenarios is still premature . the new satellites ( xmm , chandra , integral ) will play an important role in solving the problem . detailed studies of the iron line profiles in a large ( and well selected ) sample will allow to confirm the presence of cold matter very near to the black hole ( if lines are broad and redshifted ) or at large distances ( if lines are narrow ) .
variability studies will be also crucial .
an anti - correlation is expected between the ew and the x - ray continuum flux , if the jet is dominant : the observed continuum ( disk + jet ) should be more variable than the fe line ( from disk ) .
alternatively , if no correlation is found , the fe variations do not follow the continuum flux variations and a temporal delay effect ( as observed in cena case ) is present .
finally , detailed studies of the hard x - ray continuum will be able to detect any beamed radiation . at very high energies , where the seyfert - like power law drops , the jet , if intense , should emerge and become directly detectable ( see figure 3 _ left panel _ ) .
i would like to thank all the people who , working with me on the bepposax radio - galaxy project , have allowed the writing of this paper : l. maraschi , c. m. urry , e. massaro , g. matt , m. guainazzi , f. haardt , p. giommi , g.g .
palumbo , g. malaguti .
an acknowledgment also go the l. piro and the bepposax cdc team for their support to this project .
i also thank g. ghisellini , l. ferretti and c. g. perola for the useful comments and stimulating discussions and a. bazzano and a.j .
bird for critical reading of the manuscript .
i am very grateful to m. frutti for invaluable help in realizing figure 2 . | bepposax observations of broad line radio galaxies ( blrgs ) have shown that they have a considerable variety of spectral properties and important differences with respect to their radio - quiet counter - part seyfert 1s . in radio galaxies the soft photons
are often absorbed by cold material .
in contrast , in seyfert 1s the soft photons are generally absorbed by warm gas .
the iron lines , always detected in seyfert 1s , are not always present in blrgs and generally are weak .
in addition , small iron line equivalent widths seem to correspond to weak reflection components in radio galaxies .
the emerging picture of blrgs is complex .
probably several x - ray components , jet , accretion flow and molecular torus , mixed in different way in different objects , contribute to the production of their x - ray spectrum and determine the observed variety .
the weakness of the reprocessed features can be explained either by a dilution of the seyfert - like continuum from non - thermal ( jet ) radiation or by an accretion gas that is hot and geometrically thick close to the black hole and cold geometrically thin ( i.e. able to reprocess the primary x - ray radiation ) at larger radii . |
the unified model for seyfert galaxies predicts that the differences observed between type 1 and type 2 seyfert galaxies are primarily due to orientation effects @xcite . according to this model
, there is a geometrically thick torus surrounding the accretion disk around the central supermassive black hole .
type 1 seyfert galaxies are those active galactic nuclei ( agn ) seen at low inclination angles ( i.e. nearly face on ) which provide an unobstructed view of the inner central region .
broad optical emission lines , which originate close to the black holes , are observed along with the narrow ones , originating further from the source . in the x - ray regime ,
the spectra is typically a power law one ( showing sometimes a soft excess ) and affected mostly by absorption in our own galaxy .
in contrast , type 2 seyfert are those agn which are observed at high inclination angles ( i.e. nearly edge on ) and hence their central engine is obscured by the dusty torus .
the broad optical lines are not observed here but the narrow ones which originate high above in the axis of the torus , are detected .
their x - ray spectra is highly absorbed at low energies implying that the torus has a neutral column density in the range of @xmath11-@xmath12 @xmath5 .
the environment around central engine of agn is expected to be fairly complex .
for example ,
partially ionized , optical thin gas along our line of sight into the nuclei of several agn can absorb the soft x - rays .
this partially ionized material has come to be known as the `` warm absorber '' @xcite .
the geometry and the position of these warm absorbers are uncertain .
some agn show evidence for high - velocity outflows ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
moreover with the aid of high resolution , narrow band imaging ( or integral field spectroscopy ) and _ hst _ imaging , it was possible to identify extended regions of highly ionized material around the central engine also known as ` ionization cones ' with a size of a few up to 15@xmath1320 kpc which suggests an anisotropic escape of photons from agn confined to a cone by a dusty torus @xcite .
such ionization cones were studied in ngc 1068 and in several other seyfert 2 galaxies progressively using dispersive grating spectroscopy onboard _
chandra _ and _ xmm - newton _ ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
recent work on seyfert 2 ngc 4945 based on _ xmm - newton _ , _ suzaku _ , _ swift_-bat and _ chandra
_ data has revealed detailed characteristics of the circumnuclear reflector .
the study implies the distance of the reflector @xmath14 35@xmath1350 pc and the _ chandra _ imaging suggests a resolved , flattened , @xmath15 pc long clumpy structure , whose spectrum is explained by the cold reflection of the primary agn emission @xcite .
constraining the location and geometry , knowledge of the state and dynamics of these different components that make up the environment around agn , is important to understand how the central engine accretes matter and any feedback the agn may have in terms of matter outflow to the galaxy .
an interesting aspect of several seyfert 2 agn is the presence of an excess power law like emission at low energies ( @xmath16 kev ) , which may be due to scattered emission @xcite . in this scenario ,
the direct low energy photons of the source are absorbed by the torus .
there is an extended scattering medium along the axis of the torus , which scatters some of the photons to our line of sight .
significant amount of work has been done using _ asca _ , _ bepposax _ , _ xmm - newton _ and _ suzaku _ data , to address the nature of the scattering medium and its implications towards understanding the excess soft x - ray emission and geometrical structure in seyfert 2 agn .
for instance , @xcite studied in detail the importance of x - ray scattering and reflection for a sample of 25 seyfert 2 agn using _ asca _ data .
they found a case of mrk 3 whose spectrum could be explained by the presence of hot scattering gas in the soft x - ray regime with scattering fraction cover 0.02%@xmath135% . a similar study using
_ asca _ investigated the physical conditions of the scattering material in a sample of scattering - dominated agn .
the study suggested that the soft x - rays represent scattered light , similar to optical polarized broad lines @xcite .
@xcite explained the soft x - ray continuum of ngc 1068 and circinus galaxy observed by _ bepposax _ due to the combined presence of a scattered power law emission and an optically thin plasma emission and estimated column densities of the warm scatterer as @xmath17 10@xmath18 @xmath5 and @xmath7 a few @xmath19 10@xmath20 @xmath5 respectively .
@xcite studied a sample of six seyfert 2 galaxies with optical polarized broad lines using _ asca _ data .
the work revealed significant soft x - ray emission owing to the scattered light from their nuclear emission .
they further estimated a few percent larger scattering efficiency than found for seyfert 2 agn without optical polarized broad lines .
@xcite modeled the scattering emission as a secondary power law component in the _ asca _ spectra of a sample of eight seyfert 2 galaxies and a warm scattering medium in ngc 4151 was observed by @xcite .
_ suzaku _ broadband observations of obscured seyfert galaxy mcg @xmath21 revealed a soft excess below 1 kev which was explained by emission from scattered continuum photons and distant photoionized gas .
it was deduced that this source was viewed at moderate ( @xmath22 ) inclination through compton - thin matter at the edge of a compton - thick torus covering @xmath23 steradians , consistent with unified models @xcite .
further studies with _ suzaku _ showed that for two sources the scattered light constituted a small fraction ( @xmath24 ) of the nuclear emission @xcite .
this suggested a scenario in which there was a geometrically thick torus with a small opening angle and/or very small amount of gas responsible for scattering .
another detailed work studied the wide band _
suzaku _ spectrum of markarian 3 which was resolved into weak , soft power - law emission , a heavily absorbed power law component , cold reflection , and emission lines .
the weak , soft power law emission was considered to be scattered light by ionized gas with a scattering fraction of @xmath25 @xcite . the possibility of scattering from an ionized medium was discussed in the _ suzaku _ analysis of seyfert 2 ngc 4945 @xcite .
a comprehensive study by @xcite using _ xmm - newton _ observations of 32 obscured agn investigated their multi - wavelength properties in relation to the scattering fraction .
the sample covered a broad range of the scattering fraction ( @xmath26 ) ; with eight agn exhibiting low scattering fraction ( @xmath27 ) suggesting that they are buried in a geometrically thick torus with a very small opening angle .
they found no significant correlation between scattering fraction and far - infrared luminosity and a weak anti - correlation between the eddington ratio and scattering fraction .
the variability of the scattered component as compared to the primary emission can provide constraints on the geometry of the scatterer .
x - ray variability studies of mrk 3 demonstrated that unlike the hard x - ray emission , the soft x - rays did not vary over @xmath713 year baseline suggesting its origin from an extended region @xcite .
@xcite also studied mrk 3 and found no variability in the soft component over a 10 year long period although the hard component changed by a factor of 6 during 3 years .
it was suggested that if most of the soft x - rays from mrk 3 were scattered , the large time lag indicated that the distance of the scatterer from the nucleus is comparable to 6 lt - yr ( @xmath28 pc ) .
this result supported a large scattering region as also presented by @xcite .
iras 18325 - 5926 is an iras selected seyfert 2 galaxy at a redshift @xmath29 .
it was first observed in x - rays by _ ginga _
@xcite which showed that the spectral index of the source @xmath30 is steeper than normal .
the _ asca _
observation of this source revealed a quasi - periodic modulation in @xmath31 hours timescales @xcite which was not seen in later _
rxte _ or _ beppo - sax _
lightcurves @xcite .
it is one of the few agn which shows a clear broad iron line @xcite which is possibly due to reflection of an highly ionized disk @xcite .
_ beppo - sax _ observation of this source hinted at a possible roll - over at @xmath32 kev @xcite .
recently , @xcite reported the detection of warm absorbers in this source with _
hetgs spectra .
they found an intrinsic absorbing line system with an outflow velocity @xmath33 kms@xmath34 which is contributed by two warm absorbers with outflowing velocities of 340@xmath35110 kms@xmath34 and 460@xmath35220 kms@xmath34 , respectively .
@xcite also found warm absorbers ; and detected the absorption features ( in the form of a broad trough ) in the vicinity of the @xmath36 kev iron k edge .
iras 18325 - 5926 also shows a soft component which may be due to a scattered component @xcite .
the scattered emission can only respond to variations in the intrinsic photons on a timescale longer than the light crossing one .
it is fortuitous that the hard ( @xmath37 kev ) x - ray spectrum of iras 18325 - 5926 shows rapid variations on timescales @xmath38 s and hence is an ideal source where the size of the scattering emission can be constrained .
here , we analyze _ xmm - newton _ and _ suzaku _ observations of the source to investigate the timescale on which the scattered emission responds to the primary one .
the fraction of scattered photons is related to the total column density of the medium and since the scattered photons are not absorbed , the scattering medium should be highly ionized as we discuss in this work . hence
, these physical constraints allow us to estimate the the size and density of the scattering medium . in the next section
we describe the _ xmm - newton _ and _ suzaku _ observations and the details of data reduction process . in 3 , the time - averaged spectra of the source are modeled and in 4 the spectral variability in different timescales are investigated . in 5 , the size and density of the scattering medium are estimated and the results and conclusions are presented in 6 .
ccc & obsid : 0022940101 ( _ xmm - newton _ ) + epic - pn & @xmath39 & @xmath40 + + mos1 & @xmath41 & @xmath42 + + mos2 & @xmath43 & @xmath44 + & obsid : 0022940201 ( _ xmm - newton _ ) + mos1 & @xmath45 & @xmath46 + + mos2 & @xmath47 & @xmath46 + & obsid : 702118010 ( _ suzaku _ ) + pin & @xmath48 & @xmath49 + + xis 0 & @xmath50 & @xmath51 + + xis 1 & @xmath52 & @xmath53 + + xis 3 & @xmath52 & @xmath54 [ obs_det ]
two observations from _ xmm - newton _ during 2001 march 5 ( obsid : 0022940101 ) and march 6 ( obsid : 0022940201 ) and one observation from _ suzaku _ 2007 october 10 ( obsid : 702118010 ) have been analyzed in this work .
details of the observations are shown in table [ obs_det ] .
the _ epic pn _ and _ mos _ cameras were operated in the small window mode using medium filter .
data was reduced using version 12.0 of the scientific analysis software ( sas ) .
the exposure time for _ epic - pn _ camera was too small for first observation ( obsid : 0022940101 ) and the camera was not in the working mode for the second observation ( obsid : 0022940201 ) , therefore analysis were done only with the two _ epic - mos _ cameras .
the correction for background flaring is done by setting a threshold count level of 1.5 counts @xmath55 for the first observation . for the second _ xmm - newton _ observation ,
the background flares were excluded by choosing appropriate time filters which resulted in @xmath56 ks data for each _ mos _ camera . for spectral analysis ,
the data was filtered to the good x - ray events ( flag==0 ) with pattern @xmath57 12 in the energy range 0.2 - 15 kev . to extract the source spectra a circular region of radius 40@xmath58 for both _
mos1 _ and _ mos2 _ centered at the peak position of iras 18325 - 5926 were used .
the background spectra was extracted from a nearby circular region , free of sources .
the source spectra were re - binned with a minimum of 100 counts per channel .
the re - binned data were analyzed using _
xspec _ in the energy range @xmath59 . for both observations
, it was verified that the _ mos1 _ and _ mos2 _ spectra were consistent with each other and hence all spectral fittings were done on the combined data .
of the four xis instruments only three of them were in the operating mode .
so observations were taken from three xis ( xis0 , xis1 and xis3 ) and the hxd ( pin ) sensors .
the xis was in normal clocking mode .
_ suzaku _ _ ftool xisrmfgen _ and _ xissimarfgen _ is used to generate response matrices and ancillary response files respectively . for the extraction of xis spectra a circular region of @xmath60 radius centered on the source
were used , while backgrounds were extracted from the outer annulus region free of sources .
we use an appropriate scaling factor for the background to take into account the different extraction areas of the source and the background .
the source spectrum was re - binned with a minimum of 300 counts per channel for all the three xis sensors and 60 counts per channel for hxd(pin ) .
package was used for spectral analysis .
the energy range was set to @xmath61 for xis and @xmath62 for hxd .
the normalization factor between xis and hid spectra was fixed to 1.18 .
a sharp absorption line around @xmath63 was seen in the 0.6 - 10 kev energy band .
it was found that this line feature is not present in all the xis instruments .
close examination of the line reveals that the line is at the silicon k - edge and the feature could be due to incorrect response calculation .
so the energy range @xmath64 is not considered for spectral fitting . apart from this feature it was verified that all the xis spectra were consistent with each other .
so all the xis spectra were combined for the analysis .
the figure has been rebinned for display purpose only . ] to understand the spectra of the source , all three data sets were initially fitted by an an absorbed power law model .
galactic absorption was fixed to column density of @xmath65 @xmath5 . to this
an additional neutral absorption at the redshift of source was included . the first _ xmm - newton _ data with best fit model and their ratio are plotted in fig [ soft_exc ] .
the ratio plot clearly shows two excesses .
one at low energies @xmath66 which could be due to a scattered component and other at @xmath67 due to the k@xmath68 iron line emission . the iron line emission feature in both the _ xmm - newton _ and the _ suzaku _ observations can be fitted by a broad gaussian feature .
the centroid energy is at @xmath69 and is broad with @xmath70 which is consistent with the _ asca _ observations @xcite .
the width of the line suggests a relativistic origin for the line . the relativistic iron line model in _ xspec _ , `` diskline '' can also represent the feature .
however , the line is too weak to constrain most of the parameters .
for example , the first observation of _ xmm - newton _ , provides a line energy of @xmath71 and an inclination angle of @xmath72 degrees , if the other parameters ( i.e. the emissivity index @xmath73 , inner @xmath74 and outer radius @xmath75 ) are fixed at standard values . the line energy is less than what is expected from a highly ionized disk ( @xmath76 ) as modeled by @xcite
however , the present data is not good enough to undertake a detailed analysis of the complex iron line feature .
further complications are evident by the presence of an absorption edge at @xmath77 in the second _ xmm - newton _ observation .
thus , in this work we model the iron line feature using the `` diskline '' model and concentrate instead on the low energy excess .
the low energy excess is a broad feature and can not be due to presence of narrow emission lines .
we have checked the rgs data and confirmed that there are no strong narrow emission lines in the soft band as also revealed in the _ chandra _ observation @xcite .
instead , it can be modeled as an additional unabsorbed ( i.e. only affected by the galactic absorption ) power law with the spectral index tied to that of the intrinsic one .
the normalization of this additional power law is the only free parameter in the fitting @xcite .
the best fit spectral parameters are tabulated in table [ spec_par ] and the spectra with best - fit model and residuals are shown in fig .
[ spec_xmm_suz ] and fig .
[ spec_xmm_suz2 ] for the first _ xmm - newton _ and _ suzaku _ observations respectively .
the spectral fits are reasonable with no obvious systematic residuals .
lccc[htp ] & & & + @xmath78 & @xmath79 & @xmath80 & @xmath81 + @xmath82 ( kev ) & & @xmath83 &
+ @xmath84 & - & @xmath85 & + + @xmath86 ( kev ) & @xmath87 & @xmath88 & @xmath89 + @xmath90 & @xmath91 & @xmath92 & @xmath93(f ) + @xmath94 ( @xmath95 ) & @xmath96 ( f ) & @xmath96(f ) & @xmath97 + @xmath98 & @xmath99 & @xmath100 & @xmath101 + + @xmath102 & @xmath103 & @xmath104 & @xmath105 + @xmath106 & @xmath107 & @xmath108 & @xmath109 + + @xmath110 & @xmath111 & @xmath112 & @xmath113 + + @xmath114 & @xmath115 & @xmath116 & @xmath117 + [ spec_par ]
as expected , table [ spec_par ] shows that the spectrum of iras 18325 - 5926 was different in the _ suzaku _ observation of october 2007 as compared to the _ xmm - newton _ ones of march 2003 . in particular , the normalization of the primary power law changed by nearly a factor of two .
the normalization of the scattered power law emission also reveal a significant increase by @xmath118% which shows that on timescales of years the scattered emission is variable , most probably in response to the primary continuum .
the ratio of the scattered to the primary continuum also changed from @xmath119 during 2003 to @xmath120 in 2007 , suggesting that the scattering medium itself is variable on such long timescales .
the two _ xmm - newton _ observations are about a day apart and table [ spec_par ] reveals that the normalization of the primary power law does vary on this timescale , by increasing from @xmath121 to @xmath122 photons @xmath5 s@xmath34 .
however the normalization of the scattered component does not respond to this variation on @xmath123 s , remaining nearly constant at @xmath124 photons @xmath5 s@xmath34 .
the source shows significant variability on a shorter timescale of @xmath125 s , as is evident in the lightcurve of one of the _ xmm - newton _ observations shown in fig [ xmmlightcurves ] .
-@xmath126 kev for obsid 0022940101 . the solid horizontal line at @xmath127 count / s is used to divide the lightcurve into low and high flux states .
] we investigate whether the scattered continuum is variable on this timescale by dividing the data into two states defined as high ( count rate > 1 counts / s ) and low ( count rate < 1 counts / s ) flux levels .
spectral fitting of each flux level for the first _ xmm - newton _ observation , reveals that the primary power law normalization varied from @xmath128 to @xmath129 photons @xmath5s@xmath34 .
however the scattered component normalization did not seem to respond to this variation with @xmath130 ( low flux level ) to @xmath131 photons @xmath5s@xmath34 .
a similar result was obtained for the second _ xmm - newton _ observation where the primary component normalization changed from @xmath132 to @xmath133 , but there was no detectable change in the scattered one with @xmath134 ( low flux level ) to @xmath135 photons @xmath5s@xmath34 ( high flux level ) . further evidence that the scattered component does not respond to
the primary one comes from analyzing the variability as a function of energy . for this
, we extracted lightcurves for xmm - newton observations in the energy bands 0.3 - 1.0 kev , 1.0 - 2.0 kev , 2.0 - 3.5 kev , 3.5 - 5.0 kev and 5.0 - 10.0 kev , respectively and for suzaku observation , 0.3 - 1.68 kev , 1.68 - 3.0 kev , 3.0 - 5.32 kev and 5.32 - 9.5 kev , respectively . for extracting these lightcurves
we have selected the good time intervals excluding the period of high background flares . in fig [ rms ] ,
the fractional root mean square variability amplitude ( @xmath136 ) is plotted against energy for all three data sets .
@xmath136 is the square root of the excess variance which is the measured variance minus the expected variance due to poisson noise .
the measurement error on @xmath136 is computed using the formalism given by @xcite .
for all three data sets there is a sharp decrease in the variability below @xmath137 where the scattered component is a significant fraction of the total spectrum .
this is indicative of a variable primary and a constant scattered component .
one can estimate the variability that must be present in the primary component , @xmath138 to reproduce the observed @xmath136 , as @xmath139\ ] ] where @xmath140 is the total count rate detected in the energy bin and @xmath141 is the count rate in that bin due to the primary component alone .
[ rms ] shows @xmath138 as a function of energy ( triangles with dotted lines as error bars ) .
thus , a nearly energy independent variability of the primary component , @xmath138 with a non - varying scattered component can explain the total observed variability of the source .
( filled circles ) as a function of energy for ( i ) obsid : 0022940101 , ( ii ) 0022940201 of _ xmm - newton _ and ( iii ) xis spectrum of _ suzaku_. also plotted is the fractional variation that the primary component should have ( filled triangles with dotted errorbars ) , @xmath138 , in the absence of any variation in the scattered one . ]
thus the variability of the source indicates that while the primary component varies in a wide range of timescales , the scattered component responds only on timescales longer than a day ( @xmath142 s ) .
since the scattered component does not respond to the continuum on timescales of @xmath142 s , the scattering medium must be larger than @xmath143 cm .
a simple geometry for the medium is a funnel shaped volume along the axis of the accretion disk .
if the opening angle of the cone is @xmath144 , the fraction of photons that will be scattered is @xmath145 , where @xmath146 is the thomson optical depth of the medium .
thus , we can estimate the column density @xmath78 as @xmath147 where @xmath148 is the observed ratio of the normalizations of the primary to the scattered power law .
such a high column density would absorb the soft x - ray photons from the source ( instead of scattering them ) unless the medium is highly ionized . for that to happen
, the ionization parameter @xmath149 must be larger than a critical value @xmath150 .
this naturally puts an upper limit on the size of the region @xmath151 or @xmath152 the x - ray luminosity of the source in the @xmath153-@xmath154 band is @xmath155 ergs s@xmath34 for the _ xmm - newton _ observations assuming @xmath156=70 km s@xmath34 mpc@xmath34 .
the ionizing luminosity @xmath157 which includes photons with energy @xmath158 ev would be about a factor of few higher and is taken as @xmath159 ergs s@xmath34 .
thus the scattering medium can be constrained to have a column density @xmath160 @xmath5 , a size in the range @xmath161 cm and an average number density @xmath162 @xmath10 . in this interpretation ,
the scattering medium is highly ionized and hence if the source was viewed in a lower inclination angle ( like in a seyfert 1 ) , the spectrum will not be absorbed .
however , a small fraction of the gas , ( with column density @xmath163 @xmath5 ) which is furthest away from the source could be partially ionized .
this partial ionized gas could be the origin of the warm absorbers observed in many seyfert 1 .
this would imply that the warm absorber clouds are at a distance of @xmath164 cm from the black hole .
this is interesting since spectral fitting of the warm absorbers can not constrain this distance and there always has been a large uncertainty regarding their location .
the present analysis does not provide information about the dynamic nature of the scattering medium as to whether it is an outflow or inflow .
if we assume the medium to be outflowing in typical speeds of @xmath165 km s@xmath34 , the mass outflow rate can be estimated to be @xmath166 the mass accretion rate in the accretion disk can be estimated from @xmath167 as being @xmath168 g / s for a radiative efficiency , @xmath169 and bolometric luminosity @xmath170 ergs s@xmath34 , which is consistent with the results obtained by _
analysis of this source @xcite . despite the large uncertainties in these estimations
, it seems that the scattering medium would require a mass outflow rate which is a significant fraction of the accretion one . the kinetic power carried by the outflow
@xmath171 ergs s@xmath34 is small compared to the radiative luminosity .
however , these estimates depend upon the information on velocity which is unknown .
an alternate interpretation to the scattering medium described above , is that the secondary component is reflection from a highly ionized inner layer of the torus surrounding the black hole . from the lack of variability , the radius of the torus , @xmath172 cm and
for it to be highly ionized would require , the density @xmath173 .
for the torus to efficiently reflect the x - ray radiation , it must be optically thick to scattering i.e. its column density @xmath174 @xmath5 .
this implies that the width of the torus @xmath175 , making the geometry unlikely .
hence we prefer the scattering geometry above as a more viable scenario .
the spectrum of iras 18325 - 5926 is well modeled by an absorbed power law and a secondary power law component to describe the low energy excess feature @xmath176 in the form of a scattered emission .
we found that the scattered component varies on timescales longer than @xmath1 s unlike the primary emission which varies on a wide range of timescales ( @xmath0 s ) .
this indicates the size of the scattering medium to be greater than @xmath2 cm
. the observed ratio of the scattered to primary flux ( @xmath177 ) gives the column density for the scattering medium to be @xmath7 10@xmath178 @xmath5 . in order to facilitate scattering instead of absorption
, the location of such a scattering medium should be less than @xmath6 cm from the x - ray source for it to be highly ionized .
so we are also able to localize the position and extent of scattering region to be @xmath7 a few times @xmath2 cm , with an average particle density of @xmath9 @xmath10 .
hence , the location of the torus is @xmath179 10@xmath180 cm .
besides , the spectral analysis of two xmm - newton and a suzaku observation also reveal the presence of a broad iron line feature in the source . for seyfert 2 agn , with column density @xmath181 > 10@xmath182 @xmath5
, the matter is optically thick to compton scattering which makes the nucleus almost invisible .
hence the broad iron line can not be detected in this scenario .
however , for seyfert 2 agn with relatively low values of @xmath183 10@xmath178 @xmath5 , where one sees through the outer edge of the putative cold torus , central x - ray emission above a few kev can penetrate the torus making the nuclear source visible to the observer .
this picture would allow the detection of broad fe k emission line consistent with the unification scheme .
the presence of absorption edge at @xmath7 8.2 kev in the second xmm - newton observation of iras 18325 - 5926 ( obsid:0022940201 ) may suggest the possible link to the scattering medium , but it also depends on whether the x - rays are absorbed by the scattering medium or not . a speculative but consistent scenario for the agn seems to be that the source has a bi - polar outflow with a mass outflow rate comparable to the accretion one . however with a kinetic power much smaller than the radiative luminosity .
the outflow is highly ionized to a distance of @xmath184 cm , with a column density of @xmath185 @xmath5 , beyond which there may be partially ionized gas .
the source is surrounded by a torus at a distance of @xmath186 cm . a fraction of the primary x - rays scatter into our line of sight from the ionized outflow .
a more systematic study of several such sources and theoretical hydro - dynamical simulations have to be undertaken to validate this geometry .
this work is based on observations obtained with _ xmm - newton _ , an esa science mission with instruments and contributions directly funded by esa member states and nasa .
this research has made use of data obtained from the _ suzaku _ satellite , a collaborative mission between the space agencies of japan ( jaxa ) and the usa ( nasa ) .
we would like to thank the _ xmm - newton _ and _ suzaku _ mission teams for developing excellent instruments .
we thank the anonymous referee for the careful reading of the manuscript and useful comments . | we analyze suzaku and xmm - newton data of the highly variable seyfert 2 , iras 18325 - 5926 .
the spectra of the source are well modeled as a primary component described as an absorbed power law and a secondary power law component which is consistent with being scattered emission from an on - axis extended highly ionized medium .
we show that while the primary component varies on a wide range of timescales from @xmath0 s , the scattered emission is variable only on timescales longer than @xmath1 s. this implies that the extent of the scattering medium is greater than @xmath2 cm .
the ratio of the scattered to primary flux ( @xmath3 ) implies a column density for the scattering medium to be @xmath4 @xmath5 .
we argue that for such a medium to be highly ionized it must be located less than @xmath6 cm from the x - ray source .
thus we localize the position and extent of scattering region to be @xmath7 a few @xmath8 cm , with an average particle density of @xmath9 @xmath10 .
we consider the physical interpretation of these results and as an aside , we confirm the presence of a broad iron line emission in both the _ xmm - newton _ and _ suzaku _ observations . |
we are going to construct a combined effective model reproducing the equation of state of hadronic matter as obtained in recent lattice qcd simulations @xcite . the model should reproduce basic physical characteristics of processes encountered in the dense hadronic matter , from the hot qcd phase through the critical temperature region till the lower temperature hadron resonance gas phase .
it has been shown in @xcite that the equation of state derived from that time qcd lattice calculation @xcite can be reproduced by a simple hadron gas resonance model below critical temperature @xmath0 . for higher temperatures the model
is modified by introducing finite widths of heavy hadrons @xcite with a heuristic ansatz for the spectral function which reflects medium modifications of hadrons .
this mott - hagedorn type model is constructed to fit nicely the lattice data , also above @xmath0 where it does so because it leaves light hadrons below a mass threshold of @xmath1 gev unaffected .
the description of the lattice data at high temperatures is accidental because the effective number of those degrees of freedom approximately coincides with that of quarks and gluons . in order to remove this unphysical aspect of the otherwise appealing model one has to extend the spectral broadening also to the light hadrons and
thus describe their disappearance due to the mott effect while simultaneously the quark and gluon degrees of freedom appear at high temperatures due to chiral symmetry restoration and deconfinement . in the present contribution we will report first results obtained by introducing a unified treatment of all hadronic resonances with a state - dependent width @xmath2 in accordance with the inverse hadronic collision time scale from a recent model for chemical freeze - out in a resonance gas @xcite .
the appearance of quark and gluon degrees of freedom is introduced by the polyakov - loop improved nambu
jona - lasinio ( pnjl ) model @xcite .
the model is further refined by adding perturbative corrections to @xmath3 for the high - momentum region above the three - momentum cutoff inherent in the pnjl model .
one obtains eventually a good agreement with lattice qcd data , comparable with all important physical characteristics taken into account .
we introduce the width @xmath4 of a resonance in the statistical model through the spectral function @xmath5 where @xmath6 is the standard normalization factor .
the model ansatz for the resonance width @xmath4 is given by @xcite @xmath7 where @xmath8 , @xmath9 , @xmath10 and @xmath11 mev .
for simplicity , we assume @xmath12 for the strangeness number density and @xmath13 for the baryon number density . then the respective chemical potentials @xmath14 and @xmath15 always , so that the temperature is the only significant statistical parameter here .
the energy density of this model can then be cast in the form @xmath16 where @xmath17 gev and the energy density per degree of freedom with a mass @xmath18 is @xmath19 with the degeneracy @xmath20 . for mesons , @xmath21 and for baryons @xmath22 . for our hadronic gas kept at fixed volume
we have the relation @xmath23 where @xmath24 and @xmath25 are the pressure and entropy density . left : thermodynamic quantities for the old mott - hagedorn resonance gas model @xcite .
different line styles correspond to different values for the parameter @xmath6 in the range from @xmath26 ( dashed line ) to @xmath27 ( solid line ) .
lattice qcd data are from ref .
right : thermodynamic quantities for the new mott - hagedorn resonance gas where quark - gluon plasma contributions are described within the pnjl model including @xmath28 corrections ( dashed lines ) .
hadronic resonances are described within the resonance gas with finite width , as an implementation of the mott effect ( dash - dotted line ) .
the sum of both contributions ( solid lines ) is shown for the energy density ( thick lines ) and pressure ( thin lines ) in comparison with the lattice data from @xcite .
, title="fig:",scaledwidth=48.0% ] left : thermodynamic quantities for the old mott - hagedorn resonance gas model @xcite .
different line styles correspond to different values for the parameter @xmath6 in the range from @xmath26 ( dashed line ) to @xmath27 ( solid line ) .
lattice qcd data
are from ref . @xcite .
right : thermodynamic quantities for the new mott - hagedorn resonance gas where quark - gluon plasma contributions are described within the pnjl model including @xmath28 corrections ( dashed lines ) .
hadronic resonances are described within the resonance gas with finite width , as an implementation of the mott effect ( dash - dotted line ) .
the sum of both contributions ( solid lines ) is shown for the energy density ( thick lines ) and pressure ( thin lines ) in comparison with the lattice data from @xcite .
, title="fig:",scaledwidth=48.0% ] in the left panel of fig .
[ fig.1 ] the results for the pressure and energy density of the model at this stage are shown .
although providing us with an excellent fit of the lattice data , the high - temperature phase of this model is unphysical .
imposing that all mesons lighter than @xmath1 gev are stable provides us with a sb limit at high temperatures which mimicks that due to quarks and gluons in the case for three flavors @xcite . in reality , due to the chiral phase transition at @xmath0 , the quarks loose their mass and therefore the threshold of the continuum of quark - antiquark scattering states is lowered .
the light meson masses , however , remain almost unaffected by the increase in the temperature of the system .
consequently , they merge the continuum and become unbound - their spectral function changes from a delta - function ( on - shell bound states ) to a breit - wigner - type ( off - shell , resonant scattering states ) .
this phenomenon is the hadronic analogue @xcite of the mott - anderson transition for electrons in solid state physics ( insulator - conductor transition ) .
it has been first introduced for the hadronic - to - quark - matter transition in @xcite .
later , within the njl model , a microscopic approach to the thermodynamics of the mott dissociation of mesons in quark matter has been given in the form of a generalized beth - uhlenbeck equation of state @xcite , see also @xcite . as a microscopic treatment of the mott effect for all resonances is presently out of reach
, we introduce an ansatz for a state - dependent hadron resonance width @xmath2 given by the inverse collision time scale recently suggested within an approach to the chemical freeze - out and chiral condensate in a resonance gas @xcite @xmath29 which is based on a binary collision approximation and relaxation time ansatz using for the in - medium hadron - hadron cross sections the geometrical povh - hfner law @xcite . in eq .
( [ gamma ] ) the coefficient @xmath30 is a free parameter , @xmath31 is the partial density of the hadron @xmath32 and the mean squared radii of hadrons @xmath33 obtain in the medium a temperature dependence which is governed by the ( partial ) restoration of chiral symmetry . for the pion
this was quantitatively studied within the njl model @xcite and it was shown that close to the mott transition the pion radius is well approximated by @xmath34 and a pion cloud contribution @xmath35 where from the vacuum values @xmath36 fm and @xmath37 fm one gets @xmath38 fm .
a key point of our approach is that the temperature dependent hadronic radii shall diverge when hadron dissociation ( mott effect ) sets in , driven basically by the restoration of chiral symmetry . as a consequence , in the vicinity of the chiral restoration temperature all meson radii shall behave like that of the pion and all baryon radii like that of the nucleon .
the resulting energy density and pressure behavior is shown in the right panel of fig .
[ fig.1 ] .
this part of the model we call mott - hagedorn - resonance - gas ( mhrg ) .
when all hadrons are gone at @xmath39 mev , we are clearly missing degrees of freedom !
we improve the pnjl model over its standard versions @xcite by adding perturbative corrections in @xmath3 for the high - momentum region above the three - momentum cutoff @xmath40 . in the second step ,
the mhrg part is replaced by its final form , using the state - dependent spectral function for the description of the mott dissociation of all hadron resonances above the chiral transition .
the total pressure obtains the form @xmath41 where @xmath42 stands for the pressure of the mhrg model , accounting for the dissociation of hadrons in hot dense , matter .
the @xmath3 corrections can be split in quark and gluon contributions @xmath43 where @xmath44 stands for the quark contribution and @xmath45 contains the ghost and gluon contributions .
the total perturbative qcd correction to @xmath3 is @xmath46 where @xmath47 .
the corresponding contribution to the energy density is given in standard way by eq .
( [ six ] ) .
we will now include an effective description of the dissociation of hadrons due to the mott effect into the hadron resonance gas model by including the state dependent hadron resonance width ( [ gamma ] ) into the definition of the hrg pressure @xmath48 from the pressure as a thermodynamic potential all relevant thermodynamical functions can be obtained . combining the @xmath28 corrected meanfield pnjl model for the quark - gluon subsystem with the mhrg description of the hadronic resonances we obtain the results shown in the right panel of fig . [ fig.1 ] where the resulting partial contributions in comparison with lattice qcd data from ref .
@xcite are shown .
we see that the lattice qcd thermodynamics is in full accordance with a hadron resonance gas up to a temperature of @xmath49 mev which corresponds to the pseudocritical temperature of the chiral phase transition .
the lattice data saturate below the stefan - boltzmann limit of an ideal quark - gluon gas at high temperatures .
the pnjl model , however , attains this limit by construction .
the deviation is to good accuracy described by perturbative corrections to @xmath3 which vanish at low temperatures due to an infrared cutoff procedure .
the transition region @xmath50\le 250 $ ] is described by the mhrg model , resulting in a decreasing hrg pressure which vanishes at @xmath51 mev .
we have presented two stages of an effective model description of qcd thermodynamics at finite temperatures which properly accounts for the fact that in the qcd transition region it is dominated by a tower of hadronic resonances . to this end
we have further developed a generalization of the hagedorn resonance gas thermodynamics which includes the finite lifetime of hadronic resonances in a hot and dense medium by a model ansatz for a temperature- and mass dependent spectral function .
d.b . wants to thank k.a .
bugaev and k. redlich for discussions and collaboration .
this work was supported in part by the polish national science center ( ncn ) under contract no .
n n202 0523 40 and by the russian foundation for basic research grant no
. 11 - 02 - 01538-a ( d.b . ) .
s. borsanyi _ et al .
_ , jhep * 1011 * , 077 ( 2010 ) .
a. bazavov _ et al .
_ , phys .
d * 80 * , 014504 ( 2009 ) .
f. karsch , k. redlich and a. tawfik , eur .
j. c * 29 * , 549 ( 2003 ) .
f. karsch , lect .
notes phys .
* 583 * , 209 ( 2002 ) ; [ hep - lat/0106019 ] .
d. b. blaschke , k. a. bugaev , fizika b * 13 * , 491 ( 2004 ) ; + phys
. part .
* 2 * , 305 ( 2005 ) .
d. b. blaschke _ et al .
_ , phys . part .
* 8 * , 811 ( 2011 ) .
k. fukushima , phys .
b * 591 * , 277 ( 2004 ) . c. ratti , m. a. thaler , w. weise , phys .
d * 73 * , 014019 ( 2006 ) .
g. e. brown , h. a. bethe , p. m. pizzochero , phys .
b * 263 * , 337 ( 1991 ) .
d. zablocki , d. blaschke and g. rpke , _ metal - to - nonmetal transitions _ , springer series in materials science , * 132 * , 161 ( 2010 ) . | a combined effective model reproducing the equation of state of hadronic matter as obtained in recent lattice qcd simulations is presented .
+ + pacs numbers : 12.38.gc , 12.38.mh , 12.40.ee , 24.85.+p |
mean - variance hedging is one of the classical problems from mathematical finance . in financial terms , its goal is to minimize the mean squared error between a given payoff @xmath0 and the final wealth of a self - financing strategy @xmath1 trading in the underlying assets @xmath2 .
mathematically , one wants to project the random variable @xmath0 in @xmath3 on the space of all stochastic integrals @xmath4 , perhaps after subtracting an initial capital @xmath5 .
the contribution of our paper is to solve this problem via stochastic control methods and stochastic calculus techniques for the case where the asset prices @xmath2 are given by a general ( locally @xmath6-square - integrable ) semimartingale , under a natural no - arbitrage assumption .
the literature on mean - variance hedging is vast , and we do not try to survey it here ; see @xcite for an attempt in that direction .
there are two main approaches ; one of them uses martingale theory and projection arguments , while the other views the task as a linear - quadratic stochastic control problem and uses backward stochastic differential equations ( bsdes ) to describe the solution . by combining tools from both areas ,
we improve earlier work in two directions we describe the solution more explicitly than by the martingale and projection method , and we work in a general semimartingale model without restricting ourselves to particular setups ( like it processes or lvy settings ) .
we show that the value process of the stochastic control problem associated to mean - variance hedging possesses a quadratic structure , describe its three coefficient processes by semimartingale bsdes and show how to obtain the optimal strategy @xmath7 from there .
in contrast to the majority of earlier contributions from the control strand of the literature , we also give a rigorous derivation of these bsdes . for comparison ,
the usual results ( especially in settings with it processes or jump - diffusions ) start from a bsde system and only prove a verification theorem that shows how a solution to the bsde system induces an optimal strategy . apart from being more precise , we think that our approach is also more informative since it shows clearly and explicitly how the bsdes arise , and hence provides a systematic way to tackle mean - variance hedging via stochastic control in general semimartingale models .
more detailed comparisons to the literature are given in the respective sections .
the paper is structured as follows .
we start in section [ sec1 ] with a precise problem formulation and state the martingale optimality principle for the value process @xmath8 of the associated stochastic control problem . assuming that each ( time @xmath9 ) conditional problem admits an optimal strategy ,
we then show that @xmath8 is a quadratic polynomial in @xmath5 whose coefficients are stochastic processes @xmath10 that do not depend on @xmath5 .
this is a kind of folklore result , and our only claim to originality is that we give a very simple proof in a very general setting .
we also show that the coefficient @xmath11 equals the value process @xmath12 for the control problem with initial value @xmath13 and @xmath14
. motivated by the last result , we study in section [ sec2 ] the particular problem for and @xmath14 .
we impose the no - arbitrage condition that there exists an equivalent @xmath15-martingale measure for @xmath2 with @xmath6-square - integrable density and are then able to characterize the process @xmath11 as the solution of a semimartingale bsde .
more precisely , theorem [ theorem2.4 ] shows that all conditional problems for @xmath16 admit optimal strategies if and only if that bsde ( [ equ2.18 ] ) has a solution in a specific class , and in that case , the unique solution is @xmath11 and the conditionally optimal strategies can be given in terms of the solution to ( [ equ2.18 ] ) . in comparison to earlier work
, we eliminate all technical assumptions ( like continuity or quasi - left - continuity ) on @xmath2 , and we also do not need reverse hlder inequalities for our main results . section [ sec3 ] considers the general case of the mean - variance hedging problem with @xmath17 and @xmath18 .
the analog of theorem [ theorem2.4 ] is given in theorem [ theorem3.1 ] , where we describe the three coefficient processes @xmath19 by a coupled system ( [ equ3.1])([equ3.3 ] ) of semimartingale bsdes .
existence of optimal strategies for all conditional problems for @xmath20 is shown to be equivalent to solvability of the system ( [ equ3.1])([equ3.3 ] ) , with solution @xmath19 , and we again express the conditionally optimal strategies in terms of the solution to ( [ equ3.1])([equ3.3 ] ) . as mentioned above , this is stronger than only a verification result . in section [ sec4 ] ,
we provide equivalent alternative versions for our bsdes which are more convenient to work with in some examples with jumps .
this also allows us to discuss in more detail the connections to the existing literature .
finally , section [ sec5 ] illustrates the use of our results and gives further links to the literature by a number of simple examples .
we start with a finite time horizon @xmath21 and a filtered probability space @xmath22 with the filtration @xmath23 satisfying the usual conditions of right - continuity and .
let @xmath24 be an @xmath25-valued rcll semimartingale , and denote by @xmath26 the space of all predictable @xmath2-integrable processes @xmath1 , @xmath27 for short , such that the stochastic integral process @xmath28 is in the space @xmath29 of semimartingales .
our basic references for terminology and results from stochastic calculus are @xcite and @xcite . for @xmath17 and @xmath18 , the problem of _ mean - variance hedging _
( _ mvh _ ) is to @xmath30\qquad \mbox{over all $ \vartheta\in\theta$.}\ ] ] the interpretation is that @xmath2 models the ( discounted ) prices of @xmath31 risky assets in a financial market containing also a riskless bank account with ( discounted ) price 1 .
an integrand @xmath1 together with @xmath17 then describes a self - financing dynamic trading strategy with initial wealth @xmath5 , and @xmath0 stands for the ( discounted ) payoff at time @xmath32 of some financial instrument . by using @xmath33
, we generate up to time @xmath32 via trading a wealth of @xmath34 , and we want to choose @xmath1 in such a way that we are close , in the @xmath3-sense , to the payoff @xmath0 . we embed this into a _ stochastic control problem _ and define for @xmath35 and @xmath36 $ ] @xmath37,\end{aligned}\ ] ] where @xmath38 on @xmath39\!]}\}$ ] .
our goal is to study the _ dynamic value family _
@xmath40\\[-8pt ] & = & { \mathop{\operatorname{ess}\operatorname{inf}}}_{\vartheta\in\theta } e\biggl [ \biggl ( h - x - \int_t^t \vartheta_r{\,d}s_r \biggr)^2 \bigg| { { \mathcal f}}_t \biggr],\qquad t\in[0,t ] , \nonumber\end{aligned}\ ] ] in order to describe the optimal strategy for the mvh problem ( [ equ1.1 ] )
. observe that with this notation , we have the identity @xmath41\!]}t , t{]\ ! ] } } \bigr ) = { v^h}_u\bigl ( x , \psi i_{{]\!]}t , u{]\ ! ] } } \bigr)\ ] ] for @xmath42 . because the family of random variables @xmath43\ ] ] for @xmath44 is closed under taking maxima and minima ,
we have the classical _ martingale optimality principle _ in the following form ; see , for instance , @xcite for the general theory , or @xcite for a formulation closer to the present one .
[ prop1.1 ] fix @xmath18 . for every @xmath45 and @xmath36 $ ] , we have : the process @xmath46 is a @xmath6-submartingale for every .
a strategy @xmath47 is optimal for ( [ equ1.2 ] ) ( i.e. , attains the essential infimum there ) if and only if @xmath48 is a .
if the strategy @xmath49 solves ( [ equ1.1 ] ) , then @xmath50\!]}t , t{]\!]}}$ ] is optimal for @xmath51 .
for the special case @xmath14 , the fact that @xmath52 is a cone immediately gives @xmath53 = x^2 { v^0}_t(1).\ ] ] this holds for any random variable @xmath54 .
so proposition [ prop1.1 ] almost directly gives : [ cor1.2 ] for every @xmath36 $ ] , we have : the process @xmath55 is a @xmath6-submartingale for every @xmath44 .
a strategy @xmath56 is optimal for @xmath57 in ( [ equ1.3 ] ) if and only if the process @xmath58 is a @xmath6-martingale .
if @xmath59 solves ( [ equ1.1 ] ) for @xmath13 and @xmath14 , then @xmath60 since ( 1 ) and ( 2 ) are special cases of proposition [ prop1.1 ] , we only need to prove ( 3 ) .
fix @xmath36 $ ] , set @xmath61 and define @xmath62\!]}t , t{]\!]}}$ ] . by part ( 3 ) of proposition [ prop1.1 ] with @xmath16 , the strategy @xmath50\!]}t , t{]\!]}}$ ] is optimal for @xmath63 so that @xmath64\\ & & \qquad \le i_{d_t } e\biggl [ \biggl ( 1+\vartheta^*{\cdot}s_t + \int_t^t \varphi _
r{\,d}s_r \biggr)^2 \bigg| { { \mathcal f}}_t \biggr ] = 0\end{aligned}\ ] ] by the definitions of @xmath65 and @xmath66 .
this yields @xmath67 again by the definition of @xmath66 , and so we get ( [ equ1.4 ] ) . as in proposition
a.2 of @xcite or theorem 2.28 of @xcite , we also obtain : [ prop1.3 ] fix @xmath18 . for every @xmath45 ,
@xmath36 $ ] and , there exists an rcll version of the @xmath6-submartingale @xmath68 moreover , for each @xmath17 , the family @xmath69 \ } $ ] of random variables can be aggregated into an rcll process , which we again call @xmath70 in the sequel , we always choose and work with the rcll versions from proposition [ prop1.3 ] . for easier discussion of the next result
, we introduce some more terminology .
we denote by @xmath71 the ( a priori possibly empty ) set of all probability measures @xmath72 equivalent to @xmath6 on @xmath73 such that @xmath2 is a @xmath72-@xmath15-martingale and @xmath74 . assuming that @xmath71 is nonempty is one way of imposing _ absence of arbitrage _ for our financial market and also fits naturally with the fact that our basic problem is cast in quadratic terms .
the density process of @xmath72 with respect to @xmath6 is denoted by @xmath75 , and we say that @xmath76 satisfies the reverse hlder inequality @xmath77 if there is a constant @xmath78 with @xmath79 \le c ( { z^q}_\tau)^2 $ ] @xmath6-a.s . for all stopping times @xmath80 . it is well known that if there is some @xmath76 satisfying @xmath77 , then @xmath81 as well as @xmath82 for each @xmath9 are closed in @xmath3 so that both ( [ equ1.1 ] ) and ( [ equ1.2 ] ) for each @xmath9 have a solution ; see theorem 5.2 of @xcite .
moreover , for any @xmath83 and any @xmath44 , the product of @xmath84 and @xmath85 is a @xmath6-@xmath15-martingale with @xmath6-integrable supremum ; so @xmath85 is a true @xmath72-martingale , and @xmath86 a.s .
implies that @xmath87 in @xmath88 .
this is used later several times to argue that a self - financing strategy is uniquely determined by its wealth process ( i.e. , stochastic integral ) .
our main result in this section now provides the basic structure of the process @xmath8 and of the optimal strategies for ( [ equ1.2 ] ) .
[ theorem1.4 ] fix @xmath18 .
suppose that for each @xmath36 $ ] , ( [ equ1.2 ] ) has a solution @xmath89 for every @xmath17 .
suppose also that for any @xmath44 , @xmath86 a.s .
implies that @xmath90 in @xmath88
. then each @xmath91 is of the affine form @xmath92 and each @xmath93 has the quadratic form @xmath94 for rcll processes @xmath95 not depending on @xmath5 .
moreover , @xmath96 is the solution of ( [ equ1.3 ] ) , and the quadratic coefficient @xmath97 equals @xmath57 from ( [ equ1.3 ] ) and does not depend on @xmath0 .
fix @xmath36 $ ] .
denote by @xmath98 the space of all stochastic integrals on @xmath99\!]}t , t{]\!]}$ ] of @xmath44 and by @xmath100 its closure in @xmath3 .
since the problems ( [ equ1.2 ] ) with payoff @xmath0 for @xmath13 and @xmath101 have solutions ( which are given by projections ) , so does problem ( [ equ1.2 ] ) for @xmath13 and payoff by taking differences , and the latter problem is identical to ( [ equ1.2 ] ) for @xmath102 so that @xmath103 .
both here and in the next argument , we exploit our assumption that a self - financing strategy is uniquely determined by its wealth process . if @xmath104 is the projection in @xmath3 on @xmath100 , then clearly @xmath105 and so ( [ equ1.5 ] ) follows with @xmath106 and @xmath107
this gives @xmath108\\ & = & e\biggl [ \biggl ( h - \int_t^t\vartheta^{0,t}_r{\,d}s_r - x\biggl ( 1+\int_t^t \vartheta^{1,t}_r{\,d}s_r \biggr ) \biggr)^2 \bigg| { { \mathcal f}}_t \biggr],\end{aligned}\ ] ] and hence we directly obtain the expression ( [ equ1.6 ] ) with @xmath109 , \\
{ v^{(1)}}_t & = & e\biggl [ \biggl ( h-\int_t^t { \vartheta^{*,t}}_r(0 , h ) { \,d}s_r \biggr ) \biggl ( 1+\int_t^t { \vartheta^{*,t}}_r(1,0 ) { \,d}s_r
\biggr ) \bigg| { { \mathcal f}}_t \biggr]\end{aligned}\ ] ] and @xmath110 = { v^0}_t(1).\ ] ] since the families @xmath111 \}$ ] aggregate into an rcll process , the same holds for the families @xmath95 from ( [ equ1.6 ] ) .
the last assertion is clear from the above proof .
\(1 ) as mentioned above , one sufficient condition for all assumptions of theorem [ theorem1.4 ] is the existence of some @xmath76 satisfying the reverse hlder inequality @xmath77 ; see @xcite .
\(2 ) the particular choice of @xmath26 for the space of integrands is convenient and also exploited later , but not crucially important for the conclusion of theorem [ theorem1.4 ] to hold .
all we need is that there exist for all @xmath9 solutions @xmath91 for all @xmath5 , that the martingale optimality principle from proposition [ prop1.1 ] holds , and that @xmath52 [ or @xmath112 , which must be a subset of @xmath3 ] is a linear space .
of course , existence of solutions for all @xmath5 and all @xmath0 is equivalent to closedness of @xmath112 in @xmath3 ; and the key point for the martingale optimality principle is closedness under bifurcation of @xmath52 .
\(3 ) we emphasize that theorem [ theorem1.4 ] is a bit of a folklore result in the literature on mean - variance hedging , and we do not claim any great originality here .
variants in different levels of generality can be found in @xcite , @xcite , @xcite , @xcite , to name but a few .
however , we think that it is useful to have a presentation which is as general , and yet as simple , as possible . our goal in the sequel is to study the dynamics of the coefficient processes @xmath95 and use them to express the optimal strategies @xmath91
let us first simplify things a little . because @xmath114 is the solution ( minimizer ) of ( [ equ1.3 ] ) , the first order condition for that quadratic optimization problem implies that @xmath115 = 0 $ ] @xmath6-a.s .
for each @xmath36 $ ] and @xmath44 .
we note for later use that this allows us to write @xmath116.\ ] ] also for later use , we give some additional results for the coefficients @xmath117 , @xmath11 .
[ lem1.5 ] under the assumptions of theorem [ theorem1.4 ] , we have : @xmath11 is a @xmath6-submartingale with @xmath118 . @xmath119 is a @xmath6-submartingale with @xmath120 , hence of class ( d ) .
@xmath121 is a @xmath6-special semimartingale with @xmath122 of class ( d ) .
therefore @xmath121 is in @xmath123 and for its canonical decomposition @xmath124 , we have @xmath125 .
\(1 ) by theorem [ theorem1.4 ] and ( [ equ1.7 ] ) , we have @xmath126 , and this is a @xmath6-submartingale by part ( 1 ) of corollary [ cor1.2 ] ( for @xmath127 ) . because @xmath127 is in @xmath128 , we get @xmath129 directly from ( [ equ1.3 ] ) .
\(2 ) theorem [ theorem1.4 ] gives @xmath130 , and this is a @xmath6-submartingale by part ( 1 ) of proposition [ prop1.1 ] ( for @xmath131 ) and nonnegative by the definition in ( [ equ1.2 ] ) .
since @xmath127 is in @xmath52 , ( [ equ1.2 ] ) also gives @xmath132 $ ] for all @xmath9 .
\(3 ) by part ( 1 ) of proposition [ prop1.1 ] , @xmath8 is a @xmath6-submartingale , hence a semimartingale , and so are @xmath11 and @xmath119 by ( 1 ) and ( 2 ) . because @xmath133 by theorem [ theorem1.4 ] , also @xmath121 is then a @xmath6-special semimartingale .
moreover , @xmath134 for all @xmath5 due to ( [ equ1.2 ] ) implies that @xmath135 , \qquad 0\le t\le t,\ ] ] by ( 1 ) and ( 2 ) so that @xmath122 is of class ( d ) .
the rest of part ( 3 ) is then clear .
in this section , we give a description of ( the rcll version of ) the value process @xmath136 , \qquad0\le t\le t,\ ] ] of the problem ( [ equ1.3 ] ) . since this is by ( [ equ1.7 ] ) and theorem [ theorem1.4 ] the _ quadratic _ coefficient in the representation ( [ equ1.6 ] ) , we use in this section the shorter notation @xmath137 we also remark that @xmath138 coincides with the _ opportunity process _ from @xcite , although the latter is defined there with a different space @xmath52 of integrands @xmath1 for @xmath2 .
let us first prove strict positivity of @xmath138 , as well as of @xmath139 . [ lem2.1 ] suppose @xmath140 .
then @xmath138 and @xmath139 are both strictly positive , in the sense that @xmath141 and @xmath142 for @xmath143 = 1 $ ] .
if there is some @xmath76 satisfying the reverse hlder inequality @xmath77 , we even have @xmath144 @xmath6-a.s . for some constant @xmath145 . for @xmath76 with density process @xmath146 , define as in @xcite a new probability @xmath147 by @xmath148 } $ ] .
then the bayes rule gives @xmath149 \over e[z_t^2 ] } , \\
\label{equ2.3 } { z^{r;q}}_t & : = & { dr\over dq}\bigg|_{{{\mathcal f}}_t } = { e_q[z_t | { { \mathcal f}}_t ] \over e[z_t^2 ] } = { 1\over z_t } { z^{r;p}}_t.\end{aligned}\ ] ] using the bayes rule and ( [ equ2.2 ] ) , jensen s inequality , again the bayes rule and ( [ equ2.3 ] ) yields @xmath150 \\ & & \qquad = { z^{r;p}}_t e[z_t^2 ] e_r\biggl [ ( z_t^2)^{-1 } \biggl ( 1+\int_t^t \vartheta _
r{\,d}s_r \biggr)^2 \bigg| { { \mathcal f}}_t \biggr ] \\ & & \qquad \ge { z^{r;p}}_t e[z_t^2 ] \biggl ( e_r\biggl [ ( z_t)^{-1 } \biggl ( 1+\int_t^t \vartheta_r{\,d}s_r \biggr ) \bigg| { { \mathcal f}}_t \biggr ] \biggr)^2 \\ & & \qquad = { z^{r;p}}_t e[z_t^2 ] \biggl ( ( { z^{r;q}}_t)^{-1 } e_q\biggl [ ( e[z_t^2])^{-1 } \biggl ( 1+\int_t^t \vartheta_r{\,d}s_r \biggr ) \bigg| { { \mathcal f}}_t \biggr ] \biggr)^2.\end{aligned}\ ] ] but as already noted before theorem [ theorem1.4 ] , @xmath151 is a @xmath72-martingale whenever @xmath76 and @xmath44 .
so we get by using ( [ equ2.3 ] ) and ( [ equ2.2 ] ) that @xmath152 \ge { { z^{r;p}}_t e[z_t^2 ] \over ( { z^{r;q}}_t e[z_t^2])^2 } = { z_t^2 \over e[z_t^2 | { { \mathcal f}}_t]},\ ] ] and the first assertion follows since @xmath153 @xmath6-a.s . by the minimum principle for supermartingales and @xmath154 < \infty$ ] @xmath6-a.s . by the martingale
maximal inequality .
if @xmath72 satisfies @xmath77 with constant @xmath78 , we can take @xmath155 for the second claim .
strict positivity of the opportunity process and its left limits ( hence of @xmath138 and @xmath139 ) is also proved in lemma 3.10 of @xcite .
however , the above short proof seems to us more transparent .
the optimization problem in ( [ equ2.1 ] ) has a ( well - known ) dual formulation as follows . extending @xmath71 a little , we denote by @xmath156 the set of all _ signed _ measures @xmath157 on @xmath73 with @xmath158 = 1 $ ] and such that the product of @xmath2 and the density process @xmath84 of @xmath72 with respect to @xmath6 is a @xmath6-@xmath15-martingale .
we call @xmath159 _ variance - optimal _ if @xmath160 for all @xmath161 , and we say that the _ variance - optimal martingale measure _ ( _ vomm _ ) exists if @xmath162 is variance - optimal .
( in particular , @xmath163 is then by definition equivalent to @xmath6 . )
if @xmath2 is continuous , theorem 1.3 of @xcite shows that @xmath140 is sufficient for the vomm to exist ; but if @xmath2 can have jumps , the situation is more complicated
. the dynamic problem of finding the vomm has the value process @xmath164 then we have the following direct connection to @xmath12 and ( [ equ2.1 ] ) .
[ prop2.2 ] suppose @xmath165 and that the vomm exists
. then @xmath166 . we know from ( [ equ2.4 ] ) in the proof of lemma [ lem2.1 ] that for @xmath44 and @xmath76 , @xmath167 \ge 1 / e [ ( { z^q}_t / { z^q}_t ) ^2 | { { \mathcal f}}_t ] , \qquad 0\le t\le t.\ ] ] taking the ess inf over @xmath44 and the ess sup over @xmath83 implies that @xmath168 .
conversely , since @xmath169 , the martingale optimality principle in corollary [ cor1.2 ] gives @xmath170,\qquad 0\le t\le t,\ ] ] for every @xmath171 .
but if we define , as in @xcite , @xmath172 then @xmath173 is by corollary 2.9 of @xcite the closure of @xmath174 in @xmath3 , and this allows us to extend ( [ equ2.5 ] ) to every @xmath175 . indeed , for a sequence @xmath176 in @xmath177 with @xmath178 in @xmath3 , the right - hand side of ( [ equ2.5 ] ) for @xmath179 converges in @xmath180 to the right - hand side of ( [ equ2.5 ] ) for @xmath1 , and because we have @xmath181 $ ] for @xmath182 and @xmath183 , the left - hand side of ( [ equ2.5 ] ) for @xmath179 converges in probability to the left - hand side of ( [ equ2.5 ] ) for @xmath1 .
we remark that the use of corollary 2.9 in @xcite exploits that @xmath165 . by assumption , the vomm @xmath163 exists . a slight modification of the proof of lemma 2.2 in @xcite [ since @xmath2 is in @xmath123 instead of locally bounded ] yields @xmath184 for some @xmath185 and @xmath186 and thus @xmath187 = c + \int_0^t { \widetilde\vartheta}_r{\,d}s_r , $ ] @xmath188 .
applying ( [ equ2.5 ] ) with @xmath189 and using the bayes rule therefore gives @xmath190 \ge ( z^{{\widetilde q}}_t)^2 ( e_{{\widetilde q } } [ z^{{\widetilde q}}_t | { { \mathcal f}}_t ] ) ^2 { v^0}_t(1 ) = ( e [ ( z^{{\widetilde q}}_t ) ^2 | { { \mathcal f}}_t ] ) ^2 { v^0}_t(1)\ ] ] and hence @xmath191 this completes the proof . for experts on mean - variance hedging ,
proposition [ prop2.2 ] is also a kind of folklore result . for the case where the filtration is continuous ,
it can , for instance , be found in proposition 4.2 of @xcite ( with the remark that it extends to general @xmath192 if @xmath2 is continuous ) .
but we do not know a reference for the level of generality given here .
henceforth , we often use the following simple fact : @xmath193 in ( [ equ2.6 ] ) , the ( right ) superscript @xmath194 denotes the compensator or dual predictable projection .
this should not be confused with the predictable projection of a process @xmath195 which is denoted by @xmath196 , with a left superscript .
the most frequent application of ( [ equ2.6 ] ) will be for @xmath197 $ ] , where @xmath198^{{{\bolds{\mathsf p } } } } } = { \langle m\rangle}$ ] when @xmath199 is a locally square - integrable local martingale .
_ in the sequel _ , _ we focus on the case @xmath200 so that @xmath2 is one - dimensional .
_ one can obtain analogous results for @xmath201 ( and we shall comment on this later ) , but the arguments and formulations look more technical without providing extra insight . when @xmath165 so that @xmath2 is in particular a @xmath6-special semimartingale , we write @xmath202 for its @xmath6-canonical decomposition and note that @xmath203 and @xmath204 is predictable and of locally square - integrable ( or even locally bounded ) variation .
if we also have @xmath140 , then it is well known that @xmath2 satisfies the so - called structure condition , that is , that @xmath2 has the form @xmath205 with @xmath203 and @xmath206 ; see theorem 1 of @xcite .
this implies that @xmath207 = \biggl [ \int\lambda{\,d}{\langle m\rangle}\biggr ] = \sum(\lambda_s \delta{\langle m\rangle}_s)^2 = ( \lambda^2 \delta{\langle m\rangle}){\cdot}{\langle m\rangle}\ll { \langle m\rangle}.\ ] ] because @xmath204 is predictable , @xmath208 $ ] is a local @xmath6-martingale by yoeurp s lemma so that @xmath209^{{{\bolds{\mathsf p } } } } } = { ( [ m ] + [ a])^{{{\bolds{\mathsf p } } } } } = ( 1+\lambda^2\delta{\langle m\rangle}){\cdot}{\langle m\rangle}.\ ] ] now suppose that @xmath165 and @xmath140 . to describe the process @xmath210 by a bsde
, we first introduce an auxiliary operation .
suppose @xmath195 is a @xmath6-special semimartingale with canonical decomposition @xmath211 .
then @xmath212 ] = [ { n^y } , [ s ] ] + \delta b^y{\cdot}[s ] $ ] , and if @xmath213 ] $ ] is of locally @xmath6-integrable variation , we have by ( [ equ2.8 ] ) and ( [ equ2.6 ] ) that @xmath214 \bigr]^{{{\bolds{\mathsf p } } } } } = { \bigl [ { n^y } , [ s ] \bigr]^{{{\bolds{\mathsf p } } } } } + \delta b^y{\cdot}{[s]^{{{\bolds{\mathsf p } } } } } \ll { \langle m\rangle}.\ ] ] note also that the predictable stopping theorem gives @xmath215 so that @xmath216 the auxiliary quantity we need is the predictable radon nikodm derivative @xmath217 ] ^{{{\bolds{\mathsf p}}}}}_t \over d{\langle m\rangle}_t } , \qquad 0\le t\le t.\ ] ] finally , we introduce the notation @xmath218 the condition that @xmath213 ] $ ] is in @xmath219 ( and hence has a compensator ) is , for instance , satisfied if @xmath195 is bounded , hence in particular for @xmath220 . in the context of the equations we study , the operation @xmath221 in ( [ equ2.12 ] )
can sometimes be simplified .
if @xmath2 is continuous , then so are @xmath222 $ ] and @xmath223 , due to ( [ equ2.7 ] ) ; so @xmath224 and @xmath225 then both vanish and ( [ equ2.12 ] ) reduces to the expression @xmath226 .
looking ahead at ( [ equ2.18 ] ) , however , we see that we are interested in the case where @xmath227 , and so we then also get @xmath228 and hence @xmath229 . finally , if even the filtration @xmath192 is continuous , then @xmath230 in ( [ equ2.18 ] ) is continuous ; so is then @xmath195 , and we end up with @xmath231 .
our next result shows that @xmath232 is always strictly positive .
this is important since we later need to divide by @xmath233 .
[ lem2.3 ] suppose @xmath140 and @xmath165 .
if @xmath144 for some constant @xmath145 , then @xmath234\!]}$.}\ ] ] in general , we still have @xmath235\!]}$.}\ ] ] moreover , @xmath236 is locally bounded away from 0 ( uniformly in @xmath237 ) . if @xmath238 , then @xmath239 - \delta[s]$ ] is in @xmath240 and hence also @xmath241 .
but @xmath242 $ ] , hence @xmath243^{{{\bolds{\mathsf p } } } } } = ( 1+\lambda^2\delta{\langle m\rangle } ) { \cdot}{\langle m\rangle}$ ] by ( [ equ2.6 ] ) and ( [ equ2.8 ] ) , and so @xmath244 ) ^{{{\bolds{\mathsf p } } } } } - \delta(1+\lambda^2\delta{\langle m\rangle } ) { \cdot}{\langle m\rangle}\\ & = & \int\biggl ( { d { ( { q}{\cdot}[s ] ) ^{{{\bolds{\mathsf p } } } } } \over d{\langle m\rangle } } - \delta(1+\lambda ^2\delta{\langle m\rangle } ) \biggr ) { \,d}{\langle m\rangle}\in{{\mathcal a}}^+_{{\mathrm{loc}}}(p).\end{aligned}\ ] ] writing @xmath245 and @xmath246 = [ { q } , [ s ] ] $ ] and using ( [ equ2.8])([equ2.12 ] ) yields @xmath247 ) ^{{{\bolds{\mathsf p } } } } } & = & { q}_- { \cdot}{[s]^{{{\bolds{\mathsf p } } } } } + { [ { n^{{q } } } , [ s ] ] ^{{{\bolds{\mathsf p } } } } } + \delta{b^{{q}}}{\cdot}{[s]^{{{\bolds{\mathsf p } } } } } \nonumber\\ & = & \bigl ( { { } ^{{{\bolds{\mathsf p}}}}{q}}(1+\lambda^2\delta{\langle m\rangle } ) + { g}({q } ) \bigr ) { \cdot}{\langle m\rangle}\\ & = & { \mathcal n}({q } ) { \cdot}{\langle m\rangle}.\nonumber\end{aligned}\ ] ] thus we obtain @xmath248 , and this implies ( [ equ2.13 ] ) since @xmath249 . in general , setting @xmath250 ( with @xmath251 ) gives @xmath252 stationarily because @xmath253 by lemma [ lem2.1 ] , and @xmath254 on @xmath255 since @xmath256 .
the argument for ( [ equ2.13 ] ) now implies that @xmath257 holds @xmath258-a.e . on @xmath259 , and ( [ equ2.14 ] )
follows since @xmath260\!]}$ ] . for the final assertion ,
note that the preceding proof shows that @xmath261 so that the nonnegative process @xmath262 is prelocally bounded .
since @xmath263 is like @xmath236 predictable , it is therefore by @xcite , remark viii.11 also locally bounded , and this means that @xmath236 is locally bounded away from 0 .
if @xmath201 , both @xmath222 $ ] and @xmath223 have to be replaced by matrix - valued processes @xmath264)_{i , j=1 , \ldots , d}$ ] and @xmath265 .
we then take a predictable @xmath266 with @xmath267 and define the matrix - valued predictable process @xmath268 by @xmath269 ] ^{{{\bolds{\mathsf p}}}}}_t \over db_t } , \qquad 0\le t\le t.\ ] ] analogously to lemma [ lem2.3 ] , one can then prove that @xmath270 recalling the notation ( [ equ2.12 ] ) , we now consider the _ backward equation _ @xmath271
a _ solution _ of ( [ equ2.18 ] ) is a triple @xmath272 , where @xmath230 is a local @xmath6-martingale which is strongly @xmath6-orthogonal to @xmath199 , @xmath273 is in @xmath274 and @xmath211 is a semimartingale with @xmath213 ] \in{{\mathcal a}}_{{\mathrm{loc}}}(p)$ ] .
note that @xmath275 and @xmath199 come from @xmath2 via ( [ equ2.7 ] ) . with a slight abuse of terminology ,
we sometimes call @xmath195 instead of the whole triple @xmath276 a solution ; any properties then only refer to @xmath195 . denoting
the stochastic exponential started at time @xmath9 of a semimartingale @xmath277 by @xmath278 our first main result is the following description of @xmath279 via a bsde .
[ theorem2.4 ] suppose that @xmath165 and @xmath140 .
then : 1 .
the following two assertions are equivalent : a. for every @xmath36 $ ] , there exists an optimal strategy @xmath280 for ( [ equ1.2 ] ) with @xmath16 .
b. there exists a solution @xmath272 to the bsde ( [ equ2.18 ] ) having @xmath281 , @xmath282 , @xmath195 bounded and strictly positive and such that for every @xmath36 $ ] , the process @xmath283 is in @xmath29 . + if or hold , then the optimal @xmath114 is for every @xmath9 given by @xmath284 and @xmath285 is the unique bounded strictly positive solution of ( [ equ2.18 ] ) .
2 . suppose , in addition , that there is some @xmath83 satisfying the reverse hlder inequality @xmath77 .
then @xmath285 is the unique solution to the bsde ( [ equ2.18 ] ) in the class of processes satisfying @xmath286 for positive constants @xmath287 .
moreover , the optimal @xmath114 exist and are given by ( [ equ2.19 ] ) . throughout this proof ,
we write @xmath288 for @xmath114 and denote by @xmath289 a generic local @xmath6-martingale that can change from one appearance to the next .
\(1 ) for part ( 1 ) of theorem [ theorem2.4 ] , we start by deriving the bsde ( [ equ2.18 ] ) . by part ( 1 ) of lemma [ lem1.5 ] , @xmath290 is a @xmath6-submartingale , hence a @xmath6-special semimartingale with canonical decomposition @xmath291 , and @xmath292 implies that @xmath293 and @xmath294 has bounded jumps and is in @xmath295 .
the galtchouk kunita
watanabe decomposition thus allows us to write @xmath296 with @xmath297 and @xmath298 strongly @xmath6-orthogonal to @xmath199 . combining this with ( [ equ2.7 ] ) and yoeurp
s lemma then gives @xmath299 & = & m + \varphi{\cdot}[m ] + [ a , { b^{{q } } } ] \nonumber\\[-8pt]\\[-8pt ] & = & m + ( \varphi+ \lambda\delta{b^{{q}}}){\cdot}{\langle m\rangle}.\nonumber\end{aligned}\ ] ] we now apply it s formula to the process @xmath300 , @xmath301 , for @xmath17 , @xmath36 $ ] and @xmath302 .
( we sometimes omit writing the dependence of @xmath303 on @xmath9 . )
this gives @xmath304_r.\ ] ] next we apply the product rule with ( [ equ2.22 ] ) , ( [ equ2.20 ] ) , ( [ equ2.7 ] ) , ( [ equ2.21 ] ) and then use @xmath305 and @xmath306 + [ { q } , [ s ] ] = ( { q}_-+\delta{q}){\cdot}[s ] = { q}{\cdot}[s ] $ ] as well as ( [ equ2.8 ] ) , ( [ equ2.10 ] ) for @xmath138 and ( [ equ2.15 ] ) to obtain @xmath307_r \nonumber\\ & & { } + 2 \int_t^u x^\vartheta_{r- } \vartheta_r ( \varphi_r + \lambda _
r\delta{b^{{q}}}_r ) { \,d}{\langle m\rangle}_r + \int_t^u \vartheta_r^2{\,d } [ { q},[s ] ] _ r \nonumber\\[-8pt]\\[-8pt ] & = & m_u - m_t + \int_t^u ( x^\vartheta_{r-})^2 { \,d}{b^{{q}}}_r \nonumber\\ & & { } + \int_t^u \bigl ( 2 x^\vartheta_{r- } \vartheta_r ( \varphi_r + \lambda_r { { } ^{{{\bolds{\mathsf p}}}}{q}}_r ) + \vartheta_r^2 { { \mathcal n}_{r}({q } ) } \bigr ) { \,d}{\langle m\rangle}_r \nonumber\\ & = & m_u - m_t + \int_t^u f(r , x^\vartheta_{t , r- } ; \vartheta ) { \,d}c_r,\nonumber\end{aligned}\ ] ] where @xmath308 is a predictable process with @xmath309 , @xmath310 , and @xmath311\\[-8pt ] & : = & y^2 \beta_r + \bigl ( 2y\vartheta_r ( \varphi_r + \lambda_r { { } ^{{{\bolds{\mathsf p}}}}{q}}_r ) + \vartheta_r^2 { { \mathcal n}_{r}({q } ) } \bigr ) \nu_r\nonumber\end{aligned}\ ] ] is a quadratic polynomial in @xmath312 with random processes as coefficients . replacing @xmath313 by @xmath314 , we can assume that @xmath78 , as well as its continuous part @xmath315 , is strictly increasing . by corollary [ cor1.2 ]
, @xmath316 is a @xmath6-submartingale for every @xmath44 and a @xmath6-martingale for the optimal @xmath317 , if that exists .
this means that the in ( [ equ2.23 ] ) is increasing for every @xmath44 and identically 0 for @xmath318 , and the same then applies separately for the corresponding integrals with respect to the continuous and purely discontinuous parts @xmath315 and @xmath319 of @xmath78 .
similarly as in @xcite , we therefore obtain for each @xmath17 , @xmath320 the details for this step are a bit more technical and are postponed to step ( 2 ) . using the definition of @xmath321 in ( [ equ2.24 ] ) and
completing the square gives @xmath322 and we claim that for a localizing sequence @xmath323 , @xmath324\ ! ] } } \in\theta.\ ] ] indeed , @xmath236 is locally bounded away from 0 by lemma [ lem2.3 ] , and @xmath325 is bounded like @xmath138 due to lemma [ lem1.5 ] .
moreover , @xmath326 is locally bounded since it is predictable and rcll , and @xmath65 is locally in @xmath327 by construction .
thus we obtain via cauchy schwarz that both @xmath65 and @xmath275 , and then also the ratio in ( [ equ2.27 ] ) , are locally in @xmath328 , as claimed .
inserting @xmath179 into ( [ equ2.26 ] ) makes the first term in ( [ equ2.26 ] ) vanish for @xmath329 and thus yields @xmath330 plugging this into ( [ equ2.25 ] ) and integrating gives @xmath331 , and plugging that in turn into ( [ equ2.20 ] ) shows that the triple @xmath332 solves the bsde ( [ equ2.18 ] ) .
moreover , we see from lemma [ lem2.1 ] and @xmath333 that @xmath138 is strictly positive and bounded .
\(2 ) to prove ( [ equ2.25 ] ) , we use the same basic approach as in @xcite , but we must be more careful and handle jumps since @xmath2 is not continuous . for ease of notation , we sometimes omit the third argument @xmath1 of @xmath334 . we first write @xmath335 and denote by @xmath336 a sequence of stopping times exhausting the jumps of @xmath319 ( or @xmath78 ) .
each @xmath337 is predictable because @xmath78 is predictable . by corollary [ cor1.2 ]
, we then have with probability 1 that @xmath338 is rcll and simultaneously for all rational @xmath339 $ ] that @xmath340 for each @xmath44 , and for the optimal @xmath341 , the processes in ( [ equ2.28 ] ) and ( [ equ2.29 ] ) vanish identically .
indeed , ( [ equ2.29 ] ) follows from ( [ equ2.28 ] ) since the process in ( [ equ2.29 ] ) is simply the continuous part of the process in ( [ equ2.28 ] ) . for any @xmath337
, we thus have with probability 1 that @xmath342 because @xmath337 is predictable , there are stopping times @xmath343 taking only rational values and such that @xmath344 and @xmath345 on @xmath346 ; see theorem iv.77 in @xcite .
thus we obtain for @xmath6-almost all @xmath347 that @xmath348 these integrals tend to @xmath349 as @xmath329 because @xmath350 , and so we get @xmath351 which means that @xmath352 @xmath353-a.e .
, for each @xmath44 . for the optimal @xmath354 , we get the null process in ( [ equ2.28 ] ) , hence equality in ( [ equ2.30 ] ) , and so we have @xmath355 for the continuous part @xmath315 , ( [ equ2.29 ] ) gives with @xmath356 that @xmath357 $ , $ p$-a.s.}\ ] ] we claim that for each @xmath358 , @xmath359 postponing the argument for the moment , we obtain that the inequality in ( [ equ2.32 ] ) also holds for all @xmath339 $ ] , @xmath6-a.s . setting @xmath360 ,
we then get as in appendix b of @xcite via fubini s theorem that ( dropping arguments @xmath1 from @xmath334 ) @xmath361 this uses that @xmath362 by continuity of @xmath315 .
the second term on the right - hand side of ( [ equ2.34 ] ) tends to 0 as @xmath363 by corollary b.1 in @xcite .
writing @xmath364 we have @xmath365 for @xmath366 by continuity of @xmath315 and therefore @xmath367 as @xmath368 .
moreover , we have ( uniformly in @xmath369 and @xmath9 ) @xmath370 which is in @xmath3 , hence @xmath6-a.s .
finite , for @xmath44 .
the first term on the right - hand side of ( [ equ2.34 ] ) can now be estimated above by @xmath371 since @xmath372 by continuity of @xmath315 .
now we use the definition of @xmath334 in ( [ equ2.24 ] ) to obtain @xmath373 this shows that @xmath6-a.s . ,
@xmath374 for all @xmath9 as @xmath363 .
moreover , @xmath375 can be bounded uniformly in @xmath369 and @xmath9 , @xmath6-a.s .
, and using @xmath376 shows that we can apply dominated convergence to get @xmath377 as @xmath363 , @xmath6-a.s . with a similar argument
, we can prove ( [ equ2.33 ] ) .
indeed , for @xmath378 , we have @xmath379 and the first term on the right - hand side tends to 0 @xmath6-a.s . as @xmath380 by continuity of @xmath315 .
writing @xmath381 , we have @xmath382 as @xmath329 by the right - continuity of the stochastic integral and since @xmath334 from ( [ equ2.24 ] ) is continuous with respect to the second argument @xmath312 .
so ( [ equ2.33 ] ) will follow by dominated convergence as soon as we show that @xmath383 but the definition of @xmath334 in ( [ equ2.24 ] ) yields that @xmath384 and so ( [ equ2.37 ] ) follows again by ( [ equ2.35 ] ) and ( [ equ2.36 ] ) because @xmath385 @xmath6-a.s .
this establishes ( [ equ2.33 ] ) .
putting together all the results so far , ( [ equ2.34 ] ) therefore yields that with probability 1 , we have @xmath386 in @xmath387 as @xmath363 .
together with ( [ equ2.32 ] ) , this gives @xmath388 @xmath389-a.e .
, for each @xmath44 . for the optimal @xmath354 , we again get equality so that finally @xmath390 and combining this with ( [ equ2.31 ] ) yields ( [ equ2.25 ] ) .
\(3 ) we next show that @xmath288 for fixed @xmath9 is given by ( [ equ2.19 ] ) .
since @xmath391 satisfies ( [ equ2.18 ] ) , it s formula gives via ( [ equ2.22 ] ) and ( [ equ2.8])([equ2.11 ] ) like in ( [ equ2.23 ] ) for any @xmath392 that @xmath393\\[-8pt ] & & \hspace*{33pt}\qquad\quad { } + { q}_{r- } \vartheta_r^2 ( 1+\lambda_r^2 \delta{\langle m\rangle}_r ) + 2 x^\vartheta_{r- } \vartheta_r ( \varphi_r + \lambda_r \delta{b^{{q}}}_r ) \nonumber\\ & & \hspace*{95.4pt}\qquad\quad { } + \vartheta_r^2 \bigl ( \delta{b^{{q}}}_r ( 1+\lambda_r^2 \delta{\langle m\rangle}_r ) + { g}_r({q } ) \bigr ) \biggr){\,d}{\langle
m\rangle}_r \nonumber\\ & & \qquad= m_u - m_t + \int_t^u \biggl ( \vartheta_r \sqrt{{{\mathcal n}_{r}({q } ) } } + x^\vartheta_{r- } { \varphi_r + \lambda_r { { } ^{{{\bolds{\mathsf p}}}}{q}}_r \over\sqrt{{{\mathcal n}_{r}({q } ) } } } \biggr)^2 { \,d}{\langle m\rangle}_r .
\nonumber\end{aligned}\ ] ] by corollary [ cor1.2 ] , the process in ( [ equ2.38 ] ) is a martingale on @xmath394\!]}$ ] for the optimal @xmath288 , and so @xmath395\!]}t , t{]\!]}$.}\ ] ] integrating with respect to @xmath2 thus shows for @xmath13 that @xmath396 satisfies the linear sde @xmath397 for @xmath301 , and this implies that @xmath398 because @xmath288 is in @xmath52 , we have @xmath399 so that the stochastic exponential is indeed in @xmath29 ; and plugging ( [ equ2.40 ] ) into ( [ equ2.39 ] ) yields the expression ( [ equ2.19 ] ) for @xmath288 . since @xmath9 was arbitrary , we have now shown that ( a ) implies ( b ) and that we then have ( [ equ2.19 ] ) .
\(4 ) conversely , let us start from ( b ) .
again fix @xmath9 .
using the fact that @xmath400 solves the bsde ( [ equ2.18 ] ) , we obtain completely analogously as for ( [ equ2.38 ] ) for any @xmath27 that @xmath401\\[-8pt ] & & \qquad = m_u - m_t + \int_t^u \biggl ( \vartheta_r \sqrt{{{\mathcal n}_{r}(y ) } } + x^\vartheta_{r- } { \psi_r + \lambda_r { { } ^{{{\bolds{\mathsf p}}}}y}_r \over\sqrt{{{\mathcal n}_{r}(y ) } } } \biggr)^2 { \,d}{\langle m\rangle}_r \nonumber\end{aligned}\ ] ] for @xmath301 .
so @xmath402 is a local @xmath6-submartingale on @xmath394\!]}$ ] ; but since @xmath195 is bounded and @xmath403 for @xmath44 , we get that @xmath402 is actually a true @xmath6-submartingale on @xmath394\!]}$ ] so that @xmath404 gives @xmath405 $ ] for any @xmath44 .
the definition in ( [ equ2.1 ] ) thus yields @xmath406 for all @xmath36 $ ] . to prove the converse inequality , define the predictable process @xmath407 by the right - hand side of ( [ equ2.19 ] ) .
integrating then shows as for ( [ equ2.40 ] ) that and because this stochastic exponential is in @xmath29 by the assumption in b ) , we see that @xmath407 coming from ( [ equ2.19 ] ) is actually in @xmath52 .
plugging @xmath407 into ( [ equ2.41 ] ) shows by ( [ equ2.19 ] ) that the @xmath408-integral vanishes ; so @xmath409 is a @xmath6-martingale on @xmath394\!]}$ ] and hence @xmath410 \ge { v^0}_t(1 ) = { q}_t $ ] by ( [ equ2.1 ] ) .
so we obtain , hence also @xmath411 , @xmath412 , and this shows that any solution of ( [ equ2.18 ] ) with the properties in ( b ) coincides with @xmath332 , giving uniqueness .
finally , @xmath220 shows that @xmath413 is a @xmath6-submartingale on @xmath394\!]}$ ] for any @xmath44 and a @xmath6-martingale for @xmath414 ; so @xmath407 is optimal by corollary [ cor1.2 ] and in particular , an optimal @xmath415 exists .
since @xmath9 was arbitrary , we have also shown that ( b ) implies ( a ) , and part ( 1 ) of theorem [ theorem2.4 ] is proved .
\(5 ) it remains to prove part ( 2 ) .
but if there is some @xmath76 with @xmath77 , the space @xmath416 is closed in @xmath3 by theorem 5.2 of @xcite , for every @xmath9 , so that an optimal @xmath288 exists
. moreover , we then have @xmath144 by lemma [ lem2.1 ] , and so part ( 2 ) follows directly from part ( 1 ) . if @xmath201 , the backward equation ( [ equ2.18 ] ) looks more complicated .
using the notation from the remark before theorem [ theorem2.4 ] , in particular ( [ equ2.16 ] ) and ( [ equ2.17 ] ) , the equation reads @xmath417 where @xmath418 .
we do not give details .
for later use , we record the following consequence of theorem [ theorem2.4 ] .
[ cor2.5 ] under the assumptions of theorem [ theorem2.4 ] , suppose or there hold .
define @xmath419 where @xmath272 is the solution of the bsde ( [ equ2.18 ] ) , and recall the process @xmath121 from the quadratic representation ( [ equ1.6 ] ) of @xmath420 .
for every @xmath36 $ ] , we then have @xmath421 , \qquad\mbox{$p$-a.s.}\ ] ] and the process @xmath422 is a @xmath6-martingale on @xmath394\!]}$ ] . fix @xmath9 . because we have @xmath423 by ( [ equ2.40 ] ) and the definition ( [ equ2.42 ] ) of @xmath424 , ( [ equ2.43 ] ) follows directly from ( [ equ1.8 ] ) .
moreover , it is easy to check that for any semimartingale @xmath277 and any @xmath425 , we have @xmath426 @xmath6-a.s . on @xmath427 and @xmath428 @xmath6-a.s . on @xmath429 .
taking @xmath430 , @xmath42 and setting for brevity @xmath431 therefore gives the desired martingale property via @xmath432 \\ & = & e [ h { { } ^t{\mathcal e}}(\gamma{\cdot}s)_t | { { \mathcal f } } _ u ] ; \end{aligned}\ ] ] integrability holds since @xmath433 and @xmath434 by part ( 1b ) of theorem [ theorem2.4 ] .
as before , we can connect our results to the dual problem , as follows .
[ prop2.6 ] under the assumptions of theorem [ theorem2.4 ] , suppose or there hold . then the variance - optimal signed martingale measure @xmath159 is given by @xmath435 where @xmath400 is the solution of the bsde ( [ equ2.18 ] ) .
if we have , in addition , that @xmath436\\[-8pt ] & > & -1 , \qquad\mbox{$p$-a.s . for $ 0\le t\le t$,}\nonumber\end{aligned}\ ] ] then the vomm exists and is given by @xmath163 from ( [ equ2.44 ] ) . from the bsde ( [ equ2.18 ] ) and it s formula , we obtain by straightforward computation that the product @xmath437 is a local @xmath6-martingale , and it is even a true @xmath6-martingale since
@xmath195 is bounded and the stochastic exponential is in @xmath29 , and so ( [ equ2.44 ] ) defines a signed measure @xmath438 with @xmath6-square - integrable density process @xmath439 and @xmath440=1 $ ] . note for ( [ equ2.44 ] ) that @xmath404 .
another straightforward but slightly lengthier computation shows that @xmath441 is a local @xmath6-martingale so that @xmath159 .
finally , the representation ( [ equ2.44 ] ) of @xmath442 as a constant plus a `` good '' stochastic integral of @xmath2 implies that @xmath163 is variance - optimal ; see , for instance , lemma 2.1 in @xcite .
note here that the same argument as in step ( 4 ) of the proof of theorem [ theorem2.4 ] implies that the integrand @xmath443 is in @xmath52 so that @xmath444 is a @xmath72-martingale for every @xmath76 .
if ( [ equ2.45 ] ) holds , then clearly @xmath445 ; so @xmath163 is then equivalent to @xmath6 , hence in @xmath71 , and is the vomm . from ( [ equ2.43 ] ) , the proof of proposition [ prop2.6 ] and @xmath446 , we can see that under the assumptions of theorem [ theorem2.4 ] and ( [ equ2.45 ] ) , the process @xmath447 is a @xmath6-martingale with final value @xmath448 .
this implies that @xmath449 , \qquad 0\le t\le t.\ ] ]
recall from theorem [ theorem1.4 ] that the dynamic value process of the mean - variance hedging problem has the quadratic form @xmath451 our goals in this section are to describe the coefficient processes @xmath452 via backward stochastic differential equations ( bsdes ) and to give explicit expressions for the optimal strategies @xmath91 .
this will be done under the same assumptions as in section [ sec2 ] .
a general solution for the mvh problem has been given by @xcite in their theorem 4.10 and corollary 4.11 .
however , that solution involves either a process @xmath453 which is very hard to find [ see @xcite , definition 3.12 ] or the variance - optimal martingale measure [ called @xmath454 in @xcite ; see their proposition 3.13 ] which is also notoriously difficult to determine . with our approach
, we can be more explicit . to formulate our main result
, we introduce the system of bsdes , @xmath455\\[-8pt ] & & { } + { \psi^{(1)}}_s{\,d}m_s + d{l^{(1)}}_s,\qquad { y^{(1)}}_t = h,\nonumber \\ \label{equ3.3 } d{y^{(0)}}_s & = & { ( { \psi^{(1)}}_s+\lambda_s { } ^{{{\bolds{\mathsf p}}}}{y^{(1)}}_s)^2 \over{{\mathcal n}}_s({y^{(2 ) } } ) } { \,d}{\langle m\rangle}_s + d{n^{(0)}}_s,\qquad { y^{(0)}}_t = h^2.\end{aligned}\ ] ] a _ solution _ of this system consists of tuples @xmath456 , @xmath457 , @xmath458 where @xmath459 are in @xmath274 ; @xmath460 are in @xmath461 and strongly @xmath6-orthogonal to @xmath199 ; @xmath462 is a local @xmath6-martingale ; and @xmath463 are @xmath6-special semimartingales with @xmath464 ] \in{{\mathcal a}}_{{\mathrm{loc}}}(p)$ ] .
we point out that ( [ equ3.1 ] ) is the same equation as ( [ equ2.18 ] ) before theorem [ theorem2.4 ] .
note also that ( given @xmath465 ) the equation ( [ equ3.2 ] ) is linear and can therefore be solved explicitly ; and @xmath466 and @xmath462 for ( [ equ3.3 ] ) can even be written down directly . in the case where @xmath2 is continuous , this system has been obtained and studied in @xcite or ( under the additional assumption that @xmath192 is continuous ) in @xcite . for a markovian setting within a brownian filtration ,
the corresponding pdes can also be found in @xcite , with a heuristic treatment .
[ theorem3.1 ] suppose ( as in theorem [ theorem2.4 ] ) that @xmath165 and@xmath140 , and fix @xmath467 . then : 1 .
the following two assertions are equivalent : a. for every @xmath36 $ ] , there exists an optimal @xmath468 for ( [ equ1.2 ] ) for every @xmath17 .
b. for each @xmath17 , there is a solution to the bsde system ( [ equ3.1])([equ3.3 ] ) with : a. @xmath469 , @xmath470 , @xmath471 bounded and strictly positive , and with the property that for every @xmath472 $ ] , the process @xmath473 is in @xmath29 ; b. @xmath474 , @xmath475 , @xmath476 of class ( d ) , and such that for every @xmath36 $ ] , the solution @xmath477 of the linear sde @xmath478 on @xmath394\!]}$ ] is in @xmath29 ; c. @xmath466 is a true @xmath6-submartingale and ( hence ) of class ( d ) .
+ if or holds , then the value process @xmath420 from ( [ equ1.2 ] ) admits the representation @xmath479 where the processes @xmath19 satisfy the bsde system ( [ equ3.1])([equ3.3 ] ) , and for every @xmath36 $ ] , the optimal wealth process @xmath480 , @xmath301 , satisfies the sde ( [ equ3.4 ] ) and @xmath89 is given by the feedback formula @xmath481 2 .
suppose , in addition , that there is some @xmath83 satisfying the reverse hlder inequality @xmath77 .
then the value process @xmath420 from ( [ equ1.2 ] ) has the form ( [ equ3.5 ] ) , where the processes @xmath19 are those unique solutions of the bsde system ( [ equ3.1])([equ3.3 ] ) for which @xmath466 and @xmath476 are of class ( d ) and @xmath482 for constants @xmath483 .
moreover , for every @xmath36 $ ] , the optimal strategy @xmath91 for ( [ equ1.2 ] ) exists , and its wealth process @xmath484 satisfies the sde ( [ equ3.4 ] ) .
the integrability condition on the exponential in ( i ) is not really needed .
in fact , like in the proof of theorem [ theorem1.4 ] , one can argue that we have @xmath485 so that the integrability required in ( i ) follows from that in ( ii ) .
but for simpler comparison with theorem [ theorem2.4 ] , we have kept the formulation as a condition .
proof of theorem [ theorem3.1 ] as in the proof of theorem [ theorem2.4 ] , we denote by @xmath289 a generic local @xmath6-martingale that can change from one appearance to the next . ( 1 ) we first note that as in theorem [ theorem1.4 ] , the existence of optimal strategies @xmath114 ( for @xmath16 ) follows from ( a ) and is , by theorem [ theorem2.4 ] , equivalent to the solvability of ( [ equ3.1 ] ) such that ( i ) holds in ( b ) .
so let us start from ( a ) .
we note that ( [ equ3.5 ] ) holds due to theorem [ theorem1.4 ] , and first derive the bsde for @xmath121 . by lemma [ lem1.5 ] and the galtchouk
kunita watanabe decomposition
, we have @xmath486 with @xmath475 , @xmath474 strongly @xmath6-orthogonal to @xmath199 , and @xmath487 predictable and of finite variation .
exactly as for ( [ equ2.21 ] ) , this yields @xmath488 = m + \bigl({\psi^{(1)}}+ \lambda\delta{a^{(1)}}\bigr){\cdot}{\langle m\rangle}.\ ] ] now fix @xmath9 , recall @xmath424 from ( [ equ2.42 ] ) in corollary [ cor2.5 ] and write @xmath489 for brevity . then combining @xmath490 with the product rule , ( [ equ3.7 ] ) , ( [ equ2.7 ] ) , ( [ equ3.8 ] ) and ( [ equ2.10 ] ) yield @xmath491 but we know from corollary [ cor2.5 ] that @xmath492 is a @xmath6-martingale on @xmath394\!]}$ ] , and so the predictable finite variation term on the right - hand side of ( [ equ3.9 ] ) must be identically zero . with @xmath308 predictable and such that
@xmath493 , @xmath494 , we thus obtain that the process @xmath495 vanishes identically .
since @xmath496 , we can argue analogously to steps ( 1 ) and ( 2 ) in the proof of theorem [ theorem2.4 ] to get @xmath497 integrating with respect to @xmath78 gives @xmath498 and plugging this into ( [ equ3.7 ] ) shows that @xmath499 satisfies the bsde ( [ equ3.2 ] ) .
moreover , as already used , we know from lemma [ lem1.5 ] that @xmath122 is of class ( d ) , and it only remains for ( ii ) to check the last integrability property .
\(2 ) we next argue that the bsde ( [ equ3.3 ] ) has a solution , starting with a calculation that is used again later .
fix @xmath9 , take any @xmath1 in @xmath52 and consider as in the proof of theorem [ theorem2.4 ] the process @xmath300
, @xmath301 .
( again , we usually do not explicitly indicate the dependence of @xmath303 on the starting time @xmath9 , nor on @xmath5 . )
lemma [ lem1.5 ] yields @xmath500 , and as @xmath11 satisfies the bsde ( [ equ3.1 ] ) , the same computation as for ( [ equ2.38 ] ) gives , with ( [ equ2.42 ] ) , that @xmath501 finally , using the product rule , ( [ equ2.7 ] ) , the bsde ( [ equ3.2 ] ) for @xmath121 , ( [ equ3.8 ] ) and ( [ equ2.10 ] ) leads to @xmath502 \\ & = & dm + { v^{(1)}}_- \vartheta\lambda{\,d}{\langle m\rangle}- x^\vartheta_- \gamma \bigl({\psi^{(1)}}+ \lambda { } ^{{{\bolds{\mathsf p}}}}{v^{(1)}}\bigr ) { \,d}{\langle m\rangle}\\ & & { } + \vartheta\bigl({\psi^{(1)}}+ \lambda\delta{a^{(1)}}\bigr ) { \,d}{\langle m\rangle}\\ & = & dm + \bigl({\psi^{(1)}}+ \lambda{}^{{{\bolds{\mathsf p}}}}{v^{(1)}}\bigr ) ( \vartheta - \gamma x^\vartheta_- ) { \,d}{\langle m\rangle}.\end{aligned}\ ] ] using ( [ equ3.5 ] ) and adding up therefore gives @xmath503\\[-8pt ] & & { } - \int_t^u 2 \bigl ( { \psi^{(1)}}_r + \lambda_r { } ^{{{\bolds{\mathsf p}}}}{v^{(1)}}_r\bigr ) ( \vartheta_r - \gamma_r x^\vartheta_{r- } ) { \,d}{\langle m\rangle}_r \nonumber\\ & & { } + \int_t^u ( \vartheta_r - \gamma_r x^\vartheta_{r-})^2 { { \mathcal n}}_r\bigl({v^{(2)}}\bigr ) { \,d}{\langle m\rangle}_r + m_u - m_t . \nonumber\end{aligned}\ ] ] now choose @xmath101 and @xmath1 of the form @xmath504\!]}t , \varrho _ t{]\!]}}$ ] for some constant @xmath505 , where the stopping time @xmath506 is chosen such that @xmath1 is in @xmath52 ; this is possible because @xmath2 is in @xmath123 .
then @xmath507 and @xmath508 , and plugging this into ( [ equ3.10 ] ) and collecting terms gives @xmath509 by proposition [ prop1.1 ] , this process is always a @xmath6-submartingale on @xmath394\!]}$ ] .
so if we take a predictable @xmath308 with @xmath510 and @xmath511 , we obtain that the process @xmath512 for @xmath301 is , for all @xmath36 $ ] and @xmath505 , an increasing process . again arguing as in steps ( 1 ) and ( 2 ) of the proof of theorem [ theorem2.4 ] and using that @xmath513 when @xmath514 increases to @xmath515 ( used for the jumps ) or when @xmath515 decreases to @xmath514 ( used for the continuous part ) , we get @xmath516 because @xmath517 by lemma [ lem2.3 ]
, we conclude that @xmath518 this implies that @xmath519 is an increasing process , and since @xmath520 is @xmath6-integrable because @xmath119 is a @xmath6-submartingale by lemma [ lem1.5 ] , we obtain that @xmath521 < \infty.\ ] ] so if we define @xmath522 \nonumber\\[-8pt]\\[-8pt ] & = & \ ! :
{ n^{(0)}}_t + \int_0^t { ( { \psi^{(1)}}_r + \lambda_r { } ^{{{\bolds{\mathsf p}}}}{v^{(1)}}_r)^2 \over{{\mathcal n}}_r({v^{(2 ) } } ) } { \,d}{\langle m\rangle}_r , \nonumber\end{aligned}\ ] ] then clearly @xmath458 solves ( [ equ3.3 ] ) , and @xmath466 is a true @xmath6-submartingale .
this shows that there exists a solution to ( [ equ3.3 ] ) with ( iii ) , but we do not know yet if @xmath523 .
\(3 ) to finish the implication `` ( a ) @xmath524 ( b ) , '' we now want to prove that each @xmath525 satisfies ( [ equ3.4 ] ) and that @xmath523 .
we again fix @xmath9 , take @xmath526 and do the same calculation as in ( [ equ3.10 ] ) . completing the square then gives @xmath527 by proposition [ prop1.1 ] , this process must be a @xmath6-martingale on @xmath394\!]}$ ] if we plug in for @xmath1 the optimal @xmath91 . because both integral terms on the right - hand side are increasing due to ( [ equ3.11 ] ) , they must then both vanish identically , on @xmath394\!]}$ ] for every @xmath9 .
this first gives that @xmath528 and as @xmath500 is a @xmath6-submartingale , comparing ( [ equ3.12 ] ) and ( [ equ3.14 ] ) yields @xmath529 , hence @xmath523 , and so @xmath530 solves the bsde ( [ equ3.3 ] ) and also is the unique solution satisfying ( iii ) .
second , we obtain for the optimal strategy @xmath89 that @xmath531 which is ( [ equ3.6 ] ) in view of the definition ( [ equ2.42 ] ) of @xmath532 ; recall that @xmath533 solves ( [ equ2.18 ] ) . integrating with respect to @xmath2
shows that @xmath484 satisfies the sde ( [ equ3.4 ] ) on @xmath394\!]}$ ] , and since @xmath288 is in @xmath128 , the unique solution of ( [ equ3.4 ] ) is in @xmath29 .
so we have now proved that ( a ) implies ( b ) , and also that we then have ( [ equ3.5 ] ) and ( [ equ3.6 ] ) .
\(4 ) conversely , let us start with ( b ) ; then we have to prove the existence of an optimal @xmath91 .
fix @xmath9 , set @xmath534 for @xmath301 and use ( [ equ2.22 ] ) and the bsdes ( [ equ3.1])([equ3.3 ] ) for @xmath463 to compute as for ( [ equ3.10 ] ) and ( [ equ3.13 ] ) that for any @xmath44 , @xmath535\\[-8pt ] & & { } + \int_t^u \biggl ( ( \vartheta_r - \gamma_r x^\vartheta_{r- } ) \sqrt { { { \mathcal n}}_r\bigl({y^{(2)}}\bigr ) } - { { \psi^{(1)}}_r + \lambda_r { } ^{{{\bolds{\mathsf p}}}}{y^{(1)}}_r \over \sqrt { { { \mathcal n}}_r({y^{(2 ) } } ) } } \biggr)^2 { \,d}{\langle m\rangle}_r \nonumber\end{aligned}\ ] ] for @xmath301 .
so @xmath536 is a local @xmath6-submartingale on @xmath394\!]}$ ] ; but we also know from ( b ) that @xmath466 is of class ( d ) , @xmath471 is bounded , and @xmath476 is of class ( d ) . since @xmath303 is in @xmath29 for every @xmath302 , we see that @xmath536 is thus of class ( d ) , hence a true @xmath6-submartingale , and so @xmath537 = e\biggl [ \biggl ( h - x - \int_t^t \vartheta_r{\,d}s_r \biggr)^2 \bigg| { { \mathcal f}}_t \biggr]\ ] ] for any @xmath44 . this yields @xmath538 by ( [ equ1.2 ] ) .
conversely , if we take the solution @xmath477 of ( [ equ3.4 ] ) and define @xmath539 then integrating with respect to @xmath2 shows that @xmath540 equals @xmath477 , since both satisfy ( [ equ3.4 ] ) , and is in @xmath29 due to ( b ) so that @xmath407 is in @xmath52 .
moreover , plugging in @xmath407 for @xmath1 shows , similar to the argument for ( [ equ3.15 ] ) , that @xmath541 is a ( true ) @xmath6-martingale on @xmath394\!]}$ ] .
this implies that @xmath542 \ge { v^h}_t(x),\ ] ] and so we conclude that @xmath543 and that @xmath407 is optimal for ( [ equ1.2 ] ) , giving existence of @xmath544 .
this proves that ( b ) implies ( a ) and that we then also have @xmath545 for all @xmath5 , hence @xmath546 for @xmath547 .
this ends the proof of ( 1 ) .
\(5 ) finally , the assertion of part ( 2 ) follows , similarly to theorem [ theorem2.4 ] , from the proof of part ( 1 ) ; we only need to notice again that @xmath548 is closed in @xmath3 for every @xmath9 .
in this section , we give equivalent alternative versions for the bsdes obtained in sections [ sec2 ] and [ sec3 ] .
one reason is that in some models , these versions are more convenient to work with ; a second is that it allows us to discuss how our results relate to existing literature . for reasons of space
, we only look at ( [ equ2.18 ] ) or ( [ equ3.1 ] ) in detail ; this is the most complicated equation .
_ throughout this section , we assume as in theorem [ theorem2.4 ] that @xmath165 and @xmath140 . _ for convenience
, we recall that ( [ equ2.18 ] ) reads @xmath549 where @xmath550 and @xmath551 ] ^{{{\bolds{\mathsf p } } } } } \over d{\langle m\rangle } } , $ ] as in ( [ equ2.12 ] ) and ( [ equ2.11 ] ) .
a solution of ( [ equ4.1 ] ) is a priori a tuple @xmath400 with @xmath552 strongly @xmath6-orthogonal to @xmath199 , @xmath553 , and @xmath195 a @xmath6-special semimartingale such that @xmath213 ] \in{{\mathcal a}}_{{\mathrm{loc}}}(p)$ ] . in view of theorem [ theorem2.4 ] ( where @xmath195 is bounded ) , we restrict ourselves to solutions with @xmath282 and @xmath281 . for better comparison with ( [ equ3.1 ] ) ,
we really ought to write a superscript @xmath554 for @xmath555 , but we omit this to alleviate the notation . the bsde ( [ equ4.1 ] ) is written with the local @xmath6-martingale @xmath199 from the canonical decomposition @xmath557 of @xmath2 . in simple models with jumps , it is useful to split @xmath558 into its continuous and purely discontinuous local martingale parts @xmath559 and @xmath556 , respectively
. then @xmath560 , and we define the predictable processes @xmath561 we now consider the backward equation @xmath562 a _ solution _ of ( [ equ4.2 ] ) is a priori a tuple @xmath563 with @xmath564 strongly @xmath6-orthogonal to both @xmath565 and @xmath566 , @xmath567 , @xmath568 and @xmath195 a semimartingale with @xmath213 ] \in{{\mathcal a}}_{{\mathrm{loc}}}(p)$ ] . as for ( [ equ4.1 ] ) , we restrict our attention to solutions with @xmath569 and @xmath570 .
[ prop4.1 ] the bsdes ( [ equ4.1 ] ) and ( [ equ4.2 ] ) are equivalent .
more precisely , @xmath400 with @xmath282 and @xmath281 solves ( [ equ4.1 ] ) if and only if @xmath571 with @xmath567 , @xmath572 and @xmath570 solves ( [ equ4.2 ] ) , where the tuples are related by @xmath573 if @xmath400 solves ( [ equ4.1 ] ) , we use the galtchouk kunita watanabe decomposition of @xmath574 with respect to @xmath565 and @xmath566 to obtain ( [ equ4.3 ] ) and define @xmath575 ; so @xmath576 is strongly @xmath6-orthogonal to both @xmath565 and @xmath566 , and taking the covariation with @xmath199 and using @xmath577 gives @xmath578 . plugging this and ( [ equ4.3 ] ) into ( [ equ4.1 ] ) shows directly that @xmath571 solves ( [ equ4.2 ] ) .
conversely , if @xmath571 solves ( [ equ4.2 ] ) , we define @xmath579 and @xmath580 .
then plugging into ( [ equ4.2 ] ) directly shows that @xmath400 satisfies ( [ equ4.1 ] ) , and since @xmath581 due to the definitions above , @xmath230 is also strongly @xmath6-orthogonal to @xmath199 .
so @xmath400 solves ( [ equ4.1 ] ) .
equation ( [ equ4.2 ] ) is particularly convenient for models with simple jumps , as illustrated by : [ exa4.2 ] consider the _ jump - diffusion model _
@xmath582 where @xmath583 is a brownian motion , and @xmath584 , @xmath585 , is the compensated martingale of a simple poisson process with intensity @xmath586 .
the predictable processes @xmath587 satisfy @xmath588 and suitable integrability conditions , and we assume that @xmath589 to ensure that @xmath590 .
then we have @xmath591 , @xmath592 , @xmath593 , @xmath594 and @xmath595 .
because @xmath223 is continuous , so is @xmath596 due to ( [ equ4.2 ] ) ; hence @xmath597 by ( [ equ2.10 ] ) .
moreover , using @xmath598 = n$ ] gives @xmath599 ] ^{{{\bolds{\mathsf p}}}}}_t & = & { \bigl [ { \psi^c}{\cdot}{m^c}+ { \psi^d}{\cdot}{m^d}+ { l^\prime } , [ { m^d } ] \bigr]^{{{\bolds{\mathsf p}}}}}_t \\[-2pt ] & = & { \bigl [ { \psi^d}{\cdot}{m^d}+ { l^\prime } , ( s_- \eta)^2{\cdot}[n ] \bigr]^{{{\bolds{\mathsf p}}}}}_t \\[-2pt ] & = & ( s_-^3 { \psi^d}\eta^3){\cdot}{n^{{{\bolds{\mathsf p}}}}}_t \\[-2pt ] & = & ( s_-^3 { \psi^d}\eta^3 \alpha){\cdot}t\end{aligned}\ ] ] so that @xmath600 .
using the notation @xmath601 , @xmath602 and plugging in then allows us to rewrite the bsde ( [ equ4.2 ] ) after simple calculations as @xmath603 y_t & = & 1.\end{aligned}\ ] ] it depends on the choice of the filtration @xmath192 whether we can have a nontrivial @xmath570 strongly @xmath6-orthogonal to both @xmath565 and @xmath566 , or @xmath583 and @xmath604 . if @xmath192 is generated by @xmath583 and @xmath453 , then @xmath605 automatically by the martingale representation theorem in @xmath606 . for models with more general jumps , the version ( [ equ4.2 ] ) of the basic bsde ( [ equ4.1 ] ) is less useful because one can not easily express @xmath224 in terms of integrands like in the preceding example .
we therefore use semimartingale characteristics and , in particular , work with the jump measure of @xmath2 . for the required notation and results
, we refer to chapter ii of @xcite .
we take @xmath607 there so that @xmath608\times{{\mathbb r}}$ ] with the @xmath15-field @xmath609 , where @xmath610 is the predictable @xmath15-field on @xmath611 $ ] .
denote by @xmath612 the random measure associated with the jumps of @xmath2 and by @xmath613 its @xmath6-compensator . using proposition ii.2.9 of @xcite
, we have @xmath614 for a predictable increasing @xmath615 null at 0 .
moreover , ( [ equ2.7 ] ) gives @xmath616 and @xmath617+{\langle m\rangle}$ ] , and combining this with the construction of @xmath615 in @xcite and ( [ equ2.6 ] ) , we see that @xmath618 .
we introduce the predictable processes @xmath619 and note that @xmath620 = \sum(\delta m)^2 = ( x-\lambda\delta{\langle m\rangle})^2 * \mu^s $ ] implies that @xmath621 so that @xmath622 can be reformulated as @xmath623 with the notation @xmath624 , we now consider the backward equation @xmath625\\[-8pt ] & & { } + \int_0^t \varphi_s { \,d}{m^c}_s + w*(\mu^s-\nu)_t + { l^\prime}_t,\qquad y_t = 1 .
\nonumber\end{aligned}\ ] ] a _ solution _ of ( [ equ4.5 ] ) is a priori a tuple @xmath626 such that @xmath627 , [ see ( 3.62 ) in @xcite ] , @xmath570 strongly @xmath6-orthogonal to @xmath565 and to the space of stochastic integrals @xmath628 , and @xmath195 a @xmath6-special semimartingale with @xmath213 ] \in{{\mathcal a}}_{{\mathrm{loc}}}(p)$ ] . as before for ( [ equ4.1 ] ) and ( [ equ4.2 ] ) , we restrict our attention to solutions with @xmath629 and @xmath570 . in view of the next result , ( [ equ4.5 ] )
seems the natural form of the bsde ( [ equ4.1 ] ) or ( [ equ2.18 ] ) in the general case , because its generator is expressed in terms of integrands .
nevertheless , as seen in section [ sec2 ] , the form ( [ equ2.18 ] ) is more convenient for proving results via stochastic calculus .
[ prop4.3 ] the bsdes ( [ equ4.1 ] ) and ( [ equ4.5 ] ) are equivalent .
more precisely , @xmath400 with @xmath282 , and @xmath281 solves ( [ equ4.1 ] ) if and only if @xmath626 with @xmath627 , @xmath629 and @xmath570 solves ( [ equ4.5 ] ) , where the tuples are related by the equation @xmath630 if @xmath400 solves ( [ equ4.1 ] ) , we take its martingale part @xmath574 and represent this as @xmath631 with @xmath627 , @xmath629 , @xmath632 [ see @xcite , section 3.3b , pages 101 and 102 ] and @xmath633 with @xmath634 \equiv0 $ ] .
this is the so - called _ jacod decomposition _ ; see @xcite , theorem 3.75 , or theorem 2.4 in @xcite for a more detailed exposition .
we next express @xmath224 in terms of @xmath583 and @xmath613 .
using ( [ equ4.1 ] ) and ( [ equ4.6 ] ) yields @xmath635 moreover , @xmath636 \equiv0 $ ] so that we get @xmath637 ] = \sum\delta{n^y}(\delta s)^2 = \bigl ( x^2 \bigl(w(x ) - { \widehat w}\bigr ) \bigr ) * \mu^s + ( x^2 u(x ) ) * \mu^s.\ ] ] because @xmath213 ] $ ] is in @xmath219 , this implies that @xmath638 is in @xmath639 so that @xmath640 is a local @xmath6-martingale by @xcite , ( 3.73 ) .
hence we obtain @xmath599 ] ^{{{\bolds{\mathsf p } } } } } & = & { \bigl ( \bigl ( x^2 \bigl(w(x ) - { \widehat w}\bigr ) \bigr ) * \mu^s \bigr)^{{{\bolds{\mathsf p } } } } } = \bigl ( x^2 \bigl(w(x ) - { \widehat w}\bigr ) \bigr)*\nu\\ & = & \biggl ( \int x^2 \bigl(w(x ) - { \widehat w}\bigr ) f(dx ) \biggr){\cdot}b,\end{aligned}\ ] ] and so @xmath641 .
moreover , @xmath642 = [ s]^c + \sum(\delta s)^2 = { \langle m^c\rangle}+ x^2 * \mu^s\ ] ] gives @xmath643^{{{\bolds{\mathsf p } } } } } = { \langle m^c\rangle}+ x^2 * \nu= ( { \delta^c}+ \int x^2 f(dx ) b ) { \cdot}{\langle m\rangle}$ ] so that comparing with ( [ equ2.8 ] ) yields that @xmath644 and hence @xmath645\\[-8pt ] & = & { } ^{{{\bolds{\mathsf p}}}}y_t { \delta^c}_t + b_t \int x^2 \bigl ( { } ^{{{\bolds{\mathsf p}}}}y_t + w_t(x)-{\widehat w}_t \bigr ) f_t(dx ) .
\nonumber\end{aligned}\ ] ] if we now define @xmath646 , then ( [ equ4.6 ] ) gives @xmath647 but @xmath648 = [ { l^\prime } , s ] - [ { l^\prime } , \lambda{\cdot}{\langle m\rangle } ] =( x u(x ) ) * \mu^s + [ { { \widetilde l}},s ] - [ { l^\prime } , \lambda{\cdot}{\langle m\rangle } ] $ ] is a local @xmath6-martingale by yoeurp s lemma , and a similar argument as just above , using now that @xmath649 ; so @xmath650 and @xmath576 is strongly @xmath6-orthogonal to @xmath565 .
moreover , we have for all that @xmath651 = 0 $ ] since @xmath634\equiv0 $ ] , and so @xmath652 for all @xmath653 by @xcite , exercice 3.23 . finally , ( [ equ2.7 ] ) and yoeurp s lemma yield @xmath654^{{{\bolds{\mathsf p } } } } } \nonumber\\ & = & { [ w*(\mu^s-\nu ) , s ] ^{{{\bolds{\mathsf p } } } } } \nonumber\\[-8pt]\\[-8pt ] & = & { \bigl ( \bigl ( x \bigl(w(x ) -
{ \widehat w}\bigr ) \bigr ) * \mu^s \bigr)^{{{\bolds{\mathsf p } } } } } \nonumber\\ & = & \bigl
( x \bigl(w(x ) - { \widehat w}\bigr ) \bigr ) * \nu . \nonumber\end{aligned}\ ] ] taking in ( [ equ4.9 ] ) the covariation with @xmath199 and using also @xmath655 yields @xmath656 so that we get @xmath657 plugging ( [ equ4.11 ] ) and ( [ equ4.8 ] ) into ( [ equ4.1 ] ) and using ( [ equ4.9 ] ) , we see that @xmath626 solves ( [ equ4.5 ] ) .
conversely , if @xmath626 solves ( [ equ4.5 ] ) , then we define @xmath658 by ( [ equ4.11 ] ) and @xmath659 then @xmath282 , due to ( [ equ4.4 ] ) and because @xmath629 , and so @xmath281 .
moreover , equation ( [ equ4.10 ] ) , the definitions of @xmath230 and @xmath658 via ( [ equ4.11 ] ) and the definitions of @xmath660 and @xmath661 yield @xmath662 by the orthogonality properties of @xmath576 , so that @xmath230 is strongly @xmath6-orthogonal to @xmath199 . finally , the jacod decomposition applied to @xmath576 implies that the latter must have the form @xmath663 due to its orthogonality properties .
but then we obtain from ( [ equ4.5 ] ) again ( [ equ4.7 ] ) , hence also ( [ equ4.8 ] ) , and then plugging in shows that @xmath400 solves ( [ equ4.1 ] ) .
this completes the proof . just for completeness , but without any details , we give here the equivalent versions of the bsdes ( [ equ3.2 ] ) and ( [ equ3.3 ] ) for @xmath121 and @xmath119 .
they are @xmath664 and @xmath665 finally , the recursive representation for the optimal strategy in ( [ equ3.6 ] ) takes the form @xmath666 of course , this can equivalently be rewritten as a linear sde for @xmath667 as in ( [ equ3.4 ] ) , simply by integrating with respect to @xmath2 . at this point , it seems appropriate to comment on related work in the literature , where we restrict ourselves to papers that have used bsde techniques in the context of mean - variance hedging . while extending work by many authors done for an it process setting in a brownian filtration , the results in mania and tevzadze ( @xcite ) and @xcite still all assume that @xmath2 is continuous . at the other end of the scale ,
@xcite have a general @xmath165 , with @xmath140 ; but their methods do not exploit stochastic control ideas and results at all , and bsdes appear only very tangentially in their equations ( 3.32 ) and ( 3.37 ) . as a matter of fact , their opportunity process @xmath230 equals our coefficient @xmath11 , and so their equation ( 3.37 ) , which gives a bsde for @xmath230 , should coincide with our equation ( [ equ4.5 ] )
. however , @xcite give no proof for ( 3.37 ) and even remark that `` it is not obvious whether this representation is of any use . ''
moreover , a closer examination shows that ( 3.37 ) is not entirely correct ; it seems that they dropped the jumps of the fv part of @xmath230 somewhere , which explains why their equation has @xmath668 instead of ( the correct term ) @xmath669 .
the paper closest to our work is probably @xcite .
they first study the variance - optimal martingale measure as in @xcite via the problem dual to mean - variance hedging and obtain a bsde that describes @xmath670 ; see our proposition [ prop2.2 ] . for mean - variance hedging itself
, they subsequently describe the optimal strategy in feedback form with the help of a process ( called @xmath671 ) for which they give a bsde .
their assumptions are considerably more restrictive than ours because , in addition to @xmath165 and @xmath672 , they also suppose that @xmath2 is quasi - left - continuous ; and for the results on mean - variance hedging , they additionally even assume that @xmath673 is generated by integrals of @xmath674 [ and also that the vomm exists and satisfies the reverse hlder inequality @xmath77 and a certain jump condition ] .
we found it hard to see exactly why this restrictive condition on @xmath673 is needed ; the proof in @xcite for their verification result is rather computational and does not explain where the rather technical bsdes come from .
finally , a similar ( subjective ) comment as the last one also applies to @xcite .
the problem studied there is mean - variance hedging ( not the vomm ) , and the process @xmath2 is a multivariate version of the simple jump - diffusion model in example [ exa4.2 ] , with a @xmath31-dimensional brownian motion @xmath583 and an @xmath289-variate poisson process @xmath453 .
the filtration used for strategies @xmath1 and payoffs @xmath0 is generated by @xmath583 and @xmath453 ; but all model coefficients ( including the intensity of @xmath453 ) are assumed to be @xmath675-predictable .
technically speaking , this condition serves to simplify lim s equation ( [ equ3.1 ] ) , which corresponds to our equation from example [ exa4.2 ] for @xmath195 without the jump term
. it would be interesting to see also at the conceptual level why the assumption is needed . as already pointed out before theorem [ theorem3.1 ] ,
the bsde system ( [ equ3.1])([equ3.3 ] ) is less complicated than it looks .
it is only weakly coupled , meaning that one can solve ( [ equ3.3 ] ) ( even directly ) once one has the solutions of ( [ equ3.1 ] ) and ( [ equ3.2 ] ) , and that ( [ equ3.2 ] ) is linear and hence also readily solved once one has the solution of ( [ equ3.1 ] ) . in general , however , ( [ equ3.1 ] ) has a very complicated driver , and it seems a genuine challenge for abstract bsde theory to prove existence of a solution directly via bsde techniques .
we do not do that ( and do not need to ) since we only use the bsdes to describe optimal strategies ; existence of the latter ( and hence existence of solutions to the bsdes ) is proved directly via other arguments . in the special case where the filtration @xmath192 is continuous , the complicated equation ( [ equ3.1 ] ) or ( [ equ2.18 ] )
can be reduced to a classical quadratic bsde , as follows .
first of all , as already pointed out before lemma [ lem2.3 ] , the operation @xmath676 in ( [ equ2.12 ] ) reduces to @xmath231 , at least in the context of ( [ equ2.18 ] ) .
so ( [ equ2.18 ] ) becomes @xmath677 and we know from lemma [ lem2.1 ] that the solution @xmath285 is strictly positive .
if we introduce @xmath678 , apply it s formula and define @xmath679 , @xmath680 , then it is straightforward to verify that ( [ equ4.12 ] ) can be rewritten as @xmath681 this can then be tackled by standard bsde methods , if desired .
in this section , we present some simple examples and special cases to illustrate our results .
we keep this deliberately short in view of the total length of the paper .
_ throughout this section _
, _ we assume that @xmath165 and @xmath140 .
_ recall the @xmath6-canonical decomposition @xmath682 of our price process . because @xmath206 , the process @xmath683 is in @xmath684 with @xmath685 .
moreover , it is easy to check that @xmath686 is a local @xmath6-martingale so that @xmath687 is a so - called signed local martingale density for @xmath2 .
if @xmath687 is a true @xmath6-martingale and in @xmath688 , then @xmath689 with @xmath690 is in @xmath156 and called the _ minimal signed _ ( _ local _ ) _ martingale measure _ for @xmath2 ; if even @xmath691 so that @xmath689 is in @xmath71 , then @xmath689 is the _ minimal martingale measure _ ( _ mmm _ ) for @xmath2 .
the mmm is very convenient because its density process @xmath687 can be read off explicitly from @xmath2 . on the other hand ,
the important quantity for mean - variance hedging is the variance - optimal martingale measure ( vomm ) @xmath163 . by proposition [ prop2.6 ]
, we could construct a solution to the bsde ( [ equ2.18 ] ) from @xmath163 by @xmath692 } , \qquad 0\le t\le t,\ ] ] but the density process @xmath693 is usually difficult to find . an exception is the case when @xmath694 , since then @xmath695 and the above formula allows us to find an explicit expression for @xmath11 . to make this approach work , we need conditions when @xmath163 and @xmath689 coincide .
this has been studied before , and we could give some new results , but do not do so here for reasons of space .
we only mention the mmm since it comes up later in another example . in terms of complexity
, the bsde ( [ equ2.18 ] ) or one of its equivalent forms ( [ equ3.1 ] ) , ( [ equ4.2 ] ) , ( [ equ4.5 ] ) is the most difficult one .
so we focus on that equation , in the form ( [ equ4.5 ] ) , and we try to have a solution tuple @xmath697 with @xmath698 and @xmath699 .
then ( [ equ4.5 ] ) simplifies to @xmath700 which gives @xmath701 .
but @xmath702 by ( [ equ2.10 ] ) , and plugging this in above and solving for @xmath703 allows us to get @xmath704 so that ( [ equ4.5 ] ) becomes @xmath705 this is the equation for a generalized stochastic exponential , and so it is not surprising that we can find an explicit solution .
[ cor5.1 ] set @xmath706 and suppose that @xmath707 with a constant @xmath185 and a @xmath6-martingale @xmath289 which is strongly @xmath6-orthogonal both to @xmath565 and to the space of stochastic integrals @xmath628 .
then the solution of ( [ equ4.5 ] ) is given by @xmath708 , @xmath699 and @xmath709 = { { \mathcal e}}(k)_t ( c+m_t ) , \nonumber\\[-8pt]\\[-8pt ] { l^\prime}_t & = & \int_0^t { { \mathcal e}}(k)_{s- } { \,d}m_s + [ { { \mathcal e}}(k ) , m ] _ t .
\nonumber\end{aligned}\ ] ] since ( [ equ5.0 ] ) can be written as @xmath710 , defining @xmath195 and @xmath576 by ( [ equ5.1 ] ) gives by the product rule that @xmath711 satisfy ( [ equ5.0 ] ) with @xmath404 , and @xmath576 is a local @xmath6-martingale like @xmath289 by yoeurp s lemma . finally , for every @xmath712
, we have that @xmath713 \bigr ] & = & \sum\delta\bigl ( { \bar w}*(\mu^s-\nu ) \bigr ) \delta{{\mathcal e}}(k ) \delta m\\ & = & \delta{{\mathcal e}}(k ) { \cdot } [ { \bar w}*(\mu^s-\nu ) , m]\end{aligned}\ ] ] is a local @xmath6-martingale because @xmath289 is strongly @xmath6-orthogonal to @xmath714 .
hence @xmath576 is also strongly @xmath6-orthogonal to @xmath715 , and so @xmath716 is a solution to ( [ equ4.5 ] ) .
[ exa5.2 ] a special case of corollary [ cor5.1 ] occurs if the ( final ) _ mean - variance tradeoff _ @xmath717 and all the jumps @xmath718 are _
deterministic_. then @xmath719 , the solution for @xmath195 is @xmath720 [ which is adapted because @xmath721 is deterministic ] , and all other quantities in the bsdes ( [ equ2.18 ] ) or ( [ equ4.2 ] ) or ( [ equ4.5 ] ) are identically 0 . if @xmath2 or @xmath199 or even only @xmath722 is continuous , the above expression simplifies to @xmath723 now we briefly look at the special case of a model in finite discrete time @xmath724 .
our price process is given by @xmath725 , and we assume as in ( [ equ2.7 ] ) that @xmath726 with a martingale @xmath727 null at 0 .
we assume that @xmath2 is square - integrable to avoid technical complications , and we write @xmath728 for the increments of a process @xmath729 .
the doob decomposition @xmath730 is then explicitly given by @xmath731 $ ] , we have @xmath732 = { \operatorname{var}}[\delta_k s for the discrete - time version of the bsde ( [ equ2.18 ] ) , we need @xmath735 $ ] and the density @xmath736 of @xmath737 ] ^{{{\bolds{\mathsf p}}}}}$ ] with respect to @xmath223 .
but we have @xmath738 ] = \sum_j ( \delta_j { n^y } ) ( \delta_j s)^2 $ ] so that @xmath739.\ ] ] moreover , we have @xmath740 + ( e[\delta_j s
| { { \mathcal f}}_{j-1 } ] ) ^2\nonumber\\[-8pt]\\[-8pt ] & = & e [ ( \delta_j s)^2 | { { \mathcal f}}_{j-1 } ] , \nonumber\end{aligned}\ ] ] and the galtchouk
kunita watanabe decomposition @xmath741 yields @xmath742\\[-8pt ] & = & { \operatorname{cov } } ( y_j , \delta_j s | { { \mathcal f}}_{j-1 } ) .\nonumber\end{aligned}\ ] ]
hence we get @xmath743 e[y_j | { { \mathcal f}}_{j-1 } ] \bigr)^2 \\ & = & ( e[y_j \delta_j s | { { \mathcal f}}_{j-1}])^2.\end{aligned}\ ] ] writing out the discrete - time analog of ( [ equ2.18 ] ) , expanding the ratios in the first appearing sum with @xmath744 and using ( [ equ5.3])([equ5.6 ] ) then yields @xmath745)^2 \over e[y_j ( \delta_j s)^2 | { { \mathcal f}}_{j-1 } ] } \\ & & { } + \sum_{j=1}^k \psi_j \delta_j m + l_k,\qquad y_t=1 . \nonumber\end{aligned}\ ] ] but @xmath746 gives @xmath747 = y_{j-1 } + \delta_j b^y = { n^y}_{j-1 } + b^y_j,\ ] ] and the denominator in the third sum in ( [ equ5.7 ] ) therefore equals @xmath748 = e [ y_j ( \delta_j s)^2 | { { \mathcal f}}_{j-1 } ] .\ ] ] passing to increments and taking conditional expectations to make the martingale increments vanish , equation ( [ equ5.7 ] ) thus can be written as @xmath749 = e [ y_k | { { \mathcal f}}_{k-1 } ] - { ( e[y_k \delta_k s | { { \mathcal f}}_{k-1 } ] ) ^2 \over e[y_k ( \delta_k s)^2 y_t=1.\ ] ] this is exactly the recursive relation derived in equation ( 3.1 ) in theorem 1 of @xcite ; see also equation ( 3.36 ) in @xcite . under more restrictive assumptions ,
analogous equations have also been obtained in equation ( 5 ) in theorem 2 of @xcite or in equation ( 2.19 ) in theorem 1 of @xcite .
our final example serves to illustrate the relations between our work and that of @xcite , whose assumptions are rather similar to ours .
more precisely , @xcite assumes that @xmath2 ( which he calls @xmath277 ) is locally bounded , and that the vomm @xmath163 exists in @xmath71 and satisfies the reverse hlder inequality @xmath77 and a condition on the jumps of @xmath693 .
this implies of course @xmath165 and @xmath750 .
@xcite does not use bsdes , but works with a change of numeraire as in @xcite .
his numeraire is @xmath751 $ ] , and to ensure that this is positive , the existence of the vomm @xmath163 in @xmath71 is needed .
the example below illustrates that our assumptions are strictly weaker than those of @xcite .
[ exa5.3 ] we start with two independent simple poisson processes @xmath752 with the same intensity @xmath586 and define @xmath753 , @xmath188 .
we then set @xmath754 so that @xmath2 is clearly locally bounded , hence in @xmath123 , and even quasi - left - continuous .
we claim that we can choose the parameters @xmath755 such that : the variance - optimal signed martingale measure @xmath756 coincides with the minimal signed martingale measure @xmath689 , but is not in @xmath71 , which means in our terminology and that of @xcite that the vomm does not exist .
let us first argue ( 2 ) . because @xmath757 implies that @xmath758 and we have @xmath759
, we obtain @xmath760 so as soon as we have @xmath761 we get @xmath762 at jumps of @xmath763 so that @xmath764 also takes negative values . because the mean - variance tradeoff process @xmath765 , @xmath188 , is deterministic ,
the signed mmm @xmath689 is variance - optimal by theorem 8 of @xcite .
moreover , @xmath687 is clearly in @xmath688 and so @xmath766 is in @xmath156 , but not in @xmath71 .
this gives ( 2 ) . to construct an element of @xmath71 ,
start with @xmath767 , which is clearly in @xmath688 . to ensure that @xmath768 , we need @xmath769 and @xmath770 .
next , the product @xmath771 is by it s formula seen to be a local @xmath6-martingale if and only if @xmath772 , which translates into the condition @xmath773 .
this allows us to rewrite ( [ equ5.8 ] ) as @xmath774 and if we choose @xmath775 , this boils down to @xmath776 and @xmath777 . by the bayes rule
, @xmath2 is then a local @xmath72-martingale under @xmath778 with @xmath779 . if we now choose @xmath780 and @xmath781 , @xmath782 , @xmath783 , @xmath784 , one readily verifies that all conditions above are satisfied ;
hence @xmath140 since it contains @xmath72 .
if we take @xmath785 , we even keep @xmath590 since @xmath786 . by its construction , the minimal martingale density @xmath687
is always based on @xmath787 . with our above choice of model parameters
@xmath775 , this is symmetric in @xmath788 and @xmath789 and therefore risks getting negative jumps rather easily .
in contrast , writing @xmath790 with @xmath791 shows that it can be very beneficial to have some extra freedom when choosing an elmm or a martingale density .
this is quite analogous to the well - known counterexample in @xcite . | we solve the problem of mean - variance hedging for general semimartingale models via stochastic control methods . after proving that the value process of the associated stochastic control problem has a quadratic structure ,
we characterize its three coefficient processes as solutions of semimartingale backward stochastic differential equations and show how they can be used to describe the optimal trading strategy for each conditional mean - variance hedging problem . for comparison with the existing literature , we provide alternative equivalent versions of the bsdes and present a number of simple examples . , , . |
single - layer quasi two - dimensional electron gases ( 2deg ) such as those formed at the interface between two dissimilar semiconductors can be routinely driven to a many - body interaction regime by application of a strong magnetic field of several teslas perpendicular to the 2deg layer @xcite . by increasing the magnetic field ,
the 2deg first enters in the integer quantum hall regime , and then in the fractional quantum hall regime @xcite .
multilayer coupled 2deg s like trilayer systems offer much more possibilities for the theoretical @xcite and experimental @xcite search of the involved physics , even at zero magnetic field .
new single - particle effects like layer ( site ) energies , tunneling coupling between layers , and many - body effects as the inter - layer coulomb coupling should be included , with the two later effects being strongly dependent on the separation between layers .
we provide in this work an exhaustive theoretical study of trilayer systems at zero - magnetic field , within the framework of a hartree - fock mean field approximation , already used for bilayer systems @xcite , and also for a simplified version of the trilayer system considered here @xcite .
the model , schematically illustrated in fig .
[ fig1 ] , includes intra - layer and inter - layer ( hopping ) kinetic energies ( not considered in ref .
[ ] ) , site ( layer ) energies ( also not considered in ref . [ ] ) , intra - layer and inter - layer exchange energies , and the usually dominant hartree electrostatic energy .
an additional external parameter is the total density of the system : we will see that as the density decreases , the trilayer system is driven to a single - particle tunneling dominated regime . by doing a numerical minimization of the total energy of the system over all the internal variables ,
we have found a zero - temperature extremely rich phase - diagram for the possible ground - state configurations .
these full numerical results may be qualitatively understood as resulting from the diverse scaling properties of the different energy contributions to the total energy of the system , with respect to the external variables like the total density or the distance between layers .
interestingly , we have found that at zero magnetic field and for close enough layers the physics of the trilayer system becomes dominated by single - particle effects _ both _ in the high - density limit ( dominated by the intra - layer kinetic energy ) and in the low - density limit ( dominated by the hopping or tunneling and site energies ) .
the interaction dominated regime is then restricted to intermediate densities ( exchange ) and large separations between layers ( hartree ) .
the wide ground - state exploration in the available parameter space ( distance between layers , total electronic density , tunneling coupling ) provided in this work may serve as a qualitative guide for the design of experimental trilayer systems .
the rest of the work is organized in the following way : in section ii we explain the model and introduce the variational hartree - fock method we use to obtain the expression for total energy of the system .
analytical and numerical ground - states resulting from the minimization of the system total energy are presented in section iii .
section iv is devoted to the conclusions . in the two appendices we discuss some limits and the physics beyond the inter - layer exchange energy contribution ( appendix a ) , and
we explain how to model the hopping parameter dependence with the layer separation ( appendix b ) .
typically , a semiconductor trilayer system is realized experimentally by confining three _ gaas _ quantum wells among _ alas _ or _ al@xmath0ga@xmath1as _
barriers of variable width and height , which gives a control on the tunneling coupling @xmath2 among layers . usually , the central @xmath3 quantum well is designed with a larger width than the two side - wells with the purpose of populate the central well , which tends to be less populated than the two side - wells@xcite . in our model of fig . 1 , we can simulate this feature by imposing , for instance , that @xmath4 .
the model for our coulomb and tunneling coupled trilayer system is represented schematically in fig . 1 , and defined through the corresponding hamiltonian in eq .
( [ hamiltonian ] ) .
it consists of three strictly two - dimensional metallic layers , at a distance @xmath5 between them .
two other layers , located at @xmath6 along the @xmath7-axis ( @xmath8 ) provide the compensating positive charge densities .
neighboring metallic layers are coupled through the tunneling term @xmath2 ( hopping ) , while all layers are coulomb coupled through the inter - layer coulomb interaction .
schematic view of the three layer system .
each layer of area @xmath9 has a two - dimensional electronic number density given by @xmath10 , @xmath11 , and @xmath12 .
charge neutrality is provided by two positively charged layers located at @xmath13 ( not shown ) , with @xmath8 .
the layers are coupled by the hopping @xmath2 and the coulomb interaction .
@xmath14 , @xmath15 , and @xmath16 are the site or layer energies.,width=325 ] in detail , the hamiltonian of our model is as follows , @xmath17 with @xmath9 denoting the area of each layer . here ,
@xmath18 ( @xmath19 ) is a creation ( annihilation ) operator for an electron in layer @xmath20 ( @xmath21 ) , with two - dimensional momentum @xmath22 and spin @xmath23 ( @xmath24 or @xmath25 ) .
each one of the five terms in the hamiltonian represents a different physical contribution to the total energy .
the first corresponds to the sum of the layer and kinetic energy of electrons in each layer , with @xmath26 ( @xmath27 being the electron effective mass of the well - acting semiconductor ) .
the second term represents the quantum tunneling of electrons among different layers , with a hopping amplitude parametrized by @xmath2 ( @xmath28 ) ; @xmath29 , and the sum is restricted to neighboring layers .
the last three coulomb - related terms represent the attractive ion ( positive layer ) - electron interaction , the ion ( positive layer ) - ion ( positive layer ) interaction , and the repulsive electron - electron interaction , respectively .
@xmath30 ( @xmath31 ) denotes the uniform density of the two layers located at left and right of the three central layers .
the system fulfills a global neutrality condition , @xmath32 , where @xmath33 represents the total two - dimensional number density , and @xmath34 ( @xmath21 ) are the number densities of each metallic layer . finally , @xmath35 , with @xmath36 ; @xmath37 is the dielectric constant of the well - acting semiconductor ( @xmath38 12.5 for _ gaas _ ) . for @xmath39 and @xmath40
, the model reduces to the one studied previously in ref . [ ] .
the layer energies are changed at will in real samples by the application of back and front gates , which in our case are represented by the two positively charged layers at @xmath41 . in this work
we consider only the symmetric configurations , i.e. @xmath42 , and @xmath43 .
the presence of the last electron - electron interaction term in eq .
( [ hamiltonian ] ) implies that no exact solution is available for the model , and forces us to attempt its approximate solution . for this
, we will employ a hartree - fock variational approximation , widely used for the bilayer case , either at zero @xcite or with magnetic field @xcite , and also used in the previous trilayer zero - tunneling and zero - site energy study @xcite .
but before that , we find convenient to perform an exact transformation of the hamiltonian , by defining the following operators : @xmath44 with @xmath45 , and @xmath46 . the new operators satisfy fermion anticommutation relations @xmath47 , while all others anticommutators vanishes .
the transformation to the `` subband '' basis @xmath48 , @xmath49 , @xmath50 from the `` layer '' basis @xmath51 , @xmath52 , @xmath53 is defined by the two angles @xmath54 and @xmath55 . for @xmath56 ,
both basis are the same . for @xmath57 , @xmath48 , @xmath49 , and
@xmath50 are creation operators for electrons in the ground- ( symmetric , no nodes ) , first - excited ( antisymmetric , one node ) , and second - excited ( symmetric , two nodes ) subband states of the trilayer system .
the canonical transformation of eq .
( [ ct ] ) may be considered as a general rotation in a 3-dimensional pseudo - spin hilbert space of a pseudo - spin spinor pointing in the @xmath58 direction , with the layer index @xmath20 playing the role of the pseudo - spin components . in the ground - state ,
the direction of @xmath59 is determined by the total energy minimum .
the coefficients in the @xmath60 transformation matrix in eq .
( [ ct ] ) are obtained from the eigenstates of the @xmath60 matrix @xmath61 , with @xmath62 , @xmath63 and @xmath64 being the components of the @xmath60 angular momentum matrices corresponding to _ unit pseudo - spin _ ( or angular momentum ) @xcite .
write in the @xmath65 basis , the hopping term in eq .
( [ hamiltonian ] ) becomes diagonal for the choice @xmath66 , but our variational ansatz given by eq .
( [ ansatz ] ) below is more flexible and allows @xmath55 and @xmath54 to take any value within their permissible range . on the other side , the transformation of eq .
( [ ct ] ) is not the more general one , considering that the angles @xmath55 and @xmath54 may be , in principle , @xmath23 and @xmath22 dependent . for simplicity , we have not included these dependences in our calculations .
after expressing the hamiltonian in eq .
( [ hamiltonian ] ) in term of the subband operators , we have taken the expectation value of the transformed hamiltonian with the following hartree - fock variational ansatz for the ground state - vector @xcite , @xmath67 here , @xmath68 , @xmath69 , and @xmath70 are the fermi wavectors for electrons with spin @xmath23 in subbands @xmath71 , @xmath72 , and @xmath73 , respectively .
if any of the six @xmath74 , this means that the corresponding subband is empty .
we obtain , after a lengthy calculation @xmath75 \ ; , { \nonumber}\\ & = & \frac{2}{r_s^2 } \big [ e_0^{\,\text k } + e_0^{\,\text t } + e_0^{\,\text h } + e_0^{\,\text{x - intra } } + e_0^{\,\text{x - inter}}\big ] \ ; , { \nonumber}\\ \label{hf}\end{aligned}\ ] ] where @xmath76 \right . { \nonumber}\\ & + & \left . \varepsilon_1^ * \left [ \eta_a \cos^4\left(\frac{\theta}{2}\right ) + \eta_{b } \frac{\sin^2\theta}{2 } + \eta_c \sin^4\left(\frac{\theta}{2}\right ) \right ] \right . { \nonumber}\\ & + & \left . \varepsilon_3^ * \left [ \eta_a \sin^4\left(\frac{\theta}{2}\right ) + \eta_{b } \frac{\sin^2\theta}{2 } + \eta_c \cos^4\left(\frac{\theta}{2}\right ) \right ] \right\ } \ ; , { \nonumber}\\ \label{k}\end{aligned}\ ] ] @xmath77 @xmath78 { \nonumber}\\ & -&\frac{\sin^4\theta}{8}(\eta_a-2\eta_b+\eta_c)^2 \bigg\ } \ ; , \label{hartree}\end{aligned}\ ] ] @xmath79 and @xmath80 here , @xmath81 is the dimensionless two - dimensional density parameter , @xmath82 is the effective bohr radius of the well - acting semiconductor , and @xmath83 is the effective rydberg . for _ gaas _ as well - acting material , @xmath84 with @xmath85 denoting the bare electron mass , and @xmath86 , resulting in @xmath87 mev , and @xmath88 . also , @xmath89 , @xmath90 are the total and spin - discriminated subband occupation factors , @xmath91 is the distance from the central layer to the two lateral layers ( in units of the effective bohr radius @xmath92 ) , @xmath93 , and @xmath94 .
the expression for the quantities @xmath95 is given in the appendix a. these exchange integrals correspond to the overlap area between two fermi circles , with radius @xmath96 and @xmath97 , with the circles centers separated by a distance @xmath98 .
the overlap area is maximum for @xmath99 , and decreases as @xmath98 increases ; when @xmath100 the superposition area ( and the associated exchange integral ) becomes zero .
all energies in eq .
( [ hf ] ) are given in units of @xmath101 , i.e. , in units of energy per unit area .
as defined above , all @xmath102 , and @xmath103 .
@xmath104 corresponds to the sum of the intra - layer kinetic and layer energy contributions . for the case
@xmath105 , the layer energy contribution reduces to an uninteresting constant term , which only depends on the total electron density @xcite .
@xmath106 is the tunneling or inter - layer kinetic energy , and is the only term where the angle @xmath55 appears .
@xmath107 is the hartree electrostatic energy , and @xmath108 and @xmath109 are the intra - subband and inter - subband exchange - energy contributions , respectively . for a given set of external parameters (
@xmath110 , @xmath111 , @xmath112 , @xmath113 , @xmath114 , @xmath115 ) , the ground - state energy @xmath116 depends on eight variational parameters : @xmath117 . however , the neutrality condition @xmath118 allows us to eliminate one of the six subband occupation factors .
regarding the hopping term , since @xmath119 for @xmath45 , for having @xmath120 , the condition is that @xmath121 .
as noted above , the total energy in eq .
( [ hf ] ) is invariant under the exchange of the subband labels `` @xmath71 '' and `` @xmath73 '' . and
as the angle @xmath55 only enters through the hopping term @xmath106 , its value is just determined by the sign of @xmath122 . for example , by assuming that we restrict ourselves to the configurations with @xmath123 , the optimum value for @xmath55 is @xmath124 . having assumed instead that @xmath125 for all possible configurations , the optimum value for @xmath55 will be @xmath126 .
both choices are equivalent , and we have adopted here the first one : @xmath127 . under these constraints ,
@xmath116 has been minimized numerically with respect to the remaining six variational parameters : five occupation factors and the layer mixing angle @xmath54 .
the corresponding results are presented in figs . 2 - 7 .
the scaling with @xmath110 and @xmath111 of the different energy contributions deserves some discussion . in the low - density limit , @xmath110 grows since it is inversely proportional to the square root of the density and the physics is dominated by the tunneling and layer contributions , which are single - particle effects . on the other side , in the high - density limit
@xmath110 is small , and the main contributions to the total energy comes from the intra - layer kinetic and hartree terms . for large enough @xmath111 , on the other side ,
the system always becomes dominated by the hartree term , which favors an effective bilayer configuration with the central well empty , and with electrons distributed equally in the two side layers @xcite .
the intra - layer exchange interaction of eq .
( [ x - intra ] ) scales linearly with @xmath110 , is always negative , and is important in favoring spin - polarized ground - state configurations . regarding the inter - layer exchange contribution of eq .
( [ x - inter ] ) , its scaling with @xmath110 and @xmath111 is less trivial , mainly due to the presence of the ratio @xmath128 in the arguments of the exponentials .
as discussed in the appendix a , in the limit @xmath129 , and by expanding the exponentials this term acquires a leading linear dependence in @xmath111 , like the hartree electrostatic contribution . in the opposite limit @xmath130 ,
the exponential terms are small and the inter - layer exchange scales linearly with @xmath110 . according to our numerical evidence , this term is always positive and , in consequence , its optimum configuration is either @xmath131 or the spin - balanced case @xmath132 , as in both cases @xmath133 .
the total energy in the 6-parameter space is plagued by local minima , that makes the task of finding the global minimum a difficult numerical challenge .
we also have to minimize fulfilling the constraint @xmath134 , with the minimum being just at the boundary in some cases . to carry out the minimization we have partitioned the 6-parameter space in ( typically ) @xmath135 regions .
starting from a central point for each region , we find a local minimum using a simplex algorithm @xcite .
then we found the global energy minimum as the minimum among all regions .
that procedure is easy to parallelize , in particular we have implemented it using mpi ( message passing interface ) facilities @xcite . before discussing the full numerical results
, we find convenient to analyze some particular limits of the trilayer system , which admits either analytical or semi - analytical solutions
. most of the ground - state configurations will appear already in these simple limits , helping the understanding of the general cases discussed later . in the limit of small distance between layers
, the expression for the ground - state energy simplifies greatly , @xmath137 \ ; , \label{zero}\end{aligned}\ ] ] having made the choice @xmath138 , @xmath139 .
we display in fig .
[ fig2 ] the ground - state phase diagram which results from the numerical minimization of eq .
( [ zero ] ) , for the case @xmath140 @xcite .
the more prominent feature here is the growing of configuration _ b _ ( @xmath141 ) with respect to all configurations at lower values of @xmath110 , as @xmath142 increases .
this is clear physically : as @xmath131 , the configuration _
b _ is equivalent to @xmath143 , and a negative value of @xmath114 favors the central well filling .
the boundary between configurations _ a _ and _ b _ does not depend on @xmath114 , since the energy of the term proportional to @xmath114 is the same in both configurations ( see eq .
( [ zero ] ) ) .
the value @xmath144 , which is the boundary between _ a _ and _ b _ is determined then by the balance between the intra - layer kinetic energy and the intra - layer exchange term .
indeed , for @xmath145 we can apply the analytical considerations from ref .
[ ] , according to which the critical density at which a transition occurs from a configuration with @xmath146 equally occupied components to another configuration with @xmath147 equally occupied components is given by @xmath148 .
we obtain : @xmath149 , @xmath150 , @xmath151 , @xmath152 , and @xmath153 .
these analytical results are exactly the transition points at @xmath145 in fig .
[ fig2 ] , obtained numerically . for @xmath140 ( the case analyzed in fig .
[ fig2 ] ) , exists the symmetry @xmath154 .
this means , for example , that configuration _
c _ , corresponding to @xmath155 , and @xmath156 , is degenerate with the configurations @xmath157 , or @xmath158 , or @xmath159 . for finite @xmath112 , we will see how some of these degeneracies are broken .
configurations _ a _ and _ b _ are the only cases that we have found in the present work where the trilayer is actually a single - layer from the point of view of the electronic distribution , with all electrons located in the central layer . the fully spin - polarized configuration _ a _ is preferred over the unpolarized configuration _ b _ at densities such that @xmath160 by the action of the intra - subband exchange term in eq .
( [ zero ] ) , which always favors spin - polarized configurations . as most of the transitions in this work , at the boundary between
_ a _ and _ b _ configurations , the occupation factors change abruptly . as a way to understand the effect of the hopping parameter @xmath112
, we display the ground - state configurations in the parameter space @xmath161 , for @xmath162 in fig .
[ fig3 ] . from eq .
( [ zero ] ) with @xmath145 , the tunneling energy attains its minimum value with the choice @xmath163 , @xmath164 . in the low - density limit @xmath165
, the tunneling term is the dominant contribution and all the electrons are in the @xmath71-type subband ( either in the @xmath166 or @xmath167 configurations ) . in the high - density limit @xmath168 ,
the physics is dominated by the intra - subband kinetic energy term , whose energy is optimized ( lower ) by spreading electrons among the three subbands @xmath71 , @xmath72 , and @xmath73 . the intra - subband exchange energy contribution , scaling linearly with @xmath110 and favoring spin and pseudo - spin polarized states , plays an important role at intermediate densities , by favoring spin - polarized configurations over spin - unpolarized configurations as @xmath110 increases ( i.e. , @xmath169 , @xmath170 , @xmath171 ) . for increasing @xmath112 ,
the configuration @xmath167 becomes more stable and gains area in the parameter space at the expense of the remaining configurations at higher densities .
in other words , at constant @xmath110 , the occupancy of the bonding @xmath71-type subband increases with @xmath112 , for all @xmath172 , until the system falls into @xmath167 configuration . at first sight
it is not clear why the ground - state configurations for @xmath173 displayed in figs .
[ fig2 ] and [ fig3 ] are not the same .
the point here is that in fig .
[ fig2 ] ( [ fig3 ] ) we display the configurations which are the ground - states for _ finite _ values of @xmath114 ( @xmath112 ) .
the case where the three external parameters @xmath111 , @xmath112 , and @xmath114 strictly vanish is somehow ill - defined , as the total energy then reduces to just the first two terms in eq .
( [ zero ] ) , that left the angle @xmath54 undetermined .
since any value of @xmath54 is allowed , the ground - state configuration is not unique .
the same consideration applies to the results displayed in fig . 2 of ref .
ground - state phase diagram , in the @xmath174 plane , for @xmath175 , and @xmath176 . for @xmath177 ,
all configurations are the same as in fig .
[ fig3 ] . for @xmath178 , the configurations _ a _ , _ b _ , and _
f _ are the same as in fig .
[ fig2 ] . the three remaining configurations _
c _ , _ d _ , and _
e _ are similar to the corresponding ones in fig .
[ fig2 ] regarding the subband occupancies , but with @xmath179.,width=325 ] the competition between the tunneling strength parameter @xmath112 and the central layer energy @xmath114 is shown in fig .
[ fig4 ] . for @xmath178 , the @xmath114-driven configurations of fig .
[ fig2 ] are the ground - state ones ; for larger values of the tunneling parameter , the @xmath112-driven configurations of fig .
[ fig3 ] are the ones with lower energies . the boundary between configurations _ a _ and _ @xmath166 _ may be easily determined from eq .
( [ zero ] ) , and found to be @xmath180 . for @xmath181 ,
this gives @xmath182 , in agreement with the numerically calculated boundary in fig .
the boundary does not depend on @xmath110 , due to the equivalent occupation factors in configurations _ a _ ( one fully occupied spin - polarized layer ) and _ @xmath166 _ ( one fully occupied spin - polarized subband ) , and the identical quadratic scaling with @xmath110 of both tunneling and @xmath183 energy contributions .
for the same reason , the boundary between the _ b _ and _ @xmath167 _ configurations does not change with @xmath110 .
considering configurations _ c _ , _ d _ , and _
e _ , they are similar to the ones found in fig .
[ fig2 ] regarding the occupancies , but with @xmath184 .
the minimizing angle @xmath54 may be found from eq .
( [ zero ] ) , as follows .
calling @xmath185 , and given the simplicity of the @xmath186 dependence in the last terms , @xmath187 may be optimized analytically with respect to @xmath186 , @xmath188 this is quite reasonable : for @xmath189 , and given that in configurations _ c _ , _ d _ , and _
, @xmath190 , the system gains some tunneling energy by allowing that @xmath184 , although it is found numerically in all cases that the value of the minimizing angle is small .
it is worth of emphasize that at the ( _ a_,_@xmath166 _ ) boundary crossing , the trilayer system suffers an abrupt reaccommodation of the electronic charge .
this is easily seen from the following set of equations relating the occupation densities in the layer and subband basis : @xmath191 since in the _ a _ configuration @xmath192 and @xmath131 , this means in real space that @xmath193 . in the _ @xmath166 _
configuration , since instead @xmath194 and @xmath163 , this translates in real space to the distribution @xmath195 , @xmath196 . in words , at the ( _ a_,_@xmath166 _ ) transition the system passes from a monolayer to a trilayer configuration , as tunneling increases .
it is interesting to note that a `` balanced '' configuration in the @xmath197 layer space is also a `` balanced '' configuration in the @xmath198 subband space . in fig .
[ fig5 ] we display the numerical results for the ground - state configurations , in the @xmath199 plane , with the layer distance dependent hopping @xmath200 , as given by eq .
( [ tp2 ] ) in appendix b. for @xmath201 , the hopping is sizable ( @xmath202 ) , and the ground - state configurations are the same as in fig .
[ fig3 ] . for @xmath203 , the hartree term
instead becomes predominant , and the zero - tunneling and zero - site energy ground - state configurations of ref .
[ ] are obtained .
the quadratic scaling of @xmath106 with @xmath110 , however , stabilizes the tunneling - driven configurations as @xmath110 increases . in other words , for enough large values of @xmath110 , the trilayer system always enter in a tunneling dominated regime . the configuration labeled _ p2 _ , which is different from the spontaneous inter - layer coherent state _
p1 _ configuration found previously ( @xmath204 ) @xcite , results from a competition between the hopping and inter - subband exchange contributions .
p2 _ configuration may be understood as a compromise between the tunneling stabilized @xmath167 configuration , and the hartree induced _
p2 _ configuration . in passing from @xmath167 to _ p2
_ , @xmath54 vanishes and the system looses tunneling energy . in _ p2 _
, on the other side , the system keeps the gain in tunneling energy by allowing @xmath184 , but at the same time minimizes to some extent the inter - subband exchange energy contribution by inducing a more `` balanced '' subband space configuration ( one subband is occupied in @xmath167 , two subbands are occupied in _ p2 _ ) .
another interesting feature of the phase diagram in the fig .
[ fig5 ] is the behavior of the frontier between the _
( spin - polarized ) and _ @xmath167 _ ( non spin - polarized ) configurations . for @xmath175 , the limit is in the expected value of @xmath144 ( as in fig . [ fig3 ] ) .
however , when @xmath111 grows and the associated @xmath111-dependent hopping @xmath200 diminishes , the frontier between polarized and non - polarized configurations is moved to lower densities . even for the lowest density considered ( at @xmath205 ) , it is possible to find a situation with the trilayer in a paramagnetic configuration .
this is due to the @xmath111 dependence of the inter - subband exchange contribution .
as explained in appendix a , in the limit @xmath206 , one gets a linear dependence with @xmath111 for this term , as shown in eq .
( [ a5 ] ) . considering that the tunneling and hartree energies are the same in _ @xmath166 _ and _ @xmath167 _ configurations ,
the boundary between both can be found by imposing the condition that the sum of the intra - subband kinetic energy , and the intra- and inter - subband exchange energies be the same in both cases . in the limit @xmath206
, one can use the expansion in eq .
( [ a5 ] ) for the inter - subband exchange energy and obtain the equation for the boundary @xmath207 valid to linear order in @xmath111 .
the important point here is that the gain in energy of the _ @xmath167 _
configuration comes from the inter - subband exchange interaction , which in turn is a consequence that the tunneling between layers is finite , allowing a subband mixing angle @xmath184 . for even higher @xmath111 s
, the _ @xmath167 _ configuration turns into _ p2_. when the layers are more separated , the angle @xmath54 flips from @xmath208 to 0 , the system lost tunneling energy and configuration _
p2 _ becomes the one with the lowest energy .
ground - state configurations in the parameter space @xmath209 , for @xmath210 . the new configurations _
p3 _ and _ p3 _ are as follows : _ p3 _ , @xmath211 , @xmath212 ; _ p3 _ , @xmath213 .
the two empty circles correspond to the experimental sample of ref .
[ ] ( see text).,width=325 ] the only difference between fig . [ fig5 ] and fig .
[ fig6 ] is that in the former @xmath145 , while in the latter @xmath214 .
the main consequences are the disappearance of configurations _
p2 _ and _ p4 _ of fig .
[ fig5 ] in fig .
[ fig6 ] , being replaced by configurations _
p3 _ and _ p3 _ , and the growing stability of the tunneling - driven configurations on the left - hand side of the diagram .
this last feature is easy to understand : by doing @xmath215 the occupation of the central layer increases , and this favors the stability of the _ @xmath166 _ and _ @xmath167 _ configurations particularly , that when translated to layer space have a central layer with twice the occupancy of the two side layers .
this effect is also reflected in configurations _ @xmath216 _ , _ @xmath217 _ , _ @xmath37 _ , _
@xmath218 _ , although somehow in a less strength , due to the decreasing influence of tunneling and layer energies for decreasing @xmath110 .
regarding the disappearance of the _ p2 _ and _ p4 _ configurations which are present in fig .
[ fig5 ] but not in fig .
[ fig6 ] , this is due to the fact that both configurations have zero occupancy of the central layer , and this is in direct conflict with the fact that @xmath215 .
the new stable configurations _ p3 _ and _ p3 _ , on the other side , have a finite occupancy of the central layer . as in the @xmath219 case
, the non spin - polarized configuration @xmath167 is still present even at low densities @xmath220 and low and medium inter - layer distances .
however , when @xmath111 increases ( and consequently , the hopping decreases ) the configuration _ a _ is stabilized .
this configuration implies that all the electrons are in the central layer with the same spin projection .
it is stabilized by the exponential decay of the tunneling parameter with @xmath111 , that somehow recreates the situation in fig .
[ fig2 ] , in the limit of low densities . for larger @xmath111 ,
the hartree energy becomes important and a situation with the three layers occupied , like _
p3 _ , is preferred .
layer occupation factors for a trilayer with @xmath221 , @xmath214 and @xmath222 . in the upper ( lower ) panel spin discriminated ( undiscriminated ) occupation factors are shown . at the tiny dashed vertical lines ,
there is a change in the ground - state configuration.,width=325 ] finally , we display in fig .
[ fig7 ] the layer occupancies as function of @xmath110 , for fixed @xmath223 , @xmath224 and @xmath225 @xcite . for @xmath226 ,
the layer occupancies in the figure are well reproduced by the analytical estimates given in ref .
[ ] , according to which the layer occupation factors in the _ p6 _ configuration are given by @xmath227 with @xmath228 .
evaluating these expressions at @xmath221 , we obtain that @xmath229 , @xmath230 , in good agreement with the numerical values at @xmath231 in fig .
[ fig7 ] . to obtain these analytical estimates , only the intra - layer kinetic and hartree energies contributions to the total energy
are considered , as dictated by the scaling behavior of the total energy in the high - density limit .
this abrupt redistribution of charge among the layers is reminiscent of the abrupt charge transfer between the ground- and first - excited subbands of a single semiconductor quantum well , observed experimentally in ref . [ ] and discussed theoretically in ref .
[ ] . for increasing @xmath110 ,
the sequence of transitions is as follows : _ p6 _ @xmath232 _ p5 _ @xmath232 _ p3 _ @xmath232 _
_ @xmath232 _ p3 _ @xmath232 _ a_. the transitions _ p6 _
@xmath232 _ p5 _ , _ p5 _
@xmath232 _ p3 _ , and
_ p3 _ @xmath232 _ a _ are clearly discontinuous , as at each one of them a given layer occupation factor abruptly passes from a finite value to zero . the interesting `` reentrant '' sequence
_ p3 _ @xmath232 _ p3 _ @xmath232 _ p3 _ , on the other side , involves only smooth occupation changes at the boundaries .
the full sequence can be understood as a consequence of the progressive filling of the central well as @xmath110 increases .
the quadratic scaling with @xmath110 of the term proportional to @xmath114 in eq .
( [ k ] ) makes this term to become dominant for large @xmath110 , culminating with the stabilization of the _ a _ configuration , where all electrons are in the central layer , and fully spin polarized . the `` re - entrant '' sequence _
p3 _ @xmath232 _ p3 _ @xmath232 _ p3 _ is interesting , since it reflects once more the importance of the inter - layer exchange energy contribution @xmath233 , and its competition with the tunneling energy @xmath106 . in the _ p3 _ configuration , @xmath131 , as this is one of the two possible ways to cancel the positive contribution of @xmath233 , although that leads to a vanishing of the gain in tunneling energy too .
but for @xmath234 , the _
p3 _ `` balanced '' configuration becomes the one with the lowest energy . here , since @xmath235 , @xmath233 is minimized using the second option available for making it as small as possible : equal occupancy of all the subbands .
this allows that @xmath236 , and leads to a gain of the tunneling energy .
as @xmath110 increases further , the equal occupancy constraint can not be maintained any more , and the system returns to the unbalanced _
p3 _ ground - state configuration , but now with a preferential occupancy of the central layer .
it should be noted that the three parameters @xmath111 , @xmath112 , and @xmath114 used in fig .
[ fig7 ] were directly obtained from the triple semiconductor quantum well system studied experimentally in ref .
[ ] , as explained in appendix b. after adjustment of these parameters , our model reproduces qualitatively the main features of the more elaborated calculations reported in ref .
[ ] , using density - functional - theory in the local density approximation .
due to the associated computational cost , these types of calculations are usually restricted to a given particular set of parameters .
in particular , it is reported in ref . [ ] that @xmath237 at @xmath238 ( @xmath239 , corresponding to the upper empty white circle in fig .
[ fig6 ] ) , and that @xmath240 at @xmath241 ( @xmath242 , corresponding to the lower empty white circle in fig .
[ fig6 ] ) @xcite .
these lda determined layer occupation factors are in good qualitative agreement with the ones from our model , as shown in fig .
this gives us some confidence that our simple model is able to reproduce the results of more elaborate calculations , and that it may be useful for the design of real samples and for understanding the physics of trilayer semiconductor systems .
one may wonder how much of the findings of the present work can be extrapolated to other apparently similar 2deg s , as for example those formed at the interface between two band insulators , such as @xmath243 and @xmath244 @xcite .
these are fascinating systems , showing evidence of superconducting and ferromagnetic properties , simultaneously @xcite .
unfortunately , our simple theoretical model is not suitable for capturing the main physics needed for describing these 2deg s at oxide interfaces . in the first place ,
the experimental results suggest the presence of two type of carriers at the interface @xcite : one type is mobile , and presumably responsible of the superconducting features , and the other type of carriers seems to be localized , and responsible of the magnetic ( ferromagnetic ) features .
our model has only mobile carriers , which are the ones that gives all the possible ground states presented above .
other important difference is that we have assumed from the very outset that our semiconductor - based 2deg has translational invariance in the @xmath245 plane , an assumption that seems to be not justified in the case of the 2deg at oxide interfaces @xcite .
and lastly , all the evidence in this latter case points to the importance of correlation effects in describing their physics @xcite , beyond the reach of our variational hf approximation , which on the other side is reasonable for the treatment of the weakly - correlated semiconductor systems discussed here .
the possible ground - state configurations of a trilayer system have been determined , within the framework of a variational hartree - fock approximation .
the metallic layers are coulomb coupled through the inter - layer hartree and exchange interactions , and also due to the tunneling between the neighboring layers . at high - density the system becomes quite simple , as the only remaining terms in the total energy in this regime are the intra - layer kinetic energy , and the hartree classical contribution .
the low - density regime is dominated by two single - particle effects introduced in this work : the tunneling between layers and the site energies .
the inter - layer exchange interaction is found to play an important role , helping in the stabilization of the balanced configuration , in whose vicinity most of the experimental samples are designed .
it is expected that the results presented in this work for a wide range of parameters , may serve as a qualitative guide for the design of experimental samples , and also be useful for the understanding of the physics of trilayer semiconductor systems at zero magnetic field .
dm acknowledges the support of anpcyt under grant pict-2012 - 0379 .
crp and pgb thank consejo nacional de investigaciones cientficas y tcnicas ( conicet ) for partial financial support and anpcyt under grant pict-2012 - 0379 .
we acknowledge shinichi amaha for motivate us to perform the present study and for many stimulating discussions .
considering that @xmath233 is the only contribution to the total energy whose dependence with @xmath110 and @xmath111 is not fully explicit from its definition in eq .
( [ x - inter ] ) , we find convenient to analyze here some limits for it , where analytical results are available . in first place ,
the integrals @xmath95 in eq .
( [ x - inter ] ) are given by @xmath246\right .
{ \nonumber}\\ & + & \left.\eta_{\beta \sigma}\left[\cos^{-1 } \left(\frac{q - k_0}{2\sqrt{\eta_{\beta \sigma}}}\right ) -\left(\frac{q - k_0}{2\sqrt{\eta_{\beta \sigma}}}\right)\sqrt{1-\left(\frac{q - k_0}{2\sqrt{\eta_{\beta \sigma}}}\right)^2}\right ] \right\ } \ ; , { \nonumber}\\\end{aligned}\ ] ] and @xmath247 . here , @xmath248 , with @xmath249 , and @xmath250 .
for @xmath251 , the expression simplifies to @xmath252.\end{aligned}\ ] ] in the limit @xmath253 , and to the first order in @xmath128 , the term @xmath254 can be calculated explicitly : @xmath255 using the relation @xcite @xmath256 it follows from eq .
( [ x - inter - d-0 ] ) that @xmath257 { \nonumber}\\ & + & \frac{d^ { * } \sin^{4}\theta}{4}\sum_{\sigma}\left(\eta_{a\sigma}-\eta_{c\sigma}\right)^{2}. \label{a5}\end{aligned}\ ] ] from the above expression is easy to see that since in this limit @xmath254 is always positive , it is minimized either for @xmath131 or when all subbands are equally occupied . assuming the case @xmath236 , the equally occupied situation may be realized in two different ways : @xmath258 spin - polarized , with @xmath259 and all spin - down occupancies equal to zero , and @xmath260 spin - unpolarized situation with all subband occupancies for both spins equally occupied and having the value 1/6 .
owing to the presence of the remaining contributions to the total energy , the spin - polarized situation is more stable in the low - density limit , while the spin - unpolarized situation is preferred in the high - density limit .
in a tight - binding like approximation , the hopping parameter between two neighboring semiconductor quantum wells separated by a distance @xmath5 may be estimated from the expression @xcite @xmath261 here , @xmath262 and @xmath263 are the envelope normalized wavefunctions corresponding to the left and right quantum wells , respectively , and @xmath264 is the potential barrier between the two wells . consistently with our trilayer model in fig . 1
, we have approximated each quantum well by an isolated attractive delta - potential of strength @xmath166 . within this model , @xmath265 ,
@xmath266 $ ] , with @xmath166 having units of energy times length , and @xmath267 . replacing in eq .
( [ tp ] ) , and imposing the constraint that @xmath268 ( a reasonable physical choice ) we obtain @xmath269 the free parameter @xmath270 can be fixed now by imposing the second condition that for a given value of @xmath111 , the hopping parameter should be equal to some convenient value , usually taken from a more elaborate calculation . solving equation above for @xmath270 , it yields @xmath271 as an example of the use of these equations , using eq .
( 3.6 ) in ref .
[ ] , the value of @xmath112 may be estimated from the electronic subband structure of a particular trilayer .
for instance , using the local - density - approximation theoretical results in ref .
[ ] corresponding to a trilayer with @xmath272 , we obtain that @xmath273 . replacing in eq .
( [ tp3 ] ) , @xmath274 .
this gives us the dependence of the hopping parameter with the distance between layers employed in fig .
( [ fig6 ] ) .
the parameter @xmath114 was estimated by following the considerations of hanna and macdonald in ref .
[ ] , for the same trilayer system .
g. bastard , in _ wave mechanics applied to semiconductor heterostructures _
, les editions de physique , les ulis , ( 1988 ) .
s. das sarma and a. pinczuk , in _ perspectives in the quantum hall effect _ , ( wiley , new york , 1997 ) . c. b. hanna and a. h. macdonald , phys .
b @xmath275 , 15981 ( 1996 ) .
h. d. herce and c. r. proetto , solid state comm .
@xmath276 , 395 ( 2000 ) .
j. ye , phys .
b @xmath277 , 125314 ( 2005 ) . j. jo , y. w. suen , l. w. engel , m. b. santos , and m. shayegan , phys . rev .
b @xmath278 , 9776 ( 1992 ) .
t. s. lay , s. p. shukla , j. jo , x. ying , and m. shayegan , surface science * 361/362 * , 171 ( 1996 ) . s. p. shukla , y. w. suen , and m. shayegan , phys .
@xmath279 , 693 ( 1998 ) .
g. m. gusev , s. wiedmann , o. e. raichev , a. k. bakarov , and j. c. portal , phys . rev .
b @xmath280 , 161302(r ) ( 2009 ) .
s. wiedmann , n. c. mamani , g. m. gusev , o. e. raichev , a. k. bakarov , and j. c. portal , phys .
b @xmath280 , 245306 ( 2009 ) . s. wiedmann , n. c. mamani , g. m. gusev , o. e. raichev , a. k. bakarov , and j. c. portal , physica e @xmath281 , 1088 ( 2010 ) l. fernandes dos santos , b. g. barbosa , g. m. gusev , j. ludwig , d. smirnov , a. k. bakarov , and yu .
a. pusep , phys .
b @xmath282 , 195113 ( 2014 ) . c. b. hanna , d. haas , and j. c. daz - vlez , phys .
b @xmath283 , 13882 ( 2000 ) . c. r. proetto , phys .
@xmath284 , 3723 ( 1999 ) .
s. p. shukla , m. shayegan , t. jungwirth , and a. h. macdonald , phys .
@xmath284 , 3724 ( 1999 ) .
l. brey , phys .
@xmath285 , 903 ( 1990 ) .
t. jungwirth , s. p. shukla , l. smrcka , m. shayegan , and a. h. macdonald , phys .
@xmath279 , 2328 ( 1998 ) . t. jungwirth and a. h. macdonald , phys .
b @xmath286 , 035305 ( 2000 ) . s. gasiorowicz , in _ quantum physics _ ,
page 239 , ( wiley , new york , 1996 ) .
this ansatz exclude the possibility of have some kind of `` canted '' spin configuration as ground - state in the trilayer , similar to the canted phases found in bilayers in the quantum hall regime ( see for instance a. h. macdonald , r. rajaraman , and t. jungwirth , phys .
b @xmath287 , 8817 ( 1999 ) ) .
we believe that this is not a serious limitation for our zero magnetic - field study .
m. galassi _
_ , _ gnu scientific library reference manual ( 3rd ed . ) _ , isbn 0954612078 .
m. snir , s. w. otto , s. huss - lederman , d. w. walker , and j. dongara , _ mpi : the complete reference _
( mit press , cambridge , ma , 1996 ) .
the fact the we consider _ together _ that @xmath175 and @xmath140 deserves some discussion . in the first place , it is interesting to explore the possibility that the trilayer system supports an `` spontaneous interlayer coherent state '' , widely discussed in bilayers , and corresponding to the appearance of a ground - state configuration with @xmath236 _ even _ when @xmath140 . in the second place ,
as we will see later , the zero - tunneling limit is somehow recreated in the limit of large separation between the layers ( see for example how the configuration _ a _ appears in the upper right corner of fig .
[ fig6 ] ) .
a. r. goi , u. haboeck , c. thomsen , k. eberl , f. a. reboredo , c. r. proetto , and f. guinea , phys .
b @xmath285 , 121313(r ) ( 2002 ) . s. rigamonti and c. r. proetto , phys .
@xmath288 , 066806 ( 2007 ) .
starting from this zero magnetic field nearly balanced configuration , ref .
[ ] provides experimental evidence for a magnetic - field induced trilayer @xmath232 bilayer transition , due to a spontaneous charge transfer from the central layer towards the two side layers @xcite .
j. mannhart , d. h. a. blank , h. y. hwang , a. j. millis , and j .-
triscone , mrs bulletin @xmath289 , 1027 ( 2008 ) ; a. j. millis , nature physics @xmath290 , 749 ( 2011 ) .
j. a. bert , b. kalisky , c. bell , m. kim , y. hikita , h. y. hwang , and k. a. moler , nature physics @xmath290 , 767 ( 2011 ) . m. sing , g. berner , k. go , a. mller , a. ruff , a. wetscherek , s. thiel , j. mannhart , s. a. pauli , c. w. schneider , p. r. willmott , m. gorgoi , f. schfers , and r. claessen , phys .
@xmath291 , 176805 ( 2009 ) .
h. y. hwang , y. iwasa , m. kawasaki , b. keimer , n. nagaosa , and y. tokura , nature materials @xmath292 , 103 ( 2012 ) . | the ground - state electronic configuration of three coupled bidimensional electron gases has been determined using a variational hartree - fock approach , at zero magnetic field .
the layers are coulomb coupled , and tunneling is present between neighboring layers . in the limit of small separation between layers , the tunneling becomes the dominant energy contribution , while for large distance between layers the physics is driven by the hartree electrostatic energy .
transition from tunneling to hartree dominated physics is shifted towards larger layer separation values as the total bidimensional density of the trilayers decreases .
the inter - layer exchange stabilizes a `` balanced '' configuration , where the three layers are approximately equally occupied ; most of the experiments are performed in the vicinity of this balanced configuration .
several ground - state configurations are consequence of a delicate interplay between tunneling and inter - subband exchange . |
the well - known cshkp model @xcite for two - ribbon flares states that the flares result from magnetic reconnection in the corona , accompanied by acceleration of energetic electrons and the bulk heating of coronal plasma .
the nonthermal electrons produced in this process travel along the magnetic field lines toward the cold and dense chromosphere , in which they produce hard x - ray ( hxr ) emission by bremsstrahlung and heat the chromospheric plasma by coulomb collisions . in addition , direct thermal conduction from the corona can also contribute to the heating of the chromosphere .
the heated chromospheric material then expands upward to the coronal loop and finally forms the hot soft x - ray ( sxr ) flare loop .
such a process , which was first described by @xcite , is now known as chromospheric evaporation . the blue shift in extreme ultraviolet ( euv ) and sxr emission lines , evidence of this dynamic phenomenon , was observed by @xcite , @xcite , and @xcite using bent and bragg crystal spectrometer ( bcs ) on board the _ solar maximum mission _ ( _ smm _ ; @xcite ) , by @xcite , @xcite , @xcite , and @xcite using bcs on board _ yohkoh _
@xcite , by @xcite , @xcite , and @xcite using coronal diagnostic spectrometer ( cds ; @xcite ) on board the _ solar and heliospheric observatory _ ( _ soho _ ) , and recently by @xcite using the euv imaging spectrometer ( eis ; @xcite ) on _ hionde_. it is accepted that the evaporation is considered gentle when the emission lines formed in the upper chromosphere , the transition region , and the hot corona all appear to be blue shifted .
the process is said to be explosive when the lines formed in the upper chromosphere and transition region appear to be red shifted , while lines with higher formation temperatures appear to be blue shifted @xcite .
the redshifts are thought to correspond to chromospheric condensation caused by the pressure gradient induced by the violent heating of the chromosphere or the transition region by the energetic electrons .
conversion from an explosive to a gentle chromospheric evaporation during flares was reported by @xcite .
@xcite performed a numerical simulation on the radiative hydrodynamic process of the flare atmosphere by thick - target electron heating .
they found that an electron energy flux of @xmath0 ergs @xmath1 s@xmath2 acts as a threshold , above which the chromospheric evaporation becomes explosive .
when the energy flux falls below this threshold , the evaporation is gentle , in which the chromospheric plasma is depleted at a slower rate .
this theory was verified by @xcite , who found within the cds and the _ reuven ramaty high - energy solar spectroscopic image _ ( _ rhessi _ ; @xcite ) observations both a gentle chromospheric evaporation , in which the energy flux of the electron beam was below the threshold , as well as an explosive one , in which the energy flux was above the threshold .
most of the observations mentioned above were focused on energetic events , since the velocity amplitude of the evaporated mass flows is expected to be strongly related to the energy flux of nonthermal electrons .
there have been very few studies of chromospheric evaporation in small flares or microflares .
@xcite reported weak blueshifts in the line and @xcite found a clear case of gentle evaporation in microflares .
in addition , @xcite reported evidence of explosive evaporation in a small flare - like transient , although it may have been related to a nearby m - class flare .
@xcite , using _ rhessi _
data , found that the spectra of microflares can be fitted with a thermal component at low energies and a nonthermal power law component at higher energies .
furthermore , they found that the intensity enhancement at 1600 and 171 wavebands , observed by the _ transition region and coronal explorer _ ( _ trace _ ; @xcite ) , is simultaneous with the peak of the hxr light curve . @xcite
analyzed hxr and microwave emissions from microflares and corroborated the nonthermal character of these emissions .
therefore , these results suggest that the scenario of chromospheric heating predicted by the standard flare model should also apply to microflares . in this paper ,
we present evidence of explosive chromospheric evaporation in a microflare with _ goes _ soft x - ray class of b1.4 .
this implies that the electron flux in microflares is not always low but may sometimes be comparable to that in much more energetic events .
this also suggests that the same physical process exists in active events with quite different magnitudes .
we give a description of the observations of this event in [ obser ] and present our data processing method in [ data ] . the results are shown in [ result ] .
we discuss the implication of our results and present a summary in [ discu ] .
the b1.4 two ribbon flare occurred at @xmath3 4:30 ut on 2007 december 7 in noaa ar 10977 near the solar disk center .
the flare was observed by the extreme - ultraviolet imaging telescope ( eit ; @xcite ) aboard the _ soho _ spacecraft and the solar optical telescope ( sot ; @xcite ) aboard the _ hinode _ spacecraft .
the active region was scanned by eis on _ hinode _ repeatedly for the entire duration of the flare .
the eis observation of this event started at @xmath3 04:30 ut , using a 1 slit that rastered with a step of 1 and an exposure time of 30 s. the time gap between successive exposures is 2 s. the field of view ( fov ) is 5 in the scanning direction and 240 in the slit direction .
therefore , a high raster cadence of 160 s for eis was achieved , by sacrificing the spatial fov .
scanning over the active region was repeated 40 times , covering the whole evolution of the flare .
in addition , the same active region was rastered with eis from 03:27 to 04:19 ut , using the 1 slit with a step of 3 and an exposure time of 50 s. this observation corresponds to the pre - flare phase .
the fov is 180 in the scanning direction and 512 in the slit direction .
unfortunately , there is no observation by _
rhessi _ for this event .
instead , we use the derivative _ goes _ 18 light curve as a proxy of the hxr light curve , assuming that the neupert effect @xcite is applicable .
the _ goes _ 18 and 0.54 light curves and the derivative of the former are plotted in figure [ fig1 ] to estimate the impulsive phase of the flare . from this figure ,
the flare onset is at @xmath3 04:30 ut , the impulsive phase is at @xmath3 04:3504:40 ut , and the flare peak is at @xmath3 04:4004:45 ut , followed by the decay phase . to compare data sets from different wavebands , we first co - align images from different instruments .
the eit 195 image is co - aligned with the pre - flare intensity map for the @xmath4195.12 line of eis .
we track the euv bright points seen by both the instruments to obtain the offset .
the image for the h line of sot is co - aligned with the pre - flare intensity map for the @xmath4256.32 line of eis .
we use the footpoints of the loops to obtain the offset . the line - of - sight ( los ) magnetic field measured by the michelson doppler imager
( mdi ; @xcite ) aboard _ soho _ is co - aligned with the image for the h line of sot .
we then use some small magnetic structures to obtain the offset .
in addition , it is known that there exists an offset between the short waveband ( sw ) ccd , on which the @xmath4195.12 line is observed , and the long waveband ( lw ) ccd , on which the @xmath4256.32 line is observed .
the offset value commonly used is 16 in the n
s direction and 2 in the w
e direction .
finally , data from all different instruments are co - aligned to that for the line of eis .
note that the difference in the viewing angles from different instruments can be safely ignored . in this paper , we use the coordinates of the eis line for all the images from other instruments .
considering that the 2 offset between the two ccds of eis in the w
e direction is non - negligible compared to the 5 fov in the scanning direction , we only select the emission lines from the lw ccd for this study , as listed in table [ tab1 ] .
figure [ fig2 ] shows the spectrograms observed using eis for four lines in time sequence .
the vertical axis is along the slit direction and the horizontal axis is for the wavelength dispersion .
there are intensity enhancements and red shifts in the and lines , as well as co - spatial blue shifts in the and lines whose formation temperatures are much higher . to increase the signal to noise ratio , we bin the pixels in the slit direction between the two horizontal solid lines over which the line profiles do not change much .
furthermore , the binning is also done for every two columns in the scanning direction for the same purpose .
the spectra from one binned pixel are then analyzed in detail .
figure [ fig3 ] shows the eis line intensity image , the eit 195 image , the sot h image , and the los magnetic field measured by soho / mdi .
the region between the vertical lines indicates the eis fov in the scanning direction .
the cross in each panel indicates the location of the pixel to be studied in detail .
figure [ fig4 ] shows the spatial distribution of line intensity , doppler velocity , and line width for the @xmath4195.12 raster before the flare .
first , a reliable rest wavelength measurement is critical to the measurement of the doppler velocities of the evaporated plasma .
it is known that the rest centroid of lines varies because of the eis detector temperature variation during the _ hinode _ orbit .
we set the rest wavelength using the quiescent region in the eis raster from 01:15 to 03:27 ut and that from 03:27 to 04:19 ut .
the average line center over the bottom 50 in the slit direction is fitted by a sinusiodal function , which varies with time as @xmath5,\ ] ] where @xmath6 ( in ) is the average line center subtracted by the default one , i.e. , the line center listed in table [ tab1 ] , and @xmath7 ( in sec ) is the time relative to 01:15:13 ut on 2007 december 7 .
the mean fitting error is 1.1 m , corresponding to @xmath3 1.5 km s@xmath2 for the @xmath8 line .
the average line center and the fitting result for this line are shown in figure [ fig5 ] . for lines
not observed during 01:1504:19 ut , the variation curve of the @xmath4195.12 line is applied to them , since it is the strongest and cleanest line in eis observations ( harra , private communication ) .
second , line blending is also important for the measurement of doppler velocities , as well as the intensity and width of emission lines . for the lines that we select ,
the @xmath4256.32 line is blended with the @xmath4256.37 , @xmath4256.42 , and @xmath4256.41 lines @xcite .
we find that the blending is evident in the region with coronal loops ; however , the line is the dominant component at loop footpoints and in quiescent regions .
therefore , the line shows a profile that can be fitted by only one component in the region that we study .
third , for most of the emission lines that we select , adopting a one - component gaussian function can fit the line profiles very well .
however , in some cases , there appears a blue - wing asymmetry in the @xmath4264.79 and @xmath4284.16 lines .
we then need to apply a two - component gaussian fit to these lines .
finally , the function that we use to fit the line profile is @xmath9,\ ] ] where @xmath10 ( in ) is the wavelength , @xmath11 ( in ergs @xmath1 s@xmath2 sr@xmath2 @xmath2 ) is the peak value of the profile , @xmath12 ( in ) is the line center , @xmath13 ( in ) is the line width , and a and b are constants for the linear background .
when we perform a multi - component fit , the function becomes @xmath14,\ ] ] where @xmath15 is the number of gaussian components required .
we perform fits to 10 emission lines for 9 rasters covering the rise and decay phases of the flare .
the details of the physical parameters deduced from the fits are presented in the following sections .
figure [ fig6 ] shows the line profiles from the binned pixels at @xmath3 04:35 ut around the point marked by the cross in figures [ fig3 ] and [ fig4 ] .
it is clear that all the lines , except for the @xmath4284.16 line , show a symmetric gaussian profile . for the case of the line
, we need to use a two - component fit to properly account for the blue - shifted component .. note that the rest wavelength of the line can be determined through the sinusoidal fit as described in [ data ] .
] several emission lines may exist in the wavelength window of a line , e.g. , the line . in this case , a multi - component gaussian fit ( one component for each line ) is applied for the wavelength window . from figure [ fig6 ]
, we find that the line and those formed at lower temperatures show red shifts ; while the line and those formed at higher temperatures show blue shifts . the blue shift in the @xmath4284.16 line yields an upward velocity of @xmath3 100 km s@xmath2 .
the velocities for the 10 lines are plotted as a function of their formation temperature in figure [ fig7 ] .
note that the and lines are formed at nearly the same temperature ; therefore , their doppler velocities are very close , as expected . a similar behavior is found for the and lines .
a cubic polynomial and a linear function are used to fit the upward and downward flow velocities , respectively , as a function of their formation temperature .
the results show that the lines from to are red shifted by @xmath16 and those from to are blue shifted by @xmath17 , where the quantities _ v _ and _ t _ are in units of km s@xmath2 and mk , respectively .
the fitting results are shown as dashed lines in figure [ fig7 ] .
note that adopting a polynomial function of higher orders can fit all the velocity points , both positive and negative . as a test
, we find a quartic polynomial , @xmath18 , can fit all the points well except for that of .
however , doing so only makes sense mathematically without increasing the physical significance .
fortunately , the eis observations for this study cover the initial , the impulsive , and the decay phases of the flare .
we thus investigate the temporal variation of the line intensities , doppler velocities , and line widths for 9 emission lines from to .
since there are many lines in the window of the @xmath4257.26 line , it is difficult to measure and fit this line accurately in some cases .
therefore , the @xmath19 line is not selected for the study of temporal evolution behaviors . however
, this line is still included in the study of temperature - dependent velocities , since it is the only line formed at around 1.0 mk .
note that the line is blended with some other lines as discussed in [ data ] .
however , we find that the line shows a rather symmetric gaussian profile , implying the dominance of the main component , especially at the loop footpoints .
thus the line is included in the following investigation .
figure [ fig8 ] shows the temporal variation of the line intensities , doppler velocities , and line widths for the 9 emission lines .
note that for the line , the parameters for both the blue - shifted component and the stationary component are plotted .
the intensities of all the emission lines increase significantly in the impulsive phase .
this is generally consistent with the scenario predicted by the evaporation model @xcite , in which the dense chromospheric plasma is heated to temperatures similar to those in the transition region and the corona , resulting in intensity enhancements of emission lines formed at these temperatures . for the line , in particular , its two - component feature seems to contradict theoretical model that predicts no stationary component during the impulsive phase of a flare .
however , the emission of the line is dominated by the enhanced blue - shifted component , while the emission from the stationary component of the line only slightly increases during the impulsive phase .
therefore , the blue - shifted component is clearly flare - related , while the stationary component observed in the impulsive phase seems to be mostly from a non - flare emission source . in the decay phase of the flare , the intensity of each line generally decreases but remains above their pre - flare value for almost an hour .
the plots of , , , , and line intensities as a function of time reveal later second peaks during @xmath3 04:4004:50 ut ( see figure [ fig8 ] , the first and the last 4 panels in the left column ) .
this later peak in the decay phase could be induced by hotter plasma cooling back down into coronal passband .
the similarity between the and the coronal lines may be attributed to the coronal lines blended with the line , as discussed in
[ data ] . in particular , the evolution of the line is more interesting .
the later peak in this line is mainly from the stationary component and much higher than the pre - flare intensity ; the blue - shifted component contributes @xmath20 of the total intensity .
note that the stationary component is not absolutely static , as mentioned above .
it has a velocity of about 1020 km s@xmath2 .
the plasma being cooled down near the top of coronal loops , which is probably overlapped with the footpoint along the line of sight , could contribute to the enhancement of this low - velocity component . a more accurate determination of the source regions of the blue - shifted and stationary components is limited by the small fov of eis .
recently , @xcite found that the fe xv line always exhibits a blue - wing asymmetry in active regions . therefore , the blue - shifted component may be partly contributed by flows that are not related with the flare .
we should then be cautious about explaining the fitting results especially in the later stage of the flare .
however , in the impulsive phase , the intensity of the blue - shifted component is significantly enhanced .
the fitting results can reasonably reflect the plasma dynamics that is caused by the flare itself .
the widths of most of the lines follow a pattern that is well consistent with that for line intensities , i.e. , a significant increase in the impulsive phase and a gradual decrease in the decay phase .
one may notice that this pattern is not evident for the widths of the , , and lines .
note that the signal to noise ratio in these lines is lower than that in other lines , except for the time period in and shortly after the impulsive phase , when the line intensity is the most enhanced .
therefore , the measurement for these lines may be less reliable than that for other lines , as revealed by the larger error bars at some data points for the line .
nevertheless , these uncertainties do not conceal the general scenario , i.e. , the significant line broadening in the impulsive phase .
we report a case of an explosive chromospheric evaporation during the impulsive phase of a microflare .
we calculate in detail of the line intensity , the doppler velocity , and the line width for 10 emission lines formed from 0.05 to 2.5 mk . for 9 out of the 10 lines , we study further the temporal evolution .
our key findings are summarized as follows . 1
. we measure doppler velocities over an area in the flare ribbon .
the emission lines formed in the temperature range of 0.050.63 mk are red shifted and those in the range of 1.02.5 mk are blue shifted , implying a clear case of explosive chromospheric evaporation .
the upflows and the downflows show a dependence on temperature , i.e. , @xmath17 and @xmath16 , where the quantities _ v _ and _ t _ are in units of km s@xmath2 and mk , respectively .
the transition between upflows and downflows lies in a very narrow temperature range of @xmath21 mk .
3 . both a significant line intensity enhancement and a line broadening are found and correlated with the development of the flare .
the @xmath4284.16 line , the hottest line in this observation , show a blue - wing asymmetry in the impulsive phase .
thus , the line profile comprises of a relatively stationary component and a blue - shifted component .
the latter disappears in the late decay phase .
the most meaningful result is that we find explosive chromospheric evaporation in a microflare , with a division between upflows and downflows at a temperature of @xmath3 0.71.0 mk .
explosive evaporation has previously been observed in c - class and larger flares and was found to be cospatial and cotemporal with hxr emissions @xcite . in the event presented here , the upflows shown in hotter lines are definitely cospatial and cotemporal with the downflows shown in cooler lines , consistent with previous studies .
furthermore , we find that the mass flow velocities are temporally correlated with the derivative _ goes _ 18 light curve , i.e. , the proxy of hxr light curve .
the intensities of these lines are significantly enhanced in the impulsive phase and quickly decrease after the flare peak time .
therefore , the evaporation is likely induced by a nonthermal electron beam during the impulsive phase , as predicted by the theoretical electron beam heated model @xcite .
@xcite found an energy flux of @xmath0 ergs @xmath1 s@xmath2 as a threshold for the explosive evaporation , which was later verified by observations of @xcite and @xcite .
our observations imply that the energy flux of the electron beam in a microflare may be comparable to that in much bigger events , though the beam should be restricted to a rather smaller area , resulting in a very small integrated flux as observed in this b - class flare .
it is worth mentioning that @xcite found nonthermal emission in remarkably high energies ( @xmath22 50 kev ) in an a - class microflare , showing direct evidence of nonthermal electron beam heating in a microflare .
we notice that the velocities of the downflows ( 2030 km s@xmath2 ) and the upflows ( 100 km s@xmath2 ) are somewhat smaller than typical values in more energetic events
. this may partly be due to the insufficient spatial resolution of the eis scanning observations , in which the observed line profiles are in deed a convolution of that in the heated region and that in the nearby quiscent region .
owing to the smaller eis fov in this study , we are unable to estimate the exact spatial scale of the evaporation , except that it is @xmath3 10 in the n
s direction .
we suspect that the brighter flare ribbon to the west of the eis fov ( see figure [ fig3 ] , left panel ) may exhibit even more significant evaporation than what we observe in the eis fov . by comparison
, @xcite studied a microflare using eis and _ rhessi _ observations .
they found only a weak blue shift of 14 km s@xmath2 in the line in the flare ribbon without any cospatial hxr emission .
moreover , @xcite found a gentle chromospheric evaporation in the impulsive phase of a microflare , which has also no detectable hxr emission .
therefore , our event implies the possible existence of a small - scale strong electron beam in microflares . for comparison , @xcite detected a strong white - light emission , though restricted to a very small area , in a relatively small flare .
note that the white - light emission , like the explosive evaporation scenario , is also thought to be powered by a strong electron beam in the conventional point of view @xcite .
@xcite reported temperature - dependent velocities in the decay phase of an x - class flare .
they found that the upward velocities in lines of below 1.0 mk show a weaker dependence on temperature , and those in lines of higher than 1.0 mk show a stronger dependence on temperature . in our event , however , the cooler lines show downward velocities .
in addition , we observe a significant temperature dependence of upflows ( downflows ) during the impulsive phase rather than in the decay phase .
thus , the situations in these two events are quite different .
our observations are more like that of @xcite , who studied a c - class flare and found temperature - dependent velocities in the impulsive phase .
however , the transition temperature between upflows and downflows in our event is @xmath3 0.71.0 mk which is lower than the value obtained by @xcite , i.e. , @xmath3 1.52.0 mk .
theoretically , the existence of temperature - dependent velocities was predicted by @xcite , who considered a single burst of energetic electrons , and by @xcite , who adopted a continuous electron deposition in flare dynamic models .
we find that the distribution of velocities , as shown in figure [ fig7 ] , is quite consistent with the simulation results of @xcite ( see their figure 10 , panel c ) .
thus , the upward and downward velocities may be interpreted as follows .
energetic electrons deposit energy in the chromosphere , where the local pressure is enhanced to exceed that of the overlying corona .
such a pressure gradient can drive the heated plasma both upward and downward .
the short - lived euv brightenings , which last for about 5 minutes , are due to either a short duration of electron injection or a high radiative cooling rate in the condensed plasma
. we study in detail the temporal evolution of the line intensities and widths for the 9 emission lines , as well as their doppler velocities , using the high spectral resolution eis observations .
the line intensities are significantly enhanced in the impulsive phase of the flare , accompanied by the temperature - dependent velocities as mentioned above .
we find a later second peak in the decay phase ( @xmath3 04:4004:50 ut ) for intensities of coronal lines , in contrast to the intensity decrease observed in the cooler lines . the most likely explanation could be the hotter plasma with temperatures much higher than 2 mk cooling back down into the coronal passband .
unfortunately , the line is the hottest line in this observation , thus we are unable to implement a direct comparison between the light curves of coronal lines and those of lines formed in much higher temperatures .
note that such an intensity increase in the decay phase can now be cheched by the extreme ultraviolet variability experiment ( eve ; @xcite ) instrument on _ solar dynamics observatory _
( _ sdo _ ) .
we also find a significant line broadening , i.e. , the line width increases by @xmath3 60% , in association with the intensity enhancement . by checking carefully the line profiles pixel by pixel
, we confirm that the line broadening during the impulsive phase is of physical significance .
there are two interpretations of the line broadening . on one hand
, they can be interpreted as an enhancement of the nonthermal velocity induced by certain physical processes .
for example , @xcite reported that the line widths increase by 12% at the outer edge of the eit wave - associated dimming ;
@xcite interpreted the line broadenings in the cme - induced dimming as the growth of alfvn wave amplitudes . on the other hand ,
an insufficient resolution is also a possible cause .
if there exist spatially unresolved components in each pixel that are heated differently and exhibit different mass flows , the spatially integrated line profile should be more broadened than a single one in each component .
for instance , @xcite proposed an origin of the line broadening due to the multiplicity of flows .
we think that either one or both could work , i.e. , the velocity dispersion along the line of sight and/or the spatial difference lead to the observed line broadening . in summary , we report clear evidence of explosive chromospheric evaporation in a b - class microflare .
temperature - dependent upward and downward velocities and line broadenings are observed in the impulsive phase of the flare .
this event implies that the evaporated flows could possibly be induced by energetic nonthermal electrons .
the evidence , however , is not very solid , since there are no observations from _ rhessi _ for this event .
nevertheless , our findings provide a scenario in which a small flare may be locally very energetic " and possesses physical processes that were ever thought to exist only in bigger flares .
we thank the referee for constructive comments that helped to improve this manuscript .
this work was supported by nsfc under grants 10828306 and 10933003 and by nkbrsf under grant 2006cb806302 .
the authors thank ryan milligan and louise harra for insightful discussions , comments and helpful suggestions .
hinode is a japanese mission developed and launched by isas / jaxa , collaborating with naoj as a domestic partner , nasa and stfc ( uk ) as international partners .
scientific operation of the hinode mission is conducted by the hinode science team organized at isas / jaxa .
this team mainly consists of scientists from institutes in the partner countries .
support for the post - launch operation is provided by jaxa and naoj ( japan ) , stfc ( u.k . ) , nasa , esa , and nsc ( norway ) .
acton , l. w. , et al .
1980 , , 65 , 53 acton , l. w. , leibacher , j. w. , canfield , r. c. , gunkler , t. a. , hudson , h. s. , & kiplinger , a. l. 1982 , , 263 , 409 antonucci , e. , et al .
1982 , , 78 , 107 antonucci , e. , & dennis , b. r. 1983 , , 86 , 67 brosius , j. w. 2003 , , 586 , 1417 brosius , j. w. 2009 , , 701 , 1209 brosius , j. w. , & holman , g. d. 2007 , , 659 , l73 brosius , j. w. , & holman , g. d. 2009 , , 692 , 492 brosius , j. w. , & phillips , k. j. h. 2004 , , 613 , 580 carmichael , h. , in physics of solar flares , ed . w. n. hess , nasa sp-50 , washington , dc , 1964 , p451 .
chen , f. , ding , m. d. , & chen , p. f. 2010 , , 720 , 1254 culhane , j. l. , et al .
1991 , , 136 , 89 culhane , j. l. , et al .
2007 , , 243 , 19 czaykowska , a. , alexander , d. , & de pontieu , b. 2001 , , 552 , 849 delaboudinire , j .-
, et al .
1995 , , 162 , 291 ding , m. d. , watanabe , t. , shibata , k. , sakurai , t. , kosugi , t. , & fang , c. 1996 , , 458 , 391 doschek , g. a. , & warren , h. p. 2005
, , 629 , 1150 doschek , g. a. , warren , h. p. , mariska , j. t. et al .
2008 , , 686 , 1362 fisher , g. h. , canfield , r. c. , & mcclymont , a. n. 1985 , , 289 , 414 fisher , g. h. , canfield , r. c. , & mcclymont , a. n. 1985 , , 289 , 425 fisher , g. h. , canfield , r. c. , & mcclymont , a. n. 1985 , , 289 , 434 handy , b. n. , et al . 1999 , , 187 , 229 hannah , i. g. , krucker , s. , hudson , h. s. , christe , s. , & lin , r. p. 2008
, , 481 , l45 harrison , r. a. , et al . 1995 , , 162 , 233 hirayama , t. 1974 , , 34 , 323 hudson , h. s. 1972 , , 24 , 414 hudson , h. s. 1991 , , 23 , 1064 imada , s. , hara , h. , watanabe , t. , kamio , s. , asai , a. , matsuzaki , k. , harra , l. k. , & mariska , j. t. 2007 , , 59 , 793 jess , d. b. , mathioudakis , m. , crockett , p. j. , & keenan , f. p. 2008
, , 688 , l119 kahler , s. w. , meekins , j. f. , kreplin , r. w. , & bowyer , c. s. 1970 , , 162 , 293 kopp , r. a. , & pneuman , g. w. 1976 , , 50 , 85 lin , r. p. , et al . 2002 , , 210 , 3 liu , w. , petrosian , v. , & mariska , j. t. 2009 , , 702 , 1553 long , d. m. , gallagher , p. t. , mcateer , r. t. j. , & bloomfield , d. s. 2008 , , 680 , l81 mariska , j. t. , doschek , g. a. , & bentley , r. d. 1993 , , 419 , 418 mariska , j. t. 1994 , , 434 , 756 mcintosh , s. w. 2009 , , 693 , 1306 milligan , r. o. 2008 , , 680 , l157 milligan , r. o. , & dennis , b. r. 2009 , , 699 , 968 milligan , r. o. , gallagher , p. t. , mathioudakis , m. , bloomfield , d. s. , keenan , f. p. , & schwartz , r. a. 2006a , , 638 , l117 milligan , r. o. , gallagher , p. t. , mathioudakis , m. , & keenan , f. p. 2006b , , 642 , l169 neupert , w. m. 1968 , , 153 , l59 peter , h. 2010 , arxiv:1004.5403 qiu , j. , liu , c. , gary , d. e. , nita , g. m. , & wang , h. 2004 , , 612 , 530 raftery , c. l. , gallagher , p. t. , milligan , r. o. , & klimchuk , j. a. 2009 , , 494 , 1127 scherrer , p. h. , et al .
1995 , , 162 , 129 stoiser , s. , brown , j. c. , & veronig , a. m. 2008 , , 250 , 315 stoiser , s. , veronig , a. m. , aurass , h. , & hanslmeier , a. 2007 , , 246 , 339 sturrock , p. a. , in structure and development of solar active regions , ed .
, k. o. kiepenheuer , reidel , dordrecht , 1968 , p471 .
tsuneta , s. , et al .
2008 , , 249 , 167 woods , t. n. , et al .
2010 , , 3 young , p. r. , et al .
2007 , , 59 , 857 | we present a detailed study of explosive chromospheric evaporation during a microflare which occurred on 2007 december 7 as observed with the euv imaging spectrometer ( eis ) onboard _
hinode_. we find temperature - dependent upflows for lines formed from 1.0 to 2.5 mk and downflows for lines formed from 0.05 to 0.63 mk in the impulsive phase of the flare .
both the line intensity and the nonthermal line width appear enhanced in most of the lines and are temporally correlated with the time when significant evaporation was observed .
our results are consistent with the numerical simulations of flare models , which take into account a strong nonthermal electron beam in producing the explosive chromospheric evaporation .
the explosive evaporation observed in this microflare implies that the same dynamic processes may exist in events with very different magnitudes . |
electron spins in quantum dots ( qds ) are promising candidate for implementations of qubits @xcite because of their potential integration into microtechnology .
the two spin states of electron can be mapped directly to the two operational states in quantum information processing ( qip ) . a key element for spin - based qip
is the coherent manipulation of the spin states @xcite .
this qip approach requires not only rotation of unknown spin states the heart of spin - based qip , but also the spin states initialized in a known state and readout of spins
. there has been significant experimental progress in the demonstration of the key divincenzo requirements @xcite , for examples , efficient optical methods for initialization and readout of spins @xcite .
significant theoretical and experimental effort has been invested in optical manipulation of electron spin such as by using two raman - detuned laser pulses @xcite , abelian geometric phase induced by @xmath2 pulses @xcite , resonant radio - frequency pulses @xcite , and so on . using ultrafast optical pulses , press _ et al . _
reported that they have controlled and observed the spin of a single electron in a semiconductor ( over six rabi oscillations between the two spin states ) @xcite .
subsequently , the rotations of electron spins about arbitrary axes in a few picoseconds were also demonstrated in an ensemble of qds @xcite . in spin - based qip ,
in addition to preparing the spin in a precisely defined state , this state should survive long enough to allow its manipulations .
therefore a long spin coherence time is necessary .
different from the conduction electron , a valence hole has an atomic @xmath3 orbital , which has negligible overlap with the nuclei .
consequently , the suppressed hyperfine interaction leads to a longer spin coherence time than that of electron .
this may provide an attractive route to hole - spin - based applications free from the complications caused by the fluctuating nuclear spin system .
in particular , heiss _
_ reported that the spin - relaxation times of holes are up to 270 microseconds in ingaas qds embedded in a gaas diode structure @xcite .
besides long coherence times , an equally important requirement is the ability to manipulate spins coherently .
more recently , the high - fidelity hole spin initialization by optical pumping @xcite , optical control and readout of hole spin @xcite have been demonstrated experimentally .
these works promote the spin of a hole in a semiconductor qd to be the best position to be a contender for the role of a solid - state qubit .
when a quantum system governed by a hamiltonian with nondegenerate eigenstates undergoes some appropriate cyclic evolutions by adiabatically changing the controllable parameters , besides a dynamical phase , it may acquire a so - called geometric phase or berry phase @xcite .
wilczek and zee generalized the geometric phase to degenerate systems , i.e. , non - abelian geometric phase @xcite .
the geometric phase differs from the dynamic phase in that the former depends only on the geometry of the path executed , being therefore insensitive to the local inaccuracies and fluctuations .
they are thus expected to be particularly robust against noise @xcite . motivated by these work ,
we propose a method for manipulating arbitrary rotation of an unknown heavy hole ( hh ) spin state in a singly positive charged quantum dot . by applying an external magnetic field in voigt geometry
, a double tripod - shaped scheme can be configured .
most importantly , in contrast to the existing proposals based on electron spin , the hh spin rotations are realized in terms of the non - abelian berry phase , which is acquired by controlling the parameters along adiabatic loops , i.e. , stimulated raman adiabatic passage ( stirap ) @xcite and fractional stirap @xcite .
the stirap process can be used to transfer populations coherently between quantum states through dark state " which efficiently suppress relaxation .
the geometric phases accumulated during a stirap process were previously investigated for tripod systems @xcite and double-@xmath4 systems @xcite . in this paper , after briefly reviewing non - abelian geometric phase ( sec .
[ non - abelian - gp ] ) , we discuss the hole and electron energy levels of a singly positive charged qd in voigt geometry and the selection rules in sec .
[ sec - structre ] , and study the feasibility of initialization by optical pumping . in sec .
[ sec - rotations ] we show how to achieve a two - fold degenerate dark states , and how to implement the rotations about @xmath0- and @xmath1-axes by using the non - abelian geometric phase produced by the nonadiabatically coupling between the two degenerate dark states .
the fidelities of these rotations are also discussed in this section . the readout of spin state is discussed in sec .
[ sec - readout ] . in sec .
[ sec - summary ] , we end with some remarks .
our propositions of rotating the hh spin about @xmath0- and @xmath1-axes are based on non - abelian geometric phase @xcite , so we start by recalling the basic facts about non - abelian geometric phase @xcite .
we consider an @xmath5-fold degenerate eigenspace of a hamiltonian @xmath6 ( @xmath7 ) ( i.e. , the eigenspace information encoded ) depending continuously on parameters @xmath8 . based on the time - dependent schrdinger equation
, we control the parameters along loops @xmath9 in an adiabatic fashion so that the initial preparation can evolve according to @xmath10 @xmath11 denotes the path - order operator .
@xmath12 is called the gauge potential given by @xmath13 with @xmath14 being an orthonormal basis of the degenerate eigenspace .
it is worth noting that the parameter - dependent hamiltonian should evolve adiabatically so that the instantaneous state @xmath15 does not overflow the state vector space spanned by @xmath16 .
an intriguing feature of the gauge potential @xmath17 lies in that it depends only on the geometry of executed path in the space of degenerate states .
) , the stokes ( @xmath18),the driving ( @xmath19 ) fields with @xmath20 , @xmath21 and @xmath22 denoting right - hand , left - hand circular , and linear - polarized fields
. the photon bandwidth ( @xmath23 ) should be larger than the electron zeeman splitting @xmath24 but smaller than the lh zeeman splitting ( @xmath25).,width=302 ] in the scheme of rotating hh spin , the basic idea is to reconfigure a multi - level system with interacting hamiltonian possessing two - fold degenerate dark states . by changing the rabi frequencies in adiabatic fashion , we perform a loop in the parameter space . at both the beginning and the end of the cycle , we have only the hh spin states ( up and down ) . but after a loop a non - abelian geometric phase is accumulated at hh spin states
. based upon this property , arbitrary rotations of hh spin can be implemented .
we consider a singly positive charged gaas / algaas qd with growth direction @xmath1 , which can be formed naturally by interface steps in narrow quantum well ( qw ) .
the electron and holes become localized into qd regions of the qw . in the absence of magnetic field , the lowest conduction - band
( cb ) level is two - fold degenerate with respect to the spin projection @xmath26 . in the valence - band ( vb )
, the hole has a total angular momentum of 3/2 , with the projection @xmath27 ( light hole " lh ) doublet separated by more than 30 - 50 mev from the @xmath28 ( hh ) states due to confinement .
it is very large compared to the bandwidth of the picosecond and femtosecond pulsed laser , so one should be able to separate the hh and lh excitations by picosecond pulsed fields in practical applications .
the spin states of hh ( spin up and down ) trapped in the qd , which are denoted by @xmath29 and @xmath30 , are our qubit degrees of freedom .
we will perform sequentially the optical initialization , rotations of this spin by non - abelian geometrical means , and readout of a single hole spin . with @xmath21 and @xmath20 excitations ,
the only dipole allowed optical transitions from the valence hh states to the conduction electron states are @xmath31 and @xmath32 , and the light hole states can not be excited because of the frequency selection .
the angular moment restriction inhibits optical coupling between the two hh spin states .
consequently , the structure can be described by two degenerate independent two - level systems , and the spin flip raman scattering transitions are ideally dark @xcite .
the dark transitions should become bright so as to implement spin state initialization , rotations , and readout .
the problem can be solved by applying a magnetic field in the voigt geometry ( @xmath33 direction ) .
the magnetic field lifts the kramer degeneracy of the lh and electron and reconfigures the eigenstates as @xmath34 , and @xmath35 ( parallel or antiparallel to the magnetic field direction @xmath33 ) while keeping the degeneracy of the hh with a negligibly small in - plane @xmath36 factor @xcite .
the hh spin states remain unaffected by the magnetic field .
a double tripod - shaped system is therefore configured , and the energy level diagram and the associated selection rules are represented in fig .
[ level ] . in the rotations of the spin state ,
the state @xmath37 is used as an ancillary level , and it is presented by @xmath38 .
the transitions from the vb states @xmath39 , @xmath40 , and @xmath38 to the cb states @xmath41 can be excited by the pump ( @xmath21-polarized ) , the stokes ( @xmath20-polarized ) , and the driving ( @xmath22-polarized ) fields , respectively @xcite . the dark transitions therefore become bright and it is possible to initialize and rotate the hh spin states .
the photon bandwidth of the driving field ( @xmath23 ) can be controlled in a way such that it is larger than the electron zeeman splitting @xmath24 but smaller than the lh zeeman splitting ( @xmath42 ) @xcite , the excitation of state @xmath43 can be thus ignored safely . and @xmath40 as a function of time with @xmath21 continuous illumination .
different initial population distributions are considered : ( a ) @xmath44 , @xmath45 ; ( b ) @xmath46 .
the values of parameters are explained in the text.,title="fig:",width=302 ] and @xmath40 as a function of time with @xmath21 continuous illumination .
different initial population distributions are considered : ( a ) @xmath44 , @xmath45 ; ( b ) @xmath46 .
the values of parameters are explained in the text.,title="fig:",width=302 ] the initialization of hh spin can be accomplished by optical pumping @xcite .
for example , the spin up hh state can be prepared by only applying the stokes field .
the electron is excited from the vb band state to two cb band states @xmath47 or @xmath48 , and one spin up hh is left .
exciton ( electron - hole pair ) recombination occurs between the cb band electron and the vb hh with spin down or up , while the electron recombined with @xmath49 hole will be excited again .
since the pump field has not been applied , the resident hole remains in spin up state @xmath49 .
conversely , the hh can also be prepared in spin down state @xmath50 by only applying the pump field . to assess the efficiency of optical spin preparation
, we have performed numerical simulations using the liouville equation for the density matrix for the four - level system as shown in fig .
[ inilizationalize ] .
we have assumed that four exciton recombining channels ( from two cb electron states to two spin states ) proceed incoherently , and they are equal to each other ( @xmath51 ps @xcite ) . the magnetic field and
the spin - flip rates for hole and electron are taken as @xmath52 mt , @xmath53 ms @xcite , and the rabi frequencies as @xmath54 .
the hh spin state initialization with a fidelity close to 1 [ @xmath55 , we define the fidelity of the hole spin initialization as @xmath56 for @xmath20 polarization with @xmath57 ( @xmath58 ) being the population of state @xmath39 ( @xmath40 ) ] is possible to be achieved within a few times the inverse of the exciton recombining rate ( @xmath59 ns ) in the qd structure in voigt geometry .
the initialization of hh spin here is realized based on the fact that the magnetic field applied in voigt geometry reconfigures the electron eigenstates , which is different from refs .
@xcite , where the initializations are realized based on spin precession .
it is well known that , with two noncommutable rotations about two axes , any rotation can be implemented as a composite rotation @xcite . here
we design two noncommutable rotations about @xmath0- and @xmath1-axes and compose general rotations from them . in the following ,
based on nono - alelian geometric phase , we first show how to rotate the hh spin about @xmath0-axis , and then explain how to control the rotation about @xmath1-axis with the relative phase between the stokes and the driving fields . as a result , any rotations of the hh spin can be realized . in order to rotate the hh spin about @xmath0-axis
, we apply the pump field , the stokes field and the driving field to excite the corresponding transitions ( as shown in fig . [ level ] ) .
the hamiltonian in the interaction picture and in the rotating - wave approximation ( rwa ) is given by @xmath60+{\rm h.c.},\label{hamiltonian - y}\end{aligned}\ ] ] where the half rabi frequency is defined as @xmath61 with @xmath62 being the dipole moment , @xmath63 , @xmath64 denoting the pump ( the stokes and the driving ) field , and @xmath65 .
@xmath66 is the detuning and @xmath67 is the electron zeeman splitting with @xmath68 and @xmath69 representing land factor of electron and bohr magneton , respectively .
when the pump , the stokes , and the driving fields are tuned to match the conditions @xmath70 ( three fields are tuned to three - photon resonance , but they are not at the middle point of the two electron zeeman splitting levels ) and @xmath71 ( for simplicity , we choose @xmath72 and denote @xmath73 ) , one can easily find from the interaction hamiltonian ( [ hamiltonian - y ] ) that the interacting system has two degenerate dark states @xmath74 where the mixing angle @xmath75 and the additional mixing angle @xmath76 are defined as @xmath77 it is well know that the exciton recombination occurs only if there is electron excited to the cb states @xmath41 .
the two - fold degenerate dark states @xmath78 , which are known as trapped states , receive no contributions from the cb electron states [ see eqs .
( [ dark-1-y ] ) and ( [ dark-2-y ] ) ] .
hence the rotation of hh spin about @xmath0-axis is robust against the exciton recombination process , and thus leading to high fidelity operations .
it is also worth noting that , in the absence of the three fields , i.e. , all the parameters ( actually the angles @xmath75 and @xmath76 ) are fixed to zero , the previous eigenstates coincide with the two spin states @xmath79 and @xmath80 . when the three fields are applied adiabatically and hence the angles @xmath75 and @xmath76 change adiabatically , the non - abelian geometric connection components can be calculated , according to eq .
( [ gauge - potential ] ) , and we have @xmath81 with @xmath82 being the @xmath0-component pauli matrix . the related unitary operation is @xmath83 where the rotating angle @xmath84 is given by @xmath85 similar degenerate dark states and non - alelian geometric connection have been realized for ion trap @xcite and atoms @xcite .
however , in the qd structure under consideration , the non - abelian geometric connection @xmath17 is based on hh spin states . about @xmath0-axis in units of @xmath22 as a function of the delay @xmath86 .
the values of parameters are explained in the text.,width=302 ] for a quantitative analysis of the rotating angle @xmath84 about @xmath0-axis , we assume that the pump , the stokes and the driving fields have gaussian shapes as @xmath87 $ ] , @xmath88 , and @xmath89 $ ] with @xmath90 and @xmath86 being , respectively , the pulse widths and the delay . figure [ fig - gp - ry ] shows that the evolution of the rotating angle @xmath84 about @xmath0-axis in unit of @xmath22 as a function of delay @xmath86 . in the interaction , the time - dependent hamiltonian should be performed sufficiently slowly . according to the condition for adiabatic passage @xcite , we take @xmath91 , which ensures no transition between the dark and bright states ( not given in this paper ) . figure [ fig - gp - ry ] shows clearly that the rotating angle @xmath84 is delay - dependent .
it goes up successively with increasing value of the delay , and then reaches its maximum value , @xmath92 , when the delay @xmath86 is large .
thus we have shown that the rotation of hole spin about @xmath0-axis can be implemented by using the non - abelian geometric phase .
the rotation is determined only by the global property and does not depend upon the details of the evolution path in the parameters space . as suggested in ref .
@xcite , by setting @xmath93 and changing adiabatically @xmath18 and @xmath19 , the rotation about @xmath1-axis can be achieved by making use of the abelian geometric phase in our qd system . however , by involving the non - abelian geometric phase , here we suggest another method for rotating the hole spin about @xmath1-axis .
an important feature of this rotation is that the rotating angle about @xmath1-axis is not geometric phase - dependent , it is controlled by the relative phase between the stokes and the driving fields .
to do so , we set @xmath93 so that the spin down state @xmath39 is decoupled , and assume the phase of the driving field is zero , the relative phase @xmath94 is therefore the phase of the stokes field .
the time - dependent hamiltonian @xmath95 in the interaction picture and rwa takes the form @xmath96+{\rm h.c.},\end{aligned}\ ] ] where @xmath94 is the relative phase between the stokes and the drive fields . in the derivation of hamiltonian ( [ hamiltonian - z ] ) ,
the condition @xmath97 is applied . when the stokes and the driving fields are controlled to satisfy two photon resonance , only one dark state ( @xmath98 ) exists .
the hole spin can not be rotated by non - abelian geometrical means .
fortunately , however , when the two fields are tuned to the middle point of the two electron zeeman splitting levels , there is another dark state @xcite .
the two - fold degenerate dark states are @xmath99 where the mixing angle @xmath75 and the additional mixing angle @xmath76 related to the electron zeeman energy splitting @xmath100 are respectively defined as @xmath101 a similar form of double degenerate dark states in double-@xmath4 atomic scheme has been obtained @xcite . however , in our qd structure the degenerate hh spin - based dark states stem from electron zeeman splitting . in adiabatic limit ( the time derivative of the mixing angles @xmath75 and @xmath76
are small compared with the splitting of eigenvalues , given by @xmath102 ) , only the transitions between the degenerate dark states @xmath98 and @xmath103 should be taken into account , and the nonadiabatic coupling of states @xmath98 and @xmath103 to other states ( the expressions have not given in this paper ) can be safely ignored @xcite .
@xmath104 also exhibits that a nonadidabatic transition between the two - fold degenerate dark states may occur . although the dark state @xmath103 receives contribution from the cb electron states [ see eq .
( [ dark-2-z ] ) ] , we note that the electron magnetic - dependent zeeman energy splitting is independent of time . at the beginning and end of the adiabatic process , we have @xmath105 , which therefore leads to @xmath106 .
the cb electron states have no influence on the final dark state @xmath107 .
in other words , there is no electron in the cb states when the interaction is finished . ) and ( [ omegad ] ) .
( b ) the geometric phase @xmath108 in units of @xmath22 as function of the delay between two pulses @xmath86 with @xmath109 .
the values of parameters are explained in the text.,title="fig:",width=302 ] ) and ( [ omegad ] ) .
( b ) the geometric phase @xmath108 in units of @xmath22 as function of the delay between two pulses @xmath86 with @xmath109 .
the values of parameters are explained in the text.,title="fig:",width=302 ] next , we will show how to translate the phase @xmath94 into the hh spin state @xmath49 , thus leading to the rotation about @xmath1-axis with angle @xmath94 .
recalling the fact that , in the subspace , hh is prepared with spin up , namely @xmath110 .
the stokes and the driving fields are applied in the counterintuitive order , while they terminate with a constant ratio of their amplitude so that the phase @xmath94 can be introduced into the hh spin state .
this extension of stirap is called fractional stirap , which has been suggested to create the coherent atomic superpositions in a robust way @xcite . as shown in fig .
[ fig - gp - rz](a ) , the driving field consists of two parts@xmath111one with the same time dependence as the stokes field and the other coming earlier , for example @xmath112+\exp(-t^{2}/\tau^{2})\}.\label{omegad}\end{aligned}\ ] ] here @xmath113 and @xmath114 are amplitudes of the stokes and the driving fields , @xmath90 and @xmath86 are , respectively , pulse widths and delay between the two parts of driving field . during the interaction ,
the mixing angle @xmath75 varies from 0 to @xmath115 . according to the theory of non - abelian geometric phase , after the interaction
, we have @xcite @xmath116,\label{eq - rotation - d1}\end{aligned}\ ] ] where @xmath108 is given by @xmath117 equation ( [ eq - rotation - d1 ] ) exhibits that the relative phase @xmath94 translates to the hh spin state @xmath49 via fractional stirap .
if @xmath108 can be accumulated to @xmath115 by controlling the stokes and drive fields , after the cycle evolution , the hh spin will return to the spin up state with the phase @xmath94 . as a result
, the rotation of hh spin about @xmath1-axis is realized , and the varying rotating angle can be obtained by changing the relative phase @xmath94 , namely the phase of the stokes field .
it should be noted that @xmath108 is gauge invariant , it depends upon the delay @xmath86 @xcite . in fig .
[ fig - gp - rz](b ) , we plot the evolution of @xmath108 in units of @xmath22 as a function of @xmath86 .
we take the rabi frequency amplitudes as : @xmath118 ps@xmath119 , the pulse width parameter @xmath120 ps . the magnetic field is taken as @xmath52 mt ( we take the in - plane @xmath36 factor of electron as @xmath121 @xcite , then the electron zeeman splitting @xmath122 ghz ) . with these parameters ,
the adiabatic condition is hold and the governing hamiltonian evolutes adiabatically .
as shown in fig .
[ fig - gp - rz](b ) @xmath108 increases from 0 by degrees with the growing of the delay , and reaches its maximum value , @xmath115 , when the delay @xmath86 is large .
thus , the relative phase @xmath94 can be used to control the hh spin rotation about @xmath1-axis . by combining the above rotations about @xmath0- and @xmath1-axis
, any rotation can be implemented .
for example , rotation about @xmath33-axis can be realized by @xmath123 .
our rotation procedure is sensitive to the non - abelian geometric phase and the relative phase between the stokes and the driving fields .
moreover , the non - abelian geometric phase is independent of pulse areas and dependent on the ratios @xmath124 , @xmath125 , and @xmath126 .
thus it is robust against the fluctuation of the pulse shapes , pulse areas and noise .
.[tab - fedility ] fidelity of selected rotations of hh spin .
[ cols="^,^,^ " , ] based on the non - abelian geometric phase , arbitrary rotations of hh spin are possible . but how about the degree to which our approximate description matches the actual behavior of the system ?
the fidelity is a measure of how accurate the target gate is implemented and it is defined as @xmath127 , where @xmath128 is the target operation , @xmath129 is the actual operation , and the average is taken over all input spin states @xcite . by numerical simulation of the density matrix equations
, the fidelity can be calculated .
typical rotation fidelities , listed in table [ tab - fedility ] , are on the order of @xmath130 .
the rotations of hh spin are realized with high fidelity .
the accurate measurement of the spin state of each qubit is essential in a quantum computation scheme . in our scheme , the @xmath20-polarized stokes ( the @xmath21-polarized pump ) field
can only excite the hh state with spin down ( up ) .
the measurement of the hh spin states therefore can be achieved by applying a @xmath20 or @xmath21 polarized continuous laser field .
when the @xmath20-polarized field is applied , for example , if the spin is rotated to @xmath50 , the qd will emit a single photon from the @xmath131 transitions , which can be detected using a single - photon counter @xcite .
to summarize , we consider a singly positive charged quantum dot , and demonstrate sequentially the initialization , the optical rotations of hh spin with non - abelian geometrical means , and readout of a single hole spin .
together with an magnetic field applied in voigt geometry , the quantum dot system can be reconfigured as a double tripod scheme . when the pump , the stokes and the driving fields are tuned to satisfy certain conditions , the qd system has two - fold degenerate dark states .
based on the non - abelian geometric phase produced by the nonadiabatic coupling between the two dark states , not only can the hh spin be rotated about @xmath0-axis with stimulated raman adiabatic passage , but also the relative phase between the stokes and the driving fields can be translated into the hole spin state with fractional stimulated raman adiabatic passages , leading to the implementation of rotation about @xmath1-axis . therefore the key step of optical arbitrary rotations of hh spin with high fidelity for qip can be implemented by non - alelian geometrical means .
it is in principle useful for spin - based quantum information processing .
this work is supported by the national research foundation and ministry of education , singapore under academic research grant no .
wbs : r-710 - 000 - 008 - 271 .
one of the authors ( h. s. ) would also like to acknowledge the support of the national basic research program of china ( 973 program ) under grant no .
2006cb921104 , the author ( s. q. g. ) acknowledges funding from the national natural science foundation of china through grant no .
10874194 and the author ( j. m. l. ) acknowledges funding from the national natural science foundation of china under grant | a non - abelian geometric method is proposed for rotating of heavy hole spins in a singly positive charged quantum dot in voigt geometry .
the key ingredient is the delay - dependent non - abelian geometric phase , which is produced by the nonadiabatic transition between the two degenerate dark states .
we demonstrate , by controlling the pump , the stokes and the driving fields , that the rotations about @xmath0- and @xmath1-axes with arbitrary angles can be realized with high fidelity .
fast initialization and heavy hole spin state readout are also possible . |
the majority of stars reside in binary and higher multiplicity systems @xcite , and are observed to exist both in the field as well as in dense stellar environments such as globular clusters and galactic nuclei . in dense clusters the dynamical evolution of binaries can be strongly affected by encounters with others stars . in the field , encounters with other stars are rare , and do not affect the binary evolution .
such binaries may still dynamically evolve if they reside in triple ( or higher multiplicity ) systems .
the inner binary in a triple can be affected by the perturbation of the third , outer companion , and evolve through long term secular processes .
in particular , triples in which the relative inclination between the orbit of the inner binary and the outer binary is high can evolve through kozai - lidov oscillations @xcite in which the eccentricity and inclination of the inner binary significantly change in a periodic / quasi - periodic manner . in this study
we explore the evolution of binaries in the environments of smbhs which are thought to reside at the center of most galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ) ; throughout this paper we focus on galactic nuclei similar to the galactic center ( gc ) of our own galaxy .
close to the center , where the gravitational potential is dominated by the mass of the smbh , binaries are bound to the smbh and effectively form a triple system , in which the binary orbit around the smbh is the outer orbit of the triple .
binaries near a smbh may therefore evolve both through gravitational scattering by other stars in the dense nuclear cluster , as well as through secular evolution in the smbh - stellar binary triple system @xcite .
various types of binaries exist in such environments , including both main - sequence ( ms ) binaries , compact object ( co ) such as white dwarf ( wds ) , neutron star ( ns ) or stellar black hole ( bh ) binaries or mixed co - ms binaries . in this paper
we focus on co - binaries , their evolution into short periods and the way in which they may become potentially observable gw sources for the advanced laser interferometer gw observatory ( ligo ) . scattering with other stars is also accounted for , but only in the sense of the lifetime of binaries in the gc , i.e. we do not follow the evolution of binaries which are disrupted ( evaporated ) through encounters with other stars at timescales shorter than the relevant secular timescales ( kozai evolution ) . in the following we begin with a discussion of binaries in the gc and their expected properties ( section [ sec : binaries ] )
, we then discuss the relevant timescales for the evolution of stars and binaries in galactic nuclei ( section [ sec : time - scales ] ) . in section [ sec : gw - sources ]
we discuss the secular evolution of binaries near a smbh , and their potential coalescence due to a combined kozai cycles - gw loss mechanism ; we then calculate the type and properties of the gw sources which form through this process . the secular evolution of other types of binaries is briefly discussed in section [ sec : kctf ] .
we summarize our results in section [ sec : summary ] .
at present , little is known about binaries in the gc .
massive binaries can be currently detected only by observations of stellar variability @xcite , which can mainly detect eclipsing or ellipsoidal variable binaries @xcite .
the fraction of such binaries among o - stars outside the gc is @xmath2 @xcite .
@xcite conducted a variability survey of the central @xmath3 of the gc and found one such binary ( also observed by * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) among 15 massive stars ( @xmath4 ) ; a fraction consistent with the galactic one .
other binary candidates were also found : a colliding wind binary @xcite and another eclipsing binary @xcite .
although these data are still insufficient for drawing strong conclusions about the total massive binary fraction in the gc , they do suggest that it is comparable to that observed outside the gc , among massive binaries . for reference ,
the massive binary fraction in the solar neighborhood is very high ( @xmath5 ; @xcite ) , and most massive binaries have semi - major axes of up to a few au . given our poor understanding of binary formation , it is unknown whether star formation in the unusual gc environment results in similar binary properties .
no observational data exist regarding low mass binaries in the gc .
little is known about compact binaries in the gc , though the observed overabundance of x - ray sources in the central pc @xcite suggests they are not rare but their properties and origin are not known .
we focus on binaries very close to the smbh , where secular kozai evolution affects the binary dynamics . however , this population is continuously destroyed through various processes ( evaporation , mergers , disruption ) .
we therefore first discuss the binary population outside the central region near the smbh ( @xmath6pc ) , which serves as a continuous source term repopulating the binaries in the central region .
we then discuss the processes which can repopulate binaries in the central region , both from populations outside and from in - situ formation of new binaries . in the following we discuss the binary population outside the central region near the smbh ( @xmath6pc ) .
this binary population serves as a source term for binaries in the central region on which we focus , which are continuously destroyed through various processes . the binary fraction and semi - major axis distribution of observed main sequence stars in the solar neighborhood have been analyzed in many binary surveys ( e.g. @xcite and references within ) . for compact objects the observational data is still lacking
however , many theoretical analysis have been done using stellar evolution and binary synthesis codes .
most of these studies focused on close binaries , usually in the context of cvs , x - ray binaries or gravitational waves sources ( which are relatively short lived and compose only a very small fraction of the binary population ) , and did not analyze the general distribution , which is needed for our analysis .
nevertheless , the general semi - major axis distribution of cos binaries have been found for isolated ( i.e. not taking into account dynamical effects which could be important in dense stellar environments ) wd - ms binaries @xcite and bh binaries @xcite . for nss , one can use the observed pulsar binaries population as found in the atnf pulsar catalog ( at http://www.atnf.csiro.au/research/pulsar/psrcat/ ; * ? ? ?
the initial period distribution we consider for the different binary populations is shown in fig .
[ f : bin - period ] .
these distributions do not take into account dynamical evolution of binaries in the dense environment near the smbh , mainly important in evaporating wide binaries ( so called soft binaries ; @xcite ) due to distant stellar encounters .
they also do not account for the long term evolution of close compact object binaries due to gw emission .
we account for these evolutionary aspects using simplified methods .
the primordial binary fraction ( not taking into account dynamical evolution such as evaporation of the binaries due to encounters ) , @xmath7 , is different for the different binary types , ms , wd , ns or bh binaries .
it depends both on the progenitors binary fraction , @xmath8 ( i.e. bhs and nss progenitors are high mass stellar binaries whereas wd binary progenitors have lower masses ) and fraction of surviving binaries , @xmath9 , after going through their evolutionary stages . the binary fraction of massive stars ( @xmath10 ) is more than two times that of lower mass stars ( @xmath11 ; @xcite ) .
stellar evolution further changes these fractions .
thus , for obtaining @xmath7 we either take the observed binary fractions in the solar neighborhood ( in the cases of ms binaries and ns binaries ) or use the theoretical models for the survivability of the different binaries @xcite to find @xmath9 , and multiply it by the progenitors binary fractions @xmath8 ( i.e. @xmath12 ) .
we finally get @xmath13 ( for high and low mass ms stars , wds , nss and bhs respectively ) as the evolution of lower mass stars ( and wds ) changes the semi - major axis distribution of the binaries but does not result in their disruption , whereas the evolution of bhs and nss results in most cases in the binary disruption .
we adopt these initial binary fractions throughout our calculations .
the fractions of co binaries used here already account for the initial high binary fraction and the stellar evolution of these binaries .
however , these binary fractions are also expected to be reduced due to dynamical encounters with other stars and through coalescence due to gw emission ( for compact objects ) and tidal friction . in the following calculations ( specifically in section [ sec : merger_frac ] )
we assume the initial co binary fractions ( @xmath7 ) in the gc are similar to those obtained here , where the effects due to dynamical encounters and kozai evolution of compact binaries are accounted for in a simplified manner , as we discussed later on .
evolution due to tidal frction which is important for ms stars will be discussed elsewhere , and is only shortly discussed here . [ cols="^ " , ] * other replenishment processes : * various other processes may affect the replenishment rate of compact binaries and/or their progenitors in the gc . in the following we qualitatively discuss such processes , but we do not account for them in our rates estimates ( beside star formation , effectively integrated in into the assumptions used in the observations based estimates ) , as their complete analysis is beyond the scope of this paper . _ _ triple disruption _ _ * * : * * @xcite ( 2009 ; see also ginsburg and perets 2011 ) suggested that disruption of triple stars could leave behind a binary in a close orbit around the smbh ; this could serve as a continuous source of replenished binaries close to the smbh .
this study estimated a triple disruption rate which is @xmath14 % of the binary disruption rate of massive stars ( ob stars ; we do not consider disruption of triple compact objects as these are likely to be rare ) . in half of these cases a binary
is captured around the smbh , where its companion is ejected . if the s - stars observed close to the smbh in the gc
arise from a binary disruptions and/or similarly the observed population of hvss arise from such processes , then these could be used to calibrate the binary and hence triple disruption rates .
capture rate calculations by @xcite(2007 ; see also * ? ? ?
* ; * ? ? ?
* ) , suggest that the binary disruption scenario ( induced by massive perturbers scattering ) could be consistent with the observed number of the young b - stars in the gc ( and similar processes are likely to be at work in other galaxies ; @xcite ) .
if this is the case then binaries are continuously replenished into close regions around the smbh ( and into eccentric orbits around the smbh ) , and our use of the observed number of b - stars could be justified , with some important differences .
only @xmath15 % of the observed b - stars would be the result of triple disruption and would actually be binaries . on the other hand the captured binaries are likely to have smaller separations than typical binaries in the field , which are more likely to form compact binary progenitors that would efficiently merge .
taken together , it is likely that such a replenishment mechanism would produce a gw rate which is 10 - 20 times lower then the high estimate for the ns stars .
this scenario , however , is not likely to produce progenitors massive enough to form sbhs , such as the observed young o / wr stars in the gc ( which reside in a disk like configuration and have low eccentricities ) .
gw rate from bhs coalescence are therefore not likely to gain much from this process .
_ in - situ star - formation : _ * * as mentioned above , our gc contains a large population of young massive o - stars , many of which reside in a stellar disk .
these stars were most likely to form in - situ following the fragmentation of a gaseous disk formed from an in - falling gaseous clump ( * ? ? ? * and references therein ) .
such stars formation burst may occur continuously throughout the evolution of the stellar cusp around the smbh . as mentioned earlier eclipsing and close binaries are observed among the gc young stars , suggesting star formation as an additional process which can repopulate the binary population of the gc .
continuous formation of binaries in the gc is part of the assumptions made in our high gw rate estimate .
the other assumptions in the high estimate are related to the binary population properties and their distribution around the smbh .
although some pioneering work have been done on the properties of binaries formed in the gc @xcite , these are generally unknown .
another important issue is whether the orbits of binaries formed in a disk could later dynamically evolve to achieve close approach to the smbh , where efficient secular evolution as studied here can occur .
although the o - stars observed in the disks are not observed closer than 0.05 pc to the smbh , various mechanisms were suggested for their inward migration closer to the smbh .
@xcite discussed the disk migration of binaries through similar migration scenario of planets in gaseous disks ; they show that the binaries do not only migrate but also shrink their inner orbit and possibly merge before they finish their migration .
madigan et al . @xcite proposed that binaries ( and single stars ) could be excited to high eccentricities through an eccentric disk instability scenario , and their periapsis approach could become small enough for the binaries to be disrupted .
both these mechanisms suggest the existence of binaries at close approach to the smbh . in the disk migration scenario
the binaries orbit around the smbh are likely to be relatively circular , and the binaries loose their orbital energy and migrate on a short timescale of the order @xmath16 yrs ; it is not clear , however , whether the gaseous disk enabling the migration would extend close to the smbh , and/or allow for migration in the innermost regions . in the eccentric disk instability binaries remain at large semi - major axis orbits around the smbh , but gradually increase their eccentricity , and therefore come closer to the smbh only at periapsis .
the timescale for the close approach is relatively short , @xmath17 , and binaries may be left at eccentric orbits at the end of this process . _ cluster infall : _ * * nuclear stellar clusters may result from the continuous infall of globular clusters that disperse close or in the close vicinity of the smbh ( see @xcite and references therein ) .
in particular such clusters may harbor an inner core cluster of bhs that formed during the cluster evolution .
if these bhs are retained in the cluster this mechanism may contribute to the bhs population in the gc .
further work needs to be done to explore the implications of this scenario to the replenishment of binaries in the gc . *
* _ resonant relaxation : _ in our discussion on replenishment processes above we considered the replenishment of binaries through non - resonant relaxation processes in which stars diffuse through random encounters with other stars .
resonant relaxation could change the eccentricities and inclination of stars much faster than regular relaxation .
one may therefore consider the timescale @xmath18 for diffusional replenishment to be the scalar resonant relaxation timescale @xmath19 , which could be much shorter than the non - resonant relaxation timescale assumed in the low estimate ( see fig .
[ fig : times ] ) ; this could raise the low estimates by 1 - 2 orders of magnitude .
for the high estimate , resonant relaxation may support the assumption that the distribution of stars rapidly achieve a steady state thermal distribution of eccentricities , as assumed in our models .
the role of the much faster vrr , which changes the inclination of an orbit around the smbh is less clear .
it depends on the dynamical effect on the inner binary inclination in respect to its orbit around the smbh .
we may consider two extreme possibilities .
if the binary orbit conserves its mutual inclination , i.e. acts as a stiff object , than vrr would not affect the secular evolution of the inner binary . however , if the binary orbit plays the role of a gyroscope , an initial co - planer configuration ( e.g. zero mutual inclination ) , in which kozai evolution is ineffective , could be transformed into a high mutual inclination configuration at which kozai evolution plays an important role . to further complicate the problem , the timescales for vrr could be short enough as to be comparable to the kozai timescale , in which case the coupling of both processes
should be considered .
all of the above mentioned issues for vrr of binaries require a dedicated study which is beyond the scope of this paper .
in the previous sections we discussed the secular evolution of co binaries near a smbh . a similar secular evolution could be important for stellar binaries and/or ms star - co binaries . in these cases
gw emission which served to dissipate energy and induce the inspiral of the co binaries is not likely to play an important role .
however , binary stars can dissipate energy through tidal friction .
the secular evolution of the triple system , smbh - stellar binary , can therefore lead to a similar outcomes to those studied in the context of stellar triples ( see also @xcite ) .
@xcite first suggested that kozai cycles and tidal friction in triples can induce their inspiral and the formation of compact binaries .
later studies by @xcite suggested that short period binaries typically form through this process ( which could be even more effective than previously though when accounting for octupole approximation ; @xcite ) .
@xcite suggested that such process can induce mergers of binary stars and the formation of blue stragglers .
such mergers could be therefore a possible source of young stars at the galactic center @xcite .
similarly other types of close binaries and their products could be formed in a similar way .
we may therefore expect that the perturbative effect of smbhs on binaries provide an efficient channel for the formation of close binaries and their products , such as x - ray binaries as well supernovae and gamma ray bursts ( following the merger of two wds or ns , or through the formation of a wd accreting binary ; see also @xcite ) .
we note , that the binary fractions of ms stars are generally higher than those of cos , as discussed in section 2 , and tidal fraction is more efficient in inducing binary shrinkage / coalescence than is gw emission .
taken together , we therefore expect the rate of shrinkage / merger of ms stars to higher than our estimtes of co binary mergers .
since this study focuses on secular evolution of co binaries , we postpone futher quantitative discussion of the evolution of binaries with ms star components to another study .
in this study we explored the secular evolution of compact binaries orbiting a massive black hole , and considered the effects of the stellar environment of the binaries .
we have shown that the smbh can drive the binary ( inner ) orbit into high eccentricities , at which point the bhs can efficiently coalesce through gw emission , at times shorter that the evaporation times of the binaries in the hostile stellar environment of the smbh .
such coalescing binaries can have non - negligible eccentricity when they enter the ligo band , and therefore have a unique observable signature .
in addition we found that binaries at very close orbits to the smbh can interact with it on dynamical timescale and present peculiar evolution leading to several gw pulses over a short timescale .
the latter sources are likely to be rare , and not be detected with current or planned instruments .
finally , we note that similar processes can be important for the formation and merger of close stellar binaries , in which cases the gw dissipation leading to a the binary inspiral is replaced with tidal friction .
we are extremely grateful to s. mikkola who wrote the archain integrator .
we thank s. prodan for a careful reading of an earlier version of this paper .
we thank k. cannon , c. hopman , b. kocsis , d. merritt , a. mrou , n. murray , harald pfeiffer and y. wu for useful comments .
fa acknowledges partial support from harvard / cfa predoctoral fellowship during the summer 2011 . | the environment near super massive black holes ( smbhs ) in galactic nuclei contains a large number of stars and compact objects .
a fraction of these are likely to be members of binaries . here
we discuss the binary population of stellar black holes and neutron stars near smbhs and focus on the secular evolution of such binaries , due to the perturbation by the smbh .
binaries with highly inclined orbits in respect to their orbit around the smbh are strongly affected by secular kozai processes , which periodically change their eccentricities and inclinations ( kozai - cycles ) . during periapsis approach , at the highest eccentricities during the kozai - cycles ,
gravitational wave emission becomes highly efficient .
some binaries in this environment can inspiral and coalesce at timescales much shorter than a hubble time and much shorter than similar binaries which do not reside near a smbh .
the close environment of smbhs could therefore serve as catalyst for the inspiral and coalescence of binaries , and strongly affect their orbital properties .
such compact binaries would be detectable as gravitational wave ( gw ) sources by the next generation of gw detectors ( e.g. advanced- ligo ) .
about @xmath0% of such nuclear merging binaries will enter the ligo observational window while on orbit that are still very eccentric ( @xmath1 ) . the efficient gravitational wave analysis for such systems would therefore require the use of eccentric templates .
we also find that binaries very close to the smbh could evolve through a complex dynamical ( non - secular ) evolution leading to emission of several gw pulses during only a few yrs ( though these are likely to be rare ) .
finally , we note that the formation of close stellar binaries , x - ray binaries and their merger products could be induced by similar secular processes , combined with tidal friction rather than gw emission as in the case of compact object binaries . |
interest in dynamical instabilities of rotating stars is strong because of the prospects for detecting gravitational waves from the oscillations of neutron stars .
modelling suggests that for gravitational wave emission to be at a detectable level we require that the oscillations be unstable , and hence capable of growth .
one key factor under investigation is the effect of differential rotation .
differential rotation may arise in neutron stars in several circumstances .
the first is at birth ; the latest studies of rotational core collapse indicate that neutron stars will be born with strong differential rotation @xcite .
subsequent accretion of supernova fallback material or material from a companion star @xcite , may also drive differential rotation , at least in the surface layers .
another possibility is that oscillations may drive the star into differential rotation , via non - linear effects @xcite .
a differentially rotating massive neutron star may also be generated following a binary neutron star merger @xcite .
differential rotation will be relevant provided that it can be maintained in the presence of viscosity or an internal magnetic field , which will act to bring the star into uniform rotation @xcite .
differential rotation has two key effects that are relevant to the study of dynamical instabilities .
firstly , differentially rotating stars exhibit a continuous spectrum , with dynamical behaviour that is distinct from the discrete normal modes of oscillation found in uniformly rotating stars .
secondly , differential rotation can lead to the occurrence of dynamical shear instabilities , unstable oscillations that do not exist in uniformly rotating systems @xcite . before proceeding we note some points on terminology .
we state that a mode is _ co - rotating if its pattern speed @xmath1 is positive , and _
counter - rotating if its pattern speed is negative ( where the modes are assumed to behave as @xmath2 ) .
if the pattern speed of the mode matches the local angular velocity at a point in the star we state that the mode has a _ corotation point .
the range of frequencies in which modes possess corotation points is termed the _ corotation band .
all modes with corotation points are co - rotating , but not all co - rotating modes possess corotation points .
_ _ _ _ we will also make reference to the _ degree of differential rotation .
this refers to the difference between the maximum and minimum angular velocities within a differentially rotating star . near uniform rotation
the difference is small , and the degree of differential rotation is low .
the difference increases as the degree of differential rotation rises . _
the majority of studies of the oscillations of differentially rotating systems have focused on the @xmath3 f - mode ( the so - called bar mode ) . as in a uniformly rotating star ,
the f - mode is split into co - rotating and counter - rotating branches ( fig .
[ twfigure1 ] ) . as the ratio of kinetic to potential energy , @xmath4 , increases
, the counter - rotating mode pattern speed becomes less negative , passing through zero at the point @xmath5 .
it is at this point that secular instability would arise in the presence of gravitational radiation reaction . in the absence of such mechanisms
, the counter - rotating branch pattern speed continues to increase until it merges with the co - rotating branch at @xmath6 .
this merger gives rise to the well - studied dynamical bar mode instability @xcite .
the value of @xmath7 is high ( @xmath8 ) even for high differential rotation . until recently ,
the high @xmath9 bar mode instability was the main focus of attention .
recent studies by @xcite have identified dynamical bar - mode instabilities in highly differentially rotating newtonian polytropes at values of @xmath9 as low as @xmath10 .
the mechanism behind these instabilities has yet to be identified .
the presence of corotation points may be critical . in differentially rotating stars
the co - rotating branch of the f - mode traverses the corotation band as @xmath9 is increased ( fig .
[ twfigure1 ] ) . in this letter
we demonstrate that the reported features of the low @xmath9 instabilities can be explained if the co - rotating f - mode develops a dynamical shear instability at the point where it enters the corotation band .
this mechanism has been studied in detail in a simple shell model @xcite .
dynamical instabilities associated with the presence of corotation points are also well known in accretion disc theory @xcite .
this would however be the first occasion on which they have been observed in stellar models .
on @xmath9 for the @xmath3 f - modes of a differentially rotating stellar model .
the quantity @xmath11 is discussed in the text . plot ( a ) depicts low or intermediate differential rotation , plot ( b ) high differential rotation .
the corotation band , traversed by the co - rotating f - mode , gets wider as the degree of differential rotation rises.,width=321,height=188 ] in section [ cs ] we explain the nature of oscillations within the corotation band , and review the results of a simple model in which the corotation instability mechanism operates . in section
[ main ] we show how the mechanism can account for the unusual features of the low @xmath9 bar - mode instabilities , using the supporting evidence provided by @xcite .
in two previous papers @xcite we examined the nature of oscillations within the corotation band for a differentially rotating incompressible spherical shell , by solving the linearized newtonian perturbation equations .
we found that within the band there was a continuous spectrum that is physically distinct from the more familiar discrete normal modes .
in addition , we found dynamical shear instabilities that set in when modes developed corotation points . in this section
we will show that the nature of solutions within the corotation band remains the same for a more realistic stellar model .
we assume that the star is an inviscid polytrope and that the perturbations are adiabatic and non - axisymmetric with @xmath12-dependence @xmath13 .
we also assume that the perturbations have time dependence @xmath14 , @xmath15 being the frequency .
we define the pressure @xmath16 , density @xmath17 , sound speed @xmath18 , gravitational potential @xmath19 , and angular velocity @xmath20 .
we work in the inertial frame and use cylindrical coordinates ( @xmath21 ) .
commas are used to denote partial derivatives . combining the euler equations , continuity equation and a polytropic equation of state for the perturbations , the eulerian perturbation @xmath22 obeys the following equation @xcite @xmath23 \nonumber { } \\
+ \phi \left [ \bar{\sigma}^2\left(\frac { l}{c^2 } + \frac{m^2 } { \varpi^2}\right ) - \frac{2m\omega\bar{\sigma}}{\varpi}\left(\frac{\rho_{,\varpi}}{\rho } + \frac{\omega_{,\varpi}}{\omega } - \frac{l_{,\varpi}}{l}\right ) \right ] \nonumber { } \\ - \frac{\bar{\sigma}^2 l}{c^2 } \delta \phi = 0 { } \nonumber \\\end{aligned}\ ] ] where the corotation parameter @xmath24 , the lindblad parameter @xmath25 , and the vorticity @xmath26 . note that for a barotropic star @xmath27 , so that the fluid rotates on cylinders @xmath28 .
solution of equation ( [ pert ] ) for real @xmath15 is problematic at the corotation radius @xmath29 where @xmath30 , and at the lindblad radii , @xmath31 where @xmath32 . in the neighbourhood of these points
the character of the solution can be investigated using frobenius analysis .
the solution is regular at the lindblad radii . at the corotation radius , although @xmath33 is continuous its first radial derivative is not . in general , there is both a finite step and a logarithmic discontinuity in the first derivative at @xmath34 ( see @xcite for more detail ) .
the additional degree of freedom introduced by the finite step allows us to construct a continuous spectrum of `` eigenfunctions '' for all frequencies within the corotation band .
the continuous spectrum eigenfunctions found in this more realistic stellar model have the same mathematical character as those found in the shell model . as in this simple model ,
if one solves the initial value problem one finds that the collective perturbation associated with the continuous spectrum is non - singular and hence physical ( @xcite , see also @xcite ) .
the dynamical shear instabilities found in the simple shell model occurred where modes crossed into the corotation band , at the point where modes merged with the continuous spectrum . given that the nature of the continuous spectrum remains the same in the more realistic model it is legitimate to ask whether the same instability mechanism might also operate . in section
[ main ] we will argue that the low @xmath9 instabilities exhibit all of the key features of this mechanism . in order to do this we must first review some pertinent results from the shell model @xcite . as the degree of differential rotation increases ,
many of the normal mode frequencies approach the edge of the expanding corotation band . some approach the boundary tangentially and do not cross into corotation .
those that cross into corotation at low degrees of differential rotation remain stable .
their eigenfunctions generally acquire a degree of singularity , but can be distinguished from the rest of the continuous spectrum by the fact that they do not have a finite step in the first derivative . this led us to term these solutions `` zero - step solutions '' .
we then carried out numerical time evolutions and an analytical evaluation of the initial value problem .
when we produced power spectra of the time evolution data , the zero - step frequencies stood out just as clearly as the discrete normal modes outside the corotation band .
the solutions that one finds when one solves the initial value problem are in fact non - singular , proving that these solutions have physical relevance . at high differential rotation zero - step solutions are still generated where modes cross into the corotation band .
in addition , dynamically unstable ( complex frequency ) modes can be generated at the crossing point for certain rotation profiles .
the eigenfunctions of these dynamically unstable modes are perfectly regular ( recall that our earlier analysis applied only to real frequencies ) .
their pattern speeds are however in the corotation band .
we refer to this instability mechanism as the corotation instability mechanism .
in reviewing the low @xmath9 instabilities we will focus on the j - constant rotation law used by shibata et al .
the j - constant rotation law is @xmath35 the parameter @xmath36 sets the degree of differential rotation , uniform rotation being reached in the limit as @xmath37 . as @xmath38 the degree of differential rotation increases . for a given equilibrium model ,
@xmath11 is a function of the parameter @xmath9 . following the convention used by shibata et al and working in dimensionless units where @xmath39 , the corotation band occupies the range of frequencies @xmath40 note that the corotation band widens as degree of differential rotation rises .
fig . 3 of @xcite , which we reproduce in schematic form in fig .
[ twfigure2 ] , illustrates the regions of parameter space occupied by the high @xmath9 and low @xmath9 instabilities .
the following features require explanation
. bar mode and low @xmath9 instabilities .
the lines marked a , b and c are discussed in the text.,width=264 ] for low degrees of differential rotation , the low @xmath9 instabilities are not observed .
dynamical instability does not set in until high @xmath9 , where the two branches of the f - mode merge ( track a , fig .
[ twfigure2 ] ) . for intermediate degrees of differential rotation ,
the co - rotating f - mode becomes unstable at low @xmath9 .
it stabilises again at intermediate @xmath9 before merging with the other branch of the f - mode and becoming dynamically unstable again at high @xmath9 ( track b , fig.[twfigure2 ] ) . for the highest degrees of differential rotation
there is no stable region after the onset of the low @xmath9 instability ( track c , fig .
[ twfigure2 ] ) . with this in mind
, we propose the following explanation of these three feature .
the fact that no instabilities are observed for stars with low degrees of differential rotation ( track a , fig .
[ twfigure2 ] ) suggests that , as for the shell model , this type of shear instability can only develop if the f - mode crosses into the corotation band when the degree of differential rotation exceeds a certain threshold .
for the shell model we were able to determine this threshold analytically . developing a similar criterion for the more realistic model will however be more difficult .
one may , in fact , have to resort to numerical calculations . below the threshold the f - mode passes into the corotation band but remains stable .
whether the eigenfunction is then purely regular , or possesses the singular character of the zero - step solutions that we found on the shell , is a matter for further study . at intermediate degrees of differential rotation ( track b , fig.[twfigure2 ] )
the co - rotating f - mode enters the corotation band and goes dynamically unstable .
the lower bound of the low @xmath9 instability region marks the point at which the co - rotating f - mode crosses the upper edge of the corotation band .
the upper edge of the low @xmath9 unstable region marks the point at which the co - rotating f - mode exits the corotation band .
the system is then stable until the co - rotating branch merges with the counter - rotating branch at high @xmath9 ( see fig .
[ twfigure1](a ) , which illustrates the passage of the co - rotating f - mode through the corotation band and its exit prior to mode merger ) .
as the degree of differential rotation rises ( track c , fig.[twfigure2 ] ) , the corotation band gets wider . for the scaling used in fig .
[ twfigure1 ] , for the j - constant law , the upper edge of the corotation band remains fixed whilst the lower edge approaches the zero - axis .
the counter - rotating branch of the f - mode will now enter the corotation band before the co - rotating branch can emerge ( see fig .
[ twfigure1](b ) ) .
there is thus no region of stability .
this raises two interesting questions .
does the counter - rotating f - mode also develop dynamical instability when it enters the corotation band ? and what becomes of the traditional bar mode instability at high @xmath9 ?
if the co - rotating f - mode never stabilises then the traditional mode merger route is excluded , but we can not rule out the possibility of additional instabilities arising via zero - step mergers @xcite . the corotation mechanism provides a simple qualitative explanation of the location the unstable regions of parameter space .
it also leads to quantitative and qualitative predictions that we can test against the results of the numerical simulations reported by @xcite .
firstly , all of the observed low @xmath9 instabilities should have real parts of frequency that lie within the corotation band
. further communication with the authors of @xcite has confirmed that all of the low @xmath9 unstable modes found for the j - constant law lie within the corotation band ( s.karino , private communication ) . to reproduce the shape of the instability region
the co - rotating f - mode must enter the corotation band at lower values @xmath9 and exit the band at higher values of @xmath9 as the degree of differential rotation rises .
using numerical modelling we have confirmed that this is the case .
the third prediction is that the boundaries of the low @xmath9 region of parameter space represent the points at which the f - mode enters or leaves the corotation band . for models on the lower edge of the region
, we should find that @xmath41 , while for models on the upper edge of the low @xmath9 region we should find @xmath42 , where @xmath43 is the real part of frequency .
this prediction , which would provide the most clearcut confirmation of the corotation mechanism , is however the hardest to test .
the problem lies in the growth times .
although the corotation instabilities are genuine dynamical instabilities ( in that they arise in systems for which no dissipative mechanism is present ) , they only exhibit dynamical ( short ) growth times deep inside the corotation band .
the shell model was sufficiently simple that we were able to track unstable modes with @xmath44 ( @xmath45 being the imaginary part of frequency ) .
it was the ability to detect these modes with very long growth times that allowed us to pinpoint the onset of instability as being near coincident with the edges of the band .
tracking slow growing instabilities using the types of numerical codes used to investigate realistic stellar models is more difficult .
the code used to generate the results of shibata et al , for example , was sensitive to modes with @xmath46 and larger ( s.karino , private communication ) . such a limitation on the shell study
would have made it impossible for us to detect the unstable modes until they were well inside the band .
conclusive tests of the third prediction will therefore require improved sensitivity to modes with long growth times .
nonetheless those low @xmath9 instabilities for which growth times have been published do exhibit the expected behavior in that the growth times are shortest in the center of the band , getting longer towards the edges ( see for example fig .
4 of @xcite ) .
although there are other instability mechanisms associated with the corotation band @xcite , all have limitations as an explanation for the low @xmath9 bar mode instabilities . in these scenarios
the co - rotating f - mode would enter the corotation band as either a zero - step or decaying mode ( although no decay is observed ) , merging inside the band with another zero - step or decaying mode to go unstable .
what it could be merging with is however unclear : the counter - rotating f - mode is outside the band at low @xmath9 , and the r - modes never enter @xcite . in addition
it is not clear whether the unstable mode could retain a bar - like character if it arose through merger with another mode type ; the corotation mechanism , by contrast , permits bar - like instability as the f - mode would not merge with any other mode .
the corotation mechanism may also be responsible for the low @xmath9 @xmath47 spiral instabilities observed by several authors , since the unstable modes of @xcite and @xcite are all within the corotation band .
this is a topic for future work .
the presence of shear instabilities in differentially rotating stars is conceptually very interesting .
it is also possible that these instabilities will turn out to be of considerable astrophysical significance .
these instabilities may limit the range of astrophysical rotation laws .
in addition the dynamically unstable modes may lead to interesting gravitational wave signals .
it is worth noting that , while it is difficult to envisage scenarios that lead to the large value of @xmath9 required for the more familiar bar - mode instability , one can easily imagine situations where the low @xmath9 shear instabilities will operate .
we have suggested a scenario , motivated by our analysis of the differentially rotating shell , that explains the low @xmath9 f - mode instabilities observed by shibata and collaborators . in this scenario
dynamical shear instabilities arise when the co - rotating f - mode enters the corotation band and the degree of differential rotation exceeds a certain threshold value .
the corotation mechanism provides a straightforward explanation for all of the unusual features of the instabilities observed in numerical simulations , and is supported by all available data .
further tests of the mechanism will require high precision studies on the onset of the instability , a task that will require codes sensitive to dynamically unstable modes with very long growth times .
alw is a nrc rra fellow at nasa gsfc .
na acknowledges support from the leverhulme trust in the form of a prize fellowship .
dij is supported by the center for gravitational wave physics under nsf cooperative agreement phy 01 - 14375 .
we thank s.karino for helpful comments .
99 balbinski e. , 1984 , , 209 , 145 balbinski e. , 1985 , , 149 , 487 balbinski e. , 1985 , , 216 , 897 case , k.m . , 1960 , physics of fluids , 3 , 143 centrella j.m .
, new k.c.b . , lowe l.l . ,
brown d.j . , 2001 , , 550 , l193 dimmelmeier h. , font j.a .
, mller e. , 2002 , , 388 , 917 fujimoto m.y . , 1993 , , 419 , 768 goldreich p. , goodman j. , narayan r. , 1987 , , 221 , 339 karino s. , eriguchi y. , 2003 , , 592 , 1119 karino s. , yoshida s. , eriguchi y. , 2001 , , 64 , 024003 levin y. , ushomirsky g. , 2001 , , 322 , 515 luyten p.j . , 1990 , , 242 , 447 new k.c.b . , shapiro s.l . , 2001 , , 548 , 439 ott c.d . , burrows a. , livne e. , walder r. , 2004 , , 600 , 834 papaloizou j.c.b .
, pringle j.e . , 1985 , , 213 , 799 pickett b.k . , durisen r.h .
, davis g.a . , 1996 , , 458 , 714 rezzolla l. , lamb f.l . , shapiro s.l . , 2000 , , 531 , l139 shapiro s.l
. , 2000 , apj , 544 , 397 shibata m. , uryu k. , 2000 , , 61 , 064001 shibata m. , baumgarte t.w . , shapiro s.l .
, 2000 , , 542 , 453 shibata m. , karino s. , eriguchi y. , 2002 , , 334 , l27 shibata m. , karino s. , eriguchi y. , 2003 , , 343 , 619 watts a.l . , andersson n. , beyer h. , schutz b.f . , 2003 , , 342 , 1156 watts a.l . , andersson n. , williams r.l . , 2004 , , 350 , 927 yoshida s. , rezzolla l. , karino s. , eriguchi y. , 2002 , , 568 , l41 | recent numerical simulations indicate the presence of dynamical instabilities of the f - mode in differentially rotating stars even at very low values of @xmath0 , the ratio of kinetic to potential energy . in this letter
we argue that these may be shear instabilities which occur when the degree of differential rotation exceeds a critical value _ and _ the f - mode develops a corotation point associated with the presence of a continuous spectrum .
our explanation , which is supported by detailed studies of a simple shell model , offers a straightforward way of understanding all of the key features of these instabilities . |
in [ 9 ] , b.e .
johnson , r. v. kadison , and j. ringrose introduced a notion of amenability for von neuman algebras which modified johnson s original definition for banach algebras[8 ] in the sense that it takes the dual space structure of a von neumann algebra into account .
this notion of amenability was later called connes - amenability by a. ya .
helemskii [ 7 ] . in [ 10 ]
, the author extended the notion of connes - amenability to the larger class of dual banach algebras .
the concept of approximate amenability for a banach algebras was introduced by f. ghahramani and r. j. loy in [ 5 ] .
our motivation for introducing approximate connes - amenability is finding a version of these approximate forms of amenablity which sounds suitable for dual banach algebras . before proceeding
further we recall some terminology . throughout @xmath1
is a banach algebra and @xmath2 is a banach a - bimodule .
a derivation is a bounded linear map @xmath3 such that @xmath4 for @xmath5 , define @xmath6 then @xmath7 is a derivation ; maps of this form are called inner derivation . a derivation @xmath3 is approximately inner if there exists a net @xmath8 such that for every @xmath9 , @xmath10 , the limit being in norm .
we say that @xmath1 is approximately amenable if for any @xmath1-bimodule @xmath2 , every derivation @xmath11 is approximately inner .
@xmath1 is called a dual banach algebra if there is a closed submodule @xmath12 of @xmath13 such that @xmath14 .
for example if @xmath15 is a locally compact group , then @xmath16 is a dual banach algebra(with @xmath17 ) , or if @xmath1 is an arens regular banach algebra then @xmath18 is a dual banach algebra .
let @xmath1 be a dual banach algebra .
a dual banach @xmath1-bimodule @xmath2 is called normal if , for each @xmath5 , the maps @xmath19 are @xmath20-continuous .
a dual banach algebra @xmath1 is connes - amenable if , for every normal , dual banach @xmath1-module @xmath2 , every @xmath20-continuous derivation @xmath3 is inner .
let @xmath1 be a dual banach algebra , and let @xmath2 be a banach a - bimodule .
then we call an element @xmath21 a @xmath20-element if the maps @xmath22 are @xmath23 continuous .
* definition 2.1 . * a dual banach algebra a is approximately connes - amenable if for every normal , dual banach a - bimodule x ,
every @xmath20-continuous derivation @xmath24 is approximately inner . proof .
let x be the banach a - bimodule whose underlying linear space is a equipped with the following module operation : @xmath26 obviously x , is a normal dual banach a - bimodule and the identity map on a is a @xmath27 continuous derivation . since a is approximately connes - amenable
, then there exists a net @xmath8 such that @xmath28 this means that a has a right approximate identity .
similarly , one see that a has a left approximate identity.@xmath29 proof .
let @xmath35 be a @xmath20-continuous derivation for normal dual banach @xmath31-bimodule @xmath36 . by [ 5 ,
lemma 2.3 ] , @xmath37 where @xmath38 is a @xmath20-continuous derivation and @xmath39 . since @xmath40 is a normal dual banach a - bimodule , then @xmath41 and @xmath42 is approximately inner ; whence d is approximately inner .
thus @xmath31 is approximately connes - amenable .
if we define @xmath44 , then @xmath36 turns into a normal dual banach @xmath31-bimodule and @xmath45 is a @xmath20-continuous derivation .
so @xmath45 is approximately inner , and hence so is d. it follows that a is approximately connes - amenable.@xmath29 proof .
suppose x is a normal dual banach a - bimodule , and @xmath46 is the restriction map to @xmath47 .
then @xmath48 is a @xmath23 continuous homomorphism . therefore x is a normal dual banach @xmath18-bimodule with the following actions @xmath49
let @xmath3 be a @xmath23 continuous derivation .
it is easy to see that @xmath50 is a @xmath23 continuous derivation .
since @xmath18 is approximately connes - amenable , than there exists a net @xmath51 such that @xmath52 so @xmath53 let x be a neo - unital banach a - bimodule , and let @xmath56 . by [ 11 , theorem 4.4.8 ] @xmath36 is a normal dual banach @xmath18-bimodule and d has a unique extension @xmath57 .
from the approximate connes - amenability of @xmath18 we conclude that @xmath45 , and hence d is inner .
it follows that a is approximately amenable.@xmath29 * theorem 2.8 .
* let a be an arens regular banach algebra which is a right ideal in @xmath18 .
let for every @xmath18-bimodule x , @xmath86 .
if @xmath18 is approximately connes - amenable , then a is approximately amenable .
@xmath89 is a @xmath23 continuous derivation . since @xmath90 is an @xmath18-bimodule , then by assumption @xmath91 we claim that @xmath88 is a normal dual banach @xmath18-bimodule .
let @xmath92 be a net in @xmath18 such that @xmath93 .
then @xmath94 since a is a right ideal of @xmath18 and @xmath20-topology of @xmath18 restricted to a coincides with the weak topology , we have @xmath95 let @xmath96 .
then by ( 1 ) there exists @xmath9 and @xmath97 such that @xmath98 .
thus we have , @xmath99 therefore @xmath100 on the other hand by definition , @xmath101 now since @xmath18 is approximately connes - amenable , then there exists a net @xmath102 , such that @xmath103 let @xmath104 be the natural projection , then @xmath105 thus d is approximately inner .
it follow that a is approximately amenable.@xmath29 * proposition 2.9 . *
suppose that a is a dual banach algebra with identity .
then a is approximately connes - amenable if and only if every @xmath27continuous derivation into every unital normal dual banach bimodule x is approximately inner .
@xmath107 suppose @xmath24 is a @xmath108 continuous derivation into the normal dual banach bimodule e. by [ 5 , lemma 2.3 ] , we have @xmath37 where @xmath109 is a derivation and @xmath39 .
since d is a @xmath27continuous derivation and x is a normal dual banach bimodule then @xmath110 is @xmath27continuous and @xmath111 is normal .
so by assumption @xmath110 is approximately inner , and therefore a is approximately connes - amenable.@xmath29 let a be a dual banach algebra with identity , and let @xmath112 be the space of separately @xmath23-continuous bilinear functionals on a. clearly , @xmath112 is a banach a - submodule of @xmath113 and @xmath114 from [ 10 ] we have a natural a - bimodule map @xmath115 defined by letting @xmath116 . since @xmath117 and @xmath118 separates points of @xmath119 , then @xmath120 is one - to - one
. we will identify @xmath119 with its image , writing @xmath121 the map @xmath122 is defined as follow : @xmath123 since multiplication in a dual banach algebra is separately @xmath23-continuous , we have , @xmath124 so that restriction , of @xmath125 to @xmath112 turns into a banach a - bimodule homomorphism @xmath126 suppose @xmath127 and @xmath128 .
we have the notation , @xmath129 more generally given a dual banach space @xmath36 and a bounded bilinear function @xmath130 such that @xmath131 and @xmath132 are @xmath23-continuous , we define @xmath133 by @xmath134 from now on we assume that @xmath1 is a dual banach algebra with identity .
it is well known that every dual banach algebra with a normal virtual diagonal is connes - amenable [ 15 ] . in the following theorem
we extend this result to approximate connes - amenability . proof .
suppose @xmath2 is a normal dual banach @xmath1-bimodule with predual @xmath139 and @xmath24 @xmath23-continuous .
since a is unital , by proposition 3.11 we can assume that x is unital .
we follow the argument in [ 4 , theorem 3.1 ] .
the bilinear map @xmath140 is @xmath23-continuous .
thus by the above remark we may let , @xmath141 for @xmath142 , @xmath143 , we have @xmath144 @xmath145 @xmath146 therefore @xmath147 and similarly @xmath148 . we have the following relations , @xmath149 and @xmath150 where @xmath151 and @xmath152 by ( 1 ) and ( 2 ) , we have @xmath153 so @xmath154 on the other hand since @xmath155 , then @xmath156 thus @xmath157 @xmath158 so we have @xmath159 @xmath160 @xmath161 @xmath162 by assumption and ( 3),(4 ) we have @xmath163 @xmath164 thus d is approximately inner , and hence a is approximately connes - amenable.@xmath29 * definition 2.12 .
* a is called approximately strongly connes - amenable if for each unital banach a -bimodule x , every @xmath23 continuous derivation @xmath11 whose range consists of @xmath27 elements is approximately inner .
@xmath166 we follow the argument in [ 10 , theorem 4.7 ] .
since @xmath167 is @xmath23 continuous then @xmath168 is @xmath27 closed and @xmath169 clearly @xmath170 attains its values in the @xmath27 elements of @xmath168 . by the definition of approximate strong connes - amenability , there exists a net @xmath171 such that @xmath172 let @xmath173 .
it follows that @xmath174 one of the unsatisfactory sides of dealing with approximate connes - amenability for dual banach algebras is the apparent lack of a suitable intrinsic characterization in terms of approximate normal , virtual diagonals .
we saw that dual banach algebras with an approximate normal , virtual diagonal are approximately connes - amenable , but the converse is likely to be false in general .
however for a compact group g the converse is true : we now modify the definition of approximate normal , virtual diagonal and obtain the desired characterization .
let a be a dual banach algebra with predual @xmath12 and let @xmath175 be the multiplication map .
from [ 12 , corollary 4.6 ] , we conclude that @xmath176 maps @xmath12 into @xmath177[14 ] .
consequently , @xmath178 induces the homomorphism @xmath179 with these preparations made , we can now characterize approximately connes - amenable , dual banach algebras through the existence of certain approximate normal , virtual diagonals .
this is indeed an approximate version [ 12 , theorem 4.8 ] .
proof @xmath59 : first , note that @xmath182 is canonically mapped into @xmath183 . define a derivation @xmath184 since the dual module @xmath185 is normal , then it follows that d is @xmath27continuous .
clearly , d attains its values in the @xmath27closed submodule @xmath186 .
since it is a normal dual banach a - module , then there is a net @xmath187 such that @xmath188 letting @xmath173 , we see that it is the approximate @xmath0virtual diagonal for a. @xmath166 let x be a normal , dual banach a - bimodule . by proposition 2.9 we may assume that x is unital , and let @xmath3 be a @xmath27derivation .
define @xmath189 by [ 12 , lemma 4.9 ] , @xmath190 maps the predual @xmath139 into @xmath177 .
hence @xmath191 let @xmath180 be an approximate @xmath0virtual diagonal for a , and let @xmath192 .
we have the restriction map @xmath193 therefore we can choose a net @xmath194 such that @xmath195 for every @xmath142 , @xmath196 @xmath197
@xmath198 therefore @xmath199 so by assumption and ( 1 ) , we have @xmath200 this implies that a is approximately connes - amenable.@xmath29 | we introduce the notions of approximate connes - amenability and approximate strong connes - amenability .
then we characterize these two types of dual banach algebras in terms of approximate normal virtual diagonal and approximate @xmath0 virtual diagonals .
some concrete cases are also discussed . |
in this paper we present fuse observations of the supersoft x - ray binary rx j0513.9@xmath06951 ( hereafter called x0513@xmath069 ) .
this @xmath4 source was identified with a @xmath516th mag peculiar emission - line star in the large magellanic cloud ( pakull et al .
1993 , cowley et al . 1993 ) . with an absolute magnitude of m@xmath6 , x0513@xmath069 is the brightest of the supersoft x - ray binaries ( e.g. cowley et al .
it is known to show high and low optical states differing by @xmath51 mag ( e.g. alcock et al .
1996 , cowley et al .
the optical spectrum is characterized by strong , broad emission lines of he ii and weaker lines of o vi , n v , c iv , and c iii .
weak emissions flanking the strongest lines at @xmath74000 km s@xmath3 have been interpreted as arising in bi - polar jets ( e.g. crampton et al .
1996 ) .
previous spectroscopic work has shown that the emission line peaks have very small velocity amplitude ( e.g. crampton et al .
1996 , southwell et al .
if these are interpreted as motion of the compact star , the source must be seen at a low orbital inclination and the unseen secondary star must have a low mass . the small amplitude of the optical light curve ( alcock et al .
1996 ) also suggests the system is viewed from a low inclination angle .
however , using a newly refined orbital period and ephemeris , cowley et al .
( 2002 ) have shown that the phasing of the velocities derived from optical spectra differs between the high and low optical states , so that the previously inferred mass function and its interpretation are in doubt . in order to access other highly ionized lines ( particularly the o vi resonance doublet ) and investigate the far - ultraviolet continuum ,
observations were obtained with the far ultraviolet spectroscopic explorer ( fuse ) .
the fuse data presented in this paper add new information about x0513@xmath069 .
the fuse observations were taken within a 24 hour period on 2001 october 30 .
the windows during which the system can be observed , between earth occultations and saa passes , allowed 16 exposures ranging from 367 to 3310 seconds , spread fairly evenly over @xmath51.2 binary cycles .
the data were processed using calfuse version 2.0.5 .
the source is bright enough so that there were no difficulties with background subtraction or extraction windows .
analysis of the fuse fes images , taken during the acquisition and observation sequence , shows that x0513@xmath069 was in its high optical state at 16.7@xmath8 mag during the far ultraviolet observations . in general ,
the lif1a channel values are the most reliable for wavelength and flux , as this is the guide channel .
lif1b is usually well - tracked too , as it suffers little thermal cycling .
however , as the observations were continuous , the stability of all the telescope alignments is likely to be good , and the data rates for all channels showed no signs of losing the target star .
figure 1 shows the mean fuse spectrum of x0513@xmath069 , with various features marked .
we note the lack of n iii or c iii lines , although they are present in optical spectra .
the principal region of interest is the o vi doublet ( @xmath91032,1038 ) , seen as a broad emission blend , but contaminated by some sharp airglow lines .
no broad emissions were seen at other less ionized lines , such as he ii , c iii , or n iii .
there are also weak sharp emissions at the positions of the o vi lines , unfortunately blended with airglow and interstellar absorption features .
however , the relative strengths of these and other unblended airglow lines indicate that the sharp o vi peaks are present , superimposed on the broad emission .
we note that the weak o vi lines in the visible region ( crampton et al 1996 ) are seen only as sharp peaks , with no broad component detected . in the fuse spectra
he ii is only detected as a sharp emission peak , weak and blended with airglow emissions . by contrast , both a strong narrow component and broad emission wings are seen in the he ii line at 4686 . in hst
spectra , gnsicke et al .
( 1998 ) found a similarly broad structure in the n v resonance doublet ( 1239,1243 ) , also with narrow peaks ( see their figs . 1 & 4 ) .
the continuum is present throughout the entire fuse range , and there are narrow lyman absorptions , seen most clearly in lines l@xmath10 through l@xmath11 .
table 1 gives the observational details and some of the measurements made . in all cases , the new ephemeris by cowley et al .
( 2002 ) is used .
the eight different fuse spectral channels were measured separately .
those with the best signal level are reported here .
measurements of the fuse spectra were made by fitting single profiles to the o vi emission blend , and also by cross - correlation against a smoothed mean spectrum template , which had been edited to remove airglow emission and narrow interstellar absorptions .
airglow emissions were identified using the atlas in feldman et al .
( 2001 ) and also by comparing spectra with varying fractions of nighttime in the exposure .
the lyman absorption lines were measured by simple profile fitting .
the continuum variations were measured by averaging the total fluxes outside of the emission features and with airglow emissions removed .
figure 2 shows plots of the measured quantities , and figure 3 shows the o vi profile in detail and a montage of changes with orbital phase .
we discuss the broad o vi in detail in the following section .
since there was a large range in individual exposure times among the spectra listed in table 1 , we generated a set of 9 spectra combined to equalize their exposures and phase coverage , and hence enhance the s / n for measurement of the weaker features .
these were also used as a check for our major results . in all cases
these phase - averaged spectra gave similar results , so we do not quote these measures in detail .
we note , however , that there are differences in some measures between the overlapping phases from the two orbital cycles covered .
further observations would be required to fully separate the phase - dependent and irregular changes .
all measurements were run through a sine - wave fitting program to determine relative binary phases of their variations , using the cowley et al.(2002 ) ephemeris .
the results and their formal errors are shown in table 2 .
inspection of figure 2 shows that some variations are not very sinusoidal , especially the velocity curves of the broad o vi emission and lyman absorption , but the formal errors of the fits are a useful measure .
using more complex functions , such as elliptical orbit fits , does not alter the amplitudes and phasing significantly .
note that the phase errors are all quite small , with the exception of he ii emission velocities .
the o vi emission is extraordinarily broad and can not be separated into the two o vi lines at 1032 and 1038 . moreover , the whole emission feature is redshifted by some 800 km s@xmath3 or more ( @xmath53 ) from the mean wavelength expected for the blend at the velocity of the lmc .
the profile is complicated by the presence of strong c ii 1036 and other weaker interstellar absorptions , and by airglow emissions , including the strong ly@xmath12 line at 1026 .
figure 3 shows the mean o vi profile and its variation with phase .
it appears that the emission actually extends to @xmath13 km s@xmath3 , but there is a p cygni absorption that erodes the short wavelength side .
the amount of o vi absorption is evident from the difference between the profile and its reflection about zero velocity which is shown in figure 4 .
zero velocity is taken as the mean wavelength of the two o vi lines , shifted by the systemic velocity ( @xmath5 + 280 km s@xmath3 ) .
this assumes equal flux from each line ( i.e. high optical depth ) .
this means that the velocity width for each individual line is @xmath53000 km s@xmath3 .
the high velocity optical jets seen in h and he ii lines ( crampton et al . , southwell et al . ) do not appear to have a counterpart in the far ultraviolet o vi lines .
in the optical region , the jet lines stand well above the broad wings in most observations , while figure 4 shows a smooth profile with no features at @xmath14 km s@xmath3 .
however , we note that the optical jet lines are variable from epoch to epoch .
cowley et al.(1998 ) found the jet lines very weak or absent in spectra taken in 1996 during a bright optical state , whereas they had been detected easily in 1993 and 1994 spectra ( see their fig . 3 ) .
the broad o vi emission profiles do extend out to these high velocites , as do the he ii 4686wings , but the 4686 wings could not be measured accurately for radial velocity changes because of blending with the nearby n iii / c iii blend .
however , crampton et al . did report extended shortward absorption in the broad wings of h@xmath12 , indicating a similar p cygni wind profile as seen in o vi ( see figure 4 ) .
there are at least three possible causes of the measured broad emission radial velocity variations .
these include : ( a ) changing strength of the p cygni absorption , altering the mean position of the unabsorbed emission , ( b ) variation of the relative strengths of the two o vi doublet lines , corresponding to changes in the optical depth of the emission region , ( c ) velocity variations which reveal orbital motion of the emission line region . both possibilities ( a ) and ( b ) above would involve changes in the overall profile of the emission feature .
crampton et al .
report that in their 1994 data `` the h@xmath12 wind absorption appears to be strongest at phase 0.75 '' ( which is phase 0.72 using the revised ephemeris of cowley et al .
2002 ) , but from limited data they also note that the `` changes are of low significance and may not be phase related '' .
this phasing corresponds roughly to the maximum velocity of the broad o vi emission , suggesting the velocity variation might be related to changing p cygni absorption strength .
however , in the fuse broad o vi profiles we do not see such a change , but rather only changes in the overall emission flux , ( plus associated changes in the interstellar absorption depths ) .
profile changes that give rise to the measured full velocity amplitude of @xmath5240 km s@xmath3 would show up as a profile difference , significantly different on the opposite sides of the central wavelength .
this is not seen at any phase , and particularly not between the phases of maximum velocity differences , where we see the expected ` sine - wave ' difference spectrum from a moving but unchanging profile .
we thus consider that the broad profile arises in a region that appears similar from all viewing angles .
the phasing of the measured velocity curve is close to that expected for the compact star s orbital motion , assuming the optical light minimum occurs near its superior conjunction ( cowley et al .
thus , we consider the implications of possibility ( c ) . the small velocity amplitude of the peak emission velocities were discussed in this context by crampton et al . and southwell et al .
their interpretation implies either a low orbital inclination and a low - mass donor star ( @xmath50.3m@xmath15 ) .
however , as shown by cowley et al .
( 2002 ) , the new ephemeris makes this model unlikely , since maximum velocity does not occur at the expected phase .
therefore , we explore the possibility that the @xmath5120 km s@xmath3 semiamplitude of the o vi broad emission is an upper limit to the compact star s motion . if there are small profile changes that contribute to the velocity amplitude , they will reduce the contribution from the orbital motion .
thus , this measured velocity gives a maximum value to the mass function for the system .
we discuss the implications in more detail below .
the very broad o vi emission most likely arises in the innermost parts of the accretion disk , as this is where we expect the most highly ionized material to be found , if radiatively excited .
we argue below that the mass of the compact star probably lies in the range 0.7 to 1.1m@xmath15 , while the orbital inclination must be @xmath16 , given the line widths .
the orbital velocity at the surface of such a white dwarf is in the range 3100 to 5500 km s@xmath3 , which would be projected to 1900 to 3400 km s@xmath3 for an equatorial disk that reaches the stellar surface .
the emission profile from such a disk at this inclination angle would not be strongly double - peaked and the blend of two o vi line profiles ( separated by some 1700 km s@xmath3 ) together with interstellar absorptions , would further obscure any profile structures .
an alternative site for high velocities in the o vi profile is the bi - polar jets .
as noted , the jets are seen in much lower ionization lines as two separate and relatively narrow peaks , implying a high degree of collimation .
the o vi could conceivably arise in the inner accelerating part of the jets .
however , since the jets are highly collimated ( the jet emissions indicate an opening angle less than 10@xmath17 ) , this would give rise to very double - peaked structure at any point in the jet , quite unlike the profiles seen in the fuse data .
given the probable origin of the broad emission in the inner disk , and its unchanging profile with phase , we note that the p cygni absorption must indicate a considerable ( and azimuthally symmetrical ) disk wind that is seen around the viewing cone and which must be separate from the jets .
absorption within a jet would only be seen in a narrow viewing cone of about half the 10@xmath17 jet opening angle .
the continuum is detected throughout the fuse spectral range .
continuum variations were measured by averaging the total fluxes in regions outside of the emission features where the airglow emissions had been removed .
the measured flux changes are shown in figure 2 and in tables 1 and 2 .
table 1 shows values of the interpolated continuum at the o vi doublet , while the values in table 2 and figure 2 are continuum levels fitted uniformly to the wavelength regions indicated .
the agreement between the independent fuse channels is very good .
note the rising flux level to shorter wavelengths , indicating very high temperatures . expressed in magnitudes , the orbital variation is @xmath50.5 0.6 mag , but examination of overlapping phases shows this amplitude is partially due to cycle - to - cycle variations .
the fuv variation is much larger than that seen in the optical , where the mean range is only @xmath50.04 mag , but the cycle - to - cycle variations are up to @xmath8 mag ( see cowley et al .
2002 ) .
the fuv continuum and its changes must come from a region with temperatures of 30000 k or more , to be so much larger in the fuv .
we also note that the o vi broad emission flux varies with similar amplitude and phasing , so it is likely to be associated with the continuum changes .
since disk light dominates the spectrum , we are seeing azimuthally varying disk brightness .
one possibility is a thickening of the outer disk downstream of the gas stream from the donor star . illuminated by the hot inner disk
, this would be brightest when viewed from the opposite side .
we discuss below a model where the phasing of this is consistent with the radial velocity changes .
cowley et al .
( 2002 ) discuss the origin of the optical light variations , showing the phasing of the minima has persisted for at least 8 years .
thus , it is likely to be caused by a geometric occultation of a bright region fixed within the system , rather than disk structure whose position and shape may change .
this region may be the inner point in the roche lobe , where heating by the wd and x - rays is greatest .
thus , they suggest that phase 0 is at or close to superior conjunction of the compact star .
we adopt this geometry in our binary discussion below .
it was not possible to make clean measures of the o vi peak emissions because of blending with the airglow and interstellar lines .
the 1038line velocity shows a correlation with the fraction of the exposure that was in daylight ( i.e. strength of airglow ) .
if we fit a straight line through this , we find a velocity of + 295 km s@xmath3 for an airglow - free spectrum , which is close to the systemic velocity .
if we examine the deviations from this line as function of binary phase , there is a small amplitude variation ( k@xmath510 km s@xmath3 ) , with maximum close to phase 0 , as was found by cowley et al .
( 2002 ) for he ii peaks during the bright state .
since our examination of the fes image indicates x0513@xmath069 was in a high optical state during the fuse observations , it appears that the narrow o vi 1038 component probably behaves like the he ii 4686peaks .
the 1032 line is more strongly blended by the interstellar absorption on its shortward side , and the measures of this line give a scatter about apparent velocity + 380 km s@xmath3 .
both line peaks show apparent flux changes , which are also attributable to blending with airglow lines and the small velocity variations noted above for the 1038 line .
the narrow he ii emission at 1084 is weak and has airglow lines nearby .
however , we do measure a low velocity amplitude ( similar to the he ii 4686 optical line of crampton et al . ) and a phasing that is different from the broad o vi , but close to that of the o vi 1038 peak mentioned above .
table 2 includes our fits to the optical emission peak velocities from crampton et al . and southwell et al .
, for the optically bright and the transition states , using the new ephemeris .
we note that the fuse he ii peak velocities , while poorly determined , are quite consistent with the high - state optical velocity phasing and amplitude .
this is also included in our binary orbit discussion below .
the he ii peak flux shows a small change which is maximum at phase 0.62 ( see table 2 ) , if the changes are phase - related .
lyman absorptions are clearly present at all phases , and they are seen well down the series .
the stronger lyman lines have severe airglow contamination , but the airglow has a strong decrement , and the lines measured are free of any significant airglow contamination , as determined by extrapolation of the decrement .
there is a weak correlation of radial velocity with fraction of night - time in the exposure ( hence potential strength of airglow ) , but less than half the amplitude with orbital phase , and there is no measurable change in the equivalent width with phase or with night - time fraction .
the absorption mean velocity is + 226 km s@xmath3 , suggesting an outflow from x0513@xmath069 whose systemic velocity is @xmath18 km s@xmath3 ( and not consistent with contamination by zero velocity airglow ) .
we note that the lyman absorption velocity changes are closely phased with the increase in broad o vi flux .
if a broad lyman emission within the system increases during the same phases , then it could conceivably produce a line asymmetry , causing the lyman absorption to be measured more positively .
however , the lyman absorptions are only 1 wide , compared with 18 for the broad emission .
we discuss below the puzzle of where the neutral absorber may be that gives rise to these absorptions , but draw attention ( see table 2 ) to the fact that the velocity changes are close to antiphased with the broad o vi emission velocity .
here we examine the implications of assuming that the broad o vi emission velocity curve reflects the orbital motion of the compact star in x0513@xmath069 . adopting the amplitude and
phasing given in table 2 , we have adopted a weighted mean of the profile - fitting and cross - correlation measurements , using both the individual spectra and also the phase - binned spectra of equal signal .
the sketch in figure 5 shows how the measured quantities vary with binary phase . following cowley et al .
( 2002 ) , we assume phase 0 occurs at the superior conjunction of the compact star , and the phasing of the fuv measures are based on this new ephemeris . the broad o vi
emission velocity phases are consistent with them showing the orbital motion of the wd , within their uncertainties ( table 2 ) . the high state light curve of cowley et al .
has a wide minimum , centered at phase 0 as derived from a sine curve fit to the individual data points .
in the fainter , transition " state the minimum may be narrower , but also centered on phase 0 . we stress that there is an uncertainty in what is causing the optical minimum , so that assuming it occurs at conjunction may be incorrect . there
might instead be an off - centered brighter region that is partially hidden at phase 0 , but it is unlikely to be a structure in the disk , since it was stable for over 8 years .
the lack of significant profile changes suggests that the major source of broad o vi is the inner parts of the accretion disk , rather than in a stream or the inner part of the roche lobe where narrow components might be formed .
by contrast , we note that the o vi emission seen in am her , a highly magnetic cv without a disk , has a complex and varying profile , and the lines do appear to arise in other parts of the system where shock heating may ionize the o vi ( hutchings et al .
2002 ) . for x0513@xmath069 ,
the adopted semiamplitude of 117 km s@xmath3 for the compact star gives a mass function that implies the values shown in figure 6 .
the lack of any eclipse in the light curve at any wavelength indicates that the orbital inclination is less than @xmath19 .
the considerable width of the broad emission argues against a very low orbital inclination angle . on the other hand ,
the observed high velocity of the bi - polar jets , which presumably are oriented nearly perpendicular to the disk , is expected to be close to the escape velocity for the white dwarf ( e.g. southwell et al .
1996 ) . at a higher orbital inclination
, their observed velocities would be expected to be much lower ( as seen in the supersoft binary rx j0019.8 + 2156 , cowley et al .
1998 ) .
these considerations are shown in figure 6 .
the dashed lines show the locus where the projected jet velocity equals the escape velocity for the white dwarf , and also the locus of the projected orbital velocity at its equator . the jet velocity is assumed not to exceed escape velocity , and the line width not to exceed orbital velocity at the white dwarf surface .
the diagram shows a region where both conditions apply .
we consider white dwarf masses between 0.7 and 1.1m@xmath15 , and inclinations above @xmath530@xmath17 to be feasible .
we discuss below a mass ratio of @xmath52 and a donor star mass of @xmath51.6 m@xmath15 , with an orbital inclination of 35@xmath17 .
figure 5 shows a sketch of the roche lobe and center of mass for this mass ratio .
the wd orbital velocity , with an inclination angle @xmath16 , sets the donor star roche radius ( polar ) at @xmath51.7 r@xmath15 .
the main sequence radius of a star of mass 1.6 m@xmath15 is @xmath51.2 r@xmath15 .
this is also sketched in figure 5 , and it lies well within the roche lobe . on the other hand , a giant ( luminosity class iii ) of this mass will have a radius of @xmath54.5 r@xmath15 , which would engulf the companion .
thus , we suggest that the donor star is just at the end of its main sequence lifetime and expanding . this would imply a large mass transfer rate , and bright disk , as observed .
x0513@xmath069 is several times brighter visually than other supersoft x - ray sources ( sss ) in the lmc ( e.g. cal 83 and cal 87 ) , although its bolometric luminosity is similar to that of cal 83 ( l@xmath20 ergs s@xmath3 ) . both cal 83 and x0513@xmath069 also show bi - polar jets which are characteristic of a luminous disk
the absolute magnitude of a 1.5 m@xmath15 main sequence star is about + 3 ( + 1.3 for a giant ) , while the whole x0513@xmath069 system has absolute magnitude of @xmath21 .
thus , the lack of visibility of the proposed donor star is not surprising , as it is only a few percent of the disk luminosity . indeed ,
if the optical light curve is due to heating of the donor star , as in the case of hz her , the observed amplitude is about that expected for any non - eclipsing inclination . in this model , disk thickening by gas - stream impact
is seen from the far ( illuminated ) side of the disk at the phase shown for maximum fuv light .
the he ii and o vi peak velocity probably arises in or near the intersection of the mass stream and disk , as seen in other x - ray binary systems ( e.g. cyg x-1 ) .
this material lies near the system center of mass and has low velocity vector which is the sum of the stream and orbital velocity at this point . in the low optical state , when the disk is fainter ( smaller ? ) , the relative change of orbital and streaming velocity vectors leads to the observed change in phase of maximum peak velocity , as sketched .
the optical minimum may be caused by the partial occultation of the inner point of the roche lobe where the donor star will be brightest .
the he ii 1085peak flux is maximum at phase 0.62 , which is where the heated donor star is most clearly visible . however , the feature is weak and near variable airglow lines .
the lyman sharp absorption velocities are antiphased with the broad o vi , and thus they suggest association with the donor star
. however , the disk and not the donor star must be the source of the fuv continuum radiation behind the absorbing material .
also , as mentioned earlier , the absorption lines have a mean velocity of + 226 km s@xmath3 compared with the systemic velocity of + 280 km s@xmath3 obtained from the optical he ii emission lines .
thus , we may be looking through a @xmath550 km s@xmath3 outward flow of cool material that arises from the donor star and is projected against the hot continuum of the disk .
the absorbing column does not vary measurably with binary phase , but the lines are weak and the continuum does vary .
however the absorptions arise , they may still carry the orbital motion of the secondary .
it is hard to imagine another way they could be antiphased with the broad o vi emission .
it is interesting that this yields a mass ratio of the stars is 54/117=0.46 , which lies right in the region we have already noted in figure 6 .
however , the lyman absorption velocity curve is far from sinusoidal , so perhaps the origin of these lines lies elsewhere .
the masses suggested here are unusual for sss in having a more massive donor star .
as noted , it is possible that the real mass function is lower , if some of the measured broad velocity changes are due to profile changes .
however , we have suggested that such a contribution would be small , so we regard the mass discussion as an upper limit of interest .
the disk in this system is extremely luminous , and it is widely assumed that the disk brightness and jet formation occur because the mass transfer rate is high .
thus , this is consistent with a more massive donor star . as pointed out by cowley et al .
( 2002 ) , the binary model implied by the much lower velocity amplitude of the optical line peaks ( crampton et al . , southwell et al . )
is no longer consistent with the revised ephemeris from the optical light curve .
other problems with this original model were the very low inclination ( nearly pole - on ) and the low donor - star mass , if one assumed the compact star was a white dwarf .
it was originally suggested by van den heuvel et al .
( 1992 ) that the donor stars in supersoft binary systems would be expected to be about twice the mass of the white dwarf , based on evolutionary considerations . however , until the present data on the fuv o vi velocities , all previous observational evidence has indicated the donors are low mass stars ( see , for example , review by gnsicke et al .
as this system appears to differ from other sss in having a more massive donor star , it may be that the mass transfer can be driven by more than one mechanism . in this case ,
normal roche lobe overflow by evolution off the main sequence would apply , while for the less massive donor stars , an additional wind mechanism is required . 1 & 3.431 & 0.87 & 1571 & 1201 & 7.6 & 200 & 3.6 2 & 3.502 & 0.97 & 1239 & 1227 & 8.4 & 280 & 4.1 3 & 3.570 & 0.06 & 973 & 916 & 8.3 & 283 & 4.2 4 & 3.611 & 0.11 & 367 & 930 & 10.5 & 290 & 6.4 5 & 3.643 & 0.16 & 918 & 963 & 9.3 & 350 & 4.3 6 & 3.684 & 0.21 & 801 & 965 & 7.0 & 313 & 5.2 7 & 3.714 & 0.25 & 650 & 740 & 8.8 & 291 & 4.9 8 & 3.755 & 0.30 & 1230 & 934 & 6.5 & 240 & 4.4 9 & 3.827 & 0.40 & 1610 & 1147 & 6.2 & 214 & 4.0 10 & 3.906 & 0.50 & 3240 & 1153 & 4.9 & 185 & 3.1 11 & 3.976 & 0.59 & 3310 & 1217 & 4.8 & 197 & 3.0 12 & 4.045 & 0.68 & 3298 & 1156 & 5.3 & 196 & 3.0 13 & 4.114 & 0.77 & 3281 & 1026 & 5.1 & 186 & 3.1 14 & 4.184 & 0.86 & 3281 & 1190 & 4.9 & 184 & 3.2 15 & 4.257 & 0.96 & 2571 & 1099 & 5.6 & 229 & 3.3 16 & 4.328 &
0.05 & 1832 & 1159 & 6.4 & 166 & 3.4 o vi mean em & 4.00 @xmath220.05 & 126 @xmath2255 & 152 & 0.62 @xmath220.07 & 1141 @xmath2241 o vi ccor em & 4.07 @xmath220.04 & 110 @xmath2236 & 94 & 0.71 @xmath220.06 & 1094 @xmath2226 binned spec , mean & [email protected] & 121@xmath2230 & 64 & [email protected] & 1118@xmath2222adopt & & 117 & & 0.70 h lyman abs & 3.64 @xmath220.02 & 54 @xmath2210 & 23 & 0.15 @xmath220.02 & 226 @xmath227 he ii em & 3.56 @xmath220.19 & 11 @xmath22 11 & 22 & 0.05 @xmath220.25 & 268 @xmath227 * line flux & & ( @xmath23 ) & ( @xmath23 ) & & ( @xmath23 ) o vi & 3.63 @xmath220.02 & 2.0 @xmath220.4 & 1.2 & 0.14 @xmath220.03 & 6.4 @xmath220.3 he ii & 4.00 @xmath220.06 & 0.27 @xmath220.15 & 0.30 & 0.62 @xmath220.08 & 0.46 @xmath220.01 * o vi eq.w & & ( ) & ( ) & & ( ) mean & 3.51 @xmath220.05 & 2.3 @xmath220.9 & 3.0 & 0.15 @xmath220.06 & 17 @xmath221 * continuum & ( 2452210 + ) & ( @xmath24 ) & ( @xmath24 ) & & ( @xmath24 ) 920 - 965 & 3.71 @xmath220.04 & 2.0 @xmath220.4 & 1.1 & 0.24 @xmath220.03 & 7.0 @xmath220.3 1050 - 1100 & 3.69 @xmath220.03 & 1.5 @xmath220.3 & 0.9 & 0.22 @xmath220.03 & 5.6 @xmath220.2 1100 - 1180 & 3.68 @xmath220.02 & 1.3 @xmath220.3 & 0.7 & 0.20 @xmath220.03 & 4.2 @xmath220.2 * opt he ii peaks & 2449660+high opt state & [email protected] & [email protected] & 7.5 & [email protected] & 279@xmath222intermediate opt state & [email protected] & [email protected] & 8.9 & [email protected] & 299@xmath222 * * * * | we have obtained spectroscopy with the far ultraviolet spectroscopic explorer ( fuse ) of the supersoft x - ray binary rx j0513.9@xmath06951 over a complete binary orbital cycle .
the spectra show a hot continuum with extremely broad o vi emission and weak lyman absorptions .
he ii emission is weak and narrow , while n iii and c iii emissions are undetected , although lines from these ions are prominent at optical wavelengths .
the broad o vi emission and lyman absorption show radial velocity curves that are approximately antiphased and have semiamplitudes of @xmath1 and @xmath2 km s@xmath3 , respectively .
narrow emissions from he ii and o vi show small velocity variations with phasing different from the broad o vi , but consistent with the optical line peaks .
we also measure considerable changes in the fuv continuum and o vi emission line flux .
we discuss the possible causes of the measured variations and a tentative binary interpretation . |
it is by now an established fact that the universe accelerates faster than previously anticipated @xcite . hence some hitherto neglected ingredient ( in general called dark energy )
has to be responsible for this new phenomenon . to account for this phenomenon
we can introduce new physics in terms of a scalar field @xcite or modify three expressions in einstein s equations which are often equivalent to a specific model with a scalar field .
the three possibilities to account for the new physics are : the einstein tensor , the energy - momentum tensor ( this is to say , the energy momentum tensor of a fluid gets modified by the inclusion of other components , @xcite or the equation of state @xcite ) .
the first possibility encompasses a positive cosmological constant @xmath0 and higher order gravity with a more complicated einstein - hilbert action @xcite . in the the present work we choose to work with the cosmological constant as the simplest explanation for the acceleration of the universe .
we shall put forward the question if such a cosmological model has an influence on astrophysical structures .
we shall use equilibria concepts like hydrostatic equilibrium and virial equations to see how relatively low density astrophysical matter of different shapes behaves in a fast expanding universe . anticipating the results
, we can say that indeed there are some interesting effects .
often it is assumed that @xmath0 does not have any effect on astrophysical processes which take place at scales different from the cosmological ones .
indeed , looking at the scales set by @xmath0 , this assumption seems to be justified at the first glance .
the scales set by the cosmological constant are of truly cosmological order of magnitude @xcite ) .
the density scale is set by @xmath1 with @xmath2 . the length scale , @xmath3 is of the order of the hubble radius while the mass scale @xmath4 reaches up to the value of the mass of the universe .
these scales constitute the so - called coincidence problem , namely the question as to why we should live exactly at an epoch where the scales of the cosmological constant are also the scales of the universe .
neither was it so in the past nor will it be so in the future when the universe expands further .
the only astrophysical structures which match these scales are superclusters whose densities are indeed of the order of magnitude of @xmath5
. indeed , here we can almost be sure that the cosmological constant is of relevance @xcite .
however , probing into astrophysical consequences of the cosmological constant of other , smaller and denser structures like clusters of galaxies or even galaxies themselves , would look a hopeless undertaking unless we find circumstances where the effect of @xmath0 ( which in the very principle is present ) gets enhanced .
this can indeed happen through various mechanisms .
for instance , in a problem where @xmath6 combines with a much smaller length scale , say @xmath7 , the effects can be sometimes expressed as @xmath8 . in consequence , the observable in which this expression enters gets affected by @xmath0 in a way which is important at much smaller scales than @xmath6 .
a concrete example is the schwarzschild - de sitter metric where we find the parameter @xmath6 together with the much smaller length scale of the schwarzschild radius @xmath9 .
these two conspire in the form @xmath10 to define the largest extension of bound orbits as explained in the text and in @xcite .
we will discuss a very similar combination which emerges from the virial theorem defining the largest possible virialized structure with a given mass .
another possibility to enhance the effect of the cosmological constant is to consider non - spherical objects .
it then often happens that the effect of @xmath0 becomes @xmath11 where @xmath12 are two different length scales of a flattened object like a disk or an ellipsoid @xcite .
this indeed happens for many astrophysical quantities , among other the critical mass , the angular velocity and the temperature ( mean velocity of the components of the large structure ) which we will discuss in the present paper .
finally , we can vary a dimensional variable to see if this enhances the effect of the cosmological constant . as an example
let us quote the polytropic index @xmath13 in the equation of state entering also the lane - emden equation .
it is known that with growing @xmath13 ( @xmath14 ) the object described by this equation of state does not have a well - defined radius as the density goes only asymptotically to zero .
we will show that this pattern of behaviour becomes more dominant with @xmath15 . of course
, all these effects become stronger the more diluted the mass conglomeration is .
the superclusters are certainly the best candidates if we look for astrophysical effects of the cosmological constant . as a matter of fact
, they do not seem to be virialized due to the extreme low density and their pancake structure @xcite where the effect of flatness mentioned above becomes powerful @xcite .
for the next structure , the clusters of galaxies ( or groups of galaxies ) with densities between one and three orders of magnitude above the critical density @xcite , we would need one of the enhancing factors discussed above to see an appreciable effect of @xmath0 .
this is possible in various ways as shown below .
galaxy clusters can have various forms , among others oblate and prolate @xcite . and
what is more , they can even rotate @xcite .
we will show an explicit effect of @xmath0 on their angular velocity and temperature in case the angular velocity is zero .
since the effect of large eccentricity is larger for prolate than for oblate ellipsoids , it is comforting to know that clusters can assume a prolate shape .
for low - density galaxies like the low surface brightness ( lsb ) galaxies whose density is roughly four orders of magnitude above the critical density @xcite , we still find some effects . for @xmath16 and @xmath17 the solution of the lane - emden equation @xcite
is very often used as a phenomenologically valid description of the density profile ( called also plummer s law @xcite .
this is still possible as @xmath18 as @xmath19 .
however , this property vanishes for low - density galaxies and the @xmath16 case not only does not have a well - defined radius , its solution does not vanish asymptotically which thus rendering it unphysical .
the paper is organized as follows . in the second section
we will briefly review the general form of virial theorem including pressure , magnetic fields and , of course , the cosmological constant . here
we will also discuss some general results regarding @xmath0 . in the third section
we will specialize on spherical configurations .
we will show how @xmath0 sets the scale of a maximal virial radius and compare it to a result from the schwarzschild - de sitter metric .
we will also solve the lane - emden equation numerically and analytically ( for the polytropic index @xmath20 ) . in the fourth section
we will discuss non - spherical configurations .
first , we will show how @xmath0 affects the angular velocity of spheroids .
in addition , we will discuss the effects of @xmath0 for the critical mass , mean velocity and mean rotational velocity too .
the fifth section is devoted to small oscillations around equilibrium .
the cosmological constant enters the equations of newtonian limit as a consequence of its appearance in the einstein field equations .
it is through this weak field limit approximation that @xmath0 enters also in the equations describing the structure of astrophysical configurations .
it is interesting to note that all variables to be found in the virial equations , are also present in the poisson equation of the newtonian limit .
however , this is not always the reason why these terms enter the virial equations , at least in the first order .
the poisson equation for a self - gravitating system modeled as an ideal fluid is written as @xmath21 where @xmath22 is the pressure and @xmath23 is the electromagnetic energy density .
the solution of ( [ poisson ] ) at the zeroth order of @xmath24 ( from now on we set @xmath25 ) is written as @xmath26 where the dots stand for the correction terms that appear because the boundary conditions are now set at a finite distance @xcite .
these terms can be usually neglected .
the cosmological constant @xmath27 contributes to the expansion of the universe .
this fact remains partly valid in the newtonian limit where @xmath0 gives us an _
external force_. this defines the so - called ( non - relativistic ) newton - hooke spacetime @xcite .
+ the second order tensor virial equation can be derived in different ways : from a statistical point of view through the collisionless boltzmann equation , from a variational principle or by direct differentiation of the moment of inertia tensor @xmath28 in the following we use the statistical approach @xcite which also allows to derive higher order virial equations ( for instance , the first order virial equation refers to the motion of the center of mass ) . in this context , from boltzmann s equation one can derive the equation for momentum conservation ( euler s equation ) written for a self gravitating system influenced by a magnetic field as @xmath29 where @xmath30 is the gravitational potential given by ( [ psolution ] ) ( which includes @xmath0 ) and @xmath31 is the pressure tensor , @xmath22 is the pressure and @xmath32 its traceless part .
equations ( [ poisson ] ) and ( [ euler ] ) together with an equation of state @xmath33 ( @xmath34 is the entropy ) complete the description of a self gravitating fluid . by taking exterior products of @xmath35 with euler s equation and integrating over the volume of the system one obtains the second order virial equation as @xmath36 where @xmath37 is the kinetic energy tensor and @xmath38 is a generalized potential energy tensor which contains the contribution from the gravitational potential energy tensor @xmath39 and the contributions of magnetic field through @xmath40 together with @xmath41{\text{d}}s_{j},\hspace{0.5cm}{\mathcal{b}}_{ik}=\frac{1}{2}\int_{v}{b_{i}}{b_{k}}\,{{\text{d}}^{3}r},\hspace{0.5cm}{\mathcal{b}}={\rm tr}({{\mathcal{b}}_{ik}}).\ ] ] a very useful version of the virial equation can be derived by assuming an isotropic pressure tensor and taking the trace in ( [ virotro ] ) . this way we get the scalar @xmath0-virial equation @xmath42 where the total kinetic energy is written as @xmath43
the equilibrium condition is reached for @xmath44 .
this gives us the general @xmath0-virial theorem @xmath45 for rotating configurations with constant angular velocity , the kinetic term is modified as in the standard way as @xmath46 with @xmath47 the rotational kinetic energy tensor and @xmath37 is referred to motions observed from the rotating reference frame .
the @xmath0-virial theorem has been used in different contexts in @xcite . in the present work we will extend these studies .
+ the tensor virial equation is widely used in many astrophysical applications .
the inclusion of @xmath0 provides a new way to study effects of the cosmological constant ( parameters , in general ) on astrophysical objects .
the outcome depends essentially on two factors : the geometry of the configuration and the density profile .
we will explore the spherical geometry for both constant and varying density profiles and study some effects for non spherical geometry with constant density .
the consequences that can be derived from the @xmath0-virial theorem can be classified in two categories .
the first one puts an upper bound on the cosmological constant or alternatively a lower bound on density of objects in gravitational equilibrium .
provided these bounds are satisfied , we can also study in the second step the effects of @xmath0 on other properties of the astrophysical configurations like rotation , small oscillations etc .
+ the first simple consequence of the virial equation emerges if we require the system to satisfy ( [ vt ] ) .
the fact that @xmath48 implies an upper bound on the vacuum energy density @xmath49 all systems in equilibrium have to satisfy ( [ ineq0 ] ) .
note that the right hand side of this expression is a function of both the density and the geometry of the system .
hence , we must expect different bounds for different geometries and density profiles .
for instance , if we assume a constant density and @xmath50 , we can define @xmath51 and @xmath52 through @xmath53 ) becomes @xmath54 the factor @xmath55 which is only a function of the geometry ( if we neglect the contribution of magnetic fields ) , will appear in many places in the paper .
its relevance lies in the fact that it enhances the effect of the cosmological constant for geometries far from spherical symmetry when @xmath56 is large .
a useful generalization can be done for situations in which we use the tensor form of the virial equation , namely @xmath57 finally , a curious equation can be derived by eliminating @xmath0 from the tensor virial equations : @xmath58 although @xmath0 does not enter this equation , ( [ curious ] ) is only valid if the denominator is non - zero as is the case with @xmath15 . in sect .
4 we will use this equality to infer a relation between the geometry and rotational velocity of an ellipsoid .
having discussed the general form of the virial theorem , we will discuss now the effects of @xmath0 and set @xmath59 .
the tensor virial equation is trivially satisfied for spherically symmetric configuration without a magnetic field , since @xmath60 and @xmath61 .
therefore , in this section we use only the scalar form of ( [ virotro ] ) .
explicit expressions can be derived in the spherical case with constant density , with @xmath62 and @xmath63 , so that @xmath64 . in this case
the ratio @xmath65 does not get enhanced much by the geometrical factor @xmath55 .
worth mentioning is the result from general relativity .
there the upper bound for the cosmological constant comes out as @xmath66 @xcite where @xmath67 is the mean density defined by @xmath68 .
this bound is derived not only from newtonian astrophysics , but also from a general relativistic context via the tolmann - oppenheimer - volkoff equation @xcite for hydrostatic equilibrium of compact objects @xcite .
another relevant effect of the cosmological constant is the existence of a maximal virial radius of a spherical configuration which can be calculated from the @xmath0-virial equation . using the expressions for @xmath69 and @xmath70 given before , equation ( [ scalar1 ] ) yields as a cubic equation for the virial radius @xmath71 @xmath72 here we introduced the dimensionless temperature @xmath73 as @xmath74 where @xmath75 is the mass of the average member of the configuration , @xmath76 is the boltzmann constant , @xmath77 is the temperature and @xmath78 is defined by @xmath79 the length scale @xmath80 is set by the the cosmological constant as @xmath81 @xmath82 is dimensionless hubble parameter @xcite and @xmath83 is the density parameter at the present time .
the positive real root of equation ( [ cubic ] ) is given by @xmath84 where @xmath85 , is radius for the configuration at @xmath86 is given by @xmath87 and the dimensionless parameter @xmath88 is defined as @xmath89 the radius @xmath85 is the largest radius that a spherical homogeneous cloud may have in virial equilibrium ( i.e , satisfying ( [ vt ] ) ) .
the function @xmath90 can be obtained from the solution of the cubic equation and reads ( clarified in the appendix a ) @xmath91.\ ] ] figure 1 shows the behavior of @xmath90 for different values of @xmath88 .
we see that the increase of the temperature implies a decrease of the effects of @xmath0 which can be easily checked if we solve @xmath71 from the virial theorem with @xmath17 and compare it to the approximation @xmath92 in ( [ rvir ] ) : @xmath93 we can consider ( [ rvir ] ) as a radius - temperature relation for a fixed mass applied on astrophysical structures in a single state of equilibrium in the presence of @xmath0 .
that is , given @xmath88 and @xmath73 we calculate the radius .
but we can adopt another point of view for this relation .
imagine a spherical configuration characterized by a constant mass @xmath94 .
in analogy to a thermodynamical reversible process , the configuration may pass from one state of virial equilibrium to another following the curve @xmath95 , that is , satisfying the condition @xmath44 .
clearly , there must be some final temperature @xmath96 when this process ends since the temperature can not increase indefinitely .
but of course since the virial equations are not dynamical we can not know which stage is the final one .
if we _ assume _ that the effects of @xmath0 are negligible when @xmath97 , then using eqs ( [ delta ] ) and ( [ dapp ] ) we get @xmath98 this is an equation for the temperature @xmath99 @xmath100 for a hydrogen cloud ( @xmath101 ) , we then write the mass - temperature relation using eqs.([rl ] ) and ( [ x ] ) as @xmath102 note that this expression maintains the same dependence of the standard mass - temperature relation derived from the virial theorem , i.e , @xmath103 ( see @xcite or equation ( [ mvel2 ] ) of this paper ) .
however the meaning of ( [ temperature3 ] ) is different from that of typical mass temperature relations since ( [ temperature3 ] ) is associated to the temperature that a system acquires in the final stage after after going through some _
reversible processes _ which took the system through successive states of virial equilibrium with constant mass from a radius @xmath85 to a radius @xmath104 or vice versa . on the other hand , the mass - temperature relation like eq .
( [ mvel2 ] ) of this paper relates the temperature of any configuration in equilibrium with its observed mass at constant density . in that context
one considers only one equilibrium state and the cosmological constant enters just as corrections . as a final remark on eq .
( [ rvir ] ) , we discuss a result which formally coincides with the virial radius @xmath85 derived from the schwarszchild de - sitter spacetime @xcite , but whose physical meaning is quite different .
the schwarzschild - de sitter metric takes the form @xmath105 now in contrast to ( [ rs ] ) we have @xmath106 with @xmath75 the mass of the object giving rise to the schwarzschild - de sitter metric ( in eq .
( [ rs ] ) @xmath94 is the mass of the total conglomeration whereas here we consider @xmath75 as the mass of its average member ) . choosing the affine parameter as the proper time
@xmath107 the equation of motion of a test - body can be cast in a form similar to the corresponding equations from non - relativistic classical mechanics .
@xmath108 where @xmath109 is a conserved quantity and @xmath110 is the effective potential , defined by @xmath111 for simplicity we are have chosen here the angular momentum @xmath112 to be zero . with @xmath112 zero or not
, @xmath110 displays a local maximum below zero due to @xmath15 forming a potential barrier .
this is to say , the standard local minimum where we find all the bounded orbits is now followed by a local maximum after which @xmath110 goes to @xmath113 . with @xmath17 this
function approaches zero asymptotically .
one is immediately tempted to say that this barrier will occur at cosmological distances .
this is not the case and one calculates @xmath114 in other words , the combination of the large scale @xmath6 with the small scale @xmath115 gives us a distance of astrophysical relevance , namely @xmath116 .
its relevance lies in te fact that beyond @xmath116 there are no bound orbits . indeed , with @xmath75 the solar mass
, @xmath116 is of the order of a globular cluster extension ( @xmath117 pc ) ; with @xmath75 as the mass of globular cluster , @xmath116 comes out to be of the order of the size of a galaxy ( @xmath118 kpc ) , and finally taking @xmath75 to be the mass of a galaxy , @xmath116 gives the right length scale of a galaxy cluster ( @xmath119 mpc ) .
certainly , the value of the extension of a large astrophysical body is the result of a multi - body interaction .
but with the actual values of @xmath116 , it appears as if the _ length scale _ ( we emphasize that we are concerned hare about scales and not precise numbers ) of an astrophysical conglomeration is approximately @xmath116 , which apparently means that this scale does not change drastrically when going from a two body problem to a multi - body calculation .
this makes sense if the object under consideration is not too dense .
we can now say that whereas @xmath94 in @xmath85 ( via @xmath78 ) is the mass of the object , @xmath75 in eq .
( [ sca2 ] ) is the mass of its members .
clearly , we have @xmath120 , but both scales are of astrophysical order of magnitude . a result related to ( [ rvir2 ] ) derived in the framework of general relativity
can be found in @xcite .
the examination of configurations with non - constant densities can be done in two directions . knowing the density profile @xmath121
, we can set up the virial equation and evaluate the equilibrium conditions from the inequality ( [ ineq0 ] ) . in this picture , the effects of @xmath0 are included in the solution for the potential @xmath30 as in eq .
( [ psolution ] ) and the resulting term acts like an external force , as mentioned before .
+ the second option is to combine the eqs .
( [ poisson ] ) , ( [ euler ] ) and an equation of state ( e.o.s ) for which we can take a polytropic form @xmath122 . this way we obtain the lane - emden equation with @xmath0 @xcite @xmath123 where @xmath124 is the central density , @xmath125 , @xmath126 with @xmath127 , @xmath128 and @xmath129 is the associated jeans length defined as @xmath130
it is important to notice that in this picture the expected effects of @xmath0 are to be found in the behavior of the density profile since now eq .
( [ le ] ) implies that its solution is also function of the parameter @xmath131 .
some effects of @xmath0 are contained in the total mass and the radius of the configuration which is reached when @xmath132 .
hence ( [ le ] ) implies @xmath133^{1/2 } , \hspace{0.5cm}r = a\xi_{1}={r_{\lambda}}\left[\kappa(n+1)\rho_{\rm c}^{\frac{1}{n } } \left|\frac{{\text{d}}}{{\text{d}}\xi}{\left(}\xi^{2}\frac{{\text{d}}\psi}{{\text{d}}\xi } { \right)}_{\xi_{1 } } \right|\right]^{1/2}.\ ] ] note that the radius of the configuration is now proportional to @xmath80 .
this is due to the fact that @xmath0 sets a scale for length .
however , this does not mean that @xmath134 will be always of the order @xmath6 as @xmath0 is also contained in the expression in the square brackets in eq .
( [ radius ] ) .
we will show this below in a concrete example . since @xmath0 is a new constant scale the lane - emden equation loses some of its scaling properties as explained in @xcite .
the mass of the configuration can be determined as usual with , @xmath135,\ ] ] where we used eq .
( [ le ] ) for the second equality .
the total mass is then obtained by evaluating the last expression at @xmath136 .
as expected , the mass increases because the newtonian gravity has to be stronger in order for the configuration to be in equilibrium with @xmath137 .
figure 2 shows the numerical solutions for @xmath138 to @xmath16 .
we expect that the radius of the configuration is increased by the contribution of @xmath131 and find it confirmed in the figures .
however , not always is the radius of the configuration well defined , even if @xmath139 . for sizeable values of @xmath131 ( black line )
we can not find physical solutions of eq .
( [ le ] ) as the function @xmath140 acquires a positive slope .
one might be tempted to claim that the radius of the configuration could be defined in these situations as the position where @xmath140 has its first minimum , but as can be seen for @xmath141 such a radius would be smaller than the radius with @xmath142 which contradicts the behaviour shown for the other solutions where @xmath143 . as already mentioned above this
is the correct hierarchy between the radii because large @xmath144 gives rise to a large external force pulling at the matter .
the numerical solutions show that for relatively large values of @xmath131 , only @xmath138 has a well defined radius . in this case
the effect of @xmath0 is a @xmath145 increase of the matter extension as compared to @xmath146 .
as we increase the polytropic index , @xmath147 leads to non - physical solutions while the effect with bigger values of @xmath144 becomes visible only for @xmath141 .
for instance , @xmath148 results in a radius which is @xmath149 bigger than the corresponding value with @xmath142 . the combination @xmath150 and
@xmath148 also leads to a non - physical solution . whereas the radius of the case @xmath151 displays a difference of @xmath145 as compared to @xmath142 . finally , for @xmath16 , the only physical solutions are obtained for the lowest values of @xmath131 where @xmath152 this case is particularly interesting as with @xmath153 it is often used a a viable phenomenological parametrization of densities @xcite .
the solution has an asymptotic behaviour as @xmath154 which has been also found in lsb galaxies de blok at al .
with @xmath15 the @xmath16 seems less appealing as the matter is diluted .
for all values of @xmath13 , the difference between @xmath155 and @xmath156 is negligible .
+ [ lane ] analytical solutions of ( [ le ] ) can be found for @xmath157 and @xmath16 if @xmath17 . for @xmath158 ,
the polytropic equation of state reduces to the equation of state of the isothermal sphere @xmath159 . as an example , for @xmath160 , we can write the analytical solution in the case @xmath138 as @xmath161 the radius is @xmath162 , where @xmath163 is the solution of the transcendental equation @xmath164 in the first order of @xmath131 one finds @xmath165 equation ( [ rad1 ] ) also implies that there exists some @xmath166 such that for @xmath167 , we can not find a real solution for @xmath163 .
approximately this gives @xmath168 which , provided the overall density is not too big , is better that @xmath169 which is a result from the general inequality ( [ ineq ] ) for @xmath170 and spherical symmetry .
finally , we can calculate the contribution of @xmath0 to the total energy of the object .
generalizing the results found in @xcite we obtain @xmath171.\ ] ] from which we infer that the correction is very small in this case .
+ at the end of this section we would like to summarize the findings from fig .
2 . in table 1
we write the ratio @xmath172 for the same ratios @xmath131 and polytropic index as in fig .
2 . the horizontal line represents a non - defined radius .
the symbol @xmath173 indicates that the radius is defined only asymptotically in case of @xmath17 . [ cols="^,^,^,^,^,^,^",options="header " , ]
as emphasized already before , the effect of @xmath0 can get enhanced for non - spherical objects .
this happens when the vacuum energy gets multiplied by a ratio of two length scales @xmath174 and @xmath175 and we end up with expressions like @xmath176 . for instance , for ellipsoidal configurations the geometrical parameter @xmath56 entering among others the inequality ( [ ineq ] ) can be calculated from its definition ( [ agen ] ) with @xmath177 and @xmath178 given in @xcite .
we have @xcite @xmath179^{-1 } \stackrel { a_{3 } \gg a_{1 } } { \to } \frac{2}{3}{\left(}\frac{a_{3}}{a_{1}}{\right)}^{3}\left[\ln{\left(}\frac{2a_{3}}{a_{1}}{\right)}\right]^{-1},\end{aligned}\ ] ] where the eccentricity ) becomes in this way a defining equality for @xmath180 such that @xmath181 in order to maintain equilibrium . ]
@xmath182 is @xmath183 for the oblate and @xmath184 for the prolate case .
the behaviour of @xmath185 and @xmath186 is shown in figure [ ag ] .
it is clear that for very flat astrophysical objects we can gain in this way several orders of magnitude of enhancement of the effect of @xmath0 if @xmath56 is a factor of attached to @xmath187 . needless to say that we often encounter in the universe flat objects like elliptical galaxies , spiral disk galaxies , clusters of galaxies of different forms and finally superclusters which can have the forms of pancakes
of course , the more dilute the system is , the bigger the effect of @xmath0 .
we can expect sizeable effects for clusters and superclusters , even for very flat galaxies .
regarding the latter , low density galaxies like the nearly invisible galaxies are among other the best candidates . [ ag ] a convenient way to model almost all flat shaped objects is to consider ellipsoids which in the limit of flattened spheroids can be considered as disks .
there are three different kinds of elliptical configurations , characterized by three semi - axes @xmath188 , @xmath189 and @xmath190 : oblate , with @xmath191 , prolate with @xmath192 and triaxial systems with @xmath193 . here
the tensor virial equation provides a tool to determine which of these geometries are compatible with the virial equilibrium .
considering the case @xmath153 and @xmath194constant , or spheroids with confocal density distribution whose isodensity surfaces are similar concentric ellipsoids @xcite i.e. @xmath195 the oblate ellipsoid ( maclaurin ) emerges as a solution of the virial equations with a bifurcation point to a triaxial jacobi ellipsoid @xcite .
we can view the virial equation without @xmath0 as a homogeneous equation .
switching on @xmath15 this becomes an inhomogeneous equation whose right hand side is proportional to @xmath196 .
it is therefore a priori not clear if in the case of @xmath15 we can draw the same conclusions as with @xmath17 .
+ let us assume a configuration which is rotating with constant angular velocity around the @xmath197 axis . neglecting the internal motions , the @xmath0-tensor virial eq .
( [ virotro ] ) for such a configuration is @xmath198 the equations with diagonal elements yield the following identities @xmath199 this set of equations can be resolved for the angular velocity : @xmath200 the same expression holds if we make the replacement @xmath201 and @xmath202 on the right hand side of ( [ ns01 ] ) .
we can also eliminate @xmath203 from these expressions to arrive at the condition @xmath204 - 1 = \frac{8\pi}{3}g_n\rho_{\rm vac}{\mathcal{i}}_{zz}{\left(}\frac{1-\delta_i}{|{\mathcal{w}}_{yy}^{\rm n}|-|{\mathcal{w}}_{zz}^{\rm n}|}{\right)}.\ ] ] with @xmath205 .
this expression determines the possible ellipsoidal configurations .
one solution is obvious : @xmath206 together with @xmath207 .
this happens if the density is constant or for a confocal ellipsoid with the density given in ( [ density1 ] ) .
hence this is not different from the @xmath17 case .
however , the bifurcation point to a triaxial system will get affected by @xmath0 @xcite .
furthermore , the explicit calculation of this bifurcation point will now depend also on the details of the density profile even if we take the latter to be as in eq .
( [ density1 ] ) .
this is a direct consequence of a ` pre - existing ' density scale @xmath196 . for a homogeneous configuration
we have @xmath208\\ \nonumber \tilde{\delta}_i{\left(}\frac{\eta_{y}}{\eta_{x}}{\right)}{\left(}\frac{1-\eta_{x}}{1-\eta_{y}}{\right)}-1&= & { \left(}\frac{\rho_{\rm vac}}{\rho}{\right)}{\mathcal{a}}_{zz}\eta_{y } { \left(}\frac{1-\tilde{\delta}_i}{1-\eta_{y}}{\right)}.\end{aligned}\ ] ] where @xmath209 , @xmath210 and @xmath211 are defined in eq .
( [ agen ] ) . for an oblate configuration
they can be calculated in terms of the eccentricity @xmath212 to give @xcite @xmath213^{-1},\hspace{0.8 cm } { \mathcal{a}}_{zz}=\frac{8}{3}\frac{e^{2}}{\sqrt{1-e^{2}}}\left[\frac{1}{\sqrt{1-e^{2}}}-\frac{\arcsin e}{e}\right]^{-1}.\ ] ] furthermore we have @xmath214 using these expressions , we write the constant angular velocity for an oblate ellipsoid from the first line of equation eq .
( [ ns02 ] ) as @xmath215,\ ] ] where @xmath216 corresponding to the angular velocity when @xmath17 is given by the maclaurin formula @xmath217\ ] ] the function @xmath218 defined through eq . ( [ ns02 ] ) can also be calculated with explicit dependence on the eccentricity as @xmath219^{-1}.\ ] ] as is evident from the above equations , @xmath0 has a twofold effect on the angular velocity .
firstly , it reduces the angular velocity with respect to the value @xmath216 especially at the local maximum ( see figure 3 ) .
this is not a small effect and can affect even galaxies .
secondly , we see from eq .
( [ omcero ] ) that @xmath220 for @xmath221 . on the other hand @xmath222 approaching @xmath119 for a very flat oblate configuration and not too dense matter .
therefore , beyond the local maximum in @xmath223 the cosmological constant causes a steeper fall of @xmath223 towards @xmath224 .
another relevant interesting quantity which can be calculated in this context is the ratio of the rotational over the gravitational energy contributions to the scalar virial equations , i.e. , @xmath225 . in accordance with eq .
( [ omcero ] ) the latter can be written as @xmath226 , \hspace{0.8cm}\beta_{0}=\frac{3}{2e^{2}}\left[1-\frac{e(1-e^{2})^{1/2}}{\arcsin e}\right]-1.\ ] ] the effects on @xmath227 are therefore similar to the the ones encountered in @xmath223 ( see also figures 3 ) . finally , on account of @xmath228
we can infer from the virial equations with @xmath194const the following inequality @xmath229 which together with ( [ beta ] ) results in an inequality for the density of an ellipsoidal configuration : @xmath230 for @xmath231 , we have @xmath232 , while for @xmath233 , @xmath234 , and @xmath235 .
therefore the above inequality is slightly weaker than the bound given in ( [ ineq ] ) .
nevertheless it is useful since it it derived directly from a different starting point ( @xmath236 ) than ( [ ineq ] ) which is based on @xmath237 . not always the deviation from spherical symmetry guarantees that the effect of @xmath0 on observables is sizeable .
this depends on the context and also which scales we compare . if we compare @xmath196 to densities , the quantities @xmath56 and @xmath211 defined in eq .
( [ ineq ] ) are for flattened objects large enough to enhance the effect of @xmath0 . if @xmath6 is combined with the schwarzschild radius @xmath78 to give @xmath116 as in eq .
( [ rvir2 ] ) , the result is still of astrophysical relevance .
but if we had to compare one of the axes of an ellipsoid to @xmath6 ( i.e. @xmath238 ) the effect would be negligible unless the extension @xmath239 is bigger than mpc ( clusters and superclusters ) and the small ratio @xmath238 is comparable to another small quantity of the same order of magnitude entering the equation under consideration .
this happens for instance if we generalize a result ( discussed @xcite ) on a mass - weighted mean - square rotation speed @xmath240 of an ellipsoidal object to include @xmath0 .
assume that due to symmetry properties of the object the only relevant components of the tensors in the tensor virial equations are @xmath241 and @xmath242 .
we then obtain @xmath243 if the only motion is a rotation about the @xmath197 axis we have @xmath244 and we can solve for @xmath245 as @xmath246 using eqs .
( [ mapp ] ) , ( [ virotro ] ) and ( [ mof ] ) , the quantities entering our equation can be parametrized in the following way @xmath247 where @xmath248 is the mass weighted mean angular velocity , @xmath249 is the mass - weighted mean - square random velocity in the @xmath88 direction and @xmath250 measures the anisotropy in @xmath251 . if @xmath250 is of the order of one , it suffices to compare @xmath252 with @xmath253 . if both are of the same order of magnitude , the effect of @xmath0 is non - negligible .
since @xmath240 and @xmath249 are of the same order of magnitude and @xmath240 is non - relativistic , we can assume that @xmath254 .
the quantity @xmath255 can be estimated to be @xmath256 .
hence if @xmath257 , @xmath240 ( @xmath249 ) has to be truly non - relativistic and of the the order of @xmath258 to gain an appreciable effect of @xmath0 .
this improves if @xmath239 is one order of magnitude bigger which is possible for large galaxy clusters .
the velocities have to be then at most @xmath259 .
in these cases we have to keep @xmath0 and while solving ( [ para ] ) for @xmath260 one has @xmath261.\ ] ] note that if @xmath0 can not be neglected @xmath262 is not only a function of the eccentricity for ellipsoids with the density give in ( [ density1 ] ) , but depends also on the details of the matter density as the latter does not cancel .
+ using almost the same set - up as above , we can use equation ( [ curious ] ) to establish a relation between@xmath240 , the mass @xmath94 and the geometry of the object which we choose below to be oblate .
after straightforward algebra we obtain @xmath263+(1-\delta_{0}){\left(}\frac{a_{1}}{a_{3}}{\right)}^{2}-1.\ ] ] since this relation is derived from ( [ curious ] ) which in turn is based on the assumption of @xmath15 it is only valid for non - zero cosmological constant albeit the latter does not enter the expression .
note the enhancement factor @xmath264 .
[ fomega ] we can now derive other relevant quantities from the scalar @xmath0-virial theorem applied to homogeneous ellipsoidal configurations . in this section
we will not consider rotating configurations , but systems with kinetic energy coming from internal motions .
we will focus again on ellipsoidal , oblate and prolate , geometries . as in the preceding section
the relevant quantity here is the function @xmath55 written for both configurations in equation ( [ a ] ) .
_ critical mass _ : consider the stability criteria for a homogeneous cloud with mass @xmath94 and internal mean velocity @xmath265 in mechanical equilibrium with the background with pressure @xmath22 ( see @xcite ) .
the system will collapse under it s own gravity ( @xmath266 ) if its mass is greater than a critical mass @xmath267 . with cosmological constant , this critical mass
is increased with respect to its value @xmath268 when @xmath17 , which is expected , since now there is an _ external force _ associated to @xmath0 that acts against newtonian gravity and hence the collapse can be postponed . by using the scalar virial theorem ( [ vt ] ) and setting @xmath269 as the criterion for the onset of instability
, we can write for arbitrary geometry + @xmath270^{-1},\hspace{0.8cm}m_{c0 } = \frac{2\langle v^{2}\rangle}{|\tilde{{\mathcal{w}}}^{\rm n}|},\hspace{0.8cm}\rho_{c0}=\frac{m_{c0}}{v}.\ ] ] this expression is valid for any geometry .
however , we pointed out already that spherical symmetry implies @xmath271 and the effect is suppressed .
some numerical values can be considered by writing @xmath272 for a hydrogen cloud with @xmath273 k and radius @xmath274pc .
one then has @xmath275 which represents a very tiny correction to the critical mass @xmath268 for a spherically symmetric object .
on the other hand , for ellipsoidal configurations with the same temperature , we have @xmath276 in the @xmath233 approximation for oblate and prolate configurations , respectively .
we have set @xmath277 for the prolate case and @xmath278 . for an oblate ellipsoid with @xmath279 kpc , one has @xmath280 while for the prolate ellipsoid with @xmath281 , @xmath282 .
_ mean velocity and mass - temperature relation_. by virtue of the scalar @xmath0-virial theorem , we can also write down the mass - temperature relation for an astrophysical structure . note first that using @xmath283 in equation ( [ vt ] ) we have for @xmath17 the standard expression for the mean velocity @xmath284 where the second equality applies to the constant density case . clearly , with @xmath17 the mean velocity can not become zero .
let us contrast it to the case with @xmath27 .
one obtains @xmath285.\ ] ] as can been seen from ( [ mvel ] ) the mean velocity in the ellipsoidal configurations is decreased because of the @xmath0-_external force_. a drastic effect of the cosmological constant could be reached for the geometrical factor @xmath55 approaching the critical value @xmath286 which is possible for very flat objects . in the extreme
the mean velocity can go to zero ) this is somewhat different since @xmath287 as @xmath233 independently of the density .
] together with inequality ( [ ineq ] ) , the result in eq .
( [ mvel ] ) tells us that the temperature , @xmath288 of the objects is very small if the objects is above the limit to reach equilibrium ( the square bracket in eq .
( [ mvel ] ) is then very small ) .
this is a qualitative conclusion based in the presence of non - zero cosmological constant .
for ellipsoidal configurations , this defines also a maximum value for the eccentricity @xmath180 given @xmath289 or vice - versa , i.e , a minimum value for the density @xmath290 given @xmath182 , through the relation @xmath291 .
the bahavior of @xmath180 is shown in fig .
5 as a function of @xmath65 .
galactic clusters with @xmath292 may have a vanishing mean velocity for @xmath293 in the oblate case and @xmath294 in the prolate case .
if the density is smaller , a vanishing mean velocity can be reached for non so flat objects .
spherical configurations have @xmath295 for @xmath296 .
+ since the mean velocity is proportional to the temperature , @xmath297 , eq . ( [ mvel ] ) represents also a mass - temperature relation . hence the results for the mean velocity squared are also applicable to the temperature of the configuration . for instance
, we can write eq .
( [ mean ] ) for a cosmological structure , say a galactic cluster , by writing its density as resulting from a perturbation @xmath298 from the background density of the universe @xmath299 as @xmath300 , where @xmath301 .
equation ( [ mvel ] ) allows us to determine the temperature of the cluster at a given cosmic time @xmath302 as @xmath303 , \\
\nonumber \mathcal{f}(e)&=&\begin{cases } a_{1}^{2}e^{-1}\sqrt{1-e^{2}}\ln{\left(}\frac{1+e}{1-e}{\right ) } & \text{prolate},\\ a_{1}^{2}e^{-1}\sqrt{1-e^{2}}\arcsin e & \text{oblate } , \end{cases}\end{aligned}\ ] ] where @xmath304 is the scale factor and @xmath75 is the mass of average components of the cluster .
equation ( [ mvel2 ] ) assumes a flat universe @xmath305 .
this result is a generalization of a result derived in @xcite valid for spherical geometry . in that case one
recovers the typical mass temperature relation @xmath103 mantaining @xmath289 constant .
although this has the same dependence as in equation ( [ temperature3 ] ) , the meaning is different since ( [ mvel2 ] ) for @xmath271 computes the temperature of a certain galactic cluster at some redshift given its mass while equation ( [ temperature3 ] ) is associated to a reversible process where a configuration passes from @xmath306 to some final @xmath99 through states of virial equilibrium mantaining a constant mass .
the stability condition of newtonian configurations against oscillations can be also derived from the second order virial equation by expanding the periodic lagrangian perturbations @xmath307 with oscillation frequency @xmath308 . for simplicity
, we will consider a rotational configuration without internal motions . by assuming adiabatic perturbations , the variational form of the second order virial equation ( [ virotro ] )
is written as ( see @xcite ) @xmath309{{\text{d}}^{3}r}\\ \nonumber & + & \frac{1}{3}\delta_{ik}\int_{v}(5 - 3\gamma)\rho\left [ ( { \mathbf{\omega}}\cdot { \mbox{\boldmath$\xi$}})({\mathbf{\omega}}\cdot { \textbf{r } } ) -\omega^{2}({\textbf{r}}\cdot { \mbox{\boldmath$\xi$}})\right]{{\text{d}}^{3}r},\end{aligned}\ ] ] where @xmath310 is the adiabatic index governing the adiabatic perturbations .
for polytropic e.o.s we have @xmath311 , of course .
the lagrangian displacement is constrained with the boundary conditions @xmath312 , @xmath313 at @xmath314 and @xmath315 @xmath316 must be finite at the surface .
the boundary conditions are required to solve the _ equation of motion _ for @xmath316 resulting from perturbing euler s equation ( the sturm - liouville eigenvalue equation ) .
+ let us consider radial oscillations which are adequate for spherical systems and for small deviation from spherical symmetry . by assuming a general trial function for the lagrangian displacement @xmath317 satisfying the boundary conditions and taking the trace in eq .
( [ os21 ] ) , we can solve for frequency of the oscillation about equilibrium @xmath318,\ ] ] where we have assumed @xmath319 constant throughout the configuration and we have defined in analogy to eq .
( [ ineq ] ) the quantities @xmath320 in eq .
( [ defi0 ] ) @xmath321 is a parameter with units of density , and @xmath322{{\text{d}}^{3}r},\hspace{0.4 cm } i\equiv \int_{v}\rho(r ) f(r)r{{\text{d}}^{3}r},\ ] ] together with @xmath323 .
the critical adiabatic index is written as @xmath324,\ ] ] such that for @xmath325 instability sets in and the system becomes unstable while perturbed . from eq .
( [ os22 ] ) , we see that through the inclusion of the cosmological constant we ( the system ) are forced to choose a bigger adiabatic index . in the simplest case ,
when @xmath326 , we have @xmath327 , @xmath328 , @xmath329 . furthermore , for @xmath194constant , we get @xmath330 and @xmath331 .
the stability condition then becomes @xmath332.\ ] ] the spherical symmetry which we assumed compels us to write the ratio @xmath227 for low eccentricities or to parametrize it in terms of the total angular momentum of the configuration . in the first approximation , at low eccentricities
the ratio @xmath227 given in ( [ beta ] ) takes the form @xmath333 this equation is useful if we want to calculate a small , but peculiar effect . insisting that @xmath334 is very close to @xmath335 we can convert the stability condition @xmath336 into a condition on eccentricity , namely @xmath337 where @xmath338 measures a small departure from the critical value @xmath335 . for the specific case @xmath339 , we conclude that the eccentricity must be such that @xmath340 with @xmath341 which is clearly valid only for large densities , say for @xmath342 .
this imposes a range for the eccentricity @xmath343 for stability under radial perturbations .
+ for completely spherical configurations with constant density , we can write @xmath344 . figure [ gamma ] shows the behavior of the critical adiabatic index as a function of @xmath227 for different values of @xmath345 .
the largest deviation is for low densities as expected .
+ [ gamma ] for non rotating configurations , the stability criteria is simply given by @xmath346,\hspace{0.5cm}a=\frac{8\pi}{3}\varrho\frac{\int_{0}^{r } \rho(r)r^{2}f(r){\text{d}}r}{\int_{0}^{r}\rho(r)r^{2}f(r)\frac{{\text{d}}\phi^{\rm n}}{{\text{d}}r}{\text{d}}r}.\ ] ] if we work with a constant density profile , the choice of another trial function satisfying the boundary conditions mentioned before will not play a role in the above expressions since @xmath347 for any @xmath348 .
+ related results regarding the adiabatic index in cosmologies with non - zero @xmath0 have been obtained in the relativistic frame work in @xcite .
in the present work we investigated in detail the astrophysical relevance of the cosmological constant for equilibrium configurations . using the tensor and scalar virial equations and the lane - emden equation we could show that many astrophysical facets get modified by @xmath0 .
the second aspect concerns the fact that @xmath0 introduces new relevant scales ( these scales would be zero if @xmath17 ) like the maximal virial volume defined by the maximal virial radius ( [ rvir2 ] ) and the maximal extension of bound orbits given in eq . ( [ sca3 ] ) . + it is often assumed that superclusters with densities @xmath349 are not in equilibrium . with the inequality ( [ ineq ] ) we have a precise tool to quantify this statement . indeed , the pancake structure of the superclusters lead us to the conclusionthat they are even far away from the equilibrium state due to the factor @xmath55 which grows with the object s flatness . on the other hand , relatively low density objects can still reach equilibrium ( even if @xmath271 for the spherical case , the density of the object has to be only twice as large as @xmath196 , i.e , @xmath350 ) .
there is nothing which could , in pronciple , prevent a relatively low density object to be in equilibrium this is not only a result obtained from the virial equations with @xmath0 , but follows also from hydrostatic equilibrium with the inclusion of @xmath0 @xcite .
it is then natural to put forward the question how the shape of such objects affect their properties as compared to the case without @xmath0 .
for instance , if @xmath351 and @xmath352 , the angular velocity is reduced approximately @xmath353 with respect to its value with @xmath17 ( @xmath354 ) .
other effects discussed in the text are the mean velocity ( see fig . 5 ) and the adiabatic index ( see fig .
6 ) . to be specific
let us takje an example of a very low density object , @xmath355 .
this object is still virialized as long as its eccentricity is @xmath356 .
approaching @xmath357 the mean velocity ( temperature ) goes to zero until finally beyond @xmath358 the objects ceases to be in equilibrium .
+ to highlight some quantities which get affected let us mention the polytropic index @xmath13 and the angular velocity .
low - density objects with @xmath359 an index @xmath360 becomes unphysical as the the conglomeration of matter does not have a definite finite radius .
taken together with the results of section five on stability against small oscillations this implies that the adiabatic index @xmath361 from the equation of state @xmath362 gets restricted for these densities in a rather strong way , namely by @xmath363 .
for higher densities the effect grows with @xmath13 .
for instance , for @xmath364 , @xmath16 is an unacceptable solution due to the lack of a finite extension . for non - spherical configurations we could show an effect of @xmath0 on angular velocities , the mean velocity of the component of the astrophysical object and on the critical mass , again for densities one and two orders of magnitude above the critical one .
we think that the work can be generalized in various ways .
for instance , to generalize the virial equations to be able to differentiate between different models of dark energy mentioned at the beginning .
another aspect is to look into the lane - emden equation for non - spherical geometries @xcite .
here we concentrated on the equilibrium condition of already virialized matter .
the collapse of matter in an accelerated universe with dark energy has been considered in @xcite .
here we briefly show the solution for the cubic equation ( [ cubic ] ) .
let us write that expression as @xmath365 corresponding to a positive cosmological constant .
associated with this expression one defines a discriminant and a quantity @xmath134 defined as @xmath366 since @xmath367 , the roots of ( [ co4 ] ) are given in term of the auxiliary angle @xmath368 defined as @xmath369 so that the only real solution for ( [ co4 ] ) is written as @xmath370.\ ] ] the other two roots are complex numbers whose real part are negative .
if @xmath371 and @xmath372 the roots depend whether @xmath373 is bigger or smaller than zero .
if @xmath374 , the three roots are real and negative , while @xmath375 yields one real negative root and two complex roots which real parts are given as @xmath376.\ ] ] 10 krauss , l. m. 1998 , astrophys .
j. , 501 , 1998 krauss , l. m. 2003 , cosmology and fundamental physics , march 2003 perlmutter , s. et al .
1998 , nature , 391 , 51 perlmutter , s .
, turner , m. s. , white , m. j. 1999 , phys .
lett . , 83 , 670 riess , a. g. et al.2004 , astrophys .
j. , 607 , 665 allen , s. w. , schmidt , r. w , ebeling , h. , fabian , a. c. , van speybroeck , l. 2004 , astro - ph/0405340 ratra , b. & peebles , p. j. e. 1987 , phys .
rev.d , 37 , 3406 kremer , g. m. 2003 , gen .
grav . , 35 , 1459 kamenshchik , a. yu . ,
moschella , u. , pasquier , v. 2001 , phys .
b511 , 265 nojiri , s. & odintsov , s. d. 2003 , phys .
, d68 , 123512 nowakowski , m 2000 , int . j. mod .
phys . , d10 , 649 balaguera - antolnez , a. , bhmer , c. g. , nowakowski , m. , int . j. mod phys * d * , vol 14 .
( 2005 ) , gr - qc/0409004 nowakowski , m. , sanabria , j .- c . ,
garcia , a. 2001 , phys .
d66 , 023003 sussman , r. a. & hernandez , x. 2003 , mon . not .
, 345 , 871 balaguera - antolnez , a. , bhmer , c. g. , nowakowski , m. 2006 , class .
grav . , 23 , 485 , gr - qc/0511057 jaaniste , j. , tago , e. , einasto , m. , einasto , j. , andernach , h. , mller , v. 1998 , astronomy & astrophysics , 336 , 35 padmananbhan , t.1993 , structure formation in the universe ( cambridge university press ) cooray , a. r. 2000 , mon . not .
soc . , 313 , 783 kalinkov , m. , valchanov , t. , valtchanov , i. , kuneva , i. , dissanska , m. 2005 , astro - ph/0505091 de blok , w. j. g. & mcgaugh , s. s. 1996 , astrophys . j. lett . , 469 , 103 de blok , w. j. g. , mcgaugh , s. s. , bosma , a. and rubin , v. c. 2004 , astro - ph/0103102 chandrasekhar , s. 1967 , an introduction to the study of stellar structure , ( dover publications ) plummer , h. c. 1911 , mon . not .
soc . , 71 , 460 bacry , h. & levy - leblond , j. m. 1967 , possible kinematics , j. math .
, 9 , 1605 derom , j .- r . & dubois , j .-
1972 , nouvo cimento , 9b , 351 aldrovandi , r. , barbosa , a. l. , crispino , l. c. , pereira , j. g. 1999 , class .
quantum grav . , 16 , 495 chandrasekhar , s. 1961 , hydrodynamic and hydromagnetic stability , ( dover publications ) barrow , j. d. & gotz , g. 1989 , class .
, 6 , 1253 wang , l & steinhardt , p. j. 1998 , astrophys . j. , 508 , 483 bhmer , c. g. 2002 , gr - qc/0308057 , unpublished diploma thesis bhmer , c. g. 2003 , gen .
grav . , 36 , 1039 oppenheimer , j. r. & volkoff , g. m. 1939 , phys . rev .
, 55 , 374 rich , j. 2001 , fundamentals of cosmology ( springer verlag ) mak , m. k. , jr .
dobson , p.n .
, harko , t. 2001 , mod .
lett . , a15 , 2153 binney , j & tremaine , s. 1987 , galactic dynamics , ( princeton university press ) roberts , p. h. 1962 , astrophys . j. , 136 , 1108 chandrasekhar , s. 1969 , ellipsoidal figures of equilibrium ( new haven and london , yale university press ) balaguera - antolnez , a. , nowakowski , m. 2005 , in preperation padmanabhan t. 2000 , theoretical astrophysics , vol.i : astrophysical processes ( cambridge university press ) te , m. j. 1999 , phys .
, 83 , 670 friedman , j. , schutz , b. , 1978 , astrophys .
j. , 221 , 937 shapiro , s. & teokolsky , s. 1983 , black holes , white dwarfs and neutron stars ( wiley - interscience publications ) bhmer , c. g. & harko , t. 2005 , phys .
d. , 71 , 084026 lombardi , m & bertin , g. 2001 , astronomy & astrophysics , 375 , 1091 mota , d. f. & van de bruck , c. 2004 , astronomy & astrophysics , 421 , 71 | using the scalar and tensor virial equations , the lane - emden equation expressing the hydrostatic equilibrium and small oscillations around the equilibrium , we show how the cosmological constant @xmath0 affects various astrophysical quantities important for large matter conglomeration in the universe . among others we examine the effect of @xmath0 on the polytropic equation of state for spherically symmetric objects and
find non - negligible results in certain realistic cases .
we calculate the angular velocity for non - spherical oblate configurations which demonstrates a clear effect of @xmath0 on high eccentricity objects .
we show that for oblate as well as prolate ellipsoids the cosmological constant influences the critical mass and the temperature of the astrophysical object . these and other results show that the effect of @xmath0 is large for flat astrophysical bodies . _ keywords : large - scale structure of universe .
galaxies : clusters : general . instabilities . |
the galaxy cluster mass function is the most fundamental statistic of the galaxy cluster population .
it is determined by the initial conditions of the mass distribution set in the early universe in a relatively straightforward way , since the evolution of the large - scale matter distribution on scales comparable to the size of clusters and larger is linear and since the formation of clusters is governed by essentially only gravitational processes . choosing a specific cosmological scenario
provides definitive predictions about these initial conditions in a statistical sense .
the overall process of the gravitational growth of the density fluctuations and the development of gravitational instabilities leading to cluster formation is comparatively easy to understand .
it has been well described by analytical models ( e.g. , @xcite , @xcite , @xcite , @xcite , @xcite ) and simulated in numerical gravitational @xmath8-body calculations ( e.g. , @xcite ) . even though the slight deviations from the theoretical prescription found in recent simulations ( e.g. , @xcite ) seem to require further refinement in the theoretical framework , these deviations are too small to be significant for the current investigation , as will be shown later . for the precision needed here
these large simulations therefore support the validity of the model predictions . within this framework
the observed cluster mass function provides the opportunity to test different cosmological models .
the tests are particularly sensitive to the amplitude of the cosmic matter density fluctuations ( at a scale of the order of 10mpc ) as well as the normalized total matter density , @xmath9 ( e.g. , @xcite ) .
in addition to its importance in testing cosmological models the integral of the mass function yields the interesting information on the fractional amount of matter contained in gravitationally bound large - scale structures .
using one of the first attempts to construct a mass function over the mass range from giant ellipticals to massive clusters by @xcite @xcite obtain a mass fraction @xmath10 for @xmath11 ( where the mass fraction is expressed here in units of the critical density of the universe , @xmath12 ) .
this result is already close to the total matter density in some of the proposed cosmological scenarios .
therefore a precise observational determination of the integral mass function is a very important task for astronomy .
unfortunately the galaxy cluster mass is not an easily and directly observable quantity ( except for measurements of the gravitational lensing effect of clusters which may play a large role in the construction of mass functions in the future ) and one has to resort to the observation of other cluster parameters from which the cluster masses can be deduced .
x - ray astronomy has provided an ideal tool to first detect and select massive clusters by measuring their x - ray luminosity and to secondly perform mass determinations on individual clusters through x - ray imaging and x - ray spectroscopy . in this paper
we report the first rigorous application of these two approaches for the construction of the cluster mass function .
building on the rosat all - sky x - ray survey ( rass ) @xcite , which has been well studied in the search of the brightest galaxy clusters through several survey projects ( see refs in sect .
[ sample ] ) , we have compiled a new , highly complete sample of the x - ray brightest galaxy clusters ( hiflugcs , the highest x - ray flux galaxy cluster sample ) .
thanks to the numerous detailed galaxy cluster observations performed with the rosat and asca @xcite satellite observatories and accumulated in the archives we can perform a detailed mass determination for most of these clusters and obtain a good mass estimate for the few remaining objects . from these data we first establish a ( tight ) correlation of the measured x - ray luminosity and the cluster mass .
this relation assures that we have essentially sampled the most massive clusters in the nearby universe , which forms the basis of the construction of the cluster mass function .
previous local galaxy cluster mass functions have been derived by @xcite , @xcite , @xcite , and ( * ? ? ? * for galaxy groups ) .
@xcite used the galaxy richness to relate to cluster masses for optical observations and an x - ray temperature mass relation to convert the temperature function given by @xcite to a mass function .
@xcite , @xcite , and @xcite used velocity dispersions for optically selected samples to determine the mass function .
here we use a different approach and construct the first mass function for x - ray selected galaxy clusters based on the rass using individually determined cluster masses .
the mass function of this cluster sample is then used to determine the mass fraction in bound objects with masses above a minimum mass and to derive tight constraints on cosmological scenarios .
the paper is organized as follows . in sect .
[ sample ] the sample selection is described .
the details of the determination of the observational quantities are given in sect .
the results are presented in sect .
[ resul ] and discussed in sect . [ discu ] .
the conclusions are summarized in sect .
[ conclu ] . throughout a hubble constant @xmath13 , @xmath14 , @xmath15 , a normalized cosmological constant @xmath16 , and a normalized curvature index @xmath17
is used if not stated otherwise ( present day quantities ) .
we note that the determination of physical cluster parameters has a negligible dependence on @xmath9 and @xmath18 for the small redshift range used here , as will be shown later .
therefore it is justified to determine the parameters for an einstein de sitter model but to discuss the results also in the context of other models .
the mass function measures the cluster number density as a function of mass .
therefore any cluster fulfilling the selection criteria and not included in the sample distorts the result systematically .
it is then obvious that for the construction of the mass function it is vital to use a homogeneously selected and highly complete sample of objects , and additionally the selection must be closely related to cluster mass . in this work
the rass , where one single instrument has surveyed the whole sky , has been chosen as the basis for the sample construction .
using the x - ray emission from the hot intracluster medium for cluster selection minimizes projection effects and the tight correlation between x - ray luminosity and gravitational mass convincingly demonstrates that x - ray cluster surveys have the important property of being mass selective ( sect .
[ relat ] ) .
several cluster catalogs have already been constructed from the rass with high completeness down to low flux limits ( see refs below ) .
these we have utilized for the selection of candidates .
low thresholds have been set for selection in order not to miss any cluster due to measurement uncertainties .
these candidates have been homogeneously reanalyzed , using higher quality rosat ps . pointed observations whenever possible ( sect .
[ data ] ) . a flux limit well above the limit for candidate selection
has then been applied to define the new flux - limited sample of the brightest clusters in the sky . in detail
the candidates emerged from the following input catalogs .
table [ tbl : cand ] lists the selection criteria and the number of clusters selected from each of the catalogs , that are contained in the final sample .
+ 1 ) the reflex ( rosat - eso flux - limited x - ray ) galaxy cluster survey @xcite covers the southern hemisphere ( declination @xmath19 + 2.5deg ; galactic latitude @xmath20deg ) with a flux limit @xmath21 .
+ 2 ) the noras ( northern rosat all - sky ) galaxy cluster survey @xcite contains clusters showing extended emission in the rass in the northern hemisphere ( @xmath22 0.0deg ; @xmath20deg ) with count rates @xmath23 .
+ 3 ) noras ii ( j. retzlaff et al . , in preparation ) is the continuation of the noras survey project .
it includes point like sources and aims for a flux limit @xmath24 .
+ 4 ) the bcs ( rosat brightest cluster sample ) @xcite covers the northern hemisphere ( @xmath22 0.0deg ; @xmath20deg ) with @xmath25 and redshifts @xmath26 . + 5 ) the rass 1 bright sample of clusters of galaxies @xcite covers the south galactic cap region in the southern hemisphere ( @xmath27 + 2.5deg ; @xmath28deg ) with an effective flux limit @xmath29 between @xmath303 and @xmath31 .
+ 6 ) xbacs ( x - ray brightest abell - type clusters of galaxies ) @xcite is an all - sky sample of @xcite / aco @xcite clusters limited to high galactic latitudes @xmath20deg with nominal aco redshifts @xmath32 and x - ray fluxes @xmath33 .
+ 7 ) an all - sky list of abell / aco / aco - supplementary clusters ( h. bhringer 1999 , private communication ) with count rates @xmath34 .
+ 8) early type galaxies with measured rass count rates from a magnitude limited sample of @xcite have been been checked in order not to miss any x - ray faint groups .
+ 9 ) all clusters from the sample of @xcite and @xcite , where clusters had been compiled from various x - ray missions , have been checked .
lcccccccc reflex & & 0.9 & & & 1.7 & & 33 & 1 + noras & & 0.7 & & & 1.7 & & 25 & 2 + noras ii & & 0.7 & & & 1.7 & & 4 & 3 + bcs & 1.0 & & & & 1.7 & & 1 & 4 + rass 1 & & & & & & 1.0 & 0 & 5 + xbacs & 1.0 & & & & 1.7 & & 0 & 6 + abell / aco & & 0.7 & & & & & 0 & 7 + early type gx & & & 0.7 & & & & 0 & 8 + previous sat & & & & & & & 0 & 9 + the main criterion for candidate selection , a flux threshold @xmath35 , has been chosen to allow for measurement uncertainties in the input catalogs .
e.g. , for reflex clusters with @xmath36 the mean statistical flux error is less than 8% . with
an additional mean systematic error of 6% , caused by underestimation of fluxes due to the comparatively low rass exposure times , the flux threshold @xmath35 for candidate selection then ensures that no clusters are missed for a final flux limit @xmath37 .
almost none of the fluxes given in the input catalogs have been calculated using a measured x - ray temperature , but mostly using gas temperatures estimated from an @xmath38@xmath39 relation . in order to be independent of this additional uncertainty clusters
have also been selected as candidates if they exceed a count rate threshold which corresponds to @xmath40 for a typical cluster temperature , @xmath41 , and redshift , @xmath42 , and for an exceptionally high column density , e.g. , in the noras case @xmath43 .
most of the samples mentioned above excluded the area on the sky close to the galactic plane as well as the area of the magellanic clouds . in order to construct a highly complete sample from the candidate list we applied the following selection criteria that successful clusters must fulfil : + 1 ) _
redetermined _ flux @xmath44 , + 2 ) galactic latitude @xmath45deg , + 3 ) projected position outside the excluded 324deg@xmath46 area of the magellanic clouds ( see tab .
[ tab : mc ] ) , + 4 ) projected position outside the excluded 98deg@xmath46 region of the virgo galaxy cluster ( see tab . [ tab : mc ] ) .
lccc & 58 @xmath47 & @xmath48 & 0.0655 + lmc 2 & 81 @xmath49 & @xmath50 & 0.0060 + lmc 3 & 103 @xmath51 & @xmath52 & 0.0030 + smc 1 & 358.5 @xmath53 & @xmath54 & 0.0189 + smc 2 & 356.5 @xmath55 & @xmath56 & 0.0006 + smc 3 & 20 @xmath57 & @xmath58 & 0.0047 + virgo & 182.7 @xmath59 & @xmath60 & 0.0297 + milky way & 0 @xmath61 ( @xmath62 ) & @xmath63 ( @xmath64 ) & 4.2980 + these selection criteria are fulfilled by 63 candidates .
the advantages of the redetermined fluxes over the fluxes from the input catalogs are summarized at the end of sect .
[ fluxd ] . in tab .
[ tbl : cand ] one notes that 98% of all clusters in hiflugcs have been flagged as candidates in reflex , noras , or in the candidate list for noras ii ; these surveys are not only all based on the rass but all use the same algorithm for the count rate determination , further substantiating the homogeneous candidate selection for hiflugcs .
the fraction of available rosat ps . pointed observations for clusters included in hiflugcs equals 86% .
the actually used fraction is slightly reduced to 75% because some clusters appear extended beyond the ps .
field of view and therefore rass data have been used .
the fraction of clusters with published asca temperatures equals 87% . if a lower flux limit had been chosen the fraction of available ps
. pointed observations and published asca temperatures would have been decreased thereby increasing the uncertainties in the derived cluster parameters .
furthermore this value for the flux limit ensures that no corrections , due to low exposure in the rass or high galactic hydrogen column density , need to be applied for the effective area covered .
this can be seen by the effective sky coverage in the reflex survey area for a flux limit @xmath65 and a minimum of 30 source counts , which amounts to 99% .
the clear advantage is that the hiflugcs catalog can be used in a straightforward manner in statistical analyses , because the effective area is the same for all clusters and simply equals the covered solid angle on the sky . the distribution of clusters included in hiflugcs projected onto the sky is shown in fig .
[ aito ] . the sky coverage for the cluster sample equals 26721.8deg@xmath46 ( 8.13994sr ) , about two thirds of the sky .
the cluster names , coordinates and redshifts are listed in tab .
[ tbl : data1 ] .
further properties of the cluster sample are discussed in sect .
[ sample_c ] . for later analyses which do not necessarily require a complete sample , e.g. ,
correlations between physical parameters , 43 clusters ( not included in hiflugcs ) from the candidate list have been combined with hiflugcs to form an ` extended sample ' of 106 clusters .
this section describes the derivation of the basic quantities in this work , e.g. , count rates , fluxes , luminosities , and mass estimates for the galaxy clusters . these and other relevant cluster parameters are tabulated along with their uncertainties . measuring
the count rate of galaxy clusters is an important step in constructing a flux - limited cluster sample .
the count rate determination performed here is based on the growth curve analysis method @xcite , with modifications adapted to the higher photon statistics available here .
the main features of the method as well as the modifications are outlined below .
the instrument used is the rosat ps .
@xcite , with a low internal background ideally suited for this study which needs good signal to noise of the outer , low surface brightness regions of the clusters . mainly pointed observations from the have been used . if the cluster is extended beyond the ps .
field of view making a proper background determination difficult or if there is no pointed ps .
observation available , rass data have been used .
the rosat hard energy band ( channels @xmath66 ) has been used for all count rate measurements because of the higher background in the soft band .
two x - ray cluster centers are determined by finding the two - dimensional ` center of mass ' of the photon distribution iteratively for an aperture radius of 3 and 7.5arcmin around the starting position . the small aperture yields the center representing the position of the cluster s peak emission and therefore probably indicates the position where the cluster s potential well is deepest
. this center is used for the regional selection , e.g. @xmath45 . the more globally defined center with the larger aperture
is used for the subsequent analysis tasks since for the mass determination it is most important to have a good estimate of the slope of the surface brightness profile in the outer parts of the cluster .
the background surface brightness is determined in a ring outside the cluster emission . to minimize the influence of discrete sources the ring
is subdivided into twelve parts of equal area and a sigma clipping is performed . to determine the count rate the area around the global center
is divided into concentric rings . for pointed observations
200 rings with a width of 15arcsec each are used . due to the lower photon statistics a width of 30arcsec is used for rass data and the number of rings depends on the field size extracted ( @xmath67 rings for field sizes of @xmath68 ) .
each photon is divided by the vignetting and deadtime corrected exposure time of the skypixel where it has been detected and these ratios are summed up in each ring yielding the ring count rate . from this value the background count rate for the respective ring area is subtracted yielding a source ring count rate .
these individual source ring count rates are integrated with increasing radius yielding the ( cumulative ) source count rate for a given radius ( fig .
[ a2029 ] ) .
obvious contaminating point sources have been excluded manually .
the cut - out regions have then been assigned the average surface brightness of the ring .
if a cluster has been found to be clearly made up of two components , for instance a3395n / s , these components have been treated separately .
this procedure ensures that double clusters are not treated as a single entity for which spherical symmetry is assumed .
for the same reason strong substructure has been excluded in the same manner as contaminating point sources . in this work
the aim is to characterize all cluster properties consistently and homogeneously .
therefore if strong substructure is identified then it is excluded for the flux / luminosity _ and _ mass determination
. an outer significance radius of the cluster , @xmath69 , is determined at the position from where on the poissonian 1-@xmath70 error rises faster than the source count rate .
usually the source count rate settles into a nearly horizontal line for radii larger than @xmath69 .
we have found , however , that in some cases the source count rate seems to increase or decrease roughly quadratically for radii larger than @xmath69 indicating a possibly under- or overestimated background ( fig .
[ exo ] ) .
we therefore fitted a parabola of the form @xmath71 to the source count rate for radii larger than @xmath69 and corrected the measured background .
an example for a corrected source count rate profile is shown in fig .
[ exoc ] .
figure [ counts ] shows for the extended sample ( 106 clusters ) that the difference between measured and corrected source count rate is generally very small .
nevertheless an inspection of each count rate profile has been performed , to decide whether the measured or corrected count rate is adopted as the final count rate , to avoid artificial corrections due to large scale variations of the background ( especially in the large rassfields ) .
the count rates are given in tab .
[ tbl : data1 ] .
the conversion factor for the count rate to flux conversion depends on the hydrogen column density , @xmath72 , on the cluster gas temperature , @xmath73 , on the cluster gas metallicity , on the cluster redshift , @xmath74 , and on the respective detector responses for the two different ps .
the @xmath72 value is taken as the value inferred from 21 cm radio measurements for our galaxy at the projected cluster position ( @xcite ; included in the exsas software package , @xcite ; photoelectric absorption cross sections are taken from @xcite ) .
gas temperatures have been estimated by compiling x - ray temperatures , @xmath39 , from the literature , giving preference to temperatures measured with the asca satellite . for clusters where no asca measured temperature has been available ,
@xmath39 measured with previous x - ray satellites have been used .
the x - ray temperatures and corresponding references are given in tab .
[ tbl : data2 ] . for two clusters included in hiflugcs
no measured temperature has been found in the literature and the @xmath75 relation of @xcite has been used .
the relation for non cooling flow corrected luminosities and cooling flow corrected / emission weighted temperatures has been chosen . since the conversion from count rate to flux depends only weakly on @xmath73 in the rosat energy band for the relevant temperature range a cluster temperature @xmath76 has been assumed in a first step to determine @xmath77 for the clusters where no gas temperature has been found in the literature . with this luminosity the gas temperature has been estimated .
the metallicity is set to 0.35 times the solar value for all clusters ( e.g. , * ? ? ?
the redshifts have been compiled from the literature and are given in tab .
[ tbl : data1 ] together with the corresponding references . with these quantities and the count rates given in tab .
[ tbl : data1 ] fluxes in the observer rest frame energy range @xmath78 have been calculated applying a modern version of a raymond - smith spectral code @xcite .
the results are listed in tab .
[ tbl : data1 ] .
the flux calculation has also been checked using xspec @xcite by folding the model spectrum created with the parameters given above with the detector response and adjusting the normalization to reproduce the observed count rate .
it is found that for 90% of the clusters the deviation between the two results for the flux measurement is less than 1% .
luminosities in the source rest frame energy range @xmath78 have then been calculated within xspec by adjusting the normalization to reproduce the initial flux measurements . the improvements of the flux determination performed here compared to the input catalogs in general are now summarized .
1 ) due to the use of a high fraction of pointed observations the photon statistics is on average much better , e.g. , for the 33 clusters contained in reflex and hiflugcs one finds a mean of 841 and 19580 source photons , respectively . consequently the cluster emission has been traced out to larger radii for hiflugcs .
2 ) the higher photon statistics has allowed a proper exclusion of contaminating point sources ( stars , agn , etc . ) and substructure , and the separation of double clusters .
3 ) an iterative background correction has been performed .
4 ) a measured x - ray temperature has been used for the flux calculation in most cases .
simulations have shown that even for the hiflugcs clusters with the lowest number of photons the determined flux shows no significant trend with redshift in the relevant redshift range @xcite .
lrrrrrrrrrrcc a0085 & 10.4632 & @xmath79 & 0.0556 & 3.58 & 3.488 & 0.6 & 2.13 & 7.429 & 9.789 & 24.448 & p & 2 + a0119 & 14.0649 & @xmath80 & 0.0440 & 3.10 & 1.931 & 0.9 & 2.68 & 4.054 & 3.354 & 7.475 & p & 1 + a0133 & 15.6736 & @xmath81 & 0.0569 & 1.60 & 1.058 & 0.8 & 1.52 & 2.121 & 2.944 & 5.389 & p & 3 + ngc507 & 20.9106 & @xmath82 & 0.0165 & 5.25 & 1.093 & 1.3 & 0.88 & 2.112 & 0.247 & 0.326
& p & 5 + a0262 & 28.1953 & @xmath83 & 0.0161 & 5.52 & 4.366 & 3.8 & 1.48 & 9.348 & 1.040 & 1.533 & r & 6 + a0400 & 44.4152 & @xmath84 & 0.0240 & 9.38 & 1.146 & 1.1 & 1.85 & 2.778 & 0.686 & 1.033 & p & 8 + a0399 & 44.4684 & @xmath85 & 0.0715 & 10.58 & 1.306 & 5.4 & 3.18 & 3.249 & 7.070 & 17.803 & r & 6 + a0401 & 44.7384 & @xmath86 & 0.0748 & 10.19 & 2.104 & 1.1 & 3.81 & 5.281 & 12.553 & 34.073 & p & 6 + a3112 & 49.4912 & @xmath87 & 0.0750 & 2.53 & 1.502 & 1.1 & 2.18 & 3.103 & 7.456 & 16.128 & p & 1 + fornax & 54.6686 & @xmath88 & 0.0046 & 1.45 & 5.324 & 5.6 & 0.53 & 9.020 & 0.082 & 0.107 & p+r & 4 + 2a0335 & 54.6690 & @xmath89 & 0.0349 & 18.64 & 3.028 & 0.8 & 1.54 & 9.162 & 4.789 & 7.918 &
p & 10 + iiizw54 & 55.3225 & @xmath90 & 0.0311 & 16.68 & 0.708 & 7.7 & 1.27 & 2.001 & 0.831 & 1.226 & r & 11 + a3158 & 55.7282 & @xmath91 & 0.0590 & 1.06 & 1.909 & 1.5 & 1.94 & 3.794 & 5.638 & 12.779 & p & 1 + a0478 & 63.3554 & @xmath92 & 0.0900 & 15.27 & 1.827 & 0.6 & 3.12 & 5.151 & 17.690 & 49.335 & p & 6 + ngc1550 & 64.9066 & @xmath93 & 0.0123 & 11.59 & 1.979 & 5.4 & 0.71 & 4.632 & 0.302 & 0.407 & r & 13 + exo0422 & 66.4637 & @xmath94 & 0.0390 & 6.40 & 1.390 & 6.2 & 1.32 & 3.085 & 2.015 & 3.283
& r & 10 + a3266 & 67.8410 & @xmath95 & 0.0594 & 1.48 & 2.879 & 0.7 & 2.99 & 5.807 & 8.718 & 23.663 & p & 4 + a0496 & 68.4091 & @xmath96 & 0.0328 & 5.68 & 3.724 & 0.7 & 1.78 & 8.326 & 3.837 & 7.306 & p & 8 + a3376 & 90.4835 & @xmath97 & 0.0455 & 5.01 & 1.115 & 1.4 & 2.86 & 2.450 & 2.174 & 4.077 & p & 4 + a3391 & 96.5925 & @xmath98 & 0.0531 & 5.42 & 0.999 & 1.9 & 1.98 & 2.225 & 2.681 & 5.857 & p & 4 + a3395s & 96.6920 & @xmath99 & 0.0498 & 8.49 & 0.836 & 3.8 & 1.45 & 2.009 & 2.131 & 4.471 & p & 4 + a0576 & 110.3571 & @xmath100 & 0.0381 & 5.69 & 1.374 & 6.8 & 2.32 & 3.010 & 1.872 & 3.518 & r & 6 + a0754 & 137.3338 & @xmath101 & 0.0528 & 4.59 & 1.537 & 1.6 & 1.91 & 3.366 & 3.990 & 11.967
& p & 6 + hydra - a & 139.5239 & @xmath102 & 0.0538 & 4.86 & 2.179 & 0.6 & 1.66 & 4.776 & 5.930 & 11.520 & p & 13 + a1060 & 159.1784 & @xmath103 & 0.0114 & 4.92 & 4.653 & 3.3 & 0.95 & 9.951 & 0.554 & 0.945 & r & 6 + a1367 & 176.1903 & @xmath104 & 0.0216 & 2.55 & 2.947 & 0.8 & 1.55 & 6.051 & 1.206 & 2.140 & p & 8 + mkw4 & 181.1124 & @xmath105 & 0.0200 & 1.86 & 1.173 & 1.7 & 1.23 & 2.268 & 0.390 & 0.543 & p & 10 + zwcl1215 & 184.4220 & @xmath106 & 0.0750 & 1.64 & 1.081 & 1.3 & 2.55 & 2.183 & 5.240 & 11.656 & p & 19 + ngc4636 & 190.7084 & @xmath107 & 0.0037 & 1.75 & 3.102 & 7.2 & 0.39 & 4.085 & 0.023 & 0.027 & r & 13 + a3526 & 192.1995 & @xmath108 & 0.0103 & 8.25 & 11.655 & 2.2 & 1.64 & 27.189 & 1.241 & 2.238 & r & 15 + a1644 & 194.2900 & @xmath109 & 0.0474 & 5.33 & 1.853 & 5.1 & 1.85 & 4.030 & 3.876 & 7.882 & r & 8 + a1650 & 194.6712 & @xmath110 & 0.0845 & 1.54 & 1.218 & 6.6 & 3.17 & 2.405 & 7.308 & 17.955 & r & 6 + a1651 & 194.8419 & @xmath111 & 0.0860 & 1.71 & 1.254 & 1.2 & 2.03 & 2.539 & 8.000 & 18.692 & p & 22 + coma & 194.9468 & @xmath112 & 0.0232 & 0.89 & 17.721 & 1.4 & 4.04 & 34.438 & 7.917 & 22.048 & r & 8 + ngc5044 & 198.8530 & @xmath113 & 0.0090 & 4.91 & 3.163 & 0.5 & 0.56 & 5.514 & 0.193 & 0.246 & p & 24 + a1736 & 201.7238 & @xmath114 & 0.0461 & 5.36 & 1.631 & 6.3 & 2.47 & 3.537 & 3.223 & 5.682 & r & 25 + a3558 & 201.9921 & @xmath115 & 0.0480 & 3.63 & 3.158 & 0.5 & 2.11 & 6.720 & 6.615 & 14.600 & p & 1 + a3562 & 203.3984 & @xmath116 & 0.0499 & 3.91 & 1.367 & 0.9 & 2.01 & 2.928 & 3.117 & 6.647 & p & 4 + a3571 & 206.8692 & @xmath117 & 0.0397 & 3.93 & 5.626 & 0.7 & 2.35 & 12.089 & 8.132 & 20.310 & p & 21 + a1795 & 207.2201 & @xmath118 & 0.0616 & 1.20 & 3.132 & 0.3 & 2.14 & 6.270 & 10.124 & 27.106 & p & 6 + a3581 & 211.8852 & @xmath119 & 0.0214 & 4.26 & 1.603 & 3.2 & 0.64 & 3.337 & 0.657 & 0.926 & p & 28 + mkw8 & 220.1596 & @xmath120 & 0.0270 & 2.60 & 1.255 & 8.4 & 1.90 & 2.525 & 0.789 & 1.355 & r & 29 + a2029 & 227.7331 & @xmath121 & 0.0767 & 3.07 & 3.294 & 0.6 & 2.78 & 6.938 & 17.313 & 50.583 & p & 6 + a2052 & 229.1846 & @xmath122 & 0.0348 & 2.90 & 2.279 & 1.0 & 1.14 & 4.713 & 2.449 & 4.061 & p & 6 + mkw3s & 230.4643 & @xmath123 & 0.0450 & 3.15 & 1.578 & 1.0 & 1.39 & 3.299 & 2.865 & 5.180 & p & 10 + a2065 & 230.6096 & @xmath124 & 0.0721 & 2.84 & 1.227 & 6.1 & 3.09 & 2.505 & 5.560 & 12.271 & r & 6 + a2063 & 230.7734 & @xmath125 & 0.0354 & 2.92 & 2.038 & 1.3 & 2.13 & 4.232 & 2.272 & 4.099 & p & 8 + a2142 & 239.5824 & @xmath126 & 0.0899 & 4.05 & 2.888 & 0.9 & 3.09 & 6.241 & 21.345 & 64.760 & p & 6 + a2147 & 240.5628 & @xmath127 & 0.0351 & 3.29 & 2.623 & 3.2 & 1.87 & 5.522 & 2.919 & 6.067 & p & 8 + a2163 & 243.9433 & @xmath128 & 0.2010 & 12.27 & 0.773 & 1.5 & 3.15 & 2.039 & 34.128 & 123.200 & p & 31 + a2199 & 247.1586 & @xmath129 & 0.0302 & 0.84 & 5.535 & 1.8 & 2.37 & 10.642 & 4.165 & 7.904 & r & 8 + a2204 & 248.1962 & @xmath130 & 0.1523 & 5.94 & 1.211 & 1.6 & 3.29 & 2.750 & 26.938 & 68.989 & p & 6 + a2244 & 255.6749 & @xmath131 & 0.0970 & 2.07 & 1.034 & 2.1 & 2.64 & 2.122 & 8.468 & 21.498 & p & 6 + a2256 & 255.9884 & @xmath132 & 0.0601 & 4.02 & 2.811 & 1.4 & 3.09 & 6.054 & 9.322 & 22.713 & p & 6 + a2255 & 258.1916 & @xmath133 & 0.0800 & 2.51 & 0.976 & 1.2 & 3.22 & 2.022 & 5.506 & 13.718 & p & 6 + a3667 & 303.1362 & @xmath134 & 0.0560 & 4.59 & 3.293 & 0.7 & 2.81 & 7.201 & 9.624 & 24.233 & p & 1 + s1101 & 348.4941 & @xmath135 & 0.0580 & 1.85 & 1.237 & 0.9 & 1.64 & 2.485 & 3.597 & 5.939 & p & 35 + a2589 & 350.9868 & @xmath136 & 0.0416 & 4.39 & 1.200 & 1.3 & 1.46 & 2.591 & 1.924 & 3.479 & p & 37 + a2597 & 351.3318 & @xmath137 & 0.0852 & 2.50 & 1.074 & 1.2 & 1.43 & 2.213 & 6.882 & 13.526 & p & 6 + a2634 & 354.6201 & @xmath138 & 0.0312 & 5.17 & 1.096 & 1.6 & 1.79 & 2.415 & 1.008 & 1.822 & p & 6 + a2657 & 356.2334 & @xmath139 & 0.0404 & 5.27 & 1.148 & 0.9 & 1.52 & 2.535 & 1.771 & 3.202 & p & 8 + a4038 & 356.9322 & @xmath140 & 0.0283 & 1.55 & 2.854 & 1.3 & 1.35 & 5.694 & 1.956 & 3.295 & p & 4 + a4059 & 359.2541 & @xmath141 & 0.0460 & 1.10 & 1.599 & 1.3 & 1.72 & 3.170 & 2.872 & 5.645 & p & 36 + a2734 & 2.8389 & @xmath142 & 0.0620 & 1.84 & 0.710 & 2.5 & 1.74 & 1.434 & 2.365 & 4.357 & p & 1 + a2877 & 17.4796 & @xmath143 & 0.0241 & 2.10 & 0.801 & 1.2 & 1.06 & 1.626 & 0.405 & 0.714 & p & 4 + ngc499 & 20.7971 & @xmath144 & 0.0147 & 5.25 & 0.313 & 2.5 & 0.30 & 0.479 & 0.045 & 0.051 & p & 5 + awm7 & 43.6229 & @xmath145 & 0.0172 & 9.21 & 7.007 & 2.0 & 1.58 & 16.751 & 2.133 & 3.882 & r & 7 + perseus & 49.9455 & @xmath146 & 0.0183 & 15.69 & 40.723 & 0.8 & 3.30 & 113.731 & 16.286 & 40.310 & r & 9 + s405 & 58.0078 & @xmath147 & 0.0613 & 7.65 & 0.781 & 8.2 & 2.14 & 1.800 & 2.899 & 5.574 & r & 12 + 3c129 & 72.5602 & @xmath148 & 0.0223 & 67.89 & 1.512 & 5.6 & 1.61 & 10.566 & 2.242 & 4.996 & r & 10 + a0539 & 79.1560 & @xmath149 & 0.0288 & 12.06 & 1.221 & 1.3 & 1.37 & 3.182 & 1.135 & 1.935 &
p & 14 + s540 & 85.0265 & @xmath150 & 0.0358 & 3.53 & 0.788 & 5.0 & 0.84 & 1.611 & 0.887 & 1.353 & r & 4 + a0548w & 86.3785 & @xmath151 & 0.0424 & 1.79 & 0.136 & 5.4 & 0.73 & 0.234 & 0.183 & 0.240 & p & 15 + a0548e & 87.1596 & @xmath152 & 0.0410 & 1.88 & 0.771 & 1.8 & 2.12 & 1.551 & 1.117 & 1.870 & p & 15 + a3395n & 96.9005 & @xmath153 & 0.0498 & 5.42 & 0.699 & 3.9 & 1.37 & 1.555 & 1.650 & 3.461 & p & 4 + ugc03957 & 115.2481 & @xmath154 & 0.0340 & 4.59 & 0.936 & 6.0 & 0.94 & 1.975 & 0.980 & 1.531 & r & 16 + pks0745 & 116.8837 & @xmath155 & 0.1028 & 43.49 & 1.268 & 1.0 & 2.44 & 6.155 & 27.565 & 70.604 & p & 17 + a0644 & 124.3553 & @xmath156 & 0.0704 & 5.14 & 1.799 & 1.0 & 4.02 & 3.994 & 8.414 & 22.684 & p & 6 + s636 & 157.5151 & @xmath157 & 0.0116 & 6.42 & 3.102 & 4.9 & 1.18 & 5.869 & 0.341 & 0.446 & r & 18 + a1413 & 178.8271 & @xmath158 & 0.1427 & 1.62 & 0.636 & 1.6 & 2.39 & 1.289 & 11.090 & 28.655 & p & 6 + m49 & 187.4437 & @xmath159 & 0.0044 & 1.59 & 1.259 & 1.0 & 0.27 & 1.851 & 0.015 & 0.019 & p & 15 + a3528n & 193.5906 & @xmath160 & 0.0540 & 6.10 & 0.560 & 2.3 & 1.51 & 1.263 & 1.581 & 2.752 & p & 1 + a3528s & 193.6708 & @xmath161 & 0.0551 & 6.10 & 0.756 & 1.6 & 1.35 & 1.703 & 2.224 & 3.746 & p & 20 + a3530 & 193.9211 & @xmath162 & 0.0544 & 6.00 & 0.438 & 2.8 & 1.55 & 0.987 & 1.252 & 2.317 & p & 21 + a3532 & 194.3375 & @xmath163 & 0.0539 & 5.96 & 0.797 & 1.8 & 1.64 & 1.797 & 2.235 & 4.483 & p & 21 + a1689 & 197.8726 & @xmath164 & 0.1840 & 1.80 & 0.712 & 1.1 & 2.36 & 1.454 & 20.605 & 60.707 & p & 23 + a3560 & 203.1119 & @xmath165 & 0.0495 & 3.92 & 0.714 & 2.5 & 2.00 & 1.519 & 1.601 & 2.701 & p & 26 + a1775 & 205.4582 & @xmath166 & 0.0757 & 1.00 & 0.654 & 1.8 & 2.02 & 1.290 & 3.175 & 5.735 & p & 27 + a1800 & 207.3408 & @xmath167 & 0.0748 & 1.18 & 0.610 & 7.9 & 1.98 & 1.183 & 2.840 & 5.337 & r & 28 + a1914 & 216.5035 & @xmath168 & 0.1712 & 0.97 & 0.729 & 1.4 & 2.35 & 1.454 & 17.813 & 56.533 & p & 6 + ngc5813 & 225.2994 & @xmath169 & 0.0064 & 4.19 & 0.976 & 6.7 & 0.17 & 1.447 & 0.025 & 0.029 & r & 13 + ngc5846 & 226.6253 & @xmath170 & 0.0061 & 4.25 & 0.569 & 2.3 & 0.21 & 0.851 & 0.014 & 0.016 & p & 13 + a2151w & 241.1465 & @xmath171 & 0.0369 & 3.36 & 0.754 & 1.9 & 1.46 & 1.568 & 0.917 & 1.397 & p & 8 + a3627 & 243.5546 & @xmath172 & 0.0163 & 20.83 & 9.962 & 3.0 & 2.20 & 31.084 & 3.524 & 8.179 & r & 30 + triangul & 249.5758 & @xmath173 & 0.0510 & 12.29 & 4.294 & 0.7 & 2.54 & 11.308 & 12.508 & 37.739 & p & 32 + ophiuchu & 258.1115 & @xmath174 & 0.0280 & 20.14 & 11.642 & 2.0 & 2.29 & 35.749 & 11.953 & 37.391 & r & 33 + zwcl1742 & 266.0623 & @xmath175 & 0.0757 & 3.56 & 0.889 & 4.4 & 1.83 & 1.850 & 4.529 & 9.727 & r & 34 + a2319 & 290.2980 & @xmath176 & 0.0564 & 8.77 & 5.029 & 1.0 & 3.57 & 12.202 & 16.508 & 47.286 & p & 6 + a3695 & 308.6991 & @xmath177 & 0.0890 & 3.56 & 0.836 & 9.2 & 2.58 & 1.739 & 5.882 & 12.715 & r & 1 + iizw108 & 318.4752 & @xmath178 & 0.0494 & 6.63 & 0.841 & 7.3 & 2.20 & 1.884 & 1.969 & 3.445 & r & 5 + a3822 & 328.5438 & @xmath179 & 0.0760 & 2.12 & 0.964 & 7.3 & 3.18 & 1.926 & 4.758 & 9.877 & r & 1 + a3827 & 330.4869 & @xmath180 & 0.0980 & 2.84 & 0.953 & 5.8 & 1.78 & 1.955 & 7.963 & 20.188 & r & 1 + a3888 & 338.6255 & @xmath181 & 0.1510 & 1.20 & 0.546 & 2.4 & 1.52 & 1.096 & 10.512 & 30.183 & p & 23 + a3921 & 342.5019 & @xmath182 & 0.0936 & 2.80 & 0.626 & 1.7 & 2.43 & 1.308 & 4.882 & 11.023 & p & 12 + hcg94 & 349.3041 & @xmath183 & 0.0417 & 4.55 & 0.820 & 1.0 & 2.09 & 1.775 & 1.324 & 2.319 & p & 36 + rxj2344 & 356.0723 & @xmath184 & 0.0786 & 3.54 & 0.653 & 1.4 & 1.61 & 1.385 & 3.661 & 7.465 & p & 12 + a parametric description of the cluster gas density profile has been derived using the standard @xmath185 model ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
assuming spherical symmetry the model @xmath186 is fitted to the measured surface brightness profile ( ring count rates per ring area ) , where @xmath187 denotes the projected distance from the cluster center .
this yields values for the core radius , @xmath188 , the @xmath185 parameter , and the normalization , @xmath189 ( and also a fitted value for the background surface brightness , @xmath190 , since the fit is performed on the non background subtracted data ) .
the fit values have been used to construct the radial gas density distribution @xmath191 the gravitational cluster mass , @xmath192 , has been determined assuming the intracluster gas to be in hydrostatic equilibrium and isothermal . using ( [ eq : rog ] ) and
the ideal gas equation under these assumptions leads to @xmath193 where @xmath194 represents the mean molecular weight , @xmath195 the proton mass , and @xmath196 the gravitational constant .
combined @xmath8-body / hydrodynamic cluster simulations have shown that this method generally gives unbiased results with an uncertainty of 1429% ( e.g. , * ? ? ?
* ; * ? ? ?
currently there are contradictory measurements concerning the general presence of gas temperature gradients in clusters ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
if there is a systematic trend in the sense that the gas temperature decreases with increasing radius in the outer cluster parts similar to that found by @xcite then the isothermal assumption leads to an overestimation of the cluster mass of about 30% at about 6 core radii @xcite .
@xcite determined masses by employing the assumption of isothermality and also using measured cluster gas temperature profiles .
a comparison for 38 clusters included in their sample indicates that the latter masses are on average a factor of 0.80 smaller than the isothermal masses within @xmath197 ( this radius is defined in the next paragraph ) . until the final answer on this issue is given by xmm - newton ( e.g. , @xcite ; m. arnaud 2001 , private communication ) .
] , we retain the isothermal assumption .
the influence of a possible overestimation of the cluster mass on the determination of cosmological parameters is investigated in sect .
[ func_pred ] . having determined the integrated mass as a function of radius , a physically meaningful fiducial radius for the mass measurement has to be defined .
the radii commonly used are either the abell radius , @xmath197 , or @xmath198 .
the abell radius is fixed at @xmath199 .
the radius @xmath197 ( @xmath198 ) is the radius within which the mean gravitational mass density @xmath200 .
the critical cosmic matter density is defined as @xmath201 , where @xmath202 and @xmath203^{1/2}$ ] .
it has been shown that a correction for redshift is not necessary for the nearby clusters included in hiflugcs@xcite and we use the zero redshift value for all calculations , i.e. @xmath204 , unless noted otherwise .
nevertheless the influence of this approximation is tested in sect .
[ func_pred ] for the model ( @xmath205 ) , where evolution is strong . in order to treat clusters of different size in a homogeneous way we determine the cluster mass at a characteristic density but also give the mass determined formally at a fixed radius for comparison .
spherical collapse models predict a cluster virial density @xmath206 for ( @xmath205 ) , so a pragmatic approximation to the virial mass is to use @xmath197 as the outer boundary .
simulations performed by @xcite have shown , however , that isothermal x - ray mass measurements may be biased towards high masses for @xmath207 . furthermore for most of the clusters in hiflugcs ( 86% ) up to @xmath198 no extrapolation outside
the significantly detected cluster emission is necessary , i.e. @xmath208 , whereas the fraction is lower for @xmath197 ( 25% ) and @xmath209 ( 17% ) . in summary
the most accurate results are expected for @xmath210 , but for a comparison to predicted mass functions @xmath211 is the more appropriate value ( sect .
[ func_pred ] ) .
results for all determined masses and their corresponding radii are given in tab .
[ tbl : data2 ] .
masses for the cluster gas will be given in a subsequent paper .
a major source of uncertainty comes from the temperature measurements .
however , this ( statistical ) error is less than 5% for one third of the clusters , therefore also other sources of error have to be taken into account , in particular one can not neglect the uncertainties of the fit parameter values when assessing the statistical errors of the mass measurements . therefore mass errors have been calculated by varying the fit parameter values , @xmath185 and @xmath188 , along their 68% confidence level error ellipse and using the upper and lower bound of the quoted temperature ranges .
the statistical mass error range has then been defined between the maximum and minimum mass .
note that a simple error propagation applied to ( [ eq : mt ] ) would underestimate the uncertainty of @xmath212 and @xmath213 , since @xmath197 and @xmath198 also depend on @xmath73 , @xmath185 , and ( weakly ) @xmath188 .
the individual mass errors have been used in subsequent calculations , unless noted otherwise . , where @xmath214 and @xmath215 denote the upper and lower boundary of the quantity s error range , respectively . ] a mean statistical error of 23% for clusters included in hiflugcs and a mean error of 27% for the extended sample has been found .
lccccccccc a0085 & @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & 12.21 & 1 + a0119 & @xmath223 & @xmath224 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & 12.24 & 1 + a0133 & @xmath230 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & 6.71 & 9 + ngc507 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 & 1.86 & 2 + a0262 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & 3.17 & 2 + a0400 & @xmath251 & @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & @xmath257 & 4.10 & 2 + a0399 & @xmath258 & @xmath259 & @xmath260 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & 16.24 & 1 + a0401 & @xmath265 & @xmath266 & @xmath267 & @xmath268 & @xmath269 & @xmath270 & @xmath271 & 16.21 & 1 + a3112 & @xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 & @xmath278 & 10.16 & 1 + fornax & @xmath279 & @xmath280 & @xmath281 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & 3.20 & 2 + 2a0335 & @xmath286 & @xmath287 & @xmath288 & @xmath289 & @xmath290 & @xmath291 & @xmath292 & 5.76 & 2 + iiizw54 & @xmath293 & @xmath294 & ( @xmath295 ) & @xmath296 & @xmath297 & @xmath298 & @xmath299 & 6.32 & 11 + a3158 & @xmath300 & @xmath301 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath306 & 12.61 & 3 + a0478 & @xmath307 & @xmath308 & @xmath309 & @xmath310 & @xmath311 & @xmath312 & @xmath313 & 17.12 & 1 + ngc1550 & @xmath314 & @xmath315 & @xmath316 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & 2.64 & 5 + exo0422 & @xmath321 & @xmath322 & @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & 6.96 & 9 + a3266 & @xmath328 & @xmath329 & @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & 20.47 & 1 + a0496 & @xmath335 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & 6.66 & 2 + a3376 & @xmath342 & @xmath343 & @xmath344 & @xmath345 & @xmath346 & @xmath347 & @xmath348 & 13.20 & 1 + a3391 & @xmath349 & @xmath350 & @xmath351 & @xmath352 & @xmath353 & @xmath354 & @xmath355 & 10.35 & 1 + a3395s & @xmath356 & @xmath357 & @xmath358 & @xmath359 & @xmath360 & @xmath361 & @xmath362 & 15.42 & 1 + a0576 & @xmath363 & @xmath364 & @xmath365 & @xmath366 & @xmath367 & @xmath368 & @xmath369 & 10.86 & 3 + a0754 & @xmath370 & @xmath371 & @xmath372 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & 21.94 & 1 + hydra - a & @xmath377 & @xmath378 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath383 & 8.21 & 1 + a1060 & @xmath384 & @xmath385 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath390 & 6.54 & 2 + a1367 & @xmath391 & @xmath392 & @xmath393 & @xmath394 & @xmath395 & @xmath396 & @xmath397 & 8.08 & 2 + mkw4 & @xmath398 & @xmath399 & @xmath400 & @xmath401 & @xmath402 & @xmath403 & @xmath404 & 2.51 & 2 + zwcl1215 & @xmath405 & @xmath406 & ( @xmath407 ) & @xmath408 & @xmath409 & @xmath410 & @xmath411 & 14.91 & 11 + ngc4636 & @xmath412 & @xmath413 & @xmath414 & @xmath415 & @xmath416 & @xmath417 & @xmath418 & 1.24 & 4 + a3526 & @xmath419 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & @xmath424 & @xmath425 & 6.07 & 2 + a1644 & @xmath426 & @xmath427 & @xmath428 & @xmath429 & @xmath430 & @xmath431 & @xmath432 & 8.98 & 10 + a1650 & @xmath433 & @xmath434 & @xmath435 & @xmath436 & @xmath437 & @xmath438 & @xmath439 & 15.56 & 1 + a1651 & @xmath440 & @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath445 & @xmath446 & 13.01 & 1 + coma & @xmath447 & @xmath448 & @xmath449 & @xmath450 & @xmath451 & @xmath452 & @xmath453 & 18.01 & 2 + ngc5044 & @xmath454 & @xmath399 & @xmath455 & @xmath456 & @xmath457 & @xmath458 & @xmath459 & 1.87 & 2 + a1736 & @xmath460 & @xmath461 & @xmath462 & @xmath463 & @xmath464 & @xmath465 & @xmath466 & 6.22 & 1 + a3558 & @xmath467 & @xmath468 & @xmath469 & @xmath470 & @xmath471 & @xmath472 & @xmath473 & 10.56 & 1 + a3562 & @xmath474 & @xmath475 & @xmath476 & @xmath477 & @xmath478 & @xmath479 & @xmath480 & 8.10 & 3 + a3571 & @xmath265 & @xmath481 & @xmath482 & @xmath483 & @xmath484 & @xmath485 & @xmath486 & 14.04 & 1 + a1795 & @xmath487 & @xmath488 & @xmath489 & @xmath490 & @xmath491 & @xmath492 & @xmath493 & 15.46 & 1 + a3581 & @xmath494 & @xmath495 & @xmath496 & @xmath497 & @xmath498 & @xmath499 & @xmath500 & 3.30 & 5 + mkw8 & @xmath501 & @xmath502 & @xmath503 & @xmath504 & @xmath505 & @xmath506 & @xmath507 & 5.60 & 5 + a2029 & @xmath508 & @xmath509 & @xmath510 & @xmath511 & @xmath512 & @xmath513 & @xmath514 & 17.62 & 1 + a2052 & @xmath515 & @xmath516 & @xmath517 & @xmath518 & @xmath519 & @xmath520 & @xmath521 & 5.30 & 3 + mkw3s & @xmath522 & @xmath523 & @xmath524 & @xmath525 & @xmath526 & @xmath527 & @xmath528 & 7.16 & 1 + a2065 & @xmath529 & @xmath530 & @xmath469 & @xmath531 & @xmath532 & @xmath533 & @xmath534 & 20.21 & 1 + a2063 & @xmath535 & @xmath536 & @xmath537 & @xmath538 & @xmath539 & @xmath540 & @xmath541 & 6.86 & 2 + a2142 & @xmath542 & @xmath543 & @xmath544 & @xmath545 & @xmath546 & @xmath547 & @xmath548 & 19.05 & 1 + a2147 & @xmath549 & @xmath550 & @xmath551 & @xmath552 & @xmath553 & @xmath554 & @xmath555 & 7.21 & 2 + a2163 & @xmath556 & @xmath557 & @xmath558 & @xmath559 & @xmath560 & @xmath561 & @xmath562 & 34.18 & 3 + a2199 & @xmath563 & @xmath564 & @xmath565 & @xmath566 & @xmath567 & @xmath568 & @xmath569 & 8.92 & 2 + a2204 & @xmath570 & @xmath571 & @xmath572 & @xmath573 & @xmath574 & @xmath575 & @xmath576 & 14.34 & 3 + a2244 & @xmath577 & @xmath578 & @xmath579 & @xmath580 & @xmath581 & @xmath582 & @xmath583 & 14.33 & 10 + a2256 & @xmath584 & @xmath585 & @xmath586 & @xmath587 & @xmath588 & @xmath589 & @xmath590 & 19.34 & 1 + a2255 & @xmath591 & @xmath592 & @xmath593 & @xmath594 & @xmath595 & @xmath596 & @xmath597 & 17.54 & 3 + a3667 & @xmath598 & @xmath599 & @xmath600 & @xmath601 & @xmath602 & @xmath603 & @xmath604 & 12.50 & 1 + s1101 & @xmath605 & @xmath606 & @xmath607 & @xmath608 & @xmath609 & @xmath610 & @xmath611 & 6.38 & 9 + a2589 & @xmath612 & @xmath613 & @xmath614 & @xmath615 & @xmath616 & @xmath617 & @xmath618 & 7.33 & 9 + a2597 & @xmath619 & @xmath620 & @xmath621 & @xmath622 & @xmath623 & @xmath624 & @xmath625 & 9.27 & 1 + a2634 & @xmath626 & @xmath627 & @xmath628 & @xmath629 & @xmath630 & @xmath631 & @xmath546 & 7.77 & 2 + a2657 & @xmath632 & @xmath633 & @xmath634 & @xmath635 & @xmath636 & @xmath637 & @xmath638 & 6.84 & 1 + a4038 & @xmath639 & @xmath640 & @xmath641 & @xmath642 & @xmath643 & @xmath644 & @xmath645 & 5.67 & 3 + a4059 & @xmath646 & @xmath647 & @xmath648 & @xmath649 & @xmath650 & @xmath651 & @xmath652 & 8.52 & 1 + a2734 & @xmath653 & @xmath654 & ( @xmath655 ) & @xmath656 & @xmath657 & @xmath658 & @xmath659 & 7.97 & 11 + a2877 & @xmath660 & @xmath661 & @xmath662 & @xmath663 & @xmath664 & @xmath665 & @xmath666 & 6.57 & 10 + ngc499 & @xmath667 & @xmath668 & @xmath669 & @xmath670 & @xmath671 & @xmath672 & @xmath673 & 1.73 & 4 + awm7 & @xmath674 & @xmath675 & @xmath676 & @xmath677 & @xmath500 & @xmath678 & @xmath679 & 8.35 & 2 + perseus & @xmath680 & @xmath681 & @xmath682 & @xmath683 & @xmath684 & @xmath685 & @xmath686 & 12.20 & 2 + s405 & @xmath687 & @xmath688 & ( @xmath689 ) & @xmath690 & @xmath691 & @xmath692 & @xmath693 & 9.09 & 11 + 3c129 & @xmath694 & @xmath695 & @xmath696 & @xmath697 & @xmath698 & @xmath699 & @xmath700 & 11.08 & 9 + a0539 & @xmath701 & @xmath702 & @xmath703 & @xmath704 & @xmath705 & @xmath706 & @xmath707 & 6.04 & 2 + s540 & @xmath708 & @xmath709 & ( @xmath710 ) & @xmath711 & @xmath712 & @xmath713 & @xmath714 & 5.13 & 11 + a0548w & @xmath715 & @xmath716 & ( @xmath717 ) & @xmath718 & @xmath719 & @xmath720 & @xmath721 & 2.64 & 11 + a0548e & @xmath722 & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727 & @xmath728 & 4.95 & 3 + a3395n & @xmath729 & @xmath730 & @xmath358 & @xmath731 & @xmath732 & @xmath733 & @xmath734 & 15.55 & 1 + ugc03957 & @xmath735 & @xmath736 & ( @xmath737 ) & @xmath738 & @xmath739 & @xmath740 & @xmath741 & 6.35 & 11 + pks0745 & @xmath742 & @xmath743 & @xmath744 & @xmath745 & @xmath746 & @xmath747 & @xmath748 & 14.58 & 3 + a0644 & @xmath749 & @xmath750 & @xmath751 & @xmath752 & @xmath753 & @xmath754 & @xmath755 & 18.33 & 1 + s636 & @xmath756 & @xmath757 & ( @xmath758 ) & @xmath759 & @xmath760 & @xmath761 & @xmath762 & 2.93 & 11 + a1413 & @xmath763 & @xmath764 & @xmath765 & @xmath766 & @xmath767 & @xmath768 & @xmath769 & 16.03 & 3 + m49 & @xmath770 & @xmath399 & @xmath771 & @xmath772 & @xmath457 & @xmath773 & @xmath774 & 1.87 & 4 + a3528n & @xmath775 & @xmath776 & @xmath777 & @xmath778 & @xmath779 & @xmath780 & @xmath781 & 7.00 & 8 + a3528s & @xmath782 & @xmath783 & @xmath784 & @xmath785 & @xmath786 & @xmath787 & @xmath788 & 4.86 & 8 + a3530 & @xmath789 & @xmath790 & @xmath791 & @xmath792 & @xmath793 & @xmath794 & @xmath795 & 9.82 & 7 + a3532 & @xmath796 & @xmath797 & @xmath798 & @xmath799 & @xmath800 & @xmath801 & @xmath802 & 9.88 & 7 + a1689 & @xmath803 & @xmath804 & @xmath805 & @xmath806 & @xmath807 & @xmath808 & @xmath809 & 21.13 & 3 + a3560 & @xmath810 & @xmath811 & ( @xmath812 ) & @xmath813 & @xmath814 & @xmath815 & @xmath816 & 5.92 & 11 + a1775 & @xmath817 & @xmath818 & @xmath819 & @xmath820 & @xmath821 & @xmath822 & @xmath823 & 8.21 & 3 + a1800 & @xmath824 & @xmath825 & ( @xmath826 ) & @xmath827 & @xmath828 & @xmath829 & @xmath830 & 10.08 & 11 + a1914 & @xmath831 & @xmath832 & @xmath833 & @xmath834 & @xmath835 & @xmath836 & @xmath837 & 26.20 & 3 + ngc5813 & @xmath838 & @xmath839 & ( @xmath840 ) & @xmath841 & @xmath842 & @xmath843 & @xmath844 & 1.32 & 11 + ngc5846 & @xmath845 & @xmath846 & @xmath847 & @xmath848 & @xmath849 & @xmath850 & @xmath851 & 1.63 & 4 + a2151w & @xmath852 & @xmath853 & @xmath854 & @xmath855 & @xmath856 & @xmath857 & @xmath858 & 4.51 & 3 + a3627 & @xmath859 & @xmath860 & @xmath861 & @xmath862 & @xmath863 & @xmath864 & @xmath865 & 11.03 & 3 + triangul & @xmath866 & @xmath867 & @xmath868 & @xmath869 & @xmath870 & @xmath871 & @xmath872 & 19.35 & 1 + ophiuchu & @xmath873 & @xmath874 & @xmath875 & @xmath876 & @xmath877 & @xmath878 & @xmath879 & 25.32 & 2 + zwcl1742 & @xmath880 & @xmath881 & ( @xmath882 ) & @xmath883 & @xmath884 & @xmath885 & @xmath886 & 12.42 & 11 + a2319 & @xmath887 & @xmath888 & @xmath889 & @xmath890 & @xmath891 & @xmath892 & @xmath893 & 17.17 & 1 + a3695 & @xmath894 & @xmath895 & ( @xmath896 ) & @xmath897 & @xmath898 & @xmath899 & @xmath900 & 11.12 & 11 + iizw108 & @xmath901 & @xmath902 & ( @xmath903 ) & @xmath904 & @xmath905 & @xmath906 & @xmath907 & 7.47 & 11 + a3822 & @xmath908 & @xmath909 & ( @xmath910 ) & @xmath911 & @xmath912 & @xmath913 & @xmath914 & 10.29 & 11 + a3827 & @xmath915 & @xmath916 & ( @xmath917 ) & @xmath918 & @xmath919 & @xmath920 & @xmath921 & 22.44 & 11 + a3888 & @xmath922 & @xmath923 & ( @xmath924 ) & @xmath925 & @xmath926 & @xmath927 & @xmath928 & 26.85 & 11 + a3921 & @xmath929 & @xmath930 & @xmath931 & @xmath932 & @xmath933 & @xmath934 & @xmath935 & 14.37 & 3 + hcg94 & @xmath936 & @xmath937 & @xmath938 & @xmath939 & @xmath940 & @xmath941 & @xmath942 & 5.90 & 6 + rxj2344 & @xmath943 & @xmath944 & ( @xmath945 ) & @xmath946 & @xmath947 & @xmath948 & @xmath949 & 12.58 & 11 +
in this section it is shown that a tight correlation exists between the gravitational cluster mass and the x - ray luminosity .
this ensures that hiflugcs is essentially selected by cluster mass . in the second part of this section the cluster mass function
is presented , including the proper treatment of the scatter in the @xmath38@xmath192 relation . since the aim is the construction of a mass function from a flux - limited sample it is now important to test for a correlation between x - ray luminosity and gravitational mass . in fig .
[ mtlx ] @xmath38 , given in the rosat energy band , is plotted as a function of @xmath212 , showing clearly the existence of a tight ( linear pearson correlation coefficient @xmath950 0.92 ) correlation , as expected . to quantify the mass
luminosity relation , a linear regression fit in log log space has been performed .
the method used allows for intrinsic scatter and errors in both variables @xcite .
tables [ tab : lm1][tab : lm5 ] in the appendix give the results for different fit methods , where minimization has been performed in vertical , horizontal , and orthogonal direction , and the bisector result is given , which bisects the best fit results of vertical and horizontal minimization .
the fits have been performed using the form @xmath951 we find , as noted in general by previous authors ( e.g. , @xcite ) , that the chosen fitting method has a significant influence on the best fit parameter values.@xmath39 relation , one and the same fitting statistic ought to be used ( e.g. , @xcite ) . ] in this work the appropriate relation for the application under consideration is always indicated . the difference between the fit results for 63 and 106 clusters may indicate a scale dependence of the @xmath38@xmath192 relation , since the difference is slightly larger than the uncertainty evaluated with the bootstrap method .
the small number of low luminosity clusters in hiflugcs compared to the extended sample may be responsible for the less steep relation obtained using the hiflugcs clusters only .
note that only two out of the six clusters with @xmath952 are included in hiflugcs . to reliably detect any deviations from the power law shape of the @xmath38@xmath192 relation , however , more clusters with @xmath953 ( and possibly @xmath954 ) need to be sampled .
such work is in progress .
as will be seen later , in the procedure used here for the comparison of observed and predicted mass functions the precise shape of the @xmath38@xmath192 relation is not important .
when constructing the mass function the overall ( measurement plus intrinsic ) scatter in the @xmath38@xmath192 relation may become important ( sect .
[ funct ] ) . after verifying that the scatter is approximately gaussian in log space
the scatter has been measured as given in tab .
[ tab : scat ] in the appendix .
the scatter in @xmath955 , @xmath956 , and orthogonal to the best fit line is given by @xmath957 , @xmath958 , and @xmath959 , respectively .
the commonly used definition of the galaxy cluster mass function is analogous to the definition of the luminosity function ( e.g. , @xcite ) : the mass function , @xmath960 , denotes the number of clusters , @xmath8 , per unit comoving volume , @xmath961 , per unit mass in the interval @xmath962 $ ] , i.e. @xmath963 . assuming constant density the classical @xmath964 estimator ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) can be used for estimation of luminosity functions , i.e.@xmath965 .
@xmath964 is the maximum comoving volume within which a cluster with given luminosity for a given survey flux limit and sky coverage could have been detected . as mentioned in sect .
[ sample ] the hiflugcs flux limit is constant over 99% of the covered area , which simplifies the calculation of @xmath964 . in the previous section it has been shown that the x - ray luminosity is closely correlated with cluster mass .
therefore the @xmath964 estimator can also be applied to estimate the mass function ; @xmath964 then being a function of mass .
we employ different methods to correct for the scatter present in the @xmath38@xmath192 relation . if @xmath966 is used instead of @xmath967 , where @xmath968 is the luminosity estimated from the @xmath38@xmath192 relation using the determined cluster mass @xmath192 , the scatter is automatically taken into account .
this method has been widely used in the construction of x - ray temperature functions , recently , e.g. , by @xcite . if @xmath969 is used and the utilized @xmath38@xmath192 relation is assumed to be the ` true ' relation then the scatter in this relation has to be taken into account explicitly . therefore following the method employed for the temperature function by @xcite and @xcite the mass function
may also be estimated by determining @xmath970 , where the measured scatter in @xmath971 is included .
specifically we use @xmath972 where @xmath973 and @xmath974 are the best fit parameter values taken from the appropriate @xmath38@xmath192 relation of the form ( [ eq : lmform ] ) and @xmath957 is the corresponding measured standard deviation in @xmath975 given in tab .
[ tab : scat ] .
however , we can also use the measured @xmath38@xmath192 relation directly , i.e. @xmath969 , taking advantage of the fact that in our flux - limited sample there are fewer low luminosity clusters for a given mass than high luminosity ones , which results in a slightly increased normalization of the relation . therefore using this relation directly , the effect of the scatter and the resulting bias towards higher luminosity clusters is already included and thus directly accounted for .
the drawback of using @xmath966 is that a small number of clusters per mass bin possibly does not represent the true scatter well . to minimize this effect we use at least ten clusters per mass bin .
the drawback of using @xmath976 or @xmath969 , as noted , e.g. , by @xcite and @xcite , is the reliance on the validity of the measured relation over the entire mass range .
the first method and the method that accounts for the scatter explicitly ( eq . [ eq : vmax ] ) have been tested by using monte carlo simulations for a precisely known @xmath38@xmath39 relation and scatter and have been shown to give accurate estimates of @xmath977 for a large number of clusters in the study of the hiflugcs temperature function by @xcite . in fig .
[ vm_cp_ns ] hiflugcs mass functions are shown . as expected the method employing @xmath966
prompts a mass function exhibiting a larger scatter , because in this case the scatter is accounted for by the actual scatter of the ten or eleven clusters in each mass bin . for comparison the two extreme mass functions calculated using @xmath969 are shown .
extreme is meant in the sense of using the steepest ( a ) and shallowest ( b ) @xmath38@xmath192 relation for the hiflugcs sample , i.e. , ( @xmath978 ) with @xmath979 and ( @xmath980 ) with @xmath981 ( tab .
[ tab : lm2 ] ) . at the low mass end ( a ) predicts a lower luminosity for a given mass than ( b ) resulting in a smaller @xmath964 and therefore a higher @xmath982 . at the high mass side
the effect is opposite resulting in a lower @xmath982 for ( a ) .
the differences of these mass functions to the mass function calculated using @xmath966 can be understood in a similar way and are caused partly by the indication of a deviation from a power law shape of the @xmath38@xmath192 relation . using @xmath976 results in similar mass functions as shown for the open symbols in fig .
[ vm_cp_ns ] but the points lie systematically lower because the scatter is accounted for twice . for the comparison of the observational mass function to mass functions predicted by certain cosmological models @xmath966 is used because it is independent of the precise shape of the @xmath38@xmath192 relation and also because @xmath38 has a much smaller measurement uncertainty than @xmath192 .
the influence of the choice of the @xmath964 calculation on the estimation of cosmological parameters is investigated in sect .
[ func_pred ] .
a precise determination of distribution functions requires a high sample completeness . in sect .
[ sample_c ] several completeness tests for hiflugcs are discussed , indicating a high completeness .
the observed @xmath38@xmath192 relation is compared to expectations in sect .
[ relat_d ] and possible applications are indicated . the cluster mass function is compared to previous determinations and to predictions of cosmological models in sect .
[ funct_d ] . the total gravitational mass contained in galaxy clusters
is compared to the total mass in the universe in sect . [ func_dens ] .
the sample completeness is important for the accuracy of the mass function .
the selection criteria detailed in sect .
[ sample ] are met by 63 clusters with mean redshift @xmath983 and with two clusters having @xmath984 .
the sample is constructed from surveys with much deeper flux limits and high completenesses .
a possible remaining incompleteness in these surveys is likely to be present at low fluxes close to their flux limits , which therefore would not effect hiflugcs .
nevertheless four completeness tests have been performed and are described in this section ; they all indicate a high completeness of hiflugcs . the @xmath985@xmath986 and @xmath38@xmath74 diagram are compared to expectations , the luminosity function
is compared to luminosity functions of deeper surveys , and the @xmath987 test is performed .
figure [ lnls ] shows the integral number counts as a function of x - ray flux ( ` @xmath985@xmath986 ' ) .
the slope in the @xmath985@xmath986 diagram is very close to the value @xmath988 expected in a static euclidean universe for uniformly distributed clusters . due to the small number of clusters ( 4 )
the deviation is not significant for @xmath989 .
since the average redshift is smallest for the highest fluxes large scale structure is not completely washed out at the high flux end , therefore the slight bump visible around @xmath990 in fig .
[ lnls ] suggests a deviation caused by cosmic variance .
the effect of an expanding and finite universe on the @xmath985@xmath986 flattening of the slope towards low fluxes is small for the redshift range covered by the sample .
the slope consistent with @xmath988 towards the flux limit therefore indicates a high completeness of hiflugcs . in fig .
[ lz ] the x - ray luminosity is plotted as a function of redshift .
the flux limit is shown as a solid line .
for converting observer rest frame luminosities to source rest frame luminosities depends on redshift _ and _ source spectrum ( @xmath73 ) .
for source rest frame luminosities it is therefore not possible to plot the flux limit as one line in 2 dimensions ( @xmath991 ) , but rather as an area in 3 dimensions ( @xmath992 ) . for consistency
we therefore give in this 2d plot the observer rest frame luminosity ( the correction is less than 6% for 90% of the clusters anyway ) . ]
one notes the increase in rare luminous systems with increasing redshift ( volume ) . because of the seeming underdensity of clusters in the redshift range @xmath993 a comparison with the expected number of clusters as derived from @xmath8-body simulations has been performed .
an ocdm simulation , carried out for analysis of the power spectral densities of reflex clusters @xcite , adjusted to the hiflugcs survey volume in the southern hemisphere ( roughly half of the total volume sampled by hiflugcs ) has been used .
the simulation yields 39 clusters while 33 hiflugcs clusters have been detected in this region .
it is found that in fact not even one cluster with @xmath984 is expected for this volume based on this simulation and the hiflugcs subsample also does not contain any cluster with a redshift larger than 0.1 .
this is a further piece of evidence for the high completeness of the sample . in fig .
[ lxfunc ] the hiflugcs x - ray luminosity function is compared to luminosity functions of larger surveys in the southern ( reflex , @xcite ) and northern ( bcs , @xcite ) hemisphere .
these surveys have much deeper flux limits ( sect .
[ sample ] ) and contain many more clusters .
very good agreement is found , which shows the high completeness and homogeneous selection of hiflugcs . the @xmath987 test ( e.g. , @xcite , sect .
14.5 ) can be used to asses a possible sample incompleteness . assuming a uniform distribution of clusters a value 1/2
is expected on average . for hiflugcs
@xmath994 , which is consistent with the expectation and we interpret this result as a clear sign that hiflugcscovers a large enough volume for most of the @xmath38 range to be representative of the local universe with a high sample completeness .
the local nature of hiflugcsbecomes obvious by noting that the result of the comoving @xmath987 test is almost identical to the result of the equivalent test assuming a euclidean and non expanding space , i.e. , @xmath995 .
the close correlation between the x - ray luminosity and the gravitational mass found in sect . [ relat ] is not surprising .
simple self similar scaling relations predict @xmath996 and @xmath997 , where @xmath998 is a characteristic radius , e.g. , the virial radius . combined with bremsstrahlung emission , @xmath999 ( @xmath1000 ) , the relation @xmath1001 ( @xmath1002 )
is predicted ( e.g. , * ? ? ?
* ) , where the gas fraction @xmath1003 and @xmath1004 is the bolometric luminosity .
observationally from the tight correlations between x - ray luminosity and temperature ( e.g. , @xcite ) , and temperature and mass ( e.g. , @xcite ) a correlation between luminosity and mass clearly is expected .
also correlations found between x - ray luminosity and galaxy velocity dispersion ( e.g. , @xcite ) and x - ray luminosity and mean shear strength from weak lensing studies ( e.g. , @xcite ) indicate a correlation between @xmath38 and @xmath192 . the x - ray luminosity has been compared directly to gravitational mass estimates by @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite , where good correlations have been found in all of these studies . in order to compare the empirical @xmath38@xmath192 relation with predictions a quasi bolometric luminosity , @xmath1004 , has been calculated in the source rest frame energy range @xmath1005 ( for the relevant range of cluster gas temperatures at least 99% of the flux is contained in this energy range ) . in fig .
[ mtlbol ] this @xmath1004@xmath211 relation is compared to predicted relations .
the solid line shows the best fit relation for the 106 clusters in the extended sample and the triple - dot - dashed line shows the best fit relation determined using hiflugcs . here
the bisector fit results have been used in order to treat variables symmetrically , which is the appropriate method for a comparison to theory ( e.g. , * ? ? ?
the dot - dashed line shows the self - similar relation ( @xmath1006 ) normalized by the simulations of @xcite and the dashed line shows the ` pre - heated ' relation given by @xcite ( @xmath1007 ) , using a normalization taken from the simulations of pre - heated clusters by @xcite .
the idea of pre - heating is that the intracluster gas is not cold initially , as in the self similar case , but is heated by some form of non gravitational heat input , e.g. , from supernovae or agn , before or during cluster formation .
assuming the central regions of all clusters to have the same entropy yields the latter relationship .
[ mtlbol ] shows that measured and predicted relations are in rough agreement , the difference between the predicted relations being larger than the difference to the observed relations .
note , however , that the x - ray luminosity is one of the most uncertain quantities to be derived from simulations .
@xcite recently showed in a comparison of twelve different cosmological hydrodynamics codes that a factor of two uncertainty is a realistic estimate of the current accuracy . including gas cooling in simulations worsens the situation ( e.g. , * ? ? ?
the slopes of the observed relations are closer to the pre - heated relation .
observationally the effect of pre - heating can also result in a decrease of the gas mass fraction for low temperature systems .
this has actually been observed for the clusters in our sample @xcite .
the possibility that winds from , e.g. , supernovae originally invoked to explain the apparent low gas content of elliptical galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) pre - heat and dilute the central gas and thereby break the self similarity has been pointed out by various authors ( e.g. , @xcite ) .
such a process would work most efficiently on the least massive clusters ( e.g. , @xcite ) .
the @xmath38@xmath192 relation and other relevant relations between physical cluster parameters of the extended sample have been discussed more thoroughly in this context in @xcite .
one may wonder about the origin of the scatter in the @xmath192@xmath38 relation .
due to the use of pointed observations for most of the clusters and the local nature of hiflugcs the statistical errors of @xmath38 are negligible .
the logarithmic mean mass measurement uncertainty has been measured as 0.12 .
the overall scatter in log mass of the data points compared to the best fit relation is larger and has been measured as 0.21 ( tab .
[ tab : scat ] ) .
this indicates a possible contribution of intrinsic scatter to the overall scatter in the @xmath192@xmath38 relation .
an obvious candidate to cause intrinsic scatter is the central excess emission ( central surface brightness exceeding single @xmath185 model surface brightness ) present in a number of clusters .
this excess emission may have its physical origin either in cooling flows ( e.g. , * ? ? ?
* ) or in the presence of cd galaxies ( e.g. , * ? ? ?
a cooling flow analysis of the hiflugcs clusters is in progress and first results indicate that indeed clusters with a high inferred mass deposition rate lie on the high @xmath38 side of the @xmath192@xmath38 relation ( y. chen et al . , in preparation ) . in fig .
[ nmlxm ] the measured number of cluster member galaxies as taken from @xcite is compared to @xmath38 as gravitational mass tracer .
it is clearly seen that a selection by x - ray luminosity is much more efficient than a selection by abell richness in terms of mass .
even though only the x - ray surface brightness profile and neither its normalization nor the x - ray luminosity are directly used in the x - ray mass determination via the hydrostatic equation , it is nonetheless reassuring that a similar result is obtained when @xmath38 and richness are compared to masses estimated from optical velocity dispersions @xcite .
a wide range of possible applications becomes available with the quantification of the @xmath38@xmath192 relation and its scatter . for large x - ray flux - limited cluster surveys , where individual mass determinations are currently not feasible
, luminosities can be directly converted to masses .
no combination of observations , simulations and theory is then needed , like the frequently used approach of relating x - ray luminosities to x - ray temperatures by an observed relation , and converting x - ray temperatures to masses using a relation where the slope is taken from theoretical arguments and the normalization from hydrodynamical simulations ( e.g. , @xcite ) .
the observational @xmath38@xmath192 relation has first been applied directly in this sense in the power spectral analysis of reflex clusters @xcite .
an example of another direct application is given in sect .
[ lmtest ] . the @xmath38@xmath192 relation may also be applied to convert theoretical or simulated mass functions to luminosity functions for comparison with observations of x - ray flux - limited samples , which is currently being performed in the interpretation of the reflex luminosity function . at this point
it is important to note that even for the highest redshift cluster in our sample ( @xmath1008 ) the dependence of the observational determination of @xmath192 and @xmath38 on the chosen cosmological model is very weak .
for instance at @xmath1008 the increase in the luminosity distance , @xmath1009 , and the diameter distance , @xmath1010 , is less than 5% going from ( @xmath1011 ) to ( @xmath1012 ) . from ( [ eq : mt ] ) one finds that @xmath1013 and therefore @xmath1014 , implying an increase of @xmath192 by less than 5% for the two models above . for
@xmath38 one has an increase of less than 10% .
this means that the @xmath38@xmath192 relation given here can be used unchanged for various cosmological applications ( unless redshift ranges are probed where evolution becomes important , in this case a model dependent redshift correction has to be introduced ) . a similar calculation for @xmath964 shows that for the extreme case @xmath1015 the increase in @xmath964 is less than 14% , which is less than the size of the poissonian error bars in fig .
[ ps0 ] .
more detailed investigations on the shape of the relation are in progress and it is also envisaged to construct a volume - limited sample , spanning a reasonably large range in luminosity and mass , to test how much the @xmath38@xmath192 relation given here is affected by being estimated partly from a flux - limited sample . @xcite
give a mass function constructed a ) from optically selected clusters with masses determined from the galaxy richness and b ) from the cluster x - ray temperature function given by @xcite .
very good agreement is found for masses determined within @xmath209 between the @xcite and hiflugcs mass function ( fig .
[ mfunc ] ) .
@xcite constrain the cluster abundance by using published values for the abundance and median velocity dispersion of richness class @xmath1016 abell clusters .
it is not surprising that their density is significantly higher than the hiflugcs density since they have intentionally used conservative mass estimates , which are overestimates of the true cluster masses .
@xcite and @xcite have determined the cluster mass function using optically selected cluster samples with masses determined from published line - of - sight velocity dispersions of cluster galaxies . at the completeness limit given by (
* triangle in fig .
[ mfunc ] ) the cluster density given by @xcite is about a factor of two higher than the density given by @xcite .
the latter authors explain this by their on average 40% smaller mass estimates due to an improved technique for removing interlopers and the use of a surface - correction term in the virial theorem .
the value given by @xcite itself lies significantly higher than the comparable hiflugcs density .
the reason could lie in the fact that their optically estimated masses are in general slightly larger than the x - ray masses or that the external normalization for @xmath1016 ( @xmath1017 ) clusters which they applied to their mass function is intrinsically higher than the normalization obtained through hiflugcsdirectly , or both . by comparing the mass estimates for a common subsample of 42 clusters
it has been found that the virial masses determined by @xcite are on average 25% larger than the masses determined in this work .
this difference might be smaller if masses out to the abell radius were compared , since common radii would have been used in this case . increasing the x - ray masses artificially by 25% , the diamonds in fig .
[ mfunc ] shift towards higher masses , but the shift is too small to account for the difference to the triangle .
the large scatter in the @xmath1018@xmath192 diagram ( fig . [ nmlxm ] ) makes a reliable estimate of a best fit relation between these two quantities very difficult .
nevertheless , in order to get a rough idea of the mass within @xmath209 that corresponds to @xmath1019 , we have performed fits using the minimization methods outlined earlier and find @xmath1020 .
note that this range is in agreement with the ranges @xmath1021@xmath1022 and @xmath1021@xmath1023 given by @xcite and @xcite , respectively , for @xmath1024 .
this mass range corresponds to a cumulative number density obtained through hiflugcs in the range @xmath1025 .
the external density estimate applied to normalize the @xcite mass function therefore is a factor 1.26.2 higher than the corresponding estimate obtained here .
it is therefore concluded that both effects ( masses and normalization ) are important but the latter factor is responsible for a larger fraction of the discrepancy . assuming both normalizations to be determined from samples that are highly complete and representative of the local universe this may indicate that either x - ray and optical clusters are drawn from different populations or that projection effects ( e.g. , line of sight galaxy overdensities , which do not form a bound structure in three dimensions ) possibly bias optically determined normalizations high .
@xcite recently extended the mass function to loose galaxy groups , finding @xmath1026@xmath1027 , which is outside the mass range we can test .
they find that the group mass function can be described by a smooth extension of the cluster mass function by @xcite .
consistently this abundance is higher than the abundance given by @xcite at that mass scale .
@xcite have compiled and partially reestimated abundances for local cluster samples @xcite for comparison with higher redshift samples ( the ` @xmath1028 ' shows the density for a sample with higher mean redshift and therefore it can not be compared directly ) .
note that @xcite find that the reestimate of the mass limit of the @xcite sample by @xcite ( the ` + ' at @xmath1029 ) may lead to an underestimated mass limit .
in general the comparison to the hiflugcs mass function shows better agreement than the abundances given by @xcite and @xcite .
the obvious importance of the definition of the cluster outer radius for the cluster mass function can be directly appreciated by noting the large differences between the mass functions determined for hiflugcs for @xmath213 , @xmath212 , and @xmath1030 in fig .
[ mfunc ] , especially towards lower masses . since for self
similar clusters the mass scales with the third power of the characteristic radius ( sect .
[ relat_d ] ) , determination of the mass within a characteristic overdensity is the natural choice .
we mainly give the formally determined @xmath1030 mass function for the comparison with previous mass functions and recall again that especially for the low mass systems the assumption of hydrostatic equilibrium may not be justified out to @xmath209 , and therefore our mass estimates of @xmath213 and @xmath212 should be considered much more precise than the estimates for @xmath1030 .
we use the standard formalism based on the press
schechter ( ps ) prescription to predict cluster mass functions for given cosmological models ( see , e.g. , @xcite ) . to allow easier comparison with the theoretical literature on this subject in this paragraph @xmath1031 is used . the mass function
is then given by @xmath1032 ( @xcite , @xcite ; see , e.g. , @xcite for a compilation of published extensions of the ps mass function ) . here
@xmath1033 represents the halo ( cluster ) virial mass and @xmath1034 is the present day mean matter density .
the linear overdensity computed at present @xmath1035 , where the linear overdensity at the time of virialization , @xmath1036 , is computed using the spherical collapse model summarized in @xcite , for @xmath15 using ( a2 ) and for @xmath1037 using ( a6,7 ) ; the linear growth factor @xmath1038 and @xmath1039 has been defined in sect . [ massd ] . as mentioned earlier due to the low redshift range spanned by hiflugcs
, the effect of a redshift correction is very small and we therefore set @xmath1040 for all calculations , unless noted otherwise .
the variance of the cosmic mass density fluctuations @xmath1041 where @xmath1042 represents the amplitude of density fluctuations in spheres of radius @xmath1043 .
recent measurements of the cosmic microwave background ( cmb ) anisotropies indicate that the primordial power spectral index , @xmath1044 , has a value close to 1 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and is therefore set to 1 , unless noted otherwise .
for the transfer function we use the fitting formula for cold dark matter ( cdm ) cosmologies provided by @xcite for @xmath1045 @xmath1046^{-1/4}\ , , \label{eq : transf}\end{aligned}\ ] ] where the shape parameter is given by ( modified to account for a small normalized baryon density @xmath1047 , @xcite ) @xmath1048 the temperature of the cmb @xmath1049 @xcite and @xmath1050 @xcite , for the latter equation and ( [ eq : gama ] ) @xmath1051 has been used @xcite .
the comoving filter radius @xmath1052^{1/3}$ ] for the top hat filter function @xmath1053 , which is adopted in this analysis , because the hiflugcs masses have been determined with a top hat filter , too .- space filter @xcite ; see @xcite for a generalization to more realistic filter functions . ] since the ps recipe as outlined above assumes virial masses based on the spherical collapse model we use @xmath212 as approximation to the virial mass ( sect .
[ massd ] ) . similarly to the work of @xcite the statistical uncertainty in the mass determination
is incorporated in the model mass function as @xmath1054 where @xmath1055 represents the logarithmic mean mass measurement uncertainty .
note that since hiflugcs is flux - limited and not volume - limited the weighting has to be performed on the mass distribution , @xmath1056 , rather than on the mass function itself .
the effect of this weighting on the model mass function is a slight amplitude increase at the high mass end . for the modeling to be independent of the precise knowledge of the @xmath38@xmath192 relation
the quantitative comparison has been performed using a standard @xmath1057 procedure on the differential binned mass function given in fig .
[ ps0 ] ( rather than using a maximum likelihood approach on the mass distribution ) . after identifying the crude position where @xmath1057 is minimal in a large @xmath9@xmath1042 parameter space region the @xmath1057 values
have been calculated in a fine grid of 200 by 200 @xmath9@xmath1042 values in the range @xmath1058 and @xmath1059 .
a flat cosmic geometry has been assumed , i.e. @xmath1060 .
the cosmological constant enters the calculation only through @xmath1061 , however , and therefore has a negligible influence here .
the minimum and statistical error ellipses for some standard confidence levels ( c.l . )
are given in fig .
[ banana ] .
the tight constraints obtained show that with hiflugcs we can go beyond determining an @xmath9@xmath1042 relation and put limits on @xmath9 and @xmath1042 individually . it is found that @xmath1062 ( 90% c.l .
statistical uncertainty for two interesting parameters ) , indicating a relatively low value for the density parameter .
the large covered mass range , the specific region in @xmath9@xmath1042 parameter space , and the assumption of cdm cosmological models with given @xmath1063 , @xmath1064 , @xmath1044 , and @xmath1065 allow to derive these tight constraints from a local cluster sample . for comparison for a given @xmath1042 value the @xmath9 value which minimizes @xmath1057 is calculated . in the interval shown in fig .
[ banana ] these pairs can then roughly be described by a straight line in log space given by @xmath1066 in fig .
[ ps0 ] we also plot the best fit model mass functions for given @xmath1067 and @xmath1068 and one notes immediately that these value pairs give a poorer description of the shape of the mass function .
the slight differences between the best fit values for @xmath1042 and the ones expected from ( [ eq : omsig ] ) are caused by the fact that the simple power law is only an approximation to the real shape of the error ellipse and weakly by the extrapolation involved .
previous estimates have generally yielded a combination of slightly higher @xmath9@xmath1042 values ( e.g. , @xcite , sect .
it has to be noted , however , that for instance in the important work of @xcite , who find @xmath1069 , the authors explicitly state that their estimates of @xmath1042 are probably biased high due to their conservative mass estimates . before exploring sources of systematic errors
let us evaluate our treatment of the shape parameter @xmath1070 . by applying the @xcite fitting formula to calculate the transfer function we have adopted the assumption of a cdm universe , i.e. @xmath1071 . within this model @xmath1070 and @xmath9 are not independent . for a small @xmath1072 ratio @xmath1070
is given by ( [ eq : gama ] ) .
however , to test its influence we have set @xmath1070 to a fixed value of 0.19 as implied by large scale structure ( lss ) measurements of galaxies and clusters ( e.g. , * ? ? ?
* ; * ? ? ?
* see also the discussion at the end of this section ) .
the best fit values for @xmath9 and @xmath1042 change significantly , the @xmath9 value becoming smaller and the @xmath1042 value larger .
the @xmath9@xmath1042 relation changes slightly to about @xmath1073 for the @xmath9 range [ 0.040.30 ] ( note that for this range relation ( [ eq : omsig ] ) translates to about @xmath1074 , the steepening being mainly due to a strong increase of @xmath1042 for the very lowest @xmath9 values ) .
the error ellipse in general becomes much longer when @xmath1070 is set to a fixed value , preventing the possibility to obtain tight individual constraints on @xmath9 and @xmath1042 , even though @xmath1068 is still excluded at least at the 90% c.l . for all @xmath1075 .
the decreased sensitivity is due to the loss of the amplification of a change of @xmath9 through @xmath1070 ( to first order @xmath1076 , eq .
[ eq : gama ] ) , since both a larger ( smaller ) @xmath9 and a larger ( smaller ) @xmath1070 result in steeper ( less steep ) model mass functions .
having found a dependence of the results on @xmath1070 we tried to constrain @xmath1070 directly from the observations , as done by previous authors ( e.g. , * ? ? ? * ) .
we have set @xmath1070 to 31 different fixed values in the range @xmath1077 and calculated a @xmath1078 for each of those by varying @xmath9 and @xmath1042 .
the result is shown in fig .
[ chigam ] .
the conclusion is that the data do not contain enough information to provide an independent constraint on @xmath1070 .
therefore we have taken advantage of the dependence of @xmath1070 on @xmath9 within the cdm framework and have continued to calculate @xmath1070 with eq .
[ eq : gama ] .
notice that this procedure requires additional knowledge about @xmath1063 and @xmath1064 .
the influence of the uncertainty of these two parameters is tested along with various other possible systematic uncertainties in the following .
it is important to note that the best fit value for @xmath9 is not determined solely through ( [ eq : gama ] ) but also directly through @xmath1079 in ( [ eq : ps ] ) and through the filter radius @xmath1080 .
the quoted error ranges have been calculated from the @xmath1057 procedure .
most of the remaining part of this section is devoted to an identification and quantification of possible systematic uncertainties .
the quantitative results are summarized in tab .
[ tab : sys ] .
due to the large given ranges of several orders of magnitude in mass and especially density the @xmath1057 values have been determined from comparison model / data naturally in log space .
however , we have verified that the same calculation in linear space yields best fit values lying within the 68% error ellipse . in sect .
[ funct ] arguments have been given why we have used @xmath966 for the determination of the mass function . nevertheless if @xmath969 is used instead ( see fig .
[ vm_cp_ns ] ) , consistent results are obtained . since in this case
we want to estimate @xmath1081 from @xmath1033 the relation ( @xmath980 ) is the appropriate one ( e.g. , * ? ? ?
performing a fit to a mass function constructed with @xmath969 results in best fit values @xmath1082 and @xmath1083 , which is consistent with the error range given in ( [ eq : omres ] ) . the data point in fig .
[ ps0 ] that may be affected most by cosmic variance is the one at the lowest mass , since the maximum search volume is smallest for the clusters in this bin .
we therefore tested the sensitivity of the best fit results on this last point by ignoring it .
the decrease in the covered mass range of course increases the resulting error ellipse , but the best fit values vary only within the ( smaller ) 68% error ellipse . it may be worth noting that leaving out the highest mass bin or leaving out the highest _ and _ lowest mass bin changes the best fit values only within the 90% error ellipse .
simulations seem to indicate that the mass density profile of virialized halos follows the nfw profile ( * ? ? ?
* ; * ? ? ?
* but see , e.g. , @xcite ) . as test therefore gravitational masses have been recalculated assuming the dark matter density to follow an nfw profile based on the measured values for @xmath185 and @xmath188 .
the approximation @xmath1084 and @xmath1085 given by @xcite has been used combined with their eqs . 7 and 9 .
@xmath197 and @xmath212 have been redetermined based on this model and the resulting mass function has been compared to model mass functions . on average
it is found that the nfw masses are a factor of 0.90 lower , the difference being smaller for high mass clusters than for low mass ones .
the values for @xmath9 and @xmath1042 for which @xmath1057 is minimal lie within the 90% statistical error ellipse shown in fig . [ banana ] .
therefore the choice of mass calculation does not affect the results significantly . as shown in sect .
[ massd ] it is possible that the assumption of isothermality leads to an overestimate of the cluster masses on average
. therefore the robustness of the results has been tested by multiplying the isothermal cluster masses by 0.80 and recalculating the minimum . as expected values for both @xmath9 and @xmath1042
are found to be lower .
but the new minimum is contained well within the error ranges given in ( [ eq : omres ] ) .
on the other hand in sect .
[ func_comp ] it has been shown by comparison with optical virial mass measurements that masses could possibly be underestimated . after increasing all cluster masses by 25% a fit shows that both @xmath9 and @xmath1042 become slightly larger , but again they lie well within the range ( [ eq : omres ] ) .
therefore these tests indicate that the constraints obtained here are fairly insensitive against systematic uncertainties in the mass measurements .
note that systematic effects that show a mass dependence have a higher potential to affect the individual best fit values of @xmath9 and @xmath1042 than effects which simply shift all data points by about the same amount in the same direction .
however , no indications for a mass dependence of the differences between the above mass estimations have been found .
it has been mentioned that masses calculated within @xmath198 are less affected by systematic uncertainties because less extrapolation beyond @xmath69 is needed and because according to simulations the assumption of hydrostatic equilibrium can be more safely applied .
nevertheless we have used @xmath212 for the comparison to predicted mass functions because it is a better approximation to the cluster virial mass . however , for comparison the best fit values have also been determined using @xmath213 , knowing that virial masses are probably underestimated this way .
the result is that the best fit value for @xmath9 changes only marginally .
the best fit @xmath1042 value is slightly lower than allowed by the statistical error . on the other hand
the spherical collapse model implies that one may need to extrapolate even further than @xmath197 for low density universes .
for instance in a flat universe @xmath1086 implies that the virial radius is close to @xmath1087 .
therefore a mass function has been constructed using @xmath1088 as cluster masses , knowing that these masses are likely to be rather uncertain .
nevertheless the resulting best fit @xmath9 value again varies only insignificantly whereas the @xmath1042 value becomes significantly larger .
note that for both tests the best fit value for @xmath9 hardly changes . as mentioned in sect .
[ massd ] it has been shown that the small range of low redshifts covered here ensures that no redshift corrections need to be applied .
nevertheless we have tested whether or not the best fit parameter values change if @xmath212 is calculated using @xmath1089 for the extreme ( strong evolution ) case ( @xmath205 ) , i.e. , @xmath1090 for each cluster redshift . the model mass function
is then calculated for the mean redshift of hiflugcs , @xmath1091 , using the formulae outlined in the beginning of this section .
we have found that within our grid the best fit values do not change at all and also the error ellipses are almost not affected , thereby confirming that the application of redshift corrections does not affect the results .
the value @xmath1092 has been adopted for the calculation of model mass functions based on the recent combination of constraints obtained using the hubble space telescope @xcite .
setting the hubble constant to their lower limit , @xmath1093 , does not affect the best fit parameter values significantly .
using even @xmath1094 changes the results for @xmath9 and @xmath1042 only well within the 68% error ellipse . in tab .
[ tab : sys ] results have also been listed for @xmath1095 and @xmath1096 showing that even variations within this range change the @xmath9 and @xmath1042 values only within their statistical uncertainties given in ( [ eq : omres ] ) .
therefore our constraints on these cosmological parameters do not depend significantly on the specific choice of @xmath1063 . a baryon density based on measurements of the primordial deuterium to hydrogen ratio combined with standard nucleosynthesis theory has been used in this work .
the used value also agrees with recent cmb measurements ( e.g. , * ? ? ? * ) .
however , we have also calculated the minimum for the values @xmath1097 and @xmath1098 finding that the position of the minimum in each case does not deviate significantly from our original result .
the primordial power spectral index has been set to @xmath1099 based on measurements of the cmb fluctuations .
the influence of a change in this parameter has been estimated by calculating the minimum for the values @xmath1100 and @xmath1101 .
the differences in the best fit parameters are smaller than the statistical uncertainties given in ( [ eq : omres ] ) . since we have found that for the estimate of the statistical errors we need to explore ranges @xmath1102 for @xmath1103 we regarded it necessary to check if the approximation to the transfer function as given in the beginning of this section is still applicable .
recently @xcite derived a fitting function , that includes , e.g. , also the oscillations induced by the baryons , which gives a better description of transfer functions computed with cmbfast @xcite than fitting functions for zero or small baryon contribution to the total matter density derived previously .
therefore we incorporated this improved version of an analytic transfer function in the @xmath1057 procedure .
we have found that within our grid there is only a very weak shift of the minimum , the choice of the @xcite fitting function combined with the shape parameter given by @xcite therefore seems to be accurate enough for our purposes .
however , the confidence contours towards low @xmath9 are getting compressed when the @xcite fitting function is used , thereby slightly decreasing the area of the error ellipse for a given confidence level .
since we regard the latter statistical error ellipse as more realistic , we show this one in fig .
[ banana ] .
ccccc 0.116 & 0.928 & @xmath1104 & @xmath1105 & ( 1 ) + 0.137 & 0.865 & @xmath1106 & @xmath1107 & ( 2 ) + 0.150 & 0.897 & @xmath1108 & @xmath1109 & ( 3 ) + 0.092 & 1.082 & @xmath1110 & @xmath1111 & ( 4 ) + 0.083 & 1.050 & @xmath1112 & @xmath1113 & ( 5 ) + 0.150 & 0.860 & @xmath1108 & @xmath1114 & ( 6 ) + 0.107 & 0.907 & @xmath1115 & @xmath1116 & ( 7 ) + 0.132 & 1.004 & @xmath1117 & @xmath1118 & ( 8) + 0.106 & 0.826 & @xmath1119 & @xmath1120 & ( 9 ) + 0.108 & 1.242 & @xmath1121 & @xmath1122 & ( 10 ) + 0.116 & 0.959 & @xmath1104 & @xmath1104 & ( 11 ) + 0.118 & 0.956 & @xmath1123 & @xmath1124 & ( 12 ) + 0.153 & 0.861 & @xmath1125 & @xmath1126 & ( 13 ) + 0.108 & 0.991 & @xmath1121 & @xmath1127 & ( 14 ) + 0.106 & 1.001 & @xmath1119 & @xmath1128 & ( 15 ) + 0.128 & 0.923 & @xmath1129 & @xmath1130 & ( 16 ) + 0.148 & 0.874 & @xmath1131 & @xmath1132 & ( 17 ) + 0.094 & 1.056 & @xmath1133 & @xmath1134 & ( 18 ) + we have also tested whether or not the recently found deviations of the ps formalism compared to large @xmath8-body simulations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) have a significant influence on the results obtained here .
we have compared the best fit ps model ( @xmath1135 , @xmath1136 ) to the model obtained using the ` universal ' mass function ( fit to @xmath8-body simulations , @xcite ) for the same parameter values .
these two models agree well for @xmath1137 .
the differences become larger than the size of the poissonian error bars ( fig .
[ ps0 ] ) for @xmath1138 , in the sense that the @xcite mass function predicts higher cluster abundances than ps . for larger values of @xmath9
the differences become comparable to the size of the error bars at lower masses , e.g. , for @xmath1139 around @xmath1140 . to estimate the influence of these differences on the best fit values derived using the ps mass function , we adjusted the parameter values of the @xcite model to reproduce the ps mass function , finding @xmath1141 and @xmath1083 .
the value for @xmath9 becomes slightly larger but the combination of both values is still contained within the 90% error ellipse .
we therefore conclude that the differences between the model mass functions do not significantly affect the interpretation of the hiflugcs mass function .
moreover we regard this test as confirmation of the validity of the ps mass function for the accuracy needed here .
almost all identified systematics give rise to smaller uncertainties than the statistical uncertainties ( eq .
[ eq : omres ] ) .
however , it is not impossible that several systematics combined have a significant effect . in order to determine the highest @xmath9 value allowed by a conspiracy of all identified uncertainties , all positive @xmath1142 values from tab . [
tab : sys ] have been added to the positive statistical uncertainty for @xmath9 and this sum has been added to the @xmath9 value which minimizes @xmath1057 ( @xmath1135 ) . in this worst case procedure
the upper limit @xmath1143 has been found . using @xmath1144 , as derived from the mean gas mass fraction within @xmath197 for the 106 clusters in the enlarged sample , results in model mass functions very similar to the ones calculated using the baryon fraction given by @xcite .
it is therefore not surprising that the best fit values for @xmath9 and @xmath1042 vary only well within the 68% error ellipse if the former @xmath1064 determination is used .
moreover it is worth noting that combining these two measurements of @xmath1064 gives further evidence for a low value for @xmath9 by yielding an estimate @xmath1145 using @xmath1146 @xcite , where the error has been determined from the standard deviation of @xmath1064 given above .
this value for @xmath9 is an upper limit since baryons , e.g. , contained in the cluster galaxies or forming part of the dark matter have been neglected .
a low value for @xmath9 has been indicated by this method for smaller cluster samples by several works previously ( e.g. , @xcite ; but see @xcite ) .
one has to keep in mind , however , that this estimate extrapolates the gas fraction from cluster scale to cosmic scales . for the clusters in our sample we have found that the gas fraction is not constant but varies with radius and cluster mass @xcite , therefore further observational tests of this assumption may be useful .
lllc 0.30 & 1.12 & @xmath1147 & 1 + 0.30 & 1.16 & @xmath1148 & 2 + 0.30 & 0.93 & @xmath1149 & 3 + 0.30 & 1.01 & @xmath1150 & 4 + 0.30 & 0.78 & & 5 + 0.30 & 0.99 & @xmath1151 & 6 + 0.30 & 0.96 & @xmath1152 & 7 + 0.30 & 0.96 & & 8 + 0.30 & 1.02 & @xmath1153 & 9 + 0.30 & 0.91 & @xmath1154 & 10 + 0.30 & 0.68 & @xmath1155 & 11 + let us briefly discuss the role of @xmath1070 again .
the constraints ( [ eq : omres ] ) together with @xmath1051 , @xmath1156 , and eq .
[ eq : gama ] imply @xmath1157 , which is significantly lower than values for @xmath1070 determined through lss studies of galaxies ( e.g. , @xcite have found @xmath1158 ) or clusters ( e.g. , @xcite have found @xmath1159 ) . adopting the worst case upper limit on @xmath9 found above yields an upper limit @xmath1160 consistent with the lss results .
however , also many previous studies of cluster abundances have yielded low values for @xmath1070 .
for instance @xcite have found @xmath1161 ; @xcite has found @xmath1162 ; @xcite have found @xmath1163 which , e.g. , for @xmath1164 yields @xmath1165 ; judging from their fig .
3 and their best fit values for a spatially flat universe @xcite have found @xmath1166 , however , not significantly constrained .
it is unclear whether these values indicate a significant discrepancy but it is interesting to note that cluster abundances seem to favor lower values for @xmath1070 than lss studies do .
if significant this could possibly imply either unidentified systematic errors in one or both measurements from the cluster abundance side one could for instance imagine a higher merger rate for more massive clusters resulting in overestimated x - ray luminosities , temperatures , and masses for the massive clusters resulting in less steep @xmath38 , @xmath39 , and @xmath192 functions and therefore in a decreased value for @xmath1070 ( and possibly @xmath9 ) or that the description of the power spectrum using a single @xmath1070 parameter on all scales is not accurate ( e.g. , @xcite , sect .
33.3.3 ) . even though we made a conservative estimate by neglecting the possible presence of gas temperature gradients , previous estimates obtained from cluster abundances generally yielded higher values for @xmath9 and @xmath1042 ( e.g. , * ? ?
* ; * ? ? ?
* ) . for the latter two works one could have expected this already from fig . [ mfunc ] and possible reasons have been discussed in sect .
[ func_comp ] .
recently a number of constraints on @xmath9 and @xmath1042 have been obtained using different kinds of distribution functions assuming relations between the measured quantities and mass .
some of them incorporated additional dynamical information from the evolution of the corresponding distribution function .
especially intracluster gas temperature functions have been constructed and used frequently . in tab .
[ tab : sumcons ] some results , which allowed a simple comparison for a given @xmath1164 , are listed . if there was a choice , results assuming flat cosmological models and isothermal clusters are quoted .
this table is not a complete summary of recent results but merely shows that previous results yielded higher values for @xmath1042 than obtained here from a newly constructed sample ( even using @xmath1088 instead of @xmath212 would increase the @xmath1042 value only up to 0.82 ) .
the reasons may be manifold .
some have been discussed above .
other reasons may include different sources of temperature estimates .
for instance @xcite used cooling flow corrected temperatures given by @xcite , which are generally higher than temperatures derived previously . using theoretical / simulated @xmath192@xmath73 relations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) to connect observed temperature functions with theoretical mass functions will on average also result in higher values for @xmath1042 , since these relations have been shown to have higher normalizations than observed x - ray mass temperature relations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
assuming low temperature groups to be affected by non gravitational heat input , temperature functions , unlike x - ray mass functions which incorporate the additional information of the gas density profile , would show an additional artificial increase towards low temperatures .
if not taken into account , e.g. , by use of a modified @xmath192@xmath73 relation , this effect would result in artificially increased values for @xmath9 ( see fig .
[ ps0 ] ) . also the completeness and representativeness of the used cluster samples obviously play an important role for the @xmath9@xmath1042 determination . here
advantage has been taken of the currently best available local sample in the sense of homogeneous x - ray selection , completeness , size , and availability of high quality observations .
the results obtained here are , however , in good agreement with the results from the power spectral analysis of the 452 reflex clusters .
@xcite find for a given @xmath1167cdm model ( @xmath1164 ) that @xmath1168 represents the data well , which is very close to @xmath1169 expected using relation ( [ eq : omsig ] ) found here .
moreover the ( 1-@xmath70 ) range @xmath1170 ( using @xmath1171 in their eq . 18 ) quoted for @xmath9 directly is also consistent with the 90% range determined here .
furthermore @xcite , who analyzed the hiflugcs temperature function using temperatures from homogeneously reanalyzed asca data , find @xmath1172 and @xmath1173 ( 90% c.l .
statistical uncertainty ; assuming an open cosmology ) by comparison with press schechter models , which is in good agreement with our results .
moreover , after this paper had been submitted another paper appeared on astro - ph @xcite , where the authors also find @xmath9@xmath1042 values that agree with values calculated using relation ( [ eq : omsig ] ) presented here . to show consistency and the power of the @xmath38@xmath192 relation , we have also performed fits to ` mass ' functions , where masses have been estimated from the measured x - ray luminosity .
relations for the flux - limited sample have been used to get the best mass estimate for the cluster luminosities included in hiflugcs .
mass functions for the two extreme relations ( @xmath978 ) with @xmath979 and ( @xmath980 ) with @xmath981 are shown in fig .
[ mf_lm ] .
first of all one notes the fairly good agreement between the three mass functions . in detail
the differences between the two mass functions estimated from different luminosity mass relations can be understood by considering that at the low luminosity end the steeper relation predicts a higher mass for a given luminosity than the shallower relation , resulting in a shift towards higher masses of the mass function . at the high mass side
the effect is opposite , resulting in a shift towards lower masses for the steeper relation . on average the points for the steeper relation lie higher which is caused by the fact that a steeper relation results in a smaller @xmath1174 on average , which gives rise to an increased @xmath982 .
the differences to the mass function calculated using the measured masses are again understood by a similar comparison and are partially caused by a possible deviation of the shape of the @xmath38@xmath192 relation from a pure power law . despite these small differences performing an actual fit for the two relations ( @xmath980 ) and ( @xmath978 ) , @xmath1175 and @xmath1176 , respectively ( tab .
[ tab : scat ] ) , has replaced the mass measurement error , @xmath1177 , in ( [ eq : nschl ] ) . ] results in the ( @xmath9 , @xmath1042 ) values ( 0.14 , 0.85 ) for ( @xmath980 ) and ( 0.22 , 0.74 ) for ( @xmath978 ) . the first case is consistent with the error range given in ( [ eq : omres ] ) and in the second case the 90% statistical error ellipse overlaps with the 90% ellipse in fig .
[ banana ] . from the above and fig .
[ ps0 ] it is clear that using steeper luminosity mass relations results in higher values for @xmath9 and lower values for @xmath1042 . here we want to estimate @xmath1033 from @xmath1081 and therefore ( @xmath978 ) is the appropriate relation to use .
this test shows that with a comparatively easy to obtain x - ray luminosity function of a statistical cluster sample and with the knowledge of the empirical @xmath38@xmath192 relation ( even if approximated as simple power law ) and its scatter as presented here useful constraints on cosmological parameters can be set by construction of a ` quasi mass function ' . to estimate the fraction of the gravitational mass density relative to the critical density contained in galaxy clusters , @xmath1178 , the individual cluster masses divided by the corresponding maximum search volumes
have been summed up for hiflugcs , i.e. , @xmath1179 .
note that the determination of @xmath1180 is independent of the hubble constant .
the cumulative diagram for @xmath1181 is shown in fig .
[ mdens ] . in order to perform a conservative error estimate ,
hiflugcs has been split into two parts with @xmath1182 and @xmath1183 , and the results for these subsamples are also shown in the figure .
this estimate is conservative because hiflugcs is about twice as large as each subsample .
taking the second and third lowest mass clusters together with the maximum mass range given by their individual uncertainties , we obtain @xmath1184 for masses larger than @xmath7 , i.e. , the total gravitating mass contained within the virial radius of clusters amounts only to @xmath1185 percent of the total mass in a critical density universe .
combined with our best estimate @xmath1135 this implies that about 90% of the total mass in the universe resides outside virialized cluster regions above the given minimum mass .
if galaxies trace mass it also follows that by far most of the galaxies do not sit in clusters .
this result is consistent with the general presumption that clusters are rare objects , rare peaks in the density distribution field . comparing the diagram to the mass fraction @xmath1186 in clusters with masses larger than @xmath1187 given by @xcite , based on the mass function determined by @xcite ,
one finds that their estimate is a factor 45 higher .
however , the @xcite mass function is given for @xmath1030 and we get a consistent result if we calculate @xmath1180 using our formally determined cluster masses within @xmath209 .
it needs to be pointed out that at @xmath1188 we find that the typical virial radius is @xmath1189 and a mass determination at @xmath1190 based on the assumption of virial equilibrium may therefore be rather uncertain and possibly leading to overestimates of @xmath1180 .
this becomes more crucial if mass functions for @xmath1030 are extrapolated even down to galaxy masses .
this way @xcite find @xmath1191 within the mass range @xmath1192@xmath1187 , which , compared to our results from the previous section , would account already for almost all mass in the universe .
an x - ray selected and x - ray flux - limited sample comprising the 63 x - ray brightest galaxy clusters in the sky ( excluding the galactic band , called hiflugcs ) has been constructed based on the rosat all - sky survey .
the flux limit has been set at @xmath1193 in the energy band @xmath78 .
it has been shown that a high completeness is indicated by several tests . due to the high flux limit
this sample can be used for a variety of applications requiring a statistical cluster sample without any corrections to the effective survey volume .
mainly high quality pointed observations have been used to determine fluxes and physical cluster parameters .
it has been shown that a tight correlation exists between the x - ray luminosity and the gravitational mass using hiflugcs and an extended sample of 106 galaxy clusters .
the relation and its scatter have been quantified using different fitting methods .
a comparison to theoretical and numerical predictions shows an overall agreement .
this relation may be directly applied in large x - ray cluster surveys or dark matter simulations for conversions between x - ray luminosity and gravitating mass . using hiflugcs
the gravitational mass function has been determined for the mass interval @xmath1194 .
comparison with press
schechter mass functions assuming cdm power spectra has yielded tight constraints on cosmological parameters .
the large covered mass range has allowed to put constraints on relevant parameters individually .
specifically we have found @xmath2 and @xmath3 ( 90% c.l . statistical uncertainty ) .
various tests for systematic uncertainties have been performed , including comparison of the press
schechter mass function with the most recent results from large @xmath8-body simulations , almost always yielding uncertainties smaller than the statistical uncertainties .
combining all identified systematic uncertainties in a worst case scenario results in an upper limit @xmath4 . for comparison
we have also determined the best fit @xmath9 values for fixed @xmath1042 values yielding the relation @xmath5 .
the mass function has been integrated to obtain the fraction of the total gravitating mass in the universe contained in galaxy clusters .
normalized to the critical density we have found @xmath6 for cluster masses larger than @xmath7 . with the value for @xmath9 determined here
this implies that about 90% of the mass in the universe resides outside virialized cluster regions .
we are thankful for many enlightening discussions with peter schuecker , yasushi ikebe , and jrg retzlaff .
the authors acknowledge the great benefit from the use of unpublished catalogs from the reflex and noras teams .
yasushi ikebe has provided three cluster temperatures prior to publication .
volker mueller , jrg retzlaff and peter schuecker have provided results from @xmath8-body simulations .
marisa girardi has provided electronic tables of cluster masses based on galaxy velocity dispersions .
alexis finoguenov has provided electronic tables of partly unpublished x - ray mass measurements .
the bces regression software has been provided by michael akritas , tina bird , and matthew bershady .
the code for calculation of the universal mass function has been provided by adrian jenkins .
the code for calculation of the @xcite transfer function has been provided by these authors .
the authors thank the anonymous referee for useful comments on the manuscript .
the university of virginia has provided valuable computing resources .
yutaka fujita and alexis finoguenov have given helpful comments .
we acknowledge the dedicated work of the rosat and exsas hardware and software teams .
we have made extensive use of the rosat data archive of the max - planck - institut fr extraterrestrische physik ( mpe ) at garching , germany , and also of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , caltech , under contract with the national aeronautics and space administration .
this and related papers , supplementary information , as well as electronic files of the tables given in this paper are available at : and . , a. , ade , p. , bock , j. , borrill , j. , boscaleri , a. , de bernardis , p. , ferreira , p. g. , hanany , s. , hristov , v. , jaffe , a. h. , lee , a. t. , oh , s. , pascale , e. , rabii , b. , richards , p. l. , smoot , g. f. , stompor , r. , winant , c. d. , & wu , j. h. p. 2000 , apj , 545 , l1 , h. , collins , c. a. , guzzo , l. , schuecker , p. , voges , w. , neumann , d. m. , schindler , s. , chincarini , g. , de grandi , s. , cruddace , r. g. , edge , a. c. , reiprich , t. h. , & shaver , p. 2001
, a&a , submitted .
preprint : astro - ph/0106243 , h. , schuecker , p. , guzzo , l. , collins , c. a. , voges , w. , schindler , s. , neumann , d. m. , cruddace , r. g. , de grandi , s. , chincarini , g. , edge , a. c. , macgillivray , h. t. , & shaver , p. 2001
, a&a , 369 , 826 , l. n. , willmer , c. n. a. , pellegrini , p. s. , chaves , o. l. , rit , c. , maia , m. a. g. , geller , m. j. , latham , d. w. , kurtz , m. j. , huchra , j. p. , ramella , m. , fairall , a. p. , smith , c. , & lpari , s. 1998 , aj , 116 , 1 , p. , ade , p. a. r. , bock , j. j. , bond , j. r. , borrill , j. , boscaleri , a. , coble , k. , contaldi , c. r. , crill , b. p. , de troia , g. , farese , p. , ganga , k. , giacometti , m. , hivon , e. , hristov , v. v. , iacoangeli , a. , jaffe , a. h. , jones , w. c. , lange , a. e. , martinis , l. , masi , s. , mason , p. , mauskopf , p. d. , melchiorri , a. , montroy , t. , netterfield , c. b. , pascale , e. , piacentini , f. , pogosyan , d. , polenta , g. , pongetti , f. , prunet , s. , romeo , g. , ruhl , j. e. , & scaramuzzi , f. 2001 , apj , submitted .
preprint : astro - ph/0105296 , c. s. , white , s. d. m. , bode , p. , bond , j. r. , bryan , g. l. , cen , r. , couchman , h. m. p. , evrard , a. e. , gnedin , n. , jenkins , a. , khokhlov , a. m. , klypin , a. , navarro , j. f. , norman , m. l. , ostriker , j. p. , owen , j. m. , pearce , f. r. , pen , u. . , steinmetz , m. , thomas , p. a. , villumsen , j. v. , wadsley , j. w. , warren , m. s. , xu , g. , & yepes , g. 1999 , apj , 525 , 554 , a. h. , ade , p. a. r. , balbi , a. , bock , j. j. , bond , j. r. , borrill , j. , boscaleri , a. , coble , k. , crill , b. p. , de bernardis , p. , farese , p. , ferreira , p. g. , ganga , k. , giacometti , m. , hanany , s. , & et al .
2001 , phys .
lett . , 86 , 3475 , j. c. , cheng , e. s. , cottingham , d. a. , eplee , r. e. , fixsen , d. j. , hewagama , t. , isaacman , r. b. , jensen , k. a. , meyer , s. s. , noerdlinger , p. d. , read , s. m. , rosen , l. p. , shafer , r. a. , wright , e. l. , bennett , c. l. , boggess , n. w. , hauser , m. g. , kelsall , t. , moseley , s. h. , silverberg , r. f. , smoot , g. f. , weiss , r. , & wilkinson , d. t. 1994 , apj , 420 , 439 , j. r. , huchra , j. p. , freedman , w. l. , kennicutt , r. c. , ferrarese , l. , ford , h. c. , gibson , b. k. , graham , j. a. , hughes , s. m. g. , illingworth , g. d. , kelson , d. d. , macri , l. m. , madore , b. f. , sakai , s. , sebo , k. m. , silbermann , n. a. , & stetson , p. b. 2000 , apj , 529 , 786 , p. , bhringer , h. , guzzo , l. , collins , c. a. , neumann , d. m. , schindler , s. , voges , w. , de grandi , s. , chincarini , g. , cruddace , r. , mller , v. , reiprich , t. h. , retzlaff , j. , & shaver , p. 2001 , a&a , 368 , 86 , a. s. , jain , b. , matsubara , t. , scranton , r. , vogeley , m. s. , connolly , a. , dodelson , s. , eisenstein , d. , frieman , j. a. , gunn , j. e. , hui , l. , johnston , d. , kent , s. , kerscher , m. , loveday , j. , meiksin , a. , narayanan , v. , nichol , r. c. , oconnell , l. , pope , a. , scoccimarro , r. , sheth , r. k. , stebbins , a. , strauss , m. a. , szapudi , i. , tegmark , m. , zehavi , i. , sdss collaboration , & et al .
2001 , preprint : astro - ph/0107419 , w. , aschenbach , b. , boller , t. , bruninger , h. , briel , u. , burkert , w. , dennerl , k. , englhauser , j. , gruber , r. , haberl , f. , hartner , g. , hasinger , g. , krster , m. , pfeffermann , e. , pietsch , w. , predehl , p. , rosso , c. , schmitt , j. h. m. m. , trmper , j. , & zimmermann , h. u. 1999 , a&a , 349 , 389 , h. , boese , g. , becker , w. , belloni , t. , dbereiner , s. , izzo , c. , kahabka , p. , & schwentker , o. 1998 , exsas user s guide , 5th edn .
( garching , germany : mpe report july ) , http://wave.xray.mpe.mpg.de/exsas/users-guide lcccc bces(@xmath980 ) & 1.496 & 0.089 & -17.741 & 1.320 + bootstrap & 1.462 & 0.089 & -17.238 & 1.327 + bces(@xmath978 ) & 1.652 & 0.085 & -20.055 & 1.261 + bootstrap & 1.672 & 0.086 & -20.357 & 1.278 + bces - bisector & 1.571 & 0.083 & -18.857 & 1.237 + bootstrap & 1.562 & 0.083 & -18.717 & 1.239 + bces - orthogonal & 1.606 & 0.086 & -19.375 & 1.283 + bootstrap & 1.609 & 0.088 & -19.419 & 1.308 + lcccc bces(@xmath980 ) & 1.310 & 0.103 & -14.935 & 1.526 + bootstrap & 1.256 & 0.103 & -14.146 & 1.531 + bces(@xmath978 ) & 1.538 & 0.105 & -18.320 & 1.568 + bootstrap & 1.584 & 0.113 & -18.995 & 1.681 + bces - bisector & 1.418 & 0.097 & -16.536 & 1.434 + bootstrap & 1.407 & 0.096 & -16.368 & 1.427 + bces - orthogonal & 1.460 & 0.105 & -17.157 & 1.559 + bootstrap & 1.468 & 0.110 & -17.274 & 1.633 + lcccc bces(@xmath980 ) & 1.756 & 0.091 & -21.304 & 1.350 + bootstrap & 1.719 & 0.090 & -20.746 & 1.338
+ bces(@xmath978 ) & 1.860 & 0.084 & -22.836 & 1.246 + bootstrap & 1.881 & 0.084 & -23.144 & 1.250 + bces - bisector & 1.807 & 0.084 & -22.053 & 1.251 + bootstrap & 1.797 & 0.084 & -21.899 & 1.241 + bces - orthogonal & 1.835 & 0.085 & -22.473 & 1.260 + bootstrap & 1.841 & 0.085 & -22.563 & 1.270 + lcccc bces(@xmath980 ) & 1.504 & 0.089 & -17.545 & 1.298 + bootstrap & 1.469 & 0.089 & -17.042 & 1.300 + bces(@xmath978 ) & 1.652 & 0.086 & -19.708 & 1.254 + bootstrap & 1.671 & 0.086 & -19.992 & 1.260 + bces - bisector & 1.575 & 0.084 & -18.590 & 1.228 + bootstrap & 1.565 & 0.083 & -18.445 & 1.224 + bces - orthogonal & 1.609 & 0.087 & -19.075 & 1.274 + bootstrap & 1.611 & 0.088 & -19.109 & 1.290 + lcccc bces(@xmath980 ) & 2.683 & 0.215 & -35.675 & 3.219 + bootstrap & 2.606 & 0.218 & -34.525 & 3.265 + bces(@xmath978 ) & 2.488 & 0.127 & -32.761 & 1.902 + bootstrap & 2.520 & 0.130 & -33.239 & 1.946 + bces - bisector & 2.583 & 0.154 & -34.170 & 2.297 + bootstrap & 2.560 & 0.157 & -33.840 & 2.347 + bces - orthogonal & 2.513 & 0.128 & -33.137 & 1.915 + bootstrap & 2.530 & 0.132 & -33.388 & 1.975 + lccc @xmath958 & 0.21 & 0.21 & 0.21 + @xmath957 & 0.31 & 0.34 & 0.32 + @xmath959 & 0.17 & 0.18 & 0.17 + @xmath958 & 0.22 & 0.21 & 0.21 + @xmath957 & 0.29 & 0.32 & 0.30 + @xmath959 & 0.18 & 0.18 & 0.17 + | a new x - ray selected and x - ray flux - limited galaxy cluster sample is presented . based on the rosat all - sky survey the 63 brightest clusters with galactic latitude @xmath0deg and flux @xmath1 have been compiled .
gravitational masses have been determined utilizing intracluster gas density profiles , derived mainly from rosat ps .
pointed observations , and gas temperatures , as published mainly from asca observations , assuming hydrostatic equilibrium .
this sample and an extended sample of 106 galaxy clusters is used to establish the x - ray luminosity gravitational mass relation . from the complete sample the galaxy cluster mass function
is determined and used to constrain the mean cosmic matter density and the amplitude of mass fluctuations .
comparison to press
schechter type model mass functions in the framework of cold dark matter cosmological models and a harrison zeldovich initial density fluctuation spectrum yields the constraints @xmath2 and @xmath3 ( 90% c.l . ) .
various possible systematic uncertainties are quantified . adding
all identified systematic uncertainties to the statistical uncertainty in a worst case fashion results in an upper limit @xmath4 .
for comparison to previous results a relation @xmath5 is derived .
the mass function is integrated to show that the contribution of mass bound within virialized cluster regions to the total matter density is small , i.e. , @xmath6 for cluster masses larger than @xmath7 . |
quantum tunneling through nonstationary potential barriers is very unusual .
the problem was addressed in refs.@xcite and @xcite .
the method of complex classical trajectories was developed in refs.@xcite .
recent achievements in semiclassical theory are presented in refs .
@xcite . see also the related papers @xcite . in refs.@xcite
the advanced approach was developed to go beyond the method of classical trajectories and to obtain a space - time dependence of the wave function in semiclassical regime .
when a nonstationary field is very small the penetration of a particle through a barrier differs hardly from a conventional tunneling described by the semiclassical theory of wentzel , kramers , and brillouin ( wkb ) @xcite . at a relatively big nonstationary field
an over - barrier motion enters the game when a particle should absorb a number of quanta to reach the barrier top . in this paper
we consider soft nonstationary fields with a typical frequency much less than the barrier height .
this means that a particle has to absorb a big number of quanta to reach the barrier top . in the language of quantum mechanics
, this corresponds to a high order of the perturbation theory when the probability is proportional to a high power of the nonstationary field .
there are some intermediate magnitudes of a nonstationary field when neither pure tunneling nor pure over - barrier motion describes the penetration through a barrier . in this case , the real motion through a barrier is a combination of quanta absorption and tunneling .
the particle pays in probability to absorb quanta and to reach the certain higher energy level but the subsequent tunneling is easier since it occurs in a more transparent part of the barrier . that higher energy is determined by a minimization of the total probability .
this mechanism of barrier penetration in a nonstationary field is called photon - assisted tunneling .
the physics of photon - assisted tunneling has no conflict with intuition since the pay in an absorption probability is compensated by the gain in a probability of tunneling . in photon - assisted tunneling through a triangular barrier
two processes ( absorption and tunneling ) are weakly coherent which allows to consider them independently and , therefore , the total probability is a product of two partial ones .
this reminds a static tunneling through two barriers , separated in space , when quantum coherence between them is artificially destroyed by some external source . in this case ,
two processes ( tunneling and tunneling ) are also independent and the total probability is a product of partial ones .
besides quanta absorption , resulting in the increase of particle energy , also quanta emission is possible followed by tunneling with a lower energy . at the first sight
, this process can not lead to an enhancement of tunneling due to the double loss in probability ( i ) emission of quanta and ( ii ) tunneling in a less transparent part of the barrier ( with a lower energy ) .
this conclusion is based on the assumption that quanta emission and tunneling are not strongly coherent processes which allows to consider them almost independently .
the remarkable point of physics of nonstationary tunneling is that the processes of quanta emission and tunneling may be strongly _ coherent _ and can not be considered independently .
moreover , the conclusion was drawn @xcite that emission of quanta and tunneling with a lower energy may result in a strong enhancement of barrier penetration .
this conclusion is counterintuitive . indeed , intuition is sometimes useless in description of certain quantum mechanical processes .
for example , the property of non - reflectivity of some potentials @xcite can not be established on the basis of general arguments .
the strong coherence of quanta emission and tunneling is similar to the strong coherence of two stationary tunneling processes through a static double barrier potential . after tunneling through the first barrier the particle performs multiple reflections from the walls of two barriers and then tunnels through the second barrier . due to multiple coherent reflections
two tunneling processes become strongly coherent and the total penetration probability dramatically increases if the particle energy @xmath0 coincides with one of the energy levels @xmath1 in the well between two barriers .
the physical idea of this mechanism , called resonant tunneling , steams to wigner @xcite .
the strong coherence between quanta emission and tunneling also results in a resonant effect and the penetration probability , as a function of a particle energy @xmath0 , has a sharp peak at the certain energy @xmath1 determined by dynamical characteristics .
this effect is called euclidean resonance ( er ) @xcite .
the difference between a stationary resonant tunneling through a static double barrier potential and euclidean resonance in a dynamical barrier is that the former requires a long time for its formation but the latter occurs fast .
one has to emphasize that the frequency of a time - dependent field is much smaller than the barrier height and the amplitude of a time - dependent field is smaller that the static electric field . under these conditions the conventional tunneling and over - barrier transition would have very small probabilities which can be even classically small as , say , @xmath2 .
in contrast to this , under the er condition , @xmath3 , the peak in the energy dependence of probability becomes not classically small .
the phenomenon of euclidean resonance can occur in any physical process where tunneling is a substantial part .
in particular , er can be relevant in electron emission from materials .
another example is alpha decay of nuclei which occur due to tunneling of alpha particles through the coulomb potential barrier of nuclei @xcite .
an incident proton , colliding uranium nucleus , can stimulate alpha decay by its nonstationary coulomb field which leads to an energy exchange between the proton and the alpha particle @xcite .
this constitutes a new type of nuclear reactions .
the present paper addresses disintegration of atoms and molecules through tunnel barriers formed by coexisting constant and time - dependent electric fields .
there are two different channels of such disintegration , electronic and ionic .
the electronic mechanism is associated with ionization of a molecule into an electron and a positive ion .
this can occur in a wide range of frequencies of a time - dependent field between 100 mhz and infrared .
the ionic mechanism is due to dissociation of a molecule into two ions . since an ion is more massive than an electron , the necessary frequency is about of 1 mhz .
amplitudes and a frequency of applied fields can be tuned to meet er condition for the certain electron binding energy ( electronic mechanism ) or for the certain ion dissociation energy ( ionic mechanism ) .
this allows a very selective destruction of electron chemical bonds and a molecular dissociation only with respect to a given ion .
this last mechanism provides a theoretical possibility of a new method of isotope separation by radio frequencies .
the approach , used in the paper , is based on classical trajectories in imaginary time .
this method enables to determine a probability of barrier penetration only in the exponential approximation ( without a preexponential factor ) but the method provides a `` bypass '' of a complicated quantum dynamics .
suppose a particle with the energy @xmath0 to penetrate through the static potential barrier , shown in fig .
[ fig1](a ) , which is @xmath4 at @xmath5 and is zero at @xmath6 .
if the barrier is almost classical , the flux , transmitted to the right , is exponentially small and the particle mainly reflects from the barrier as shown by two arrows to the left from the barrier in fig .
[ fig1](b ) .
this relates to damping of the wave function in a classically forbidden region corresponding to the dominant branch in fig .
[ fig1](b ) . to provide only the outgoing wave at @xmath7 , shown by the arrow to the right from the barrier in fig .
[ fig1](b ) , the subdominant branch in fig .
[ fig1](b ) enters the game @xcite .
the total wave function under the barrier is a sum of two branches .
the both branches have the same order of magnitude at the exit point from under the barrier .
the barrier penetration depends on two remarkable parameters , the barrier height @xmath8 and the classical time @xmath9 of an under - barrier motion . for a weakly transparent barrier
the tunneling probability is small and is given by the wkb formula @xcite @xmath10\ ] ] where @xmath11 the time of an under - barrier motion , for the considered triangular barrier , is @xmath12
suppose now that the barrier is not static since its slope is modulated in time @xmath13 and @xmath14 at @xmath6 .
the amplitude @xmath15 is positive .
when the frequency is big , @xmath16 , the probability of the barrier penetration becomes not exponentially small due to over - barrier excitation of a particle @xcite .
a mostly interesting case corresponds to relatively small frequencies @xmath17 which are considered in the paper . according to quantum mechanical rules @xcite ,
the probability of absorption of one quantum of the field @xmath18 is proportional to @xmath19 where @xmath20 is of the order of the de broglie wave length .
a particle can absorb @xmath21 quanta and tunnel with a higher energy where the barrier is more transparent ( photon - assisted tunneling ) .
the probability of this process is a product of two ones @xmath22\ ] ] where the energy transfer is @xmath23 .
the simple separation ( [ 6 ] ) of the total quantum process into two independent mechanisms , generally speaking , is not accurate due to quantum interference between absorption and tunneling .
nevertheless , in the case of triangular barrier this separation is correct @xcite . at a very small amplitude @xmath15 of the nonstationary field eq .
( [ 6 ] ) can be considered as a perturbative contribution to the total penetration probability . since @xmath24 ,
the probability of tunneling , assisted by one photon , has the form @xmath25^{2}\exp\left[-a^{(0)}(e)\right]\ ] ] when @xmath26 the photon - assisted tunneling ( [ 6a ] ) dominates the conventional one ( [ 1 ] ) and multi - photon processes ( @xmath27 ) have to be accounted . in this case , to calculate a physical probability , one should find a maximum of the expression ( [ 6 ] ) with respect to @xmath28 .
the maximum relates to the balance between a lose of probability due to absorption and a gain of probability due to tunneling at a higher energy .
proper calculations are very simple .
an optimal number of absorbed quanta depends on the amplitude @xmath15 @xcite .
the optimal energy transfer @xmath28 increases with @xmath15 . under the condition ( [ 6b ] )
, @xmath28 reaches the value @xmath8 and the probability takes the form @xmath29\ ] ] the expression ( [ 7 ] ) is simply the probability to absorb @xmath30 quanta to reach the barrier top .
it is remarkable that at big @xmath31 tunneling processes are strongly influenced by the small amplitude ( [ 6b ] ) of a nonstationary field @xcite .
it is very instructive to formulate photon - assisted tunneling in terms of wave function .
an exact analytical solution of schrdinger equation with the nonstationary potential ( [ 3 ] ) is impossible .
nevertheless , in the limit of small frequency ( [ 4 ] ) one can express the wave function as a semiclassical series with respect to @xmath32 @xmath33\exp\left[\frac{i}{\hbar}s\left(x , t\right)\right].\ ] ] @xmath34 is the classical action which mainly determines the wave function if the semiclassical conditions @xmath35 hold @xcite .
this method was applied to investigation of photon - assisted tunneling in ref .
the result is unusual .
the semiclassical description ( [ 8 ] ) is valid almost at all times and the wave function is of the type shown in fig . [ fig1](b ) for the static case .
dramatic phenomena occur close to the moments @xmath36 when the potential barrier ( [ 3 ] ) in fig .
[ fig1](a ) is mostly narrow .
close to @xmath37 the semiclassical approach ( [ 8 ] ) breaks down at the dominant branch in fig .
[ fig3 ] within the dashed circle . during the short time interval @xmath38 , which can not be accounted by the semiclassical approximation ( [ 8 ] ) ( instant violation of semiclassical condition ) ,
the new branch of the wave function is formed shown in fig .
[ fig3 ] by the thick solid curve .
then , after an almost instant formation of the new branch , the semiclassical approach ( [ 8 ] ) recovers and the new branch in fig . [ fig3 ] corresponds to the wave packet moving towards big @xmath39 .
this process can be treated as an under - barrier instability caused by a nonstationary field . with no weak quantum effects of smearing , the maximum amplitude of the outgoing wave packet conserves and the packet moves as a classical particle . in the exponential approximation ( without a preexponential factor ) , the outgoing flux of particles is determined by a maximal value of the new branch in fig .
so , the maximum of the new branch plays an important role .
the outgoing flux consists of a sequence of wave packets , created at the moments ( [ 9 ] ) in the barrier region , which propagate towards big @xmath39 as classical particles .
the duration of each wave packet is shorter than @xmath40 but longer than the de broglie period of the order of @xmath38 @xcite .
this enables to describe wave packets by the energy @xmath41 , where @xmath28 is the energy transfer introduced in sec .
[ sec : phassisted ] a. analogously to results of ref .
@xcite , the maximum of the new branch in fig .
[ fig3 ] coincides with the value calculated on the basis of simple physical arguments given in sec .
[ sec : phassisted ] a. these simple arguments work only for truncated potential barriers as one given by eq .
( [ 3 ] ) @xcite . as one can see , the quantum dynamics of photon - assisted tunneling is not trivial .
a violation of the semiclassical approximation ( all terms in the series ( [ 8 ] ) become of the same order ) is unavoidable during short ( @xmath38 ) intervals of time even for a slow nonstationary field under the condition ( [ 4 ] ) . in this situation
it would be extremely useful to have some method of calculation of only a maximum of an outgoing wave packet , avoiding complications of the full quantum dynamics .
let us restrict ourselves only by a calculation of a maximal value of outgoing wave packets which relates to an extreme value of the classical action @xmath42 in eq .
( [ 8 ] ) .
according to classical mechanics , an extreme action relates to a classical trajectory .
but there are no conventional under - barrier classical trajectories since the positive value @xmath43 would be negative under the barrier .
the famous method to avoid this problem is to use imaginary time @xmath44 when @xmath45 becomes negative .
the method of complex time for photon - assisted tunneling was developed in refs.@xcite . in our case
one should find a classical trajectory @xmath46 connecting the points @xmath47 and @xmath48 as shown in fig .
[ fig3 ] by the dashed curve .
the classical trajectory satisfies newton s equation @xmath49=-{\cal e}_{0}-{\cal e}\cosh\omega\tau\ ] ] with the boundary conditions @xmath50 the three conditions ( [ 11 ] ) for second order differential equation ( [ 10 ] ) define the solution and the under - barrier time @xmath51 .
the moment @xmath52 relates to the point @xmath47 in fig .
the outgoing flux , associated with each generated wave packet , is proportional to the square of its amplitude and is given by the relation @xmath53 where the action has the form @xcite @xmath54-e\right\}\ ] ] the last term in eq .
( [ 13 ] ) corresponds to the return at @xmath6 from the time @xmath55 to @xmath56 according to the expression @xmath57 since at @xmath6 the nonstationary field does not act @xcite .
so , the action @xmath58 connects wave functions between the physical times @xmath56 and @xmath56 . the solution of eq .
( [ 10 ] ) should be substituted into eq .
( [ 13 ] ) .
the field @xmath18 in imaginary time goes over into @xmath59 which increases exponentially and influences the trajectory only close to @xmath51 .
the parameter @xmath31 is supposed to be not small .
so , the main under - barrier motion ( the dashed line in fig . [ fig2 ] ) is free with the energy @xmath41 . only close to @xmath51 ( in the vicinity of @xmath60 in fig . [ fig2 ] ) the particle looses its energy down to @xmath0 within a short interval of imaginary time .
for this reason , the process of energy transfer weakly contributes to the action ( [ 13 ] ) which is mainly collected from the dashed under - barrier line in fig .
[ fig2 ] where one can neglect the nonstationary field . in the expression ( [ 13 ] ) in the value of energy @xmath61 the first part relates to the free action @xmath62 and @xmath28 is an extra term . the action ( [ 13 ] )
can be written in the form @xmath63 the form ( [ 14 ] ) also follows from the exact equations ( [ 10 ] ) , ( [ 11 ] ) , and ( [ 13 ] ) which give in the limit ( [ 6b ] ) the time of under - barrier motion @xmath64 the energy transfer @xmath28 , given by the exact equations , coincides with the value obtained from minimization of the expression ( [ 14 ] ) @xcite .
as one can see , eqs .
( [ 12 ] ) and ( [ 14 ] ) are equivalent to eq .
( [ 6 ] ) followed from simple physical arguments .
an importance of a small time - dependent field , as in eq .
( [ 6b ] ) , simply follows from the imaginary time formalism where @xmath65 .
one can conclude from above that quantum dynamics of photon - assisted tunneling is complicated ( instant violation of semiclassical condition , generation of wave packets , etc . ) .
nevertheless , a calculation of probability of this process in the main approximation can be done on the basis of the relatively simple method of trajectories in imaginary time .
the trajectory provides a `` bypass '' ( the dashed curve in fig .
[ fig3 ] ) of the complicated quantum dynamics .
according to the under - barrier quantum dynamics , a creation of outgoing wave packets occurs at the moments @xmath37 when the nonstationary field reaches its maximal value .
an amplitude of wave packets at those moments also has maximum and can be calculated by the method of classical trajectories as an extreme value .
but besides the extremes at @xmath37 ( maxima of the field ) there are also other extremes of the nonstationary field at @xmath66 ( minima of the field ) where @xmath67 at the moments @xmath68 the barrier is mostly thick in accordance with eq .
( [ 3 ] ) .
one can put a reasonable question : what happens at the moments @xmath68 ?
close to the moments @xmath68 the nonstationary field ( [ 3 ] ) is opposite to the static component @xmath69 and the particle , moving under the barrier , loses its energy . at the first sight ,
the particle emits ( not absorbs as in fig . [ fig2 ] ) @xmath21 quanta with the probability @xmath70 and then tunnels with a lower energy . in this case
the probability would be given by eq .
( [ 6 ] ) with the negative @xmath71 where @xmath72 .
but this total probability has no extreme with respect to @xmath73 since the both mechanisms , an emission of quanta and tunneling with a lower energy , become less probable under the increase of @xmath73 . therefore the scenario , analogous to fig .
[ fig2 ] but with emission instead of absorption , is not realized .
this scenario would correspond to a weak interference between tunneling and an interaction with the nonstationary field . as follows from above
, one has to look for another mechanism related to a strong interference between quanta emission and tunneling . in the vicinity of the moments @xmath68
the similar phenomenon occurs as in the vicinity of @xmath56 @xcite .
close to the moments @xmath68 the semiclassical approximation ( [ 8 ] ) breaks down within the dashed circle in fig .
in contrast to the moments @xmath56 , this happens not at the dominant branch but at the subdominant one as shown in fig .
[ fig4 ] . the new branch , denoted in fig . [ fig4 ] by the thick solid curve ,
is also formed fast during the short time interval @xmath38 and then it starts to move away of the barrier as a wave packet .
the maximum value of the new branch in fig .
[ fig4 ] also plays an important role since it determines an outgoing flux with the exponential accuracy .
but now there is no simple arguments , as at @xmath56 , to determine that maximum since there is no representation of the type ( [ 6 ] ) . the maximum of the new branch in fig .
[ fig4 ] can be calculated directly from the classical action @xmath42 in the expansion ( [ 8 ] ) as it was done in @xcite for the different shape of a nonstationary field .
it is more convenient to determine the maximum of the new branch using the method of trajectories in imaginary time . the imaginary time technique for the moments @xmath68 is analogous to one developed in sec . [ sec : phassisted ] c. the classical trajectory @xmath74 now satisfies the equation @xmath75=-{\cal e}_{0}+{\cal e}\cosh\omega\tau\ ] ] with the boundary conditions ( [ 11 ] ) .
the classical trajectory has to be substituted into the classical action @xmath76-e\right\}\ ] ] which is analogous to ( [ 13 ] ) and connects wave functions between the physical times @xmath68 and @xmath68 .
the solution of equation ( [ 16 ] ) @xmath77 provides a connection of the points @xmath47 ( @xmath52 ) and @xmath48 shown by the dashed curve in fig .
[ fig4 ] . with the solution ( [ 18 ] )
the action ( [ 17 ] ) reads @xmath78\ ] ] the first of the boundary conditions ( [ 11 ] ) defines the under - barrier time @xmath51 and has the form @xmath79 the outgoing flux , associated with wave packets generated at the moments @xmath68 , has the form analogous to eq .
( [ 12 ] ) @xmath80 it is useful to look at the limit of big @xmath31 . in this case
the under - barrier time @xmath51 , given by eq . ( [ 18b ] ) , coincides , within the logarithmic accuracy , with the expression ( [ 14a ] ) .
analogously to sec .
[ sec : phassisted ] c , at big @xmath31 the process of the energy transfer occurs within a short interval of imaginary time and weakly contributes to the action @xmath81 which can be written , as ( [ 14 ] ) , in the form @xmath82 the energy transfer @xmath83 is determined by the condition @xmath84 where the exit point is @xmath85 .
this @xmath73 coincides with one followed from a minimization of the action ( [ 19 ] ) with respect to @xmath73 @xmath86 according to classical mechanics , the right - hand side of eq .
( [ 20 ] ) is a time of free under - barrier motion with the energy @xmath87 .
the energy transfer , followed from eq .
( [ 20 ] ) , is @xmath88\ ] ] in the limit of big @xmath31 , after a simple calculation within the logarithmic accuracy , one can obtain for the probability ( [ 18c ] ) @xmath89\ ] ] in eq .
( [ 22 ] ) the new remarkable energy is introduced @xmath90 which is called the resonant energy .
another form of eq .
( [ 22 ] ) is @xmath91 eq . ( [ 22a ] ) can be interpreted as a probability to reach a top of the new reduced barrier of the height @xmath1 . when the energy of the incident particle @xmath0 coincides with @xmath1 , the energy transfer , as follows from eqs .
( [ 21 ] ) and ( [ 23 ] ) , is @xmath92 and the exit point from under the barrier is determined by the relation @xmath93 under the condition @xmath3 the position of the circled point in fig .
[ fig4 ] is @xmath94 .
the probability ( [ 22 ] ) exhibits a new feature which is absent in photon - assisted tunneling ( the moments @xmath56 ) . at the energy
@xmath0 close to the resonant value @xmath1 the probability to penetrate the barrier becomes not exponentially small .
this phenomenon is called euclidean resonance since the motion occurs in imaginary time .
this name relates to special relativity where the metrics @xmath95 becomes euclidean in imaginary time @xmath44 .
eq.([22 ] ) is valid at @xmath96 , otherwise one has to use an approximation generic with a multi - instanton one .
the resonance condition @xmath97 can be considered from another point of view , namely , when it determines the certain amplitude of the time - dependent field @xmath98 as a function of a frequency and a static field @xmath69 . as follows from eqs .
( [ 18a ] ) and ( [ 18b ] ) , the resonant value of the amplitude @xmath99 , defined by the condition @xmath100 , in the limit of big @xmath31 is @xmath101\omega\tau_{00 } \exp\left(-\omega\tau_{00}\sqrt{3}\right)\ ] ] eqs .
( [ 21 ] ) - ( [ 23 ] ) are valid for amplitudes @xmath102 when @xmath15 is not very far from @xmath99 . at @xmath103
a multi - instanton approximation should be used in the vicinity of the resonant value @xmath99 . with the nonstationary amplitude ( [ 24 ] ) , the probability ( [ 22 ] ) becomes of the form @xmath104 $ ] that allows to estimate the resonance width in energy @xmath105
@xmath106 @xmath105 is small compared to the resonant energy @xmath1 .
the total outgoing flux of penetrated particles consists of wave packets created at the moments @xmath56 and @xmath68 .
this results in the certain flux @xmath107 , averaged in time , which can be experimentally measured as a steady outgoing flux .
close to the energy @xmath1 , @xmath107 is mainly determined by the wave packets created at the moments @xmath68 and therefore , within the exponential approximation used , one can write down @xmath108 . the probability @xmath107 is shown in fig .
[ fig5 ] . in semiclassical approximation , @xmath109 can not be less than a few of @xmath110 .
hence , the peak value of @xmath107 in fig .
[ fig5 ] is estimated , by means of eq .
( [ 22a ] ) , as @xmath111 where @xmath112 is a number of the order of unity .
the above semiclassical method does not allow to calculate the number @xmath112 accurately .
as one can see , the peak value of the probability ( [ 26 ] ) is small but not semiclassically small as the expression ( [ 1 ] ) . according to eq .
( [ 24 ] ) , at big @xmath31 the nonstationary amplitude @xmath113 is small .
therefore , it is better to choose a not very big @xmath31 . in this case
, one has to use the full equations ( [ 18a ] ) and ( [ 18b ] ) . at @xmath114
it follows that @xmath115 and @xmath116 . at @xmath117
one can obtain @xmath118 and @xmath119 .
one dimensional tunneling through a nonstationary barrier can be formulated as a numerical problem .
but performing the numerical calculation encounters a serious obstacle which is specific for euclidean resonance . since the new solution ( the thick curve in fig . [ fig4 ] ) is generated on the exponentially small branch , discrete steps in a numerical calculation should be chosen also exponentially small .
otherwise a calculation accuracy does not allow to resolve the exponentially small value , from which the instability develops , and the effect would be lost .
the choice of extremely small steps tremendously increases a calculation time which makes the numerical calculation non - realistic . for a more transparent barrier
this exponential value is not too small .
but , in such a not very semiclassical case , it is unclear whether or not the above semiclassical mechanism works .
to understand processes underlying photon - assisted tunneling and euclidean resonance it is useful to consider first the static problem of tunneling through the double barrier shown in fig .
the incident flux goes from the left and there is only the outgoing wave to the right from the barriers . if the both potential barriers would be not transparent there are discrete energy levels in the well between them . with a finite barrier transparency
each level goes over into a quasi - level with a finite width due to tunneling through the barriers .
suppose that in the space between two barriers in fig .
[ fig6]a some inelastic processes ( phonons , etc . ) influence the particle .
this results in a loss of coherence between two tunneling processes across barriers which become independent . in this case
the wave function @xmath120 of the intermediate incoherent state @xmath121 between two barriers in fig .
[ fig6]b is connected with the wave function of the initial state @xmath48 by the wkb tunneling condition @xmath122 . in the same way the state @xmath47 is linked with @xmath121 as @xmath123 .
the total probability is @xmath124 this total probability to overcome the both barriers is simply a product of two partial probabilities according to the independence of the both processes .
the situation is similar to photon - assisted tunneling , shown in fig .
[ fig2 ] , where the transition @xmath125 occurs due to quanta absorption ( the second term in ( [ 14 ] ) ) and the transition @xmath126 relates to tunneling ( the first term in ( [ 14 ] ) ) . as established in sec .
[ sec : phassisted ] , the both processes can be treated as incoherent ones and the total probability is a product of two partial probabilities .
the intermediate incoherent state @xmath121 in fig .
[ fig2 ] corresponds to some position at the barrier border with the exit energy . whereas the wave function for the static problem in fig .
[ fig6 ] exists in reality , in the nonstationary case this treatment relates to virtual states which arise in imaginary time . the process of absorption and subsequent tunneling in fig .
[ fig2 ] is mapped on the motion in imaginary time shown in fig .
[ fig3 ] by the dashed curve .
the point @xmath48 in figs .
[ fig2 ] and [ fig3 ] corresponds to the time @xmath68 and @xmath60 .
the point @xmath121 corresponds to the time @xmath127 and @xmath60 . the path @xmath125 in figs .
[ fig2 ] and [ fig3 ] relates to quanta absorption .
the point @xmath47 corresponds to @xmath68 and the path @xmath126 relates to tunneling .
the probability of photon - assisted tunneling is given by eq .
( [ 12 ] ) where , according to the form ( [ 14 ] ) , @xmath128 and @xmath129 .
so , two incoherent real processes of tunneling in fig .
[ fig6]b are analogous to virtual incoherent processes of absorption and tunneling which occur in imaginary time . when a particle motion through the static barriers in fig .
[ fig6]a is not influenced by any external source , a barrier penetration is very peculiar .
if a particle energy @xmath0 is close to one of the discrete levels @xmath1 between the barriers , the total tunneling probability strongly increases and can even reach unity in the case of symmetric barriers .
this phenomenon is called resonant tunneling . the idea of static resonant processes in quantum mechanics steams to wigner @xcite .
two tunneling processes across the barriers in fig .
[ fig6]a become very coherent and the total probability is not reduced to a product of partial ones .
let us consider the wave function shown in fig .
[ fig6]c . under the resonance condition @xmath3 , the static wave function in between the barriers is enhanced as in fig .
[ fig6]c giving rise to the resonant coherent state denoted as @xmath112 .
the resonant state @xmath112 can be connected with the wave function at the barrier entrance @xmath48 in the form @xmath130 .
analogously , the connection between @xmath112 and the exit point @xmath47 is @xmath131 .
the total probability of penetration the both barriers is @xmath132 positive parameters @xmath133 and @xmath133 are determined by an exact solution of the static quantum mechanical problem . away from the resonance , @xmath134 is a negative big value and the probability ( [ 27 ] ) is exponentially small .
close to the resonance @xmath3 the parameter @xmath134 reduces leading to a strong increase of the probability ( [ 28 ] ) . for symmetric barriers under the resonance condition @xmath135
, the total tunneling probability equals unity . in this case
the static resonant phenomenon can be formulated ( besides the direct way of energies coincidence @xmath3 ) as the equality of squared amplitudes @xmath136 for the transitions @xmath137 and @xmath138 in fig .
[ fig6]c .
a phenomenon of euclidean resonance has an an analogy with static resonant tunneling . in both cases
the tunneling probability has a sharp peak as a function of a particle energy at a resonant value @xmath1 . for resonant tunneling through a static barrier , @xmath1
is determined by energy levels in the potential well but for euclidean resonance @xmath1 is of a dynamical origin .
the coherent resonant state @xmath112 in fig .
[ fig7 ] is analogous to the state @xmath112 between two static barriers in fig .
[ fig6]c where the wave function is enhanced compared to the points @xmath48 and @xmath47 .
the resonant state @xmath112 in fig .
[ fig7 ] also corresponds to an enhanced wave function from where one can reach the point @xmath48 by emission of quanta with the probability @xmath139 or one can reach the point @xmath47 by tunneling with the probability @xmath140 $ ] .
the total probability in the case of euclidean resonance is given by eq .
( [ 28 ] ) which coincides with the eqs .
( [ 18c ] ) and ( [ 19 ] ) . as one can see , the condition of euclidean resonance is of the same type as for two static barriers , the coincidence of squared amplitudes @xmath137 and @xmath138 in fig .
[ fig7 ] .
the dashed curve in fig . [ fig4 ] relates to the processes in fig .
[ fig7 ] analogously to the case of photon - assisted tunneling .
the resonant state @xmath112 is reached at the moment @xmath127 and points @xmath48 and @xmath47 correspond to the moment @xmath68 .
whereas for the static tunneling in fig .
[ fig6]a the resonant state @xmath112 exists in reality , in the case of euclidean resonance the resonance state @xmath112 is virtual and exists in imaginary time .
the static resonant tunneling requires a long time for its formation ( before it becomes steady ) due to a slow process of filling in the inter - barrier space by tunneling leakage .
in contrast , in a dynamical barrier the er wave function is created fast .
the phenomenon of euclidean resonance can provide a method of ionization of atoms and molecules into an electron and a positive ion by applying a constant and a time - dependent fields with properly tuned amplitudes and a frequency .
the theory , developed above , can describe an electron tunneling from a bound state in an atom or a molecule through a potential barrier formed by applied electric fields . in the exponential approximation
the tunneling probability is the same for both the reflection problem in fig .
[ fig1 ] and for decay of an electron metastable state if it would situate at @xmath6 in fig .
[ fig1 ] ( instead of incident and reflected waves ) . in both cases
only the electron energy @xmath0 to the left of the barrier is relevant .
the main physical processes occur inside the triangular barrier far from an initial localized bound state of a particle . for this reason , a detailed barrier shape close to a molecule ( the region in the vicinity of @xmath60 ) is not very important and one can use the formalism developed above for a triangular barrier . as in the end of sec . [
sec : eucl]b , we choose the parameters so that @xmath117 . in this case , according to sec .
[ sec : eucl]b , the constant @xmath69 and nonstationary @xmath15 fields obey the relations ( the frequency @xmath141 is measured in ghz ) @xmath142 the electromagnetic energy flux @xmath143 , the under - barrier time @xmath51 , and the exit point @xmath144 from under the barrier are @xmath145 the distance @xmath144 plays a role of the under - barrier path giving the minimal size of an experimental setup .
the ionization rate , with the exponential accuracy ( without a preexponential factor ) , corresponds to the expression ( [ 26 ] ) which is generic with the exponent ( [ 22 ] ) .
the preexponential factor can be roughly estimated as the inverse time of under - barrier motion @xmath146 . by means of eq .
( [ 30 ] ) , the ionization rate becomes of the form @xmath147 the conditions of euclidean resonance in electron tunneling for different frequencies are collected in table [ tabl1 ] .
the under - barrier time @xmath51 and the width of euclidean resonance @xmath105 ( [ 25 ] ) are also included in table [ tabl1 ] .
the width @xmath105 has only a meaning when it exceeds a thermal smearing of the electron level in the well , otherwise the thermal smearing plays a role of the resonance width .
formation of euclidean resonance can be prevented by processes of energy loss under the barrier ( friction ) .
the main friction mechanism is a noise of a device which produces electric fields .
a small noise is not destructive . if the noise is not small it has to be of a relatively low frequency in order to prevent a destruction of euclidean resonance . in this case
the typical noise frequency of the electric field should be less than the inverse under - barrier time @xmath146 which are listed in table [ tabl1 ] for each frequency .
it is remarkable that under the above conditions , but without a nonstationary field , the tunneling probability ( [ 1 ] ) would be of the order of @xmath148 as for a typical classical barrier . in the opposite case of only a nonstationary field ( @xmath149 ) , the probability to overcome the barrier would be determined by the process of multi - quanta absorption to reach the barrier top . according to eq .
( [ 7 ] ) , the probability of this process is also of the same type as the above classically small value .
the cooperative action of the both fields , due to the strong quantum coherence , leads to the big enhancement of barrier penetration .
this is the mostly peculiar feature of euclidean resonance .
in big molecules various chemical bonds have , generally speaking , different binding energies of electrons @xmath8 . according to eq .
( [ 23 ] ) , at a fixed amplitude and a frequency of the time - dependent field , a resonant value of the constant field @xmath69 depends on the binding energy @xmath8 only .
this allows to adjust @xmath69 to provide solely a destruction of a certain bond via electron tunneling with no violation of other electron bonds in the molecule .
tunneling mechanism may result in dissociation of molecules into ions if to put a molecule in a constant electric field as shown in fig .
if to additionally switch on a nonstationary field one can encounter conditions for euclidean resonance .
let us consider a particular example of dissociation of a molecule nacl into two ions , @xmath150 and @xmath151 in fig .
[ fig8 ] . the ionization energy , which should substitute @xmath8 in the electron tunneling problem , is approximately @xmath152 ev . the reduced mass of the system of two ions is @xmath153 g. as in the previous case of electron tunneling
, one can take @xmath117 that leads to the estimates @xmath154 the electromagnetic power flux @xmath143 , the under - barrier time @xmath51 , and the exit point @xmath144 from under the barrier are given by the relations @xmath155 in eqs .
( [ 33 ] ) and ( [ 34 ] ) the frequency @xmath141 is measured in ghz .
( [ 33 ] ) and ( [ 34 ] ) allow to define a position of euclidean resonance for various frequencies of an external nonstationary field . for @xmath156 mhz ( radio frequency )
one can obtain @xmath157 v / cm , @xmath158 v / cm , @xmath159 , and @xmath160 cm .
vibrational levels in the potential well in fig .
[ fig8 ] for the molecule of nacl are separated by the energy @xmath161
ev@xmath162520 k. the above estimate relates to the ground state level .
suppose that an amplitude and a frequency of the nonstationary field are fixed and only a constant field @xmath69 can vary . then under a variation of @xmath69 one can satisfy the condition of euclidean resonance with respect to other energy levels corresponding to @xmath163 in eq .
( [ 23 ] ) .
this gives the positions of peaks of the dissociation rate in fig .
[ fig9 ] @xmath164 in our case the positions of er peaks of the dissociation rate are @xmath165 different amplitudes of the peaks are due to different thermal occupations of the vibrational levels . the magnitudes of the peaks in dissociation rate , can be roughly estimated as in the analogous problem of electronic ionization in the form @xmath166 . in our case
it turns into @xmath167 the width of peaks @xmath168 in fig .
[ fig9 ] can be defined from the relation @xmath169 .
the energy width @xmath170 k is given by eq .
( [ 25 ] ) and is much less than the natural width @xmath171 of vibrational levels determined by thermal processes .
therefore , the peak width in fig .
[ fig9 ] is given by @xmath172 .
in complicated molecules it is possible to tune field amplitudes and a frequency in order to destroy a certain vibrational molecular mode via ionic tunneling through the barrier . this is an ionic selective destruction of molecules in addition to the electronic mechanism considered in sec .
[ sec : ion ] .
one should emphasize that the radio frequency quantum used is @xmath173 times smaller than the distance @xmath174 between vibrational levels in fig .
the considered effect of dissociation is not due to a redistribution of levels occupation in the potential well in fig .
[ fig8 ] but due to under - barrier processes in presence of the time - dependent field . in the absence of a nonstationary field
the tunneling probability is classically small as @xmath175 .
the same can be concluded about the probability to absorb @xmath176 quanta to reach the barrier top in the absence of a constant field .
there are various methods of isotope separation based on diffusion , centrifugal forces , electromagnetic forces , laser irradiation , and chemical processes .
a phenomenon of euclidean resonance adds a new theoretical possibility of isotope separation ( by radio frequency waves ) to the previous list .
the method is based on induced dissociation of molecules considered in sec.[sec : molec ] . if the isotope mass shift is @xmath177 , then the relative shift of resonant values of the constant field in fig .
[ fig9 ] is @xmath178 . when @xmath179 the isotope shift of er peaks in fig .
[ fig9 ] is approximately 1 v / cm .
this enables to tune the system to ionize only the isotope of the mass @xmath180 .
the action of a nonstationary field on quantum tunneling is very non - trivial .
as soon as the nonstationary field is small , one can use a perturbation theory with respect to its amplitude and the total effect is simply reduced to an absorption or an emission of a few quanta and a subsequent tunneling .
a physical scenario becomes completely different with the increase of an amplitude of the nonstationary field .
methods of perturbation theory break down but indicate that multiquanta processes become relevant . in this case , as shown in the paper , the barrier penetration is realized as an emission of outgoing wave packets at the moments of time when the barrier becomes of an extremal width .
the mostly thin barrier relates to photon - assisted tunneling and the mostly thick one corresponds to euclidean resonance .
the quantum dynamic of the both physical processes is similar to some extend .
due to an under - barrier instability the wave packet is quickly formed and then it moves out as a classical particle .
nevertheless , there is an essential difference between the two processes .
suppose the static barrier to be classically non - transparent with the tunneling probability of the type @xmath2 .
then due to photon - assisted tunneling a particle increases its energy and exits from under the barrier with the probability , say , @xmath181 .
due to euclidean resonance a particle looses its energy but penetrates the barrier with the probability given by eq .
( [ 26 ] ) which is small but not classically small as the above numbers .
the regime of euclidean resonance is extremely unusual since a classical potential barrier can be penetrated by particles acted by a time - dependent electric field of a frequency which is much less compared to the barrier height .
moreover , an electron looses its energy and , at the first sight , seems to travel a longer way under the barrier .
this contradicts to the traditional imagination of under - barrier motion , a quanta absorption and a subsequent tunneling , based on a weak coherence between these processes .
the reason of this contradiction is that the phenomenon of euclidean resonance relates to a _
strong _ coherence between electromagnetic interaction and tunneling when a separation of a total process into almost independent emission and tunneling is impossible .
the method of classical trajectories in imaginary time , developed in the paper , enables to determine the magnitude of the effect with no detailed quantum mechanical calculations .
this method provides a `` bypass '' of the complicated quantum dynamics including the regions where semiclassical approximation breaks down . in the paper two tunneling mechanisms , electronic and ionic , of molecule destruction
are considered . the electronic destruction ( ionization )
relates to electron tunneling through a potential barrier , formed by a constant electric field , in presence of a time - dependent one . the ionic destruction ( dissociation )
is associated with the analogous process participated by an ion .
an ion is more massive than an electron and moves slower under the barrier .
therefore a smaller frequency ( roughly 1 mhz ) is required for the ionic dissociation than for the electronic ionization ( more than 100 mhz ) . in a complicated molecule
one can selectively destroy a particular electronic chemical bond since er peak position depends on a binding electron energy .
in addition to this , a selective ionic dissociation of molecules is possible by tuning external constant and time - dependent fields to a certain ionic motion .
since a position of er peak depends sensitively on an ion mass this provides a theoretical possibility of a new method of isotope separation if to tune properly field magnitudes and a frequency .
this method is unusual since radio frequencies of the order of 1 mhz have never been used for isotope separation .
generally speaking , the phenomenon of euclidean resonance is applicable to a variety of phenomena where quantum tunneling is a substantial part of physical processes .
in particular , one can consider a possibility of er application in scanning tunneling microscopy ( stm ) @xcite , in molecular electronics @xcite , nanoscience , for tunneling chemical reactions @xcite , and for decay of zero - voltage state of josephson junctions @xcite . on the basis of euclidean resonance
a prediction of a new type of nuclear reaction has been made @xcite .
i am grateful to g. berman , l. bulaevskii , j. engelfried , v. gudkov , s. gurvitz , m. kirchbach , j. knight , m. kunchur , l. levitov , r. prozorov , a. redondo , j. seminario , s. trugman , and v. tsifrinovich for valuable discussions . | the phenomenon of euclidean resonance ( a strong enhancement of quantum tunneling through a nonstationary potential barrier ) is applied to disintegration of atoms and molecules through tunnel barriers formed by applied constant and time - dependent electric fields .
there are two different channels for such disintegration , electronic and ionic .
the electronic mechanism is associated with the ionization of a molecule into an electron and a positive ion .
the required frequencies are in a wide range between 100 mhz and infrared .
this mechanism may constitute a method of selective destruction of chemical bonds .
the ionic mechanism consists of dissociation of a molecule into two ions .
since an ion is more massive than an electron , the necessary frequency is about 1 mhz .
this provides a theoretical possibility of a new method of isotope separation by radio frequency waves
. 0.5 cm |
let there be a set @xmath1 of points @xmath2 a geometrical object @xmath3 is a subset @xmath3 of points @xmath2 of @xmath1 , @xmath4 .
geometry @xmath5 is an infinite set @xmath6 of prepositions @xmath7 on properties of geometrical objects @xmath8 . if one can choose a finite or countable set @xmath9 of prepositions in such a way , that the infinite set @xmath6 of all prepositions @xmath7 on properties of geometrical objects @xmath8 can be obtained as a result of logical reasonings of prepositions @xmath10 , the geometry @xmath5 may be axiomatized , and the set @xmath11 of prepositions @xmath12 forms axiomatics ( the system of axioms ) of the geometry @xmath5 .
it is evident , that not any geometry can be axiomatized .
however , there are geometries , which can be axiomatized .
for instance , the proper euclidean geometry can be axiomatized .
let us consider a geometry @xmath5 , which can be axiomatized partly .
it means that a part @xmath13 of all propositions @xmath12 of geometry @xmath5 may be obtained as a result of logical reasonings of propositions @xmath10 , where @xmath11 is the set of axioms @xmath9 .
the set of axioms @xmath11 is supposed to be complete in the sense , that any supposition @xmath14 may be deduced from the set @xmath11 of axioms .
it is supposed that the set @xmath15 of remaining propositions of geometry @xmath5 can not be deduced from the set @xmath11 of axioms .
the godel s theorem proves the same , but godel starts from the supposition , that the complete axiomatization of geometry is possible , and he deduced essentially that the complete axiomatization is not always possible .
the real problem of the geometry construction lies in the fact , that we know no other method of the geometry construction except for the deduction of the geometry propositions from the geometry axioms .
euclid had deduced his euclidean geometry from the system of axioms .
it was proved @xcite , that the euclidean axioms are compatible between themselves , and the proper euclidean geometry is a consistent geometry .
mathematicians try to repeat the euclidean experience for construction of other ( non - uniform ) geometries .
they use the euclidean method of the geometry construction . as a result
one can not be sure that one deduces the full geometry , but not only its part .
we start from the general supposition that the complete axiomatization of a geometry is not always possible . in this case
the result of the godel s theorem is evident .
however , to adduce such a supposition , one is to have an alternative method of the geometry construction , because a formal consideration of the set of the geometrical propositions is not constructive without a method of these propositions obtaining .
in reality there is an alternative method of the geometry construction .
it is based on two statements . 1 .
geometry is described completely by its metric ( distance between any two points of the space ) .
2 . the proper euclidean geometry is a true geometry , which can be described completely by its metric .
we shall refer to geometry , which is completely described by its metric , as the physical geometry .
the geometry , which is deduced from the system of some axioms , will be referred to as an axiomatic geometry , or a mathematical geometry .
then the alternative method of the physical geometry construction is formulated as follows . _
any physical geometry is obtained from the proper euclidean geometry as a result of its deformation .
_ it means that all propositions of the proper euclidean geometry are expressed via euclidean metric , and the euclidean metric in all propositions is replaced by the metric of the geometry in question .
there is a theorem , which proves that all propositions of the proper euclidean geometry can be expressed via euclidean metric @xcite . as a result
one obtains all propositions of the geometry in question , expressed via its metric .
usually one supposes , that the term `` metric '' means the distance , satisfying the triangle axiom . in the previous presentation
the term `` metric '' means the distance simply ( which is not satisfies the triangle axiom , in general ) . to avoid a confusion ,
one uses the world function @xmath16 instead of the metric @xmath17 .
it is more convenient from technical point of view .
besides , in the case of indefinite geometry ( for instance , in geometry of minkowski ) the world function @xmath0 is real , although @xmath18 may be imaginary .
if one knows the proper euclidean geometry , the expression of the proper euclidean propositions via the euclidean world function is a purely technical problem .
but there are some subtleties in this problem .
the fact is that , the proper euclidean geometry has specific euclidean properties and general geometric properties .
all propositions of the proper euclidean geometry are to be expressed only via general geometric properties . only in this case
the expression via euclidean world function may be deformed and used in the deformed geometry .
the fact is that , the specific properties of the proper euclidean geometry are different in the euclidean spaces of different dimensions .
any specific property of the proper euclidean geometry contains a reference to the dimension @xmath19 of the euclidean space .
the general geometric propositions of the proper euclidean geometry do not refer to the dimension @xmath19 of the euclidean space .
thus , only the euclidean propositions , which do not contain a reference to the dimension @xmath19 , may be deformed to obtain corresponding relation of the deformed geometry .
we consider a simple example .
vector @xmath20 is an ordered set of two points @xmath21 and @xmath22 .
the scalar product @xmath23 of two vectors @xmath24 and @xmath25 is defined by the relation@xmath26where @xmath0 is the world function@xmath27definition ( [ a1.1 ] ) of the scalar product of two vectors coincides with the conventional scalar product of vectors in the proper euclidean space .
( one can verify this easily ) . the relation ( [ a1.1 ] )
does not contain a reference to the dimension of the euclidean space and to a coordinate system in it .
hence , the relation ( [ a1.1 ] ) is a general geometric relation , which may be considered as a definition of the scalar product of two vectors in any physical geometry .
equivalence ( equality ) of two vectors @xmath24 and @xmath25 is defined by the relations@xmath28where @xmath29 is the length of the vector @xmath24@xmath30 in the developed form the condition ( [ a1.3 ] ) of equivalence of two vectors @xmath24 and @xmath25 has the form@xmath31 if the points @xmath32 , determining the vector @xmath24 , and the origin @xmath33 of the vector @xmath25 are given , we can determine the vector @xmath25 , which is equivalent ( equal ) to the vector @xmath24 , solving two equations ( [ a1.5 ] ) , ( [ a1.6 ] ) with respect to the position of the point @xmath34 . in the case of the proper euclidean space
there is one and only one solution of equations ( [ a1.5 ] ) , ( [ a1.6 ] ) independently of its dimension @xmath19 . in the case of arbitrary physical geometry
one can guarantee neither existence nor uniqueness of the solution of equations ( [ a1.5 ] ) , ( a1.6 ) for the point @xmath34 .
number of solutions depends on the form of the world function @xmath0 .
this fact means a multivariance of the property of two vectors equivalence in the arbitrary physical geometry . in other words
, single - variance of the vector equality in the proper euclidean space is a specific property of the proper euclidean geometry , and this property is conditioned by the form of the euclidean world function . in other physical geometries
this property does not take place , in general .
the multivariance is a general property of a physical geometry .
it is connected with a necessity of solution of algebraic equations , containing the world function .
as far as the world function is different in different physical geometries , the solution of these equations may be not unique , or it may not exist at all . in the proper euclidean geometry the equality of two vectors @xmath24 and @xmath25 may be defined also as equality of components of these vectors at some rectilinear coordinate system .
( it is the conventional method of definition of two vectors equality ) .
let the dimension of the euclidean space be equal to @xmath19 .
let us introduce @xmath35 linear independent vectors @xmath36 , @xmath37 .
linear independence of vectors @xmath36 means that the gram s determinant @xmath38we construct rectilinear coordinate system with basic vectors @xmath36 , @xmath37 in the @xmath19-dimensional euclidean space .
covariant coordinates @xmath39 and @xmath40 of vectors @xmath24 and @xmath25 in this coordinate system have the form@xmath41equality of vectors @xmath24 and @xmath25 is written in the form@xmath42and according to ( [ a1.1 ] ) it may be written in terms of the world function ( metric ) . the points @xmath43 are supposed to be given .
the point @xmath34 is to be determined by equations ( [ a1.9 ] ) .
equations ( [ a1.9 ] ) determine equality of vectors @xmath24 and @xmath25 only in the @xmath19-dimensional euclidean space @xmath44 . already in the @xmath45-dimensional euclidean space
@xmath46 @xmath19 equations ( [ a1.9 ] ) do not determine equality of vectors @xmath24 and @xmath25 , because in @xmath45-dimensional euclidean space @xmath46 one needs @xmath45 equations of the form ( [ a1.9 ] ) to define equality of vectors .
there may be no dimension in the physical geometry , or the dimension of the space may be different at different points . in this
cases conventional conditions ( [ a1.9 ] ) of two vectors equality can not be used also . from formal point of view the equations ( [ a1.9 ] )
define some geometrical object , or a set of geometrical objects , whose points are described by means of running point @xmath34 .
this geometrical object depends on parameters @xmath47 .
how can one interpret this object ?
it is quite unclear . from formal viewpoint
the relations ( [ a1.5 ] ) , ( [ a1.6 ] ) describe some geometrical object by means of the running point @xmath34 .
this object depends on parameters @xmath48 , and one interprets this as a set of vectors @xmath25 , which are equivalent to vector @xmath24 .
one may consider , that the relations ( a1.5 ) , ( [ a1.6 ] ) describe some geometrical object by means of the running point @xmath49 .
this object depends on parameters @xmath50 , and one interprets this as a set of vectors @xmath24 , which are equivalent to vector @xmath25 .
it is to note that the proper euclidean geometry is a degenerate geometry in the sense that the same geometrical object may be described by different ways in terms of the world function .
for instance , the circular cylinder @xmath51 is defined by the relation@xmath52where @xmath32 are two different points on the axis of the cylinder and @xmath22 is some point on the surface of the cylinder .
the quantity @xmath53 is the area of the triangle with the vertices at the points @xmath54 .
the area of triangle may be calculated by means of the hero s , expressing the triangle area via length of the triangle sides , or by the formula@xmath55which may be expressed in terms of the world function by means of ( [ a1.1 ] ) . in the proper euclidean geometry
the circular cylinder @xmath51 depends only on its axis @xmath56 , passing through the points @xmath32 , but not on positions of the points @xmath32 on the axis @xmath56 .
the axis of the cylinder @xmath56 is described by the relation @xmath57where @xmath58 means that the vectors @xmath24 and @xmath59 are collinear ( linear dependent ) , which means mathematically , that @xmath60 thus , if the point @xmath61 , @xmath62 , @xmath63 , then the straight line @xmath64 in the euclidean space , but @xmath65 in the physical geometry , in general .
then@xmath66 in the proper euclidean geometry .
but in general,@xmath67 in other words , the circular cylinder of the proper euclidean space is split into many different cylinders after deformation of the euclidean space .
a more accurate statement is as follows .
cylinders @xmath68 and @xmath69 are different , in general .
but in the proper euclidean geometry they may coincide , even if @xmath70 but @xmath71 .
it means , that in the proper euclidean geometry different geometrical objects may coincide , because of very high symmetry of the euclidean geometry . in other words
, a deformation of the euclidean geometry may violate its symmetry , and coincidence of different geometrical objects ceases .
in application to the property of equivalence the multivariance property looks as follows . in general , there are many vectors @xmath25,@xmath72 which are equivalent to vector @xmath24 and are not equivalent between themselves .
this situation is common for all physical geometries .
however , there are possible such geometries , where the set of vectors @xmath25,@xmath72 degenerates into one vector . in the proper euclidean geometry
we have such a degeneration for all vectors @xmath24 and any point @xmath33 , which is an origin of the vector @xmath25 , equivalent to the vector @xmath24 .
it means that the multivariance of the equivalence property is a general property of a physical geometry , whereas single - variance of the equivalence property in the proper euclidean geometry is a specific property of the euclidean geometry , which is conditioned by the form of the euclidean world function .
the multivariance is a new property of physical geometry .
multivariance properties have not been yet investigated properly .
multivariance of the equivalence property generates intransitivity of the vector equivalence . in other words , if @xmath24eqv@xmath25 and @xmath25eqv@xmath73 , then , in general , @xmath24 is not equivalent to @xmath73 .
the intransitive equivalence is difficult for investigation .
it is reasonable to consider physical geometries , which do not contain multivariance , or contain multivariance in the minimal degree .
the form of the world function is a unique characteristic of a physical geometry .
one can change a physical geometry only changing its world function . to obtain the riemannian geometry we are to impose on the world function
the following constraint @xmath74it is the triangle axiom .
its meaning is as follows .
the segment @xmath75 } $ ] of the straight between the points @xmath76 is described by the relation @xmath77 } = \left\ { r|f_{3}\left ( p_{0},r , p_{1}\right ) = 0\right\ } \label{a1.18}\]]in general , the equation @xmath78describes some surface @xmath79 around some volume @xmath80 in the space @xmath1 .
the external points @xmath81 with respect to the volume @xmath80 satisfy the relation @xmath82 .
the internal points @xmath81 inside the volume @xmath80 satisfy the relation @xmath83 . if the triangle axiom ( [ a1.17 ] ) is fulfilled , the volume @xmath80 is empty , and the segment @xmath75 } $ ] of straight is one - dimensional , because the surface @xmath79 does not contain any points inside .
on the other hand , the segment ( [ a1.18 ] ) of the straight can be described by the relation @xmath77 } = \left\ { r|\mathbf{p}_{0}\mathbf{p}_{1}\upuparrows \mathbf{p}_{0}\mathbf{r\wedge } \left\vert \mathbf{p}_{0}\mathbf{r}\right\vert \leq \left\vert \mathbf{p}_{0}\mathbf{p}_{1}\right\vert \right\ } \label{a1.20}\]]where @xmath84 is the condition of parallelism of vectors @xmath24 and @xmath85 , which is described by the relation @xmath86 it easy to verify that two definitions ( [ a1.18 ] ) , ( [ a1.19 ] ) and ( a1.20 ) , ( [ a1.21 ] ) are equivalent because of the identity@xmath87where@xmath88 thus , if the world function satisfies the triangle axiom ( [ a1.17 ] ) , any segment ( [ a1.20 ] ) is single - variant ( one - dimensional ) , and equivalence ( [ a1.3 ] ) of two vectors @xmath24 and @xmath85 is single - variant , provided they have a common origin @xmath89 .
it means that it follows from the relation @xmath24eqv@xmath85 and ( [ a1.17 ] ) * , * that @xmath90 .
note , that the single - variance of the two vectors equivalence takes place only for the proper riemannian geometry , when the world function is nonnegative and all terms in the relation ( [ a1.17 ] ) are real for any points @xmath91 . for the pseudo - riemannian geometry , when the world function @xmath0 may have any sign , the equivalence of two vectors is multivariant , in general , even if the vectors have a common origin .
investigation shows , that the equivalence of two vectors is single - variant only in the flat proper riemannian space , i.e. in the proper euclidean space .
note , that in @xcite one investigated multivariance of parallelism of two directions , the concept of the vector equivalence had not yet been introduced .
however , there is a single - valued connection between the multivariance of two vectors equivalence and multivariance of the two direction parallelism .
there are some special cases of the geometry and of vectors , when the equivalence property is single - variant .
for instance , in the proper riemannian space , when the world function is nonnegative and the triangle axiom ( [ a1.17 ] ) takes place , equivalence of two vectors @xmath24 and @xmath92 is single - variant , provided that @xmath93 .
the physical geometry , whose world function satisfies the triangle axiom ( [ a1.17 ] ) will be referred to as the @xmath0-riemannian geometry .
the riemannian geometry distinguishes from the @xmath0-riemannian geometry by the method of its construction .
the @xmath0-riemannian geometry is constructed by means of the deformation principle , and it is defined completely by its world function . the @xmath0-riemannian geometry may be discrete , or continuous .
this circumstance is of no importance , because the construction of physical geometry by means of the deformation principle does not need introduction of a coordinate system .
the @xmath19-dimensional riemannian geometry is constructed as an internal geometry of the @xmath19-dimensional smooth surface @xmath95 in @xmath96-dimensional proper euclidean space @xmath97 @xmath98 .
one introduces a curvilinear coordinate system @xmath99 in @xmath97 with coordinates @xmath100 @xmath101 , where @xmath102 , @xmath103 .
coordinates are introduced in such a way , that coordinates @xmath104 describe points of the @xmath19-dimensional surface @xmath95 .
besides , coordinates @xmath105 are supposed to be chosen in such a way , that any vector @xmath106 at a point @xmath107 is orthogonal to any tangent to the surface @xmath95 vector @xmath108 , taken at the same point @xmath107 .
let @xmath109 , @xmath110 be the metric tensor in the proper euclidean space @xmath97 in the coordinate system @xmath99 .
then@xmath111where @xmath112 @xmath113 are cartesian coordinates of the point @xmath105 in @xmath97 .
curvilinear coordinates are chosen in such a way , that on the surface @xmath95 they satisfy the conditions@xmath114 the line element @xmath115 on the surface @xmath95 is described by the relation@xmath116where the sign of sum is omitted and the summation over repeated indices is produced from @xmath117 to @xmath19 .
this rule is used further .
one can determine geodesics on the surface @xmath95 by means of the relations@xmath118where @xmath119 is the cristoffel symbol in the coordinate system @xmath120 on the surface @xmath95@xmath121
@xmath122let for simplicity there be only one geodesic segment @xmath123 } \subset s_{n}$ ] , connecting any two points @xmath124 .
we can define the world function @xmath125 on @xmath95 by means of the relation@xmath126 } } \sqrt{g_{ik}\left ( \mathbf{x}\right ) dx^{i}dx^{k}}\right ) ^{2},\qquad p_{0},p_{1}\in s_{n } \label{a4.5}\]]according to definition ( [ a4.5 ] ) the world function @xmath127 satisfies the triangle axiom ( [ a1.17 ] ) .
it means , that the world function @xmath125 of the riemannian geometry may coincide with the world function of the @xmath0-riemannian geometry , if the set @xmath1 is identified with the surface @xmath95 .
then we may repeat construction of the riemannian geometry on the surface @xmath95 by means of the deformation principle . in this case
we obtain the above obtained results .
we obtain the line element in the form ( [ a4.3 ] ) and equation for the straight ( geodesic ) in the form ( [ a4.4 ] ) .
all obtained single - variant results of the riemannian geometry may be obtained by means of the deformation principle from the world function ( [ a4.5 ] ) .
two vectors @xmath128 and @xmath129 at two different points @xmath130 and @xmath131 are equal in the euclidean space @xmath97 , if their cartesian components coincide@xmath132where @xmath133 and @xmath134 are components of vectors @xmath128 and @xmath129 respectively in the curvilinear coordinate system @xmath99 . if vectors @xmath128 and @xmath129 are vectors of the internal geometry in @xmath95 , they are tangent to the surface @xmath95 , i.e.@xmath135 let the vector @xmath128 , satisfying the first relation ( [ a4.8 ] ) be fixed .
then the vector @xmath129 , satisfying conditions ( [ a4.7 ] ) , ( [ a4.8 ] ) does not exist , in general , if the points @xmath49 and @xmath136 are different , and @xmath137but the multivariant relations of the @xmath0-riemannian geometry can not be obtained in the riemannian geometry , because the riemannian geometry does not contain multivariant relations in principle .
thus , in general , we can not take from the proper euclidean geometry the concept of the remote vectors equality in the internal geometry of the surface @xmath95 .
if the riemannian geometry is considered as an abstract logical construction , one may at all not introduce equality of two remote vectors and their parallelism . in this case
we obtain a geometry , which has only some concepts of the proper euclidean geometry , but not all of them .
the obtained geometry appears to be more poor in concepts , than the proper euclidean geometry . however , if the ( pseudo-)riemannian geometry pretends to be used for description of the space - time geometry , one is forced to introduce these concepts , because a construction of particle dynamics is impossible without a conception of equality of remote vectors .
the parallel transport is introduced in the riemannian geometry as follows .
let @xmath138 @xmath37 be a vector on the surface @xmath95 .
simultaneously the vector @xmath139 is a vector in the proper euclidean space @xmath97 .
its coordinates in the coordinate system @xmath99 have the form @xmath140 @xmath141 @xmath142 @xmath143 .
let @xmath144 be an infinitesimal vector of displacement on the surface @xmath95 , @xmath145 .
let us transport the vector @xmath146 in @xmath147 from the point @xmath148 into the point @xmath149 .
the transport is produced in such a way , that @xmath150 .
it is always possible in the proper euclidean space @xmath97 . as far as @xmath151
is infinitesimal quantity , in the coordinate system @xmath99 the vector @xmath152 has the form @xmath153where @xmath154 is an infinitesimal quantity of the order @xmath155 . as far as @xmath156 , @xmath143
, we obtain @xmath157and @xmath158 coincide with @xmath159 to within @xmath160 .
let us project the vector @xmath152 onto the surface @xmath95 .
it means that we set @xmath161 .
calculation of @xmath162 gives@xmath163where the christoffel symbol @xmath164 is given on the surface @xmath95 by the relation ( [ a4.6 ] ) .
the relation ( [ a4.12 ] ) contains only tangential components @xmath165 of the vector @xmath166 .
it means that the relation ( [ a4.12 ] ) is a relation of the internal geometry on the surface @xmath95 .
this relation may be described in the form , containing only quantities of the internal geometry .
the vector @xmath167is considered to be in parallel with the vector @xmath168 , @xmath37 .
it is well known relation for the parallel transport of a vector in the riemannian geometry .
the parallel transport from point @xmath169 to the point @xmath170is produced as follows .
the points @xmath169 and @xmath171 are connected by some line @xmath172 .
the line is divided into infinitesimal segments .
the parallel transport is produced step by step along all segments by means of the formula ( [ a4.14 ] ) .
result of the parallel transport depends on the path @xmath172 .
essentially the result of the parallel transport is multivariant , but each version of the parallel transport is connected with some path @xmath172 of the transport . the parallel transport in @xmath0-riemannian geometry was investigated in sec.6 of paper @xcite . here
we present only the result of this investigation .
the parallel transport ( parallelism of two vectors ) is defined by the relation ( [ a1.21 ] ) . in the coordinate form
it is written as follows@xmath173where @xmath174 is the world function of the @xmath0-riemannian geometry between the points with coordinates @xmath175 , @xmath176 and @xmath177 , @xmath176 @xmath178prime at the index means that this index corresponds to the point @xmath179 .
@xmath180 is a contravariant vector at the point @xmath181 .
@xmath182 is a contravariant vector at the point @xmath183 .
if vectors @xmath184 and @xmath185 are in parallel , they satisfy the relation ( [ a4.15 ] ) .
the relation ( [ a4.15 ] ) has been obtained for infinitesimal vectors @xmath180 and @xmath182 .
but the relation ( [ a4.15 ] ) is invariant with respect to a change of lengths of vectors @xmath180 and @xmath186 , and it appears to be valid also for finite vectors @xmath187 , @xmath182 . at the deduction of the relation ( [ a4.15 ] ) one uses the fact that the @xmath0 is the world function , defined by the relation ( [ a4.5 ] ) , i.e. it is a world function of the riemannian space . let the vector @xmath188 be infinitesimal .
then @xmath189where @xmath190 and @xmath191 are infinitesimal quantities , and the relation ( [ a4.15 ] ) may be transformed to the form@xmath192where @xmath193it is easy to verify , that the conventional parallel transport ( [ a4.14 ] ) @xmath194satisfies the relation ( [ a4.18 ] ) .
however , the solution ( [ a4.20 ] ) is unique , if direction @xmath144 of transport coincides with the direction of the vector @xmath195 . in the general case ,
when the curvature of the surface @xmath95 does not vanish , the set of solutions @xmath196 of equation ( [ a4.18 ] ) forms a cone ( the collinearity cone ) .
this cone degenerates into one - dimensional line , if @xmath195 is in parallel with @xmath144 * * , * * or if the curvature of the surface @xmath95 vanishes . in these cases
the conventional parallel transport ( [ a4.20 ] ) is the unique solution of ( [ a4.18 ] ) .
one should expect , that the world function determines geometry uniquely . at any rate the euclidean world function determines the euclidean geometry uniquely .
if the world function satisfies the triangle axiom ( [ a1.17 ] ) , it determines the @xmath0-riemannian geometry uniquely . however , there is in addition the riemannian geometry which does not coincide with the @xmath197-riemannian geometry in some details .
the riemannian geometry pretends to be a true geometry , describing the real space - time . at any rate ,
most of contemporary mathematicians consider the riemannian geometry as a true geometry and ignore the @xmath0-riemannian geometry .
we compare the @xmath0-riemannian geometry and the riemannian geometry , firstly , as logical constructions and , secondly , as possible space - time geometries .
the riemannian geometry is rather special geometry .
it is described only in terms of coordinates .
the riemannian space is isometrically embeddable in the euclidean space of rather large dimension .
the riemannian geometry uses the triangle axiom ( [ a1.17 ] ) , taken in the form ( [ a4.5 ] ) , as internal constraint of a geometry .
it is internal in the sense , that some basic concepts of the riemannian geometry ( concept of a curve ) can not formulated at all without the triangle axiom .
the concept of the world function is a secondary concept in the riemannian geometry .
the concept of the world function , defined by the relation ( [ a4.5 ] ) , can not be formulated without a reference to the concept of a curve ( geodesic ) .
it is impossible to obtain a generalization of the riemannian geometry in terms of its basic concepts .
the @xmath0-riemannian geometry is a special case of the physical geometry , when the world function is restricted by the triangle axiom ( a1.17 ) .
the world function is a primary concept of the @xmath0-riemannian geometry .
the @xmath0-riemannian geometry is not constrained by such conditions as a use of coordinates and isometric embeddability in the euclidean space .
the @xmath0-riemannian geometry does not use such non - metrical concept as the concept of a curve .
the triangle axiom is an external constraint in the @xmath0-riemannian geometry in the sense , that it is not used in construction of the geometry . avoiding the triangle axiom , we obtain a more general geometry , whose concepts are constructed without a reference to the triangle axiom .
the @xmath0-riemannian geometry is a more general construction than , the riemannian one in the sense , that imposing some constraints on the @xmath0-riemannian geometry , one may to obtain the riemannian geometry .
for instance , in the physical geometry there are two sorts of straights @xmath198and@xmath199 the straights of the type ( [ a4.21 ] ) are single - variant ( one - dimensional ) in the @xmath0-riemannian geometry , whereas the straight of the type ( a4.22 ) are multivariant , in general .
the straight of the type ( [ a4.21 ] ) is a special ( degenerate ) case of the straight ( [ a4.22 ] ) , when the point @xmath33 coincides with the point @xmath89 . in the proper euclidean geometry
the straights of both types are single - variant ( one - dimensional ) . to obtain the riemannian geometry from the @xmath0-riemannian geometry
, one needs to remove the multivariant straights of the type ( [ a4.22 ] ) and to use only degenerate one - dimensional straights of the type ( [ a4.21 ] ) .
thereafter the degenerate straights ( [ a4.21 ] ) are used for introduction of the concept of a geodesic .
the world function is constructed as a secondary concept on the basis of the infinitesimal line element and of the geodesic .
the world function is not used at the construction of the riemannian geometry .
the riemannian geometry is constructed as an internal geometry of a surface in the proper euclidean space .
eliminating multivariant straights ( [ a4.22 ] ) from the riemannian geometry and leaving only degenerate straights , one can not be sure that the way of riemannian geometry construction , based on a use of only degenerate straights ( geodesics ) , is consistent .
a use of geodesics in the construction of geometry leads to the fact , that two - dimensional euclidean plane with a hole can not be isometrically embedded to the euclidean plane without a hole .
besides , in the riemannian geometry the absolute parallelism is absent , although it takes place in the @xmath0-riemannian geometry .
thus , it seems , that the riemannian geometry is only a part of the full physical geometry .
the remaining part of the full geometry is cut by the constraint , that the degenerate straights ( [ a4.21 ] ) form the complete set of straights .
this constraint is an internal constraint , which is used at the construction of the riemannian geometry .
it can not be removed without destruction of the riemannian geometry . in application to the space - time the @xmath0-riemannian geometry is more effective , than the riemannian geometry .
the @xmath0-riemannian geometry is a more general geometry . to obtain a more general space - time geometry from the @xmath0-riemannian geometry , it sufficient only to remove the triangle axiom , which is an external constraint .
we obtain immediately a space - time geometry of a general form . in the riemannian geometry
the triangle axiom is an internal constraint , which is used at the construction of the riemannian geometry . removing the triangle axiom ,
we lose the method of the riemannian geometry construction . using the riemannian geometry as a space - time geometry
, one can not imagine , that the space - time geometry may be responsible for quantum effects r91 .
one can not imagine that the space - time geometry may be responsible for a limited divisibility of physical bodies @xcite .
the unlimited divisibility , used in the riemannian space - time geometry , generates an independence of the particle dynamics on the space - time geometry . in reality
, the true space - time geometry must determine the particle dynamics @xcite .
yu . a.rylov tubular geometry construction as a reason for new revision of the space - time conception .
( printed in _
what is geometry ?
_ polimetrica publisher , italy , pp.201 - 235 http://www.polimetrica.com/polimetrica/406/ ) , see also http;//afxiv.org / abs / physics/0504031 | not any geometry can be axiomatized .
the paradoxical godel s theorem starts from the supposition that any geometry can be axiomatized and goes to the result , that not any geometry can be axiomatized .
one considers example of two close geometries ( riemannian geometry and @xmath0-riemannian one ) , which are constructed by different methods and distinguish in some details .
the riemannian geometry reminds such a geometry , which is only a part of the full geometry . such a possibility is covered by the godel s theorem . |
with the conclusive evidence for a significant component of cold dark matter ( cdm ) in the universe , there is considerable interest , both at the theoretical and experimental level , to identify this cdm and analyse its properties . in particular ,
if the cdm consists of a new weakly interacting massive particle ( wimp ) , for example the lightest neutralino in supersymmetric models , the next generation of colliders has good prospects to discover it and determine its properties .
if the parameters of the underlying model can be measured with sufficient precision , it may be possible to predict the annihilation cross sections hence the thermal relic density of the cdm candidate .
the aim is to reach an accuracy comparable to the one from cosmological measurements to be able to perform consistency checks between a particular model of new physics and cosmology @xcite .
the recent data from wmap and sdss imply a value for the relic density of cold dark matter @xmath6
@xcite at 2@xmath7 .
range as of 2003 , @xmath8 @xcite and refer to this as the wmap range .
note that the exact value and uncertainty of @xmath0 extracted from cosmological data depend on both the precise data sets used and the assumptions about other parameters @xcite . ]
most of the studies that have either analysed the constraints on the parameter space of the mssm arising from the relic density measurement , or that have analysed how to extract precise information on the model parameters at future colliders have assumed that cp is conserved .
however cp - violating phases are generic in the mssm
. furthermore cp violation may help to generate the correct baryon asymmetry in the universe in scenarios of electroweak baryogenesis @xcite .
cp phases can have a strong impact on the relic density of neutralino dark matter , both due to modifications in the sparticle couplings and due to changes in the physical masses . here
we highlight the impact of cp phases in a few typical scenarios for which the lsp is a ` good ' cdm candidate @xcite . some of the largest effects are in fact due to kinematics .
this should be expected as the relic density is often very sensitive to masses , in particular to the exact mass difference between the lsp and nlsp in coannihilation processes , or in the case of annihilation near a @xmath9-channel higgs resonance , to the difference between twice the lsp mass and the mass of the higgs . in these scenarios , setting apart the purely kinematic effects hence somewhat tames the huge effects due to phases found in some of the early studies @xcite . on the other hand
, there are other cases where the phase dependence of masses and couplings work against each other .
taking out the kinematic effects actually enhances the phase dependence of the number density . since what are relevant to experiments are rather the physical observables ( masses , branching ratios , etc . ) than the underlying parameters , we take special care to disentangle effects arising from changes in the couplings from purely kinematic effects .
in the mssm , the parameters that can have cp phases are the gaugino and higgsino mass parameters , @xmath10 and @xmath11 , and the trilinear sfermion - higgs couplings , @xmath12 . for the relic density of the lsp
the relevant phases are those of the neutralino sector ( only @xmath13 and @xmath14 since @xmath15 can be rotated away ) and the phase of the trilinear coupling @xmath16 which affects the higgs sector .
the parameters that will be allowed to vary are those of the mssm defined at the weak scale @xmath17 where universality at the gut scale for the gaugino masses is assumed .
@xmath18 is the common mass for third generation sfermion , the common mass for the sfermions of the first and second generation is set to @xmath19 tev and all trilinear couplings with the exception of @xmath16 are set to @xmath20 tev . the phase of the selectron coupling @xmath21 , although irrelevant for the relic density has to be taken into account since it contributes to electric dipole moments ( edms ) .
allowing for cp - violating phases induces a mixing between the two cp - even states @xmath22 , @xmath23 and the cp - odd state @xmath24 of the mssm .
the resulting mass eigenstates @xmath25 ( @xmath26 ) are no longer eigenstates of cp , therefore it is preferable to use the charged higgs mass , @xmath27 , as an independent parameter .
the lagrangian for the interaction of the lightest neutralino with higgs bosons which governs the neutralino annihilation cross section via higgs exchange , writes @xmath28 with the scalar and pseudoscalar couplings corresponding to the real and imaginary part of the same expression , see @xcite . these couplings depend on the neutralino mixing , hence on phases in the neutralino sector , as well as on phases that enter the higgs mixing , for example @xmath29 .
indeed , in the mssm , the higgs cp mixing is induced by loops involving top squarks and is proportional to @xmath30 @xcite .
thus a large mixing is expected when @xmath31 is large as compared to the square of the stop masses .
the results presented here have been obtained with the new implementation of the cpv - mssm within 2.0 @xcite . in this code , cp phases are taken into account consistently in all annihilation and coannihilation channels .
the computation of masses , mixings and effective couplings in the higgs sector relies on cpsuperh @xcite .
standard routines are used to calculate the relic density of dark matter @xcite .
the cross sections for the annihilation and coannihilation processes will depend on phases , and so will the thermally - averaged cross section @xcite
. part of this is due to changes in the physical masses , leading to huge variations in the relic density especially when coannihilation processes are important or when annihilation occurs near a resonance .
we will therefore take special care to disentangle the effects from kinematics and couplings . at vanishing relative velocity
, @xmath32 , neutralino annnihilation through @xmath9-channel scalar exchange is @xmath33-wave suppressed ; the annihilation proceeds strictly through pseudoscalar exchange .
nevertheless when performing the thermal averaging , the scalar exchange can not be neglected altogether . in the mssm with real parameters
it can amount to @xmath34 of the total contribution . in the presence of phases
, all the neutral higgs bosons can acquire a pseudoscalar component ( that is @xmath35 ) and hence significantly contribute to neutralino annihilation even at small @xmath36 .
there is a kind of sum rule that relates the couplings squared of the higgses to neutralinos .
therefore , for the two heavy eigenstates which are in general close in mass , we do not expect a large effect on the resulting relic density from higgs mixing alone .
a noteworthy exception occurs when , for kinematical reason , only one of the resonances is accessible to neutralino annihilation .
that is for example the case when @xmath37 .
we now turn to the numerical analysis and present results for two typical scenarios for which the relic density is in agreement with wmap : the mixed bino - higgsino lsp that annihilates into gauge bosons and the rapid annihilation through a higgs resonance , the so - called higgs funnel .
more comprehensive results can be found in @xcite .
we always impose a constraint from the eedm , @xmath38 @xcite , including one- and two - loop contributions . when @xmath39 is not too large and sfermions of the first and second generations are heavy , this constraint usually forces @xmath40 .
in addition , the free parameter @xmath21 is adjusted so that this constraint is satisfied . when all scalars except the light higgs are heavy , @xmath41 tev , the most efficient annihilation of a pair of neutralinos with a mass of the order of 100 gev is through its higgsino component . in the real mssm
, one needs a higgsino admixture of roughly 25%30% for the relic density to be within the wmap range @xcite . in terms of fundamental mssm parameters
this means @xmath42 .
figure [ fig : m1mumh1000]a displays the allowed 2@xmath7 wmap band in the @xmath43 plane .
the main annihilation mechanisms then are @xmath44 through @xmath45-channel chargino and neutralino exchange , as well as @xmath46 when kinematically allowed . the latter proceeds through @xmath9-channel @xmath47 or @xmath48 exchange .
the lsp higgsino fraction determines the size of the annihilation cross - section because it directly enters both the @xmath49 and @xmath50 vertices .
wmap bands for @xmath51 , @xmath52 tev a ) in the @xmath53@xmath54 plane for @xmath55 tev and for all phases zero ( blue ) , for @xmath56 ( or @xmath57 ) and all other phases zero ( dashed red lines ) and for arbitrary phases ( green ) . , ( b ) @xmath53@xmath13 plane for @xmath58 gev , @xmath59 tev , @xmath60 . in the yellow region
@xmath0 is below the wmap bound , and in the white region it is too large .
superimposed are contours of constant lsp mass ( dashed lines ) and contours of constant lsp higgsino fraction ( dash - dotted lines).,title="fig : " ] wmap bands for @xmath51 , @xmath52 tev a ) in the @xmath53@xmath54 plane for @xmath55 tev and for all phases zero ( blue ) , for @xmath56 ( or @xmath57 ) and all other phases zero ( dashed red lines ) and for arbitrary phases ( green ) . ,
( b ) @xmath53@xmath13 plane for @xmath58 gev , @xmath59 tev , @xmath60 . in the yellow region
@xmath0 is below the wmap bound , and in the white region it is too large .
superimposed are contours of constant lsp mass ( dashed lines ) and contours of constant lsp higgsino fraction ( dash - dotted lines).,title="fig : " ] when allowing all phases to vary arbitrarily , thus increasing the number of free parameters , it is natural to expect a widening of the allowed band , see the green ( light grey ) band in fig .
[ fig : m1mumh1000]a where we display models for which all constraints are satisfied for at least one combination of phases .
since the eedm constraint strongly constrains @xmath14 , the widening of the allowed band is mostly due to @xmath13 .
figure [ fig : m1mumh1000]b shows the wmap band in the @xmath53@xmath13 plane for @xmath58 gev , @xmath61 tev and all other phases set to zero . also shown
are contours of constant lsp mass as well as contours of constant lsp higgsino fraction @xmath62 .
we can make several observations .
first , the mass of the lsp increases with @xmath13 . on the one hand
this induces a decrease in the lsp pair - annihilation cross - sections . on the other hand ,
since the chargino mass is independent of @xmath13 , the nlsp lsp mass splitting is reduced , making coannihilation processes with @xmath63 ( and also @xmath64 ) more important .
second , the higgsino fraction decreases with increasing @xmath13 .
this modifies the lsp couplings to gauge bosons and leads to a decrease in the dominant @xmath44 cross - sections and thus a higher value for the relic density .
this phase dependence in the couplings is predominantly what determines the shift in the value of the relic density , to wit the almost perfect match between contours of constant @xmath0 and those of constant @xmath62 in fig .
[ fig : m1mumh1000]b .
the deviation near @xmath65 comes from chargino / neutralino coannihilations . to isolate the effect that comes solely from modifications in the couplings , we display in fig . [
fig : mlsp140 ] the variation of @xmath0 as function of @xmath13 for constant lsp mass , as compared to the variation of @xmath0 for constant @xmath53 .
the former is more pronounced reaching almost an order of magnitude . in this scenario with a mixed bino - higgsino lsp , the dependence of masses and couplings on @xmath13 work against each other , so that taking out the kinematic effects actually enhances the variation of @xmath0 . as a function of @xmath13 for the parameters of fig .
[ fig : m1mumh1000]b ; the dashed line is for fixed @xmath66 gev , while for the full line @xmath53 is adjusted such that @xmath67 gev .
the green ( grey ) band shows the @xmath68 wmap - allowed range . ] in the higgs sector , nonzero phases can induce scalar - pseudoscalar mixing as well as important changes in the masses .
moreover , the scalar and pseudoscalar couplings of the higgses to the lsp depend on the phases in the neutralino sector .
one can therefore expect large differences between the real and complex mssm in the higgs - funnel region . since the relic density is very sensitive to the mass difference @xmath69 @xcite , it is important to disentangle the phase effects in kinematics and in couplings . for a case study of the higgs funnel
, we choose @xmath70 leading to a dominantly bino lsp and small mixing in the higgs sector for @xmath71 . in this scenario with relatively small @xmath54 ,
varying the phase of @xmath29 does not have a large impact since the scalar - pseudoscalar mixing never exceeds 8% .
the main effect of @xmath29 on the value of the relic density can be explained by shifts in the physical masses and position of the resonance . on the other hand ,
the phase @xmath13 changes the neutralino masses and mixing , and hence also the neutralino
higgs couplings , eq .
( [ eq : gs ] ) . in fig .
[ fig : at]a we show the wmap - allowed regions in the @xmath27@xmath13 plane for @xmath72 . for a given value of @xmath27 , when increasing @xmath13 , the relic density drops
this is because the mass of the neutralino increases slowly , resulting in a smaller @xmath73 .
adjusting @xmath74 or @xmath75 such that the mass difference stays constant , we find rather that the relic density increases with @xmath13 , fig .
[ fig : at]b .
the maximum deviation which comes purely from modifications in the couplings can reach 70% .
this can be readily understood from the phase dependence of the couplings of @xmath76 to the lsp . for @xmath77 ,
the coupling of @xmath78 is predominantly pseudoscalar and thus gives the dominant contribution to neutralino annihilation . for @xmath79 ,
the coupling of @xmath80 has a large pseudoscalar component and also contributes to the annihilation .
when increasing @xmath13 further ( up to @xmath81 ) , @xmath78 exchange again dominates , however with a coupling to neutralinos smaller than for @xmath77 .
thus one needs a smaller mass splitting @xmath82 for @xmath0 to fall within the wmap range , see fig .
[ fig : at]a .
wmap bands ( green / dark grey ) in the @xmath27@xmath83 planes for the parameters of eq .
( [ eq : funnelpar ] ) and @xmath72 . contours of constant mass differences @xmath84 are also displayed . in the yellow ( light grey ) region
, @xmath0 is below the wmap range .
b ) @xmath0 as a function of @xmath13 for @xmath85 gev and the value of @xmath53 adjusted so that @xmath82 stays constant ( full line ) . for comparison ,
the variation of @xmath0 for fixed @xmath86 gev is shown as a dashed line.,title="fig : " ] wmap bands ( green / dark grey ) in the @xmath27@xmath83 planes for the parameters of eq .
( [ eq : funnelpar ] ) and @xmath72 .
contours of constant mass differences @xmath84 are also displayed . in the yellow ( light grey ) region
, @xmath0 is below the wmap range .
b ) @xmath0 as a function of @xmath13 for @xmath85 gev and the value of @xmath53 adjusted so that @xmath82 stays constant ( full line ) . for comparison , the variation of @xmath0 for fixed @xmath86 gev
is shown as a dashed line.,title="fig : " ] when the mixing is large in the higgs sector , say for @xmath87tev , the phase @xmath29 affects the scalar - pseudoscalar mixing as well as the mass splitting between @xmath88 .
the latter can reach several gev s .
this means that @xmath29 can induce very large shifts in the relic density which very sensitively depends on the mass difference between the lsp and the predominantly pseudoscalar higgs .
isolating the phase dependence of @xmath0 due to the scalar - pseudoscalar mixing by keeping @xmath89 constant , we found an increase in @xmath0 relative to the @xmath72 case by almost an order of magnitude @xcite .
this is however far less than the huge shifts of several orders of magnitude found in ref .
@xcite for fixed values of @xmath27 when a higgs pole is passed .
we have shown here that in the cpv - mssm , the relic density could be quite different as compared to that in the mssm with vanishing phases , the variations in @xmath0 often exceeding the @xmath90 range of the wmap bound . in many cases , a large part of the phase dependence
can be explained by changes in the masses of the involved particles .
however , in some cases disentangling the kinematic effects also leads to an enhancement of the phase dependence .
this happens , for instance , in the case of a mixed bino - higgsino lsp , where we have found effects of almost an order of magnitude from modifications in the couplings due to nonzero cp phases .
when aiming at a precise prediction of the neutralino relic density from collider measurements , it is clear that one does not only need precise sparticle spectroscopy one also has to precisely measure the relevant couplings and this certainly includes the determination of possible cp phases .
whether parameters of the cpv - mssm can be determined with sufficient precision at the lhc or at a future linear collider requires careful investigation .
this work was supported in part by gdri - acpp of cnrs and by grants from the russian federal agency for science , ns-1685.2003.2 and rfbr-04 - 02 - 17448 . s.k .
is supported by an apart ( austrian programme for advanced research and technology ) grant of the austrian academy of sciences . | in the mssm , the presence of cp phases can have a significant impact on the neutralino relic abundance .
these phase effects are on the one hand due to shifts in the masses , on the other hand due to modifications of the couplings .
typical variations in @xmath0 solely from modifications in the couplings are @xmath1(10%100% ) .
lapth - conf-1161/06 + cern - ph - th/2006 - 187 * neutralino dark matter in the mssm with cp violation * + g. blanger@xmath2 , f. boudjema@xmath2 , s. kraml@xmath3 , a. pukhov@xmath4 , a. semenov@xmath5 + _ 1 ) lapth , 9 chemin de bellevue , b.p .
110 , f-74941 annecy - le - vieux , france + 2 ) cern , dept . of physics , theory division , ch-1211 geneva 23 , switzerland + 3 ) skobeltsyn inst . of nuclear physics , moscow state univ . ,
moscow 119992 , russia + 4 ) joint institute for nuclear research ( jinr ) , 141980 , dubna , russia _ + |
energy - momentum is an important conserved quantity whose definition has been a focus of many investigations .
unfortunately , there is still no generally agreed definition of energy and momentum in general relativity ( gr ) .
this dilemma in gr is highlighted in an important paper by penrose@xcite in the following way : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ it is perhaps ironic that _ energy conservation _ , paradigmatic physical concept arising initially from galileo s ( 1638 ) studies of the motion of bodies under gravity , and which now has found expression in the ( covariant ) equation @xmath0 a cornerstone of einstein s ( 1915 ) general relativity should nevertheless have found no universally applicable formulation , within einstein s theory , incorporating the energy of gravity itself .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ indeed , conservation laws of energy - momentum together with the equivalence principle , played a significant role in guiding einstein s search for his generally covariant field equations .
the numerous attempts aimed at finding an expression for describing energy - momentum distribution due to matter , nongravitational and gravitational fields gave rise to a large number of energy - momentum complexes whose physical meaning have been questioned ( see @xcite ) .
the absence of a generally accepted definition of energy distribution in curved space - times has led to doubts regarding the idea of energy localization . over the past two decades considerable effort
has been put in trying to define an alternative concept of energy , the so - called quasi - local energy .
the idea in this case is to determine the effective energy of a source by measurements on a two - surface .
these masses are obtained over a two - surface as opposed to an integral spanning over a three - surface of a local density as is the case for pseudocomplexes . to date a large number of definitions of quasi - local mass
have been proposed ( see in brown and york @xcite and hayward@xcite ) .
bergqvist@xcite considered quasi - local mass definitions of komar , hawking , penrose , ludvigsen - vickers , bergqvist - ludvigsen , kulkarni _ et al .
_ , and dougan - mason and concluded that no two of these definitions give agreed results for the kerr as well as reissner - nordstrm space - times . on the contrary ,
the pioneering work of virbhadra and his collaborators ( notably , nathan rosen of einstein - rosen gravitational waves fame ) , and others have demonstrated with several examples that for a given spacetime , the energy - momentum complexes of einstein , landau and lifshitz , papapetrou , and weinberg ( ellpw ) , show a high degree of consistency in giving the same and acceptable energy and momentum distributions ( see for instance @xcite ) .
_ et al__@xcite showed that the ellpw energy - momentum complexes `` coincide '' for any metric of kerr - schild class .
the stationary axially symmetric and asymptotically flat electrovac kerr - newman ( kn ) solution , characterized by constant parameters mass @xmath1 , charge @xmath2 , and angular momentum @xmath3 , is the most general black hole solution to the einstein - maxwell equations .
virbhadra @xcite and then following him cooperstock and richardson @xcite showed ( up to the third order and seventh order , respectively , of rotation parameter ) that the energy - momentum complexes of einstein and landau - lifshitz give the same and reasonable energy distribution in kn space - time .
further aguirregabiria _
@xcite performed exact computations for the energy distribution in kn space - time in kerr - schild cartesian coordinates .
they showed that the energy distribution in the prescriptions of einstein , landau - lifshitz , papapetrou , and weinberg ( ellpw ) gave the same result . in this paper
we investigate the energy - momentum distribution of the kerr - newman space - time using the bergmann - thomson prescription and compare the results obtained by aguirregabiria _
in this section we first write the energy - momentum complex formulated by bergmann and thomson and then use it to compute energy and momentum distributions in kerr - newman spacetime . the energy - momentum complex of bergmann - thomson is@xcite @xmath4 where @xmath5 with @xmath6_{,m } \ . \label{uikl}\ ] ] the latin indices take values @xmath7 to @xmath8 .
the bergmann - thomson energy - momentum complex satisfies the local conservation laws @xmath9 in any coordinate system ; however , @xmath10 itself does not transform as a tensor under a general coordinate transformation .
the energy and energy current ( momentum ) density components are respectively represented by @xmath11 and @xmath12 .
energy and momentum components @xmath13 are given by @xmath14 using gauss s theorem one obtains energy @xmath15 and momentum components @xmath16 ( @xmath17 takes values @xmath18 to @xmath8 ) as follows : @xmath19 the kerr - newman space - time in kerr - schild cartesian coordinates @xmath20 is expressed by the line - element@xcite : @xmath21 ^ 2 \text{,}\label{knmetric}\end{aligned}\ ] ] where @xmath22 is defined by @xmath23 note that @xmath22 becomes the usual spherical radial coordinate @xmath24 for rotation parameter @xmath25 .
now we use equations @xmath26 and @xmath27 to obtain energy and momentum distributions in kerr - schild cartesian coordinates . in these coordinates , all the components of the metric tensor for the kerr - newman metric are nonvanishing and also have lengthy expressions .
it is extremely difficult to perform these calculations manually .
therefore , we used _
mathematica 4.0_@xcite for computations .
the energy and momentum components inside a 2-surface with constant @xmath22 are respectively given by @xmath28 \label{enenrgy}\ ] ] and @xmath29 it is striking that bergmann - thomson complex gives the same result as found by aguirregabiria _
et al._@xcite ) who used definitions proposed by einstein , landau and lifshitz , papapetrou , and weinberg .
the results found by us further support the importance of the energy - momentum complexes , as these yield consistent and meaningful results .
now for our convenience we define @xmath30 and plot @xmath31 against @xmath32 and @xmath33 for different values of @xmath34 ( see fig .
1 ) . for the same value of @xmath32 and @xmath33
, @xmath31 is higher for lower values of @xmath34 .
linear momentum components @xmath35 remain zero for all values of @xmath36 , and @xmath32 .
this is also shown in the same figure .
note that the kerr - newman represents gravitational fields of black holes for @xmath37 and of naked singularities for @xmath38 .
the main weakness of energy - momentum complexes is that these restrict one to make calculations in ` cartesian coordinates ' .
the alternative concept of quasi - local mass appears more attractive because these are not restricted to the use of any special coordinate system .
there is a large number of quasi - local masses .
it has been shown@xcite that for the kerr as well as reissner - nordstrm space - times many quasi - local mass formulations do not give agreed results . on the other hand aguirregabiria , chamorro , and
virbhadra@xcite showed that the energy - momentum complexes of einstein , landau - lifshitz , papapetrou , and weinberg give the same result for any metric of kerr - schild class if the computations are carried out in kerr - schild cartesian coordinates .
the well - known spacetimes of the kerr - schild class are for example the schwarzschild , reissner - nordstrm , kerr , kerr - newman , vaidya , dybney _ et al .
_ , kinnersley , bonnor - vaidya and vaidya - patel .
virbhadra@xcite also showed that for a general nonstatic spherically symmetric spacetime of the kerr - schild class , the ellpw give the same energy distribution as the penrose quasi - local mass definition if calculations are performed in kerr - schild cartesian coordinates . in the previous section we showed that the bergmann - thomson energy - momentum complex gives the same energy and momentum distribution in the kerr - newman spacetime as the ellpw prescriptions .
these results and the important paper of chang _ _ et al__@xcite , which dispels doubts expressed about the physical meaning of these energy - momentum complexes , give confidence to the use of energy - momentum complexes to compute energy - momentum distribution in a given spacetime .
the energy expression @xmath39 in eq .
@xmath40 can be treated as the effective gravitational mass that acts on a neutral ( electric charge zero ) test particle situated at a coordinate distance @xmath22 from the centre of the kerr - newman black hole .
the repulsive effects of the electric charge and rotation parameters are obvious . for sufficiently small @xmath22
, @xmath39 can be negative for large values of @xmath41 and @xmath3 .
for reissner - nordstrm metric , @xmath42 for @xmath43 . as expected , the total energy ( @xmath44 ) of the kerr - newman metric comes to be the well - known adm mass @xmath1 .
it is independent of the other two ( electric charge and rotation ) parameters of the charged rotating black hole . a most important implication of our result ( expressed in eq .
@xmath40 ) is that it supports _ cooperstock hypothesis _
@xcite for energy localization , which essentially states that energy in a curved space - time is localized in the region of nonvanishing energy - momentum tensor @xmath45 of matter and non - gravitational fields . for the kerr black hole space - time @xmath46 and
therefore energy is confined to exterior of the black hole ; however , for the kerr - newman black hole metric @xmath47 and the energy is distributed to the interior and exterior of the black hole .
i am very grateful to nrf of s. africa for financial support and professor t. dube of the university of zululand for encouragement during the work .
r. penrose , _ proc .
. london _ * a381 * , 53 ( 1982 ) .
l. d. landau and e. m. lifshitz , _ the classical theory of fields _ , pergamon press , p. 280
a. papapetrou , _ proceedings of the royal irish academy a _ , * 52 * , 11 ( 1948 ) .
s. weinberg , _ gravitation and cosmology : principles and applications of general theory of relativity _ , ( john wiley and sons , inc . , new york , 1972 ) p. 165 .
p. g. bergmann and r. thomson , _ phys .
_ * 89 * , 400 ( 1953 ) .
mller , c. , _ annals of physics _
( ny ) * 4 * , 347 ( 1958 ) .
s. chandrasekhar and v. ferrari , _ proc .
* a435 * , 645 ( 1991 ) .
j. d. brown and j. w. york , jr .
rev . _ * d47 * , 1407 ( 1993 )
. s. a. hayward , _ phys .
rev . _ * d49 * , 831 ( 1994 ) .
g. bergqvist , _ class .
quantum gravit . _ * 9 * , 1753 ( 1992 )
. k. s. virbhadra , _ pramana - j .
_ * 38 * , 31 ( 1992 ) ; _ ibid . _
* 44 * , 317 ( 1995 ) ; _ ibid . _
* 45 * , 215 ( 1995 ) ; int .
a12 , 4831 ( 1997 ) .
n. rosen and k. s. virbhadra , _ gen .
_ * 25 * , 429 ( 1993 ) ; a. chamorro and k. s. virbhadra , _ pramana - j .
phys . _ * 45 * , 181 ( 1995 ) ; _ int . j. mod
* d5 * , 251 ( 1996 ) ; k. s. virbhadra and j. c. parikh , _ phys .
lett . _ * b317 * , 312 ( 1993 ) ; _ phys . lett . _ * b331 * , 302 ( 1994 ) .
s. s. xulu , _ int . j. mod
. phys . _ * a15 * , 2979 ( 2000 ) ; _ ibid . _
* a15 * , 4849 ( 2000 ) ; _ int .
_ * 39 * , 1153 ( 2000 ) ; gr - qc 0304081 .
s. s. xulu , _ mod .
lett . _ * a15 * , 1511 ( 2000 ) .
i. radinschi , _ mod .
_ * a15 * , 2171 ( 2000 ) ; _ ibid . _ * a16 * , 673 ( 2001 ) ; _
nuovo cim .
_ * b115 * , 501 ( 2000 ) ; i - c yang and i. radinschi , _ mod.phys.lett . _
* a17 * 1159 ( 2002 )
. t. bringley , _ mod .
_ * a17 , * 157 ( 2002 ) .
r. m. gad , gr - qc 0306101 , gr - qc 0307010 .
e. c. vagenas , gr - qc 0303009 , hep - th 0307162 j. m. aguirregabiria , a. chamorro and k. s. virbhadra , _ gen .
_ * 28 * , 1393 ( 1996 ) .
k. s. virbhadra , _ phys . rev .
_ * d41 * , 1086 , ( 1990 ) ; _ ibid _ * d42 * , 1066 ( 1990 ) ; _ ibid . _ * d42 * , 2919 ( 1990 ) ; _ ibid _ * d60 * , 104041 ( 1999 ) .
cooperstock , f. i. and richardson , s. a. , in _ proc .
4th canadian conf . on general relativity and relativistic astrophysics _ ( world scientific , singapore , 1991 ) .
g. c. debney , r. p. kerr , and a. schild , j. math .
* 10 * , 1842 ( 1969 ) .
s. wolfram , mathematica 4.0 .
c. c. chang , j. m. nester and c. chen , _ phys .
* 83 * , 1897 ( 1999 ) .
cooperstock , f. i. , _ mod .
phys lett .
_ * a14 * , 1531 ( 1999 ) ; _ annals of phys .
_ * 282 * , 115 ( 2000 ; faraoni , v. and cooperstock , f. i. , _ astrophys . j. _ * 587 * , 483 ( 2003 ) ; cooperstock , f. i. and tieu , s. , _ gr - qc/0302020_. | we obtain the energy distribution in the kerr - newman metric with the help of bergmann - thomson energy - momentum complex . we find that the energy - momentum definitions prescribed by einstein , landau - lifshitz , papapetrou , weinberg , and bergmann - thomson give the same and acceptable result and also support the _ cooperstock hypothesis _ for energy localization in general relativity
. the repulsive effect due to the electric charge and rotation parameters of the metric is also reflected from the energy distribution expression . |
cellular membranes are two - dimensional fluid interfaces that consist of a large variety of components .
they form the boundary between the cell and the outside world , and , for eukaryotic cells , separate the inside of the cell into numerous compartments known as organelles . in order for biological processes like cell division , vesicular trafficking and endo / exocytosis to occur
, cellular membranes have to reshape constantly .
consequently , membranes exhibit a variety of morphologies , from a simple spherical liposome to bewildering complex structures like interconnected tubular networks as found in mitochondria and the endoplasmic reticulum ( er ) , or connected stacks of perforated membrane sheets in the golgi apparatus@xcite . there are different mechanisms by which membranes achieve these structures , the most important of which is through the interplay between membrane lipids and various proteins@xcite .
a biological membrane is home to different types of proteins that are adhered to or embedded in it .
these proteins deform the membrane and , consequently , they can either repel or attract each other @xcite . spatial organization of such proteins in biological membranes is essential for stabilizing the membrane and for the dynamic behaviour of cellular organelles @xcite .
recently , it has been experimentally @xcite and theoretically @xcite revealed that membrane - curving particles , like colloids or identical proteins , adhered to a membrane self - assemble into striking patterns .
for instance it has been shown that colloids adhered to a spherical membrane form linear aggregations @xcite . in all of the studies to date , the global shape of the membrane is selected from one of three options : planar , spherical , or tubular .
these global membrane shapes impose a homogeneous background curvature , which is considered to be conserved throughout the process under investigation .
outside factors changing the membrane have not yet been included in the study of membrane - mediated interactions .
membranes in cellular compartments such as in er and the golgi complex are however dynamic entities and possess peculiar shapes forming regions with high local curvature and regions with less curvature @xcite .
forming and stabilizing such shape inhomogeneities is necessary for cellular functions like sensing and trafficking @xcite .
it is therefore warranted to investigate how the interactions between membrane inclusions are affected by anisotropies in the membrane curvature . in this study , through a numerical experiment
, we investigate the interactions between colloids adhered to a quasi - ellipsoidal membrane with a varying curvature .
we also include all other factors from earlier studies such as surface tension , adhesion energy ( required for colloids to adhere to the membrane ) and constant volume effects .
we use a dynamic triangulation network to model the membrane , and computationally minimize the total energy of the membrane via a monte carlo algorithm .
firstly , we show that the interaction between two colloids adhered to spherical vesicles is significantly affected by the vesicle curvature .
secondly , we demonstrate that linear aggregates of colloids exploit the curvature anisotropy and adjust their orientation to minimize the total energy on a quasi - ellipsoidal membrane . using umbrella sampling ,
we further show that the total energy of the membrane favors two colloids to attract each other at the mid - plane of a prolate ellipsoid that is perpendicular to its major axis .
finally , we investigate how the various terms in the total energy of the membrane affect the strength of the interactions .
our results show that the variation in the membrane shape can play a crucial role in a variety of cellular functions that require macromolecular assembly or membrane remodeling .
the conformation of a fluid membrane can be described as the shape minimizing the classical helfrich energy functional @xcite : + @xmath0 where @xmath1 is the mean curvature at any point on the surface of the membrane and geometrically is defined as the divergence of the normal vector to the surface , @xmath2 . in our computational scheme , we discretize the membrane by a triangulated network , whose triangles represent course - grained patches of the membrane @xcite . using a discretized form of the helfrich energy ,
we define the curvature energy as : @xmath3 where @xmath4 and @xmath5 are the normal vectors to any pair of adjacent triangles @xmath6 and @xmath7 , respectively .
the summation runs over all pairs of such triangles . in order to guarantee the fluidity of the membrane
, we cut and reattach the connection between the four vertices ( which we label with and refer to as beads ) of any two neighboring triangles .
the membrane in our system does not undergo any topological changes and we can thus ignore the gaussian curvature contribution in the bending energy .
we impose the conservation of membrane surface area ( @xmath8 ) and enclosed volume ( @xmath9 ) by adding the terms @xmath10 and @xmath11 to the energy during the minimization process , with @xmath12 and @xmath13 the target values of the membrane s area and enclosed volume .
the corresponding constants are chosen such that both the area and volume deviate less than @xmath14 from their target values . to enable colloids to adhere to the membrane
, we introduce an adhesion potential , @xmath15 , between colloids and the membrane , where @xmath16 is the strength of the adhesion energy and , @xmath17 and @xmath18 are , respectively , the center to center distance and the minimum allowed separation between colloids and membrane beads .
finally , we need to give the membrane an anisotropic shape for which we deform our spherical membrane into a prolate ellipsoid . in order to do so ,
we introduce two weak ( compared to the strength of the adhesion energy ) spring - like potentials between two small areas of the vesicle ( the two poles of the ellipsoid ) and the center of the vesicle , @xmath19 ; @xmath20 , @xmath21 and @xmath22 are the potential strength , the major axis of the ellipsoid and the length of any line connecting the beads situated at the poles of the ellipsoid to the center , respectively .
since the adhesion energy is stronger than the applied harmonic potential , colloids effectively do not feel any difference between the energy cost for bending the membrane at these two areas and at the regions belonging to the rest of the ellipsoid .
we verified this claim by considering a spherical membrane , and find that there is no significant difference between the case of including @xmath23 with @xmath21 being the radius of the vesicle , and the case we do not include such a potential . + having defined all the contributions to the total energy of the membrane ( @xmath24 ) , we perform monte carlo simulations to reach the equilibrium shape of an ellipsoid containing an arbitrary number of colloids .
to do so , we implement the metropolis algorithm , in which we have three types of moves : we can modify the position of a random bead of the membrane , impose a rearrangement in the connections of beads , or move the colloids around .
the first two moves are energetically evaluated based on the total energy , while any changes in the position of colloids are only based on the adhesion energy .
+ during the simulations , we keep the number of particles constant and set all the relevant parameters as : @xmath25 , @xmath26 , @xmath27 , @xmath28 and @xmath29 , where @xmath30 is the thermal energy and @xmath31 is the diameter of the beads constructing the membrane . the diameter of the colloids is set to @xmath32 .
first , we analyze the interaction between two colloids adhered to the surface of two vesicles of different sizes .
we keep the size of colloids and beads the same in both cases .
we use umbrella sampling @xcite to calculate the excess energy of the membrane as a function of the distance between the colloids . in effect
, we apply a harmonic potential @xmath33 , as our biased potential , between the two colloids directed along the coordinate of interest in order to restrain the system to sample around each distance @xmath34 .
having performed the sampling process , we use the weighted histogram analysis method ( wham ) for obtaining the optimal estimate of the unbiased probability distribution , from which we can calculate the free energy of the system . the free energy is calculated with respect to the initial position of the colloids @xmath35 .
the excess area of the membrane available for colloids to adhere to is equal in both vesicles . as illustrated in fig .
[ fig : twovesicles ] , the depth of the excess energy of the membrane with a smaller radius is significantly larger . in contrast , for the larger vesicle after a short distance colloids do not feel each other and the energy becomes flat . as the only difference between two test cases is the curvature , we conclude that this effect is due to vesicles being of different radii .
next , we examine the interaction between two colloids on the surface of a quasi - ellipsoidal membrane .
we position the colloids symmetrically along the major axis of the ellipsoid ( see fig . [
fig : fig2]d(_i _ ) ) .
we repeat the sampling procedure for different aspect ratios , @xmath36 , of the ellipsoid .
since the volume is conserved during the shape evolution , one can easily calculate the semi - minor axis , @xmath37 , as : @xmath38 . as depicted in fig .
[ fig : fig2]a , along the major axis colloids attract each other in order to minimize both the adhesion and curvature energies . decreasing the asphericity of the ellipsoid ( @xmath39 ) in this direction
enhances the deformable area ( i.e. the number of accessible beads ) , hence the strength of the attraction energy increases .
similarly , particles that are situated along the semi - minor axis ( as depicted in fig .
[ fig : fig2]d(_ii _ ) ) attract each other .
there is , however , an important difference between the two directions .
in contrast to the previous case , decreasing the asphericity of the ellipsoid makes the attraction force between colloids weaker . since the number of membrane beads adhered to each colloid
remains the same , this behavior can not be explained by the adhesion energy of the membrane . to illuminate the reason that colloids select the direction along the minor axis to attract each other , we investigate the energy of a pair of colloids along a different coordinate . as shown in fig .
[ fig : fig2]d(_iii _ ) , we rotate a pair of colloids , that are constrained at a fixed distance to their center , along the angle spanning the space between the semi - major and -minor axes . as fig .
[ fig : fig2]c depicts , the most energetically favorable configuration is when the colloids are aligned with the direction perpendicular to the major axis .
this itself introduces a mechanism by which , without involving any other factors , two colloids find the mid - plane perpendicular to the symmetry axes of the ellipsoid , as it minimizes the total energy of the membrane .
in contrast , in the case of having a perfect spherical membrane , it is not possible to predict the localization of colloidal aggregates as it will be randomly chosen .
increasing the major axis of the membrane ( making @xmath40 larger ) , drives the colloid reorientation stronger .
one should be careful about the values for the bending moduli and adhesion coefficients during the simulations , as it can cause an effect where colloids are arrested and prevented from diffusing on the surface of the membrane @xcite .
in addition , a very high value of @xmath41 , in addition to influencing the adhesion energy between the colloids and the membrane , would also pull two tubes out of the vesicle .
+ although the above results quantitatively show different behavior in two directions , the dominant contribution in the total energy of the membrane causing this effect is not yet clarified . in order to approximately determine it ,
we proceed as follows : we pick a vesicle with @xmath42 and constrain the position of the colloids with a strong potential . here , in contrast to earlier , we do not use the sampling method .
instead we let the system explore possible configurations of the membrane after reaching equilibrium , and then take the average of the energies for all those configurations . as depicted in fig .
[ fig : fig3 ] , both the adhesion energy and the curvature of the membrane decrease when the angle between the line connecting two colloids and the semi - major axis of the ellipsoid ( fig .
[ fig : fig3]d ) approaches @xmath43 .
the bending energy , as quantified in fig .
[ fig : fig3]b , has a larger contribution to the total energy than the adhesion energy ( fig .
[ fig : fig3]c ) .
+ putting all the results together , we expect that when we have more than two colloids they will initially attract each other to form linear aggregates ( to minimize the adhesion energy ) , and afterwards these aggregations change their orientation to align with the minor axes of the ellipsoid .
this is indeed what we observe in our simulations .
[ fig : fig4 ] depicts the equilibrium shape of the membrane for different numbers of colloids . in all the test cases colloids tend to form a ring - like structure in the mid - plane of the ellipsoid . with a sufficiently large number of colloids ( figs .
[ fig : fig4]c and [ fig : fig4]d ) , they form a full ring in this plane ( see also the supplemental movie sm1 ) .
it is important to mention that these patterns are quite stable during the whole simulation .
in contrast , in spherical vesicles there is no preferred direction for the aggregation of particles . although colloids attract each other on a spherical membrane ( fig .
[ fig : twovesicles ] ) , there is no preference for the direction of the attraction .
this means that even in case of forming a perfect ring on a vesicle , particles self - assemble in an arbitrary direction on the membrane . + as the final experiment , we look at the movement of particles on a vesicle having negatively curved regions .
to do so , we first overstretch the springs ( by which we give quasi - ellipsoidal shape to vesicles ) and form negatively curved regions in a big vesicle ( @xmath44 ) .
having inserted a dimer in the system , we then look at the migration of the dimer . as shown in fig .
[ fig : fig5 ] in this case the dimer does not stay at the mid - plane of the vesicle .
it instead spends much of its time during mc simulation at the regions that are negatively curved .
since the springs are overstretched , in the regions close to poles there is no excess area for the dimer to adhere to and therefore the dimer can not explore that area ( see also supplemental movies sm2 - 4 ) .
this type of dimer migration toward the areas having higher deviatoric curvature is also experimentally observed in ref .
+ the type of pattern formation we observe in our simulations is reminiscent of recruiting proteins by the membrane during different biological processes . it has been shown that , for example , dynamin proteins form a ring like structure during exocytosis to facilitate membrane scission @xcite and that ftsz proteins self - assemble into rings during the last step of bacterial cell division , namely cytokinesis @xcite . because most of the proteins in biological cells are either anchored to or embedded in the membrane , their interaction is a response to the deformation of the membrane they themselves impose . as in our simulation
the varying curvature is a determining factor that drives the pattern formation , we can relate our results to those membrane trafficking machinery functions .
although in this study we adjusted the included harmonic potential strength @xmath45 such that it would not affect the interaction of the colloids with the membrane , it has been proven that during the cell division we have the same situation .
cytoplasmic dynein , as a multi - subunit molecular motor , generates the force that is exploited by the cell to direct the orientation of the division axis by mitotic spindles @xcite .
our results show that curvature inhomogeneity and anisotropy can at least facilitate the process of protein self - assembly in the mid - plane of the cell .
+ although we have only investigated the interaction between identical isotropic inclusions , our results can explain the behavior of a system containing anisotropically shaped inclusions as well .
based on the local deformation of an ellipsoid , we expect that anisotropic inclusions adhered to a spherical membrane attract each other in the direction of negative curvature ( with respect to the curvature of the membrane ) .
this situation corresponds to having an isotropic inclusion embedded in a membrane with an anisotropic shape , which is the case we have studied here .
we studied the role of curvature heterogeneity and anisotropy on the interaction between colloids adhered to a membrane .
first , we showed that the strength of the interaction between two colloids on the surface of a spherical vesicle is altered by changing the size of the vesicle .
next , we focused on such interactions on a membrane with an ellipsoidal shape .
we revealed that the interaction on such an inhomogeneously shaped membrane depends on direction .
for example , decreasing the asphericity of an ellipsoidal membrane makes the attraction between the colloids stronger along the semi - major axis and weaker in the semi - minor direction .
similarly , it has been previously shown in simulations that , on an elastic cylindrical membrane , colloids assemble perpendicularly to its major axis in the regime dominated by the bending energy @xcite . in case of fluid membranes , through an analytical framework
, it has also been shown that inclusions _
`` embedded '' _ in a tubular membrane can attract each other in a transversal direction@xcite . simulating a vesicle containing many colloids
, we showed how they form a ringlike structure around the mid - plane of the ellipsoid .
while the cluster of colloids freely explores all the surface of a spherical membrane , less curved area energetically is more favorable for colloids on an ellipsoid .
our results suggest that forming regions of different curvatures on membrane vesicles can control pattern formation of inclusions , and this can be important from both nanotechnological application and biological points of view .
this work was supported by the netherlands organisation for scientific research ( nwo / ocw ) , as part of the frontiers of nanoscience program . | cellular membranes exhibit a large variety of shapes , strongly coupled to their function .
many biological processes involve dynamic reshaping of membranes , usually mediated by proteins .
this interaction works both ways : while proteins influence the membrane shape , the membrane shape affects the interactions between the proteins . to study these membrane - mediated interactions on closed and anisotropically curved membranes , we use colloids adhered to ellipsoidal membrane vesicles as a model system .
we find that two particles on a closed system always attract each other , and tend to align with the direction of largest curvature .
multiple particles form arcs , or , at large enough numbers , a complete ring surrounding the vesicle in its equatorial plane .
the resulting vesicle shape resembles a snowman .
our results indicate that these physical interactions on membranes with anisotropic shapes can be exploited by cells to drive macromolecules to preferred regions of cellular or intracellular membranes , and utilized to initiate dynamic processes such as cell division .
the same principle could be used to find the midplane of an artificial vesicle , as a first step towards dividing it into two equal parts . |
topologically ordered phases are fascinating states of quantum matter characterized by a ground - state degeneracy which depends on the genus of the considered topology @xcite .
they are also known to exhibit exotic elementary excitations , dubbed anyons , which have nontrivial braiding statistics @xcite .
recently , the interest for the study of topologically ordered phases has been triggered by the possibility to use anyons as an essential ingredient either to perform quantum computation or to provide a quantum memory @xcite .
to understand these properties , much focus has been drawn on two - dimensional quantum spin systems especially since kitaev has proposed two exactly solvable models , the toric code @xcite and the honeycomb model @xcite , in which such anyons are present . in the present work ,
we provide an extension of the kitaev honeycomb model for which many experimental realizations in optical lattices @xcite , josephson junctions arrays @xcite , or solid state systems with strong spin - orbit coupling @xcite have been proposed . up to now
, many variations of this model have already been investigated , but almost all of them rely on changing the geometry of the underlying lattice @xcite . here , we choose an alternative route by keeping the honeycomb geometry but by changing the coupling configuration .
doing so , we propose a set of exactly solvable models which , as we will see , can give rise to surprising properties as compared to the original one @xcite .
this paper is organized as follows . in sec .
[ sec : themodel ] we recall the basic properties of the kitaev honeycomb model and we explain how to extend it by considering different dimer coverings . in sec .
[ sec : phasediag ] , we focus on three simple coupling patterns for which we compute the ground - state ( vortex - free ) phase diagrams using majorana fermionization and analyze the influence of a three - spin interaction term . the following sec .
[ sec : vortexprop ] is dedicated to sparse vortex configurations .
there , we use a perturbative treatment to show that the nature ( attractive or repulsive ) of vortex - vortex interaction is very sensitive to the dimer configuration .
the kitaev honeycomb model describes a set of spins 1/2 located at the vertices of a hexagonal lattice which interact via the following hamiltonian @xmath0 where @xmath1 are the usual pauli matrices at site @xmath2 . in the original model @xcite , kitaev considered the case where the type of links ( @xmath3 , or @xmath4 ) depends only on its orientation as depicted in fig .
[ fig : lattice1 ] . initially considered by kitaev @xcite with two sites per unit cell .
vectors @xmath5 and @xmath6 define the unit cell .
red , green , and blue links represent @xmath7 , and @xmath4 links respectively . ]
so far , extensions of this model were based on a different lattice geometry @xcite or a different underlying algebra @xcite .
sticking to a spin-1/2 model , the exact solvability is preserved only if the following two constraints are satisfied : @xmath8 each site of the system must be trivalent and @xmath9 the three links connected to a given site have to be of different types ( @xmath10 , @xmath11 , and @xmath4 ) .
thus , the hamiltonian ( [ eq : ham ] ) can be defined and solved on any trivalent graph . here ,
by contrast , we study the kitaev model ( [ eq : ham ] ) in the original honeycomb lattice but for different link distributions or dimer coverings ( we shall use the expressions link distributions or dimer coverings as synonyms , so dimer coverings do not have the same meaning here as they have in the context of quantum dimer models ) . more precisely , we focus on the three simple dimer coverings shown in figs . [
fig : lattice1 ] and [ fig : lattice2 ] .
these configurations are among the simplest periodic decorations one may build on such a lattice . indeed , the honeycomb lattice has two sites per unit cell so that any periodic decoration will have @xmath12 sites per unit cell .
configurations shown in figs .
[ fig : lattice1]-[fig : lattice2 ] correspond to @xmath13 , and 3 respectively , and the corresponding models will be consequently referred to as @xmath14 , @xmath15 , and @xmath16 . in this terminology , the original kitaev model introduced in ref .
is model @xmath14 . with four sites per unit cell .
right : dimer covering @xmath16 with six sites per unit cell .
conventions are the same as in fig .
[ fig : lattice1 ] and labels ( a ) , ( b ) , and ( c ) refer to the different kinds of plaquettes whose corresponding operators are given in eqs .
( [ eq : wp_ii]-[eq : wp_iii ] ) . ]
as shown below , the phase diagram strongly depends on the dimer covering and displays very different properties when switching on a three - spin interaction term already introduced by kitaev @xcite to break the time - reversal symmetry explicitely . such a term is built in the following way : if two dimers @xmath17 of type @xmath18 and @xmath19 of type @xmath20 share a common site @xmath21 , then the three - spin term is defined by @xmath22 where @xmath23 is the direction orthogonal to @xmath18 and @xmath20 . the hamiltonian considered in the present work
can thus be written as @xmath24 here , we only consider positive couplings @xmath25 and @xmath26 . a crucial feature of hamiltonian ( [ eq : hamk ] ) is that it commutes with plaquette operators defined , on each plaquette @xmath27 , as @xmath28 where @xmath2 runs over the set of six spins around the plaquette @xmath27 , and where the notation @xmath29 means the `` outgoing '' direction at site @xmath2 , with respect to the plaquette s contour @xcite . since @xmath30 , plaquette operators @xmath31 have eigenvalues @xmath32 . consequently , there is one conserved @xmath33 degree of freedom per plaquette .
following kitaev @xcite , we will say that there is a vortex on a plaquette if @xmath34 and no vortex if @xmath35 .
since @xmath36=0 $ ] , one can block - diagonalize @xmath37 in each vortex sector given by a configuration of the @xmath38 s .
the three dimer coverings considered here differ by the number and the type of plaquette operators @xmath39 .
more precisely , for covering @xmath14 there is only one type of plaquette operator @xmath40 where @xmath41 to @xmath42 are the six spins numbered clockwise around any plaquette @xmath27 , spin number @xmath41 being at the bottom of the plaquette .
covering @xmath43 has two types ( a and b ) of plaquette operators ( see fig .
[ fig : lattice2 ] ) @xmath44 covering @xmath45 has three types ( a , b and c ) of plaquette operators ( see fig .
[ fig : lattice2 ] ) @xmath46 one can readily see that the structure of these operators is similar for coverings @xmath14 and @xmath43 since they contain two pauli matrices of each kind ( @xmath7 and @xmath4 ) whereas , for covering @xmath45 , each plaquette operator is made of identical pauli matrices .
for the three dimer coverings introduced in sec .
[ sec : themodel ] , we shall now determine the ground - state phase diagram . as shown by lieb in a different context
@xcite , the ground state , at least for @xmath47 , lies in the vortex - free sector ( @xmath38=@xmath48 for all @xmath27 ) on which we focus thereafter .
undoubtedly , one of the most remarkable properties of the hamiltonian ( [ eq : hamk ] ) is that , in each vortex sector , it can be mapped onto a free majorana fermion problem in a static @xmath33 gauge field . as detailed in ref .
, @xmath37 can be written in terms of majorana fermion operators @xmath49 at site @xmath21 , and reads @xmath50 where @xmath51 . in the latter expression , @xmath52 is the coupling on the link @xmath19 and @xmath53 is a gauge field which describes the vortex configuration through the relation @xmath54 where @xmath21 belongs to the black sublattice and @xmath55 to the white one ( see figs .
[ fig : lattice1 ] and [ fig : lattice2 ] ) . in eq .
( [ eq : trafohami ] ) , the sum runs over all sites and @xmath56 is only nonvanishing when @xmath2 and @xmath21 are nearest neighbors .
thus , for a given vortex configuration , one simply has to diagonalize the matrix @xmath57 to get the spectrum of the corresponding sector .
obviously , such a task is easy for periodic configurations but difficult otherwise . in particular ,
when one is interested in sparse vortex configurations , one needs to diagonalize large systems and perform some finite - size analysis ( see sec . [
sec : vortexprop ] ) .
to determine the phase diagram , we computed the fermionic gap for the three dimer coverings @xmath14 , @xmath43 , and @xmath45 . for symmetry reasons ,
all results below are displayed in the plane @xmath58 which is further parametrized in the following way @xmath59 with @xmath60 .
the dimer covering shown in fig .
[ fig : lattice1 ] is completely symmetric under the exchange of the three directions @xmath61 , and @xmath4 .
thus , for convenience and without loss of generality , let us assume @xmath62 .
the fermionic gap as a function of @xmath26 evolves as follows .
+ @xmath63 for @xmath47 , one has @xcite
@xmath64 so that one has to distinguish between a gapped phase @xmath57 for @xmath65 and a gapless phase @xmath66 for @xmath67 .
+ @xmath63 in the small @xmath26 limit , a gap opens up in the @xmath66 phase whereas , in the @xmath57 phase , the gap does not depend on @xmath26 . for all @xmath26 ,
the spectrum remains gapless on the line @xmath68 @xcite .
+ @xmath63 in the large @xmath26 limit , the gap in the @xmath66 phase reaches the following asymptotic value @xmath69 the evolution of the phase diagram when varying @xmath26 is given in fig .
[ fig : square0 ] . as a function of @xmath26 .
gapped regions at @xmath47 correspond to phases a and the gapless domain is identified with phase b. for finite @xmath26 the phase b becomes gapped and the gap saturates in the large @xmath26 limit.,title="fig : " ] as a function of @xmath26 .
gapped regions at @xmath47 correspond to phases a and the gapless domain is identified with phase b. for finite @xmath26 the phase b becomes gapped and the gap saturates in the large @xmath26 limit.,title="fig : " ] + as a function of @xmath26 .
gapped regions at @xmath47 correspond to phases a and the gapless domain is identified with phase b. for finite @xmath26 the phase b becomes gapped and the gap saturates in the large @xmath26 limit.,title="fig : " ] as a function of @xmath26 .
gapped regions at @xmath47 correspond to phases a and the gapless domain is identified with phase b. for finite @xmath26 the phase b becomes gapped and the gap saturates in the large @xmath26 limit.,title="fig : " ] as shown by kitaev , phase a is abelian , whereas phase b contains non - abelian excitations in the presence of a three - spin term @xcite . in both cases ,
anyonic excitations are vortices living on plaquettes .
contrary to covering @xmath14 , covering @xmath43 is not symmetric under the exchange of the three directions @xmath61 , and @xmath4 , but it is only symmetric under the exchange @xmath70 . in this case , the evolution of the fermionic gap with @xmath26 is the following .
+ @xmath63 for @xmath47 , the phase diagram is exactly the same as for dimer covering @xmath14 .
since model @xmath43 is equivalent to model @xmath14 when @xmath71 or @xmath72 , one can conclude that the statistics of the excitations in phases @xmath57 and @xmath66 will also correspond to the statistics for covering @xmath14 .
the same argument will hold for model @xmath45 .
+ @xmath63 for @xmath73 , the situation is more complex and one must distinguish between two regions . + - for @xmath65 , the gap is independent of @xmath26 and is still given by eq .
( [ eq : gap10 ] ) .
+ - for @xmath67 , the gap has a nontrivial @xmath26-dependence .
it vanishes on the lines @xmath74 @xmath63 in the large @xmath26 limit , the gap for @xmath65 remains unchanged whereas the gap in the @xmath66 phase is made up of two planes which intersects on the line @xmath75 ( see fig .
[ fig : hii ] ) . as a function of @xmath26 . for @xmath47 ,
the gap is the same as for covering @xmath14.,title="fig : " ] as a function of @xmath26 . for @xmath47 ,
the gap is the same as for covering @xmath14.,title="fig : " ] + as a function of @xmath26 . for @xmath47 ,
the gap is the same as for covering @xmath14.,title="fig : " ] as a function of @xmath26 . for @xmath47 ,
the gap is the same as for covering @xmath14.,title="fig : " ] this covering is completely symmetric under the exchange of the three directions @xmath61 , and @xmath4 .
the fermionic gap as a function of @xmath26 is however very different from previous cases , and is illustrated in fig .
[ fig : kagome0 ] .
+ @xmath63 for @xmath47 , the gap is given by @xmath76 as can be easily checked , it only vanishes at the isotropic point @xmath72 ( as besides for any possible covering of the honeycomb lattice at @xmath47 ) . in other words , for this covering there is no gapless phase but just a gapless point .
+ @xmath63 for @xmath77 , a gapped non - abelian phase starts to develop around the isotropic point delimited by a gapless circle .
the gap in the abelian region decreases when increasing @xmath26 . + as a function of @xmath26 . for @xmath47
, there is a single gapless point and the whole system is in an abelian phase .
by contrast , in the large @xmath26 limit , the system is in a non - abelian phase and the gap is constant.,title="fig : " ] as a function of @xmath26 . for @xmath47 ,
there is a single gapless point and the whole system is in an abelian phase .
by contrast , in the large @xmath26 limit , the system is in a non - abelian phase and the gap is constant.,title="fig : " ] + as a function of @xmath26 .
for @xmath47 , there is a single gapless point and the whole system is in an abelian phase .
by contrast , in the large @xmath26 limit , the system is in a non - abelian phase and the gap is constant.,title="fig : " ] as a function of @xmath26 . for @xmath47 ,
there is a single gapless point and the whole system is in an abelian phase .
by contrast , in the large @xmath26 limit , the system is in a non - abelian phase and the gap is constant.,title="fig : " ] - for @xmath78 , the gapless circle reaches the middle of the triangle sides and isolates each abelian gapped phase .
+ - for @xmath79 , the gapless line finally reaches the triangle corners so that only the non - abelian phase remains .
+ @xmath63 in the large-@xmath26 limit , the gap becomes constant and equal to @xmath80 in the whole parameter range considered and the excitations are non - abelian anyons .
let us now discuss the one- and two vortex sectors .
our main motivation for such a study is to unveil some qualitative differences of the vortex properties for the different coverings .
such differences are already present in the abelian phase ( @xmath81 ) and for @xmath47 which is the parameter range to which we restrict in the following .
furthermore , for the sake of simplicity , we set @xmath82 and @xmath83 . along this line ,
vortex properties of coverings @xmath14 and @xmath43 are identical but differ from those of covering @xmath45 . for this study
, we have also used majorana fermionization , but contrary to the vortex - free sector , we now have to deal with non - periodic configurations and to perform some finite - size analysis on large systems . here , we treated system sizes up to @xmath84 .
one important remark is that , using periodic boundary conditions , vortices can only be created in pairs @xcite .
thus , to compute the one - vortex gap , one has first to consider the two - vortex gap @xmath85 of a pair of vortices on plaquettes @xmath2 and @xmath21 defined by @xmath86 the right - hand side is the energy difference between the ground - state energy @xmath87 of the considered two - vortex sector and @xmath88 which is the ground - state energy of the vortex - free sector .
then we use the fact that , in the abelian phase , vortex excitations are gapped @xcite so that the interaction of two distant vortices decreases exponentially .
this allows us to define the one - vortex gap as @xmath89 where it is essential that both vortices are of the same type , _ i. e. _ a or b for model @xmath43 , and a or b or c for model @xmath45 , see sec .
[ sec : themodel ] ) .
practically , results are converged for @xmath90 as long as the vortex gap is not too small , namely , away from the transition points ( see sec .
[ sec : single_vortices ] ) . to have a better understanding of these few - vortex properties
, we have also developped an alternative approach .
we computed the spectrum of @xmath37 perturbatively up to high orders using the perturbative continuous unitary transformation ( pcut ) method @xcite around the isolated - dimer limit , where one of the couplings is much larger than the others .
such a strategy has been shown to be very successful in the original kitaev model @xcite although restricted to abelian phases . here
, we proceed along the same line as in ref . and we skip all technical details which can be found in this reference . by replacing @xmath4-dimers by effective sites ( filled black circles ) .
the effective lattice for coverings @xmath14 ( left ) and @xmath43 ( middle ) is a square " lattice . for covering @xmath45 ( right ) ,
one gets a kagome lattice .
gray - shaded areas show the different types of effective plaquette identified by the same labels as in fig .
[ fig : lattice2 ] .
red ( green ) bonds denote @xmath10 ( @xmath11 ) links . ] as discussed in refs .
, in the limit @xmath91 it is convenient to analyze the problem by replacing @xmath4-dimers by effective sites with four degrees of freedom corresponding to the four possible dimer states .
then , depending on the dimer configuration , one can face different structures ( see fig .
[ fig : eff_lattices ] ) . for covering @xmath14 ,
the effective lattice is a square " lattice in which all plaquettes are identical . for covering @xmath43
, one still has a square lattice but one must distinguish between two kinds of plaquettes ( a and b ) which reflect the operator structure ( [ eq : wp_ii ] ) .
by contrast , the effective lattice for covering @xmath45 is a kagome lattice with three different types ( a , b and c ) of plaquettes : one hexagonal plaquette with alternating @xmath10- and @xmath11-links and two triangular plaquettes with either only @xmath10- or only @xmath11-links .
it is worth noting that such an effective lattice was also obtained in ref . , starting from a different initial geometry . as already explained , in the abelian phase the kitaev honeycomb model contains two different kinds of excitations : low - energy vortices and high - energy fermions .
the pcut method around the isolated - dimer limit provides ( order by order in @xmath92 and in the thermodynamic limit ) an effective hamiltonian which commutes with the number of fermions .
thus , the low - energy spectrum is obtained by considering the zero - fermion ( quasiparticle ) block usually denoted 0qp . in this model ,
the effective hamiltonian in the 0qp sector is particularly simple because it can be fully expressed in terms of the conserved @xmath33 plaquette operators @xmath93 .
its general structure reads @xmath94 where @xmath95 is a constant and where @xmath96 denotes a set of @xmath97 plaquettes .
the sum runs over all possible plaquette numbers and configurations . in the present work ,
we have computed the corresponding coefficients @xmath98 of this multi - plaquette expansion up to order 10 .
the ground - state energy of any vortex configuration is then readily obtained by replacing @xmath39 by @xmath99 if there is a vortex and by @xmath48 otherwise . before presenting the results
, we would like to point out that the validity range of the pcut results does _ not _ depend on the vortex filling .
however , this perturbative treatment is only valid as long as all fermionic gaps are finite .
for instance , this implies that although the gapless phase is reduced to a single point for dimer covering @xmath45 , this perturbative approach breaks down well before this point is reached since some fermionic gaps in other vortex sectors vanish @xcite .
for the three dimer coverings considered here , we are led to conjecture that there always exists a vortex configuration such that the fermionic gap vanishes for @xmath68 . for covering @xmath14 and @xmath43 , this
is achieved in the vortex - free sector but , for covering @xmath45 , this is found for more complex vortex configurations @xcite . as a consequence , we restrict our analysis to @xmath100 $ ] . on the line considered here ( @xmath82 ) ,
coverings @xmath14 and @xmath43 are identical . however , for covering @xmath45 , one must distinguish between two different vortex gaps according to the kind of plaquette which is excited ( triangle or hexagon of the effective kagome lattice ) . as a function of @xmath101 for coverings @xmath14 and @xmath43 ( top ) and for covering @xmath45 ( middle and bottom ) .
we give the bare series expansion computed up to order 10 , some pad approximants of these series ( except for the hexagonal plaquette for which we do not get enough terms in the series to provide reliable approximants ) , and the exact diagonalization results .
insets show the excited plaquette considered together with the effective lattice.,title="fig : " ] as a function of @xmath101 for coverings @xmath14 and @xmath43 ( top ) and for covering @xmath45 ( middle and bottom ) .
we give the bare series expansion computed up to order 10 , some pad approximants of these series ( except for the hexagonal plaquette for which we do not get enough terms in the series to provide reliable approximants ) , and the exact diagonalization results .
insets show the excited plaquette considered together with the effective lattice.,title="fig : " ] as a function of @xmath101 for coverings @xmath14 and @xmath43 ( top ) and for covering @xmath45 ( middle and bottom ) .
we give the bare series expansion computed up to order 10 , some pad approximants of these series ( except for the hexagonal plaquette for which we do not get enough terms in the series to provide reliable approximants ) , and the exact diagonalization results .
insets show the excited plaquette considered together with the effective lattice.,title="fig : " ] the single - vortex gaps obtained by majorana fermionization and by the pcut are displayed for these three different cases in fig .
[ fig : single_vex_full ] . an excellent agreement between both approaches is evidenced .
the most striking result is that the value of this vortex gap depends strongly on the type of effective plaquette .
this can be understood easily by noting that the leading order in the perturbative expansion is determined by the `` perimeter '' of the plaquette ( 3 for the triangle , 4 for the square , and 6 for the hexagon ) . in this sense
, one may say that the mass of the vortex is strongly sensitive to the dimerization pattern . for illustration , we give the analytical expression of these various gaps computed up to order 10 @xmath102 @xmath103 @xmath104 finally , let us remind that in this abelian phase , vortices behave as abelian anyons or hard - core bosons depending on their relative position ( see ref . for a detailed discussion ) . in the present case , for covering @xmath43 , plaquette excitations ( a ) and ( b ) are mutual semions whereas they behave , individually , as hard - core bosons .
for covering @xmath45 , triangular and hexagonal plaquette excitations are mutual semions whereas triangular ( hexagonal ) plaquette excitations behave individually as hard - core bosons . to conclude this section ,
let us discuss the nature of vortex - vortex interactions @xmath105 between two vortices on plaquettes @xmath2 and @xmath21 . the interaction energy between two such vortices is defined as @xmath106 where the energy of each type of vortex is subtracted .
of course , such an interaction depends on the relative position of the vortices . once again , on the line considered here ( @xmath107 ) this interaction is the same for coverings @xmath14 and @xmath43 .
for these two coverings , we found that this interaction is always attractive ( this is also true for @xmath108 ) whereas , for covering @xmath45 , there are some positions for which it is repulsive .
these results are illustrated qualitatively in fig .
[ fig : vortex_interactions ] . for coverings @xmath14 and @xmath43 ( left ) and for covering @xmath45 ( right ) .
a green ( red ) plaquette interacts attractively ( repulsively ) with a vortex placed on the central blue plaquette .
the color code is such that a darker green corresponds to a stronger attractive interaction.,title="fig : " ] for coverings @xmath14 and @xmath43 ( left ) and for covering @xmath45 ( right ) .
a green ( red ) plaquette interacts attractively ( repulsively ) with a vortex placed on the central blue plaquette .
the color code is such that a darker green corresponds to a stronger attractive interaction.,title="fig : " ] in addition , fig .
[ fig : ia_full ] shows a comparison between exact diagonalizations and pcut results for some special cases where vortices are placed on nearest - neighbor effective plaquettes .
again , both approaches are in very good agreement and confirm that some repulsion can be found in covering @xmath45 .
this interesting finding can be readily extracted from the series expansion obtained from pcut .
indeed , the leading order of the interaction between two effective triangular plaquettes is given by the repulsive term @xmath109 while the dominant contribution to the interaction between a triangular and a hexagonal plaquette is attractive and reads @xmath110 .
for two adjacent effective plaquettes as a function of @xmath101 for coverings @xmath14 and @xmath43 ( top ) and for covering @xmath45 ( middle and bottom ) . as in fig .
[ fig : single_vex_full ] , we do not show pad approximants when hexagonal plaquettes are involved since the corresponding series starts at order 6 so that there is not enough terms to get reliable results when resumming .
insets show the nearest - neighbor excited plaquettes for which the two - vortex interaction is computed.,title="fig : " ] for two adjacent effective plaquettes as a function of @xmath101 for coverings @xmath14 and @xmath43 ( top ) and for covering @xmath45 ( middle and bottom ) . as in fig .
[ fig : single_vex_full ] , we do not show pad approximants when hexagonal plaquettes are involved since the corresponding series starts at order 6 so that there is not enough terms to get reliable results when resumming .
insets show the nearest - neighbor excited plaquettes for which the two - vortex interaction is computed.,title="fig : " ] for two adjacent effective plaquettes as a function of @xmath101 for coverings @xmath14 and @xmath43 ( top ) and for covering @xmath45 ( middle and bottom ) . as in fig .
[ fig : single_vex_full ] , we do not show pad approximants when hexagonal plaquettes are involved since the corresponding series starts at order 6 so that there is not enough terms to get reliable results when resumming .
insets show the nearest - neighbor excited plaquettes for which the two - vortex interaction is computed.,title="fig : " ]
we have extended the kitaev honeycomb model to arbitrary dimer coverings satisfying matching rules .
this results in a family of models which can still be solved exactly by majorana fermionization . although these models are equivalent at the isotropic point , they strongly differ at any other point in parameter space . in the present work , we focused on three different coverings ( @xmath14 , @xmath43 , and @xmath45 ) having the smallest possible unit cells .
for these three models , we have studied the ground - state phase diagram as a function of the three - spin interaction strength @xmath26 .
we found that there are many similarities between the phase diagrams of @xmath111 and @xmath112 .
however , the phase diagram of @xmath113 is dramatically different . in particular ,
the gapless phase at @xmath47 is reduced to a single point and , for large @xmath26 , no abelian phase exists .
we have furthermore studied one- and two - vortex properties in these different coverings in the abelian phase at @xmath47 and we have shown that they also depend on the covering considered .
these findings have been analytically confirmed by perturbative expansions around the isolated - dimer limit inside the abelian phase . finally , let us point out two routes left for the future .
firstly , it would be interesting to investigate other vortex congurations since , as recently observed in the original kitaev model @xcite , vortex interactions could lead to new topological phases characterized , notably , by exotic chern numbers @xcite .
secondly , it would be worth studying the effect of disordered dimer and/or vortex configurations to see if an anderson - type transition can occur in such systems .
note that other kinds of disorder have recently been considered in refs . .
we hope that the present work will stimulate further investigations in a near future . | we consider an extension of the kitaev honeycomb model based on arbitrary dimer coverings satisfying matching rules .
we focus on three different dimer coverings having the smallest unit cells for which we calculate the ground - state phase diagram .
we also study one- and two - vortex properties for these coverings in the abelian phases and show that vortex - vortex interactions can be attractive or repulsive .
these qualitative differences are confirmed analytically by high - order perturbative expansions around the isolated - dimer limit .
similarities and differences with the original kitaev honeycomb model are discussed . |
for the three classes of states , all entangled states are successfully detected without the need to perform full tomography .
there are , however , other entangled states that would escape detection , and all separable states can only give inconclusive results
. regarding entangled states , we recall that about @xmath66 of the random pure states and about @xmath67 of the random mixed states are not detected by the six witness families without the separability check of scheme c. hence , to confirm the efficiency and accuracy of the tomographic scheme , we collect ic data even if the state is known to be entangled before all six witness - family measurements are done . using the technique of ml estimation @xcite
, we then infer the state from the data and calculate the fidelity @xmath68 between the true and the estimated state .
the average fidelities are @xmath69 , @xmath70 , and @xmath71 for the respective true states of rank one , two , and four . indeed ,
if all the six family measurements fail to detect the entanglement in the state , one can use the tomographic information to reliably reconstruct the unknown input state and then determine its separability numerically . | we report an experiment in which one determines , with least tomographic effort , whether an unknown two - photon polarization state is entangled or separable .
the method measures whole families of optimal entanglement witnesses .
we introduce adaptive measurement schemes that greatly speed up the entanglement detection .
the experiments are performed on states of different ranks , and we find good agreement with results from computer simulations . * introduction * entangled states play an important role in the manipulation of quantum information , be it for present - day quantum key distribution or future quantum computation .
one may need to verify if a certain quantum state perhaps emitted by a source of quantum - information carriers or obtained as the output of a quantum computation is entangled or not .
for this purpose , the expectation value of an _ entanglement witness _ is telling : the state is surely entangled if a negative value is obtained . a positive value , however , is inconclusive the unknown state could be entangled or separable , the witness can not tell , but other witnesses might be able to .
the concept of a witness was first used by m. , p. , and r. horodecki @xcite , and the term `` witness '' was introduced by terhal @xcite ; for reviews that cover all important aspects of entanglement witnesses , see refs .
@xcite .
how many witnesses , then , does one need to measure until a conclusion is reached ?
the answer to this question is given in ref .
@xcite : if one solely relies on the expectation values of the witnesses one by one , one may never get a conclusive answer ; if , however , the expectation values of suitably chosen witnesses are jointly used for an estimation of the quantum state , @xmath0 witnesses suffice for a @xmath1-dimensional quantum system . the number can be further reduced by exploiting all the information gathered when determining the expectation values of the witnesses measured in succession . in the case of a two - qubit state ( @xmath2 ) , for instance , one never has to measure more than six witnesses , rather than @xmath3 .
this can be demonstrated by an experiment such as the one proposed in ref .
@xcite ; we are here reporting its laboratory realization .
* witnesses and witness families * a hermitian observable @xmath4 is an entanglement witness if @xmath5 for _ all _ separable states @xmath6 and @xmath7 for at least one entangled state @xmath8 . for each entangled state
, there are some witnesses that detect it ( `` @xmath9 '' ) , but many other witnesses will give an inconclusive result ( `` @xmath10 '' ) .
we shall concern ourselves with two - qubit systems in the experiment , they are polarization qubits of a down - converted photon pair and focus on optimal decomposable witnesses @xcite of the form @xmath11 , where @xmath12 is the ket of an entangled pure two - qubit state , and @xmath13 denotes the partial transposition on the second qubit .
this witness is optimal in the sense that no other witness can detect some entangled states in addition to the states already detected by @xmath4 .
the generic example is @xmath14 with @xmath15 , on which all other @xmath12s can be mapped by local unitary transformations . for any @xmath16 ,
the eigenkets of the resulting witness , @xmath17 are the same : the two product kets @xmath18 and @xmath19 as well as the two bell kets @xmath20 .
this whole family of entanglement witnesses can , therefore , be measured by the projective measurement of their common eigenstates the _ witness basis _ of the family . for a separable state ,
@xmath21 for all @xmath16 , and this requirement implies the witness - family criterion @xcite @xmath22 here , @xmath23 and @xmath24 are the probabilities for the two product states , and @xmath25 and @xmath26 are those for the bell states ; we estimate these probabilities from the observed frequencies .
consequently , once the frequency data are obtained from the measurement , a negative value of @xmath27 reveals that the unknown state @xmath28 is entangled .
three remarks are in order .
( i ) the expectation value of one witness @xmath29 is a linear function of the @xmath30s , whereas @xmath27 is a quadratic function .
this is reminiscent of , yet different from , the `` nonlinear entanglement witness '' of ref .
@xcite , which requires a joint measurement on two copies of the unknown state .
our _ witness - family measurement _ uses only one copy at a time .
( ii ) also , the nonlinear witnesses of ref .
@xcite are quite different ; their evaluation requires complete or almost complete knowledge of the state .
( iii ) the witnesses @xmath29 can also be measured by other schemes , such as that of barbieri _ et al .
_ @xcite who extracted the expectation value of @xmath31 from local measurements that examine the two qubits individually .
by contrast , we perform a joint measurement on both qubits , which implements the projective measurement in the eigenbasis of @xmath29 and so realizes the most direct measurement of the witness .
see the appendix for more details about ( ii ) and ( iii ) .
our witness - family measurement provides estimates for _ three _ two - qubit observables ( four @xmath30s with unit sum ) , whereas the expectation value of a single witness is only one number . as discussed in ref .
@xcite , this can be exploited for quantum - state reconstruction after measuring six witness families , related to each other by the six local unitary transformations of table [ tbl : six ] : each witness family provides the expectation values of one ( of six ) single - qubit observables and of two ( of nine ) two - qubit observables .
therefore , a measurement of all six witness families constitutes an informationally complete ( ic ) measurement for full tomography of an unknown two - qubit state ; thereby , all six single - qubit parameters and six two - qubit parameters are obtained once , while three two - qubit parameters are determined twice .
this offers the possibility of measuring an ic set of witness families such that , if all families give an inconclusive result ( @xmath32 ) , a full state estimation can be performed for identifying @xmath28 . with @xmath28 then at hand , its separability can be determined straightforwardly by , for example , checking the peres - horodecki criterion @xcite . .[tbl : six]the six witness families that enable full tomography of the two - qubit state .
the single - qubit unitary operators @xmath33 and @xmath34 transform the first family into the other five families .
the pauli operator @xmath35 permutes @xmath36 and @xmath37 ; the clifford operator @xmath38 permutes the three pauli operators cyclically . [ cols="^,^,^",options="header " , ] the angles @xmath16 , @xmath39 , and @xmath40 , for which the various @xmath41s are realized , are reported in table [ stbl : six ] . as an example , we consider @xmath42 and verify that @xmath43 , @xmath44 , and @xmath45 are correct choices .
indeed , they are : @xmath46 since the global phase factor is irrelevant . * hom interferometer * to implement the most direct measurement of the witness bases , we make use of a hom interferometer . for its optimization
, we first remove the bss in the output ports of the pbss ; see fig . [
fig : exp ] .
next , we address the outputs of the two detectors in the transmission ports of the two pbss to a coincidence unit , where we record the coincidence rate for the input two - photon state @xmath47 . a translation stage with a step size of @xmath48 is used to control the temporal overlap between the photons , and the spatial overlap is controlled by adjusting the mirrors that direct the two beams to interfere on the bs .
the interferometer is optimized at where the coincidence rate is minimal , as shown in fig .
[ sfig : homdip ] .
the visibility of the hom dip is @xmath49 .
imperfections in the bs ratios , the wps , and the polarization controllers limit the maximum experimental achievable visibility of the dip
. nevertheless , the visibilities of the hom dips obtained in our experiment exceed @xmath50 for all the data collected for the construction of the histograms shown in fig .
[ fig : data ] .
an example of a hom dip obtained in our experiment for the state @xmath51 .
the visibility , @xmath52 , of the hom dip above is @xmath49 ; other hom dips observed for different polarization states are similar to this one . ] * data analysis * here we illustrate how data were analyzed to obtain fig . 3 in the main text .
for the witness family measurement , the first family is chosen at random among the six families .
if the measurement of this family gives a negative value of @xmath27 , the state is detected to be entangled and no further measurement is needed . if the result is inconclusive , then one uses the adaptive scheme to choose the next family , until a conclusive result is obtained .
the result is @xmath53 , the number of witness families that have to be measured in order to detect the entanglement ; see table [ tbl : expl ] for illustrations .
c@ccc@cstate & & @xmath53 + ' '' '' @xmath54 & family 3 & family 4 & family 2 & 3 + @xmath55 & @xmath56 & @xmath57 & @xmath58 & + @xmath59 & family 1 & & & 1 + @xmath60 & @xmath61 & & & + @xmath62 & family 4 & family 2 & & 2 + @xmath63 & @xmath64 & @xmath65 & & + |
thanks to recent technological advances in producing high - quality polarized monochromatic photon beams , and in developing polarized nucleon targets , it becomes possible to measure a sufficiently large amount of single- and double - polarization observables in pion and kaon photoproduction from the nucleon . as a result ,
a status of complete quantum mechanical information of meson photoproduction comes within reach .
measurements are complete whenever they enable one to determine unambiguously all amplitudes of the underlying reaction process at some specific kinematics .
we consider the @xmath0 reaction as a prototypical example of pseudoscalar - meson photoproduction from the proton .
the transversity amplitudes @xmath1 express the transition matrix elements in terms of the @xmath2 and @xmath3 spinors ( with quantization axis perpendicular to the reaction plane ) and of linear photon polarizations @xmath4 .
we propose @xcite to use normalized transversity amplitudes ( nta ) @xmath5 to perform an amplitude analysis of the single- and double- polarization observables .
the nta provide complete information after determining the differential cross section .
the corresponding polarization observables can be expressed in terms of linear and nonlinear equations of bilinear products of the @xmath6 . for a given kinematical setting determined by the meson angle @xmath7 and the invariant mass @xmath8 , the @xmath9
are fully determined by six real numbers conveniently expressed as three real moduli @xmath10 and three real relative phases @xmath11 .
all observables are invariant under a transformation of the type @xmath12 , with @xmath13 an arbitrarily chosen overall phase . in fig .
[ f1 ] we show predictions for the nta at @xmath14 mev and various @xmath7 .
the adopted model for @xmath15 is the regge - plus - resonance ( rpr ) approach in its most recent version rpr-2011 @xcite .
the model has a reggeized @xmath16-channel background and the @xmath17-channel resonances @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 , and @xmath24 .
the rpr approach provides a low - parameter framework with predictive power for @xmath25 and @xmath26 photoproduction on the proton and the neutron @xcite .
an obvious advantage of using the transversity amplitudes is that linear equations connect the moduli @xmath28 of the nta to the single - polarization observables @xmath29 @xmath30 accordingly , a measurement of @xmath31 at given @xmath32 allows one to infer the moduli @xmath33 of the nta .
the graal collaboration provides @xmath34 data for @xmath35 at 66 @xmath36 combinations in the ranges @xmath37 gev ( @xmath38 mev ) and @xmath39 ( @xmath40 ) .
figure [ f2 ] shows the extracted @xmath28 at three @xmath7 intervals along with the rpr-2011 predictions .
for a few kinematic points the @xmath41 could not be retrieved from the data .
this occurs whenever one or more arguments of the square roots in eq .
( [ eq:1 ] ) become negative due to finite experimental error bars .
the rpr-2011 model offers a fair description of the @xmath8 dependence of the extracted @xmath41 except for the most forward angles at @xmath42 gev .
furthermore , the data confirm the predicted dominance of the @xmath43 .
inferring the nta phases @xmath44 from data requires measured double asymmetries .
complete sets of the first kind , which involve seven observables ( e.g. @xmath45 ) , lead to the following set of nonlinear equations for the phases @xcite @xmath46 where @xmath47 and @xmath48 .
solutions to the above set of nonlinear equations gives the phases @xmath49 for given moduli @xmath50 .
we stress that single - polarization observables are part of any complete set as they provide the information about the moduli .
double polarization observables are required to get access to the phases . in all practical situations one has @xmath51 .
finite error bars introduce a bias for the choices made with regard to the reference phase ( here , @xmath52 ) for the above equations .
a consistent set of estimators @xmath53 for the independent phases ( insensitive to choices made with regard to the reference phase ) has been proposed in ref .
@xcite .
to date , the published double polarization observables for @xmath15 do not allow one to extract the phases of the nta .
we have conducted studies with pseudo - data generated by the rpr-2011 model for @xmath0 .
we have considered ensembles of 200 pseudo - data sets each containing samples of 50 events for the asymmetries @xmath54 .
the pseudo - data are drawn from gaussians with the rpr-2011 prediction as mean and a given @xmath55 as standard deviation .
the retrieved @xmath56 do not necessarily comply with the input amplitudes .
there are various sources of error : ( i ) imaginary solutions for the moduli ; ( ii ) imaginary solutions for the phases ; ( iii ) incorrect solutions which stem from the fact that @xmath57 can not be exactly obeyed for data with finite errors .
we find that the amount of incorrect and imaginary solutions is much larger for the phases than for the moduli .
the frequency of finding imaginary solutions can be dramatically reduced by improving on the experimental resolution @xmath55 .
we have sketched a possible roadmap for reaching a status of complete information in pseudoscalar - meson photoproduction .
we suggest that the use of transversity amplitudes is tailored to the situation that experimental information about @xmath58 is more abundant ( and most often more precise ) than for the double polarization observables .
linear equations connect @xmath59 to the moduli @xmath28 of the nta .
an analysis of @xmath60 data for @xmath61 from graal allowed us to extract the @xmath28 in the majority of considered @xmath62 combinations .
extracting the nta independent phases is far more challenging as they are connected to the double asymmetries by means of nonlinear equations .
it has been suggested @xcite that over - complete sets which involve more than seven polarization observables may provide a solution to tackle the problem of extracting the relative phases of the amplitudes from the data .
this work is supported by the research council of ghent university and the flemish research foundation ( fwo vlaanderen ) .
0 t. vrancx , j. ryckebusch , t. van cuyck , p. vancraeyveld , _ phys .
c _ * 87 * , 055205 ( 2013 ) . l. de cruz , j. ryckebusch , t. vrancx , p. vancraeyveld , _ phys .
c _ * 86 * , 015212 ( 2012 ) . l. de cruz , t. vrancx , p. vancraeyveld , j. ryckebusch , _ phys .
lett . _ * 108 * , 182002 ( 2012 ) .
p. vancraeyveld , l. de cruz , j. ryckebusch , t. vrancx , _ nucl .
_ * a897 * , 42 ( 2013 ) .
a. lleres _ et al . _
( graal collaboration ) , _ eur .
phys . j. _
* a31 * , 79 ( 2007 ) and * a 39 * , 149 ( 2009 ) .
wen - tai chiang and f. tabakin , _ phys .
c _ * 55 * , 2054 ( 1997 ) . | a possible roadmap for reaching a status of complete information in @xmath0 is outlined . |
all @xmath0-ray agns are identified as flat - spectrum radio sources .
the third catalog of high - energy @xmath0-ray sources detected by the energetic gamma ray experiment telescope ( egret ) on the compton gamma ray observatory ( cgro ) includes 66 high - confidence identifications of blazars and 27 lower confidence potential blazar identifications @xcite .
this offers a good sample for the explorations on the radiative mechanisms of @xmath0-rays from agns .
the violent variations in very short time - scales imply that the @xmath0-ray emission is closely related with the relativistic jets in blazars .
there are two kinds of models , namely , leptonic models and hadronic models , proposed for @xmath0-ray emission in blazars ( see mukherjee 2001 for a review ) . according to the different origins of the soft photons , the leptonic models can be classified as two groups : synchrotron self - compton ( ssc ) models and erc models ( see sikora @xmath7 madejski 2001 for a review ) . in the frame of ssc models ,
the synchrotron photons are both produced and compton up - scattered by the same population of relativistic electrons in the jets of @xmath0-ray blazars .
the synchrotron radiation is responsible for the low energy component in radio bands , and the synchrotron photons are compton up - scattered to @xmath0-ray photons by the same population of relativistic electrons in the jets @xcite .
however , ssc models meet difficulties with the observed rapidly variable fluxes in mev - gev for some blazars .
it has been realized that the processes other than ssc may occur at least in some @xmath0-ray blazars .
one possibility is soft seed photons being from the external radiation fields outside the jets , namely , erc models .
the origins of soft seed photons may include the cosmic microwave background radiation , the radiation of the accretion disk ( including photons from the disk scattered by surrounding gas and dust ) , infrared emission from the dust or / and a putative molecular torus , and broad - line regions , etc @xcite .
recently , @xcite proposed that the external radiation is from the blrs for gev @xmath0-ray blazars with flat @xmath0-ray spectra , while the near - ir radiation from the hot dust is responsible for mev @xmath0-ray blazars with steep @xmath0-ray spectra . in their model ,
the electrons are assumed to be accelerated via a two - step process and their injection function takes the form of a double power law with a break at the energy that divides the regimes for two different electron acceleration mechanisms .
the doppler boosting factor @xmath3 of the jets is a crucial quantity in studying the physical processes at work in the regions near the massive black holes in agns .
several different approaches are proposed to estimate the doppler factors @xmath3 of the jets in agns .
@xcite derived the synchrotron self - compton doppler factor @xmath8 of the jets in agns from the vlbi core sizes and fluxes , and x - ray fluxes , on the assumption of the x - ray emission being produced by the ssc processes in the jets .
@xcite assumed energy equipartition between the particles and the magnetic fields in the jet components , and derived the equipartition doppler factor @xmath9 .
more recently , the variability doppler factor @xmath10 is derived on the assumption that the associated variability brightness temperature of total radio flux density flares are caused by the relativistic jets @xcite . in this paper , we use broad - line and @xmath0-ray emission data to derive the doppler boosting factors @xmath11 for a sample of @xmath0-ray blazars in the frame of the erc model .
the cosmological parameters @xmath12=75 km s@xmath13mpc@xmath13 and @xmath14=0.5 have been adopted in this work .
in this work , we calculate the power of @xmath0-ray emission from blazars by the external compton processes mainly following the calculations of @xcite . we then extend it to the case of the soft seed photons being from the blrs of the blazars . assuming that blobs are moving with a constant velocity at an angle @xmath15=cos @xmath16 with respect to the line of sight , we have the inverse compton scattered @xmath0-ray flux density @xmath17 @xcite : @xmath18 where @xmath11 is the doppler factor , @xmath19 is the @xmath0-ray photon energy spectral index , @xmath20 is the cosmological redshift , @xmath21 is the energy density of the soft seed photons in the blob s frame in units of ergs @xmath22 , @xmath23 is the luminosity distance , @xmath24 is the dimensionless monochromatic mean soft photon energy in units of the electron rest mass energy , @xmath25 is the number density of electron in the blob , @xmath26 is the volume of the blob , and @xmath27 is the thomson cross section .
an optically thin , homogenous spherical blob is assumed to be located very near the central black hole , so most of the soft seed photons from the blrs only experience one or a few inverse compton scatters in the blob . the total cross section @xmath28 , where @xmath29 is the radius of the blob . for the homogenous spherical blob , we have @xmath30 , where @xmath31 is the optical depth of the blob .
as we do not know the value of the electron density @xmath25 in the blob , the parameter @xmath32 , which is required @xmath33 , is adopted to describe the optical depth of the blob in our calculations . for @xmath0-ray blazars ,
the jets are believed to move in the direction very close to the line of sight , so @xmath34 is taken and eq .
( 1 ) becomes @xmath35 the soft seed photons are assumed to be from the different line emission .
the observed @xmath0-ray flux density is the sum of inverse compton scattered flux densities contributed by different emission lines : @xmath36 the size of the blob can be estimated from the observed @xmath0-ray variability time - scale .
@xcite derived the size of the blobs @xmath37 ( @xmath38 ) from the time - scale for substantial @xmath0-ray variations of 3c279 , if a @xmath39 black hole is present in 3c279 .
@xcite derived the upper limit on the blob - to - disc distance from the observed variability timescale @xmath40 of the @xmath0-ray fluxes .
as the timescale @xmath41 , the radii of the blobs are around @xmath42 cm . as the typical variable timescale , @xmath43 day
, we can infer the radii of the blobs are @xmath44 cm if the doppler factor @xmath45 is adopted . for typical radio - loud quasars having masses of black holes around @xmath46 ,
so the radii @xmath29 of the blobs are around @xmath47 .
@xcite employed the erc model to explain @xmath0-ray emission from the blazars .
the comparisons of their model with the observed spectrum of 3c273 give the radius of the blob @xmath48 cm .
the central black hole mass of 3c273 is around @xmath49 @xcite .
thus , we get @xmath50 for the blob in 3c273 . in this paper , we take @xmath51 in our calculations .
the observed flux of the broad emission line is @xmath52 where @xmath53 is the luminosity of a single broad emission line .
the energy density of this emission line in the blob is given by @xmath54 where @xmath21 is monochromatic seed photon energy density of the broad line in the blobs and @xmath55 is the radius of the blr . the total soft seed photon energy density @xmath56 is the sum of all emission lines
( 5 ) is only valid for the blob near the central black hole . at further distance away from the central black hole ,
the energy density of blr photons in the blob is anisotropic and can be treated analogously to the approach for the dust torus @xcite . in this paper , however , we only consider the case of the blob being located near the central black hole .
based on the long - time observations of 17 nearby seyfert 1 galaxies and 17 palomar - green quasars , @xcite found an empirical relation between the size of the blr @xmath1 of h@xmath57 and the monochromatic continuum luminosity at 5100 @xmath58 : @xmath59 @xcite suggested that mgii is a low - ionization line as h@xmath57 , so that mgii is expected to be produced in the same region as h@xmath57 .
for a sample of 22 objects , they found a correlation between the fwhm of mgii and h@xmath57 : @xmath60 .
it is therefore reasonable to expect that mgii and h@xmath57 are produced in the same region . using the same sample and the size of the blr derived by @xcite
, they obtained a correlation between @xmath1 and the monochromatic continuum luminosity at 3000 @xmath61 : @xmath62 which has been corrected to the cosmological parameters used in this paper .
@xcite suggested a relation between the black hole mass @xmath63 and the monochromatic continuum luminosity at 1350 @xmath61 .
the black hole mass @xmath63 is the central black hole mass based on the calibration of the single - epoch uv spectral measurements .
they derived @xmath64 where @xmath65 is the full width at half - maximum of the line civ .
the relation between @xmath1 and the monochromatic continuum luminosity at 1350 @xmath61 is therefore available : @xmath66 now , we can derive the sizes of broad emission lines h@xmath57 , mgii , and civ from the optical or uv continuum luminosity by using the empirical relations ( 6 ) , ( 7 ) and ( 9 ) , respectively .
we can estimate the central black hole masses of agns on the assumption that the motions of the gases in blrs are virilized from the width of the emission line , if the size of the blr is available .
we use the empirical relation between the size of the blr @xmath1 and the optical continuum luminosity @xmath67 given in last subsection to estimate the size of the blr @xmath1 .
the black hole mass is given by @xmath68 if the fwhm of any broad - line of h@xmath57 , mgii , and civ is available @xcite .
the doppler factor @xmath11 is given by @xmath69 using the observed @xmath0-ray flux , the doppler factor @xmath11 is available from eq . (
11 ) , if the soft seed photon energy density @xmath21 in the blob , the radius of the blob @xmath29 , and @xmath32 are available . as the parameter @xmath32 is required to be less than unit
, we can obtain the lower limit of @xmath11 , if @xmath70 is adopted .
we search the literature and collect all available data of broad emission lines of @xmath0-ray agns in egret catalog iii .
the selection criterion is that the sources of which the fwhm of at least one of the following broad emission lines h@xmath57 , mgii and civ and the fluxes are available .
this leads to 36 sources , of which 30 sources are the high - confidence identification blazars and 6 sources are the lower confidence potential blazar identifications listed in @xcite . for the source 0954@xmath71658 ,
only the fwhm and the flux of the broad line h@xmath19 are available .
we estimate the blr size of h@xmath19 using the empirical relation @xmath72 proposed by @xcite . in principle , all broad emission line fluxes are needed to calculate the total soft seed photon energy in the blobs contributed by different broad emission lines . usually , the fluxes of only one or several broad emission lines are available for most sources in our sample due to the restriction of redshift . in our calculations , we consider fluxes for the following lines : h@xmath57 , mgii , civ , ciii , ly@xmath19 and h@xmath19 , which contribute about 60 per cent of the total broad line emission .
we use the line ratios presented by @xcite , in which the relative strength of ly@xmath19 is taken as 100 , to calculate the total broad line emissions .
@xcite added the contribution from the flux of h@xmath19 @xmath73 , with a value of 77 .
this gives a total relative flux @xmath74 ( narrow lines are not included ) .
for the lines h@xmath57 , mgii and civ , we use eq .
( 6 ) , ( 7 ) and ( 9 ) to estimate their radii of the blrs .
for other lines of which the fwhm is available , we compare their fwhm with that of any one of those three lines h@xmath57 , mgii and civ to estimate the radii of the blrs .
we use eq .
( 5 ) to calculate the soft photon energy density in the blob for the lines with available fwhm and flux data .
we calculate the total soft seed photon energy density @xmath75 in the blob as @xmath76 where @xmath77 is the photon density in the blob contributed by the lines with available emission line data , @xmath78 is the sum of relative fluxes of these lines .
using the method described in @xmath79 2 , we derive the parameters for all the sources in our sample .
all the observational data and the results are listed in tables @xmath80 .
the @xmath0-ray emission is violently variable . for each source , we derived two values of the doppler factor @xmath11 corresponding to the lowest and highest values of the observed @xmath0-ray fluxes . in figs .
[ 1 ] and [ 2 ] , we compare the doppler factors @xmath11 for the lowest and highest @xmath0-ray flux cases with @xmath10 derived from the variability time - scale of radio emission , respectively @xcite .
it is found that the values of @xmath10 are higher than that of @xmath11 for most sources in our sample . in fig . [ 3 ] , we plot the relation between the core dominance parameter @xmath6 and the doppler factor @xmath81 derived from the lowest @xmath0-ray flux .
the core dominance parameters @xmath6 are taken from @xcite , except those of the sources of 0454@xmath82234 , 0521@xmath82365 and 1741@xmath82038 taken from @xcite . a correlation between @xmath6 and @xmath83
is found at a significant level of 97.5 per cent ( spearman correlation coefficient @xmath84 ) .
the results are plotted in fig . [ 4 ] for the doppler factor @xmath85 derived from the highest @xmath0-ray flux .
the significant level of the correlation becomes 95.2 per cent .
the distribution of the central black hole mass @xmath86 in @xmath0-ray blazars is plotted in fig . [ 5 ] .
the black hole masses @xmath86 are in the range of @xmath87 with an average of around @xmath88 .
we plot the relation of the black hole mass @xmath86 with the lowest and highest @xmath0-ray luminosity @xmath89 in figs . [ 6 ] and [ 7 ] , respectively .
the correlations between them are found at the significant levels of 97.8 and 96.6 per cent for the lowest and highest @xmath0-ray emissions , respectively .
in this work , we estimate the size of the blr from the optical or uv continuum luminosity . the observed optical / uv continuum emission from @xmath0-ray blazars is a mixture of the emission from the disks and jets .
the optical / uv continuum emission may be strongly beamed to us , since the viewing angles of the relativistic jets in these blazars are small .
the emission from the jets may dominate over that from the disks at least in some blazars .
so , the blr size derived from the observed optical / uv continuum luminosity may be over estimated , and the derived black hole mass is a upper limit @xcite .
the size of the blob @xmath29 is assumed to be 120@xmath90 in our calculations ( see discussion in section @xmath79 2.2 ) . as the black hole mass @xmath86 is derived from the size of the blr and the line width , the size of the blob @xmath29 is proportional to the blr size @xmath1 . from eq .
( 5 ) , we know that the energy density in the blob @xmath91 .
the derived doppler factor @xmath11 in this work is therefore independent of the blr size @xmath1 ( see eq . ( 11 ) ) .
it indicates that the overestimate of the blr size caused by the contamination of the optical / uv continuum emission has not affected the values of derived doppler factor @xmath11 . in our calculations
, we adopted a parameter @xmath32 to describe the optical depth of the blob , since we do not know the exact value of the actual number density @xmath25 of the electrons in the blob . from eq .
( 11 ) , it is found that the doppler factor @xmath92 . for a typical value of @xmath93
, we found that the derived @xmath11 becomes about twice of its initial value , if the parameter @xmath32 is changed from 1 to 0.01 .
the uncertainty of @xmath32 in our calculations would not affect our results significantly .
we only consider the photons from the blrs as the soft seed photons .
other sources of soft seed photons ( e.g. , emission from the disks , synchrocyclotron emission in the blobs , etc . ) have not been included in the calculations .
the doppler factor @xmath11 derived in this work are in the range of 3.04 and 17.00 , if @xmath94 is adopted .
it is generally consistent with the results derived by other methods @xcite .
it may imply that the soft seed photons are indeed mainly from the blrs in these sources .
the variability doppler factors @xmath10 of 20 @xmath0-ray blazars in our sample have been derived from the radio variable time - scales are available @xcite .
comparing with the doppler factors @xmath11 derived in this work , we find that the variability doppler factor @xmath10 are higher than @xmath11 derived in this work for most sources ( see figs . [ 1 ] and [ 2 ] ) .
this may be due to the fact that we adopt @xmath94 and the derived doppler factor @xmath11 are then the lower limits .
the value of @xmath32 may possibly be estimated roughly by letting @xmath95 .
it can be found that the values of @xmath32 can be low as @xmath96 for some sources ( see eq . ( 11 ) , figs . [ 1 ] and [ 2 ] ) . the core dominance parameter @xmath97 , where @xmath98 and @xmath99 are the core and extended flux densities at 5 ghz in the rest frame of the source .
the core dominance parameter @xmath6 is believed to be a good indicator of the doppler factor of the jet .
we find a significant correlation between the core dominance parameter @xmath6 and the derived doppler factor @xmath11 .
this implies that the derived doppler factors @xmath11 are reliable for the @xmath0-ray blazars in this sample and the soft seed photons are mainly from the blrs .
it was argued that the ssc radiative mechanism is dominant in bl lac objects @xcite .
however , we can not find significant differences between bl lac objects and quasars in our investigations .
the reason may be that the bl lac objects in our sample have relative stronger broad line emission compared with other bl lac objects listed in egret catalog iii , since we only select the @xmath0-ray blazars with broad - line emission data as our sample .
so these bl lac objects in our sample are more like quasars as they have rather strong broad - line emission @xcite . for these bl lac objects with relative strong broad - line emission ,
the erc radiative mechanism may probably be important as that in quasars .
the black hole masses of @xmath0-ray blazars in this sample are in the range of @xmath100 .
this is consistent with some previous results for radio - loud agns @xcite .
we have found significant correlations between the black hole mass @xmath86 and the @xmath0-ray luminosity @xmath89 .
the reason may be that the @xmath0-ray flux depends sensitively on the doppler factor @xmath11 and the values of the doppler factor of the @xmath0-ray agns in our sample spread over a large range ( @xmath4 ) .
on the other hand , it is expected that the @xmath0-ray flux is independent of the black hole mass in erc mechanisms with the soft seed photons being from the blrs ( see eq .
( 2 ) and discussion in the second paragraph of this section ) .
this work is supported by nsfc(no .
10173016 ) and the nkbrsf ( no .
g1999075403 ) .
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 aeronautic and space administration .
arbeiter , c. , pohl , m. , & schlickeiser , r. 2002 , , 386 , 415 baldwin , j. a. 1975 , , 201 , 26 baldwin , j. a. , wampler , e. j. , & gaskell , c. m. 1989 , , 338 , 630 cao , x. , & jiang , d. r. 2001 , , 320 , 347 celotti , a. , padovani , p. , & ghisellini , g. 1997 , , 286 , 415 corbin , m. r. 1992 , , 391 , 577 corbin , m. r. , & boroson , t. a. 1996 , , 107 , 69 dermer , c. d. , sturner , s. j. , schlickeiser , r. 1997 , , 109 , 103 dermer , c. d. , & schlickeiser , r. 2002 , , 575 , 667 dondi , l. , & ghisellini , g. 1995 , , 273 , 583 francis , p. j. , hewett , p. c. , foltz , c. b. , chaffee , f. h. , weymann , r. j. , morris , s. l. 1991 , , 373 , 465 fricke , k. j. , kollatschny , w. , & witzel , a. 1983 , , 117 , 60 ghiselini , g. , padovani , p. , celotti , a. , & maraschi , l. 1993 , , 407 , 65 ghisellini , g. , & madau , p. 1996
, , 280 , 67 gu , m. f , cao , x. , & jiang , d. r. 2001 , , 327 , 1111 gijosa , a. , & daly , r. a. 1996 , , 461 , 600 hartman , r. c. , et al .
1996 , , 461 , 698 hartman , r. c. et al . 1999 , , 123 , 79 hunstead , r. w. , murdoch , h. s. , & shobbrook , r. r. 1978 , , 185 , 149 jackson , n. , & browne , w. a. 1991 , , 250 , 414 kaspi , s. , smith , p. s. , netzer , h. , maoz , d. , jannuzi , b. t. , & giveon , u. 2000 , , 533 , 631 lhteenmki , a. , & valtaoja , e. 1999 , , 521 , 493 laro , a. 2000 , , 543 , 111 lawrence , c. r. , zucker , j. r. , readhead , a. c. s. , unwin , s. c. , pearson , t. j. , & xu , w. 1996 , , 107 , 541 maraschi , l. , ghisellini , g. , & celotti , a. 1992 , , 397 , l5 mariziani , p. , sulentic , j. w. , dultzin - hacyan , d. , calvani , m. , & moles , m. 1996 , , 104 , 37 montigny , v. c. , et al .
1997 , , 483 , 161 mukherjee , r. 2001 , in aip conf .
558 , high energy gamma - ray astronomy , ed . by f. a. aharonian & h. j. vlk ( new york : aip ) , 324 mclure , r. j. , & dunlop , j. s. 2001 , , 327 , 199 mclure .
r. j. , & jarvis . , m. j. 2002 , , 337 , 109 osmer , p. s. , porter , a. c. , green , r. f. 1994 , , 436 , 678 rudy , r. j. 1984 , , 284 , 33 scarpa , r. , & falomo , r. 1997 , , 325 , 109 sikora , m. , & madejski , g. 2001 , in aip conf . proc .
558 , high energy gamma - ray astronomy , ed . by f. a. aharonian & h. j. vlk ( new york : aip ) , 275 sikora , m. , blazejowki , m. , moderski , r. , & madejski , g. m. 2002 , , 577 , 78 stickel , m. , fried , j. w. , & khr , h. 1989 , , 80 , 103 stickel , m. , & khr , h. 1993 , , 101 , 521 stickel , m. , khr , h. , & fried , j. w. 1993 , , 97 , 483 stockton , a. , & mackenty , j. w. 1987 , , 316 , 584 ulrich , m. h. , et al .
1980 , , 192 , 561 vron - cetty , m. p. , & vron , p. 2000 , , 10 , 81 vestergaard , m. 2002 , , 571 , 733 wilkes , b. j. 1986 , , 218 , 331 wills , b. j. , et al .
1995 , , 447 , 139 cccccccc 0119@xmath71041@xmath101 & 0.637 & h@xmath57 & 4444 & g01&8.855@xmath102 & 46.309 & 46.909 + 0208@xmath82512@xmath103 & 1.003 & h@xmath57 & 4070 & w86 & 9.208 & 47.916 & 48.497 + 0336@xmath82019 & 0.852 & h@xmath57 & 4875 & g01 & 9.459@xmath102 & 47.903 & 48.580 + 0414@xmath82189 & 1.536 & mgii & 1713 & h78 & 8.075 & 47.310 & 47.996 + 0420@xmath82014 & 0.915 & h@xmath57 & 3000 & g01 & 9.510@xmath102 & 46.987 & 47.826 + 0440@xmath82003 & 0.844 & mgii & 3900 & b89 & 8.813 & 47.071 & 47.908 + 0454@xmath82234@xmath103 & 1.009 & mgii & 7891 & s89 & 9.173 & 46.823 & 47.082 + 0454@xmath82463 & 0.858 & mgii & 2700 & f83 & 8.626 & 46.587 & 47.204 + 0458@xmath82020 & 2.286 & civ & 5200 & b89 & 8.662 & 47.878 & 48.734 + 0521@xmath82365@xmath104@xmath101 & 0.055 & h@xmath57 & 6758 & sk93 & 8.708 & 44.584 & 44.889 + 0537@xmath82441@xmath103 & 0.894 & mgii & 3200 & w86 & 8.709 & 47.228 & 47.970 + 0539@xmath82057@xmath101 & 0.839 & mgii & 8598 & s93 & 9.411 & 48.015 & 48.015 + 0804@xmath71499@xmath101 & 1.430 & mgii & 2259 & l96 & 8.634 & 47.522 & 47.781 + 0836@xmath71710 & 2.172 & mgii & 3062 & l96 & 9.450 & 47.723 & 48.312 + 0851@xmath71202@xmath103 & 0.306 & h@xmath57 & 3441 & s89 & 8.919 & 46.215 & 46.427 + 0954@xmath71556 & 0.901 & civ & 10503 & w95 & 9.718 & 46.990 & 47.851 + 0954@xmath71658@xmath103 & 0.368 & h@xmath19 & 2084 & l96 & 8.526 & 46.182 & 46.618 + 1127@xmath82145@xmath101 & 1.187 & mgii & 2750 & w86 & 8.850 & 47.168 & 47.964 + 1222@xmath71216 & 0.435 & h@xmath57 & 2197 & sm87 & 8.435 & 46.235 & 47.078 + 1226@xmath71023 & 0.158 & h@xmath57 & 3416 & k00 & 9.298 & 45.268 & 46.026 + 1229@xmath82021 & 1.045 & mgii & 5000 & b89 & 9.118 & 46.698 & 47.199 + 1253@xmath82055@xmath103 & 0.538 & h@xmath57 & 3100 & g01 & 8.912@xmath102 & 46.770 & 48.229 + 1331@xmath71170 & 2.084 & civ & 5740 & c92 & 9.171 & 47.472 & 48.348 + 1334@xmath82127@xmath103 & 0.539 & mgii & 7077 & sk93 & 9.294 & 46.188 & 46.753 + 1424@xmath82418 & 1.522 & mgii & 9144 & s89 & 9.652 & 47.860 & 48.418 + 1504@xmath82166@xmath101 & 0.876 & mgii & 5139 & h78 & 9.130 & 47.450 & 47.754 + 1510@xmath82089 & 0.361 & h@xmath57 & 3180 & g01 & 9.130@xmath102 & 46.233 & 46.827 + 1611@xmath71343 & 1.401 & h@xmath57 & 5600 & g01 & 10.05@xmath102 & 47.716 & 48.276 + 1633@xmath71382 & 1.814 & mgii & 9657 & l96 & 10.09 & 48.332 & 48.861 + 1725@xmath71044 & 0.296 & h@xmath57 & 2090 & g01 & 8.545@xmath102 & 46.001 & 46.251 + 1739@xmath71522 & 1.375 & mgii & 2559 & l96 & 8.866 & 47.410 & 48.076 + 1741@xmath82038 & 1.054 & mgii & 11574 & s89 & 9.566 & 47.231 & 47.851 + 2230@xmath71114 & 1.037 & civ & 3340 & w95 & 9.164 & 47.217 & 47.847 + 2251@xmath71158 & 0.859 & h@xmath57 & 2800 & g01 & 9.644@xmath102 & 47.467 & 48.141 + 2320@xmath82035 & 1.411 & civ & 3300 & b89 & 8.079 & 47.605 & 48.273 + 2351@xmath71456 & 1.922 & mgii & 5118 & l96 & 9.190 & 47.867 & 48.427 + ccccccc 0119@xmath71041 & h@xmath57 & jb91 & @xmath821.896 + 0208@xmath82512 & mgii , ciii & w86 , s97 & @xmath822.461 + 0336@xmath82019 & h@xmath57 , mgii , ciii & b89 , jb91 & @xmath822.794 + 0414@xmath82189 & mgii , civ , ciii & h78 & @xmath710.026 + 0420@xmath82014 & mgii , ciii & b89 , s97 & @xmath822.924 + 0440@xmath82003 & mgii & b89 & @xmath822.069 + 0454@xmath82234 & mgii , ciii & s89 & @xmath821.872 + 0454@xmath82463 & mgii & f83 & @xmath821.837 + 0458@xmath82020 & civ & b89 & @xmath821.702 + 0521@xmath82365 & h@xmath57 & sk93 & @xmath822.722 + 0537@xmath82441 & mgii & w86 , s97 & @xmath822.234 + 0539@xmath82057 & mgii & s93 & @xmath821.390 + 0804@xmath71449 & mgii , civ , ciii & l96 & @xmath710.411 + 0836@xmath71710 & mgii , civ , ciii , ly@xmath19 & l96&@xmath710.258 + 0851@xmath71202 & h@xmath57 & s89 & @xmath824.109 + 0954@xmath71556 & civ , ly@xmath19 & w95 & @xmath822.063 + 0954@xmath71658 & h@xmath19 & l96 & @xmath824.425 + 1127@xmath82145 & mgii , civ , ciii & w86 , w95 & @xmath820.317 + 1222@xmath71216 & h@xmath57 & sm87 & @xmath822.109 + 1226@xmath71023 & h@xmath57 , mgii , civ , ciii , ly@xmath19 , h@xmath19 & k00 , u80 , cb96 , b87 & @xmath822.067 + 1229@xmath82021 & mgii & b98 & @xmath821.729 + 1253@xmath82055 & h@xmath57 , mgii , civ , ciii , ly@xmath19 & w95 , m96 & @xmath821.797 + 1331@xmath71170 & civ & c92 & @xmath821.716 + 1334@xmath82127 & mgii & sk93 & @xmath822.370 + 1424@xmath82418 & mgii , ciii & s89 & @xmath822.194 + 1504@xmath82199 & mgii & h78 & @xmath821.898 + 1510@xmath82089 & h@xmath57 & o94 & @xmath822.595 + 1611@xmath71343 & h@xmath57 , civ , ciii , ly@xmath19 & w95 & @xmath823.075 + 1633@xmath71382 & mgii , civ , ciii , ly@xmath19 & l96 & @xmath822.537 + 1725@xmath71044 & h@xmath57 & r84 & @xmath822.819 + 1739@xmath71522 & mgii , civ , ciii & l96 & @xmath821.446 + 1741@xmath82038 & mgii & s89 & @xmath821.816 + 2230@xmath71114 & civ , ciii , ly@xmath19 & w95 & @xmath821.856 + 2251@xmath71158 & h@xmath57 , civ , ciii , ly@xmath19 & w95 & @xmath822.632 + 2320@xmath82035 & civ & b89 & @xmath821.653 + 2351@xmath71456 & mgii , civ , ciii & l96 & @xmath821.480 + cccccccc 0119@xmath71041 & 2.63 & 5.1 & 20.3 & 5.98 & 7.23 & ... & 15.85 + 0208@xmath82512 & 1.99 & 35.2 & 134.1 & 7.76 & 9.71 & ... & ... + 0336@xmath82019 & 1.84 & 37.4 & 177.6 & 7.27 & 9.56 & 19.01 & 31.62 + 0414@xmath82189 & 3.25 & 10.2 & 49.5 & 7.63 & 9.19 & ... & ... + 0420@xmath82014 & 2.44 & 9.3 & 64.2 & 6.61 & 8.75 & 11.72 & 251.3 + 0440@xmath82003 & 2.37 & 12.5 & 85.9 & 8.52 & 11.3 & 11.46 & 19.95 + 0454@xmath82234 & 3.14 & 8.1 & 14.7 & 5.94 & 6.39 & ... & 5.00 + 0454@xmath82463 & 2.75 & 5.5 & 22.8 & 7.15 & 8.65 & ... & ... + 0458@xmath82020 & 2.45 & 9.5 & 68.2 & 10.1 & 13.5 & 17.80 & 5.01 + 0521@xmath82365 & 2.63 & 15.8 & 31.9 & 4.46 & 4.91 & ... & 1.00 + 0537@xmath82441 & 2.41 & 16.5 & 91.1 & 9.03 & 11.6 & ... & 199.6 + 0539@xmath82057 & 2.00 & 66.5 & 66.5 & 4.60 & 4.60 & ... & 7.61 + 0804@xmath71499 & 2.15 & 8.3 & 15.1 & 3.82 & 4.20 & 26.21 & 151.9 + 0836@xmath71710 & 2.62 & 8.6 & 33.4 & 3.19 & 3.85 & 10.67 & 34.93 + 0851@xmath71202 & 2.03 & 9.7 & 15.8 & 9.62 & 10.4 & 18.03 & 781.0 + 0954@xmath71556 & 2.12 & 6.5 & 47.2 & 3.44 & 4.73 & 4.63 & 10.81 + 0954@xmath71658 & 2.08 & 6.6 & 18.0 & 14.5 & 17.0 & 6.62 & 110.5 + 1127@xmath82145 & 2.70 & 9.9 & 61.8 & 4.87 & 6.24 & ... & 50.12 + 1222@xmath71216 & 2.28 & 6.9 & 48.1 & 7.12 & 9.57 & 8.16 & ... + 1226@xmath71023 & 2.58 & 8.5 & 48.3 & 3.04 & 3.87 & 5.71 & 6.31 + 1229@xmath82021 & 2.85 & 4.9 & 15.5 & 5.45 & 6.33 & ... & 3.43 + 1253@xmath82055 & 1.96 & 9.3 & 267.3 & 5.18 & 9.14 & 16.77 & 12.59 + 1331@xmath71170 & 2.41 & 4.4 & 33.1 & 6.20 & 8.34 & ... & ... + 1334@xmath82127 & 2.62 & 5.5 & 20.2 & 4.71 & 5.64 & ... & 12.87 + 1424@xmath82418 & 2.13 & 15.3 & 55.3 & 5.28 & 6.48 & ... & 2.38 + 1504@xmath82166 & 2.00 & 16.5 & 33.2 & 5.58 & 6.27 & ... & 19.95 + 1510@xmath82089 & 2.47 & 12.6 & 49.4 & 5.84 & 7.11 & 13.18 & 31.62 + 1611@xmath71343 & 2.42 & 19.0 & 68.9 & 6.36 & 7.68 & 5.04 & 25.12 + 1633@xmath71382 & 2.15 & 31.8 & 107.5 & 5.22 & 6.33 & 8.83 & 79.44 + 1725@xmath71044 & 2.67 & 13.3 & 23.7 & 8.68 & 9.39 & 2.46 & 125.8 + 1739@xmath71522 & 2.42 & 9.7 & 44.9 & 6.72 & 8.40 & 12.12 & 50.1 + 1741@xmath82038 & 2.42 & 11.7 & 48.7 & 4.44 & 5.47 & 8.92 & 4.00 + 2230@xmath71114 & 2.45 & 12.1 & 51.6 & 5.92 & 7.31 & 14.23 & 25.12 + 2251@xmath71158 & 2.21 & 24.6 & 116.1 & 5.97 & 7.61 & 21.84 & 15.85 + 2320@xmath82035 & 2.00 & 8.2 & 38.2 & 12.1 & 15.6 & ... & ... + 2351@xmath71456 & 2.38 & 11.8 & 42.8 & 6.37 & 7.71 & ... & ... + | the sizes of the broad - line regions ( blrs ) in @xmath0-ray active galactic nuclei ( agns ) are estimated from their optical continuum luminosity by using the empirical relation between @xmath1 and @xmath2 . using the broad emission line data
, we derive the photon energy density in the relativistic blobs near the massive black holes in agns .
we calculate the power of the broad - line photons compton up - scattered by the relativistic electrons in the blobs .
compared with observed @xmath0-ray emission data , the doppler factors @xmath3 of the blobs for a sample of 36 @xmath0-ray agns are derived , which are in the range of @xmath4 .
we estimate the central black hole masses of these @xmath0-ray agns from the sizes of the blr and the broad line widths .
it is found that the black hole masses are in the range of @xmath5 .
a significant correlation is found between the doppler factor @xmath3 and the core dominance parameter @xmath6 .
the results are consistent with the external radiation compton ( erc ) models for @xmath0-ray emission from agns .
the soft seed photons are probably from the broad - line regions . |
this is our second paper of a series devoted to the thermal properties of a pure gauge theory within a five - dimensional framework nowadays known as ads / qcd . we explore further the model proposed in the first paper @xcite .
it is based on the following euclidean background metric @xmath0 where @xmath1 is a periodic variable of period @xmath2 such that @xmath3 , with @xmath4 , the temperature . at zero temperature
, we have in fact the slightly deformed @xmath5 metric .
such a deformation is notable because it results in a regge - like spectrum .
this fact allows one to fix the value of @xmath6 from the @xmath7 meson trajectory . for our purposes ,
we use the estimate of @xcite @xmath8 it is worth noting that the metric contain any free fit parameter . moreover , the only dimensionful parameter of our model is the regge slope of meson trajectories .
it is quite unusual for qcd , where it is @xmath9 .
thus , evaluations of thermodynamic quantities and the polyakov loop we are going to undertake can be considered as a further consistency check of our model . for the case of interest ,
let us briefly point out a couple of facts .
\(i ) the free energy of a heavy ( static ) quark - antiquark pair at finite temperature is expressed in terms of a correlator of two polyakov loops @xcite @xmath10 with @xmath11 and @xmath12 being a point in @xmath13 . in the free energy
is defined up to a normalization constant @xmath14 which is related to the infinite self - energy of the quarks .
there is a subtle point here . in the literature @xmath15
is often called the heavy quark potential at finite temperature .
apparently , such a definition discards the entropy contribution .
an order parameter for the confinement - deconfinement phase transition is the expectation value of the polyakov loop .
after the normalization of , it is then @xmath16 \(ii ) in discussing a wilson line within ads / cft ( qcd ) @xcite , one first chooses a contour @xmath17 on a four - manifold which is the boundary of a five - dimensional manifold .
next , one has to study fundamental strings on this manifold such that the string world - sheet has @xmath17 for its boundary .
the expectation value of the loop is schematically given by the world - sheet path integral @xmath18 where @xmath19 denotes a set of world - sheet fields .
@xmath20 is a world - sheet action . in principle , the integral can be evaluated semiclassically in terms of minimal surfaces that obey the boundary conditions .
the result is written as @xmath21 where @xmath22 means a regularized minimal area whose relative weight is @xmath23 .. ] the paper is organized as follows . in section 2 , we develop the framework we will work . in section 3 , we present , on the basis of ads / qcd , a few results on a heavy quark - antiquark pair in a thermal medium . in particular , we exhibit the free energy , the string tension , and the entropy at low temperatures as well as the polyakov loop expectation value .
we conclude in section 4 with a discussion of some open problems .
our basic approach will be as follows .
it is believed that the correlator of the two polyakov loops has a path integral representation like that of the wilson loop .
given the background metric , we can attempt to evaluate the values of the regularized areas @xmath22 .
if we then discard quantum fluctuations of strings , in the case of one dominant exponent we will get that the free energy is simply proportional to a proper @xmath24 @xmath25 note that the regularized areas are defined up to normalization constants .
since these constants are in fact due to the infinite self - energy of the quark sources , we omit @xmath14 .
moreover , @xmath26 can be absorbed into a proper normalization constant too . in this paper
we will use the approximation .
but before going on , let us shortly pause here to gain some intuition about the problem at hand . as in @xcite , we introduce the notion of an effective string tension depending on the fifth coordinate @xmath27
. it is given by @xmath28 now consider the behavior of a string bit in the effective potential @xmath29 .
an important observation is that the form of @xmath30 is temperature dependent .
indeed , a short algebra shows that there is a special value of temperature @xmath31 . below @xmath32
the effective potential has local extrema at @xmath33 while above this temperature the potential is just a decreasing function of @xmath27 as shown in fig.1 .
such a behavior clearly has some of the suspected properties of gauge theory at finite temperature . for temperatures below @xmath32 the string ended on the heavy quark - antiquark pair set at @xmath34 can not get deeper than @xmath35 in @xmath27 direction because a repulsive force prevents it from doing so
this gives rise somewhat of a wall located at @xmath36 .
the large distance physics of the string is determined by this wall .
since the value of the effective potential ( string tension ) at its minimum is not vanishing , the quark - antiquark free energy has a dominant linear term whose coefficient is proportional to @xmath37 .
so , this can indeed be interpreted as the low temperature phase . on the other hand , for temperatures above
@xmath32 the string can get deeper and finally reach the horizon @xmath38 .
the large distance physics is now determined by the near horizon geometry .
the crucial point is that the effective potential vanishes on the horizon . as a result
, there is no linear term in the quark - antiquark free energy .
this can be interpreted as the deconfined phase .
now let us go back to the approach and choose a pair of the contours ( polyakov loops ) living on the boundary ( @xmath34 ) of our five dimensional space .
we set @xmath39 next we want to look for static configurations of the world - sheet action . to this end , we make use of the nambu - goto action equipped with the background metric @xmath40 there are basically two types of configurations to be considered .
one type describes connected surfaces whose boundaries are the polyakov loops , while another describes disconnected surfaces
. our first goal will be to analyze connected surfaces . in this case
the world - sheet coordinates @xmath41 s can be chosen as @xmath42 and @xmath43 .
with such a choice , the action takes the form @xmath44 where @xmath45 .
a prime denotes a derivative with respect to @xmath46 .
it is easy to find the equation of motion for @xmath27 @xmath47 as well as its first integral @xmath48 on symmetry grounds , we have @xmath49 .
this allows us to express the integration constant @xmath50 via the value of @xmath27 at @xmath51 .
so , we get @xmath52 where @xmath53 is defined by and @xmath54 .
next we perform the integral over @xmath55 $ ] of @xmath56 . by virtue of
, it is given by @xmath57 where @xmath17 is a contour in @xmath27 plane .
we look for solutions that obey the following condition @xmath58 .
the reason for this is that , on general grounds , the string ended on the quarks set at @xmath59 reaches the deepest point in @xmath27 direction at @xmath51 . for temperatures below @xmath32 ,
there are two possibilities for @xmath60 and , as a result , we have @xmath61 but otherwise @xmath62 is complex . after a short inspection we find that @xmath63 is a continuously growing function of @xmath64 on the interval @xmath65 $ ] .
moreover , it equals to zero at @xmath66 and goes to infinity as @xmath67 . the function @xmath68 is , unlike @xmath63 , decreasing .
it goes from its maximum ( finite ) value @xmath69 at @xmath70 to zero at @xmath71 ( on the horizon ) .
thus , the second solution contributes only at distances smaller than @xmath69 . to complete the picture
, we present the plots of @xmath72 in fig.2 .
when temperature is increased , the interval @xmath73 $ ] becomes smaller and finally disappears at @xmath74 . for @xmath75
, a simple analysis leads to the picture that differs noticeably from that of fig.2 .
the point is that both the solutions now contribute only at distances smaller than some finite @xmath76 .
this is illustrated in fig.3 .
now we move on to the second type that describes disconnected surfaces . in this case
a surface contains two pieces each of which has a topology of a cylinder .
the cylinders are stretched from the polyakov loops on the boundary to the horizon .
if we use @xmath42 and @xmath77 as the world - sheet coordinates , the nambu - goto action is then @xmath78 where a dot stands for a derivative with respect to @xmath27 .
the equation of motion for @xmath46 is @xmath79 it is obvious that it has a trivial solution @xmath80 that represents a straight string stretched between the boundary and the horizon . we will call the choice the solution @xmath81 . since this solution makes the dominant contribution , as seen from the integrand in , we will not dwell on other solutions here .
having discussed the solutions , we can now see what happens with the corresponding areas .
we begin with the solution @xmath82 . to this end , we use the first integral to reduce the integral over @xmath46 to that over @xmath27 in . since the integral is divergent at @xmath34 due to the factor @xmath83 in the background metric , in the process we regularize it by imposing a cutoff @xmath84 . at the end of the day
, we have @xmath85 subtracting the @xmath86 term we find a finite result @xmath87+\frac{c(t)}{t } \,,\ ] ] where @xmath14 stands for a normalization constant . for the second solution @xmath88
, it is a little bit tricky because the minimal surface is built by sewing together two pieces .
the first piece comes from @xmath89 defined on the interval @xmath90 $ ] . at @xmath91 ,
it is sewn with the second piece coming from @xmath88 .
the integral is divergent at @xmath34 , so we regularize it by imposing the cutoff @xmath84 as before . after subtracting the divergency , we get @xmath92 note that both the areas are regularized in the same way , so the corresponding normalization constants coincide . for the third solution
, the minimal area can be read off from eq .. one drops the third term as coming from the solution @xmath88 and replaces @xmath93 with @xmath94 in the remaining terms .
the area is then @xmath95 now let us look at the @xmath96 s as functions of @xmath62 .
we begin with sufficiently low temperature .
since the solution @xmath88 is defined only for distances smaller than @xmath69 , it makes sense to first probe asymptotic behaviors near the point @xmath97 .
a short inspection shows that @xmath98 is not bounded from below , while the others are bounded .
this provides a bit of evidence in favor of dominance of @xmath98 at small distances .
further numerical calculations show that this is indeed the case
. moreover , the first solution turns out to be dominant at physical distances too .
that is of primary importance for phenomenology . ]
we present the plot of the regularized areas in fig.4 .
can be fixed from the slope of the heavy quark potential at zero temperature .
we use the estimate of @xcite @xmath99 , here and below . ] from this figure it is also clear that there exists a critical distance @xmath100 such that for larger distances @xmath101 becomes dominant .
there is an apparent reason for this . at very large distances
the quark and the anti - quark decouple from each other . usually , instability occurs in models with dynamical quarks , where a string breaks . in the case of interest
this occurs due to emission of closed strings ( glueballs ) .
as known , in ads / qcd a free particle is described by a straight string stretched between the boundary and the horizon that is nothing but our solution @xmath89 .
when temperature is increased , two effects are seen : ( 1 ) the minimal difference between the functions @xmath98 and @xmath102 becomes smaller .
( 2 ) the critical distance @xmath100 is decreasing . for temperatures close to @xmath74 ,
@xmath98 is no longer dominant at physical distances and it is time to account for the other solutions in the series .
in this section we will consider some applications of the developed formalism .
we will focus on the cases where the approximation might be applicable .
we begin with temperatures which are sufficiently smaller than @xmath32 . in this case
the free energy can be evaluated by using with @xmath103 .
first , we need to fix the normalization constant @xmath14 . in doing so ,
we follow @xcite and look for the small @xmath62 expansion of the free energy . since small distances correspond to small deviations in @xmath27 direction , we need to study the expressions and near @xmath66 .
the asymptotic behavior of @xmath104 is given by @xmath105 where @xmath106 .
in a similar way we find the behavior of @xmath15 .
it is @xmath107 combining this with , we get @xmath108 where @xmath109 and @xmath110 . having derived the small distance expansion of the free energy , we can now get that of the singlet free energy gauge theory . ]
@xmath111 and compare it with the small distance expansion of the heavy - quark potential at zero temperature . to leading order , we have the same coulomb term as in @xcite . to get an agreement at next - to - leading order , we choose @xmath112 finally , we fix the value of @xmath113 by matching it to the constant term of the cornell potential @xcite .
this yields @xmath114 actually , our result shows that at sufficiently small distances the singlet free energy of the pair is temperature independent .
so , the agreement with the lattice data @xcite is very satisfactory at this point .
our next goal is to analyze the long distance behavior of @xmath15 .
as noted in section 2 , large @xmath62 corresponds to @xmath115 .
so , we need to study the behavior of and near @xmath116 . in this case
a crucial observation is that the integrals are dominated by the upper limits , where they take the form @xmath117 .
such an integral may be found in @xcite . at the end of the day
, we have @xmath118 where @xmath119 and @xmath120 .
this means that at long distances the free energy of the pair shows the desired confining behavior @xmath121 with the string tension @xmath122 here @xmath53 denotes a tension at zero temperature .
explicitly , it is given by @xmath123 @xcite .
we have also introduced the critical temperature @xmath124 obtained from the spatial string tension in @xcite . as expected , the string tension @xmath125 is a decreasing function of t. in fig.5 we have plotted @xmath126 against @xmath127 .
we conclude this subsection by making a few remarks .
( i ) it is quite interesting that the tension shows _ very _ little dependence on temperature up to @xmath128 .
( ii ) unlike the coefficient @xmath125 of the linear term in the large distance expansion of @xmath15 , the coefficient @xmath129 of the linear term in the small distance expansion turns out to be independent of temperature .
( iii ) the free energy of the pair is written in parametric form given by eqs . and .
since we do not know how to eliminate the parameter @xmath64 and find @xmath15 as a function of @xmath62 , we present the result of numerical calculations . in fig.6 the free energy @xmath15 versus @xmath62 for a temperature below @xmath32 is shown .
we have already mentioned that in the literature the entropy contributions are sometimes ignored .
it is therefore of great importance to address this issue .
having understood the behavior of the free energy , we can easily investigate the properties of the entropy .
since the free energy shows no temperature dependence at small distances , it makes sense to focus on long distances , where its behavior is given by . in this case
a straightforward calculation leads to the following expression @xmath130 here we treat @xmath131 as a function of two variables .
then , using the relations and , we can compute the entropy density of the pair @xmath132 a closer look at this expression shows that the entropy density grows as the temperature increases .
this is the expected result . to complete the picture
, we plot the entropy density against @xmath127 in fig .
7 . from this figure
, we see that the entropy density is close to zero up to temperatures of order @xmath133 .
thus , in this temperature range , the entropy contributions are indeed negligible .
it is obvious that the situation changes drastically with the temperature growth .
therefore , a natural question to ask is whether the entropy contributions play a major role at finite temperature . to answer this question , we consider the internal energy of the pair . as usual
, it is given by @xmath134 .
it is clear from above that @xmath135 is independent of temperature at small distances , where it coincides with the heavy - quark potential at zero temperature . at large distances ,
the internal energy has a complex dependence on temperature together with a linear growth with @xmath62 .
explicitly , it is given by @xmath136 with the tension @xmath137 the difference with the string tension is due to the second factor that is nothing but the entropy contribution .
a direct but lengthy calculation shows that @xmath138 is a growing function of temperature .
this is in contrast to the behavior of the string tension @xmath139 discussed in section 3.1 .
thus , the entropy contributions _ do _ play a major role at finite temperature .
we conclude the discussion with a few short comments : ( i ) a temperature growth of the singlet internal energy at large distances was observed in the lattice calculations of @xcite .
( ii ) the use of thermal ads as the background metric in the low temperature phase results in zero entropy . , so gives zero .
] this leads to the picture that is completely inconsistent with physics of a pure @xmath140 gauge theory at finite temperature . gauge theories . ]
( iii ) in fig.8 we plot @xmath138 versus temperature as provided by the expression .
\(iv ) it is tempting to see to what extent our predictions for the low temperature behaviors of @xmath139 and @xmath138 are in agreement with the lattice data of @xcite . unfortunately , the lattice data are only available for temperatures close to its critical value .
as noted earlier , at very large separation the quark and anti - quark become free . in this case the dominant exponent is given by @xmath101 .
so , we are in a situation in which the approximation might be applicable .
when we use it to find the expectation value of the polyakov loop , we get @xmath141 with @xmath101 the minimal area given by . our normalization is stated in and . with this choice
, the polyakov loop expectation value takes the form shown in fig.9 . certainly , it has the desired behavior : the expectation value of @xmath142 is zero at low temperatures , while it is nonzero at high temperatures .. ] one sees there is a phase transition from the confined phase to the deconfined phase , as expected .
numerically , the transition temperature is of order @xmath143 this is roughly twice smaller than the value found in @xcite from the spatial string tension .
thus , the numerical consistency is not good enough at this point .
we chose the normalization constant in the form @xmath144 .
the meaning of the coefficients is the following : @xmath145 specifies the value of @xmath142 at @xmath146 @xmath147 @xmath113 specifies the form of @xmath142 .
indeed , the expectation value of the loop is a continuously growing function of temperature for @xmath148 , while it has a local maximum for @xmath149 .
the position of the maximum is given by a solution to equation @xmath150 for our set of parameters a numerical analysis of results in @xmath151 .
finally , we can compare the temperature dependence of the polyakov loop expectation value as provided by our model with the lattice results of @xcite .
here there is a subtle point .
the lattice data are only available for the range @xmath152 .
this makes it difficult to see what exactly happens in the low temperature phase as well as for high temperatures .
we now fix the normalization by fitting to the data given in table 1 of @xcite near @xmath153 . in fig.10
we have plotted @xmath142 against @xmath154 .
we find that the temperature dependence is in good agreement for @xmath155 .
what we have learned is that the 5-dimensional effective model we proposed to study the properties of a heavy quark - antiquark pair at finite temperature turns out to be remarkably consistent with the qualitative expectations of thermal gauge theory .
moreover , in certain cases it provides the analytic results which are in good agreement with the lattice data .
it is also worth noting that our model predicts the behavior of the thermodynamic quantities for low temperatures where field theory is unreliable and lattice data are missing .
apparently , the model suffers from a lack of numerical self - consistency : the value of the critical temperature as seen from the spatial string tension @xcite turns out to be at least twice bigger than the value we found from the analysis of the polyakov loop in section 3.3 .
there are two lines of thought on this problem .
the first is to account for one - loop corrections in the world - sheet path integral .
this is a complicated problem . indeed
, it is a challenge to theorists to find the world - sheet formulation of string theory on warped geometries like ads spaces . among many things
, it requires world - sheet fermions and even a world - sheet theta angle that is a two - form field @xmath156 with an arbitrary value of @xmath157 @xcite .
but if it was resolved , it could help us to sum the series and hence refine the estimates .
there is another interesting problem here .
what were considered in @xcite are some low temperature corrections to the string tension
. it will be interesting to find similar corrections in our model .
the second line of thought is to somehow modify the background .
it may include a slight revision of the metric or more radical changes like additional background fields .
for instance , a recent proposal of @xcite is even to include the tachyon background .
if a modification is made , it would produce a number of additional free fit parameters that makes it less attractive for phenomenology . to escape the problem , a clever mechanism for reducing the number of parameters must be invented .
part of the interest of ads / cft(qcd ) stems from attempts to understand the physics of rhic .
the reasons are the following : first , a quark - gluon plasma is strongly coupled , so the perturbative qcd is of limited utility .
second , the lattice data do not always provide good information like , for instance on transport properties . on the other hand , ads / cft(qcd )
offers an alternative way of dealing with this real - world problem
. it would be particularly interesting to see if our model can shed some light on this subject .
o.a . would like to thank o. kaczmarek , p. petreczky , m.i .
polikarpov , and e. shuryak for useful communications and conversations .
the work of o.a . was supported in part by max - planck - gesellschaft , deutsche forschungsgemeinschaft , and russian basic research foundation grant 05 - 02 - 16486 .
99 o. andreev and v.i .
zakharov , the spatial string tension , thermal phase transition , and ads / qcd , hep - ph/0607026 .
o. andreev , phys.rev.d 73 , 107901 ( 2006 ) .
mclerran and b. svetitsky , phys .
lett.b 98 , 195 ( 1981 ) .
o. kaczmarek , f. karsch , p. petreczky , and f. zantow , nucl.phys .
129 , 560 ( 2004 ) .
the literature on the wilson loops at finite temperature is vast .
the following is an incomplete list : + e. witten , adv.theor.math.phys . 2 , 505 ( 1998 ) ;
rey , s. theisen , and j .-
yee , nucl.phys.b 527 , 171 ( 1998 ) ; + a. brandhuber , n. itzhaki , j. sonnenschein , and s. yankielowicz , jhep 9806 , 001 ( 1998 ) ; + d.j .
gross and h. ooguri , phys.rev.d 58 , 106002 ( 1998 ) ; + h. dorn and h.j .
otto , jhep 9809 , 021 ( 1998 ) ; + s.a .
hartnoll and s. p. kumar , phys.rev.d 74 , 026001 ( 2006 ) ; + s .- j .
sin and i. zahed , ampere s law and energy loss in ads / cft duality , hep - ph/0606049 ; + h. boschi - filho , n.r.f .
braga , and c.n .
ferreira , phys.rev.d 74 , 086001 ( 2006 ) ; + k. kajantie , t. tahkokallio , and j .-
yee , thermodynamics of ads / qcd , hep - ph/0609254 ; + h. boschi - filho and n.r.f .
braga , ads / cft correspondence and strong interactions , hep - th/0610135 .
d.a . kirzhnits and a.d .
linde , phys.lett.b 42 , 471 ( 1972 ) ; + l. dolan and r. jackiw , phys.rev.d 9 , 3320 ( 1974 ) ; + s. weinberg , phys.rev.d 9 , 3357 ( 1974 ) .
o. kaczmarek , f. karsch , p. petreczky , and f. zantow , phys.lett.b 543 , 41 ( 2002 ) . o. andreev and v.i .
zakharov , phys.rev.d 74 , 025023 ( 2006 ) .
e. eichten , k. gottfried , t. konoshita , k.d .
lane , and t .- m .
yan , phys.rev.d 17 , 3090 ( 1978 ) ; 21 , 203 ( 1980 ) .
i.s . gradshteyn and i.m .
ryzhik , table of integrals , series , and products , academic press , 1994 .
e. witten , as cited in @xcite ; + c.p .
herzog , a holographic prediction of the deconfinement temperature , hep - th/0608151 .
see the discussion of e. witten in @xcite .
pisarski and o. alvarez , phys.rev.d 26 , 3735 ( 1982 ) .
c. csaki and m. reece , toward a systematic holographic qcd : a braneless approach , hep - ph/0608266 .
s. gubser , relativistic heavy ion collisions and string theory , lectures given at the pitp 2006 program , + http://www.admin.ias.edu/pitp/2006files/lecture notes / gubser lecture notes revised.pdf . | in this paper we explore some of the features of a heavy quark - antiquark pair at finite temperature using a five - dimensional framework nowadays known as ads / qcd . we shall show that the resulting behavior is consistent with our qualitative expectations of thermal gauge theory .
some of the results are in good agreement with the lattice data that provides additional evidence for the validity of the proposed model .
+ pacs : 12.38.lg , 12.39.pn , 12.90.+b mpp-2006 - 162 |
development of experimental capabilities dealing with nanostructures brings necessity of appropriate theoretical description of quantum transport ( charge , spin , and heat ) in mesoscopic junctions to the forefront of research.@xcite indeed , a lot of work has been done in this direction . in particular ,
many approaches are based on the landauer expression for current through such junctions in elastic tunneling regime.@xcite one of specific features of molecular transport junctions , the focus of molecular electronics , is flexibility of the molecules , which results in inelastic features being much more pronounced in transport through such junctions as compared e.g. to semiconductor quantum dots .
inelastic features are used as a diagnostic tool , helping to assure presence of the molecule and study its characteristics in the junction within inelastic electron tunneling spectroscopy both in the off - resonance ( iets)@xcite and resonant tunneling ( riets)@xcite situations .
detailed discussion of the inelastic transport in molecular junctions can be found in refs .
. theoretical description of iets is well established today both within simple models@xcite and more realistic calculations.@xcite ability to predict quantitatively experimental findings is a sign of maturity of the field . from theoretical perspective
this success is caused by ability to use well - established nonequilibrium perturbation ( in electron - vibration coupling ) technique .
indeed , in the off - resonant situation electron - vibration coupling @xmath0 is an effectively small parameter , @xmath1 with @xmath2 being resonant off - set and @xmath3 characterizing strength of molecule - contacts coupling .
this allows expansion of the evolution operator in powers of @xmath0 , and truncation at low ( @xmath4 ) order , the born approximation , is usually sufficient to get quantitatively correct predictions of iets signal in molecular junctions .
resonant tunneling situation , @xmath5 , provides richer physics .
while weak electron - vibration coupling , @xmath6 , is treated within perturbation theory also here,@xcite the last fails in the opposite situation , @xmath7 , when e.g. formation of polaron on the molecule becomes possible .
theoretically this case till now mostly was treated either within scattering theory ( or isolated molecule ) approach,@xcite or within quasiclassical ( rate or generalized rate equations ) scheme.@xcite while the first treats electron - vibrational interaction ( numerically ) exactly , it disregards fermi populations in the contacts , as well as dynamical features due to their presence,@xcite and may lead to erroneous predictions.@xcite the last disregards quantum correlations , and as such is applicable to either high temperature , @xmath8 , ( truly classical ) situations , or ( for generalized rate equations approach ) quantum situations when correlations in the system die much quicker than electron transfer ( between contact and molecule ) time ( @xmath9 ) .
besides , these schemes lack formal procedure for improvement of their results similar to taking into account higher order terms in perturbative expansion .
recently we proposed nonequilibrium equation - of - motion ( eom ) approach perturbative in molecule - contact coupling capable of dealing with riets situation.@xcite the approach is formulated for simple resonant level model , but can be easily generalized for more realistic situations.@xcite while it incorporates contacts ( and hence nonequilibrium character of the junction ) into consideration , the price to pay is ( generally ) more approximate level of description of electron - vibration interaction on the bridge .
alternatively , schemes exploring particular parameter regions ( slow vibration @xmath10,@xcite small , @xmath11,@xcite or big , @xmath12,@xcite bias ) were proposed at a model level .
another point especially important for resonant transport ( both elastic and inelastic ) , when actual oxidation or reduction of the molecule takes place , is necessity to speak in the language of many - body molecular states contrary to single - particle molecular orbitals ( the last is used in most ab initio transport calculations today ) .
this includes electronic structure reorganization upon charging , state dependent vibrational modes , anharmonicities , and non - born - oppenheimer couplings .
first schemes trying to treat transport in a many - body molecule states language were recently proposed.@xcite difficulties in describing riets stem from absence of well - established _
nonequilibrium atomic limit _ for molecule in the junction .
indeed , contacts play role of boundary conditions responsible for establishing nonequilibrium state of the molecule .
the last is a complicated mixture of different charge ( and excitation ) states .
approaches developed in molecular electronics community so far either disregard boundary conditions and treat molecule as an equilibrium object ( scattering theory and isolated molecule treatments ) , or establish this nonequilibrium state ( mostly nonequilibrium green function , negf , approaches ) being unable to map it into separate charge ( or more exactly state ) constituents .
note , density matrix based schemes capable of such mapping were developed recently.@xcite such schemes however , miss time correlations , which may become important in e.g. noise spectrum calculations .
hubbard operators is a natural language to talk about system ( in our case subsystem - molecule ) in terms of its states .
thus one seems to be interested in utilizing nonequilibrium hubbard operator green function technique for description of situations similar to riets , when ability to establish nonequilibrium atomic limit of system is desirable
. the approach should be capable to provide a systematic way to take correlations into account ( similar to perturbative expansion in standard diagrammatic techniques ) .
such approach originating from kadanoff and baym functional derivative eom scheme , was developed in the form of equilibrium hubbard operator gfs by sandalov and coworkers for materials with strong electron correlations ( magnets with localized and partly localized moments , mott insulators , kondo lattices , heavy fermion systems , and high - tc superconductors).@xcite the method so far has been applied to model elastic transport through quantum dots@xcite and lowest states of double quantum dots,@xcite and is completely ignored in the molecular electronics community . the goal of our present consideration is to introduce inelastic transport description in the coulomb blockade regime within proper non - equilibrium atomic limit , and to attract attention of the molecular electronics community to the proper nonequilibrium approach capable of speaking in the language of many - particle states ( rather than single - particle orbitals ) and taking into account both molecular charge state dependent normal modes ( presently largely ignored in simulations ) and non - adiabatic couplings .
note , that kondo physics is beyond the scope of current consideration .
the approach takes into account only on - the - molecule correlations in a way that generalizes previous considerations.@xcite note , that including many - body molecular states into consideration of transport potentially allows for : 1 . much more accurate molecular structure simulation out of equilibrium than in current ab initio schemes , due to possibility to employ equilibrium quantum chemistry methods as a starting point for self - consistent procedure .
2 . proper treatment of oxidation / reduction and corresponding electronic and vibrational molecular structure changes , as well as non - adiabatic couplings , 3 . ability to deal with general form of electron - vibration interaction as long as it is localized in space ( in the spirit of ref .
but in addition retaining many - body character of the junction ) , 4 .
calculation of noise spectrum of the junction due to preserved time correlations ( see e.g. ref . for detailed discussion ) , 5 .
proper treatment of degenerate situations due to preserved space correlations ( see e.g. ref . for discussion ) .
structure of the paper is the following . in section [ method ]
we briefly describe the method in terms of many - electronic states of the system , and compare it to previously proposed generalized master equation scheme .
section [ results ] presents numerical examples of its application to transport with discussion .
section [ conclude ] concludes .
here we introduce model of molecular junction , briefly review the basics of nonequilibrium hubbard green function technique , and compare it to previously proposed generalized master equation approach . as usual we consider molecular junction consisting of 3 parts : left ( l ) and right ( r ) contacts and the molecule ( m ) .
the contacts are assumed to be reservoirs of free electrons each at its own equilibrium .
molecule ( or supermolecule if inclusion of parts of contacts is required ) is the nonequilibrium part of the system . besides
, any external potential , e.g. gate voltage probe , or additional contacts can be added to the picture if necessary .
the hamiltonian of the system is @xmath13 where @xmath14 ( @xmath15 ) is hamiltonian for contact @xmath16 @xmath17 @xmath18 is hamiltonian of the isolated molecule , and @xmath19 is coupling between the subsystems @xmath20 @xmath21 ( @xmath22 ) and @xmath23 ( @xmath24 ) are creation ( annihilation ) operators for electron on the molecule and in the contacts , respectively .
their indices @xmath25 and @xmath26 denote electronic state in some chosen single - particle basis , and incorporate all the necessary quantum indices ( e.g. site and spin ) .
now we want to consider molecular subsystem in the basis of _ many - electron states _
@xmath27 , where @xmath28 stands for molecular charge ( number of electrons or excess electrons on the molecule ) and @xmath29 numerates different ( e.g. excitation ) states within the same charge state block . generally these states should not be orthonormal , and consequences of overlap between different molecular states ( as well as overlap of molecular and contact states ) were considered in several papers.@xcite in what follows however we chose the states @xmath27 to be orthonormal @xmath30 in order to keep notation as simple as possible .
hubbard operators are introduced as usual @xmath31 in terms of these many - electron states transfer hamiltonian becomes @xmath32 where @xmath33 denotes transition of the system from state @xmath34 to state @xmath27 , while @xmath35 stands for the backward transition .
transfer matrix element is @xmath36 and @xmath37 .
often many - electron states are chosen as eigenstates of isolated molecule , in this case @xmath38 with @xmath39 energies of the isolated molecule states .
following derivation by meir and wingreen@xcite one gets usual expression for the current at interface @xmath15 @xmath40 \nonumber\end{aligned}\ ] ] which for steady - state situation simplifies to @xmath41\ ] ] the only difference from the standard negf expression is that @xmath42 $ ] in ( [ ik_t ] ) and ( [ ik_e ] ) goes not over single - electron basis , but over basis of _ single - electron transitions _
@xmath43 , eq.([m ] ) , between many - particle states of the molecule .
self - energies @xmath44 in ( [ ik_t ] ) and ( [ ik_e ] ) are defined on the keldysh contour as @xmath45_{\mathcal{m}\mathcal{m } ' } \equiv \sum_{k\in k } v_{\bar{\mathcal{m}}k}\,g_k(\tau,\tau')\ , v_{k\mathcal{m}'}\ ] ] with @xmath46 gf for free electrons in the contacts .
ses projections are @xmath47_{\mathcal{m}\mathcal{m } ' } & = i \gamma^k_{\mathcal{m}\mathcal{m}'}(e ) f_k(e ) \\ \left[\sigma_k^{>}(e)\right]_{\mathcal{m}\mathcal{m } ' } & = -i \gamma^k_{\mathcal{m}\mathcal{m}'}(e ) \left[1-f_k(e)\right]\end{aligned}\ ] ] where @xmath48 and @xmath49 is the fermi distribution in contact @xmath16 .
gfs in ( [ ik_t ] ) and ( [ ik_e ] ) are hubbard operator gfs defined on the keldysh contour as @xmath50 note , that operators in ( [ gk ] ) and ( [ gmm ] ) are in heisenberg representation .
note also , that @xmath43 and @xmath51 in ( [ gmm ] ) may be ( in principle ) arbitrarily far away from one another in the charge space .
gf ( [ gmm ] ) represents correlation between different single - electron molecular many - body state transitions due to coupling to the same bath ( contacts ) . in practice
however , it seems unreasonable to go beyond correlations between nearest charge space blocks . in order to show connection of the present formalism to previously proposed generalized rate equation ( master equation in the fock space ) approach
, we have to realize that the last misses correlations both in space and time .
so to reduce present gf description to the master equation in the fock space , we need to make several simplifications : 1 .
_ diagonal approximation .
_ we have to stick to diagonal elements of gf only @xmath52 with @xmath53 .
markov approximation .
_ we have to consider only gfs of equal times @xmath54 . in order to reduce gfs of different times entering ( [ ik_e ] ) to equal times quantities we use approximation @xmath55\ , g_{\mathcal{m}\mathcal{m}}(t - t)\ ] ] where @xmath56 now , noting that @xmath57 are probabilities to find molecule in state @xmath27 and @xmath34 respectively , we get from ( [ ik_e ] ) @xmath58\ , p^{n+1}_j \right ) \nonumber\end{aligned}\ ] ] if now one restricts attention only to particular charge space block @xmath59 and its nearest neighbors , one gets eqs . ( 6 ) and ( 7 ) of ref . .
now , when expression for the current is established , we need a procedure to calculate hubbard operators gf ( [ gmm ] ) .
note , that standard diagrammatic techniques are inapplicable here , due to lack of the wick s theorem ( since hubbard operators are many - particle operators ) .
an alternative to diagrammatic expansion in the form of functional derivative equation - of - motion technique was developed in ref . .
here we briefly review steps needed to obtain eom for gf we will use in our numerical simulations .
following ref .
we start by writing eom for hubbard operator @xmath60 , where @xmath61 .
this leads to @xmath62 \hat x_\mathcal{m}(\tau ) = \nonumber \\ & \sum_{k\in\{l , r\};\ell}\left ( -v_{k(n+1,j;n+2,\ell)}\,\hat c_k^\dagger(\tau)\,\hat x_{(n , i;n+2,\ell)}(\tau ) \right.\nonumber\\ & -v_{k(n-1,\ell;n , i)}\,\hat c_k^\dagger(\tau)\,\hat x_{(n-1,\ell;n+1,j)}(\tau ) \\ & + v_{(n+1,j;n,\ell)k}\,\hat x_{(n , i;n,\ell)}(\tau)\,\hat c_k(\tau ) \nonumber \\ & \left .
+ v_{(n+1,\ell;n , i)k}\,\hat x_{(n+1,\ell;n+1,j)}(\tau)\,\hat c_k(\tau ) \right ) \nonumber\end{aligned}\ ] ] in what follows we disregard first @xmath63 terms on the right - hand - side , since they describe simultaneous transfer of @xmath63 electrons between contact and molecule , which is beyond the scope of this consideration .
it is clear that when writing eom for gf ( [ gmm ] ) terms in the right - hand - side of ( [ x_eom ] ) will produce correlation functions of the form @xmath64 which can not be factorized into product of single - excitation gf ( [ gmm ] ) and contact single - electron gf ( [ gk ] ) due to lack of the wick s theorem . in order to make this separation a trick with auxiliary fields
@xmath65 is employed .
we need to introduce additional disturbance potential @xmath66 and corresponding generating functional @xmath67\ ] ] then defining gf of @xmath63 arbitrary operators @xmath68 and @xmath69 in the presence of auxiliary fields @xmath70 @xmath71 one easily can get the following identity @xmath72 g_{ab}(\tau,\tau ' ) \nonumber\end{aligned}\ ] ] eq.([trick ] ) allows to express correlation function ( [ corr ] ) in terms of single - excitation gf and its functional derivatives relative to auxiliary fields .
note , that putting ( at the end ) auxiliary fields to be zero turns ( [ gab ] ) into a standard definition of gf .
so , introducing auxiliary fields as in eqs .
( [ disturb ] ) and ( [ generat ] ) and using expression ( [ trick ] ) , one gets general eom for hubbard operator gf ( [ gmm ] ) in the form @xmath73 g_{\mathcal{m}\mathcal{m}'}(\tau,\tau ' ) \nonumber \\ & -\sum_\ell\left(\mathcal{u}_{(n+1,j;n+1,\ell)}(\tau)\ , g_{(n , i;n+1,\ell)\mathcal{m}'}(\tau,\tau ' ) \right .
\nonumber \\ & \qquad \left .
-\mathcal{u}_{(n,\ell;n , i)}(\tau)\ , g_{(n,\ell;n+1,j)}(\tau,\tau')\right ) \nonumber \\ & = \delta(\tau,\tau')\,p_{\mathcal{m}\mathcal{m}'}(\tau ) \nonumber \\ & + \sum_\ell\left ( \left[<t_c\,\hat x_{(n , i;n,\ell)}(\tau)>_\mathcal{u } + i\frac{\delta}{\delta\mathcal{u}_{(n , i;n,\ell)}(\tau)}\right ] \right . \\
& \qquad\times \sum_{\mathcal{m}''}\int_c d\tau''\ , \sigma_{(n,\ell;n+1,j)\mathcal{m}''}(\tau,\tau'')\ , g_{\mathcal{m}''\mathcal{m}'}(\tau'',\tau ' ) \nonumber \\ & + \left[<t_c\,\hat x_{(n+1,\ell;n+1,j)}(\tau)>_\mathcal{u } + i\frac{\delta}{\delta\mathcal{u}_{(n+1,\ell;n+1,j)}(\tau)}\right ] \nonumber \\ & \left .
\qquad\times \sum_{\mathcal{m}''}\int_c d\tau''\ , \sigma_{(n , i;n+1,\ell)\mathcal{m}''}(\tau,\tau'')\ , g_{\mathcal{m}''\mathcal{m}'}(\tau'',\tau ' ) \right ) \nonumber\end{aligned}\ ] ] where @xmath74 is defined in ( [ delta0 ] ) and @xmath75 eq.([gmm_fderiv ] ) is a general equation for hubbard operator gf , representing an alternative to standard diagrammatic technique approaches .
role of expansion in small parameter here play functional derivatives in auxiliary fields .
level of approximation is defined by order of the derivative used in evaluation of gf . at the end of differentiations auxiliary fields
are put to be zero , and resulting expression is equation for gf at particular level of approximation . the simplest approximation , hubbard i ( hi ) , is obtained from ( [ gmm_fderiv ] ) by keeping only diagonal averages , omitting all functional derivatives , and @xmath76 @xmath77 g_{\mathcal{m}\mathcal{m}'}(\tau,\tau ' ) = \delta(\tau,\tau')\,\delta_{\mathcal{m}\mathcal{m}'}\,p_\mathcal{m } \\ & + p_\mathcal{m}\sum_{\mathcal{m}''}\int_c d\tau '' \ , \sigma_{\mathcal{m}\mathcal{m}''}(\tau,\tau'')\ , g_{\mathcal{m}''\mathcal{m}'}(\tau'',\tau ' ) \nonumber\end{aligned}\ ] ] following most of the papers , employed the method so far,@xcite in our consideration we go one step further - we take one functional derivative to get so called first loop approximation .
note , here we take derivative of the gf only , disregarding fluctuations of the spectral weight @xmath78 . after performing the differentiation
once more we keep only diagonal averages , omit all functional derivatives , and @xmath76 . since the procedure was described in details in many papers ( see e.g. refs . ) , here we present only final result @xmath79 g_{\mathcal{m}\mathcal{m}'}(\tau,\tau ' ) -i\sum_{\kappa,\ell}\sum_{\mathcal{m}''}\int_c d\tau '' \nonumber \\ & \left [ \sigma_{(n,\kappa;n+1,j)\mathcal{m}''}(\tau,\tau'')\ , d_{\mathcal{m}''(n-1,\ell;n , i)}(\tau'',\tau+)\ , \right .
\nonumber \\ & \qquad \times g_{(n-1,\ell;n,\kappa)\mathcal{m}'}(\tau,\tau ' ) \nonumber \\ & -\sigma_{(n,\kappa;n+1,j)\mathcal{m}''}(\tau,\tau'')\ , d_{\mathcal{m}''(n,\kappa;n+1,\ell)}(\tau'',\tau+)\ , \nonumber \\ & \qquad \times g_{(n , i;n+1,\ell)\mathcal{m}'}(\tau,\tau ' ) \nonumber \\ & + \sigma_{(n , i;n+1,\kappa)\mathcal{m}''}(\tau,\tau'')\ , d_{\mathcal{m}''(n,\ell;n+1,\kappa)}(\tau'',\tau+)\ , \nonumber \\ & \qquad \times g_{(n,\ell;n+1,j)\mathcal{m}'}(\tau,\tau ' ) \\ & -\sigma_{(n , i;n+1,\kappa)\mathcal{m}''}(\tau,\tau'')\ , d_{\mathcal{m}''(n+1,j;n+2,\ell)}(\tau'',\tau+)\ , \nonumber \\ & \qquad \times \left .
g_{(n+1,\kappa;n+2,\ell)\mathcal{m}'}(\tau,\tau ' ) \right ] \nonumber \\ & = \delta(\tau,\tau')\,\delta_{\mathcal{m}\mathcal{m}'}\,p_\mathcal{m } \nonumber \\ & + p_\mathcal{m}\sum_{\mathcal{m}''}\int_c d\tau '' \ , \sigma_{\mathcal{m}\mathcal{m}''}(\tau,\tau'')\ , g_{\mathcal{m}''\mathcal{m}'}(\tau'',\tau ' ) \nonumber\end{aligned}\ ] ] where @xmath61 and @xmath80 is so called full locator , which ( in the first loop approximation ) obeys the same equation ( [ firstloop ] ) as gf but without spectral weight @xmath81 multiplying delta function in the right - hand - side .
expressions for gfs @xmath82 and @xmath80 ( first loop approximation ) in the shorthand ( matrix in both fock space and keldysh contour variables ) notation can be written as @xmath83 where @xmath84 \\ \delta_\mathcal{m } & = \delta_\mathcal{m}^{0}+\delta\delta_\mathcal{m}\end{aligned}\ ] ] and @xmath85 is given by the second term in the left - hand - side of eq.([firstloop ] ) , it is responsible for shifts of transition energies in the molecule due to contacts induced correlation .
one sees that eq.([shorthand_d ] ) has usual structure of the dyson equation , which is obtained in standard diagrammatic expansion .
the only difference is dressing of se @xmath86 by spectral weight @xmath78 .
thus formally one can use all the standard equations , using dressed se @xmath86 everywhere , to get desired projections of gf @xmath80 .
when @xmath80 is known , @xmath82 is obtained by simple matrix multiplication @xmath87 note , side of matrix dressing by spectral weight @xmath78 is different for @xmath86 and @xmath82 , compare eqs .
( [ d_1 ] ) and ( [ gdp ] ) .
note also , that the scheme is self - consistent , since both transition energies shift @xmath88 and spectral weights @xmath78 depend on gf @xmath82 , while the last depends on these quantities .
in particular ( @xmath61 ) @xmath89 here @xmath90 and @xmath91 are probabilities to find system in the state @xmath27 and @xmath34 respectively .
here we present results of simulations within first loop approximation . in order to speed up calculations we employed also diagonal approximation . may result in unphysical behavior .
in particular , retarded and advanced ses and gfs are not hermitian conjugates of one another .
while the issue does not arise in the diagonal approximation implemented for calculations here , this is a problem for a more general consideration .
a simple workaround is to use average of two dyson - like expressions : one with @xmath92 , eq.([d_1 ] ) , applied from the left and one from the right .
this leads to set of equations for gfs , which under markov approximation reduce to widely employed equations for dm in dissipative environment . ]
we consider transport through quantum dot and double quantum dot , discuss obtained data , and compare it to previously published results . in the case of quantum dot
molecular hamiltonian is @xmath93 where @xmath94 indicates spin projection and @xmath95 .
full fock space of the molecular part of the system ( without vibrations )
consists of one empty state ( @xmath96 ) , two single - electron states , ( @xmath97 and @xmath98 ) , and one doubly occupied state ( @xmath99 ) . transitions between these states to be considered are spin up electron transfers ( @xmath100 and @xmath101 ) and spin down electron transfers ( @xmath102 and @xmath103 ) .
writing eq.([firstloop ] ) in the basis of these transitions one gets equations obtained in ref . .
( color online ) conductance map for elastic transport through quantum dot .
see text for parameters . ]
figure [ fig_qd ] presents conductance map for elastic transport through quantum dot .
parameters of the calculation are @xmath104 k , @xmath105 ev , @xmath106 ev ( @xmath107 and @xmath15 ) , and @xmath108 ev . as usually one has areas of blockaded transport ( inside part of diamonds ) with fixed population on the dot ( @xmath109 , @xmath110 , and @xmath63 from right to left ) , and transition areas ( between the diamonds ) where population on the dot is noninteger ( see e.g. ref . for more detailed discussion ) . in order to treat inelastic transport
we add to @xmath18 molecular vibration linearly coupled to electron(s ) on the dot @xmath111 such model is frequently used to describe inelastic transport in molecular junctions . in a sense it is similar to marcus theory , and describes shift of the molecular vibration when molecule is charged due to electron transfer from / to the contacts .
in general , model with non - diagonal electron - vibration coupling also can be considered within the formalism .
after small polaron ( lang - firsov or canonical ) transformation@xcite linear coupling term is eliminated , while energy level position @xmath112 and hubbard repulsion @xmath113 are renormalized ( @xmath114 and @xmath115 ) , and transfer matrix elements in @xmath19 , eq.([ht ] ) , are dressed with shift operators ( @xmath116 ) @xmath117\ ] ] where @xmath118 . in what follows
we disregard renormalization of @xmath112 and @xmath113 , assuming that it was included in definition of these parameters .
now molecule is characterized by direct product of electronic and vibrational spaces , so its state should be indicated by additional index @xmath119 showing state of the vibration , i.e. molecular subspace is spanned by states @xmath120 , @xmath121 , @xmath122 , and @xmath123 , where @xmath124 .
one has to consider the same electronic transitions as in the case of elastic transport , but in addition all possible transitions between states of the vibration have to be included .
transitions between these states ( within the model ) are possible only by electron transfer between molecule and contacts .
due to shift operators ( [ xvib ] ) appearing in @xmath19 ses ( [ sigma ] ) are now dressed with corresponding vibrational overlap integrals ( @xmath125 ) @xmath126 with @xmath127^{1/2 } l_{v_{min}}^{v_{max}-v_{min}}(\lambda^2 ) \nonumber\end{aligned}\ ] ] where @xmath128 ( @xmath129 ) is minimal ( maximal ) of @xmath119 and @xmath130 , @xmath131 is step function , and @xmath132 is laguerre polynomial . note an important formal difference between the present approach and the one presented in ref . . while in the last we had to consider separately electron and phonon dynamics , which leads to convolution of electron gf ( electron dynamics ) with franck - condon factors ( phonon dynamics ) , here the situation is different .
since we consider generalized fock space ( product of electronic and vibrational ones ) , within the formalism strictly speaking we do not have inelastic processes at all . instead one has to consider elastic scattering events between electron - vibrational states . as a result the role played previously by the franck - condon factors ( to introduce vibrational dynamics ) now is included into hubbard gf of the generalized fock space .
( color online ) conductance map for inelastic transport through quantum dot .
linear coupling model ( [ hm_qd_vib_x ] ) .
see text for parameters . ]
figure [ fig_vib_x ] presents conductance map for inelastic transport through quantum dot within linear coupling model ( [ hm_qd_vib_x ] ) .
parameters of the calculation are @xmath133 ev and @xmath134 ev , all the other parameters are as in fig .
[ fig_qd ] . within the calculation we restricted vibrational subspace to 4 lowest levels ( @xmath135 ) .
as expected , besides elastic peaks in the conductance map we get additional resonant vibrational features corresponding to inelastic processes .
this figure is equivalent to fig .
3a of ref . , where calculation was done within perturbative ( in coupling to electrodes )
nonequilibrium eom approach for the same model .
as previously,@xcite within the model ( see also discussion below ) distance between the diamond edges ( elastic peak ) and vibrational sidebands is defined by the oscillator frequency . increase in electron - vibrational coupling would result in both more pronounced vibrational features and suppression of transport in the low source - drain voltage region due to franck - condon blockade . while increase in temperature would produce also vibrational sidebands corresponding to phonon absorption ( features inside the diamond ) .
finally , we want to demonstrate capabilities of the present scheme , which go beyond approaches previously used to treat inelastic transport .
suppose our molecule is small enough , so that upon charging it changes its normal modes essentially .
suppose also , that from all the normal modes of the molecule only one is coupled to tunneling electron .
inelastic transport in this case can be modeled by assigning different vibration frequencies to different charge states of the molecule . in our quantum dot model
this corresponds to situation , when vibrational frequencies for @xmath120 , @xmath136 , and @xmath123 states are different - @xmath137 , @xmath138 , and @xmath139 respectively .
self - energies due to electron transfer between molecule and contacts once more have to be dressed by overlap integrals between different vibrational wavefunctions , as is shown in eq.([ovlp_x ] ) . however this time ( when vibrational frequencies change ) the integrals should be calculated in the way discussed in refs .
( color online ) conductance map for inelastic transport through quantum dot .
charge state dependent frequencies .
see text for parameters . ]
figure [ fig_vib ] presents conductance map for inelastic transport through quantum dot , when vibrational frequency depends on charge state of the dot .
parameters of the calculation are @xmath140 ev , @xmath141 ev , @xmath142 ev and shift vector for both transitions was taken to be @xmath143 ( see refs . for detailed explanation )
, all the other parameters are as in fig .
[ fig_qd ] one sees that result of calculation is counter intuitive at the first sight .
naively one could expect to see inelastic peaks at each diamond edge ( each charge state of the quantum dot ) being separated by frequency corresponding to the neighboring charge state ( riets probes frequencies of the intermediate ion ) .
real picture is more complicated however .
let consider electron transfer between @xmath63 particular charge states of the quantum dot , say between states @xmath144 and @xmath145 upon electron transfer from contact to molecule .
in this case change in the subsystem energy , which will be observed in transport as inelastic peak in conductance , is @xmath146 ( we omit here change in elastic electronic energy for simplicity , this will define only position of the elastic peak in conductance ) . since @xmath147 and @xmath148 in principle can be any non - negative numbers , it is clear that one can observe a a progression of frequencies .
note , that in this progression one can see inelastic peaks in conductance , separated from the elastic one by frequency which does not exist in the system at all ( e.g. @xmath149 ) .
note also , that due to overlap factors involved the lowest frequencies of the progression will be observed better in riets signal .
non - adiabatic couplings can be included in calculation in a similar way .
molecular hamiltonian for double quantum dot is @xmath150 where @xmath151 numbers sites and @xmath152 stands for spin projection , @xmath21 ( @xmath22 ) is creation ( annihilation ) operator , @xmath153 , and @xmath154 .
we assume that site @xmath110 is coupled to the left contact , while site @xmath63 - to the right .
we chose many - body states for molecular subsystem in the form @xmath155 . unlike choice of refs .
these are not eigenstates of the molecular hamiltonian . as a result eom for hubbard operator gfs couples
them also by hopping @xmath156 . besides this
all the treatment presented in section [ method ] remains the same .
there are @xmath157 states ( @xmath110 , @xmath158 , @xmath159 , @xmath158 , and @xmath110 states for @xmath109 , @xmath110 , @xmath63 , @xmath160 , and @xmath158 electrons in the system respectively ) and @xmath161 single - electron transitions ( @xmath157 for each spin block ) to be considered .
( color online ) conductance map for elastic transport through double quantum dot .
see text for parameters . ]
figure [ fig_dqd ] shows conductance map for elastic transport through double quantum dot .
parameters of the calculation are @xmath104 k , @xmath162 ev , @xmath163 ev , @xmath164 ev , @xmath165 , @xmath166 ev , @xmath167 ev , and @xmath168 ev . as usually one sees pattern of blockaded and allowed transport regions . however , here this pattern is more complicated than in the case of quantum dot .
figure [ fig_dqd_cut ] demonstrates this pattern for a horizontal cut of fig . [ fig_dqd ] at source - drain voltage @xmath169 . shown
are current ( a ) and probabilities to find the molecular subsystem in different occupation states .
( color online ) elastic transport quantum dot at fixed source - drain voltage @xmath169 : ( a ) current vs. gate voltage , ( b ) probability to find double quantum dot empty ( dotted line , black ) , singly- ( dashed line , green ) , doubly- ( solid line , red ) , triply- ( dash - dotted line , blue ) , or fully - occupied ( dash - double - dotted line , magenta ) .
parameters are the same as in fig .
[ fig_dqd ] . ]
approaches based on renormalization group technique ( nrg , dmrg , real time rg , etc . ) are routinely used to treat quantum impurity systems , where e.g. correlations between localized impurity ( quantum dot or molecule ) and delocalized states ( contacts ) leads to kondo effect .
mostly these approaches were applied to description of strongly correlated systems in equilibrium .
description of transport is more problematic , since one needs to know spectral function at finite temperature , where all excitations may contribute.@xcite nevertheless first approaches dealing with transport started to appear as well.@xcite the main complication with implementation of these methods ( besides dmrg ) for ab initio calculation is their complexity , so that all the calculations done so far are restricted to simple models only . in the case of molecular junctions
most of ab initio calculations done today are performed at the mean mean field level of treatment with effective single particle orbitals used in place of molecular states .
such approach clearly breaks down in the resonance tunneling regime , where actual reduction / oxidation of the molecule leading to corresponding electronic and vibrational structure change become possible .
necessity of treating this regime in the language of many - body molecular states , thus incorporating on - the - molecule correlations , was realized and first approaches like e.g. generalized master equation approach@xcite where proposed . here
we generalize this consideration by incorporating many - body molecular states language into nonequilibrium green s function framework .
the main formal problem here is that many - body states language makes the wick s theorem inapplicable , and thus standard nonequilibrium diagrammatic techniques can not be used .
a workaround based on functional derivative equation - of - motion technique for hubbard operator gfs was developed by sandalov and coworkers @xcite for equilibrium case .
the method so far has been applied to model elastic transport through quantum dots@xcite and lowest states of double quantum dots,@xcite and is completely ignored in the molecular electronics community . here
we employ the approach to deal with inelastic transport through molecular junctions in nonequilibrium atomic limit .
we formulate the method within basis of charged states of the molecule .
we demonstrate its ability to deal with transport situation in the language of these states ( rather than effective single - electron orbitals ) , as well as take into account charge - specific normal modes as well as nonadiabatic couplings .
capabilities of the technique are illustrated with simple model calculations of transport through quantum dot and double quantum dot .
extension to realistic calculations is the goal of our future research .
is indebted to igor sandalov for numerous illuminating discussions , and thanks karsten flensberg , jonas fransson and ivar martin for helpful conversations .
m.g . gratefully acknowledges the support of the ucsd startup fund and lanl director s postdoctoral fellowship .
thanks the israel science foundation , the us - israel binational science foundation and the german - israel foundation for financial support . m.r .
thanks the nsf / mrsec for support , through the nu - mrsec .
this work was performed , in part , at the center for integrated nanotechnologies , a u.s .
department of energy , office of basic energy sciences user facility .
los alamos national laboratory , an affirmative action equal opportunity employer , is operated by los alamos national security , llc , for the national nuclear security administration of the u.s .
department of energy under contract de - ac52 - 06na25396 .
99 a.nitzan and m.a.ratner , science * 300 * , 1384 ( 2003 ) .
r.landauer , ibm j. res . dev . * 1 * , 223 ( 1957 ) .
b.c.stipe , m.a.rezaei , and w.ho , phys .
* 82 * , 1724 ( 1999 ) .
n.b.zhitenev , h.meng , and z.bao , phys .
lett . * 88 * , 226801 ( 2002 ) .
m.galperin , m.a.ratner , and a.nitzan , j. phys . : condens .
matter * 19 * , 103201 ( 2007 ) .
m.galperin , m.a.ratner , a.nitzan , and a.troisi , science * 319 * , 1056 ( 2008 ) .
r.lake and s.datta , phys . rev .
b * 45 * , 6670 ( 1992 ) .
r.lake and s.datta , phys .
b * 46 * , 4757 ( 1992 ) .
s.tikhodeev , m.natario , k.makoshi , t.mii , and h.ueba , surf .
sci . * 493 * , 63 ( 2001 ) .
t.mii , s.tikhodeev , and h.ueba , surf .
sci . * 502 - 503 * , 26 ( 2002 ) .
t.mii , s.g.tikhodeev , and h.ueba , phys .
b * 68 * , 205406 ( 2003 ) .
m.galperin , m.a.ratner , and a.nitzan , j. chem . phys . * 121 * , 11965 ( 2004 ) .
n.lorente , m.persson , l .- j.lauhon , and w.ho , phys .
lett . * 86 * , 2593 ( 2001 ) .
l.bocquet , h.lesnard , and n.lorente phys .
lett . * 96 * , 096101 ( 2006 ) .
t.frederiksen , m.brandbyge , n.lorente , and a .- p.jauho , phys .
* 93 * , 256601 ( 2004 ) .
m.paulsson , t.frederiksen and m.brandbyge , nano lett .
* 6 * , 258 ( 2006 ) n.lorente and m.persson , phys .
lett . * 85 * , 2997 ( 2000 ) a.pecchia , a. di carlo , a.gagliardi , s.sanna , t.frauenheim , and r.gutierrez , nano lett . * 4 * , 2109 ( 2004 ) .
p.hyldgaard , s.hershfield , j.h.davies , and j.w.wilkins , ann . phys . * 236 * , 1 ( 1994 ) .
a.mitra , i.aleiner , and a.j.millis , phys . rev .
b * 69 * , 245302 ( 2004 ) . n.s.wingreen , k.w.jacobsen , and j.w.wilkins , phys . rev .
* 61 * , 1396 ( 1988 ) . n.s.wingreen , k.w.jacobsen , and j.w.wilkins , phys .
b * 40 * , 11834 ( 1989 ) .
j.bona and s.a.trugman , phys
lett . * 75 * , 2566 ( 1995 ) .
a.s.alexandrov and a.m.bratkovsky , phys .
b * 67 * , 235312 ( 2003 ) .
b.muralidharan , a.w.ghosh , and s.datta , phys .
b * 73 * , 155410 ( 2006 ) .
l.siddiqui , a.w.ghosh , and s.datta , phys .
b * 76 * , 085433 ( 2007 ) .
k.flensberg , phys . rev .
b * 68 * , 205323 ( 2003 ) .
m.galperin , a.nitzan , and m.a.ratner , phys .
b * 73 * , 045314 ( 2006 ) .
r.h " artle , c.benesch , and m.thoss , arxiv:0801.3602 ( 2008 ) .
m.galperin , m.a.ratner , and a. nitzan , nano lett .
* 5 * , 125 ( 2005 ) ; j. phys . : condens .
matter to be published ( 2008 ) .
a. la magna and i.deretzis , phys .
rev . lett . * 99 * , 136404 ( 2007 ) .
a.m.kuznetsov , j. chem . phys . * 127 * , 084710 ( 2007 ) .
a.mitra , i.aleiner , and a.j.millis , phys .
lett . * 94 * , 076404 ( 2005 ) .
d.mozyrsky , m.b.hastings , and i.martin , phys .
b * 73 * , 035104 ( 2006 ) .
j.p.bergfield and c.a.stafford , arxiv:0803.2756 ( 2008 ) .
h.schoeller , lecture notes in physics * 544 * , 137 ( 2000 ) .
j.rammer , a.l.shelankov , and j.wabnig , phys . rev .
b * 70 * , 115327 ( 2004 ) .
i.sandalov , j.johansson , and o.eriksson , int .
94 * , 113 ( 2003 ) .
j.fransson , o.eriksson , and i.sandalov , phys .
lett . * 88 * , 226601 ( 2002 ) .
j.fransson , phys .
b * 72 * , 075314 ( 2005 ) . i.sandalov and r.g.nazmitdinov , phys . rev .
b * 75 * , 075315 ( 2007 ) .
j.fransson , o.eriksson , and i.sandalov , photonics and nanostructures * 2 * , 11 ( 2004 ) .
j.fransson and o.eriksson , j. phys . :
matter * 16 * , l85 ( 2004 ) .
j.fransson , phys .
b * 69 * , 201304(r ) ( 2004 ) . j.fransson and o.eriksson , phys .
b * 70 * , 085301 ( 2004 ) .
m.galperin , a.nitzan , and m.a.ratner , phys .
b * 74 * , 075326 ( 2006 ) .
t.s.rahman , r.s.knox , and v.m.kenkre , chem . phys .
* 44 * , 197 ( 1979 ) .
j.fransson , o.eriksson , and i.sandalov , phys .
b * 66 * , 195319 ( 2002 ) . y.meir and n.s.wingreen , phys . rev .
* 68 * , 2512 ( 1992 ) . h.haug and a.jauho , _ quantum kinetics in transport and optics of semiconductors _ springer - verlag : berlin , heidelberg ( 1996 )
m.galperin , a.nitzan , and m.a.ratner , phys .
b * 76 * , 035301 ( 2007 ) .
g.d.mahan , _ many - particle physics _ ,
kluwer academic / plenum publishers : new york ( 2000 ) .
p.t.ruhoff , chem . phys .
* 186 * , 355 ( 1994 ) .
p.t.ruhoff and m.a.ratner , int .
77 * , 383 ( 2000 ) .
r.bulla , t.a.costi , and t.pruschke , rev .
80 * , 395 ( 2008 ) .
k.a.al-hassanieh , a.e.feiguin , j.a.riera , c.a.busser , and e.dagotto , phys .
b * 73 * , 195304 ( 2006 ) .
j.e.han , phys .
b * 73 * , 125319 ( 2006 ) .
p.s.cornaglia , g.usaj , and c.a.balseiro , phys .
b * 76 * , 241403(r ) ( 2007 ) .
f.b.anders , phys .
* 101 * , 066804 ( 2008 ) . | method developed by sandalov and coworkers [ int . j. quant . chem . * 94 * , 113 ( 2003 ) ] is applied to inelastic transport in the case of strong correlations on the molecule , which is relatively weakly coupled to contacts .
ability of the approach to deal with the transport in the language of many - body molecular states as well as take into account charge - specific normal modes and nonadiabatic couplings is stressed .
we demonstrate capabilities of the technique within simple model calculations , and compare it to previously published approaches . |
the azimuthal anisotropy of particles created in ultra - relativistic heavy - ion collisions is commonly measured by expanding the particle s azimuthal momentum distribution with respect to the reaction plane in terms of a fourier series @xcite .
the obtained coefficients @xmath3 for increasing order @xmath4 characterize the distribution in more and more detail . in non - central collisions
sufficient rescattering will drive the initial spatial anisotropy of the system into a state where this spatial anisotropy is diminished . during this evolution
an azimuthal anisotropy in momentum space is built up . due to the self - quenching nature of this process
the signal of anisotropic flow is sensitive to an early stage in the evolution of the system .
this picture holds for all different harmonics , even though flow of different order might test different time scales .
the second fourier coefficient @xmath5 , so - called elliptic flow , is studied in detail at rhic @xcite .
this is mainly due to its large magnitude , which allows for a precise measurement of the reaction plane .
it was realized that this high resolution of the second order reaction plane allowed for the measurement of higher order anisotropies as well @xcite . while the qualitative agreement of hydrodynamical model predictions with anisotropic flow measurements below transverse momentum of about @xmath6gev/@xmath2
@xcite is good , at higher momenta this picture breaks down .
instead , quark coalescence models provide a very intriguing explanation of the observed particle species dependence in the intermediate @xmath0-region up to @xmath7gev/@xmath2 @xcite . in this article the measurements of elliptic flow @xmath5 and higher order flow by the star experiment at rhic
will be reviewed .
the data presented come from au+au collisions at @xmath8gev .
about two millions events in the star main time projection chamber ( tpc ) with a pseudorapidity coverage of @xmath9 were analyzed .
in the low-@xmath0 region up to @xmath10gev/@xmath2 elliptic flow shows an almost linear increase with transverse momentum . this feature is present for all measured particle species , as can be seen in fig . [ v2pt_mass_ordering ] : charged pions , charged and neutral kaons , protons and lambdas and their anti - particles all follow this general behavior .
a clear mass ordering effect is visible : particles with a higher mass show a smaller @xmath5 at a given @xmath0 than particles with lower mass .
hydrodynamical model calculations @xcite can successfully describe these observations in this @xmath0-region to a level of 20 30% , attributing the mass ordering to an underlying common transverse velocity field .
these models assume a local thermal equilibrium of the particle source .
since the mechanism of generating anisotropic flow is self - quenching , the development of local thermal equilibrium at an early stage is supported .
nevertheless , since these calculations ca nt predict the full shape of the event , neither the observed anisotropies at forward / backward rapidities nor other harmonics than @xmath5 , the conclusion of thermal equilibrium is not at all solid .
it is important to note that the aforementioned hydrodynamical models fail to describe the mass and momentum dependence of elliptic flow as long as they do nt invoke a phase transition from partonic to hadronic degrees of freedom @xcite .
this sensitivity to the equation of state justifies the importance of elliptic flow for the understanding of hot and dense nuclear matter , including the possible creation of a quark - gluon plasma ( qgp ) .
at transverse momentum 1.5 @xmath1 @xmath0 @xmath1 5gev/@xmath2 the elliptic flow of different particle species starts to deviate significantly from the mass ordering schematics discussed above , see fig .
interestingly , different particle species seem to exhibit a different value of saturation for @xmath5 , while they fall on top of each other if grouped into mesons and baryons .
while at low @xmath0 the lighter mesons have a larger @xmath5 for a given @xmath0 , at intermediate @xmath0 the elliptic flow of mesons is smaller than that of baryons .
quark coalescence models provide an elegant explanation for this observed feature .
these models assume that in this @xmath0 range particle production is dominated by coalescing quarks : two quarks moving with the same momentum make up a meson with twice the momentum of the original quarks , three quarks coalesce into a baryon with three times the quark momentum @xcite .
therefore , by scaling the observed @xmath5 signal and the transverse momentum by the number of constituent quarks @xmath4 one obtains the underlying quark flow . the results , as shown in fig .
[ v2_scaled ] , are very intriguing . in the region @xmath11gev/@xmath2 of scaled particle momentum ,
elliptic flow for different particles is literally the same .
only the pions deviate from the universal curve , which can be explained by feed - down from resonances @xcite .
the success of these quark coalescence models in describing the measured @xmath5 is a strong hint that the observed large anisotropic flow builds up at the partonic stage of the system s evolution already . on one hand this suggests that the system evolves through a state of partonic degrees of freedom , which on the other hand
suffer a lot of rescattering and therefore should be close to local thermal equilibrium .
in order to measure harmonics higher than @xmath12 we use the knowledge of the very well determined second order reaction plane like @xmath13 , @xmath14 are accessible . ] @xcite . as shown in fig .
[ v4 ] , the coefficient for the fourth harmonic is significantly lower than that for the second harmonic , but is non - zero for all @xmath0 . within errors
@xmath15 is equal to zero .
it was suggested that @xmath3 could be proportional to @xmath16 , as long as the @xmath17 distribution is a smooth , slowly varying function of @xmath18 @xcite .
the ratio of @xmath13 over @xmath19 is shown in fig .
[ v4v22ratio ] and is close to 1.2 . even though a straight line might not be the best fit , the ratio is clearly larger than 1 . according to this mean ratio the two dashed lines in fig .
[ v4 ] were drawn .
this very good agreement between the scaled @xmath5 and the measurements of higher order can be also seen in the picture of parton coalescence . assuming a simple model @xcite one obtains for the ratio @xmath20 for mesons and @xmath21 for baryons . as shown ,
this ratio is experimentally determined to be 1.2 , which means that the fourth - harmonic flow of quarks @xmath22 must be greater than zero .
one can go one step further and assume that the observed scaling of the hadronic @xmath5 actually results from a similar scaling occurring at the partonic level @xcite . in this case @xmath23 and the hadronic ratio @xmath24 then equals @xmath25 for mesons and @xmath26 for baryons , respectively . again , since this value is measured to be 1.2 , even the partonic @xmath22 must be greater than simple scaling and quark coalescence models predict .
elliptic flow of identified particles was shown to exhibit a constituent - quark - number dependence at intermediate transverse momentum 1.5@xmath1@xmath0@xmath15gev/@xmath2 .
the quantitative measurements of @xmath5 support model predictions for quark - number scaling which imply hadronization by coalescing partons . within statistical uncertainties observations of higher order momentum anisotropies for charged hadrons support these model predictions .
the quantitative deviations of only about 20% compared to model assumptions can be easily explained in the light of simplifications in the models while going from hadronic to partonic scaling of @xmath13 vs. @xmath5 .
it has to be stressed that the observed mass ordering of elliptic flow measurements at low @xmath0 is nevertheless present and stands in no contradiction to the described behavior at intermediate transverse momentum . with the large data
set taken during rhic run iv the star experiment will greatly enhance the statistical significance of the presented measurements .
we will be able to test constituent - quark - number scaling with much higher precision .
it will be interesting to see the outcome of these new measurements compared to recent developments in the theoretical understanding of the underlying process of hadronization .
99 poskanzer a m and voloshin s a 1998 _ c _ * 58 * 1671 see for example adams j ( star collaboration ) 2004 _ arxiv _ nucl - ex/0409033 ( to be published in _ c _ ) and references therein kolb p f 2003 _ c _ * 68 * 031902(r ) sorensen p 2003 _ arxiv _ nucl - ex/0309003 adler c ( star collaboration ) 2001 182301 adler c ( star collaboration ) 2002 132301 adler s s ( phenix collaboration ) 2003 172301 adams j ( star collaboration ) 2004 052302 huovinen p , kolb p f , heinz u w , ruuskanen p v and voloshin s a 2001 _ b _ * 503 * 58kolb p f 2004 _ aip conf . proc .
_ * 698 * 694 huovinen p 2004 private communication sorensen p 2004 _ arxiv _ nucl - ex/0405001 kolb p f and heinz u 2003 _ quark gluon plasma 3 , eds .
hwa r c and wang x - n ( world scientific , singapore ) _ and _ arxiv _ nucl - th/0305084 molnar d and voloshin s a 2003 092301 hwa r c and yang c b 2003 _ c _ * 67 * 064902 fries r j , mller b , nonaka c and bass s a 2003 _ c _ * 68 * 044902 greco v , ko c m and levai p 2003 _ c _ * 68 * 034904 greco v , ko c m and levai p 2003 202302 fries r j , mller b , nonaka c and bass s a 2003 202303 adams j ( star collaboration ) 2004 _ arxiv _ nucl - ex/0409033 ( to be published in _ c _ ) castillo j ( for the star collaboration ) 2004 s1207 greco v and ko c m 2004 _ c _ * 70 * 024901 dong x , esumi s , sorensen p , xu n and xu z 2004 _ b _ * 597 * 328 adams j ( star collaboration ) 2004 062301 ollitrault j - y 2003 private communication poskanzer a m ( for the star collaboration ) 2004 s1225 kolb p f , chen l - w , greco v , and ko c m 2004 _ c _ * 69 * 051901 chen l - w , ko c m , and lin z - w 2004 _ c _ * 69 * 031901 | measurements of anisotropic flow at low ( @[email protected]/@xmath2 ) and intermediate ( 1.5 @xmath1 @xmath0 @xmath1 5gev/@xmath2 ) transverse momentum from the star collaboration are reviewed .
while at low @xmath0 an ordering of elliptic flow strength with particle mass is observed , the measured signals appear to follow number - of - constituent quark scaling at intermediate @xmath0 .
the observations of higher harmonics support this picture qualitatively , and are sensitive to specific model assumptions . |
during the last years the development of different and powerful laser cooling techniques @xcite has lead to the experimental observation of several striking matter wave phenomena for cold atoms .
these phenomena resemble the usual effects observed in standard optics , and therefore they have been the subject of study of a fast developing research field properly called _
atom optics_. some atom optics examples could be the diffraction of atoms @xcite , interferometers@xcite and lenses @xcite for atoms , and very recently even the atom optics equivalent of a laser source , or atom
laser @xcite .
atomic mirrors have also been constructed using evanescent laser fields @xcite or , more recently , with magnetic fields @xcite . both mirrors , although based on different physical effects , share the same physical background , i.e. the atom feels a repulsive potential , and if its initial velocity is sufficiently small , it reaches a turning point , being reflected . note that after the turning point the atomic density of probability follows an exponential decay , resembling the case of an electric field impinged onto a metallic surface .
therefore , these atomic mirrors can be considered the atom optics counterpart of metallic mirrors in light optics . on the other hand , in the last few years
, the behavior of a cold atom in periodic laser potentials ( light lattices ) has aroused great interest , in particular the resemblance of this physical situation and solid state physics @xcite . in this sense
, bragg scattering has been analysed in the context of atomic waves @xcite , and so has the use of such effect to construct atomic beam splitters and interferometers @xcite .
bragg reflection is just a particular case of a more general situation , i.e. the so called photonic band gap
structures ( or pbgs ) @xcite . in dielectric periodic structures some electromagnetic waves can not propagate ( basically because their corresponding energies are within an energetic gap produced by the periodic structure ) , and therefore are reflected . in previous papers @xcite
we have proposed a laser arrangement that acts as an atomic band gap structure , in the sense that it resembles a pbgs but for atomic waves instead of electromagnetic ones .
the atoms are reflected if their incoming kinetic energy lies within a gap produced by the laser periodicity .
this effect produces a band like momentum spectrum of the reflected atoms .
we have shown that the atoms can be reflected not only by repulsive laser potentials , but also by attractive ones , allowing the atom optics equivalent of a multilayer dielectric mirror in light optics . among the different solid state like phenomena
recently reported , is of special interest the analysis of the behavior of cold atoms accelerated in a periodic potential . in this sense
, the well known bloch oscillations ( bo ) , and their stationary counterpart , i.e. the wannier stark ( ws ) ladders , have been experimentally observed in the context of atom optics @xcite . in the present paper
we demonstrate that an atomic beam dropped onto a laser potential as that of @xcite can undergo under certain conditions multiple oscillations inside the laser region .
we show that this effect can be physically explained using an image of energetically forbidden spatial regions , what we call spatial gaps .
in particular the effect is produced due to a combination of partial landau
zener tunneling through sufficiently narrow spatial gaps , and a partial reflection on them . in this sense
the laser arrangement behaves as an atomic wave fabry - prot interferometer @xcite .
we have also analysed the similarities and differences between these oscillations and bloch oscillations . due to these similarities and differences
we have called the effect bloch like oscillations .
we show that the effect can be easily observed analysing either the reflected momentum spectrum ( where wannier stark like resonances appear ) or the temporal evolution of an atomic wavepacket .
we prove that the multiple reflections appear also for attractive potentials , i.e. quantum multiple reflections are also possible .
the scheme of the paper is as follows . in sec .
[ sec : theo ] we briefly review the theoretical model already presented in @xcite . sec .
[ sec : spgaps ] is devoted to the development of the spatial gaps image .
[ sec : fabry ] uses the spatial gaps image to explain the appearance of resonancies in the momentum spectrum of the reflected atoms , and compare the effect with a fabry
prot interferometer .
[ sec : bows ] discuss the similarities and differences between the presented oscillations and bloch oscillations . in sec .
[ sec : bloch ] we present the temporal evolution of an atomic wavepacket for the case of a repulsive potential , whereas in sec.[sec : qmr ] the case of quantum multiple reflections is considered .
we finalize in sec.[sec : conclu ] with some conclusions .
in this section we briefly discuss the theoretical model we use in the paper . for a more detailed discussion see @xcite .
let us consider the same laser arrangement as that of @xcite , formed by the interference of two laser beams ( fig .
[ fig:1 ] ) of the same intensity , polarisation , and with the same gaussian profile , but with respective wave vectors : @xmath0 and @xmath1 . in the zone very close to @xmath2 the electric field , which is assumed to be linearly polarised in the @xmath3-direction , can be written in the form : @xmath4 where @xmath5 is the amplitude of each laser , @xmath6 is the center of the laser region , and @xmath7 is the halfwidth of the gaussians . let us assume that the laser frequency @xmath8 is quasi
resonant with some atomic transition between the ground state and an excited state ( of energies @xmath9 and @xmath10 , respectively ) , and therefore we can treat the atom as a two - level system . in this paper , we assume a large internal detuning ( @xmath11 ) ) , in such a way that the adiabatic approximation will be valid @xcite . in order to obtain a scalar schrdinger equation to describe the atomic interaction with the laser , we follow the standard formalism developed in @xcite , except for the inclusion of the gravitational field .
the scalar equation takes the form : @xmath12\psi(z ) , \label{scheq}\ ] ] where @xmath13 with @xmath14\cos ( k(z - z_{c})\sin\phi)$ ] . the sign in eq .
( [ vz ] ) depends on the detuning : the `` '' sign corresponds to @xmath15 , and the `` + '' corresponds to @xmath16 . the coupling is given by @xmath17 , which is the rabi frequency associated with each laser , where @xmath18 , with @xmath19 the transition dipole .
@xmath20 is the @xmath21-momentum in @xmath22 , sufficiently far away from @xmath6 to consider that the field in @xmath22 is negligible . to make the model as realistic as possible one could include the diffuse scattering resulting from spontaneous emission by adopting the lossy vector schrdinger equation method of @xcite , i.e. @xmath23 , where @xmath24 is the spontaneous emission frequency .
however a large detuning is considered in the paper to avoid the effects of the spontaneous emission . using convenient units of length ( @xmath25 ) , momentum ( @xmath26 ) and frequency ( @xmath27 ) )
, we can rewrite the schrdinger equation in a dimensionless form : @xmath28 in which the tilde denotes dimensionless units . in eq .
( [ scheq2 ] ) , we find the function @xmath29 .
@xmath30 is a gravitational parameter which introduces some differences depending on the mass of the atom . since we shall assume that @xmath31 , we can define a parameter @xmath32 which determines the strength of the laser potential
this can be easily shown by introducing a taylor expansion on the right hand side of eq.([scheq2 ] ) : @xmath33 where @xmath34 note that the potential is formed by the product of a cosine squared function and a gaussian envelope .
in all the figures throughout the paper we analyse the case of @xmath35@xmath36 @xmath37li transition , whose parameters are @xmath38 nm , @xmath39 s@xmath40 , @xmath41 s@xmath40 and @xmath42 . the position of the center of the laser region is at @xmath43 mm , and the halfwidth of the laser gaussians is given by @xmath44 @xmath45 m .
in this section we present a model which allows us to understand the physics behind the results we will observe in the following sections . from eq .
( [ vz2 ] ) we observe that the potential @xmath46 is quasi - periodic , except for the gaussian envelope , its periodicity given by @xmath47 . in order to better understand the effects of periodicity ,
let us remove gravitation and spontaneous emission , and for the time being let us forget the smooth gaussian dependence . with these assumptions ,
the laser potential is a simple cosine squared potential , whose amplitude ( @xmath48 ) can be positive ( if @xmath49 ) or negative ( if @xmath50 ) , depending on which dressed state is reached .
it is well known that a periodic potential leads to an energy structure of allowed and forbidden bands @xcite .
only the incoming momentum components @xmath20 whose associated kinetic energy lies within an allowed band can propagate inside the potential .
if @xmath20 does not satisfy the band condition ( i.e. if @xmath51 lies in a gap ) , then the atoms with this incoming momentum can not propagate inside the laser region and are therefore reflected , reflection bands being formed .
however , this image is excessively simple . in order to obtain physical insight into the numerical results , obtained by direct resolution of the schrdinger eq.([scheq3 ] )
, we must take into account the gaussian laser envelope .
the analysis is greatly simplified if , as we consider in this paper , the width of the gaussian envelope is very large compared to the cosine squared periodicity .
in particular , in the cases analysed below the width of the gaussian envelope of the potential at @xmath52 is @xmath53 times the cosine squared period .
if this condition is satisfied we can consider that within small intervals of the envelope we have a large number of cosine squared oscillations of approximately constant amplitude .
we can therefore define an energy band structure for each of these intervals considered as the band structure calculated for an infinitely extended periodic potential with this constant amplitude .
we can extend this reasoning and define a local band structure for each position @xmath54 inside the laser region , or in other words for each value of the gaussian envelope ( @xmath55 ) .
this local band structure is calculated for a cosine squared potential of infinite number of periods with constant amplitude @xmath55 .
therefore each position within the laser region is linked with an energetic spectrum of allowed and forbidden energies , a spatially dependent band structure being formed .
[ fig:2]a shows such spatial band structure for the case of a repulsive potential of @xmath56 .
in particular , for certain momenta some spatial regions are energetically forbidden ( white regions ) .
we will call these regions _ spatial gaps_. fig .
[ fig:2]b shows the reflectivity for different incoming momentum components @xmath20 , obtained by direct resolution of eq .
( [ scheq3 ] ) .
the spatial gaps image allows an intuitive understanding of the physical processes behind these numerical results . if an atom with some momentum component finds during its travel inside the laser region a spatial gap , then the propagation through it is energetically forbidden , and consequently the atom is reflected
this point becomes clear by comparison between figs .
[ fig:2]a and [ fig:2]b . in @xcite
we also show that this model explains very well the effects of the spontaneous emission and gravitation . in this paper
we will assume a sufficiently large detuning to neglect the spontaneous emission effects .
the gravitational effects will be also negligible for the momentum components of interest in this paper and for the laser arrangement considered .
in sec . [ sec : spgaps ] we have observed that an atom is reflected if it finds a spatial gap inside the laser region .
therefore the spatial gap acts as a potential barrier .
note that : * if the spatial gap is sufficiently narrow then transmission via tunneling becomes possible .
this tunneling produced from one allowed spatial region to another is no more than the well known landau
zener tunneling between allowed energetic bands . *
the form of the spatial gap depends on the specific incoming momentum one considers , and hence this effective potential barrier has the interesting property that it changes its form and width depending on the incoming atomic momentum . for certain interval of momenta ( around @xmath57 in fig .
[ fig:2]b ) , several peaks in the momentum spectrum of reflection can be observed .
these peaks are depicted in detail in fig .
[ fig:3]b .
the physical process behind these peaks can be well understood by using the spatial gaps image of sec .
[ sec : spgaps ] . in fig .
[ fig:3]a we present in detail the region in fig .
[ fig:2]a corresponding with the region depicted in fig.[fig:3]b .
as we observe in this figure the spatial gap , i.e. the effective potential barrier , is very narrow in this region .
this fact allows the possibility of tunneling of the wavefunction through the spatial gap ( from point @xmath58 to point @xmath59 ) , and therefore a partial reflection is produced .
since the envelope is symmetric , the transmitted part reaches again an spatial gap ( point @xmath60 ) , from where the atoms can be partially transmitted again by tunneling until point @xmath61 , or can be reflected back to point @xmath59 , from where they can be partially reflected and so on .
this process leads to multiple oscillations between points @xmath59 and @xmath60 , i.e. the region from @xmath59 to @xmath60 acts as fabry
prot cavity @xcite , where @xmath59 and @xmath60 act as lossy mirrors . as we indicated previously , since the form of the effective barrier depends on the momentum , the length of the `` cavity '' depends also on the momentum .
following the optical analogy , this is the same case we would have if the length of a fabry prot cavity was different depending on the wavelength of the light .
the peaks we find in the momentum spectrum of the reflection can therefore be explained as resonances of this special fabry
prot cavity .
other way to understand these resonances in the reflection spectrum is provided by a well known phenomenum of solid state physics , namely the wannier
stark ( ws ) ladders @xcite .
this effect is simply a stationary counterpart ( i.e. in the frequency domain ) of another effect in the time domain called bloch oscillations ( bo ) .
let us briefly review this concept .
bo appear when particles within a periodic potential are affected by a constant acceleration .
we must point out that although this effect was initially observed in solid state context @xcite , several recent experiments @xcite have reported the same phenomena in atom optics .
bo can be easily understood for weak potentials @xcite : due to the acceleration , the momentum of the particles increases linearly according to newton s law until it reaches a critical value satisfying the bragg condition , then the atomic wave is reflected and its momentum is reversed .
the atom travels again under newton s law until it reaches other bragg condition and then it is reflected again .
then , the bo can be understood as multiple reflections between two bragg reflections . for larger potentials
the bo can be understood using the spatial gaps image ( as pointed out previously , bragg reflection is a particular case of the band gap structures for weak potentials ) .
let us consider an infinitely extended periodic potential of constant amplitude affected by an external force .
let us suppose that this force is linear in the spatial coordinate .
hence the band structure varies in the space leading to tilted allowed and forbidden spatial regions ( fig .
[ fig:4]a ) .
let us suppose a particle inside an allowed region .
this particle moves until it reaches a forbidden region ( point @xmath58 ) , from where it is reflected .
then it travels again within the allowed region until it finds another forbidden region ( point @xmath59 ) , from where it is reflected and so on , leading to multiple oscillations . note that in absence of external force , no tilting is present ( fig.[fig:4]b ) and hence there are not multiple oscillations .
we observe that the oscillations reported in sec .
[ sec : fabry ] are certainly similar to the bo . in particular
, both are due to the same reason , i.e. multiple reflections between forbidden regions .
however several important differences can be pointed out : * the physical process which produces the `` cavity '' is basically different : in the bo is the tilting of the bands due to an external force , while in our case is the symmetry of the gaussian envelope . *
whereas in the bo the reflection is produced between two different spatial gaps , in the oscillations reported here both sides of the `` cavity '' are produced by the same spatial gap , which is curved due to the gaussian shape of the laser envelope .
* in the bo the atom is initially inside the allowed region between two forbidden ones . in our case
the atom enters from outside , and therefore needs to tunnel a narrow spatial gap to enter into the internal allowed region where it oscillates . due to
these similitudes and differences , we call the resonances in the reflection spectrum wannier stark like resonances .
we have therefore interpreted the resonances appearing in fig .
[ fig:3]b as wannier stark like resonances . as for standard ws
ladders , the wannier
stark like resonances are linked in the time domain with what we call bloch like oscillations . in order to observe this effect
we have numerically calculated the evolution of an atomic wavepacket through the laser arrangement . in order to achieve this
we have solved the schrdinger eq.([scheq3 ] ) evaluating the wavefuction inside and outside the laser region for different incoming momentum components @xmath20 .
note that at this point turns very important the possibility to calculate the wavefunction inside the laser region , which can not be calculated by using a transfer matrix method as that of @xcite .
we have employed a finite differencing numerical method previously presented in @xcite , which allows us to know the wavefunction also inside the laser region .
once we have calculated the spatial behavior of the wavefunction for the different momentum components @xmath62 , we obtain the temporal evolution of an atomic wavepacket by applying a one dimensional fourier transform : @xmath63 where @xmath64 is an initial gaussian momentum distribution of the form : @xmath65 where @xmath66 is the central momentum component of the atomic wavepacket .
[ fig:5 ] shows the evolution of an atomic wavepacket with @xmath67 , @xmath68 through a laser arrangement as that of fig .
[ fig:3 ] .
the initial position of the centre of the atomic wavepacket is @xmath69 .
[ fig:5]a shows the snapshot at @xmath70 , whereas figs .
[ fig:5]b , c and d show respectively the snapshots at @xmath71 , @xmath72 and @xmath73 , with @xmath74 . in all the figures
the gaussian form of the laser envelope is represented in dashed lines for comparative purposes . in fig .
[ fig:5]b a partial reflection and tunneling is clearly produced . the reflected wavepacket travels to the left dissapearing , while the transmitted wavepacket travels within the laser region . in fig.[fig:5]c
the previously transmitted wavepacket is again splitted into two parts : a transmitted part which travels to the right dissapearing , and a reflected part which travels back in the laser region . again in fig .
[ fig:5]d a new splitting is produced .
the observation of successive multiple reflections becomes more difficult due to the wavepacket spreading and due to the losses in the different partial transmissions .
we show in this section that the previously reported multiple oscillations can also be observed for the case of attractive laser potential , i.e. it is possible to achieve multiple reflections in a laser potential that classically allows none .
[ fig:6 ] shows a detail of the reflection spectrum for the case of an attractive potential with @xmath75 .
we clearly observe the appearance of several resonances as those of fig .
[ fig:3 ] .
these distortions are no more than the previously analysed ws like resonances . as for the repulsive case
, a narrow spatial gap appears leading to the already analysed effect of partial landau
zener tunneling , and therefore to multiple reflections , whose effect in the momentum domain is the appearance of the ws like resonances . as in the previous case ,
bloch like oscillations can be observed if we monitorize the evolution of an atomic wavepacket through the laser region .
fig.[fig:7 ] shows this evolution for the case of the same atomic wavepacket of fig.[fig:5 ] , but now the laser is that of fig.[fig:6 ] .
the first snapshot ( fig . [ fig:7]a ) is at @xmath70 , whereas figs .
[ fig:7]b , c , d are respectively obtained at @xmath76 , @xmath77 , @xmath78 , with @xmath74 , where we observe different wavepacket splittings due to partial tunneling and reflection . as in the repulsive case
multiple oscillations are possible , i.e. quantum multiple reflections also appear .
in this paper we have analysed the reflection of an atomic beam dropped onto a laser with a periodic profile modulated by a gaussian envelope , formed in the interference region of two gaussian laser beams .
we have numerically calculated the atomic reflection on such arrangement by direct resolution of the corresponding schrdinger equation , and explained the band like character of the reflection momentum spectrum using a band theory model of energetically forbidden spatial regions , or spatial gaps .
we have probed that due to a partial landau
zener tunneling through sufficiently narrow spatial gaps , and also due to the gaussian symmetry of the laser potential , an atom can undergo multiple reflections inside the laser region , in a process which resembles a fabry
prot interferometer but for atomic waves .
these multiple oscillations have been compared with the well - known bloch oscillations of solid state physics , analysing the similarities and differences between both oscillations . in particular
although both processes are due to multiple reflections between energetically forbidden spatial regions , in the effect presented in this paper the spatial gap is curved due to the gaussian symmetry and the atom oscillates between two barriers which are actually part of the same spatial gap , contrary to the bloch oscillations in which the atom oscillates between two different spatial gaps .
also , the reported multiple oscillations are not due to an external force as bloch oscillations but due to the gaussian form of the envelope . due to these similarities and differences
we have called these oscillations bloch like oscillations .
we have analysed the effect both in the frequency domain ( where wannier stark like resonances appear ) and in the time domain ( observing the bloch like oscillations ) .
we have proved that the effect appears not only for repulsive laser potentials , but also for attractive ones , i.e. quantum multiple reflections are also possible .
* acknowledgments . *
partial support from the spanish direccin general de investigacin cientfica y tcnica ( grant no .
pb950955 ) and from the junta de castilla y len ( grant no .
sa 16/98 ) is acknowledged .
s. chu , nobel lecture , rev .
phys . * 70 * , 685 ( 1998 ) , c. cohen tannoudji , nobel lecture , _
ibid_. , 707 ; w. d. phillips , nobel lecture , _
ibid_. , 721 .
d. w. keith , m. l. schattenburg , h. i. smith and d. e. pritchard , phys .
lett . * 61 * , 1580 ( 1988 ) .
o. carnal and j. mlynek , phys .
* 66 * , 2689 ( 1991 ) .
o. carnal , m. sigel , t. sleator , h. takuma and j. mlynek , phys .
lett . * 67 * , 3231 ( 1991 ) .
m. r. andrews , c. g. townsend , h. j. miesner , d. s. durfee , d. m. kurn and w. ketterle , science * 275 * , 637 ( 1997 ) .
v. i. balykin , v. s. letokhov , yu . b. ovchinnikov and a. i. sidorov , pisma zh .
. fiz . * 45 * , 282 ( 1987 ) [ jetp lett .
* 45 * , 353 ( 1987 ) ] ; v. i. balykin , v. s. letokhov , yu . b. ovchinnikov and a. i. sidorov , phys .
lett . * 60 * , 2137 ( 1988 ) . t. m. roach , h. abele , m. g. boshier , h. l. grossman , k. p. zetie and e. a. hinds , phys . rev . lett . *
75 * , 629 ( 1995 ) .
m. wilkens , e. schumacher and p. meystre , phys .
a * 44 * , 3130 ( 1991 ) .
p. j. martin , b. g. oldaker , a. h. miklich and d. e. pritchard , phys .
lett . * 60 * , 515 ( 1988 ) ; e. m. rasel , m. k. oberthaler , h. batelaan , j. schmiedmayer and a. zeilinger , phys . rev . lett .
* 75 * , 2633 ( 1995 ) .
e. m. rasel , m. k. oberthaler , h. batelaan , j. schmiedmayer and a. zeilinger , phys .
lett . * 75 * , 2633 ( 1995 ) ; d. m. giltner , r. w. mcgowan and siu au lee , phys .
lett . * 75 * , 2638 ( 1995 ) .
e. yablonovitch , phys .
* 58 * , 2059 ( 1987 ) ; e. yablonovitch , j. opt .
b * 10 * , 283 ( 1993 ) .
l. santos and l. roso , phys .
a * 58 * , 2407 ( 1998 ) ; l. santos and l. roso , j. phys . b. : at . mol
phys . * 30 * , 5169 ( 1997 ) .
m. b. dahan , e. peik , j. reichel , y. castin and c. salomon , phys .
lett . * 76 * , 4508 ( 1996);q .
niu , x. g. zhao , g. a. georgiakis and m. g. raizen , phys .
* 76 * , 4504 ( 1996 ) ; s. r. wilkinson , c. f. bharucha , k. w. madison , q. niu and m. g. raizen , phys .
lett . * 76 * , 4512 ( 1996 ) .
also a two gaussian laser arrangement can behave as a fabry prot like interferometer for atoms , see l. santos and l. roso , phys .
a * 57 * , 432 ( 1998 ) .
r. deutschmann , w. ertmer and h. wallis , phys .
a * 47 * , 2169 ( 1993 ) .
w. zhang and b. c. sanders , j. phys .
b * 27 * , 795 ( 1994 ) .
n. w. ashcroft and n. d. mermin , solid state physics ( saunders , philadelphia , 1976 ) .
m. born and e. wolf , principles of optics , pergamon , london , 1980 .
g. h. wannier , phys . rev . * 117 * , 432 ( 1960 ) .
e. e. mendez and g. bastard , phys . today * 46 * , no 6 , 34 ( 1993 ) .
e. peik , m. b. dahan , y. bouchoule , y. castin and c. salomon , phys .
a * 55 * , 2989 ( 1997 ) .
s. m. tan and d. f. walls , phys .
a * 50 * , 1561 ( 1994 ) . | we show that under certain circumstances an atom can follow an oscillatory motion in a periodic laser profile with a gaussian envelope .
these oscillations can be well explained by using a model of energetically forbidden spatial regions .
the similarities and differences with bloch oscillations are discussed .
we demonstrate that the effect exists not only for repulsive but also for attractive potentials , i.e. quantum multiple reflections are also possible . |
understanding interference has been seminal in optics .
more than two centuries ago , young presented his bakerian lecture which contained an experimental demonstration of the general law of interference of light@xcite .
fifteen years later , fresnel and arago studied the effect of the polarization state of light beams in the phenomena of interference@xcite .
thus , through interference , evidence of the transverse wave nature of light was brought forward . in the second half of the twentieth century
there were many studies on the interference of polarized light for the undergraduate laboratory .
various interferometric methods were proposed to carry out such experiments using ( i ) a young double slit experiment , covering both slits with different polarizing filters@xcite ; ( ii ) the polarizing properties of the ordinary and extraordinary axis in a double refracting crystal of calcite @xcite or a nematic liquid crystal@xcite ; ( iii ) a ronchi grating as beam splitter placing polarizing filters afterwards the grating and analyzing at the conjugate plane of the grating through a focusing element @xcite ; ( iv ) a mach - zehnder interferometer with polarizing filters at each arm @xcite .
just to list a few schemes .
in the same period of time , a sound mathematical description for the young double slit experiment with polarized light was presented characterizing the resulting field by the stokes parameters@xcite . also , an alternate mathematical formulation and a proposal for using this scheme in image processing involving wollaston prisms was discussed@xcite .
recently , this was used to understand the concept of a quantum eraser , where the polarization of two macroscopic fields is manipulated to loose or restore interference fringes @xcite . in all of the aforementioned studies , discussions focused on understanding the resulting intensity pattern , which is directly related to the fresnel - arago laws .
to our knowledge , little has been said about the polarization state of the resulting field ; it is possible to obtain a @xmath0-linearly polarized field from the superposition of two collinear fields with right and left circular polarization , with the given angle @xmath1 $ ] degrees defined by the phase difference between the superposing fields@xcite .
we believe the analysis of the superposition of two light fields could go beyond the study of fresnel - arago laws .
our motivation comes from the fact that a rich spatially dependent polarization structure , known as a polarization grating , arises from the superposition of two non - collinear polarized light beams with different polarizations .
this polarization grating can be used to simplify the fabrication of three dimensional periodic microstructres eliminating complicated procedures to optimize contrast and polarization@xcite .
it can be recorded in polarization sensitive materials , like bacteriorhodopsin , to perform phase shifting shearing interferometry without neither mechanical moving parts nor mounted gratings@xcite .
it can also be used to simultaneously rotate or control the orientation of multiple microscopic birefringent particles , which might be useful for biomedical applications@xcite .
these are just a few examples involving polarization gratings in modern research .
our experimental proposal uses a slightly modified mach - zehnder interferometer to produce light with spatially varying polarization .
our theoretical description of the experiment is based on jones calculus@xcite and stokes parameters@xcite for polarized monochromatic light .
the analysis is complemented with three specific examples involving the superposition of combinations of linearly and circularly polarized light showing some specific polarization gratings that can be produced with the proposed scheme .
the presence of such polarization structures is experimentally confirmed through qualitative analysis of the resulting light field with a linear polarizer .
we encourage the reader to reproduce the experiment and calculate the values for the stokes parameters of the resulting field@xcite . for readers who may not be familiar with jones calculus and
stokes parameters we suggest the textbooks by simmons and guttmann@xcite or by born and wolf@xcite in addition to the articles mentioned above@xcite . for those interested in an advanced treatment ,
we recommend as a starting point the articles by tervo _ et .
al_@xcite and roychowdhury and wolf@xcite , as well as the introductory book by wolf@xcite .
their analyses deal with polarization and coherence degree of superposed arbitrary electromagnetic fields in three dimensional space .
figure [ fig : fig2 ] illustrates a simplified version of the superposition scheme .
we focus on the plane of incidence defined by the @xmath2-plane so the @xmath3-coordinate will be obviated .
two polarized monochromatic plane waves of light intersect with a small angle @xmath0 , such that @xmath4 in radians , at some point @xmath5 on the detection line @xmath6 .
such light fields are described by the equations @xmath7 where the distances @xmath8 are the distances from the @xmath9-th beam source to the point @xmath10 , e.g. @xmath11 .
the counterclockwise rotation of the polarization state of the second beam about the @xmath3-axis is introduced by means of the rotation matrix @xmath12 the unitary polarization vector state @xmath13 is , up to a phase constant , a jones vector @xmath14 with parameters in the ranges @xmath15 $ ] and @xmath16 $ ] . the symbols @xmath17 and @xmath18 are the unitary vectors in the @xmath19- and @xmath3directions .
the stokes parameters for the total field @xmath20 , at a point @xmath10 on the detection line @xmath6 are given by the expression @xmath21 .
\end{aligned}\ ] ] the angle brackets are shorthand notation for time averaging over the detection interval , which is large compared to the period associated with the optical radiation frequency , @xmath22 for plane waves ( asterisk meaning complex conjugation ) .
the symbol @xmath23 for @xmath24 denotes the pauli matrices @xmath25 the parameters @xmath26 are the stokes parameters for the @xmath27-th polarization vector @xmath28 @xmath29 finally , the phase difference parameter @xmath30 can be approximated as @xmath31 with the initial phase difference between the sources of the beams given by @xmath32 . in our experimental scheme
the source for both beams is the same laser so the initial phase difference is null , @xmath33 .
the experimental setup uses small angles , @xmath34 radians , that for practical purposes , @xmath35 and @xmath4 .
these experimentally feasible restrictions allow us to consider the polarization state of the second field beam @xmath36 in its own reference frame and in the general reference frame almost equal ; i. e. the jones vectors @xmath37 and @xmath38 are almost parallel , @xmath39 these approximations simplify the theoretical treatment , thus the real parts involved in the last term of eq .
are given by @xmath40 as usual , the stokes parameter @xmath41 is useful for discussing the intensity profile at the detection line such as discussed by pescetti@xcite or collet@xcite , while the latter three parameters , @xmath42 to @xmath43 , relate to the polarization state of the field . our purpose is to understand the polarization properties of the total field . in order to do so ,
let us consider the interfering beams carrying orthogonal polarizations , that is @xmath44 .
notice that two orthogonal polarization vectors can be written as @xmath45 and @xmath46 if we relax the restrictions on the domain of @xmath47 .
the stokes parameters for orthogonal polarization vectors fulfill the condition @xmath48 for @xmath49 ; the points on the polarization sphere that represent these vectors being antipodes .
it is also possible to parametrize the amplitudes of the fields as @xmath50 in a range @xmath51 $ ] , such that the corresponding normalized stokes parameters for the total electromagnetic field on the detection line are @xmath52 it is possible to write the latter three normalized stokes parameters , @xmath53 to @xmath54 , as the vector , @xmath55 with the vector @xmath56 defining a great circle on the @xmath57-plane of the polarization sphere , @xmath58 and the rotation matrices given in the traditional way @xmath59 equation [ eq : orthogonaleq ] implies that the parameter @xmath47 generates a counterclockwise rotation around the @xmath60-axis , the parameter @xmath61 acts as a scaling factor and a @xmath62-translation on the great circle @xmath56 , and the parameter @xmath63 as a counterclockwise rotation around the @xmath64-axis .
+ the counterclockwise rotations were expected .
the great circle @xmath56 is obtained from the superposition of the fields emitted by two sources with equal amplitudes of emission and horizontal / vertical linear polarizations .
the rotation @xmath65 transforms the two sets of stokes parameters that map the orthogonal polarization pair @xmath66 , @xmath67 into those mapping any other orthogonal pair @xmath45 and @xmath46 .
figure [ fig : parameters ] shows an example of the effect of the set of parameters @xmath68 on the behavior of the polarization for the total field at the detection line .
an interactive demonstration , where the user can input any combination of these parameters and obtain the corresponding stokes parameters on the polarization sphere , is provided online@xcite .
we present the experimental realization and discussion of three cases that can shed more light on the problem when working in the undergrad laboratory . in the first two cases ,
beams are used with polarization states orthogonal to each other , eq .. in the third case , the general treatment is used , eq .. the experimental setup is shown in fig . [
fig : expsetup ] . using a mach - zehnder interferometer ,
two beams are superposed at the back aperture of a microscope objective .
the image of the beams superposition is formed at the focal region of the objective .
each of the beams is given a specific polarization state through a suitable retarder at the corresponding arm of the interferometer .
characterization of the superposition polarization state is performed by placing a linear polarizer as analyzer behind the focus of the objective .
images are captured for angles of @xmath69 , @xmath70 , @xmath71 , and @xmath72 radians of the linear polarizer axis with respect to the horizontal axis .
a continuous wave solid state laser , emitting at a @xmath73 nm wavelength with a linear vertical output polarization state , ratio 100:1 , is used as a source .
the beam splitters , bs1 and bs2 , are non - polarizing and one of them , bs2 , is mounted on a linear displacement stage in order to control the angle of interference , @xmath0 in fig .
[ fig : fig2 ] . in the interferometer , each beam reflects once from a beam splitter and once from a mirror , thus canceling the phase shift introduced by each single reflection .
a @xmath74 microscope objective is used as an imaging element . a black and white 1/2-ccd , located just after the focal distance of the imaging element ,
is used to capture the images .
the theoretical results shown here are calculated to fit the results at the ccd , whose range of detection is approximately @xmath75 $ ] mm .
an intersection angle given by @xmath76 radians is used .
a general interactive demonstration where the user can set at will all of the two beams parameters is provided online@xcite .
this demonstration generates the corresponding stokes parameters on the detection line , their representation on the polarization sphere , and the polarization ellipse for a given position on the detection line . in the following paragraphs polarization states
will be denoted using the notation : circularly right / left , @xmath77 , elliptically right / left , @xmath78 , linearly @xmath0 , @xmath79 . in this case , p1 is a half - wave plate , whose fast axis is placed at @xmath70 radians with respect to the @xmath19-axis , producing horizontal linear polarization and p2 is removed to keep vertical linear polarization .
figure [ fig : case1a ] shows the intensities captured by the ccd camera at the four analyzer orientations for this configuration alongside the corresponding theoretical intensities .
the vertical lines crossing both columns of figures are presented as reference marks to relate experimental with theoretical results .
figure [ fig : case1b ] shows that the analysis is consistent with a polarization of the resulting field varying on a meridian of the polarization sphere crossing the @xmath85 linear polarization states ; i.e. polarization varies periodically with the cycle : @xmath86 @xmath87 @xmath88 @xmath87 @xmath89 @xmath87 @xmath90 @xmath87 @xmath86 , with intermediate states of elliptical polarization .
the alphabetical labels ( a , b , c , d ) in both figures identify the intensities of the analyzed polarization states in fig .
[ fig : case1a ] , as well as the related stokes parameters , their corresponding mapping on the polarization sphere , and the polarization state they represent in fig .
[ fig : case1b ] , for the field at a given point @xmath19 .
thus , the labels a / b / c / d show where the field has a @xmath86/ @xmath90/@xmath89/@xmath91 polarization state . in this configuration , p1 and p2
are quarter - wave plates whose fast axes are placed at @xmath94 radians with respect to the @xmath19-axis , thus their fast axes are perpendicular to each other , producing right circular and left circular polarizations .
figure [ fig : case2a ] shows the experimental and theoretical intensities corresponding to the four analyzer positions .
the vertical lines are used to relate both results .
figure [ fig : case2b ] shows that the analysis is consistent with a polarization of the resulting field varying on the equator of the polarization sphere ; i.e. polarization is always linear with orientation angle varying periodically from @xmath95 to @xmath96 radians .
it is shown that `` the superposition of the right and left circularly polarized light yields linearly polarized light but the direction of the polarization depends on the phase angle between the two beams '' @xcite .
the alphabetical labels in both figures identify the intensities of the analyzed polarization states in fig .
[ fig : case2a ] , as well as the related stokes parameters , their corresponding mapping on the polarization sphere , and the polarization state they represent in fig .
[ fig : case2b ] , for the field at a given point @xmath19 .
thus , the label a / b / c / d shows where the field has a @xmath97/ @xmath91/@xmath98/@xmath90 polarization state .
finally , here p1 is a quarter wave plate , whose fast axis is placed at @xmath70 radians with respect to the @xmath19-axis , producing right circular polarization and p2 is a half - wave plate , whose fast axis is placed at @xmath70 radians with respect to the @xmath19-axis , producing horizontal linear polarization .
figure [ fig : case3a ] shows the intensities captured by the ccd camera at the four analyzer orientations alongside the corresponding theoretical intensities .
figure [ fig : case3b ] shows that the analysis is consistent with a polarization of the field varying on some circle on the polarization sphere ; i.e. polarization varies periodically being elliptically polarized but for two points where it is linearly @xmath85 polarized .
the alphabetical labels show elliptical polarization states with varying eccentricity and inclination of the major axis , for this configuration .
their intensities , fig .
[ fig : case3a ] and their related stokes parameters , their corresponding mapping on the polarization sphere , and the polarization state they represent , fig . [ fig : case3b ] , for the field at a given point @xmath19
we have presented an experimental scheme that an undergraduate student can use for analyzing the polarization state of the superposition of two slightly non - collinear polarized light beams .
the equations modeling the stokes parameters for this experiment have been presented . the explicit case of interfering orthogonal polarizations
was discussed and complemented with two particular configurations to help elucidating this scheme ; a third experimental configuration involving a general case , the interference of two non - orthogonal polarization beams , was also presented .
it has been shown that the polarization state of light is spatially dependent in all cases due to the spatially dependent phase between the beams introduced by the impinging angle between them .
our experimental scheme can be implemented into an optical tweezer to demonstrate the transfer of intrinsic angular momentum to birefringent particles@xcite using polarization structures@xcite .
this could also attract the attention to polarization , interference and mechanical properties of light at the undergraduate level .
b. m. r. l. acknowledges financial support from unam - dgapa .
i. r. v. acknowledges financial support from conacyt .
both authors thank r. juregui and s. hacyan for fruitful discussion .
authors are particularly grateful to k. volke - seplveda .
t. young,``experiments and calculations relative to physical optics '' , phil .
. lond . * 94 * , 116 ( 1804 ) .
d. f. j. arago and a. j. fresnel , `` on the action of rays of polarized light upon each other '' , ann .
* 2 * , 288304 ( 1819 ) .
j. l. hunt and g. karl , `` interference with polarized light beams '' , am .
* 38 * 12491259 ( 1970 ) .
d. pescetti , `` interference of elliptically polarized light , '' am .
* 40 * , 735740 ( 1972 ) .
s. mallick , `` interference with polarized light '' , am .
* 41 * , 583584 ( 1973 ) c. pontiggia , `` interference with polarized light '' , am . j. phys . * 39 * , 679 ( 1971 ) . j. l. ferguson , `` a simple , bright demonstration of the interference of polarized light '' , am . j. phys . * 52 * , 11411142 ( 1984 ) . l. z. cai and x. l. yang , `` interference of circularly polarized light : contrast and application in fabrication of three dimensional periodic microstructures '' , opt
. laser technol . *
34 * , 671674 ( 2002 ) .
f. gori , `` measuring stokes parameters by means of a polarization grating '' , opt .
* 24 * , 584586 ( 1999 ) .
b. schaefer _ et .
`` measuring the stokes polarization parameters '' , am . j. phys . * 75 * , 163169 ( 2007 ) .
j. tervo _ et .
al . _ , `` degree of coherence for electromagnetic fields '' , opt
. exp . * 10 * , 11371143 ( 2003 ) .
roychowdhury and e. wolf , `` young s interference experiment with light of any state of coherence and of polarization '' , opt
252 * , 268 - 274 ( 2005 ) .
b. m. rodrguez - lara and i. ricardez - vargas , `` stokes parameters for the superposition of two slightly noncollinear light beams with orthogonal polarizations '' from the wolfram demonstrations project http://demonstrations.wolfram.com/stokesparametersforthesuperpositionoftwoslightlynoncollinear/ b. m. rodrguez - lara and i. ricardez - vargas , `` stokes parameters for superposition of two slightly noncollinear polarized beams '' from the wolfram demonstrations project http://demonstrations.wolfram.com/stokesparametersforsuperpositionoftwoslightlynoncollinearpol/ | using a scheme based on a mach - zehnder interferometer , we propose an analysis of the superposition of polarized laser beams at a given angle .
the focus of our study is the spatially varying polarization state of the resulting field , also known as a polarization grating , generated by this setup .
our proposal combines a theoretical description of the resulting field in terms of its stokes parameters with an experimental demonstration of the existence of such a polarization grating due to the effects of polarization on beam interference experiments . |
` cern - ph - th/2008 - 120 ` + qcd has a rich phase diagram as a function of quark chemical potential @xmath1 and tem - perature @xmath0 , with at least 3 regimes : confining ( low @xmath8 ) , quark - gluon plasma ( high @xmath0 ) and color superconducting ( high @xmath1 ) .
lattice simulations have provided clear evidence that , for @xmath9 , the temperature - driven `` transition '' between the first 2 regimes is actually a crossover @xcite .
then , the commonly expected @xmath10 phase diagram , depicted fig . 1 ( left ) , contains a critical point caused by the @xmath11 crossover turning into a first - order transition .
this critical point is the object of both experimental search and theoretical lattice investigation .
if located at small enough chemical potential , it could be discovered in high - energy , not - so - heavy ion collisions at rhic or lhc .
in fact , it has been predicted to lie at @xmath12 mev in a celebrated lattice study @xcite . and temperature @xmath0 .
dotted lines denote crossovers , solid lines first - order phase transitions .
the qcd critical point ( end of first - order line ) , marked by a @xmath13 on the left , is absent on the right .
, title="fig:",scaledwidth=49.0%,height=207 ] and temperature @xmath0 .
dotted lines denote crossovers , solid lines first - order phase transitions .
the qcd critical point ( end of first - order line ) , marked by a @xmath13 on the left , is absent on the right . , title="fig:",scaledwidth=49.0%,height=192 ] however , the numerical method of ref .
@xcite , which consists of reweighting results obtained from @xmath11 simulations to @xmath14 , is known to fail and give wrong results without warning ( see , e.g. , @xcite ) when the overlap between the @xmath11 monte carlo ensemble and the target @xmath14 ensemble becomes insufficient .
therefore , the abrupt change observed in @xcite as a function of @xmath1 ( see fig . 1 there ) might be caused by an abrupt breakdown of the numerical approach .
in addition the needed reweighting factors to the @xmath14 target ensemble are not always positive , because of the notorious `` sign problem '' affecting the @xmath14 determinant .
this makes the statistical errors more difficult to control .
finally , the lattices used in @xcite have @xmath15 time - slices , corresponding to a very coarse lattice spacing @xmath16 fm .
as acknowledged in @xcite already , the results may change considerably after extrapolation to the continuum limit @xmath17 .
therefore , the landmark result of @xcite should not be considered the final word on the qcd critical point .
a more cautious crosscheck like the one we present here is warranted . on the theoretical side , the expectation of a critical point as in fig . 1 ( left )
can be traced back to two assumptions : @xmath18 qcd with @xmath19 massless flavours , at @xmath9 , has a second - order temperature - driven phase transition in the @xmath20 universality class @xcite .
this implies , for this theory , the existence of a tricritical point at @xmath21 , and for the @xmath22 theory that of another tricritical point at @xmath23 .
@xmath24 these two tricritical points are analytically connected by a line in the @xmath25 space . however , assumption @xmath18 depends on the strength of the axial @xmath26 symmetry breaking at @xmath27 . and assumption @xmath24 has no other basis than simplicity .
this motivates us to take a careful look at the phase diagram of @xmath22 qcd , generalized to arbitrary quark masses @xmath28 and @xmath29 .
we present our @xmath11 , then @xmath30 findings below . ) , as a function of the light quark mass @xmath31 and the strange quark mass @xmath29 . on the left , theoretical expectations @xcite . on the right ,
@xmath15 lattice results for light quarks , from @xcite .
the straight line corresponds to @xmath32 ( @xmath33 ) .
the curve is the expected scaling behaviour from a tricritical point at @xmath34 mev .
the physical point , marked by a @xmath35 , is in the crossover region .
, title="fig:",scaledwidth=43.0% ] ) , as a function of the light quark mass @xmath31 and the strange quark mass @xmath29 . on the left , theoretical expectations @xcite . on the right ,
@xmath15 lattice results for light quarks , from @xcite .
the straight line corresponds to @xmath32 ( @xmath33 ) .
the curve is the expected scaling behaviour from a tricritical point at @xmath34 mev .
the physical point , marked by a @xmath35 , is in the crossover region .
, title="fig:",scaledwidth=58.0% ]
to protect ourselves against the pitfalls of reweighting to @xmath14 , we consider the effect of an _ imaginary _ chemical potential @xmath36 @xcite
. then the fermion determinant is positive and no overlap problem occurs because we do not reweight .
the drawback is that , in order to analytically continue an observable , its @xmath1-dependence is fitted to low - order taylor series about @xmath11 , thus introducing a truncation error . as a safeguard
, we also calculate the coefficients without truncations @xcite .
note that these derivatives can be expressed , and measured , as non - local observables in the @xmath11 ensemble .
this is the strategy pursued successfully in @xcite for the pressure .
our approach is computationally more economical , and we extend it to the critical line . the first step is to determine , at @xmath11 , the line in the @xmath37 plane for which the temperature - driven phase transition is second order . for lighter quarks , the transition becomes first order as in the chiral limit ; for heavier quarks , lattice simulations show a crossover .
theoretical expectations are displayed in fig . 2 ( left ) .
our determination of the critical line on an @xmath15 ( @xmath16 fm ) lattice is shown fig . 2 ( right ) , and matches expectations , including the possible scaling behaviour in the vicinity of a tricritical point at ( @xmath38 mev ) .
also , the physical point ( with real - world @xmath39 and @xmath40 meson masses ) lies , as expected , in the crossover region .
we now choose a point on the critical line just determined , and monitor the effect of a small @xmath1 .
this exercise has been performed , with similar results , at the two points marked by arrows in fig . 2 ( right ) , with higher accuracy for the @xmath41 case . if conventional wisdom is right , then the surface spanned by the critical line as @xmath1 is turned on will bend towards the crossover region , so as to produce a critical point at some small @xmath1 value for physical quark masses .
this is depicted fig . 3 ( left ) .
on the contrary , an opposite curvature will indicate the absence of a critical point at small @xmath1 , where our truncated taylor expansion is a good approximation .
this is illustrated fig .
3 ( right ) .
the issue is thus to determine whether the @xmath11 second order transition becomes first order or crossover for a small @xmath1 .
we have now confirmed our first study @xcite by a second one using higher statistics and an improved numerical method @xcite .
we measure the change in the binder cumulant @xmath42 , with @xmath43 . at the second order transition
, @xmath44 takes the value 1.604 dictated by the @xmath45 ising universality class .
reweighting with a small imaginary @xmath1 induces small changes in @xmath44 , which can be measured accurately because statistical errors largely cancel between the reweighted and original ensembles .
our current result for the @xmath41 ( @xmath46 ) theory is @xmath47 supporting the unconventional fig .
3 ( right ) , both in the leading and subleading term . is turned on .
depending on the curvature of this surface , a qcd critical point is present ( left ) or absent ( right ) . for heavy quarks , the curvature ( right )
has been determined in @xcite .
the arrows indicate the effect of a finer lattice : the distance between the physical quark masses and the critical surface increases , driving the critical point ( left ) to larger @xmath1 . , title="fig:",scaledwidth=55.0% ] ( -174.9,88)@xmath48 ( -182,88)*@xmath49 * ( -180,49 ) @xmath48 ( -195,49)*@xmath50 * is turned on . depending on the curvature of this surface ,
a qcd critical point is present ( left ) or absent ( right ) . for heavy quarks , the curvature ( right )
has been determined in @xcite .
the arrows indicate the effect of a finer lattice : the distance between the physical quark masses and the critical surface increases , driving the critical point ( left ) to larger @xmath1 . , title="fig:",scaledwidth=55.0% ] ( -180.5,49 ) @xmath48 ( -195,49)*@xmath50 * it is now mandatory to repeat this study on a finer lattice , in order to gain some control over the continuum limit .
@xmath51 ( @xmath52 fm ) simulations are in progress .
the first step , at @xmath11 , already reveals an important shift of the critical line towards the origin : for the @xmath53 theory , the pion mass , measured at @xmath54 with the critical quark mass , decreases from @xmath55 to @xmath56 @xcite . as shown fig . 3
, this considerably increases the distance of the critical surface to the physical point , pushing the critical point to larger @xmath1 ( left ) , or requiring larger higher - order terms of the right sign to bend the critical surface `` back '' ( right ) .
regardless of the sign of the curvature in the continuum limit , this trend alone makes a qcd chiral critical point at small @xmath57 unlikely .
10 y. aoki et al . ,
nature * 443 * ( 2006 ) 675 [ arxiv : hep - lat/0611014 ] .
z. fodor and s. d. katz , jhep * 0404 * ( 2004 ) 050 [ arxiv : hep - lat/0402006 ] .
i. m. barbour et al . , nucl .
suppl . * 60a * ( 1998 ) 220 [ arxiv : hep - lat/9705042 ] .
r. d. pisarski and f. wilczek , phys .
d * 29 * ( 1984 ) 338 .
e. laermann and o. philipsen , ann .
nucl . part .
* 53 * ( 2003 ) 163 [ arxiv : hep - ph/0303042 ] .
p. de forcrand and o. philipsen , jhep * 0701 * ( 2007 ) 077 [ arxiv : hep - lat/0607017 ] .
p. de forcrand and o. philipsen , nucl .
b * 642 * ( 2002 ) 290 [ arxiv : hep - lat/0205016 ] ; nucl .
b * 673 * ( 2003 ) 170 [ arxiv : hep - lat/0307020 ] .
c. schmidt , c. r. allton et al .
d * 68 * ( 2003 ) 014507 [ arxiv : hep - lat/0305007 ] .
de forcrand , s. kim and o. philipsen , pos * lat2007 * ( 2007 ) 178 [ arxiv:0711.0262 [ hep - lat ] ] .
s. kim et al . , pos * lat2005 * ( 2006 ) 166 [ arxiv : hep - lat/0510069 ] . | the phase diagram of qcd , as a function of temperature @xmath0 and quark chemical potential @xmath1 , may contain a critical point @xmath2 whose non - perturbative nature makes it a natural object of lattice studies .
however , the sign problem prevents the application of standard monte carlo techniques at non - zero baryon density .
we have been pursuing an approach free of the sign problem , where the chemical potential is taken as imaginary and the results are taylor - expanded in @xmath3 about @xmath4 , then analytically continued to real @xmath1 . within this approach
we have determined the sensitivity of the critical chemical potential @xmath5 to the quark mass , @xmath6 .
our study indicates that the critical point moves to _ smaller _ chemical potential as the quark mass _
increases_. this finding , contrary to common wisdom , implies that the deconfinement crossover , which takes place in qcd at @xmath4 when the temperature is raised , will remain a crossover in the @xmath1-region where our taylor expansion can be trusted . if this result , obtained on a coarse lattice ,
is confirmed by simulations on finer lattices now in progress , then we predict that no _ chiral _ critical point will be found for @xmath7 mev , unless the phase diagram contains additional transitions . |
the precise knowledge of the @xmath0he(@xmath1)@xmath2be reaction rate is essential in two distinct fields of nuclear astrophysics : in the hydrogen - burning of the sun ( and similar main sequence stars ) and in the big - bang nucleosynthesis . in the pp - chain of solar hydrogen burning
the @xmath0he(@xmath1)@xmath2be reaction is the staring point of the 2nd and 3rd branches of the chain from where the high energy @xmath2be and @xmath5b neutrinos are originated .
owing to the fast technical development of neutrino detectors , the study of the solar neutrinos has entered a high precision era .
the flux of the @xmath5b neutrinos has already been measured by the sno and superkamiokande neutrino detectors @xcite with a precision of 3.5% and similar precision is foreseen for the @xmath2be neutrino flux in the near future @xcite . as the solar neutrino puzzle has been solved by the experimental observation of neutrino oscillation @xcite ,
the precise knowledge of the solar neutrino fluxes allows a more detailed test of the prediction of the standard solar model ( ssm ) .
the ssm uses a number of input parameters to predict the solar neutrino fluxes .
these input parameters include on one hand some solar properties ( like the luminosity , radiative opacity , diffusion , elemental composition ) and on the other hand the rates of nuclear reactions involved in the pp - chain .
the uncertainty in the input parameters translates directly into uncertainties in the neutrino flux prediction .
therefore , the reduction of the input parameter uncertainties is required to carry out a more stringent test of ssm . from the nuclear physics parameters the rate of the @xmath0he(@xmath1)@xmath2be reaction has the highest uncertainty .
its 9% error @xcite at solar energies ( which was reduced to about 5% by a recent work @xcite ) contributes 8% @xcite to the uncertainty in the predicted fluxes for solar @xmath2be and @xmath5b neutrinos . from the very precise measurement of the anisotropy of the cosmic microwave background radiation by the wmap satellite ,
the barion - to - photon ratio in our universe is known with high accuracy @xcite .
since this ratio is the only free parameter in big - bang nucleosynthesis calculations , the abundances of the primordial elements ( d , @xmath0he , @xmath6he , @xmath2li ) can be calculated unambiguously and the results can be compared with astronomical observations . while there is a perfect agreement between the observations and calculations for the d abundance and reasonably good agreement for @xmath6he ,
the predicted abundance of @xmath2li is a factor 2 to 3 higher than the observed one . at the measured barion - to - photon ratio @xmath2li
is produced almost exclusively by the @xmath0he(@xmath1)@xmath2be reaction followed by the @xmath7-decay of @xmath2be .
therefore , the abundance of primordial @xmath2li depends on the rate of this reaction .
the precise knowledge of this reaction rate is the necessary basis of the search for possible solutions to the @xmath2li problem .
the rate of the @xmath0he(@xmath1)@xmath2be can be determined from the reaction cross section at energies relevant for the astrophysical site in question .
hydrogen burning in the core of the sun takes place at roughly 15@xmath810@xmath9k temperature corresponding to a relevant energy range of 10 to 35 kev for the @xmath0he(@xmath1)@xmath2be reaction .
owing to the higher temperatures involved , the relevant energy range for the big - bang nucleosynthesis spans from 160 to 380 kev . at these deep sub - coulomb energies
the reaction cross section has very low values from a few hundreds of nanobarn at the highest big bang energies down to the attobarn range for the pp chain energies .
while direct measurement of the cross section is possible at big - bang energies , the cross section at solar energies can only be calculated by extrapolating the high energy data towards low energies using theoretical considerations .
the extrapolation is normally done with the help of the astrophysical @xmath4 factor which does not contain the trivial energy dependence of the coulomb barrier penetration and which has a finite value at zero energy .
therefore , the @xmath4 factor is usually extrapolated to zero energy instead of a ( temperature dependent ) finite energy of astrophysical relevance . since the @xmath0he(@xmath1)@xmath2be reaction at low energies proceeds through the direct capture process , different theoretical approaches lead to very similar energy dependence of the @xmath4 factor , though small deviations can be found .
the only free parameter in the extrapolations for a certain theoretical model is the absolute magnitude of the @xmath4 factor .
this parameter can be obtained by normalizing the theoretical energy dependence to measured values at energies where data are available .
therefore , it is important to measure the cross section of @xmath0he(@xmath1)@xmath2be with high precision and at energies preferably relevant directly to the big - bang nucleosynthesis and low enough that the extrapolation to zero energy could be done with low uncertainty despite the small differences in various theoretical energy dependences .
the mechanism of the @xmath0he(@xmath1)@xmath2be provides two possibilities for the measurement of the cross section via @xmath3-spectroscopy .
the reaction has a q value of 1.586 mev and at low energies the direct capture can lead to the ground state and first excited state in @xmath2be with the emission of two single @xmath3-lines with energies of e@xmath10+q and e@xmath10+q-429kev . in the latter case the first excited state of @xmath2be decays to the ground state with a 429kev @xmath3-emission .
the detection of these three @xmath3-lines provides one possibility for the cross section measurement ( prompt-@xmath3 or on - line method ) .
the other method for the cross section determination is the measurement of the produced @xmath2be activity . by the measurement of the 478kev @xmath3-radiation following the electron - capture decay of @xmath2be
can be used to determine the total number of fusion reactions ( activation method ) . until the 1980
s the @xmath0he(@xmath1)@xmath2be cross section has been measured seven times @xcite with the on - line method and three times with activation @xcite .
the experiments have been carried out in a wide range of energies and using very different experimental techniques .
the lowest energy reached with the on - line method was @xmath11=107kev .
this means that on - line experiments cover the energy relevant for the big - bang nucleosynthesis but are rather far from the solar energy . with activation , however , the lowest measured energy was 897kev @xcite , far above the astrophysical energies .
each data sets have been used to extrapolate the @xmath4 factor to zero energy .
the extrapolations led to an ambiguous result when considering the on - line and activation experiments separately .
activation experiments obtained higher @xmath4 factor values than the on - line measurements resulting in a discrepancy of about 13% ( roughly 2.5@xmath12 deviation ) at zero energy .
for the solution of this discrepancy several possible hidden experimental errors have been proposed .
these include the not precisely known angular distribution of the prompt-@xmath3 radiation , a possible weak monopole transition in the reaction hidden for the on - line experiments , or the presence of parasitic reactions leading to @xmath2be production in the activation case . as more and more precise data became available for both the primordial @xmath2li abundance and solar neutrino fluxes , and as the precision of other reaction rate determinations relevant for bbn and solar hydrogen burning
has been dramatically increased , the need for a more precise @xmath0he(@xmath1)@xmath2be reaction rate measurement became pressing .
realizing the need for new , high precision cross section data , b. s. nara singh _ et al .
_ has performed an activation experiment for the @xmath0he(@xmath1)@xmath2be reaction @xcite .
the cross section was measured at four energy points between 420 and 950 kev .
although these results were the lowest energy activation points ever measured , they were still above the astrophysically relevant energies . moreover , an activation - only experiment whatever precise it is is not able to examine the discrepancy between the on - line and activation data .
therefore , an experiment measuring the @xmath0he(@xmath1)@xmath2be @xmath4 factor simultaneously with both techniques is highly needed .
such a measurement should be carried out with high precision and at energies as low as possible in order to have a reliable zero energy extrapolation of the @xmath4 factor .
.[tab : results ] results of the luna activation experiments [ cols="<,^,^,^,^",options="header " , ] figure[fig : all_act_fit ] shows the measured @xmath4 factor data from the four experiments considered above .
the r - matrix calculation by descouvemont _
rescaled to the average of the luna and nara singh _ et al .
_ data is also plotted .
the perfect agreement of the two recent experiments and the systematically higher values obtained in the older measurements is clearly visible
. the next step should be the inclusion of the on - line cross section data in the @xmath13 analysis .
the high precision luna on - line data are already available @xcite but the analysis of these data with the comparison of other on - line experiments and the activation results is beyond the scope of the present paper .
this work was supported by infn and in part by the european union ( tari rii3-ct-2004 - 506222 ) , the hungarian scientific research fund ( k68801 and t49245 ) , and the german federal ministry of education and research ( 05cl1pc1 - 1 ) .
10 aharmim b__et al . _ _ 2005 phys . rev .
c * 72 * 055502 hosaka j__et al .
_ _ 2006 phys . rev .
d * 73 * 112001 alimonti g _
_ 2002 astropart .
* 16 * 205 ahmad q r _ et al .
_ 2001 phys .
lett . * 87 * 71301 eguchi k _ et al .
_ 2003 phys .
lett . * 90 * 21802 adelberger e _
_ 1998 rev .
mod . phys . * 70 * 1265 nara singh b s , hass m , nir - el y and haquin g 2004 phys .
lett . * 93 * 262503 bahcall j n and pinsonneault m h 2004 phys .
* 92 * 121301 spergel d _ et al .
_ 2003 astrophys .
148 * 175 holmgren h
d and johnston r l 1959 phys . rev . * 113 * 1556 parker p d and kavanagh r w 1963 phys . rev . *
131 * 2578 nagatani k , dwarakanath m r and ashery d 1969 nucl .
a * 128 * 325 krwinkel h _
_ 1982 z. phys .
a * 304 * 307 osborne j l _ et al . _
1982 phys .
lett . * 48 * 1664 alexander t k , ball g c , lennard w n , geissel h and mak h b 1984 nucl . phys .
a * 427 * 526 hilgemeier m , becker h w , rolfs c , trautvetter h p and hammer j w 1988 z. phys .
a * 329 * 243 robertson r g h _ et al . _ 1983 phys .
c * 27 * 11 volk h , krwinkel h , santo r and wallek l 1983 z. phys . a * 310 * 9l bemmerer d _ et al .
_ 2006 phys .
. lett . * * 9**7 122502 gyrky gy _
_ 2007 phys .
c * 75 * 035805 confortola f _ et al .
_ 2007 phys .
c * 75 * 065803 kajino t and arima a 1984 phys .
lett . * 52 * 739 cst a and langanke k 2000 few - body systems * 29 * 121 descouvemont p _
et al . _ 2004 atomic data and nuclear data tables * 88 * 203 kajino t , toki h , austin s m 1987 astrophys .
j. * 319 * 531 angulo c _ et al .
_ 1999 nucl .
a * 656 * 3 | recently , the luna collaboration has carried out a high precision measurement on the @xmath0he(@xmath1)@xmath2be reaction cross section with both activation and on - line @xmath3-detection methods at unprecedented low energies . in this paper the results obtained with the activation method are summarized .
the results are compared with previous activation experiments and the zero energy extrapolated astrophysical @xmath4 factor is determined using different theoretical models . |
the top quark has been discovered only recently @xcite due to its very large mass : @xmath1 .
indeed , the top quark is easily the heaviest particle in the standard model ( sm ) .
this peculiar property brings several reasons to study it .
first , the yukawa coupling of the top quark is nearly one which could be a sign that it plays a special role in the origin of mass beyond the standard model .
second , radiative corrections due to top quark loops are often dominant in the prediction of precision observables like the @xmath0 or higgs boson mass .
finally , its large mass implies it has a very short lifetime ( @xmath2s ) , about an order of magnitude smaller than the hadronization time , which provides a unique opportunity to study a bare quark .
the top quarks are produced by the tevatron , a @xmath3 collider operating at a center - of - mass energy of 1.96 tev for the current period of data - taking ( run ii ) .
they are produced predominantly in pairs ( @xmath4 ) via the strong interaction with a predicted cross - section of @xmath5 pb @xcite .
because the @xmath6 ckm matrix element is nearly one under the 3 families sm assumption , the top quarks are predicted to decay @xmath7 99.8% of the time to a real @xmath0 boson and a @xmath8-quark . the two resulting @xmath0 bosons in turn decay either hadronically or leptonically , defining the three channels of @xmath4 events ( with branching ratios in parenthesis ) : `` all - hadronic '' for two hadronic decays ( 46% ) , `` lepton+jets '' for one leptonic and hadronic decays ( 29% ) and `` dilepton '' for two leptonic decays ( 5% ) . by `` leptons ''
we refer only to electrons and muons , taus being generally too difficult to identify to be usually considered in top quark properties measurements .
the @xmath4 events reconstruction are performed by the cdf and d detectors that are described in detail elsewhere @xcite .
typical event selections include identification of isolated leptons with high transverse momentum ( @xmath9 ) , large transverse missing energy ( @xmath10 ) due to the undetected neutrino(s ) from @xmath0 boson decays , several high-@xmath9 jets and identification of @xmath8-jets ( @xmath8-tagging ) generally using secondary vertex tagging .
the integrated luminosity ( @xmath11 ) used for the measurements presented here vary from @xmath12 200350 pb@xmath13 .
the current information on @xmath14 comes from measurements performed during run i of the tevatron ( with a luminosity of @xmath12 100 pb@xmath13 ) .
the world average is @xmath15 @xcite .
the final goal in run ii is to collect between 4 - 8 fb@xmath13 and reduce the uncertainty on @xmath14 to @xmath12 @xmath16 . a precise measurement of @xmath14 is strongly motivated in the sm because the radiative corrections to many precision observables are dominated by top quark loops .
a famous example is the indirect constraint on the higgs boson mass ( @xmath17 ) .
the current constraint is @xmath18 @xcite .
more detailed discussions on the precision electroweak observables can be found in these proceedings ( see contribution from p. renton ) . before describing specific analyses ,
we first discuss general considerations .
one challenge of the @xmath14 measurement is to solve the combinatorics problem in the event reconstruction .
for example there are four jets in the final state in the lepton+jets channel ( @xmath19 ) , resulting in 12 possible jet - parton assignments ( since the two @xmath0 daughter jets are interchangeable in the reconstruction ) .
the problem is simplified by tagging @xmath8-jets as described above .
another limitation of the measurement is the large uncertainty in the modeling of the jet energy response , referred to as the jet energy scale uncertainty ( jes ) .
this is generally the dominant systematic uncertainty , and is currently about @xmath20 for cdf and 5@xmath21 for d .
cdf recently reported the most precise measurement of @xmath14 to date .
it has been performed in the lepton+jets channel using @xmath11 = 318 pb@xmath13 of data .
one novelty of this analysis is the usage of the hadronic @xmath0 boson decays ( @xmath22 ) to improve the jet energy scale uncertainty .
templates of reconstructed top quark and @xmath0 boson mass are created in mc events as a function of the true top quark mass and the jet energy scale and compared with the data .
the fit to the data yields @xmath23 where both the statistical and jes uncertainties are included .
the systematic uncertainties not including the jes are small ( @xmath24 ) .
the final result is @xmath25 .
figure [ f : fitconsistencymreco ] shows the reconstructed top quark mass in the data with the best fits from the mc templates overlaid .
analyses using the template method in the lepton+jets channel have also been performed by the d collaboration .
the best measurement so far corresponds to a luminosity of 230 pb@xmath13 and yields @xmath26 .
a different class of analyses uses a matrix - element method that consists of computing a probability for each event to be signal with a given top quark mass .
the probability is computed using the full sm production and decay matrix - elements .
this method has been shown to be very powerful statistically , as demonstrated by the best measurement in run i @xcite .
the cdf collaboration has recently reported a measurement using this method : @xmath27 .
the result is in very good agreement with the template analysis above .
a similar measurement is expected to be released by the d collaboration very soon .
the top quark mass can also be measured in the dilepton channel but with larger statistical uncertainties due to the smaller branching ratio and the two neutrinos in the final state that complicate the event reconstruction .
the cdf collaboration has recently reported the best measurement in this channel to date using the matrix - element method : @xmath28 ( 340 pb@xmath13 ) .
the best measurement at d so far is : @xmath29 ( 230 pb@xmath13 ) .
it is very important to study in detail the properties of the top quark since it might be more closely related to new physics than other sm particles due to its large mass .
for instance , one can study the polarization of the @xmath0 boson inside the top decays . in the sm , 70% , 30% and 0% of @xmath0 bosons
are predicted to have a longitudinal , left - handed and right - handed helicity , respectively .
the @xmath0 helicity is sensitive to the angle between the charged lepton in @xmath0 rest frame and the @xmath8-jet angle , denoted as @xmath30 .
the d collaboration measured the fraction of right - handed @xmath0 bosons in top decays : @xmath31 at 95% confidence level ( c.l . )
figure [ f_whel_d0 ] shows the distribution @xmath30 in data .
the @xmath0 helicity is also sensitive to the charged lepton @xmath9 spectrum from @xmath0 boson decays that has been used at cdf to measure the fraction of longitudinal @xmath0 bosons : @xmath32 .
the assumption that the top quark decays nearly 100% of the time to @xmath33 is checked by measuring @xmath34 .
this is done by counting the number of @xmath8-tags in the sample .
such a measurement was performed at d and yielded @xmath35 at 95% c.l .. a consistent result has been obtained by cdf : @xmath36 at 95% c.l . @xcite .
other properties of the top quark will be measured in run ii as more data is accumulated like the electric charge , @xmath4 spin correlation and rare decays .
the precise determination of top quark properties is already well underway in run ii at the tevatron .
the best measurement of the top quark mass to date has been performed at cdf : @xmath25 .
a very competitive measurement is expected to be released by d very soon .
measurements of other properties of top quarks show a good agreement with the sm using only the small @xmath4 samples collected so far ( @xmath37 events per experiment ) .
we can expect these datasets to grow by a factor of ten or so by the end of run ii , which will greatly improve our knowledge of the peculiar particle that is the top quark .
we note that there was no time to present all analyses of top quark properties performed at the tevatron .
more details can be found on the public web pages of the cdf @xcite and d @xcite experiments . | we present recent analyses of top quark properties performed at run ii of the tevatron .
measurements of the top quark mass , branching ratios and @xmath0 boson helicity inside top quark decays are covered .
address = department of physics , university of toronto , 60 st .
george st . ,
toronto , m5s 1a7 , on , canada |
since more than one hundred years , photoelectron spectroscopy plays an important role in physics .
einstein s explanation of the photo - effect , probably the most well - known experiment in the field , was a crucial step in the development of quantum mechanics . whereas the photo - effect revolutionized the understanding of light , the main aim of modern photoelectron spectroscopy is to understand electronic and ionic structures from solids down to single atoms . especially in the context of nanoscale materials
photoelectron spectroscopy is one of the most important experimental tools since it is almost the only method that provides access to the electronic and ionic structures of these materials .
the direct observation of the electronic shell structure in sodium clusters @xcite is just one example for the power of the method .
another application is the determination of the ionic structure of , e.g. , clusters . since the electronic structure , and thus the photoelectron spectrum ( pes ) , depends on the ionic configuration , comparing the measured pes with the results from first - principle calculations allows the identification of the ionic structure .
this interplay between theory and experiment has already been used successfully in many cases @xcite .
clearly , the just mentioned method can only work if reliable calculations for the system of interest can be performed .
since most of the measured systems consist of many electrons , density - functional theory ( dft ) is an especially well - suited tool due to its low numerical costs .
unfortunately , evaluating the pes from a kohn - sham ( ks ) dft calculation is not an easy task since only the highest occupied ks eigenvalue has a rigorous connection to the pes : it is equal to the ionization potential @xcite
. thus , it yields the position of the first peak in the pes .
the most common approach to obtain the other peaks in the pes from a dft calculation is based on the density of states of the occupied ks orbitals . in this approach the ks eigenvalue spectrum of the ` mother ' system , i.e.
, the system which still contains the photoelectron , is directly compared to the experimental pes . in many situations
the resulting spectra reproduce the experimental ones quite well @xcite .
another way to extract the information related to the pes from a dft calculation is via the excitation energies of the ` daughter ' system , i.e. , the system with one electron less . since time - dependent dft ( tddft ) @xcite , in principle , allows to calculate the excitation energies of a system exactly , tddft can be used to calculate the positions of the pes peaks accurately .
this approach is followed in ref . and
ref .. the basic idea of this approach is also used in ref . , but in combination with configuration - interaction and not tddft calculations .
we finally want to mention a third method how the pes can be obtained from a dft calculation . in this approach ,
the time - dependent ionization process is simulated in real - time and the kinetic energy spectrum of the outgoing components of the ks slater determinant are connected to the pes @xcite . in ref . ,
the pes resulting from na@xmath0 , na@xmath1 , and na@xmath2 ( among others ) irradiated by an xecl excimer laser ( @xmath3 ev ) were measured and compared with the corresponding ks eigenvalue spectra . although the agreement between the theoretical results and the measured pes was generally reasonable , a systematic discrepancy was found .
namely , the width of the theoretical spectrum , i.e. , the difference between the energy of the energetically highest and lowest occupied ks eigenvalue , was too large by about 0.2 - 0.4 ev . in ref .
the reason for this discrepancy was examined .
it was shown that technical aspects , e.g. , the treatment of the pseudopotential , could not explain the differences . furthermore , it was demonstrated that using the exchange - only optimized - effective potential ( oep ) @xcite reduced the width of the ks spectrum but not to an extent that would bring the spectrum in agreement with experiment .
in addition , it was also shown that using slater s transition state concept could also not improve the theoretical results .
thus , the question arises whether a different method to extract the pes from a dft calculation leads to a better agreement with the experiment .
it is the aim of the present manuscript to answer this question by extracting the pes from the excitation energies of the corresponding ` daughter ' systems .
before discussing the results let us sketch the theoretical background of the presented calculations in more detail . -electron system .
right : the photoelectron has already been detected and the remaining @xmath4-electron system is left in an excited state . the link between the kinetic energy of the photoelectron and the energy of the excited state of the @xmath4-electron system
is provided by energy conservation .
[ fig : pesschematic],width=306 ] fig .
[ fig : pesschematic ] schematically shows two approaches how the peak positions in the pes can be calculated . on the left hand side ,
the process is described as an excitation process from the ground state to an energetically high - lying state with continuum contributions .
since ks eigenvalue differences are zeroth - order approximations to excitation energies @xcite , the ks dos of the @xmath5-electron system can be used to obtain an approximate pes .
in addition to this argument , chong _ et al . _ @xcite
have given well founded arguments that ks eigenvalues can be interpreted as approximations to relaxed vertical ionization potentials . on the right hand side of fig .
[ fig : pesschematic ] , the situation after the photoelectron has been detected is considered . in this case , the remaining system is left in an energetically low - lying excited state of the @xmath4-electron system . to connect the excitation energies of this system to the pes ,
energy conservation is used .
before the photon is absorbed the total energy is given by @xmath6 , where @xmath7 is the ground - state energy of the ` mother ' system containing @xmath8 electrons and @xmath9 is the photon energy . after the detection of the photoelectron
the total energy is given by the kinetic energy of the photoelectron @xmath10 and the energy of the remaining ` daughter ' system .
since the total energy is conserved , it follows that @xmath11 must hold . here
, @xmath12 is the ground - state energy of the ` daughter ' system and @xmath13 are its excitation energies @xcite . for the first peak in the pes
the kinetic energy of the photoelectron is maximal . in this case , the ` daughter ' system is in its ground - state , i.e. , @xmath13 is zero and the peak position is at @xmath14 . to obtain the excitation energies from time - dependent dft
the full linear density - response function of the interacting system can be used .
this function provides access to the excitation energies of the system since it has poles at these energies .
the crucial observation now is that the interacting linear density - response function can be expressed in terms of the ks response function and the exchange - correlation ( xc ) kernel @xcite .
nowadays , most applications use the matrix equation of casida @xcite to obtain these excitation energies .
alternatively , the excitation energies can be extracted from a spectral analysis of the time - dependent density coming from a real - time propagation @xcite . in this approach
, the xc kernel is not needed , but instead , the time - dependent ks equations are solved without explicit linearization . to illustrate this approach , imagine we have created a time - dependent density @xmath15 of an interacting system by , e.g. , a laser excitation . assuming that the system is confined by the same time - independent potential before and after the laser pulse we can write the excited density in terms of the eigenstates @xmath16 of the interacting system in the time - independent potential .
it reads @xmath17 where @xmath18 is the eigenvalue corresponding to @xmath16 and @xmath19 is the density operator . assuming that the time - dependent state @xmath20 is dominated by the ground state , i.e. , @xmath21 we can write @xmath22 here
, @xmath23 is the ground - state density of the system .
if we now calculate the fourier transform of @xmath15 we will get peaks at the exact excitation energies of the system . since time - dependent dft in principle provides us with the exact time - dependent density , this is an easy method to obtain the excitation energies of the interacting system from a time - dependent dft calculation . in a practical calculation
, two problems must be solved to get the excitation energies from this scheme .
first , one has to create a time - dependent density which is dominated by the ground - state density and , in addition , contains the excited states of interest .
the second problem is how to extract the excitation energies from the time - dependent density in practice . since the density in every space point at all times
can not be stored , a full fourier transform of eq .
( [ eq : density ] ) giving @xmath24 is not possible . to overcome this problem
several possibilities exist .
one is to evaluate @xmath24 only for some points in space @xcite , e.g. , in the center of the cluster .
a different method is to fourier transform certain moments of the density distribution .
typically , the dipole moment is used for this purpose @xcite . obviously , some excitations are filtered out by this procedure because the fourier spectrum of the dipole moment only shows excitation energies of states which can be coupled to the ground state via the dipole operator .
non - dipole active excitations can be taken into account by recording higher moments . in the following
we will use the time - dependent dipole and quadrupole moments obtained from a real - time propagation to obtain excitation energies for the systems of interest .
for the ionic ground - state configurations of the ` mother ' systems we used optimized structures obtained with the parsec @xcite program package
. the generalized - gradient approximation of perdew _ et al . _ ( pbe ) @xcite was employed for the geometry optimization .
the ionic cores were treated consistently with norm conserving non - local pseudopotentials @xcite .
the time - dependent ks equations were solved on a real - space grid in real time . for this
we implemented the necessary algorithms into a modified version of the parsec code . in detail
, we implemented a fourth - order taylor approximation to the propagator in combination with a higher - order finite - difference formula for the kinetic part of the ks hamiltonian .
for the calculations , we used a time step of 0.003 fs and the total propagation time was 75 fs .
the ionic cores were again described by norm conserving non - local pseudopotentials .
furthermore , the ionic structures were fixed during the propagation . the grid spacing was 0.7 @xmath25 and the grid radius varied between 20 and 23 @xmath25 depending on the system .
the time - dependent density was created by applying a boost @xmath26 to the ground - state ks orbitals .
the total excitation energy of the system was @xmath27 ev , i.e. , a boost strength @xmath28 was applied to each ks orbital ( with @xmath29 being the electron mass and @xmath4 the number of electrons ) .
in addition , the calculations were repeated with a boost strength reduced by a factor of @xmath30 .
using these two small boost strengths allowed us to check whether the created time - dependent density was dominated by the ground - state density ( see below ) .
instead of applying the same boost vector @xmath31 to all ks orbitals , and thus creating a coherent velocity field , we varied the boost direction for different ks orbitals .
this is necessary since applying the same boost direction to all ks orbitals corresponds to first order in @xmath31 to a dipole excitation of the system , i.e. , from the resulting time - dependent density it is only possible to retrieve the excitation energies of ` dipole - active ' states . by applying different boost directions to different ks orbitals we modeled a general excitation mechanism creating a time - dependent density containing excited states with different symmetry properties . in detail
, we randomly chose a boost direction ( no symmetry axis of the considered cluster ) for the first orbital and then chose our coordinate system such that this direction was the first diagonal ( for the remaining rotational degree of freedom a random angle was chosen ) .
after this we boosted the second orbital in the opposite direction of the first boost .
the third orbital was then boosted in the direction of the second diagonal of the chosen coordinate system , the forth again in the opposite direction and so on . for na@xmath32 ,
the ninth orbital was boosted again in the same direction as the first orbital .
since the only purpose of this procedure was to create a time - dependent density without any particular symmetries , we do not consider the relative orientation of the cluster with respect to the boost directions to be of special importance .
finally , we used the time - dependent local - density approximation ( tdlda ) @xcite for the xc potential for the propagation .
since the linear response of the homogeneous electron gas is the same in this approximation and in the pbe functional , the differences in the resulting excitation energies can be expected to be small in the low - energy regime .
unfortunately , the exact - exchange orbital functional can not be used for comparative calculations since it requires a solution of the time - dependent optimized - effective potential equation @xcite and no method for this exists in real - time at the moment @xcite .
[ fig : na3dipspec ] shows the dipole power spectra of na@xmath34 resulting from two boost strengths differing by a factor of @xmath35 .
the dipole power spectrum is given by @xmath36 with @xmath37 being the fourier transform of the dipole moment @xmath38 where @xmath39 corresponds to the cartesian coordinate @xmath40 , @xmath41 to @xmath42 , and @xmath43 to @xmath44 .
resulting from an incoherent boost excitation .
the result obtained from a total excitation of @xmath45 ev is labeled ` strong excit . ' whereas the label ` weak excit . '
corresponds to a boost reduced by a factor of @xmath46 .
clearly , the dipole power spectrum scales quadratically with the boost strength indicating that the peak positions correspond to excitation energies between the ground state and excited states .
[ fig : na3dipspec],width=306 ] for small momentum boosts first - order perturbation theory predicts a linear dependence of the expansion coefficients @xmath47 in eq .
( [ eq : densexp ] ) on the boost strength . as a consequence , reducing the boost strength by a factor of @xmath48 suppresses peaks corresponding to energy differences between two excited eigenstates by a factor of @xmath49 in the power spectrum . since peaks corresponding to transitions between the ground state and an excited eigenstate
are only suppressed by a factor of @xmath50 , changing the boost strength allows one to distinguish between these two kinds of excitations .
as one can see in fig .
[ fig : na3dipspec ] , the results for the two boost strengths are almost identical except for the predicted factor of @xmath51 .
thus , we conclude that all the peak positions in the dipole power spectrum of fig .
[ fig : na3dipspec ] correspond to energy differences between the ground state energy and the energy eigenvalues of the excited eigenstates .
the situation is different for the power spectrum resulting from the quadrupole moments . in fig .
[ fig : na3quadspec ] we plot .
in contrast to the dipole power spectrum some peaks vanish with reduced boost strength indicating that they correspond to energy differences between excited states . [
fig : na3quadspec],width=306 ] @xmath52 for the same two excitation boosts . in this equation
, @xmath53 is the fourier transform of the quadrupole moment @xmath54 and the sum only runs over the independent components of the quadrupole tensor .
clearly , the quadrupole spectra for the different excitation strengths differ considerably .
for instance , the three large peaks at around @xmath55 , @xmath56 , and @xmath57 ev vanish almost completely .
thus , we conclude that they belong to transition energies between different excited states .
indeed , one can see that these energies are exactly equal to the energy differences between the first excited state and the other excited states from the dipole spectrum .
the reason why there are no peaks at these energies in the dipole spectrum can easily be understood if one takes the geometry of na@xmath34 into account . since na@xmath34 has a linear ionic configuration , the ground state has even parity .
thus , the dipole spectrum only shows excited states with odd parity . since
two states with odd parity can not be coupled by the dipole operator , transitions between these states do not show up in the dipole spectrum .
after the identification of the true excitation energies we can now compare the results with the measured pes . in fig .
[ fig:1na3exp ] the excitation energies of na@xmath34 , the ks dos of na@xmath33 and the measured pes ( of na@xmath33 ) are plotted .
the positions of the occupied ks eigenvalues are indicated by arrows , long bars indicate excitation energies from the dipole spectrum and shorter bars excitation energies from the quadrupole moments . in addition , excitation energies leading to peaks below the strongest bound experimental peak are reduced in their overall height . for better comparison , the ks dos and the excitation spectrum are both rigidly shifted in such a way that the most weakly bound peak coincides with the experimental one .
( ` exp . ' ) and theoretical pes obtained from the excitation energies of na@xmath34 .
excitation energies from the dipole spectrum are labeled ` dip . excit . ' whereas ` quad . excit . '
labels excitation energies deduced from the quadrupole moments .
arrows indicate the result obtained from the ks dos .
upper part : results obtained from the ionic ground - state configuration at zero temperature .
lower part : results obtained from an ionic configuration with a larger bond length to simulate a higher temperature . for most peaks
the agreement with the experimental pes is clearly improved .
[ fig:1na3exp],title="fig:",width=306 ] ( ` exp . ' ) and theoretical pes obtained from the excitation energies of na@xmath34 .
excitation energies from the dipole spectrum are labeled ` dip . excit . '
whereas ` quad . excit . ' labels excitation energies deduced from the quadrupole moments .
arrows indicate the result obtained from the ks dos .
upper part : results obtained from the ionic ground - state configuration at zero temperature .
lower part : results obtained from an ionic configuration with a larger bond length to simulate a higher temperature . for most peaks the agreement with the experimental pes is clearly improved .
[ fig:1na3exp],title="fig:",width=306 ] as the upper part of fig .
[ fig:1na3exp ] shows the peak positions that one obtains from the ks dos are close to the experimental peak positions . unlike in the case of larger na clusters ,
the width is slightly smaller than the energy difference between the two large experimental peaks but it is still reasonable .
however , since there are only two occupied ks orbitals in na@xmath33 the ks dos picture fails completely to describe the higher lying peaks in the measured spectrum . as one expects from eq .
( [ eq : econst ] ) the pes obtained from the excitation energies shows a much richer structure than the ks dos .
one striking feature for instance is the second excitation around @xmath58 ev .
it seems that the energy difference between this peak and the one at @xmath59 ev is too large in the calculation and that they are merged to one peak in the experimental pes .
however , in general , the dynamically calculated excitation energies and the energies obtained from the experimental pes are close to each other even for the stronger bound peaks . to see
if the remaining discrepancy can be further reduced by taking temperature effects into account we have repeated our calculations with a larger bond length . due to the net negative charge of the cluster
one can expect that other geometry changes , e.g. , bending , only play a minor role in the case of na@xmath33 .
we have used a new bond length of approx . @xmath60 instead of @xmath61 @xmath25 .
this new value for the bond length @xmath62 of the cluster has been obtained from an estimate for the thermal expansion at @xmath63 k. it is based on the formula @xmath64 for the linear thermal expansion coefficient @xmath65 which we have roughly estimated by @xmath66 ( see ref . ) , where @xmath67 is the bulk value for crystalline sodium at room temperature .
the result can be seen in the lower part of fig .
[ fig:1na3exp ] . for most peaks
one can observe a small shift towards lower absolute binding energies .
except for the peak at @xmath59 ev the agreement between the experimental and theoretical spectrum is slightly improved by the increased bond length .
especially the broader peak at around @xmath68 ev is nicely reproduced in this case .
all in all , both calculations show that for na@xmath69 the main advantage of the ` excitation picture ' is the reproduction of the deeper bound structures in the pes .
[ fig:2na5exp ] shows the experimental pes of na@xmath0 , the ks dos , and the pes obtained from the excitation energies of na@xmath70 .
the labeling is the same as in the corresponding previous figures . as for na@xmath33
, the ks dos is in acceptable agreement with the first large peaks , although the strongest bound large peak has a too negative binding energy in the ks dos .
as one can see these peaks are also well described by the excitation energies of the ` daughter ' system with the additional advantage that the last peak at @xmath71 ev is better reproduced . in this approach
it consists of four close - lying excitations . but for na@xmath0 .
although both , the ks dos and the pes from the excitation energies , describe the measured pes acceptable the large peak at @xmath71 ev is much better reproduced by the excitation energies from the ` daughter ' system .
the experimental data is taken from ref . .
[ fig:2na5exp],width=306 ] beyond the peak at @xmath71 ev the comparison with the experimental measurement is difficult since no clear peak structures can be observed .
perhaps the accumulation of excited states around @xmath72 ev can be associated with the measured peak in this region , but for the reasons given below , one has to be very cautious in making comparisons in this part of the spectrum .
as one can see from the results for na@xmath1 and na@xmath2 discussed below the problem of comparing the deeper lying part of the measured pes with calculated excitation energies is not specific to na@xmath0 .
in general , the density of excited states grows with the excitation energy , i.e. , more and more states appear in the theoretical calculation .
on the other hand , from the point of view of first - order perturbation theory the pes depends not only on the positions of the excited states but also on the matrix element of the perturbing operator @xmath73 between the initial and the final state . taking the ground state of the ` mother ' system for the initial state and a product state consisting of one photoelectron with momentum @xmath74 for the final state one obtains matrix elements of the form @xmath75 .
it is intuitively clear that these matrix elements are much larger for low - lying states than for energetically high - lying ones which in an independent - particle picture would correspond to removing one particle and exciting a second one above the fermi level .
especially , in the case of truly independent particles this process can not happen if the perturbing operator is a one - particle operator like the dipole operator .
thus , many energetically high - lying eigenstates of the ` daughter ' system are hardly or even not at all excited in the experiment .
since the mentioned matrix elements depend on the interacting many - particle wavefunctions , calculating these exactly is close to being impossible .
especially , retrieving these matrix elements from a tddft propagation of the ` daughter ' system is not trivial because the propagation only provides information about matrix elements between excited states and the @xmath4-particle ground state and not the @xmath5-particle ground state .
however , as the presented calculations show , the matrix elements do not play a very important role in the part of the spectrum that we are mainly interested in .
nevertheless , the calculations also clearly indicate that one has to consider them if the deeper lying parts of the spectrum are of interest . a possible method how this can be done in a tddft calculation
can be found at the end of sec .
[ sec : sumcon ] .
the results for na@xmath1 are shown in fig .
[ fig:3na7exp ] . as said previously , in the region below @xmath76 ev
, it is difficult to compare theory and experiment due to the great number of close lying transitions . as in na@xmath0
, the ks dos describes the strongest bound large peak worst . in this case
it is already off by @xmath77 ev .
in contrast , the peak position obtained from the tdlda excitation energies is considerably closer to the experimental peak .
it is only off by @xmath78 ev .
thus , the overestimation of the spectrum s width by the ks dos @xcite is not observed in the result obtained from the tdlda calculation .
the remaining difference of @xmath78 ev between the width of the theoretical and the experimental spectrum can be easily caused by technical aspects like the employed pseudopotential and xc potential @xcite .
in addition , thermal effects like bond elongation and structural isomerization can shift the obtained width by @xmath78 ev @xcite . considering that the experimental pes were obtained from clusters with a temperature of around 250 - 300 k ,
the difference between the theoretical result at zero temperature and the experimental result is hardly surprising . at these temperatures
the larger anionic sodium clusters behave liquidlike @xcite . consequently , many different ionic configurations are present in the experiment and show up in the measured pes . but for na@xmath1 .
especially , the peak at @xmath79 ev is much more accurately described by the excitation energies than by the ks dos . in the weakly bound region
thermal effects play an significant role in the case of na@xmath1 .
this explains the rather poor agreement between the theoretical values calculated at zero temperature and the measured curve between @xmath80 and @xmath59 ev . the experimental data is taken from ref . .
[ fig:3na7exp],width=306 ] this aspect must also been kept in mind if the theoretical and experimental results are compared in the region between @xmath80 and @xmath59 ev . in this region
both the zero temperature ks dos result and the zero temperature result from the excitation energies do not describe the measured pes very accurately .
especially , the excitation peak at @xmath81 ev does not fit very well .
however , from ref . it is known that the agreement between the experimental and the ks dos result in this energy region is significantly improved if different ionic structures are taken into account via born - oppenheimer langevin molecular - dynamics @xcite .
therefore , one can expect that the agreement between the experimental and the tdlda result is also improved if different ionic structures are taken into account . due to the more complicated structures and
the growing number of isomers the inclusion of the temperature influence on the ionic structures of larger clusters is much more involved than in the case of na@xmath33 .
additionally , combining born - oppenheimer langevin molecular - dynamics with the calculation of excitation energies is substantially more expensive than combining such a molecular dynamics scheme with a ks dos calculation .
thus , including thermal effects in the present study is beyond the scope of the present work and is a future project .
finally , the theoretical results for na@xmath2 are compared in fig.[fig:4na9exp ] with the measured pes .
this cluster is the first one which has a clear peak in the range between the highest and lowest occupied ks eigenvalue which is completely absent in the ks dos , i.e. , the experimental pes shows six clear peaks whereas the ks dos consists of only five peaks : the peak around @xmath79 ev is completely missing in the ks dos .
in addition , the strongest bound peak in the ks dos is off by @xmath82 ev . in other words ,
the ks dos result is inaccurate to the extent of being useless below @xmath71 ev . as for na@xmath1
, the splitting of the large peak around @xmath83 ev is reproduced if different ionic structures are used @xcite .
in contrast to the ks dos result , the pes obtained from the excitation energies is close to the measured curve over the whole range . below the lowest lying peak at @xmath84 ev
the comparison is again difficult without knowing the matrix elements mentioned above .
for the two peaks at @xmath84 and @xmath79 ev the theoretical values are off by @xmath78 ev .
especially since na@xmath2 , in contrast to na@xmath1 , is not a closed - shell cluster , one can expect that such energy differences can be easily caused by ionic structure modifications induced by finite temperatures . as expected from ref .
, the splitting of the peak at @xmath83 ev is also not reproduced by the zero temperature tdlda calculation .
all in all , the experimental result in the weaker bound part of the spectrum is described equally well by the ks dos and the excitation energies of the ` daughter ' system . however , in the stronger bound part the time - dependent calculation yields a much more realistic description of the pes than the ks dos .
since this emerges as a general observation for all systems studied in this manuscript , we discuss it on general grounds in the following section .
but for na@xmath2 .
as in na@xmath1 , especially the stronger bound part of the spectrum is described more accurately by the calculated excitation energies than by the ks dos .
the experimental data is taken from ref . .
[ fig:4na9exp],width=306 ]
using tddft we have calculated the excitation energies of small neutral sodium clusters .
the energies of the excited states were retrieved from the dipole and quadrupole moments of the time - dependent density via spectral analysis .
the time - dependent density was created by an incoherent boost of all ks ground - state orbitals and then propagated in real - space and real - time . to discriminate between true excited state energies and energies corresponding to energy differences between excited states we did two calculations for all systems using different excitation energies . for comparison with measured pes of the anionic clusters
the excitation energies were calculated in the ionic configuration of the anions . in general ,
the pes for all clusters studied in this manuscript can be divided into three parts .
the first part consists of large , ` weakly ' bound peaks , the second of large , ` strongly ' bound peaks , and finally a ` less structured ' region below the lowest large experimental peak . except for na@xmath33 no comparisons between our theoretical and
the experimental results can be made in the third region .
as discussed in subsection [ subsecna5 ] the main reason for this is the missing access to the transition matrix elements between the ground and the excited states . since the number of excited states can grow very rapidly
, one can expect that the omission of the transition matrix elements can cause severe problems if more complex systems are examined . a possible way to overcome
this problem is by including the information of the matrix elements in the initial density , i.e. , by creating an initial density that only includes the states which are really excited in the ionization process .
work in this direction is under way . in the middle part of the spectrum the results obtained from the tdlda excitation energies are clearly superior to the results from the ks dos .
especially , the position of the strongest bound large peak is much better reproduced by the tdlda calculation than by the ks dos .
thus , using the tdlda cures the main problem that plagues theoretical results obtained from the ks dos for sodium clusters , namely the prediction of a significantly too large width of the spectrum .
in addition , the pes from the tdlda excitations can describe an experimental peak in the pes of na@xmath2 which is completely missing in the ks dos .
the remaining differences between the experimental and our theoretical results are all small enough to be explainable by technical details or the finite temperature ( 250 - 300 k ) of the ionic structures in the experiment .
in particular the finite temperature can be expected to be responsible for the difference since the considered clusters behave liquidlike at this temperature and thus , the measured pes result from many different ionic structures / isomers which differ from the theoretical zero temperature ground - state structures used for the calculations . finally , in the most weakly bound part of the spectrum we find that the tdlda result and the one from the ks dos are very similar . since the ks dos at finite temperature is in very good agreement with the experimental result @xcite , it is extremely likely that also the tdlda excitation energies calculated from higher temperature ionic structures will describe the experimental pes very well in this region .
generally , our findings are in line with earlier results @xcite which report a worse agreement between the ks dos results and the experimental values for stronger bound levels . in addition
, our results clearly show that the agreement between the theoretical and the experimental spectrum is considerably improved for small sodium clusters if the pes is extracted from the true excitation energies of the ` daughter ' system and not the ks dos .
this shows the importance of taking effects beyond the independent - particle picture into account in the interpretation of photoelectron spectra .
this work is supported by the deutscher akademischer austauschdienst in the ppp germany - finland .
we acknowledge stimulated discussions with m. walter and h. hkkinen which became possible through this support .
we thank b. v.issendorff for providing us the raw data of the measured photoelectron spectra .
s. k. also acknowledges financial support from the deutsche forschungsgemeinschaft .
ionic relaxation processes are neglected here , i.e. , we assume that the ions do not move during the time interval between the photon absorption and the photoelectron detection . as a consequence , the excitation energies of the ` daughter ' system are calculated in the ionic configuration of the ` mother ' system . | we calculate the excitation energies of small neutral sodium clusters in the framework of time - dependent density - functional theory . in the presented calculations , we extract these energies from the power spectra of the dipole and quadrupole signals that result from a real - time and real - space propagation . for comparison with measured photoelectron spectra
, we use the ionic configurations of the corresponding single - charged anions .
our calculations clearly improve upon earlier results for photoelectron spectra obtained from static kohn - sham eigenvalues . |
the main motivation of this paper is the discrete - time approximation of the following system of reflected backward stochastic differential equations ( bsdes ) @xmath3 \,{\mathrm{d}}k ^i_t=0 , & \quad$i \in { \mathcal{i}}$,}\ ] ] where @xmath4 , @xmath5 , @xmath6 and @xmath7 are lipschitz functions and @xmath8 is the solution of a forward stochastic differential equation ( sde ) .
these equations are linked to the solutions of _ optimal switching problems _ , arising , for example , in real option pricing . in the particular case where @xmath5 does not depend on @xmath9 ,
a first study of these equations was made by hamadne and jeanblanc @xcite .
they derive existence and uniqueness of solution to this problem in dimension 2 .
the extension of this result to optimal switching problems in higher dimension is studied by djehiche , hamad ` ene and popier @xcite , carmona and ludkovski @xcite , porchet , touzi and warin @xcite or pham , ly vath and zhou @xcite for an infinite time horizon consideration . in this last paper ,
the resolution of optimal switching problems relies mostly on their link with systems of variational inequalities . considering deterministic costs , hu and tang @xcite derive existence and uniqueness of solution to this type of bsde and
relate it to optimal switching problems between one - dimensional bsdes .
extensions developed in @xcite and @xcite cover , in particular , the existence of a unique solution to the bsde ( [ eqbsdecorintro ] ) .
recently two of the authors related in @xcite the solution of ( [ eqbsdecorintro ] ) to corresponding constrained bsdes with jumps .
as presented in @xcite , this type of bsde can be numerically approximated combining a penalization procedure with the use of the backward scheme for bsdes with jumps ; see @xcite .
unfortunately , no convergence rate is available for this algorithm .
we present here a more natural discretization scheme based on a geometric approach . for any @xmath10 ,
all the components of the @xmath11 process are interconnected , so that the vector @xmath11 lies in a random closed convex set @xmath12 characterized by the cost functions @xmath7 .
the vector process @xmath13 is thus obliquely reflected on the boundaries of the domain @xmath14 and we approximate these continuous reflections numerically . as in @xcite , we first introduce a discretely reflected version of ( [ eqbsdecorintro ] ) , where the reflection occurs only on a deterministic grid @xmath15 , and , for @xmath16 and @xmath17 , @xmath18 where @xmath19 is the oblique projection operator on @xmath12 , for @xmath10 . extending the approach of hu and tang @xcite , we observe that the solution to ( [ bsdedorintro ] ) interprets as the value process of a one - dimensional optimal bsde switching problem with switching times belonging to @xmath20 .
this allows us to prove a key stability result for this equation .
we control the distance between @xmath21 and @xmath9 in terms of the mesh of the reflection grid . due to the obliqueness of the reflections , the direct argumentation of @xcite does not apply . using the reinterpretation in terms of switching bsdes
, we first prove that @xmath22 approaches @xmath13 on the grid points with a convergence rate of order @xmath23 , @xmath24 uniformly in @xmath25 , whenever the cost function is lipschitz and @xmath5 is bounded in @xmath26 ( see theorem [ thmerrordiscretize ] ) . imposing more regularity on the cost functions , we control the convergence rate of @xmath27 to @xmath28 ( see theorem [ thmerrordiscretize2 ] ) .
we then consider a euler type approximation scheme associated to the bsde ( [ bsdedorintro ] ) defined on @xmath29 by @xmath30 and , for @xmath31 , @xmath32 , \vspace*{2pt}\cr { { \widetilde y^{\re,\pi}}}_{{{t_i } } } : = \mathbb{e}[{y^{\re,\pi}}_{{t_{i+1}}}\mid\mathcal{f}_{{{t_i } } } ] + ( { { t_{i+1}}}- { { t_i } } ) f({x^{\pi}}_{{t_i}},{{\widetilde y^{\re,\pi}}}_{{t_i}},{{\bar z^{\re,\pi}}}_{{{t_i } } } ) , \vspace*{2pt}\cr { y^{\re,\pi}}_{{{t_i } } } : = { { \widetilde y^{\re,\pi}}}_{{t_i}}\mathbf{1 } _ { \ { { { t_i}}\notin\re\ } } + { \mathcal{p}}({x^{\pi}}_{{t_i } } , { { \widetilde y^{\re,\pi}}}_{{t_i}})\mathbf{1 } _ { \ { { { t_i}}\in\re\ } } , } \ ] ] where @xmath33 is the euler scheme associated to @xmath8 .
it is now well known ( see , e.g. , @xcite ) , that the convergence rate of the scheme ( [ schemeintro ] ) to the solution of ( [ bsdedorintro ] ) is controlled by the regularity of @xmath34 through the quantities @xmath35 \quad\mbox{and}\quad \mathbb{e}\biggl[\sum_{i < n } \int_{{{t_i}}}^{{{t_{i+1 } } } } |z^{\re}_{t } - \bar z^{\re}_{{{t_i}}}|^2 \,{\mathrm{d}}t\biggr]\ ] ] with @xmath36 $ ] , for @xmath37 . using classical malliavin differentiation tools , we prove a representation for @xmath38 , extending the results of @xcite to the system of discretely reflected bsdes ( [ bsdedorintro ] ) .
we deduce the expected regularity results on @xmath39 and , using the techniques of @xcite , chapter 3 , we obtain in a very general setting the convergence of ( [ schemeintro ] ) to ( [ bsdedorintro ] ) .
however , due to the obliqueness of the reflections , the projection operator @xmath40 is only @xmath41-lipschitz with @xmath42 , leading to a convergence rate controlled by @xmath43 is the number of points in the reflection grid @xmath20 .
the term @xmath44 can be very large even for small @xmath45 and leads to a poor logarithmic convergence rate when passing to the limit @xmath46 for the approximation of ( [ eqbsdecorintro ] ) . in the particular case where @xmath5 does not depend on @xmath26 , we are able to get rid of the @xmath47 term .
our innovative approach relies on the use of comparison results to get a control of the involved quantities : * we interpret the solution of ( [ bsdedorintro ] ) as a value process of an optimization problem , which allows us to get a control of the distance between the continuously and discretely reflected bsdes ; * we introduce a convenient auxiliary process dominating both solutions ( [ bsdedorintro ] ) and ( [ schemeintro ] ) to get a control of the distance between these quantities .
combining the previous estimates , we deduce the convergence of the discrete time scheme ( [ schemeintro ] ) to the solution of ( [ eqbsdecorintro ] ) with a convergence rate of order @xmath23 , @xmath24 , on the grid points , whenever @xmath48 and @xmath5 is independent of @xmath0 .
whenever the cost functions are constant , all the previous estimates hold true with @xmath49 .
we want to emphasize that all these results are obtained without any assumption on the nondegeneracy of the volatility matrix @xmath50 .
the rest of the paper is organized as follows . in section [ sedisc ]
, we introduce the notion of discretely obliquely reflected bsdes , connect it with optimal switching problems and give the fundamental stability result .
section [ secreg ] focuses on the regularity of the solution to this new type of bsde .
this analysis leads to precious estimates allowing us to deduce the convergence of the associated discrete time scheme ( see section [ subsecscheme ] ) .
afterward , section [ secontinuous ] focuses on the extension to the continuously reflected case and provides a convergence rate of the discretely reflected bsde to the continuously one , whenever the driver @xmath5 is bounded in the variable @xmath0 .
the global error of the scheme is provided at the end of this section .
some a priori estimates are reported in the . throughout this paper
we are given a finite time horizon @xmath51 and a probability space @xmath52 endowed with a @xmath53-dimensional standard brownian motion @xmath54 .
the filtration @xmath55 generated by the brownian motion is supposed to satisfy the usual conditions .
here , @xmath56 denotes the @xmath50-algebra on @xmath57\times\omega$ ] generated by @xmath58-progressively measurable processes .
any element @xmath59 with @xmath60 will be identified to a column vector with @xmath61th component @xmath62 and euclidean norm @xmath63 . for @xmath64
, @xmath65 denotes the scalar product of @xmath66 and @xmath67 , and @xmath68 denotes the transpose of @xmath66 .
we denote by @xmath69 the component by component partial ordering relation on vectors .
@xmath70 denotes the set of real matrices with @xmath71 lines and @xmath53 columns .
we denote by @xmath72 the set of functions from @xmath73 to @xmath74 with continuous and bounded derivatives up to order @xmath75 . for a function @xmath76
, @xmath77 denotes the jacobian matrix of @xmath5 with respect to @xmath66 .
for ease of notation , we will sometimes write @xmath78 $ ] instead of @xmath79 $ ] , @xmath80 $ ] . in the following , we shall use the notation without specifying the dimension nor the dependence in @xmath81 when it is clearly given by the context . finally , for any @xmath82
, we introduce the following : * the set @xmath83 of real - valued _ _ cdl ` ag _ _ @xmath56-measurable processes @xmath84 satisfying @xmath85 ^{1/p } < \infty$ ] . * the set @xmath86 of @xmath73-valued @xmath56-measurable processes @xmath87 such that @xmath88^{1/p } < \infty$ ] . * the closed subset @xmath89 of @xmath83 consisting of nondecreasing processes @xmath90 satisfying @xmath91 . + in the sequel
we denote by @xmath92 a constant whose value may change from line to line but which depends only on @xmath93 .
we use the notation @xmath94 whenever it depends on some other parameter @xmath95 .
in the beginning of this section we define and study discretely obliquely reflected bsdes in a general setting . in particular , we show how their solutions relate to the solutions of one - dimensional optimal switching problems , where the switching times are restricted to lie in a discrete time set .
this allows us to prove a stability result for obliquely rbsdes which will be used several times in the paper .
a discretely obliquely reflected bsde is a reflected bsde where the reflection is only allowed on a discrete time set .
we thus consider a grid @xmath96 of the time interval @xmath57 $ ] satisfying @xmath97 such that @xmath98 belongs to @xmath99 for @xmath100 and satisfies the structure condition @xmath101 we introduce a random closed convex set family associated to @xmath102 : @xmath103 and the oblique projection operator onto @xmath104 , denoted @xmath105 and defined by @xmath106 which is @xmath107-measurable .
[ repropp ] ( i ) it follows from the structure condition ( [ coststructureabs ] ) that @xmath108 is increasing with respect to the partial ordering relation @xmath69 , where @xmath109 means @xmath110 for all @xmath111 .
\(ii ) an easy calculation leads to @xmath112 we observe that the constant @xmath113 is optimal in our setting taking , for example , @xmath114 and @xmath115 .
thus @xmath116 is @xmath41-lipschitz continuous with @xmath117 .
finally , we are also given a random variable @xmath118^d$ ] valued in @xmath119 , representing the terminal value of the bsde and a random function @xmath120\times{\mathbb{r}}^d\times{\mathcal{m}}^{d , q}\rightarrow{\mathbb{r}}^d$ ] which is @xmath121-measurable and satisfies the lipschitz property @xmath122 for all @xmath123\times({\mathbb{r}}^d)^2\times({\mathcal{m}}^{d , q})^2 $ ] , @xmath124-a.s .
we shall also assume that the component @xmath61 of @xmath125 depends only on the component @xmath61 of the vector @xmath67 and on the row @xmath61 of the matrix @xmath26 , that is , @xmath126 .
given this set of data @xmath127 , a discretely obliquely reflected bsde , denoted @xmath128 , is a triplet @xmath129 satisfying @xmath130 , and defined in a backward manner , for @xmath131 and @xmath17 , by @xmath132 this rewrites equivalently for @xmath133 $ ] as @xmath134 observe that @xmath135 , since @xmath98 is nonnegative and valued in @xmath99 , for any @xmath136 .
we shall also use the following integrability condition for some @xmath137 : @xmath138 is a positive random variable satisfying @xmath139\le c_{l}$ ] .
importantly , @xmath140 does not depend on @xmath20 .
the proof of the following a priori estimates is postponed until the .
[ estprioedsrdrabs ] assume that ( [ cp ] ) holds for some given @xmath141 , there exists a unique solution @xmath142 to ( [ eqdebsdedor1abs ] ) and it satisfies @xmath143 in this subsection , we interpret the solution of the discretely obliquely rbsde ( [ eqbsdedorabs ] ) as the value process of a corresponding optimal switching problem , where the possible switching times are restricted to belong to the grid @xmath20 .
our approach relies on similar arguments as the one followed by hu and tang @xcite in a framework with continuous reflections .
a switching strategy @xmath144 is a nondecreasing sequence of stopping times @xmath145 , combined with a sequence of random variables @xmath146 valued in @xmath147 , such that @xmath148 is @xmath149-measurable , for any @xmath150 .
we denote by @xmath151 the set of such strategies . for @xmath152
, we introduce @xmath153 the ( random ) number of switches before @xmath51 as @xmath154 to any switching strategy @xmath155 , we associate the current state process @xmath156}$ ] and the compound cost process @xmath157}$ ] defined , respectively , by @xmath158 for @xmath159 . for @xmath160\times{\mathcal{i}}$ ] , the set @xmath161 of admissible strategies starting from @xmath61 at time @xmath162 is defined by @xmath163 < \infty\}.\ ] ] similarly , we introduce @xmath164 , the restriction to @xmath25-admissible strategies @xmath165 and denote @xmath166 . for @xmath160\times{\mathcal{i}}$ ] and @xmath167 , we consider as in @xcite the associated one - dimensional switched bsde defined by @xmath168\\[-8pt ] & & { } - a^a_t + a^a_u , \qquad t\le u \le t.\nonumber\end{aligned}\ ] ] theorem 3.1 in @xcite interprets each component of the solution to the continuously reflected bsde ( [ eqbsdecorintro ] ) as the snell envelope associated to switched processes of the form ( [ equa ] ) , where the switching strategies @xmath144 are not restricted to lie in the reflection grid @xmath20 . the next theorem is a new version of this snell envelope representation adapted to the context of discretely obliquely reflected bsde ( [ eqbsdedorabs ] ) .
[ thswr ] assume that is in force . for any @xmath111 and @xmath80 $ ] ,
the following hold : the process @xmath169 dominates any @xmath20-switched bsde , that is , @xmath170 define the strategy @xmath171 recursively by @xmath172 and , for @xmath173 , @xmath174\cap\re \bigm| ( { { \widetilde y}}_s^{\re})^{\alpha_{j-1}^ * } \le\max_{k\neq\alpha^*_{j-1 } } \ { ( { { \widetilde y}}^\re_s)^{k } - c^{\alpha^*_{j-1}k}_s \ } \bigr\},\\[-2pt ] \alpha^*_j & : = & \min\bigl\ { \ell\neq\alpha^*_{j-1 } \bigm| ( { { \widetilde y}}^\re_{\theta_j^*})^\ell- c^{\alpha^*_{j-1}\ell}_{\theta ^*_{j } } = \max_{k\neq\alpha^*_{j-1 } } \ { ( { { \widetilde y}}^{\re}_s)^k - c^{\alpha^*_{j-1 } k}_{\theta_{j}^ * } \ } \bigr\}.\end{aligned}\ ] ] then , we have @xmath175 and @xmath176 the following `` snell envelope '' representation holds : @xmath177 observe first that assertion ( iii ) is a direct consequence of ( i ) and ( ii ) .
let us fix @xmath80 $ ] and @xmath111 .
_ step _ 1 .
we first prove ( i ) .
set @xmath178 and the process @xmath179 defined , for @xmath180 $ ] , by @xmath181\\[-9pt ] z^a_s & : = & \sum_{k\ge0 } ( z_s^{\re})^{\alpha_{k } } { \mathbf1}_{\ { \theta_k \le s < \theta_{k+1}\ } } .\nonumber \ ] ] observe that these processes jump between the components of the discretely reflected bsde ( [ eqdebsdedor1 ] ) according to the strategy @xmath144 , and , between two jumps , we have @xmath182 & & { } + ( k ^{\re } _ { \theta _ { k+1}-})^{\alpha_{k } } - ( k ^{\re}_{\theta_{k}})^{\alpha _ { k}}\nonumber\\[-9pt]\\[-9pt ] & = & { { \widetilde y}}^{a}_{\theta_{k+1 } } + \int_{\theta_{k}}^{\theta_{k+1 } } f^{a_s}(s,{{\widetilde y}}^a_s , z^a_s ) \,{\mathrm{d}}s - \int_{\theta_{k}}^{\theta_{k+1 } } z^a_s \,{\mathrm{d}}w_s + ( k ^{\re}_{\theta_{k+1}-})^{\alpha_{k}}\nonumber\\[-2pt ] & & { } - ( k ^{\re}_{\theta _ { k}})^{\alpha_{k } } + \bigl((y^{\re}_{\theta_{k+1}})^{\alpha_{k } } - ( { { \widetilde y}}^{\re}_{\theta _ { k+1}})^{\alpha_{k+1}}\bigr ) , \qquad k\ge0.\nonumber\end{aligned}\ ] ] introducing @xmath183 & & \hspace*{25.8pt}{}+\mathbf{1}_{\ { \theta _ { k+1}\leq s \}}\bigl((y^{\re}_{\theta_{k+1}})^{\alpha_k}-({{\widetilde y}}^{\re } _ { \theta_{k+1}})^{\alpha_{k+1}}+c^{\alpha_{k}\alpha_{k+1}}_{\theta _ { k+1}}\bigr)\biggr]\end{aligned}\ ] ] for @xmath180 $ ] , and summing up ( [ babel ] ) over @xmath75 , we get , for @xmath184 , @xmath185 using the relation @xmath186 for all @xmath187 , we check that @xmath188 is increasing . since @xmath189 solves ( [ equa ] ) , we deduce by a comparison argument ( see @xcite , theorem 1.3 ) that @xmath190 . since @xmath144 is arbitrary in @xmath164 , we deduce ( [ equpperboundyu ] ) . _
we now prove ( ii ) .
consider the strategy @xmath191 given above as well as the associated process @xmath192 defined as in ( [ defyaza ] ) . by definition of @xmath191
, we have @xmath193 which gives @xmath194 for all @xmath195 .
we deduce from ( [ coststructureabs ] ) that @xmath196 hence , @xmath197 and @xmath198 are solutions of the same bsde and @xmath199 . to complete the proof ,
we only need to check that @xmath200 , that is , @xmath201 . by definition of @xmath191 on @xmath202 $ ] and the structure condition on the cost ( [ coststructureabs ] ) , we have @xmath203 which gives @xmath204 \leq c_{l}$ ] . combining @xmath205 with the lipschitz property of @xmath206 and the fact that @xmath207 ( recall proposition [ estprioedsrdrabs ] ) , we get the square integrability of @xmath208 and the proof is complete .
[ rethswrone ] although the optimal strategy @xmath191 depends on the initial parameters @xmath162 and @xmath61 , we omit the script @xmath209 for ease of notation . combining the previous representation with the a priori estimates of proposition [ estprioedsrdrabs ] and the structure condition ( [ coststructureabs ] ) ,
we deduce the following estimates , whose proof is postponed until the .
[ corhata ] assume that ( [ cp ] ) holds for some given @xmath141 , then @xmath210 } |u^{a^*}_{s}|^p+ \biggl(\int _
{ t}^t|v^{a^*}_{u}|^2\,{\mathrm{d}}u\biggr)^{p/2 } + |a^{a^*}_{t}|^p + @xmath211 , @xmath212\times { \mathcal{i}}$ ] .
we now study the dependence on the solution with respect to the parameters of the bsde . in the `` abstract '' setting considered , we obtain precious estimates for the analysis of the regularity of the solution to the discretely obliquely reflected bsde as well as the convergence of the discrete - time scheme .
we consider two discretely reflected bsdes , with the same reflection grid @xmath20 but different parameters . for @xmath213 , we consider an @xmath214-measurable random terminal condition @xmath215 , a random @xmath93-lipschitz continuous map @xmath216 , satisfying ( hf ) and a matrix of continuous cost processes@xmath217 satisfying the structural condition ( [ coststructureabs ] ) . we suppose that the coefficients satisfy the integrability condition ( c@xmath218 ) . for @xmath213 ,
we denote by @xmath219 the solution of the obliquely discretely reflected bsde @xmath220 . defining @xmath221 , @xmath222 , @xmath223 , @xmath224 together with @xmath225 $ ] , we prove the following stability result .
[ prstaby ] assume that holds .
then we have , for any @xmath80 $ ] , @xmath226 + \mathbb{e}[|{\delta\widetilde y}^{\re}_t|^2 ] + \frac { 1}{\kappa } \mathbb{e}\biggl[\int_{t}^{t } |{{\delta z}}^{\re}_{s}|^2 \,{\mathrm{d}}s \biggr ] \\ & & \qquad\le c_l \biggl(\mathbb{e}\biggl[\int_t^t |{{\delta f}}_s|^2_\infty\,{\mathrm{d}}s + |{{\delta\xi}}|^2\biggr ] + \mathbb{e}\bigl[\sup_{r\in\re}|{{\delta c}}_{r}|^4_\infty\bigr]^{1/2}\biggr).\end{aligned}\ ] ] the proof is divided into three steps and relies heavily on the reinterpretation in terms of switching problems . we first introduce a convenient dominating process and then provide successively the controls on the @xmath227 and @xmath228 terms
. _ step _ 1 . _
introduction of an auxiliary bsde .
_ let us define @xmath229 , @xmath230 and @xmath102 by @xmath231 .
observe that @xmath206 satisfies ( hf ) , @xmath102 satisfies the structure condition ( [ coststructureabs ] ) and that ( c@xmath218 ) holds for the data @xmath232 .
we denote by @xmath233 the solution of the discretely obliquely reflected bsde @xmath128 , recalling ( [ eqdebsdedor1abs ] ) . using ( hf ) , the definition of @xmath206 and the monotonicity property of @xmath108 [ see remark [ repropp](i ) ] , we easily obtain by a comparison argument on each interval @xmath234 , @xmath235 , that @xmath236 recalling theorem [ thswr ] , we introduce the switched bsdes associated to @xmath237 , @xmath238 and @xmath22 and denote by @xmath239 the optimal strategy related to @xmath22 starting from a fixed @xmath240 $ ] . therefore , we have @xmath241 _ step _ 2
. _ stability of the @xmath242 component . _ since @xmath243 , we deduce from theorem [ thswr](iii ) that @xmath244 where @xmath245 is the process of cumulated costs @xmath246 associated to the strategy @xmath247 . combining this estimate with ( [ eqdisccomp1 ] ) and ( [ eqdom0 ] ) , we derive @xmath248 since both terms on the right - hand side of ( [ eqtropfacile ] ) are treated similarly , we focus on the first one and introduce the continuous processes @xmath249 and @xmath250 . applying it s formula , we compute , for all @xmath251 , @xmath252 \nonumber\\ & & \qquad \le\mathbb{e}_{t } \biggl[|\gamma^{{{{\check{a}}}}}_t - { } ^{1}\gamma^{{{{\check{a}}}}}_t|^2\\ & & \hspace*{48pt } { } + 2 \int_u^t ( \gamma^{{{{\check{a}}}}}_s - { } ^{1}\gamma^{{{{\check{a}}}}}_s)[f^{{{{\check{a}}}}_s}(s , u^{{{{\check{a}}}}}_s,{}^{1}v^{{{{\check{a}}}}}_s)-{}^{1}f^{{{{\check{a}}}}_s}(s,{}^{1}u^{{{{\check{a}}}}}_s,{}^{1}v^{{{{\check{a}}}}}_s ) ] \,{\mathrm{d}}s\biggr ] . \nonumber\end{aligned}\ ] ] since @xmath253 and @xmath254 is lipschitz continuous , we also get @xmath255 ) that @xmath256\ ! + \
! \sup _ { t\le s \!\le\ ! t } \mathbb{e}_{t } [ |a^{{{{\check{a}}}}}_s\!-\ ! { } ^{1}a^{{{{\check{a}}}}}_s|^2 ] \biggr).\hspace*{-40pt}\end{aligned}\ ] ] moreover , using the inequality @xmath257 for @xmath258 and the convexity of the function @xmath259 , we compute @xmath260 \nonumber\\ & & \qquad= \mathbb{e}_{t } \biggl[\biggl| \sum_{k= 1}^{{{{n^{{{\check{a } } } } } } } } [ ^2c^{{{{\check{\alpha}}}}_{k-1}{{{\check{\alpha}}}}_{k}}\wedge^1c^{{{{\check{\alpha}}}}_{k-1}{{{\check{\alpha}}}}_{k}}-^1c^{{{{\check{\alpha}}}}_{k-1}{{{\check{\alpha}}}}_{k } } ] ( { { { \check{\theta}}}}_{k } ) \mathbf { 1 } _ { \ { { { { \check{\theta}}}}_{k } \le s \ } } \biggr|^2 \biggr]\\ & & \qquad\le \mathbb{e}_{t } \bigl[|n^{{{{\check{a}}}}}| \sup_{r \in\re}|{{\delta c}}_{r}|_{\infty}^2\bigr],\qquad t
\le s \le t.\nonumber\end{aligned}\ ] ] plugging [ eqcontrola ] in ( [ eqcontrolgamma ] ) and recalling the definition of @xmath261 and @xmath262 , we get @xmath263.\ ] ] the exact same reasoning leads to the same estimate for @xmath264 .
therefore , we deduce from ( [ eqtropfacile ] ) and the cauchy schwarz inequality that @xmath265\nonumber\\[-8pt]\\[-8pt ] & & \qquad\le c_l \biggl(\mathbb{e}[|n^{{{{\check{a}}}}}|^2]^{1/2 } \mathbb{e}\bigl[\sup_{r \in\re }
|{{\delta c}}_{r}|_{\infty}^4\bigr]^{1/2 } + \mathbb{e}\biggl[{\int_t^t |{{\delta f}}_{s}|_{\infty}^{2 } \,{\mathrm{d}}s + |{{\delta\xi}}|^2}\biggr]\biggr).\hspace*{-20pt}\nonumber\end{aligned}\ ] ] using proposition [ corhata ] , we compute , since @xmath61 is arbitrary , @xmath266 \nonumber\\[-8pt]\\[-8pt ] & & \qquad\le c_l \biggl ( \mathbb{e}\biggl[\int_t^t |{{\delta f}}_s|^2_\infty\,{\mathrm{d}}s + |{{\delta\xi}}|^2\biggr ] + \mathbb { e}\bigl[\sup_{r\in\re}|{{\delta c}}_{r}|^4_\infty\bigr]^{1/2 } \biggr).\nonumber\end{aligned}\ ] ] _ step _ 3 . _ stability of the @xmath0 component . _ applying it s formula to the _ cdlg _ process @xmath267 and noting @xmath268 , we obtain @xmath269\\ & & \qquad = \mathbb{e}\biggl[|{\delta\widetilde y}^{\re}_{t}|^2 + 2 \int_{t}^t { { \delta y}}^{\re}_{s } { { \delta f}}_{s } \,{\mathrm{d}}s + 2 \int_{t}^t { { \delta y}}^{\re}_{r } \,{\mathrm{d}}{\delta\widetilde{k}}^{\re } _ { r } \biggr],\end{aligned}\ ] ] where we used the fact that @xmath270 . since @xmath271 is a pure jump process , we compute @xmath272 \le \mathbb{e}\biggl[\alpha\sum_{t < r \le t , r\in\re } |{{\delta y}}^{\re}_{r}|^2 + \frac{1}{\alpha } \sum_{t < r \le t}|\delta{\delta\widetilde{k}}^{\re}_{r}|^2 \biggr],\qquad \alpha>0,\ ] ] which , for @xmath273 large enough and using standard arguments , leads to @xmath274\\ & & \qquad \le c_l \biggl ( \mathbb { e}[|{{\delta\xi}}|^2 ] + \mathbb{e}\biggl[\int_t^t |{{\delta f}}_{s}|^2_\infty\,{\mathrm{d}}s + \sum_{t < r \le
t , r\in\re } |{{\delta y}}^{\re}_{r}|^2\biggr ] \biggr).\end{aligned}\ ] ] since ( [ eqcontrolstaby ] ) holds true for any @xmath80 $ ] , we deduce @xmath274\\ & & \qquad \le c_l \kappa\biggl ( \mathbb { e}[|{{\delta\xi}}|^2 ] + \mathbb{e}\biggl[\int_t^t |{{\delta f}}_{s}|^2_\infty\,{\mathrm{d}}s \biggr ] + \mathbb { e}\bigl[\sup_{r \in\re}|{{\delta c}}_{r}|_{\infty}^4\bigr]^{1/2 } \biggr),\end{aligned}\ ] ] which concludes the proof of the proposition .
this section is dedicated to the derivation of regularity properties for the solution of discretely reflected bsdes .
these results are obtained in a markovian diffusion setting .
this means that the randomness of the parameter @xmath232 is due to a state process @xmath8 , which is the solution of a stochastic differential equation ( sde ) . in this framework , we focus on the @xmath275-regularity of the @xmath38 component of the solution of the bsdes .
the main results are retrieved by means of kernel regularization and malliavin differentiation arguments . finally , we extend this result to the case where the diffusion @xmath8 is replaced by its euler scheme .
let @xmath8 be the solution on @xmath57 $ ] to the following sde : @xmath276 where @xmath277 and @xmath278 are @xmath93-lipschitz functions . under the above assumption ,
the following estimates are well known ( see , e.g. , @xcite ) : @xmath279}|x_t|^p\bigr ] & \le & c_l^p \quad\mbox{and}\nonumber\\[-8pt]\\[-8pt ] \sup_{s\in[0,t ] } \bigl(\mathbb{e}\bigl[\sup_{u\in[0,t],|u - s|\le h}|x_{s}-x_{u}|^p\bigr ] \bigr ) ^{1/p } & \le & c^p_l \sqrt{h}\nonumber\end{aligned}\ ] ] for any @xmath95 . in the sequel
, we shall denote by @xmath280 a positive random variable , which may change from line to line , but which depends only on @xmath281 } |x_{t}|$ ] and which satisfies @xmath282\le c^p_{l}$ ] for all @xmath283 .
importantly , @xmath280 does not depend on @xmath20 .
observe that , as in @xcite and contrary to @xcite , we make no uniform ellipticity condition on @xmath50 .
this allows us to treat the case of nonhomogenous diffusion by setting , for example , @xmath284 , @xmath133 $ ] . in this context , we are given a matrix valued maps @xmath285 where @xmath286 , are @xmath93-lipschitz continuous and satisfy @xmath287 we then introduce a family @xmath288 of closed convex domains : @xmath289\\[-8pt ] & & \eqntext{\mbox{where } { \mathcal{i}}:=\ { 1,\ldots , d \}.}\end{aligned}\ ] ] we introduce the oblique projection operator @xmath290 onto @xmath291 defined by @xmath292 finally , we are given : an @xmath93-lipschitz function @xmath293 such that @xmath294 for all @xmath295 , a generator function , that is , an @xmath93-lipschitz map @xmath296 . from now on , we shall appeal to the following assumption : the component @xmath61 of @xmath297 depends only on the component @xmath61 of the vector @xmath67 and on the column @xmath61 of the matrix @xmath26 , that is , @xmath298 .
we denote by @xmath233 the solution of the discretely reflected bsde @xmath299 which reads on each interval @xmath300 , for @xmath301 @xmath302 or equivalently on @xmath57 $ ] as @xmath303 from ( [ eqcontrolx ] ) , it follows that the data @xmath304 satisfies the integrability condition ( [ cp ] ) for all @xmath141 .
we thus deduce from the proofs of propositions [ estprioedsrdrabs ] and [ corhata ] , the following estimate on @xmath233 and their associated optimal switched bsdes , recalling theorem [ thswr ] .
[ estprioedsrdr ] there exists a unique solution @xmath142 to ( [ eqdebsdedor1 ] ) and it satisfies @xmath305 } |{{\widetilde y}}^{\re}_{s}|^p + \biggl(\int_{t}^t|z_{s}^{\re}|^2\,{\mathrm{d}}s\biggr)^{p/2 } + |k^{\re } _ { t}-k^{\re}_{t}|^p \biggr ] \le\beta ^x\qquad \forall t \le t.\hspace*{-28pt}\ ] ] moreover , for all @xmath160\times{\mathcal{i}}$ ] , the optimal strategy @xmath306 satisfies @xmath307}|u^{a^*}_{s}|^p + \biggl(\int _ { t}^t|v^{a^*}_{s}|^2\,{\mathrm{d}}s\biggr)^{p/2}+ |a^{a^*}_{t}|^p + |n^{a^*}|^p \biggr ] \le\beta^x.\ ] ] we shall sometimes use the following regularity assumption on the coefficients : the coefficients @xmath308 , @xmath50 , @xmath6 @xmath5 and @xmath309 are @xmath310 in all their variables , with the lipschitz constants dominated by @xmath93 .
we denote by @xmath311 the set of random variables @xmath312 which are differentiable in the malliavin sense and such that @xmath313 , where @xmath314 denotes the malliavin derivative of @xmath312 at time @xmath10 . after possibly passing to a suitable version , an adapted process belongs to the subspace @xmath315 of @xmath275 whenever @xmath316 for all @xmath317 and @xmath318 . for a general presentation on malliavin calculus for
stochastic differential equations , the reader may refer to @xcite .
[ remarkmalliavindiffx ] under ( hr ) , the solution of ( [ eqforward ] ) is malliavin differentiable and its derivative satisfies @xmath319 and we have @xmath320\\[-9pt ] & & \qquad\le c_l^p | t - u |^{1/2}\nonumber\end{aligned}\ ] ] for any @xmath321 .
let @xmath322 .
since @xmath8 belongs to @xmath315 under ( hr ) and @xmath108 is @xmath41-lipschitz continuous , we deduce that @xmath323 . using lemma 5.1 in @xcite , we compute @xmath324 & & \qquad= \sum_{j = 1}^{d } \bigl(d_{s}g^{j } - d_{s}c_{ij}(x_{t})\bigr)\mathbf{1 } _ { \ { g^{j } - c^{ij}(x_{t } ) >
\max_{\ell < j}(g^{\ell}-c^{i \ell}(x_{t } ) ) \ } } \\[-2pt ] & & \qquad\quad\hspace*{13pt}{}\times\mathbf{1 } _ { \ { g^{j } - c^{ij}(x_{t } ) \geq\max_{\ell > j}(g^{\ell}-c^{i \ell}(x_{t } ) ) \ } } .
\nonumber\end{aligned}\ ] ] combining ( [ eqmalderivcp ] ) , proposition 5.3 in @xcite and an induction argument , we obtain that @xmath233 is malliavin differentiable and that a version of @xmath325 is given by @xmath326 & & { } + \int_t^{{{r_{j+1 } } } } \nabla_{x } f^i ( x_s,({{\widetilde y}}^{\re } _ s)^i,(z^{\re}_s)^{i\cdot } ) d_ux_s\,{\mathrm{d}}s \nonumber\\[-9pt]\\[-9pt ] & & { } + \int_t^{{{r_{j+1 } } } } \nabla_{y^i } f^i ( x_s,({{\widetilde y}}^{\re } _ s)^i,(z^{\re}_s)^{i\cdot } ) d_u({{\widetilde y}}^{\re}_s)^i \,{\mathrm{d}}s\nonumber\\[-2pt ] & & { } + \int_t^{{{r_{j+1 } } } } \nabla_{z } f^i ( x_s,({{\widetilde y}}^{\re}_s)^i,(z^{\re}_s)^{i\cdot } ) d_u(z^{\re}_s)^{i\cdot } \,{\mathrm{d}}s\nonumber \ ] ] for @xmath327 and @xmath301 . here
, @xmath328 denotes @xmath329 , recalling ( hf ) . for @xmath330
, we introduce the process @xmath331 defined by @xmath332\\[-9pt ] & & \hphantom{\exp\biggl\ { } { } - \int_t^s \biggl ( \frac { 1}{2 } |\nabla_{z } f^{a_r}(x_r,{{\widetilde y}}^{\re}_r , z^{\re}_r)|^2 - \nabla _ { y}f^{a_r}(x_r,{{\widetilde y}}^{\re}_r , z^{\re}_r )
\biggr ) { \mathrm{d}}r \biggr\}\hspace*{-28pt}\nonumber\end{aligned}\ ] ] for @xmath333 . for later use
, we remark @xmath334 and deduce from the dynamics of @xmath335 that @xmath336\\[-8pt ] & & \qquad\le c^p_l \sqrt{t - u } , \qquad u\le t\le t , p\ge2 .\nonumber \ ] ] [ proprepresz ] under , there is a version of @xmath38 such that @xmath337\nonumber\end{aligned}\ ] ] for @xmath160 $ ] , with @xmath338 the optimal strategy given in theorem [ thswr ] and recalling ( [ eqdefna ] ) .
we fix @xmath339 and , observing that the process @xmath191 is constant on the interval @xmath340 , we deduce from ( [ eqdy ] ) and it s formula that @xmath341 \ ] ] for @xmath342 . combining ( [ eqmalderivcp ] ) and the definition of @xmath343 given in theorem [ thswr](ii ) ,
we compute @xmath344 for @xmath339 . plugging the second equality into the first one and summing up over @xmath345
concludes the proof .
we conclude this section by providing a `` weak '' regularity property of @xmath38 in the general lipschitz setting . in order
to get rid of the previous assumption ( hr ) , we make use of kernel regularization arguments . since this procedure is very classical , we do not detail it here precisely ( see , e.g. , the proofs of proposition 4.2 in @xcite or proposition 3.3 in @xcite ) . [ przbounded ] there is a version of @xmath38 satisfying @xmath346 \le c_{l } |t - s| , \qquad s\le t \le t.\vadjust{\goodbreak}\ ] ] combining ( [ eqmajodx ] ) , with ( [ eqlambdasup ] ) , ( [ eqrepresz ] ) and doob s inequality , we observe that @xmath347 } \|z^{\re}_{t}\|_{_{\mathbf{l}^p}}\leq c_l^p,\qquad p\ge2,\ ] ] holds under ( hr ) .
therefore ( [ reguweakz ] ) is satisfied under ( hr ) . as in the proof of proposition 4.2 in @xcite ,
the stability results of proposition [ prstaby ] allow us to use classical kernel regularization arguments . since the previous estimate holds uniformly for the sequence of approximating regularized bsde , the proof is complete .
we consider a grid @xmath348 on the time interval @xmath57 $ ] , with modulus @xmath349 , such that @xmath350 .
we want to control the following quantities , representing the @xmath351-regularity of @xmath352 : @xmath353 \quad\mbox{and}\quad \mathbb{e}\biggl[\int_0^t \bigl| z^{\re}_t - \bar{z}^{\re}_{\pi ( t ) } \bigr|^2 \,{\mathrm{d}}t\biggr],\ ] ] where @xmath354 is defined on @xmath57 $ ] as the projection to the closest previous grid point of @xmath355 and @xmath356 , \qquad i\in\{0,\ldots , n-1 \}.\ ] ] [ rebestapprox ] observe that @xmath357 is interpreted as the best @xmath358-approximation of the process @xmath38 by adapted processes which are constant on each interval @xmath359 , for all @xmath360 .
[ prregy ] the following holds : @xmath361 \le\sup_{t \in[0,t ] } \mathbb{e}\bigl[\bigl|{{\widetilde y}}^{\re}_t - { { \widetilde y}}^{\re}_{\pi(t)}\bigr|^2 \bigr ] \le c_{l } |\pi|.\ ] ] observe first that @xmath362 \le\mathbb { e}\biggl[\biggl| \int_{\pi(t)}^t f(x_s,{{\widetilde y}}^{\re}_s , z^{\re}_s ) \,{\mathrm{d}}s + \int_{\pi(t)}^t z^{\re}_s \,{\mathrm{d}}w_s \biggr|^2 \biggr],\qquad 0\le t \le t.\ ] ] the proof is concluded combining this estimate with ( [ eqcontrolx ] ) , propositions [ estprioedsrdr ] and [ przbounded ] .
we now turn to the study of the regularity of the process @xmath38 .
[ thmreguz ] the process @xmath38 satisfies @xmath363 \le c_l ( |\pi|^{1/2 } + \kappa|\pi| ) .\vadjust{\goodbreak}\ ] ] a regularization argument as in proof of proposition [ przbounded ] allows us to work under ( hr ) . from remark [ rebestapprox ] , it is clear that @xmath364 \le\mathbb{e}\biggl[\int_0^t \bigl| z^{\re}_s - z^{\re}_{\pi(s ) } \bigr|^2 \,{\mathrm{d}}s \biggr].\ ] ] for @xmath365 and @xmath366 , @xmath367 , we define @xmath368 by @xmath369.\end{aligned}\ ] ] we now fix @xmath367 and denote , for @xmath370 , by @xmath371 the optimal strategy associated to the representation of @xmath372 , recalling ( ii ) in theorem [ thswr ] .
observe that , by definition , we have @xmath373 fix @xmath360 , and deduce from proposition [ proprepresz ] and ( [ eqstrategy ] ) that @xmath374 & = & \mathbb { e}[|v^ { a^{t}}_{t , t } - v^ { a^{{{t_i}}}}_{t_i , t_i}|^2 ] \nonumber\\[-8pt]\\[-8pt ] & \le&2 ( \mathbb{e}[|v^ { a^{{{t_i}}}}_{t , t } - v^ { a^{{{t_i}}}}_{t_i , t}|^2 ] + \mathbb{e}[|v^ { a^{{{t_i}}}}_{t_i , t } - v^ { a^{{{t_i}}}}_{t_i , t_i}|^2])\nonumber\end{aligned}\ ] ] for @xmath375 . combining ( hr ) , ( [ eqmajodx ] ) , ( [ controlregudsxt ] ) , ( [ eqlambdasup ] ) , ( [ eqlambdacontrol1 ] ) and cauchy schwarz inequality with the definition of @xmath376
, we deduce @xmath377 \le c_{l } | \pi|^{1/2 } , \qquad { { t_i}}\le t \le{{t_{i+1 } } } , i\le n .\ ] ] since @xmath378 is a martingale on @xmath379 $ ] , we obtain @xmath380 & \le & \mathbb{e}[|v^ { a^{{{t_i}}}}_{{{t_i}},{{t_{i+1 } } } } - v^ { a^{{{t_i}}}}_{{{t_i}},{{t_i}}}|^2 ] \nonumber\\ & \le & \mathbb{e}[|v^ { a^{{{t_i}}}}_{{{t_{i+1}}},{{t_{i+1}}}}|^2 - |v^ { a^{{{t_i}}}}_{{{t_i}},{{t_i}}}|^2 ] + \mathbb{e}[|v^ { a^{{{t_i}}}}_{{{t_i}},{{t_{i+1}}}}|^2 - |v^ { a^{{{t_i}}}}_{{{t_{i+1}}},{{t_{i+1}}}}|^2 ] \\ & \le & \mathbb{e}[|v^ { a^{{{t_i}}}}_{{{t_{i+1}}},{{t_{i+1}}}}|^2 - |v^ { a^{{{t_i}}}}_{{{t_i}},{{t_i}}}|^2 ] + c_{l } |\pi|^{1/2 } , \qquad { { t_i}}\le t\le{{t_{i+1}}},\nonumber\end{aligned}\ ] ] where the last inequality follows from ( [ eqcontrolz2bis ] ) . combining ( [ eqcontrolz1 ] ) ,
( [ eqcontrolz2bis ] ) , ( [ eqcontrolz3 ] ) and summing up over @xmath61 , we obtain @xmath381 \\ & & \qquad\le c_{l}|\pi|^{1/2 } + |\pi| \biggl ( \mathbb{e}[|v^ { a^{r_{\kappa-1}}}_{t , t}|^2 - |v^ { a^{0}}_{0,0}|^2 ] + \sum_{j=1}^{\kappa-1 } ( |v^ { a^{{{r_{j-1}}}}}_{{{r_j}},{{r_j}}}|^2 - |v^ { a^{{{r_j}}}}_{{{r_j}},{{r_j}}}|^2 ) \biggr).\end{aligned}\ ] ] combined with ( [ eqmajodx ] ) and ( [ eqlambdasup ] ) , this concludes the proof since @xmath382 is arbitrary .
we shall approximate the process @xmath8 by its eulerscheme @xmath383 , with dynamics @xmath384 classically , we have the following upper - bound , uniformly in @xmath355 : @xmath385^{1/p } \le c_{l}^p , \qquad p\geq2.\ ] ] the control of the error between @xmath8 and its euler scheme @xmath33 is well understood ( see , e.g. , @xcite ) and we have @xmath386^{1/p } \le c_{l}^p | \pi|^{1/2 } , \qquad p\geq2.\ ] ] in this context , we denote by @xmath387 the unique solution of the reflected bsde @xmath388 .
our main result here is the counterpart of proposition [ prregy ] and theorem [ thmreguz ] when @xmath8 is replaced by @xmath33 .
[ prregyeze ] the following hold : @xmath389 \le c_l z^{eu}_{s } |^2 \,{\mathrm{d}}s \biggr ] \le c_l ( |\pi|^{1/2 } + \kappa|\pi we only sketch the main step of the proof since it follows exactly the same arguments as the one used to obtain proposition [ prregy ] and theorem [ thmreguz ] .
_ step _ 1 .
we use a kernel regularization argument which allows us to work under ( hr ) . in this case , we observe that @xmath33 belongs to @xmath315 and satisfies @xmath390 for @xmath391 .
one then checks ( see @xcite , remark 5.2 , for details ) that @xmath392 it is also straightforward that @xmath387 is malliavin differentiable and satisfies ( [ eqdy ] ) with @xmath33 instead of @xmath8 .
_ step _ 2 . in order to retrieve the results of the proposition ,
one then follows exactly the same steps and arguments as the ones used in the previous sections [ sserepz ] and [ sseregz ] .
we present here a discrete time scheme for the approximation of the solution of the discretely obliquely reflected bsde ( [ eqdebsdedor1 ] ) .
recall that @xmath348 is a grid on the time interval @xmath57 $ ] , such that @xmath350 and @xmath393 . in the sequel , the process @xmath8 is approximated by its euler scheme @xmath383 ( see section [ sseeuler ] for details ) .
we introduce a euler - type approximation scheme for the discretely reflected bsdes .
starting from the terminal condition @xmath394 we compute recursively , for @xmath395 , @xmath396 , \vspace*{2pt}\cr { { \widetilde y^{\re,\pi}}}_{{{t_i } } } = \mathbb{e}[{y^{\re,\pi}}_{{t_{i+1}}}\mid\mathcal{f}_{{{t_i } } } ] + ( { { t_{i+1}}}- { { t_i } } ) f({x^{\pi}}_{{t_i}},{{\widetilde y^{\re,\pi}}}_{{t_i}},{{\bar z^{\re,\pi}}}_{{{t_i } } } ) , \vspace*{2pt}\cr { y^{\re,\pi}}_{{{t_i } } } = { { \widetilde y^{\re,\pi}}}_{{t_i}}\mathbf{1 } _ { \ { { { t_i}}\notin\re\ } } + { \mathcal{p}}({x^{\pi}}_{{t_i}},{{\widetilde y^{\re,\pi}}}_{{t_i}})\mathbf{1 } _ { \ { { { t_i}}\in\re\ } } .}\ ] ] this kind of backward scheme has been already considered when no reflection occurs ( see , e.g. , @xcite ) and in the reflected case ( see , e.g. , @xcite ) .
see also @xcite for a recent survey on the subject
. combining an induction argument with the lispchitz - continuity of @xmath5 , @xmath6 and the projection operator , one easily checks that the above processes are square integrable and that the conditional expectations are well defined at each step of the algorithm .
\(i ) this so - called `` moonwalk '' algorithm is given by an implicit formulation , and one should use a fixed point argument to compute explicitly @xmath397 at each grid point .
( ii ) in the two - dimensional case , hamadne and jeanblanc @xcite interpret @xmath398 as the solution of a doubly reflected bsde .
it is worth noticing that the solution of the corresponding discrete time scheme developed by @xcite for the approximation of doubly reflected bsde exactly coincides with @xmath399 derived here . for later use
, we introduce the piecewise continuous time scheme associated to the triplet @xmath400 . by the martingale representation theorem
, there exists @xmath401 such that @xmath402+\int_{{t_i}}^{{t_{i+1}}}{z^{\re,\pi}}_u \,{\mathrm{d}}w_u , \qquad i\le n-1,\ ] ] and by the it s isometry , for @xmath395 , @xmath403.\ ] ] we set @xmath404 for @xmath80 $ ] , define @xmath397 by @xmath405\\[-8pt ] & & { } - \int_t^{{t_{i+1}}}{z^{\re,\pi}}_u\,{\mathrm{d}}w_u,\qquad { { t_i}}\le t\le{{t_{i+1 } } } , i\in { \mathcal{i}},\nonumber\end{aligned}\ ] ] and introduce @xmath406 on @xmath57 $ ] by @xmath407 this can be rewritten as @xmath408 we finally provide a useful a priori estimate for the solution of the discrete time scheme whenever @xmath5 does not depend on @xmath26 , whose proof is postponed until appendix [ secap2 ] .
[ 4bushceme ] if @xmath5 does not depend on @xmath26 and @xmath409 , the following bound holds : @xmath410 \leq c_{l}^p,\qquad p\geq2.\ ] ] recall that @xmath411 neither depends on @xmath20 nor on @xmath355 .
the next proposition provides a control on the error between the discrete - time scheme ( [ scheme ] ) and the solution of the discretely reflected bsde ( [ eqdebsdedor1 ] ) .
[ thmerrorscheme ] the following holds : @xmath412 } \mathbb{e}[|{{\widetilde y}}^{\re}_t - { { \widetilde y^{\re,\pi}}}_t|^2 + & & \qquad\le c_{l } |l_{\mathcal{p}}|^{2\kappa } ( |\pi|^{1/2}+\kappa|\pi|),\nonumber\end{aligned}\ ] ] where we recall that @xmath413 is the lipschitz constant of the projection operator @xmath108 . as in section [ sseeuler ] ,
we consider @xmath387 the unique solution of the reflected bsde @xmath388 .
using proposition [ prstaby ] , the lipschitz property of @xmath5 , @xmath6 , @xmath414 and ( [ eqerroreuler ] ) , we obtain @xmath415 } \mathbb{e}[|{{\widetilde y}}^{\re}_t - { { \widetilde y}}^{eu}_t|^2 + \mathbb{e}\biggl[\int_0^t|z^{\re}_s - z^{eu}_{s } |^2 \,{\mathrm{d}}s\biggr]\nonumber\\[-8pt]\\[-8pt ] & & \qquad \le c_l |\pi|.\nonumber\end{aligned}\ ] ] using the same arguments as in the proof of proposition 3.4.1 , step 1.a in @xcite , for example , we get the following inequality : @xmath416 \nonumber\\[-8pt]\\[-8pt ] & & \qquad \le c_l \biggl ( \mathbb{e}\biggl[|y^{eu}_{{t_{i+1}}}- { y^{\re,\pi}}_{{t_{i+1}}}|^2\nonumber\\ & & \qquad\hphantom{\le c_l \biggl ( \mathbb{e}\biggl [ } { } + \int_{{t_i}}^{{t_{i+1}}}\bigl(\bigl|{{\widetilde y}}^{eu}_s - { { \widetilde y}}^{\re}_{\pi(s)}\bigr|^2 + \bigl|z^{eu}_s - \bar z^{\re } _ { \pi(s)}\bigr|^2\bigr ) \,{\mathrm{d}}s\biggr ] \biggr).\nonumber\end{aligned}\ ] ] there are two differences with the proof of proposition 3.4.1 in @xcite .
first , @xmath108 here depends both on @xmath66 and @xmath67 : but this is not a problem since @xmath417 and @xmath418 are parametrized by the same forward process @xmath33 .
second , @xmath108 is not @xmath419-lipschitz but only @xmath41-lipschitz , with @xmath420 , in its @xmath67 component .
this explains the term @xmath421 in ( [ eqthmerrorscheme ] ) .
indeed , we have , for @xmath422 , @xmath423 this leads , using an induction argument ( see , e.g. , step 1.b in the proof of proposition 3.4.1 in @xcite ) , to @xmath424 } \mathbb{e}[|{{\widetilde y}}^{eu}_t - { { \widetilde y^{\re,\pi}}}_t|^2 + |y^{eu}_t - { y^{\re,\pi}}_t|^2 ] + \mathbb{e}\biggl[\int_0^t |z^{eu}_s - { { \bar z^{\re,\pi}}}_{s } |^2 \,{\mathrm{d}}s \biggr]\\ & & \qquad \le c_l |l_{\mathcal{p}}|^{2 \kappa } \biggl ( |\pi| + \int_0^t\bigl(\bigl|{{\widetilde y}}^{eu}_s - { { \widetilde y}}^{\re } _ { \pi(s)}\bigr|^2 + \bigl|z^{eu}_s - \bar z^{\re}_{\pi(s)}\bigr|^2\bigr ) \,{\mathrm{d}}s \biggr).\end{aligned}\ ] ] combining the last inequality with proposition [ prregyeze ] and ( [ eqpourromu1 ] ) completes the proof
. the term @xmath421 , even when @xmath45 is small can be very large .
moreover , we shall see in the next section that it yields to a poor convergence rate for continuously reflected bsdes .
this term is due to the `` geometric '' approach , used in the proof of proposition [ thmerrorscheme ] , and the fact that @xmath108 is only @xmath41-lipschitz with @xmath425 . we obtain below a better control , using the stability results proved at the end of section [ sedisc ] but unfortunately under the assumption that @xmath5 does not depend on @xmath26 .
the optimal choice for @xmath45 in terms of @xmath426 is discussed in section [ subsecconvergence ] below .
[ thcvsh ] if @xmath5 does not depend on @xmath26 , the following holds : @xmath427 } \mathbb{e}[|{{\widetilde y}}^{\re}_{t}-{{\widetilde y^{\re,\pi}}}_{t}|^2 + |y^{\re } _
{ t}-{y^{\re,\pi}}_{t}|^2 ] & \le & c_l|\pi|,\\[-2pt ] \mathbb{e}\biggl[\int_{0}^{t}|z^{\re}_{t } - { { \bar z^{\re,\pi}}}_{t}|^2 \,{\mathrm{d}}t\biggr ] & \le & c_{l}(\kappa|\pi| + |\pi|^{1/2 } ) \end{aligned}\ ] ] for @xmath426 small enough .
we use here the stability results of proposition [ prstaby ] setting @xmath428 with @xmath429 and @xmath430 , with @xmath431 . combining ( [ 4propubscheme ] ) and proposition [ estprioedsrdr ] with the lipschitz property of @xmath5 ,
it is clear that ( c@xmath218 ) holds .
applying proposition [ prstaby ] and ( [ eqerroreuler ] ) , we derive , for @xmath80 $ ] , @xmath432\\[-9pt ] & & \qquad\le c_{l}\biggl(|\pi|+ \int_{t}^t{\mathbb{e}}\bigl|{{\widetilde y}}^{\re}_{s } -{{\widetilde y}}^{\re}_{{\pi ( s)}}\bigr|^2 \,{\mathrm{d}}s+\int_{t}^t{\mathbb{e}}\bigl|{{\widetilde y^{\re,\pi}}}_{\pi(s)}-{{\widetilde y}}^{\re}_{{\pi(s)}}\bigr|^2\,{\mathrm{d}}s\biggr).\nonumber\end{aligned}\ ] ] applying the discrete version of gronwall s lemma to estimate ( [ eqconvscheme1 ] ) rewritten at time @xmath433 , we deduce @xmath434\\[-9pt ] \eqntext{0\le t\le t_{j}\le t , t_j\in\pi.}\end{aligned}\ ] ] plugging this estimate into ( [ eqconvscheme1 ] ) , we compute @xmath435 which combined with proposition [ prregy ] leads to the first claim of the theorem .
observe from the representations ( [ defbarz ] ) and ( [ eqrepbzp ] ) that @xmath436 \\ & & \qquad\le c_{l } \biggl(\mathbb{e}\biggl[\int_{0}^{t } |z^{\re}_{t } - \bar{z}^{\re } _ { t}|^{2 } \,{\mathrm{d}}t\biggr ] + \mathbb{e}\biggl[\int_{0}^{t } |z^{\re}_{t } - { z^{\re,\pi}}_{t}|^{2 } \,{\mathrm{d}}t\biggr ] \biggr).\end{aligned}\ ] ] plugging ( [ eqreguz1 ] ) , estimate ( [ eqconvscheme1 ] ) written at time @xmath437 and the first claim of this theorem into this expression concludes the proof .
in this section , we extend the convergence results of the scheme ( [ scheme ] ) to the case of continuously reflected bsdes . to this end , we show that the error between discretely and continuously obliquely reflected bsdes is controlled in a convenient way . in the sequel , we shall use the following assumption on @xmath5 : the function @xmath5 is bounded in its last variable : @xmath438 and the following assumption on the cost @xmath414 . for @xmath439 ,
the function @xmath440 is equal to @xmath441 , with @xmath442 is @xmath443 with bounded first and second derivatives and @xmath444 is a convex function with bounded first derivative .
this last assumption is needed to retrieve some regularity on the reflecting process @xmath90 ( see lemma [ lereguk ] below ) .
we denote by @xmath445 the solution of the continuously obliquely reflected bsde @xmath446,c(x),f(x,\cdot ) , g(x_{t}))$ ] defined by @xmath447 \,{\mathrm{d}}k ^i_t=0 , \qquad \mbox{$i\in{\mathcal{i}}$}.}\ ] ] under the assumption on @xmath5 , @xmath6 and @xmath414 , the existence and uniqueness of such a solution is given in @xcite .
the solution of ( [ eqbsdecor ] ) has also a representation property in term of switched bsdes , recalling ( [ equa ] ) . here
, of course , the switching times of the strategy are not restricted to take their values in @xmath20 .
we refer to @xcite for more details .
[ thsummary ] there exists , for any fixed initial condition @xmath448\times{\mathcal{i}}$ ] , an optimal switching strategy @xmath449 , such that @xmath450 we deduce from ( [ snellcontinuous ] ) , theorem [ thswr](iii ) , the monotonicity property of @xmath108 and ( [ eqbsdecor ] ) @xmath451 moreover , most of the estimates presented in section [ sedisc ] for discretely reflected bsdes hold true for continuously reflected bsdes . for reader
s convenience , we collect them in the following proposition .
the proof itself is postponed to appendix [ secap3 ] .
[ prstabcont ] the following a priori estimates hold . for any @xmath141 , @xmath452 + \mathbb{e}_{t } [ |k _ { t}-k _ { t}|^p ] \le\mathbb{e}_{t } [ \beta^x ] , \nonumber\\[-8pt]\\[-8pt ] \eqntext{0 \le t \le t,}\ ] ] and , for all @xmath453\times{\mathcal{i}}$ ] , the optimal strategy @xmath454 satisfies @xmath455 } |u^{\dot a}_{s}|^p \bigr ] + \mathbb{e}_{t } [ |n^{\dot a}|^p ] \le\mathbb
{ e}_{t } [ \beta^{x}].\ ] ] we first provide a control of the error on the grid points of @xmath20 between the solutions of the obliquely discretely and continuously reflected bsdes ( [ eqbsdedor ] ) and ( [ eqbsdecor ] ) .
[ thmerrordiscretize ] under , the following holds : @xmath456 \le c_l^{\varepsilon}|\re|^{1-{\varepsilon } } , \qquad { \varepsilon}>0.\ ] ] moreover , if the cost functions are constant , the last inequality holds true with @xmath457 .
the proof of this result relies mainly on the interpretation in terms of switched bsdes provided in section [ subsediscopt ] . for a fixed @xmath458\times{\mathcal{i}}}$ ] , we associate to the optimal strategy @xmath459 not restricted to lie in the grid @xmath20 , the corresponding `` discretized '' strategy @xmath460 defined by @xmath461 _ step _ 1 .
we first derive two key controls on the distance between @xmath462 and @xmath463 .
we fix @xmath464 and , since @xmath465 , @xmath466 , we compute @xmath467\\[-8pt ] & & \qquad \le c_{l}^p\int_{t}^t \biggl| \sum_{k=1}^{n^{\dot a } } [ c^{\dot\alpha_{k-1}\dot\alpha_k}(x_{\theta_{k}})-c^{\dot\alpha _ { k-1}\dot\alpha_k}(x_{\dot\theta_{k } } ) ] { \mathbf1}_{\theta_k \le s } \biggr|^p\,{\mathrm{d}}s \nonumber\\ & & \qquad\quad { } + c_{l}^p\biggl(\int_{t}^t \biggl| \sum_{k=1}^{n^{\dot a } } c^{\dot\alpha_{k-1}\dot\alpha_k}(x_{\dot\theta_{k}}){\mathbf 1}_{\dot \theta_k \le s
< \theta_{k } } \biggr|^2 \,{\mathrm{d}}s\biggr)^ { p/2}.\nonumber\end{aligned}\ ] ] using the convexity inequality @xmath468 , we obtain @xmath469\\[-8pt ] & & \qquad\leq c^p_{l } \bigl(1+\sup_{t\in [ 0,t]}|x_{t}|^p\bigr)|n^{\dot a}|^p|\re|^{p/2}.\nonumber\end{aligned}\ ] ] using once again the same convexity inequality with @xmath470 , the lipschitz property of the maps @xmath7 and the definition of @xmath471 and @xmath472 , we get @xmath473 { \mathbf1}_{\theta_k \le s } \biggr|^p \,{\mathrm{d}}s & \leq & c^p_{l}|n^{\dot a}|^{p-1 } \sum_{k=1}^{n^{\dot a}}|x_{\theta _ { k}}-x_{\dot\theta_{k}}|^p \\ & \leq & c^p_{l}|n^{\dot a}|^{p } \chi^{|\re|,p},\end{aligned}\ ] ] where @xmath474}|x_{r}-x_{r_{k}}|^p$ ] . plugging this estimate and ( [ 4maj1deltaa ] ) in ( [ 4decompdeltaa ] )
, we deduce @xmath475\\[-8pt ] & & \qquad\le c^p_l |n^{\dot a}|^p\bigl(\bigl(1+\sup_{s\in[0,t]}|x_{s}|^p \bigr)|\re that , for @xmath476 , we have @xmath477 which gives @xmath478\\[-8pt ] & \le & c_l |n^{\dot a}|^p \chi^{|\re|,p}.\nonumber\end{aligned}\ ] ] _ step _ 2 .
we now prove the main result of the theorem .
we introduce the processes @xmath479 and @xmath480 , so that @xmath481 on @xmath202 $ ] , using gronwall s lemma and the lipschitz property of @xmath5 , we obtain @xmath482\\[-9pt ] & & \qquad\le c_l { \mathbb{e}}_{t}\biggl[\int_{t}^t elevating this expression to the power @xmath483 , we deduce @xmath484 & & \qquad\le c^p_l { \mathbb{e}}_t\biggl[\biggl(\int _ { t}^t |[f^{\dot a_s } - f^{a_s}](x_{s},u^{\dot a}_s , v^{\dot a}_s)|^2 \,{\mathrm{d}}s\biggr)^{p/2}\\[-2pt ] & & \qquad\hspace*{100.6pt } { } + \biggl ( \int_{t}^t | a^{\dot a}_s - a^a_s|^2 \,{\mathrm{d}}s\biggr)^{p/2}\biggr].\nonumber\end{aligned}\ ] ] combining the definition of @xmath485 with the lipschitz property of @xmath5 and ( hz ) , we compute @xmath486(x_{s},u^{\dot a}_s , v^{\dot a}_s)|^2 \,{\mathrm{d}}s\\[-2pt ] & & \qquad = \int_{t}^t \biggl|\sum_{k=1}^ { n^{\dot a } } f^{\alpha _ { k-1}}(x_s , u^{\dot a}_s , v^{\dot a}_s ) ( { \mathbf1}_{\dot\theta_{k-1 } \le s <
\dot\theta_k } - { \mathbf1}_{\theta_{k-1 } \le s < \theta_k } ) \biggr|^2 \,{\mathrm{d}}s\\[-2pt ] & & \qquad \le c_l |n^{\dot a}|^2 \sup_{s \in[0,t ] } ( 1+|x_s|^2 + | u^{\dot a}_s|^2 ) |\re|.\end{aligned}\ ] ] plugging the last inequality and ( [ eqcontrolba - a1 ] ) in ( [ eqteubee2 ] ) , we deduce @xmath487 } ( 1+|x_s|^p + | u_s^{\dot a}|^p)|\re + \chi^{|\re|,p}\bigr)\bigr].\ ] ] restricting to the case where @xmath488 , we deduce from ( [ eqcontrolba - a2 ] ) and ( [ eqteubee1 ] ) that @xmath489 } ( 1+|x_s|^p + | y _
s|^p ) \bigr]^{2/p } |\re| + { \mathbb{e}}_t\bigl[|n^{\dot a}|^p |\chi^{|\re|,p}\bigr]^{2/p } \bigr).\end{aligned}\ ] ] using cauchy
schwarz inequality and proposition [ prstabcont ] with the last inequality , we obtain @xmath490^{1/p } \bigr).\vadjust{\goodbreak}\ ] ] again using cauchy
schwarz inequality and defining @xmath491 $ ] , we get @xmath492 \le c^p_l \bigl(|\re| + \mathbb{e}\bigl[\sup_{t\in[0,t ] } |m_{t}|^{2/ p } \bigr]^{1/2}\bigr ) .\ ] ] combining burkholder
gundy and convexity inequalities with ( [ eqcontrolx ] ) , we compute @xmath493}|m_{t}|^{2/p}\bigr ] & \leq & c_{l}^p ( |m_{0}|^{2/p}+ { \mathbb{e}}[|m_{t}|^2 ] ^{1/p } ) \leq c_{l}^p { \mathbb{e}}\bigl[\bigl|\chi^{|\re|,p}\bigr|^4\bigr]^{1/p}\\ & \leq & c_{l}^p |\kappa|^{4/p } |\re|^{2}.\end{aligned}\ ] ] plugging this expression in ( [ 4relmiparcourt ] ) , we deduce ( [ 4errordisccontcasgen ] ) from the condition @xmath494 and the arbitrariness of @xmath61 .
_ step _ 3 .
we finally consider the particular case where the cost functions are constant . following the same arguments as in step 1
, we observe that ( [ eqcontrolba - a1 ] ) turns into @xmath495}|x_{s}|^p\bigr)|\re|^{p/2},\ ] ] and that @xmath496 for @xmath497 .
the same reasoning as in step 2 then leads to @xmath498 } ( 1+|x_s|^p + | \dot y_s|^p)\bigr]^{2/p } the proof .
we now present the main result of this section , which allows us to control the error between the solutions of the continuously and the discretely obliquely reflected bsde at any time between @xmath499 and @xmath51 . [ thmerrordiscretize2 ] under , the following holds : @xmath500}\mathbb{e}[|y _
t-{{\widetilde y}}^{\re}_t|^2+|y _ t - y^{\re } _ t|^2]+
\mathbb{e}\biggl[\int_0^t|z _ s- z^{\re}_s|^2 \,{\mathrm{d}}s\biggr]\\ & & \qquad \lec^\varepsilon_{l}|\re|^{1/2 - \varepsilon } , \qquad \varepsilon>0.\nonumber\end{aligned}\ ] ] if , furthermore , the cost functions are constant , the previous estimate holds true for @xmath457 . in order to prove this theorem , we first state the following lemma discussing the regularity of @xmath501 .
[ lereguk ] under , there exists some positive process @xmath502 satisfying @xmath503 and such that , for all @xmath504 , @xmath505 in the sense of random measure .
we follow here the main idea of the proof of proposition 4.2 in @xcite and divide the proof in three steps . _
fix @xmath439 .
we first observe using it
tanaka formula , that , under ( hc ) , @xmath506 where @xmath507 is an increasing process and @xmath508 we then introduce @xmath509 . using once again it tanaka formula , we compute @xmath510^{+ } & = & [ \gamma^{ij}_{0}]^{+ } + \int _ { 0}^{t}\bigl(-f^{i}(x_{s},y ^{i}_{s},z ^{i}_{s})+f^{j}(x_{s},y ^{j}_{s},z ^{j}_{s } ) + b^{ij}_{s}\bigr)\mathbf{1 } _ { \ { \gamma^{ij}_{s}>0 \ } } \,{\mathrm{d}}s \\ & & { } + \int_{0}^{t}(\nu^{ij}_{s}+z ^{i}_{s}-z ^{j}_{s})\mathbf{1 } _ { \ { \gamma^{ij}_{s}>0 \ } } \,{\mathrm{d}}w_{s}\\ & & { } + \int_{0}^{t}\mathbf{1 } _ { \ { \gamma^{ij}_{s}>0 \ } } ( - { \mathrm{d}}k ^{i}_{s } + { \mathrm{d}}k ^{j}_{s } - { \mathrm{d}}\delta^{ij}_{s } ) + \frac12 \int_{0}^{t}{\mathrm{d}}l^{ij}_{s}\end{aligned}\ ] ] for @xmath511 , where @xmath512 is the local time at @xmath499 of the continuous semi - martingale @xmath513 . since @xmath514 and @xmath507 , @xmath512 are increasing processes , we compute @xmath515\\[-8pt ] & \le & c_{l}\bigl(1 + |x_{s}|+ \sup_{\ell\in{\mathcal{i}}}|y ^{\ell}_{s}| + \sup _ { \ell , k \in{\mathcal{i}}}|b^{\ell k}_{s}|\bigr ) \,{\mathrm{d}}s\nonumber\\ & & { } + \mathbf{1 } _ { \ { \gamma ^{ij}_{s}=0 \ } } \,{\mathrm{d}}k ^{j}_{s } \nonumber\end{aligned}\ ] ] for @xmath516 , where we used ( hz ) in order to obtain the last inequality . _ step _ 2 . we now prove that @xmath517 in the sense of random measure .
we first observe that @xmath518 with @xmath519 . indeed ,
if @xmath520 were a positive random measure on @xmath57 $ ] , this would contradict the minimality condition ( [ eqbsdecor ] ) for @xmath501 .
suppose the existence of a stopping time @xmath521 smaller than @xmath51 , such that @xmath522 by definition of the projection @xmath108 , we have @xmath523 where @xmath524 takes value in @xmath147 .
moreover , @xmath525 , which leads , combined with ( [ eqcontra0 ] ) and ( [ eqcontra1 ] ) , to @xmath526 and then contradicts ( [ coststructure ] ) .
thus , @xmath527 for any stopping time @xmath521 smaller than @xmath51 and we deduce that @xmath528 is undistinguishable from @xmath499 , which proves ( [ eqdkj=0 ] ) .
_ step _ 3 . to conclude , using once again the minimality condition for @xmath90 in ( [ eqbsdecor ] ) , observe that @xmath529 with @xmath530 which satisfies @xmath531 , recalling ( [ eqcontrolx ] ) , ( [ eqestimateia ] ) and ( [ eqmajobij ] ) .
proof of theorem [ thmerrordiscretize2 ] fix @xmath80 $ ] and introduce @xmath532 , @xmath533 , @xmath534 and @xmath535 . applying it s formula to the _ cdlg _ process @xmath536 , we get @xmath537\\[-8pt ] & & \qquad= |{\delta{{\widetilde y}}}_t|^2 - 2\int _ { ( t , t]}{\delta{{\widetilde y}}}_{s-}\,{\mathrm{d}}{\delta{{\widetilde y}}}_s - \sum_{t < s \le t}|{\delta{{\widetilde y}}}_s -{\delta y}_{s}|^2.\nonumber\end{aligned}\ ] ] recalling that @xmath538 , @xmath539 } { \delta y}_s\,{\mathrm{d}}k ^{\re}_s \ge0 $ ] and the lipschitz property of @xmath5 , standard arguments lead to @xmath540 & \le & c_{l}\mathbb{e}\biggl[\int_t^t { \delta y}_s\,{\mathrm{d}}k _ s \biggr]\nonumber\\[-8pt]\\[-8pt ] & \le & c_{l } \sum_{j<\kappa } \mathbb{e}\biggl[\int_{{{r_j}}}^{{{r_{j+1 } } } } { \delta y}_{s } \,{\mathrm{d}}k _ { s}\biggr].\nonumber\end{aligned}\ ] ] using the expression of @xmath541 and lemma [ lereguk ] , we obtain @xmath542 combining ( hz ) , ( [ eqcontrolx ] ) , ( [ controlcontref ] ) , ( [ eqestimateia ] ) and the fact that @xmath531 , we deduce @xmath543 & \le & \mathbb{e}\biggl[\sum_{j<\kappa } \int_{{{r_j}}}^{{{r_{j+1 } } } } \!\!\int _ { s}^{{{r_{j+1}}}}({\delta f}_{u } + \eta_{u})\,{\mathrm{d}}u \,{\mathrm{d}}k _ { s}\biggr ] \\ & & { } + \mathbb { e}\biggl[\sum_{j<\kappa}\int_{{{r_j}}}^{{{r_{j+1}}}}{\delta y}_{{{r_{j+1}}}}\,{\mathrm{d}}k _ { s}\biggr ] \\ & \le & c_{l}|\re| + \mathbb{e}\bigl[k _ { t}\sup_{r \in\re}|{\delta y}_{r}|\bigr].\end{aligned}\ ] ] plugging this expression in ( [ eqcontroldydzref0 ] ) and using cauchy
schwarz inequality together with ( [ 4errordisccontcasgen ] ) and proposition [ estprioedsrdrabs ] concludes the proof . combining the previous results with the control of the error between the discrete - time scheme and
the discretely obliquely reflected bsde derived in section [ subsecscheme ] , we obtain the convergence of the discrete time scheme to the solution of the continuously obliquely reflected bsde . in the next theorem ,
we detail the corresponding approximation error for different optimal choices of reflection time step @xmath544 with respect to the discrete time step @xmath426 .
the following hold : if holds , taking @xmath545 for @xmath24 , we have @xmath424 } \mathbb{e}[|y _ t - { { \widetilde y^{\re,\pi}}}_t|^2 + |y _ t - { y^{\re,\pi}}_t|^2 ] + \mathbb{e}\biggl[\int^t_{0}|z _ s - { { \bar z^{\re,\pi}}}_{s } |^2 \,{\mathrm{d}}s\biggr]\\ & & \qquad \le\frac { c_{l}^{\varepsilon}}{[-{\log } ( |\pi|)]^{1/2-{\varepsilon}}}.\end{aligned}\ ] ] if @xmath5 does not depend on @xmath26 and @xmath409 , taking similar grids @xmath546 , we have @xmath547 \le c_l^{\varepsilon}|\pi|^{1-{\varepsilon } } , \qquad { \varepsilon}>0.\ ] ] moreover , under , @xmath548 } \mathbb{e}[|y _ t - { y^{\re,\pi}}_t|^2 + |y _ t - { { \widetilde
y^{\re,\pi}}}_t|^2 ] \le c^{\varepsilon}_l |\pi|^{1/2 - { \varepsilon } } , \qquad { \varepsilon}>0.\ ] ] under , if @xmath5 does not depend on @xmath26 and @xmath549 , taking @xmath550 , we get @xmath551 \le c_{l}^{{\varepsilon } } furthermore , for constant cost functions , and hold true with @xmath552 . for @xmath24 ,
setting @xmath20 such that @xmath553 , we obtain , combining proposition [ thmerrorscheme ] and theorem [ thmerrordiscretize2 ] , that @xmath424 } \mathbb{e}[|y _ t - { { \widetilde y^{\re,\pi}}}_t|^2 + |y _ t - { y^{\re,\pi}}_t|^2 ] + \mathbb{e}\biggl[\int^t|z _ s - { { \bar z^{\re,\pi}}}_{s } |^2 \,{\mathrm{d}}s\biggr]\\ & & \qquad\le c^{{\varepsilon}}_{l } \biggl [ \biggl(\frac{-1}{\log(|\pi|)}\biggr)^{1/2 -{\varepsilon } } \vee|\pi|^{1/2 - { \varepsilon } } \biggr].\end{aligned}\ ] ] therefore , ( i ) is proved .
furthermore , ( ii ) , ( iii ) and ( iv ) are direct consequences of theorems [ thcvsh ] and [ thmerrordiscretize ] or [ thmerrordiscretize2 ] .
proof of proposition [ estprioedsrdrabs ] observing that on each interval @xmath554 , @xmath555 solves a standard bsde , existence and uniqueness follow from a concatenation procedure and @xcite . the rest of the proof divides in two steps controlling separately @xmath169 and @xmath556 . since @xmath569 are solutions to standard nonreflected bsdes , usual arguments lead to @xmath570 where the positive random variable @xmath571 satisfies classically @xmath572 \le c_{l}$ ] , under condition ( [ cp ] ) for a given @xmath137 .
we fix @xmath10 and applying it s formula to the _ cdlg _ process @xmath573 on @xmath574 $ ] to derive @xmath575}|z^{\re}_s|^2 \,{\mathrm{d}}s\\ & & { } + \sum_{s\le t}(|{{\widetilde y}}^{\re}_s|^2-|{{\widetilde y}}^{\re}_{s-}|^2- 2 { { \widetilde y}}^{\re}_{s- } \delta y ^{\re}_s).\end{aligned}\ ] ] since the last term on the right - hand side is nonnegative , we deduce that @xmath576}{{\widetilde y}}^{\re}_{s- } \,{\mathrm{d}}k ^{\re}_s + 2 \int_t^t ( z^{\re } _ s{{\widetilde y}}^{\re}_{s } ) \,{\mathrm{d}}w_s.\end{aligned}\ ] ] using standard arguments , together with ( [ eqespsupyp ] ) and ( [ cp ] ) for a fixed @xmath137 , we compute @xmath577 combining ( [ eqapco2c ] ) and ( [ eqapco2cbis ] ) we obtain @xmath579\\[-8pt ] & & { } + c_{l } \int_t^t ( z^{\re}_s{{\widetilde y}}^{\re}_s)\,{\mathrm{d}}w_s\nonumber\end{aligned}\ ] ] for any @xmath24 . elevating the previous estimate to the power @xmath580 , it follows from burkholder
gundy inequality that @xmath581\\ & & \qquad\le c^{p}_l \biggl ( { \varepsilon}^{-{p/2 } } \mathbb{e}_{t } [ \bar { \beta } ] + { \varepsilon}^{p/2}\mathbb{e}_{t } \biggl[\biggl(\int_t^t|z^{\re}_s|^2\,{\mathrm{d}}s\biggr)^ { p/2}\biggr ] \\ & & \hspace*{109.5pt } { } + \mathbb{e}_{t } \biggl[\biggl(\int_t^t |z^{\re}_s{{\widetilde y}}^{\re}_s |^2 \,{\mathrm{d}}s\biggr)^{p/4}\biggr]\biggr)\\ & & \qquad \le c^{p}_l \biggl({\varepsilon}^{-{p/2}}\mathbb{e}_{t } [ \bar { \beta } ] + { \varepsilon}^{-{p/2}}\mathbb{e}_{t } \bigl[\sup_{s\in[t , t]}|{{\widetilde y}}^{\re }
_ s|^{p}\bigr ] \\ & & \hspace*{81.7pt } { } + { \varepsilon}^{p/2}\mathbb{e}_{t } \biggl[\biggl(\int_t^t|z^{\re}_s|^2\,{\mathrm{d}}s\biggr)^ { p/2}\biggr ] \biggr).\end{aligned}\ ] ] using ( [ eqespsupyp ] ) and ( [ cp ] ) , we deduce , for @xmath582 small enough , @xmath583 \le c_l^p\mathbb{e}_{t } [ \bar{\beta } ] .\ ] ] taking ( [ eqapco2cbis ] ) up to the power @xmath584 , and combining burkholder davis gundy inequality with ( [ eqcontrolintzt - t ] ) yields @xmath585 \le c_l^p\mathbb{e}_{t } [ \bar{\beta}]$ ] , which concludes the proof of the proposition , recalling ( [ cp ] ) . proof of proposition [ corhata ] fix @xmath160\times{\mathcal{i}}$ ] and @xmath137 . according to the identification of @xmath586 with @xmath192 , obtained in the proof of theorem [ thswr ] , we deduce from proposition [ estprioedsrdrabs ] the expected controls on @xmath587 and @xmath588 .
writing the equation satisfied by @xmath589 and using standard arguments for bsdes , we observe that @xmath590 \le c^p_{l } \biggl ( \mathbb{e}_{t } \biggl[\sup_{s\in[t , t]}|u^{a^*}_{s}|^p + \biggl(\int_{t}^t|v^{a^*}_{s}|^2\,{\mathrm{d}}s \biggr)^{p/2}\biggr ] + proof of proposition [ 4bushceme ] the proof follows exactly the same arguments as in step 1 of the proof of proposition [ estprioedsrdrabs ] above .
the only difficulty is the use of a comparison argument for euler scheme that we provide below in lemma [ propcompascheme ] . for @xmath594 ,
let @xmath595 be a square integrable random variable and @xmath596 an @xmath93-lipschitz generator function .
we suppose that @xmath597 and @xmath598 on @xmath599 . for a time grid @xmath355 ,
we denote by @xmath600 the discrete - time scheme starting from the terminal condition @xmath601 and computing recursively , for @xmath602 , @xmath603 + ( { { t_{i+1}}}- { { t_i } } ) \psi _ { k}({x^{\pi}}_{{t_i}},y^{\pi , k}_{{t_i } } ) . \ ] ] since the result holds true on the grid point @xmath606 and follows from a backward induction on @xmath355 , we just prove @xmath607 . using ( [ schemethmcomp ] )
, we compute @xmath608 + ( t-{t_{n-1}})\lambda_{n-1 } ( y^{\pi , 1}_{t_{n-1}}-y^{\pi,2}_{t_{n-1}})\nonumber\\[-8pt]\\[-8pt ] & & { } + \delta_{n-1},\nonumber\end{aligned}\ ] ] where @xmath609 and @xmath610 since @xmath611 is @xmath93-lipschitz , the condition @xmath409 , implies @xmath612 . plugging this estimate , @xmath613 and @xmath597 and @xmath611 in ( [ 4deltayappendix ] )
, the proof is complete .
proof of proposition [ prstabcont ] the proof of ( [ eqestimateia ] ) is a direct adaptation of the proof of proposition [ estprioedsrdrabs ] . the only difference is in step 1 : we approximate @xmath614 by a sequence of penalized bsdes ( see the proof of theorem 2.4 in @xcite or step 3 in the proof of theorem [ thmerrordiscretize2 ] ) which are bounded by @xmath615 and @xmath564 .
estimate ( [ eqestimateib ] ) follows from the exact same arguments as the one used in the proof of proposition [ corhata ] . | in this paper , we study the discrete - time approximation of multidimensional reflected bsdes of the type of those presented by hu and tang [ _ probab . theory related fields _ * 147 * ( 2010 ) 89121 ] and generalized by hamadne and zhang [ _ stochastic process .
appl . _
* 120 * ( 2010 ) 403426 ] . in comparison to the penalizing approach followed by hamadne and jeanblanc [ _ math .
oper .
res . _
* 32 * ( 2007 ) 182192 ] or elie and kharroubi [ _ statist .
probab .
lett . _
* 80 * ( 2010 ) 13881396 ] , we study a more natural scheme based on oblique projections .
we provide a control on the error of the algorithm by introducing and studying the notion of multidimensional discretely reflected bsde . in the particular case where the driver does not depend on the variable @xmath0 ,
the error on the grid points is of order @xmath1 , @xmath2 . , and . |
the nernst law or third law of thermodynamics comes in two versions . the weak version , which states that zero temperature can only be reached asymptotically , is uncontroversial . in constrast , there is an ongoing discussion about the status of the strong version , originally formulated by planck , which states that entropy goes to zero at zero temperature .
see for example @xcite for contrasting views on this .
with regard to gauge / gravity duality , there is a natural tension between condensed matter systems , where the strong version is believed to apply generally or at least generically , and bps and other extremal black hole solutions , where a regular , and hence normally finite , horizon is associated with a finite , and typically large entropy .
this raises the question whether and how systems obyeing the strong version of the nernst law can be modelled by gravitational counterparts .
extremal black brane solutions obeying the strong version of the nernst law have been found in a variety of theories @xcite , including four - dimensional fi - gauged supergravity @xcite , where they were dubbed nernst branes .
more recently , a two - parameter family of nernst branes parametrized by temperature @xmath4 and a chemical potential @xmath5 was found in @xcite .
asymptotically , these solutions approach hyperscaling violating lifshitz ( ` hvlif ' ) geometries both at infinity and at the horizon , and therefore are interesting in the context of gauge / gravity duality with hyperscaling violation @xcite , see also @xcite and references therein . four - dimensional nernst branes share the typical problems of hvlif geometries , in that they exhibit curvature singularities @xcite . moreover the scaling properties of the geometry at infinity suggested an entropy temperature relation of the form @xmath6 , while the high - temperature asymptotics extracted from the equation of state was found to be @xmath7 . since in addition the scalars showed runaway behaviour at infinity , and the relation @xmath6 is valid for ads@xmath0 , it was conjectured in @xcite that the inconsistent uv behaviour of four - dimensional nernst branes signals a dynamical decompactification , and that the above problem would be cured by lifting the solutions to five dimensions .
this follows the general idea that scale covariant vacua can be obtained by dimensional reduction of scale invariant vacua @xcite .
in this paper we will verify this proposal and study the relation between five - dimensional and four - dimensional nernst branes in detail .
the five - dimensional nernst branes will be constructed within fi - gauged minimal five - dimensional supergravity with an arbitrary number of vector multiplets , by dimensional reduction to an effective three - dimensional euclidean theory and using the special geometry of the scalar sector .
we will show that the singular asymptotic behaviour of four - dimensional nernst branes is a compactification artefact , and that the ground state geometry is ads@xmath0 . a crucial role in understanding the relation between five- and four - dimensional nernst branes
is played by the compactification circle , whose size changes dynamically along the direction transverse to the brane .
the behaviour of this circle also solves another puzzle , namely the origin of the four - dimensional chemical potential .
five - dimensional nernst branes turn out to be boosted ads schwarzschild branes , depending on two continuous parameters , the temperature @xmath4 , and the linear momentum @xmath8 .
since momentum turns into electric charge upon dimensional reduction , one might naively expect that four - dimensional nernst branes depend on one continuous parameter , temperature @xmath4 and on one discrete parameter , electric charge @xmath9 .
however , the solutions of @xcite depend on an additional continuous parameter , the chemical potential @xmath5 .
as it turns out , its origin can be traced to the fact that the compactification circle grows towards infinity and towards the horizon , and has a minimum in between .
this minimum introduces a new scale , and since the minimal value of the radius can be varied continuously , this provides an additional continuous parameter .
the boosted ads schwarzschild metric we obtain by solving the five - dimensional equations of motion is an einstein metric and can be brought to carter - novotn - horsk form .
such metrics describe the near horizon regions of dimensionally reduced d - branes and m - branes with superimposed pp - waves @xcite .
this does , however , not immediately provide us with a string theory embedding of our solutions , unless we switch off the vector multiplets .
the solution we find is valid for an arbitrary number of vector multiplets , and depends on the choice of the prepotential and on the choice of an fi - gauging through parameters @xmath10 and @xmath11 .
while the scalars are constant , they still have to extremize the scalar potential , and therefore these parameters determine the effective cosmological constant , and enter into the various integration constants of our solution
. we will give explicit expressions in the paper .
fi - gauged five - dimensional @xmath12 supergravity has so far only been obtained as a consistent truncation of a higher - dimensional supergravity in a very limited number of cases .
the case without vector multiplets , that is pure gauged five - dimensional @xmath12 supergravity , can be obtained by reduction of iib supergravity on sasaki - einstein manifolds @xmath13 @xcite . the stu - model and consistent truncations thereof
can be obtained as consistent reductions of eleven - dimensional supergravity @xcite .
other consistent truncations involve hypermultiplets or massive vector multiplets and consequently have different types of gauging @xcite .
the dimensionally reduced boosted d3-branes of @xcite which lead to the same five - dimensional metric should be considered as solutions of five - dimensional gauged @xmath14 supergravity , which can be obtained by reduction of iib supergravity on @xmath15 . in this case
the five - dimensional cosmological constant is simply determined by the d3-charge , and we can not account for the parameters @xmath16 of an fi - gauged supergravity theory with vector multiplets .
but while there is no obvious string theory embedding of our solutions , the five - dimensional metric is still the same as for boosted d3-branes .
therefore it is reasonable to assume that at least the same type of dictionary between geometry and field theory will apply .
we will come back to this in the conclusions . throughout the paper we keep a strictly five - dimensional perspective and work with relations between geometric and thermodynamical quantities without using any higher - dimensional or stringy input .
our work includes a detailed study of the thermodynamical properties of five - dimensional nernst branes .
using the quasilocal energy momentum tensor we construct expressions for the mass @xmath17 and momentum @xmath8 . from the near horizon behaviour of the solution we obtain the entropy @xmath18 and through the surface gravity of the killing horizon , the temperature @xmath4 .
we verify the validity of the first law , as well as the strong version of the third law .
the solution is shown to be thermodynamically stable .
we obtain an equation of state and show that the relation between entropy and temperature interpolates between @xmath6 at high temperature and @xmath19 at low temparture .
this asymptotic behaviour agrees with the literature on boosted d3-branes@xcite and verifies the prediction of @xcite .
one subtlety is that the metric admits a reparametrisation , which naively removes the integration constant corresponding to the temperature ( as long as temperature is non - zero ) from the solution .
however , as the detailed analysis shows , when properly setting up thermodynamics using the quasi - local stress energy tensor , temperature is defined by a reparametrisation invariant expression .
the additional input that thermodynamics requires is the choice of the norm of the static killing vector field , which should be considered as part of the choice of the ads@xmath0 groundstate .
besides the trivial extremal limit , which is global ads@xmath0 , the solution admits a non - trivial double scaling limit , where temperature goes to zero , and the boost parameter goes to infinity while the momentum ( density ) is kept fixed .
this limit was studied ( in different coordinates ) in @xcite , where it was shown to result in a homogenoeus einstein space of kaigorodov type , which is 1/4 bps .
this analysis applies to our solution and implies that it supports 2 out of a maximum of 8 killing spinors . in the extremal limit
we also recover the five - dimensional extremal nernst branes of @xcite .
there are interesting parallels as well as differences between boosted ads schwarzschild black branes and rotating black holes .
like for a kerr black hole , boosted branes have an ` ergoregion ' that is a region before the event horizon where observers can not stay static any more , but have to co - translate with the brane .
also , the euclidean continuation of such a brane looks very similar to that of a kerr black hole , and allows to derive the temperature by imposing the absence of a conical deficit . in other aspects
the analogy breaks down , however .
while supersymmetric rotating black holes can not have an ergosphere @xcite , the ergoregion of an infinitely boosted black brane remains .
we show that this is consistent with supersymmetry , because the killing vector obtained as a killing spinor bilinear is null rather than timelike
. since we are interested in how the lift to five dimensions affects the curvature singularities of four - dimensional nernst branes , we work out explicit expressions for the five- and four - dimensional curvature in our preferred coordinate systems .
part of these results have been obtained in the previous literature , and where results can be compared , we find agreement .
our own contribution is to explicitly demonstrate how singular four - dimensional asymptotic hvlif metrics , when combined with the four - dimensional scalars encoding the dynamics of an additional compact direction , lift consistently to a geometry asymptotic to ads@xmath0 .
this does not only show that the hvlif singularities are artefacts resulting from , as one might say , a ` bad slicing ' of ads@xmath0 , but the mechanism is dynamical in the sense that the run - away of the four - dimensional scalars encodes the decompactification of the fifth dimension .
this removes all the sp curvature singularities and run - away behaviour of scalars that four - dimensional nernst branes exhibit at asymptotic infinity .
in addition , four - dimensional nernst branes also have pp curvature singularities which lead to infinite tidal forces acting on freely falling observers .
these occur at the horizon , and only at zero temperature .
they are again accompanied by run - away behaviour of the scalar fields , which encodes the dynamical decompactification of a fifth dimension .
but in contrast to what happens at asymptotic infinity under decompactification , the pp singularity of the asymptotic hvlif space @xcite is not removed but lifted to the pp singularity of a five - dimensional kaigorodov - type space - time @xcite .
this shows that pp singularities and infinite tidal forces are intricately related to the vanishing of the entropy ( density ) , thus bringing us back full circle to the third law .
we will continue this discussion in the final section . in section 2 we review five - dimensional @xmath20 fi gauged supergravity with vector multiplets and its dimensional reduction to three euclidean dimensions . in section 3 we obtain five - dimensional nernst branes by solving the three - dimensional effective equations of motion and lifting the solution back to five dimensions .
we solve the full second order equations of motion but observe that imposing regularity conditions reduces the number of parameters by one half , so that the solution will satisfy a unique set of first order equations , despite being non - extremal . by a coordinate transformation
the solution can be brought to the form of a boosted ads schwarzschild black brane , and further to a metric of carter - novotn - horsk type .
we work out the thermodynamics in full detail , investigate the extremal limit , compare geometrical properties to those of rotating black holes , and analyse the behaviour of curvature . in section 4 we perform a reduction to four dimensions and show that we recover the four - dimensional nernst branes of @xcite .
the relations between the geometrical and thermodynamical properties of five - dimensional and four - dimensional nernst branes is worked out in detail . in section 5
we interpret the results , obtain a consistent picture which ties together five- and four - dimensional nernst branes , and discuss its interpretation in the context of the gauge / gravity correspondence .
we also come back to the question of a higher - dimensional string theory embedding , and use the fact that our solutions have the same metric as reduced boosted d3-branes to set up a gauge / gravity dictionary .
we briefly explain how the method used in this paper to generate solutions can be applied to find more general solutions in the future .
finally we discuss open questions regarding the fate of pp curvature singularities and of the third law .
various technical details have been relegated to appendices .
appendix a derives certain re - writings of the scalar potential , which are used in the main text .
appendix b contains the details of computing thermodynamic quantities using the quasi - local energy - momentum tensor .
while in section 3 the hawking temperature is obtained from the surface gravity of the killing horizon , appendix c presents an alternative derivation using the euclidean approach .
this allows to again compare boosted branes to rotating black holes .
appendices d and e give the details for computing tidal forces in five and in four dimensions , respectively .
these details have been included to give , in combination with the main text , a full and self - contained account of curvature in five and four dimensions . in appendix
f we spell out the details of a ` well known ' fact about the normalization of vector potentials , for completeness , and because we are not aware of an easily digestible and sufficiently detailed explanation in the literature .
we start with the five - dimensional lagrangian for @xmath20 gauged supergravity coupled to @xmath21 vector multiplets @xcite .
our conventions for the ungauged sector follow those of @xcite , albeit with the opposite sign for the einstein - hilbert term : @xmath22 with @xmath23 . here
@xmath24 are five - dimensional lorentz indices , while @xmath25 label the five - dimensional gauge fields .
we use a formulation of the theory where the @xmath21-dimensional scalar manifold @xmath26 is parametrised by @xmath27 scalar fields @xmath28 which are subject to real scale transformations .
this formulation is natural in the context of the superconformal calculus and will turn out to be helpful for finding solutions .
the construction of the theory of five - dimensional vector multiplets coupled to supergravity using the superconformal calculus can be found in @xcite , while the superconformal method in general is reviewed in @xcite .
we will in addition use the formulation of special real geometry developed in @xcite . as explained in more detail in @xcite , the scalars @xmath28 are special coordinates on an open domain @xmath29 , which is invariant under the action of the group @xmath30 by scale transformations .
the manifold @xmath31 is the scalar manifold of an auxiliary theory of @xmath27 superconformal vector multiplets , from which a theory of @xmath21 vector multiplets coupled to poincar supergravity is obtained by gauge fixing .
@xmath32 is a so - called conic affine special real ( casr ) manifold .
this means that it carries a hessian metric which transforms with weight 3 under the @xmath33-action .
when choosing special coordinates , which are affine coordinates with respect to the hessian structure and transform with weight 1 under scale transformations , the hesse potential is a homogeneous cubic polynomial , @xmath34 . in this way
one recovers the original definition of @xcite .
the physical scalar manifold of the supergravity theory can be identified with the hypersurface @xmath35 defined by @xmath36 note that the @xmath30-action is transverse to this hypersurface , so that we can identify @xmath37 .
this is a real version of the superconformal quotients for four - dimensional vector multiplets and for hypermultiplets .
the manifold @xmath38 will be referred to as a projective special real ( psr ) manifold . in the lagrangian ( [ 5dlagr ] )
we use the special coordinates @xmath28 , but it is understood that the constraint ( [ hs ] ) has been imposed . within the superconformal calculus
this constraint is the ` d - gauge ' which gauge fixes the local dilatations of the auxiliary superconformal theory in order to obtain the associated poincar supergravity theory in its conventional form . in ( [ 5dlagr ] ) this is reflected by the einstein - hilbert term having its dimension - full prefactor @xmath39 , rather than being multiplied by a conformal compensator to make it scale invariant . in ( [ 5dlagr ] )
we have chosen to express both the scalar and the vector couplings using the symmetric , positive definite tensor field @xmath40 here we use a notation which suppresses indices which are summed over : @xmath41 , @xmath42 , etc . the tensor @xmath43 is a positive definite hessian metric with hesse potential @xmath44 on @xmath32 . while it is different from the conical hessian metric @xmath45 , which has lorentz signature , with the negative eigendirection along the orbits of the @xmath33-action , the pullbacks of both metrics to the hypersurface @xmath38 agree , so that one can use either to obtain the positive definite metric @xmath46 which encodes the self - couplings of the @xmath21 physical scalars .
the couplings of the @xmath27 physical vector fields are given by the restriction of the positive definite metric @xmath47 to @xmath38 .
the scalar potential @xmath48 in results from an fi - gauging parametrized by @xmath27 gauging parameters @xmath11 . using the expressions of @xcite and @xcite we find @xmath49 g_i g_j.\ ] ]
we have fixed a convenient normalisation of the gauging parameters @xmath11 by comparing the dimensional reduction of to the four - dimensional scalar potential of @xcite , evaluated for a `` very special prepotential '' @xmath50 that is a prepotential which can arise by reduction from five to four dimensions .
, @xmath51 , and @xmath52 . ]
we remark that while on the physical scalar manifold we have to impose the constraint @xmath53 , we have kept factors of @xmath54 explicit in and .
this is useful in keeping track of the scaling weights of fields , and thus checking that expressions are consistent with the scaling properties of the corresponding gauge - equivalent superconformal theory .
we have chosen our conventions such that the scaling weights of the fields used in ( [ 5dlagr ] ) match with @xcite .
in particular , we take @xmath55 as a quick check , note that the lagrangian has scaling weight 5 , so that the resulting action has scaling weight 0 .
further , provided that we include the appropriate factors of @xmath54 , the functions @xmath47 and @xmath56 are homogeneous in @xmath28 , even in presence of the dimension - full factors @xmath57 and @xmath11 , which appear after imposing d - gauge .
note that throughout the remainder of this paper , we shall set @xmath58 .
we now want to reduce the five - dimensional theory to three ( euclidean ) dimensions .
we make the metric ansatz @xmath59 where all fields depend only on the coordinates of the three - dimensional space .
in addition we choose to switch off all of the five - dimensional gauge fields @xmath60 , i.e. we look only for uncharged five - dimensional solutions .
the presence of the kaluza - klein one - form @xmath61 indicates that we are looking for non - static five - dimensional solutions . upon compactification of the @xmath62 circle
this will give rise to a non - trivial electric charge for the corresponding four - dimensional solution .
note that whilst the killing vector @xmath63 is always space - like in five dimensions , @xmath64 can be either time - like , space - like , or null , depending on the magnitude of @xmath65 .
however , after performing the dimensional reduction over @xmath62 , the @xmath66 direction will always be time - like in four dimensions , and so we are able to use the same dimensional reduction technique as in @xcite , i.e. we reduce over both a space - like and time - like direction .
the resulting three - dimensional action is given by @xmath67 where the three - dimensional scalar potential is given by @xmath68g_i g_j.\ ] ] in order to solve the equations of motion resulting from it turns out to be convenient to introduce the variables @xmath69 , @xmath70 and @xmath71 via @xmath72 so that the three - dimensional lagrangian becomes @xmath73 the scalar potential is given in terms of the new fields by @xmath74g_i g_j \nonumber \\ & = & 3\left[\hat{g}^{ij}(y)+4y^i y^j\right ] g_i g_j \ , .\end{aligned}\ ] ] the explicit steps used in getting to the second line are carried out in appendix [ app : scalpot ] .
we note that the lagrangian has no explicit dependence on the field @xmath70 appearing in the metric ansatz . this reflects the fact that when taking the rescaled scalar fields @xmath71 as independent variables , the field @xmath70 becomes dependent , and can be recovered from the equation @xmath75 which follows from the hypersurface constraint @xmath76 . in terms of the new fields
@xmath69 and @xmath70 , the five - dimensional metric ansatz becomes @xmath77 the independent three - dimensional variables are : the metric @xmath78 , the scalars @xmath71 which encode the @xmath21 independent five - dimensional scalars together with the kaluza - klein scalar @xmath70 , the second kaluza - klein scalar @xmath69 , and the scalar @xmath79 which is dual to the kaluza - klein vector from the reduction over @xmath66 .
the metric on the scalar submanifold parametrized by the @xmath71 , @xmath80 is , up to a constant factor , isometric to the positive definite hessian metric on the manifold @xmath81 .
as shown in @xcite this metric is isometric to the product metric @xmath82 on @xmath83 . from it
is manifest that the scalar manifold @xmath84 of the three - dimensional lagrangian carries a product metric , with the first factor parametrized by @xmath71 and the second factor parametrized by @xmath69 and @xmath79 . by inspection ,
the second factor is locally isometric to the metric of the pseudo - riemannian symmetric space @xmath85 , which can be thought of as the ` indefinite signature version ' of the upper half plane ( equivalently , of the unit disk ) @xmath86 . to be precise @xmath69 and @xmath79 parametrise an open subset which can identified with the iwasawa subgroup of @xmath87 , or , in physical terms , with the static patch of @xmath88 .
the combined scalar manifold parametrized by @xmath89 , @xmath90 has dimension @xmath91 .
if we perform the reduction of five - dimensional supergravity with @xmath21 vector multiplets to three euclidean dimensions without any truncation , then the resulting scalar manifold is a para - quaternionic khler manifold @xmath92 of dimension @xmath93 @xcite .
the submanifold @xmath84 is obtained by a consistent truncation and therefore it is a totally geodesic submanifold of @xmath92 .
we remark that @xmath84 is a ( totally geodesic ) submanifold of the @xmath94-dimensional totally geodesic para - khler manifold @xmath95 described in @xcite , @xmath96 it was shown there how to obtain explicit stationary non - extremal solutions of four- and five - dimensional ungauged supergravity by dimensional reduction over time .
as we will see in the following , it is still possible to obtain explicit solutions in the gauged case , where the field equations of the three - dimensional scalars are modified by a scalar potential .
while we will retrict ourselves to the submanifold @xmath84 in this paper , the higher dimensional para - khler submanifold @xmath95 will be relevant when the present work is extended to more general , charged solutions , including the solutions found in @xcite .
we now turn to the three - dimensional equations of motion coming from .
the equations of motion for @xmath97 and @xmath98 read : @xmath99 @xmath100 @xmath101 where we have introduced the christoffel symbols for the metric @xmath102 : @xmath103 meanwhile , the einstein equations read @xmath104g_i g_j = 0.\end{aligned}\ ] ] we now proceed to solving the equations of motion , and make the following brane - type ansatz for our three - dimensional line element : @xmath105 where @xmath106 is some function to be determined , and @xmath107 is a radial coordinate which parametrizes the direction orthogonal to the world - volume of the brane .
this is the same brane - like ansatz for the three - dimensional line element as in @xcite .
moreover we will impose that all of the fields @xmath71 , @xmath98 and @xmath69 depend only on @xmath107 .
this coordinate has been chosen such that it is an affine curve parameter for the curve @xmath108 on the scalar manifold @xmath84 .
the ricci tensor has components @xmath109 from which we find that the einstein equations become @xmath110 for @xmath111 , and @xmath112 for @xmath113 , where we have used .
we will now consider the equations of motion for each of @xmath98 , @xmath69 and @xmath71 in turn .
the equation of motion for @xmath98 can be brought to the form @xmath114 which is solved by @xmath115 for some integration constant @xmath116 , where we have chosen the factor for later convenience .
once we solve the equation of motion for @xmath69 we will further integrate to obtain an expression for the kaluza - klein vector @xmath117 appearing in the five - dimensional metric .
substituting in to the equation of motion for @xmath69 we find @xmath118 introducing the variable @xmath119 , this becomes @xmath120 by differentiation we obtain the necessary condition @xmath121 , which can be integrated to @xmath122 , where @xmath123 is a real constant .
would yield a solution periodic in @xmath107 , which we discard . ] parametrizing the general solution as @xmath124 with arbitrary constants @xmath125 , and substituting back into the original equation we find the constraint @xmath126 which imposes one relation between the four constants @xmath127 .
it will turn out to be useful in what follows to consider @xmath128 , @xmath123 and @xmath129 to be the independent quantities , and to write everything in terms of these . in particular
, we then have @xmath130 .
we are also now in a position to further integrate , which we write as @xmath131 + for simplicity we will chose the negative sign in , and will not carry through the corresponding positive solution . since @xmath98 is dual to a kaluza - klein vector , this means that we have fixed the sign of the ` charge ' that the solution carries . substituting in and integrating ,
we find @xmath132-{\zeta}^0_\infty,\ ] ] for some integration constant @xmath133 , which can be fixed by imposing a suitable physicality condition on the solution . at this point
we anticipate that a horizon , if it exists , will be located at @xmath134 .
moreover , as we will show in section [ section:4d ] , upon dimensional reduction we obtain a four - dimensional stationary ( in fact static ) solution with a killing horizon .
such horizons admit , for finite temperature , an analytic continuation to a bifurcate horizon @xcite . in order
that the four - dimensional one - form @xmath135 is well defined , it must vanish at the horizon @xcite , see also appendix [ vanishing ] .
this fixes @xmath136 and therefore the kaluza - klein one - form is given by @xmath137 the equation of motion for the @xmath71 becomes @xmath138 to proceed , we first contract with the dual scalar fields @xmath139 and make use of the identity @xmath140 which follows from the fact that @xmath102 is homogeneous of degree @xmath141 in the @xmath71 .
we thus find @xmath142 which upon using becomes @xmath143 given that @xmath144 , we can integrate to find @xmath145 for some integration constant @xmath146 , where the factor has been chosen for later convenience .
writing @xmath147 we can integrate further to obtain @xmath148 for an integration constant @xmath149 . again
the prefactor has been chosen for later convenience .
we now return to the hamiltonian constraint . using and this becomes : @xmath150 we then have the following picture .
the solutions @xmath151 to should satisfy the constraints and constraint .
one way to proceed , which is valid for generic five - dimensional models , is to set all of the @xmath71 proportional to one another , i.e. we put @xmath152 for some constants @xmath153 , which satisfy @xmath154 note that since the ( constrained ) scalar fields @xmath28 can be recovered from the @xmath71 via @xmath155 , we see that this ansatz will result in constant five - dimensional scalar fields .
using we obtain : @xmath156 eliminating the quantity @xmath157 from we obtain an equation for the function @xmath158 : @xmath159 this is precisely the same equation as was found in @xcite , and so can be solved in the same way by @xmath160 for some integration constants @xmath161 and @xmath162 , where the quantity @xmath163 is given by @xmath164 from this , we can integrate to find @xmath165 for some constant @xmath166 , and hence the @xmath71 are given by @xmath167 where we have defined @xmath168 .
we finally need to ensure that the solution satisfies the original equations of motion .
this fixes @xmath169 in terms of the gauging parameters @xmath11 and other integration constants as @xmath170 therefore the function @xmath70 appearing in the line element is given by @xmath171 the signs in should be chosen such that the function @xmath172 is real and positive for all @xmath173 . at this stage
we have six independent integration constants @xmath174 which are _ a priori _ yet to be determined .
however , following @xcite we choose to set @xmath175 in what follows so that the asymptotic region is at @xmath176 and the near horizon region at @xmath177
. we can then scale @xmath107 to set @xmath178 . in order for our solution to make sense as a black brane in five dimensions , we need to impose some physicality constraints . in particular , we require that the five - dimensional solution generically has a finite entropy density .
is infinite , the entropy itself will diverge .
by ` generic ' we mean that we allow that the solution has a limit , which hopefully will coincide with the zero temperature limit , where the entropy density becomes zero . ] combining the five - dimensional and three - dimensional metric anstze and we see that finite entropy density corresponds to a finite value of @xmath179 as @xmath177 ( i.e. at the horizon ) . to leading order
we find @xmath180 in order that this be finite and non - zero we therefore require @xmath181 which , given , is equivalent to @xmath182 , further resulting in @xmath183 . hence , the physicality constraint further reduces the number of independent integration constants by one . before moving on to study properties of the solution , we summarise the story so far . the functions appearing in the five - dimensional line element are given by @xmath184 whilst the scalar fields @xmath28 parametrising the casr manifold are constant and given by @xmath185 we have therefore found a family of solutions to the equations of motion depending on three non - negative parameters @xmath186 .
since the field equations for the three - dimensional scalars @xmath187 are of second order , and our ansatz amounts to three independent scalar fields ( since the @xmath71 have been taken to be proportional ) , we should a priori have expected six independent integration constants . however , as we have seen , physical regularity conditions imposed on the lifted , higher - dimensional solution reduces the number of integration constants by one half .
this is consistent with physical solutions being uniquely characterised by a system of first order flow equations , despite that the equations of motion are of second order , as has been observed for other types of solutions before @xcite .
we further remark that since the physical five - dimensional scalar fields have turned out to be constant , their only contribution is to generate an effective cosmological constant , whose value is determined by the value of the scalar potential at the corresponding stationary point . since no five - dimensional gauge fields have been turned on , our solution , which is valid for any five - dimensional vector multiplet theory , can therefore be obtained from an effective action , which only contains the einstein - hilbert term together with a cosmological constant , while the gauge fields and scalar fields have been integrated out .
we introduce a new ` radial ' ( more accurately : transversal ) coordinate @xmath188 via @xmath189 so that the near horizon region is at @xmath190 , and the asymptotic region is at @xmath191 .
hence we can use @xmath188 to analytically continue the solution to the region @xmath192 beyond the horizon . in terms of @xmath188
we find @xmath193 where we have defined @xmath194 .
moreover , we have @xmath195 and @xmath196 introducing the notation @xmath197 the five - dimensional line element becomes @xmath198 let us now turn to an investigation of the properties of the solutions constructed in section [ ssec : eoms ] , which we recall depend on three independent parameters : @xmath128 , @xmath123 and @xmath199 .
it is instructive to look at the cases @xmath200 and @xmath201 separately .
moreover , we focus first on the situation @xmath202 , and will comment on the @xmath203 case later . in this situation
it is convenient to introduce the notation : @xmath206 after a suitable scaling of the boundary coordinates , and introducing the new radial coordinate @xmath207 , we can bring the five - dimensional line element to the form @xmath208 here @xmath209 is defined by @xmath210 and , as we will see below , is the radius of an asymptotic @xmath211 space , whilst @xmath212 in order to interpret our solution , as well as to read off the various thermodynamic quantities associated with it , it is useful to introduce coordinates in terms of which the line element becomes manifestly asymptotically @xmath211 .
we observe that the solution is invariant under the combined rescalings @xmath213 of parameters , with @xmath123 invariant .
note that @xmath214 is invariant , so that for @xmath205 we obtain a two - parameter family of solutions parametrized by @xmath123 and @xmath214 .
the coordinate transformation @xmath215 absorbs @xmath128 and brings the metric to the form of a boosted ads schwarzschild black brane : @xmath216 .\ ] ] the constants @xmath217 satisfy @xmath218 and parametrise a boost along the @xmath219-direction . by taking
@xmath220 one sees that indeed asymptotes to @xmath211 with radius @xmath209 .
the constant @xmath214 parametrizes the boost of the brane , while @xmath123 , as we will show below , is a non - extremality parameter and therefore related to temperature .
this metric can be rewritten by making the following co - ordinate transformation : @xmath221 to obtain @xmath222 this metric is the 5-dimensional generalized carter - novotn - horsk metric constructed in @xcite .
we further remark that the line element ( [ 5dmetric4 ] ) can be further simplified by setting @xmath223 , @xmath224 , @xmath225 , @xmath226 , @xmath227 .
this rescaling corresponds to formally setting @xmath228 in the function @xmath229 in ( [ 5dmetric4 ] ) , thus fixing the coordinate of the horizon to @xmath230 .
however , this reparametrization obscures the fact that @xmath231 in ( [ 5dmetric4 ] ) encodes the temperature , which , as we will show later , is defined in a reparametrization invariant way .
let us now look at the case where we take @xmath201 , so that @xmath232 in .
in this case , after suitably rescaling the boundary coordinates and introducing the radial coordinate @xmath233 as before , we find that the five - dimensional line element becomes @xmath234 making the coordinate redefinition @xmath235 we can bring the metric to the form of a boosted ads schwarzschild black brane .
the boost parameters are given by @xmath236 where the quantity @xmath237 is defined via @xmath238 as in the case @xmath205 we obtain a two - parameter family of black brane solutions .
for @xmath201 the parameters can be taken to be @xmath123 ( equivalently @xmath231 ) and @xmath199 .
we remark that while both the cases @xmath205 and @xmath201 can be mapped to two - parameter families of black branes , both families can not be related smoothly by taking @xmath239 .
if we take @xmath240 in then the region @xmath241 contracts to @xmath242 , which suggests that this limit is the extremal limit .
we will show later that @xmath203 does indeed correspond to vanishing surface gravity , and , hence vanishing hawking temperature , and , moreover , that the solution is a bps solution .
for any value ( zero or non - zero ) of @xmath128 we can then bring the metric to the form @xmath243.\ ] ] this solution agrees with the five - dimensional extremal nernst branes found in @xcite . we can equivalently obtain this form of the metric from the boosted black brane by taking the limits @xmath244 in the extremal limit @xmath199 can be interpreted as a boost parameter .
the vacuum @xmath211 solution is obtained by taking the zero boost limit @xmath245 .
thus in the extremal limit @xmath199 determines the mass , or more precisely the mass per worldvolume or tension of the brane .
the precise expressions for the mass and thermodynamic quantities will be calculated in section [ subsec : thermodynamics5d ] .
the metric ( [ extmetric ] ) displays an interesting scaling behaviour in the limit @xmath246 . to display it ,
we introduce coordinates @xmath247 by these coordinates agree with @xmath62 and @xmath66 in the extremal limit .
moreover , the near - horizon limit preserves the symmetry that allows to set @xmath248 . ]
@xmath249 then the metric becomes @xmath250 \;.\ ] ] dropping terms which are subleading in the ` near horizon limit ' @xmath246 we obtain @xmath251 \;.\ ] ] this
metric is invariant under the scale transformations : @xmath252 thus the asymptotic metric shows a scaling invariance similar to a lifshitz metric with scaling exponent @xmath253 ( and no hyperscaling violation , @xmath254 ) .. ] the only difference is that the coordinate @xmath255 has scaling weight @xmath256 rather than @xmath257 .
this type of generalized scaling behaviour was observed in @xcite , where the metric ( [ nh_limit ] ) was obtained by taking a particular limit of boosted d3-branes .
we will come back to this in section [ conclusions ] , where we discuss the dual field theory interpretation of our solutions .
the boosted black brane has similarities with kerr - like black holes , with the linear momentum related to the boost playing a role analogous to the angular momentum .
it is instructive to work this out in some detail , following the discussion of the kerr solution in @xcite .
let us first look for the existence of static observers , who remain at constant @xmath258 and as such have velocities parallel to the killing vector field @xmath259 . therefore static observers exist in regions where @xmath259 is timelike , and the limit of staticity is at the value of @xmath233 where @xmath260 this `` ergosurface '' is always located outside the event horizon , with the trivial exception of globally static ( unboosted )
spacetimes for which @xmath261 and the two surfaces overlap completely .
this is different to the rotating case where ergosurface and event horizon always coincide at the north and south pole .
+ beyond the limit of staticity there still exist stationary observers which are co - moving ( more precisely , but less elegantly ` co - translating ' ) with the brane .
observers which have fixed @xmath262 and a constant velocity in the @xmath219-direction have world lines tangent to killing vector fields @xmath263 where the quantity @xmath264 will be referred to as the velocity .
such co - moving observers exist in regions where @xmath265 is time - like . killing vector fields of the form
@xmath265 become null for values of @xmath233 where @xmath266 thus there is a finite range of velocities @xmath70 , given by @xmath267 , which co - moving observers can attain .
note that at the limit of staticity , where @xmath268 , we find that @xmath269 . therefore @xmath70 must be negative once the limit of staticity has been passed .
the limit for co - moving observers is reached when @xmath270 , which happens at the point where @xmath271 it is straightforward to verify that this happens at the same value @xmath231 of @xmath233 where @xmath272 .
the limiting velocity @xmath273 is given by @xmath274 and can be interpreted as the boost - velocity of the surface @xmath275 . since @xmath272 implies that @xmath276 , it follows that on this surface outgoing null congruences have zero expansion , see @xcite for the analogous case of a rotating black hole .
consequently @xmath275 is an apparent horizon , and since the solution is stationary , an event horizon .
moreover this event horizon is a killing horizon for the vector field @xmath277 and we can interpret @xmath273 as the boost - velocity of this horizon . observe that the limit of staticity and the limit of stationarity are in general different , and only agree in the unboosted limit @xmath278 where we recover the ads schwarzschild black brane .
we note that there is frame dragging in our solutions , since the metric is non - static for @xmath279 .
indeed , since the metric coefficients are independent of @xmath280 and @xmath219 , the covariant momentum components @xmath281 and @xmath282 are conserved .
but even when setting @xmath283 , particles have a non - vanishing contravariant momentum component @xmath284 in the @xmath219-direction .
the boost velocity of the metric varies between the horizon and infinity .
it can be read off by writing the metric in the form @xmath285 where the omitted terms involve @xmath286 and @xmath287 .
an observer at fixed @xmath288 is co - moving with the space - time if their velocity is @xmath289 . bringing the metric to the above form one finds @xmath290 with limits @xmath291 { } - \frac{u_z}{u_t } = w \geq - 1\;,\ ] ] and @xmath292 { } 0 \;.\ ] ]
it is straightforward to check that for @xmath293 the boost speed @xmath294 is strictly monontonically increasing from @xmath295 at infinity to @xmath296 at the horizon .
thus the boost speed is bounded by the speed of light and can only reach it at the horizon and in the extremal limit .
note that the asymptotic ads space at infinity is not co - moving .
this is different from kerr - ads , where the asymptotic ads space is co - rotating , with implications for the black brane thermodynamics @xcite . in particular
, we will not need to subtract a background term , corresponding to the asymptotic ads space , from our expressions for the boost - velocity in order to have quantities satisfying the first law of thermodynamics .
we will come back to this later when verifying the first law . in this section ,
we consider the properties of the extremal solution in further detail .
we begin by considering the solution ( [ extmetric ] ) . on making the co - ordinate transformation @xmath297 the metric ( [ extmetric ] ) becomes @xmath298 the
metric ( [ kaig1 ] ) is a five - dimensional generalized kaigorodov metric , constructed in @xcite , which describes gravitational waves propagating in ads@xmath299 .
the supersymmetry of this solution was investigated in @xcite , where it was shown that this solution preserves 1/4 of the supersymmetry .
furthermore , after making some appropriate co - ordinate transformations , this solution can be shown to correspond to a class of supersymmetric solutions which appears in the classification of supersymmetric solutions of minimal five - dimensional gauged supergravity constructed in @xcite .
it is straightforward to show that the null killing vector which is obtained as a spinor bilinear is given by @xmath300 , in the co - ordinates of ( [ extmetric ] ) .
the fact that this killing vector is null rather than timelike is related to an interesting feature which distinguishes these bps solutions from five - dimensional rotating bps black holes , namely the existence of an ergoregion , i.e. a region outside the horizon where it is not possible for observers to remain static .
note that in the bps limit ( [ extlimit ] ) the limit of staticity is at @xmath301 which is outside the horizon at @xmath242 ( unless we switch off momentum , @xmath302 , and go to global ads ) .
therefore the ergoregion persists in the bps limit . for stationary bps black holes the killing vector obtained as a spinor bilinear is the standard static killing vector field @xmath259 , which is timelike at infinity . for rotating black holes @xmath259
is different from the ` horizontal ' killing vector field @xmath303 , which becomes null on the horizon .
an ergoregion exists when @xmath259 becomes space - like outside the horizon .
however if @xmath259 is a bilinear formed out of killing spinors , supersymmetry implies that it must be either time - like or null .
hence , rotating bps black holes can not have an ergoregion .
moreover it can be shown that the event horizon of a rotating bps black hole must be non - rotating @xcite . as we have shown above , this is different for the extremal limit of an ads - schwarzschild black brane , which is a bps wave solution in ads@xmath0 : the ergoregion persists in the bps limit , and the ( degenerate limit of the ) horizon moves with the speed of light , since @xmath304 .
this is consistent with the solution being bps , because the killing vector obtained as a killing spinor bilinear is not @xmath259 , which is timelike at asymptotic infinity and becomes spacelike before the horizon , but @xmath305 , which is null everywhere for the bps solution .
moreover , the horizon turning into a purely left - moving wave is consistent with the familiar string theory description of a bps state as a state with massless excitations moving in one direction only .
we turn to an investigation of the thermodynamics of the black brane solutions of section [ ssec : properties ] .
the hawking temperature is related to the surface gravity by @xmath306 , where the surface gravity @xmath57 of a killing horizon is given by @xmath307 evaluating this for @xmath308 , we find the hawking temperature @xmath4 : @xmath309 we remark that the same result can be obtained by imposing that the euclidean continuation of the solution does not have a conical singularity at the horizon , see appendix [ app : euclideanisation ] . in the zero boost limit @xmath261 , @xmath278
, we obtain the hawking temperature of an ads schwarzschild black brane . in the extremal limit ( [ extlimit ] ) , where the boost parameters go to infinity @xmath310 , while @xmath311 , the hawking temperature becomes zero , @xmath312 , irrespective of whether we keep @xmath199 finite or not .
since our solutions are not asymptotically flat , but rather asymptotic to ads@xmath0 , we can not apply the standard adm prescription to compute the mass and linear momentum of our branes .
instead , we use the method based on the quasilocal stress tensor @xcite , see also @xcite for a review in the context of the fluid - gravity correspondence . here
we simply present the result , and relegate explicit calculational details to appendix [ app : fluidgrav ] . to leading order in @xmath313
we find that the quasilocal stress tensor takes the form @xmath314 where @xmath315 is the minkowski metric on @xmath316 which denotes the hypersurface @xmath317 of our space - time , with coordinates @xmath318 .
as indicated we have omitted terms subleading in @xmath313 , since we are ultimately interested in expressions which are finite on the boundary @xmath319 of space - time .
note that @xmath320 takes the form of the stress energy tensor of a perfect ultra - relativistic fluid ( equation of state @xmath321 , where @xmath188 is the energy density and @xmath322 is the pressure ) , with pressure proportional to @xmath323 .
the proportionality between @xmath231 and @xmath4 is the same behaviour as for large ads - schwarzschild black holes . in the absence of a boost
, it is known that ads - schwarzschild black branes behave thermodynamically like large ( rather than small ) ads - schwarzschild black holes @xcite .
having obtained the quasilocal stress tensor , mass and linear momentum can be computed as conserved charges associated to the killing vectors of our solution .
again , the details are relegated to the appendix [ app : fluidgrav ] .
the mass , which is the conserved charge associated with time translation invariance , is @xmath324 where @xmath325 is the spatial volume of the brane , computed with the standard euclidean metric @xmath326 . due to the infinite extention of the brane ,
the mass is infinite , and to obtain a finite quantity we must either compactify the world volume directions or define densities .
we will do the latter by consistently splitting off a factor @xmath327 from all extensive quantities .
next we calculate the momentum in the @xmath219-direction , which is the conserved charge associated to @xmath219-translation invariance .
the result is @xmath328 and vanishes as expected in the zero boost limit @xmath278 , @xmath261 .
notice that these charges satisfy @xmath329 , which resembles the motion of a non - relativistic body of mass @xmath17 , moving at velocity @xmath330 .
finally , we calculate the bekenstein - hawking entropy of the solution by integrating the pull back of the metric over the horizon . recalling that we are working in units where @xmath331
, we find @xmath332 where @xmath333 denotes the pullback of the metric to the surface @xmath334 .
using these , we can check that the thermodynamic variables satisfy the first law : @xmath335 we remark that obtaining is a non - trivial consistency check for the correctness of the definition of the thermodynamical quantities , which are initially ambiguous because they require background subtractions corresponding to renormalization of the boundary cft @xcite , see also @xcite for a discussion in the context of rotating black holes in higher than four dimensions . as noted before we do not need to apply a background subtraction for the translation velocity @xmath273 , since the asymptotic ads@xmath0 background is not co - translating .
this is different for ads - kerr - type black holes , where the subtraction of the background rotation velocity is crucial for obtaining the correct thermodynamic relations @xcite .
we also note that @xmath336 which are all defined in a reparametrization invariant way , depend on the parameter @xmath231 , which is therefore a physical parameter , despite the fact that it could be absorbed into the coordinates in the line element ( [ 5dmetric4 ] ) .
moreover , without the ability of varying this parameter , one could not obtain the temperature / entropy term in the first law .
we refer to appendix [ mmcc ] for further details on this technical point .
the extremal limit of these quantities can be reached by taking @xmath311 and @xmath337 with @xmath338 fixed . in this case
we find that the entropy density @xmath339 vanishes in the extremal limit , @xmath340 as @xmath341 .
therefore our solutions satisfy the strong version of the nernst law , and will be referred to as nernst branes .
moreover , since in this case @xmath342 , we find @xmath343 , which is of course the saturation of the bps bound , as it must be given the results of section [ subsection : bps ] .
as already remarked earlier , in the extremal limit the boost parameter @xmath199 controls the mass , and @xmath344 is the limit where the solution becomes globally ads@xmath0 .
we can eliminate the quantities @xmath231 and @xmath345 in favour of the thermodynamical variables @xmath4 and @xmath273 via @xmath346 in terms of @xmath4 and @xmath273 the mass of the solution is given by @xmath347 hence , we see that the heat capacity @xmath348 is positive , and the solution is thermodynamically stable .
this is as expected , at least in the absence of a boost , since it is well known that ads - schwarzschild black branes behave thermodynamically like large ads - schwarzschild black holes @xcite .
as we see from , the introduction of a boost does not introduce thermodynamic instablility . expressing the entropy in terms of @xmath349 we find @xmath350 note that turning off the boost @xmath278 , which corresponds to @xmath351 , we have @xmath6 , which is the scaling behaviour expected for an @xmath352 schwarzschild black brane
. indeed we can use to investigate the behaviour of @xmath18 as a function of @xmath4 in both the high temperature and low temperature limits .
the limit of high temperature ( equivalently small boost velocity ) is @xmath353 this corresponds to @xmath354 , and so we see from that @xmath6 .
the limit of low temperature ( equivalently boost velocity approaching the speed of light ) is the extremal limit @xmath355 in this case , one can see that @xmath356 , and so the entropy scales like @xmath19 .
this is the behaviour predicted for five - dimensional lifts of four - dimensional nernst branes @xcite .
we will comment further on the thermodynamic properties of our solutions in section [ conclusions ] .
one motivation of the present work is the singular behaviour of the four - dimensional nernst branes found in @xcite .
we will show in section [ section:4d ] that the five - dimensional nernst branes found above are dimensional lifts of these four - dimensional nernst branes . to investigate the effect of dimensional lifting on such singularities
, we now examine the behaviour of curvature invariants and tidal forces of the five - dimensional solutions . from both
the gravitational point of view , and with respect to applications to gauge - gravity dualities , one would like the solutions to have neither naked singularities , nor null singularities ( singularities coinciding with a horizon ) , while the presence of singularities hidden behind horizons is acceptable . in practice , the presence of large curvature invariants or large tidal forces will also be problematic , given that the supergravity action we start with needs to be interpreted as an effective action .
therefore large curvature invariants or tidal forces are indications that this effective description breaks down due to quantum or , assuming an embedding into string theory , stringy corrections .
this might also limit the applicability of gauge - gravity dualities to only part of the solution , where the corrections remain sufficiently small . for our five - dimensional metric
we compute the kretschmann scalar and ricci scalar to be @xmath357 note that these only depend on the temperature @xmath358 and the curvature radius @xmath209 of the ads@xmath0 ground state .
for the extremal solution ( @xmath359 ) both curvature invariants take constant values which agree with those for global @xmath352 : @xmath360 for the non - extremal solution the curvature invariants tend to the @xmath211 values asymptotically , but will blow up as @xmath361 . since this is behind the horizon , there are no naked or null singularities related to the curvature invariants of five - dimensional nernst branes .
even if all scalar curvature invariants are finite , there might still be curvature singularities related to infinite tidal forces .
such curvature singularities can be found by computing the components of the riemann tensor in a ` parallely - propogated - orthonormal - frame ' ( ppon ) associated with the geodesic motion of a freely - falling observer .
following @xcite they are called pp singularities , in contrast to sp singularities , where a scalar curvature invariant becomes singular .
while such singularities are often considered milder than those associated to curvature invariants , they are nevertheless genuine singularities and have drastic physical effects ( ` spaghettification ' ) on freely falling observers . the details of this construction for the five - dimensional extremal solution
are relegated to appendix [ app:5dtidalforces ] .
we only need to consider the extremal solution , since non - extremal solutions are manifestly analytic at the horizon @xmath362 . from table
[ riemanntable ] in appendix [ app:5dtidalforces ] we observe that the non - zero components of the riemann tensor in the ppon all have near horizon behaviour of the form @xmath363 with @xmath364 for all but one independent non - vanishing component .
hence , as the observer approaches the horizon of the extremal brane ( @xmath361 ) these components will diverge , resulting in infalling observers being subject to infinite tidal forces .
this is the same behaviour as observed in four dimensions @xcite , and seems to be the price for having zero entropy .
it is an interesting question whether stringy or other corrections could lift this singularity , and if so , whether it is possible to maintain zero entropy .
we now want to dimensionally reduce our five - dimensional nernst branes and compare the resulting four - dimensional spacetimes to those found in previous work @xcite .
let us therefore review the relevant features , emphasising the problems that we want to solve .
four - dimensional nernst branes depend on three parameters : the temperature @xmath4 , the chemical potential @xmath5 and one electric charge @xmath9 . due to
dirac quantisation charge is discrete , and the solution depends on two continuous parameters .
the asymptotic geometries , both at infinity and at the horizon , are of hyperscaling violating lifshitz ( hvlif ) type : @xmath365 here @xmath280 is time , @xmath233 parametrizes the direction transverse to the brane , and @xmath366 , @xmath367 are the directions parallel to the brane ( with @xmath368 in @xmath116 spacetime dimensions ) . for @xmath254
the line element ( [ hvlif ] ) is invariant under rescalings @xmath369 the parameter @xmath219 , which measures deviations from ` relativistic symmetry ' ( due to time scaling different from space ) is called the lifshitz exponent . for @xmath370
the metric is not scale invariant but still scales uniformly , and @xmath371 is known as the hyperscaling violating exponent . for four - dimensional nernst
branes the geometry at infinity is independent of the temperature . for finite chemical potential it takes the form of conformally rescaled ads@xmath372 , which is of the above type , with @xmath373 , @xmath374 .
moreover , the curvature scalar goes to zero @xmath375 , while the scalar fields @xmath376 run off to infinity @xmath377 .
in contrast , for infinite chemical potential the geometry at infinity is asymptotic to hvlif with @xmath253 and @xmath378 .
the behaviour of curvature and scalars is precisely the opposite as previously : the curvature scalar diverges @xmath379 , while the scalar fields go to zero @xmath380 .
the geometry at the horizon is independent of the chemical potential , but depends on the temperature . for zero temperature
the asymptotic geometry is again hvlif with @xmath253 , @xmath381 , but approaching the ` opposite end ' of this geometry , so that the curvature scalar goes to zero @xmath382 .
while there is no sp curvature singularity there remains a pp curvature singularity at the horizon , that is , freely falling observers experience infinite tidal forces .
simultanously the scalar fields go to infinity @xmath383 .
this type of behaviour can be considered as a generalized form of attractor behaviour @xcite .
for finite temperature the geometry takes the expected form for a non - extremal black brane , the product of two - dimensional rindler space with @xmath384 .
the scalars and the curvature take finite values , so that the solutions are regular at the horizon for non - zero temperature .
the only element of hvlif holography that we will use is the entropy temperature relation @xmath385 valid for field theories with hyperscaling violation @xcite .
since the low - temperature asymptotics of the exact entropy temperature relation of four - dimensional nernst branes is @xmath19 , which matches the scaling properties of the asymptotic zero temperature near horzion hvlif geometry with @xmath381 , @xmath253 , gauge / gravity duality implies the existence of a corresponding three - dimensional non - relativistic field theory with this scaling behaviour . as the solution is charged , the global solution should describe the rg flow starting from a uv theory corresponding to asymptotic infinity , and ending with this ir theory . identifying this uv theory
turned out to be problematic : the solution at infinity jumps discontinously between finite and infinite chemical potential , and is singular in either case .
the more likely candidate ( having no curvature singularity ) is the conformally rescaled ads@xmath372 , which still does not look like a ground state , due to the run - away behaviour of the scalars .
moreover , while the geometric scaling properties indicate an entropy temperature relation of the form @xmath6 , the high temperature asymptotic of the four - dimensional nernst brane solution is @xmath7 . as discussed before ,
this lead to the conjecture that the solution decompactifies at infinity , and needs to be understood from a five - dimensional perspective . to relate five - dimensional nernst branes to four dimensions , we make the spacelike direction @xmath62 compact , i.e. we identify @xmath387 .
clearly then , to understand the four - dimensional properties , it is crucial to first understand the behaviour of the @xmath62 circle . writing as since it involves the variable @xmath388 .
however , using , it is possible to write in terms of a general kaluza - klein vector , valid for both @xmath205 and @xmath201 .
this then allows the reduction of both cases in parallel , leading to . ]
@xmath389 with @xmath390 we find the four - dimensional line element @xmath391 after identifying @xmath392 . from we can read off the behaviour of the physical ( geodesic ) length @xmath393 of the compactification circle : @xmath394 thus the geodesic length of the compactification circle @xmath386 varies dynamically along the transverse direction , parametrized by @xmath233 , of the four - dimensional spacetime , as shown in figure [ fig : s1 ] .
notice from that for @xmath205 there are two competing terms , resulting in decompactification both for @xmath395 and for @xmath361 .
the latter decompactification is only reached in the extremal limit , since otherwise we encounter the horizon at @xmath362 .
this implies that in the non - extremal case the near horizon solution will still depend on the parameter @xmath128 , while in the extremal case the near horizon solution becomes independent of @xmath128 .
the insensitivity of the extremal near horizon solution to changes of parameters which determine the asymptotic behaviour at infinity , in our case @xmath128 , can be viewed as a version of the black hole attractor mechanism . making the solution non - extremal results in the loss of attractor behaviour by making the near horizon solution sensitive to the asymptotic properties of the solution at infinity . of course
, the four - dimensional scalars run off to infinity instead of approaching finite fix - point values , but they do so in a particular , fine - tuned way , which leads to a consistent lifting of the near horizon geometry five dimensions .
a remarkable feature of solutions with @xmath205 is the existence of a critical point , @xmath396 , where the compactification circle reaches a minimal size at @xmath397 .
in contrast , for @xmath201 , this critical point does not exist and so , whilst the circle continues to decompactify as @xmath361 in the extremal case , it now shrinks monotonically with increasing @xmath233 , ultimately becoming a null circle goes to zero in this limit .
] of zero size for @xmath395 .
this fundamentally different behaviour of the @xmath386 means we must treat the dimensional reduction of the @xmath205 and @xmath201 cases separately in what follows .
additionally , we clearly see that @xmath128 is the parameter responsible for the asymptotic behaviour at infinity from a five - dimensional point of view .
this resembles the role played by the parameter @xmath398 in the four - dimensional solutions of @xcite : this connection will be made manifest in the following subsections . in the case
@xmath205 , the compactification introduces a new continuous parameter , the parametric radius @xmath399 of the circle .
we now observe that the identification @xmath400 breaks the scaling symmetry ( [ scaling ] ) , which made the parameter @xmath128 irrelevant for five - dimensional ( uncompactified ) solutions .
for @xmath205 there is a circle of minimal size at @xmath401 , with geodesic size @xmath402 given by @xmath403 the size of this minimal circle depends only on the combination @xmath404 and is therefore invariant under any increase in @xmath128 that is compensated for by a reduction in @xmath405 and vice - versa .
this ability to trade @xmath405 for @xmath128 means that @xmath128 can be used as the physical parameter controlling the minimal circle size , whilst @xmath405 becomes redundant .
it is natural to set @xmath406 , as this is precisely what is needed such that the expression for the four - dimensional charge , @xmath9 , calculated later in , is independent of the compactification radius , which is natural for a quantity which was defined in @xcite in a purely four - dimensional context . in the case
@xmath201 , there is no such invariant length and we can see this in a number of ways .
firstly , the @xmath407 limit pushes @xmath408 and so no minimal circle exists .
secondly , with @xmath201 , the geodesic size of the compactification circle is found from to be @xmath409 and depends only on @xmath199 ; since this is already a parameter of the five - dimensional solution , there is nothing else to be accounted for and no need for additional parameters .
one can try to obtain an invariant length from the size of the circle on the horizon , @xmath410 , which , assuming non - extremality , will at least be finite .
however , it is clear from that this will be a function of both @xmath199 and @xmath231 , which again are already existing parameters of the five - dimensional @xmath201 solution . in @xcite
a family of four - dimensional nernst branes was found , which depend on one electric charge @xmath9 and two continuous parameters @xmath411 and @xmath398 , which can be expressed alternatively in terms of temperature @xmath412 and chemical potential @xmath5 .
it was also observed that the four - dimensional solutions with finite chemical potential exhibited a specific singular behaviour in the asymptotic regime , which suggested to be interpreted as a decompactification limit .
given the behaviour of the compactification circle , the natural candidate for a lift of four - dimensional nernst branes with finite chemical potential is the @xmath205 family of five - dimensional nernst branes .
we begin by comparing the four - dimensional nernst brane solutions with finite chemical potential ( @xmath413 ) as found in @xcite to the four - dimensional metric in obtained by dimensionally reducing our five - dimensional solution with @xmath205 . setting @xmath414 in ( 3.30 ) of @xcite
gives : @xmath415 where @xmath416 and @xmath417 = : c { \cal h}_0(r ) \;.\ ] ] here @xmath9 parametrizes the four - dimensional electric charge , the continuous parameter @xmath418 corresponds to a chemical potential @xmath5 , with @xmath419 , and the continuous parameter @xmath420 corresponds the temperature @xmath421 .
the constant @xmath422 is determined by the choice of a prepotential and a gauging of the four - dimensional theory .
more precisely , it is determined by the cubic coefficients @xmath10 and gauging parameters @xmath11 , but since we are assuming that this solution can be lifted to five - dimensions , these are the same parameters that enter into our five - dimensional theory in
. the precise form of @xmath422 can be read off from the unnumbered equation between ( 3.30 ) and ( 3.31 ) in @xcite . at this point
we anticipate that the functions @xmath423 and @xmath229 in the four- and five - dimensional solutions can be identified , which allows us to drop the superscrips ` 4d ' on @xmath123 and @xmath4 .
since we can no longer rescale the coordinate @xmath233 , matching the coefficients of @xmath287 between the metrics and fixes the relation between the functions @xmath424 and @xmath425 to be @xmath426 then the remaining metric coefficients match if we rescale @xmath427 by constant factors involving @xmath209 . into @xmath233 , but then by comparing the functions @xmath229 we will conclude that the respective parameters @xmath123 differ by a factor @xmath428 .
given the relation of @xmath123 to the position of the event horizon and to temperature , we prefer not to do this . ] writing out the functions @xmath429 and @xmath26 and comparing , we obtain : @xmath430 while the five - dimensional line element is non - static , the four - dimensional one is static , but as an additional degree of freedom we have a kaluza - klein gauge field , given by @xmath431 here we use the definitions and conventions of section [ 5dto3d ] , and with regard to four - dimensional quantities , we use the conventions of@xcite , which were also used in @xcite . by matching the expression for @xmath432 given by with the @xmath107-derivative of ( 3.38 ) of @xcite , we can identify the kaluza - klein vector with the four - dimensional gauge field provided that @xmath433 from this we can find @xmath434 which expresses the four - dimensional parameters @xmath435 in terms of the five - dimensional parameters @xmath436 . comparing ( [ match1 ] ) to ( [ match2 ] ) we find that these relations are mutually consistent provided that @xmath437 this equations relates the overall normalizations of metrics ( [ 4dmetric ] ) and ( [ 4dnernstbrane ] ) and of the underlying vector multiplet actions .
the four - dimensional chemical potential is given by the asymptotic value of the gauge field @xmath438 , which is chosen such that @xmath439 , as explained in appendix [ vanishing ] .
having matched the five - dimensional kaluza - klein vector with the four - dimensional gauge field of @xcite , the corresponding expressions for the chemical potential must also match .
for reference , we provide the following expression in terms of both four- and five - dimensional parameters , @xmath440 = \frac{\delta}{2 q_0 a } = -\frac{\sqrt{3}}{a } \sqrt { \frac{\delta}{\delta + 2 b_0 a}}\;,\ ] ] where we used ( [ match2 ] ) . notice from that @xmath441 which then forces @xmath442 by , which is consistent with the remark in @xcite that @xmath443 .
moreover we observe that @xmath441 implies @xmath444 .
this reflects the correlation in the signs of the charge @xmath9 and of the chemical potential @xmath5 .
we have , for concreteness and simplicity , restricted ourselves to solutions where @xmath445 , which have turned out to correspond to negative charge and negative chemical potential .
conversely , solutions with @xmath446 will carry positive charge and positive chemical potential .
this is consistent with the fact that in relativistic thermodynamics the chemical potentials of particles and antiparticles differ by a minus sign . for completeness
we note a few further signs which are implied by our decisision to focus on solutions with @xmath446 ( and , hence , @xmath447 ) . from
we deduce that the four - dimensional constant @xmath422 must be negative , @xmath448 , which explains the minus sign in .
furthermore , it is clear from that @xmath449 such that the harmonic function @xmath450 , which we need in order that the roots of @xmath451 , which appear in our expression for the solution , are real .
since the reduction is carried out over the @xmath62 direction , it is instructive to calculate the killing charge associated to the killing vector @xmath452 . for @xmath205 ,
tells us this is related to the killing vector @xmath453 of the five - dimensional spacetime via @xmath454 since the charge associated with @xmath453 is the brane momentum , the killing charge corresponds to momentum in the @xmath62 direction , and can be determined as follows @xmath455 where we have omitted @xmath327 and @xmath209 for simplicity .
the periodicity of the @xmath62 direction implies that momentum takes discrete values , @xmath456 where we have taken into account that @xmath457 . rearranging this as @xmath458 and comparing to
, we see explicitly how the quantization of the internal momentum implies the quantization @xmath459 of the four - dimensional charge .
note that while the spectrum of @xmath460 changes with the radius @xmath461 of the compactification circle , the four - dimensional electric charge @xmath9 is independent of it .
as already mentioned before , @xmath460 and @xmath9 being negative results from choosing @xmath79 positive , and solutions with positive @xmath460 and @xmath9 can be obtained by flipping signs in ( [ zeta_0_dot ] ) .
our choice of signs is consistent with the choices made in @xcite , in particular the same anti - correlation between the signs of @xmath462 and @xmath9 can be observed in the equation above ( 3.38 ) of @xcite .
let us end this discussion by comparing the number of parameters describing the nernst branes in different dimensions .
five - dimensional nernst branes are parametrized by three continuous paramters @xmath463 , but for @xmath205 we have the scaling symmetry ( [ scaling ] ) , which tells us that @xmath128 is redundant , and that we can parametrize solutions by the two independent and continuous parameters @xmath464 , which then correspond to temperature and boost momentum . upon compactification
a new length scale is introduced that breaks the scaling symmetry present in five dimensions .
consequently , the four - dimensional solution picks up an extra parameter ; we need to specify the three independent and continuous parameters @xmath465 in order to completely define the metric . in terms of physical parameters ,
the four - dimensional solution depends on temperature , charge and chemical potential @xmath466 .
these are all independent but , as we have seen , since the momentum has a component in the direction we compactify over , it becomes discrete , which corresponds directly to the discretization of four - dimensional electric charge . as such
, the five - dimensional solution involves two independent and continuous thermodynamic parameters whilst the four - dimensional solution has three independent parameters , two of which are continuous and one of which is discrete .
the two parameter family of four - dimensional nernst branes found in @xcite exhibits discontinuities in the asymptotic behaviour of both the geometry and the scalar fields when taking the limit @xmath467 , or equivalently , @xmath468 .
this discontinuity can be accounted for by the discontinuous asymptotic behaviour of the compactification circle in the limit @xmath407 as seen in figure [ fig : s1 ] .
we should therefore expect that the infinite chemical potential four - dimensional solutions of @xcite with @xmath469 can be recovered from the @xmath201 five - dimensional solution with one dimension made compact . to demonstrate this relationship we take the four - dimensional nernst brane metric obtained in @xcite and set @xmath469 in which reduces the function @xmath470 to @xmath471 substituting this back into gives the following metric @xmath472 on the other hand , the dimensional reduction of the @xmath201 class of five - dimensional nernst branes gives @xmath473 where we have used with @xmath201 . again we identify the functions @xmath423 and @xmath229 appearing in the above metrics , which means the parameters @xmath123 and @xmath4 will be the same in both cases . as before , this prevents rescaling of the coordinate @xmath233 and then , by comparing @xmath287 terms in and , we establish the following relationship between four- and five - dimensional quantities @xmath474 again , the remaining metric coefficients can be made to match by rescaling @xmath427 by constant factors involving @xmath209 . following the same procedure as in section [ 4dmetricmatchsec ]
, we match the gauge field and kaluza - klein vector by comparing expressions for @xmath432 .
specifically , we match with the @xmath107-derivative of ( 3.38 ) in @xcite .
the two are equivalent provided that @xmath475 which expresses the four - dimensional electric charge in terms of the five - dimensional boost parameter @xmath199 .
this is a much simpler expression than in the @xmath205 case and we observe that it matches the @xmath407 limit of .
considering the discontinuities we have encountered previously when taking @xmath407 limits , this seems at first surprising but just reflects that @xmath9 is a well defined paramater for the four - dimensional solutions of @xcite , for any choice of @xmath5 and @xmath4 .
having established @xmath441 , we see from that @xmath407 corresponds to @xmath476 , and thus from that @xmath477 .
lastly , we can substitute into to find the relationship between the overall normalizations of the metrics and , @xmath478 clearly this requires @xmath448 as before and , in fact , is exactly the same relationship as for the @xmath205 case in , which is expected since @xmath422 and @xmath209 are only sensitive to the four- and five - dimensional multiplet actions respectively , and these are indpendent of @xmath128 . again , since we have matched the gauge fields by comparing @xmath479 , the chemical potentials must match and this is indeed the case ; using the asymptotic value of with @xmath201 , we find @xmath480 which agrees with the negatively charged , @xmath469 solutions in @xcite . the parameter counting becomes simpler in the @xmath201 case .
five - dimensional nernst branes are parameterized by two independent and continuous parameters @xmath481 , or equivalently temperature and momentum .
however , as we have seen in section [ s1evolutionsection ] , no new length scale is introduced by the reduction and consequently , the four - dimensional solution obtained via dimensional reduction also depends on exactly two independent parameters , @xmath481 , which are sufficient to completely determine since @xmath201 is fixed . using , these are equivalent to @xmath482 with @xmath480 .
the difference between the five - dimensional and four - dimensional parameters is that the @xmath386 causes charge quantization .
this means that whilst both @xmath123 and @xmath199 are continuous in five dimensions , reducing to four dimensions forces one parameter , namely @xmath483 , to become discrete .
one difference between the @xmath201 solution and the @xmath205 solution is that for the @xmath201 solution the compactification circle has no critical value .
therefore we can not relate the momentum @xmath460 to the electric charge @xmath9 using @xmath405 as a reference scale .
this is not a problem since we could relate @xmath9 to five - dimensional quantities through ( [ match5 ] ) , and , moreover , we have seen that the relation between @xmath9 and five - dimensional quanities has a well defined limit for @xmath484 .
a related feature of the @xmath201 solution is that compactification circle has no minimal size , and contracts to zero for @xmath485 .
that means that there is a region in this solution , where the circle has sub - planckian , or sub - stringy size .
while this is problematic for an interpretation as a four - dimensional solution , the lifted five - dimensional solution is simply ads@xmath0 , and can be decribed consistently within five - dimensional supergravity .
the four - dimensional solutions with @xmath205 and @xmath201 , obtained in sections [ 4da>0dimredsec ] and [ 4da=0dimredsec ] , exactly match the @xmath486 and @xmath469 solutions of @xcite respectively . in @xcite these four - dimensional solutions
were observed to be hyperscaling - violating lifshitz metrics .
it is known from @xcite that such solutions suffer from various curvature singularities , and we shall now investigate this by computing the singular behaviour of the metrics and . as with the five - dimensional spacetimes in section [ ssec:5dsingular ] we can determine the presence of curvature singularities of our four - dimensional solutions by looking at the kretschmann scalar and ricci scalar associated to the metrics and . indeed , since any singular behaviour in the curvature
will already be present for the extremal solutions , we will concentrate only on the case @xmath359 .
the curvature invariants are calculated ( using maple ) to be @xmath487 for @xmath205 , or equivalently @xmath419 , we find that the ricci scalar behaves as @xmath488 for large @xmath233 , and @xmath489 for @xmath246 , whilst the kretschmann scalar scales as @xmath490 and @xmath491 in these respective regions .
hence , the curvature invariants will remain finite along the solution .
however , for the @xmath201 solution we will still have the same behaviour at @xmath246 , but asymptotically we find @xmath489 and @xmath491 .
we therefore have a naked curvature singularity as we approach the boundary of the spacetime .
in order to investigate whether the four - dimensional solutions of @xcite admit infinite tidal forces in the near - horizon regime we will follow the analysis of @xcite , albeit considering a slightly simpler set - up in which the infalling observer is moving only in the radial direction i.e. has zero transverse momentum .
the technical details of this procedure can be found in appendix [ app:4dtidalforces ] .
our results in tables [ 4daneq0tidalforces ] and [ 4da=0tidalforces ] show that , for both @xmath205 and @xmath201 , there exist components of the riemann tensor , as measured in the ppon , that diverge as @xmath246 .
this indicates that the radially infalling observer will experience infinite tidal forces at the extremal horizon , @xmath359 . as before
, tidal forces will remain finite on non - extremal horizons , @xmath362 .
a summary of the singular behaviour of our four- and five - dimensional solutions can be found in tables [ 4dtable ] and [ 5dtable ] .
notice that since @xmath123 and @xmath128 control the near horizon and asymptotic geometries respectively , we can use these to catalogue any singularities .
we will now explain how the singularities present in the four - dimensional hyperscaling - violating lifshitz solutions of section [ 4dsingularsection ] , except those related to infinite tidal forces at extremal horizons , can be removed by dimensional lifting to the asymptotically ads solutions of section [ ssec:5dsingular ] .
.summary of singular behaviour of four - dimensional nernst brane . [ cols="^,^,^,^,^ " , ]
dimensional reduction relates the five - dimensional ricci scalar to its four - dimensional counterpart by .
the appearence of the second power of the dilaton reflects the fact that the kretschmann scalar is quadratic in the curvature . ]
@xmath492 as can be seen from table [ 4dtable ] and table [ 5dtable ] , the only situation where we encounter a curvature singularity is the asymptotic regime of the four - dimensional solution with @xmath469 , or equivalently @xmath201 . in this instance
we have @xmath493 from whilst @xmath494 from resulting in @xmath495 being asymptotically constant and exactly equal to the value of global @xmath352 as seen in section [ ssec:5dsingular ] . recalling that the dilaton @xmath496 measures the geodesic length of the @xmath62 circle , we can now account for the presence of an asymptotic curvature singularity in this class of four - dimensional nernst branes .
specifically , the four - dimensional , @xmath480 , asymptotic curvature singularity emerges from a ` bad slicing , ' of the parent @xmath352 hyperboloid by a circle that gets pinched at infinity .
it was shown previously , that the independent four - dimensional scalars are all proportional to each other , see formula ( 3.29 ) in @xcite .
it was also observed that for infinite chemical potential , these scalars approach zero asymptotically . from the five - dimensional point of view , the single profile of the four - dimensional scalars determines the profile of the kaluza - klein scalar .
therefore the four - dimensional scalars approaching zero corresponds to the shrinking of the compactification circle .
when combining this with the singular behaviour of the four - dimensional metric , we obtain @xmath352 . in the @xmath205 , or equivalently @xmath497 ,
case the four - dimensional solution of @xcite is asymptotically conformal to @xmath498 , or @xmath499 for short .
we see from that the curvature invariants of @xmath499 behave as @xmath500 and vanish asymptotically . at the same time , this is compensated by @xmath501 from , meaning the circle now blows up at large @xmath233 such that @xmath495 remains asymptotically constant and equal to @xmath502 .
thus , in this case the asymptotic behaviour of the four - dimensional metric and scalars is reversed compared to the @xmath201 case , but still leads to the same five - dimensional asymptotic geometry after lifting . as can be seen from tables [ 4dtable ] and [ 5dtable ] ,
tidal forces are asymptotically irrelevant and [ app:4dtidalforces ] for reasons why . ]
and so we are only concerned with the situation near the horizon .
it is clear that infinite tidal forces are present at the horizon of the extremal nernst brane in four - dimensions , and are not removed by dimensional lifting .
this seems to be the price for obtaining the strong version of nernst s law .
let us summarize and discuss our results .
starting from fi - gauged five - dimensional supergravity with an arbitrary number of vector multiplets , we have obtained a two - parameter family of nernst branes , labelled by temperature and momentum .
these solutions interpolate between ads@xmath0 and an event horizon , and have an entropy temperature relation interpolating between @xmath6 at high temperature / low boost and @xmath19 at low temperature / high boost .
the relation @xmath6 is consistent with the scaling properties of ads@xmath0 . given that we are working within five - dimensional gauged @xmath12 supergravity , the dual uv field theory should be a conformally invariant four - dimensional @xmath503 field theory .
since the metric is the same as in the duality between gauged @xmath14 supergravity and @xmath504 super yang mills , one might expect it to be a conformally invariant @xmath503 super yang mills theory or a deformation thereoff , but without having a higher dimensional embedding which allows one to understand the role of the parameters @xmath10 and @xmath11 of the gauge theory , we ca nt say much more .
we have seen how the five - dimensional lift of four - dimensional nernst branes removes all the singularities at asymptotic infinity as well as the mismatch between geometrical and thermodynamic scaling relations . to understand the variation of the compactification circle along the transverse direction , which from the four - dimensional point of view is encoded in the scalar fields , is crucial .
the apparently singular behaviour of the four - dimensional geometry is exactly compensated for by the singular behaviour of the scalars , or , put differently , by the behaviour of the circle one has to add to obtain asymptotically ads@xmath0 .
moreover , the compactification circle also accounts for the four - dimensional chemical potential , which has no counterpart in the un - compactified five - dimensional solution .
however , once we decide to make the boost direction compact the dynamics forces the circle to expand at both ends , and the resulting minimum introduces a new parameter which we can relate to the chemical potential . as proposed in @xcite , we can interpret the apparently singular uv behaviour of four - dimensional nernst branes as a dynamical decompactification limit , which tells us that the description as a four - dimensional system breaks down and has to be replaced by a five - dimensional one .
the five - dimensional solution admits a non - trivial extremal limit , where the boost parameter is sent to infinity , while the momentum ( density ) is kept fixed .
the resulting extremal near horizon geometry should define a field theory with entropy - temperature relation @xmath19 . in the context of boosted d - branes and m - branes ,
the proposed interpretation is a conformal field theory in the infinite momentum frame , which carries a finite momentum density @xcite .
moreover , it was proposed in @xcite that the compactification of the direction along the boost corresponds to discrete light cone quantisation . in this respect
it is interesting to look at the asymptotic scaling symmetries of the five- and four - dimensional extremal solutions near the horizon . in five dimensions
the metric looks like a lifshitz metric with @xmath253 and @xmath254 , except that the direction along the boost has weight @xmath256 instead of @xmath257 . upon reduction to four dimensions , the asymptotic geometry , and
if we go to infinite chemical potential even the global geometry , is a hyperscaling violating lifshitz geometry with @xmath253 and @xmath381 @xcite .
that is , by reduction over the boost direction one trades the non - trivial scaling of this direction for an overall scaling of the metric .
following @xcite we propose to associate a four- and a three - dimensional field theory to the near - horizon five- and four - dimensional geometries , respectively , with the three - dimensional theory encoding the zero mode sector of the discrete light cone quantisation of the four - dimensional theory .
both theories are non - relativistic with lifshitz exponent @xmath253 , and supersymmtric with two supercharges .
the four - dimensional theory is scale invariant and arises by deforming a four - dimensional relativistic @xmath503 supersymmetric theory by a finite momentum density , while the three - dimensional theory is scale covariant . from a strictly gravitational point of view
, one should still worry about the pp curvature singularities which persist in the extremal limit irrespective of whether we consider four - dimensional or five - dimensional nernst branes .
while sometimes considered to be ` mild , ' they are genuine curvature singularities which make the solution geodesically incomplete .
moreover , they are not cured by stringy @xmath505-corrections @xcite , and strings probing pp singularities get infinitely excited @xcite .
while at finite temperature there is technically no singularity , near extremality objects falling towards the event horizon will still experience very large tidal forces @xcite .
this behaviour is , if not an inconsistency , at least a sign that the singularity has physical relevance .
moreover , the pp singularity is clearly caused by the way the metric complies with the strong version of nernst s law , namely through a warp factor which scales any finite piece of the world volume , which we use to refer extensive quantities to ` unit world volume . '
] to zero volume .
it is not obvious at all how pp singularities could be removed while keeping the strong version of nernst s law . for small bps black holes ,
@xmath506-corrections remove null curvature singularities , by making the area finite @xcite . but
as these singularities are of the sp type , it is not clear what this implies for pp singularities .
one example where a pp singularity is removed is the d6 brane of type iia supergravity , using an m - theory embedding @xcite .
the effect of higher curvature corrections on pp type singularities has been investigated in @xcite .
one can also approach the problem from the field theory side .
for example , in @xcite they study the infinite momentum frame cft dual to a boosted brane and find evidence that the cft resolves the geometric singularity . in our case
it would be interesting to understand the dual four- , or possibly , the three - dimensional ir field theory , and to investigate whether it is non - singular , and whether its ground state is unique or degenerate . and
if the ground state is unique , one would need to understand whether this means that ( i ) pp - singularities are acceptable , ( ii ) they are not , but the dual field theory can be used to construct a ` quantum geometry ' of some sort , ( iii ) or if there is some kind of breakdown of gauge / gravity duality in the extremal limit . points ( i ) and ( iii ) are not necessarily mutually exclusive , since one might invoke the process version of the third law to assure that the extremal limit can not be reached by any physical process .
this paper is part of a series of papers where explicit , non - extremal solutions of five- and four - dimensional ungauged and gauged supergravity have been constructed using time - like dimensional reduction in combination with special geometry @xcite . as explained in section 2 , solutions correspond to curves on a particular submanifold of the para - quaternionic khler manifold obtained by reduction to three dimensions , which satisfy the geodesic equation deformed by a potential . as part of the solution
we have obtained an explicit expression for a stationary point of the five - dimensional scalar potential , corresponding to an ads@xmath0 vacuum , for an arbitrary number of vector multiplets and general fi - gauging .
while we initially obtain solutions to the full second order field equations , with the corresponding number of integration constants , we have seen that once we impose regularity of the lifted five - dimensional solution at the horizon the number of intergration constants is reduced by one half , so that the solution satisfies a unique set of first order equations .
such behaviour has been observed before , and been interpreted as a remnant of the attractor mechanism @xcite . for our five - dimensional solutions
the scalars are constant , so that the only sense in which we have attractor behaviour is that the scalars sit at a stationary point of the scalar potential . however , from the four- and three - dimensional perspective we have scalar fields which need to exhibit a particular , fine - tuned , asymptotic behaviour at the horizon in order to make the five - dimensional solution regular .
this is very similar to attractor behaviour , and the effect of reducing the number of integration constants by one half is the same .
such universal features of scalar dynamics deserve further study . in the present paper
we have made a very particular choice of the ansatz , which was tailored to obtaining the five - dimensional lift of the four - dimensional nernst branes of @xcite . in the future
we will study systematically other choices , which will lead to other and more general solutions .
already in @xcite a four - dimensional magnetic solution was found , and we expect that it is possible to obtain dyonic solutions as well
. it would also be interesting to revisit the issue of embeddings into ten- and eleven - dimensional supergravity .
the work of jg is supported by the stfc grant st/1004874/1 .
the work of tm is partially supported by stfc consolidated grant st / g00062x/1 .
the work of de was supported by stfc studentship st / k502145/1 and by the scottish international educational trust .
tm would like to thank gabriel cardoso and suresh nampuri for extensive and helpful discussions about their work on nernst branes .
no additional research data beyond the data presented and cited in this work are needed to validate the research findings in this work .
our goal in this appendix is to obtain a workable expression for the scalar potential @xmath327 appearing in .
let us concentrate on the term @xmath507 .
this is to be interpreted as the matrix inverse to @xmath508 in the sense that @xmath509 now , using the expression for @xmath102 : @xmath510 we have @xmath511 .\ ] ] we now introduce the dual scalars @xmath512 via @xmath513 hence , becomes @xmath514.\ ] ] in other words , the quantity @xmath507 is just the inverse of the term in square brackets in .
thankfully , the latter is easily invertible .
indeed , we find @xmath515\cdot \frac{2}{3}\left[\hat{g}_{jk}(y)+4y_j y_k\right ] = \delta^i_k \,.\ ] ] hence we can rewrite @xmath516 so that the scalar potential term in becomes @xmath517g_i g_j.\ ] ]
we use the form of our five - dimensional line element given in , which can be rewritten as @xmath518 where @xmath519 . note that @xmath520 so we can interpret this as a velocity vector . following the procedure of @xcite we want to calculate the quasilocal stress tensor @xmath521 associated with the metric .
given a timelike surface @xmath316 at constant radial distance @xmath233 we define the metric @xmath522 on @xmath316 via the adm - like decomposition @xmath523 we define the extrinsic curvature @xmath524 via @xmath525 where @xmath526 is the outward - pointing normal vector to the surface @xmath316 . for solutions asymptoting to @xmath352 the procedure of @xcite
tells us that the quasilocal stress tensor is then given by . ]
@xmath527 where @xmath528 is the trace of the extrinsic curvature , and @xmath529 is the einstein tensor for @xmath522 .
for the case at hand we see that the metric decomposes according to with @xmath530 the unit normal vector @xmath526 to a surface of constant @xmath233 is given by @xmath531 from which we find the extrinsic curvature @xmath532 in order to calculate the trace of this we need an expression for the inverse metric @xmath533 , which is given by @xmath534,\ ] ] where @xmath535 , etc .
this can be used to compute the trace of the extrinsic curvature @xmath536 .\ ] ] putting all this together , and noting that @xmath537 , we can use to find the resulting gravitational stress - energy tensor induced on the boundary @xmath538 , @xmath539 where the dots represent terms which are subleading in the limit @xmath485 .
the quasilocal stress tensor can be used to compute well - defined mass and other conserved charges for the spacetime .
let @xmath540 be a spacelike hypersurface in @xmath541 and make the adm decomposition @xmath542 where @xmath543 are coordinates spanning @xmath540 , which has metric @xmath544 .
let @xmath545 be the timelike unit normal to @xmath540 .
then for any isometry of @xmath546 , which we take to be generated by a killing vector @xmath547 , we can define a conserved charge @xmath548 by @xmath549 in particular , the mass of the solution is given by taking @xmath550 , whilst the momentum in the direction @xmath551 is given by taking @xmath552 . for the boosted black brane we can make the adm decomposition of the metric with @xmath553 the timelike unit normal to @xmath540 has components @xmath554 using these expressions , as well as the components of the quasilocal stress tensor
, we can calculate the mass and linear momentum associated with the boosted black brane .
taking @xmath550 and @xmath555 we obtain the expressions and for the mass and linear momentum respectively .
finally , let us add some further comments on the fact that @xmath231 , and hence temperature , is a physical parameter despite that it can be absorbed by rescaling coordinates in ( [ 5dmetric4 ] ) . from ( [ conservedcharge ] ) , ( [ kappa_squared ] ) , ( [ s ] ) it is manifest that all quantities entering into the first law are geometric quantities ( norms of vectors fields , and integrals of functions over submanifolds using the induced metric ) which are independent of the choice of coordinates . applying the coordinate transformation @xmath556 to these expressions
, it is straightforward to see that the parameter @xmath231 is not eleminated , but scaled out as an overall prefactor . in particular @xmath557
while @xmath558 so that irrespective of our choice of coordinates @xmath358 , @xmath559 , @xmath560 and @xmath561 .
it is precisely this @xmath231-dependence of the thermodynamic quantities that gives rise to the correct temperature / entropy term in the first law . put differently
, when working in the rescaled coordinates @xmath562 the parameter @xmath231 is hidden in the choice of the vector field @xmath547 and the volume @xmath327 .
as is well known from the study of kerr black holes , obtaining the hawking temperature by euclidean methods is much more subtle for non - static spacetimes . for this reason
, we find it useful to give an explicit demonstration of how this works in the case of boosted ( non - static ) black branes .
the treatment of the linear case given below will be parallel to the analysis of the kerr black hole in @xcite .
a euclidean continuation of the boosted black brane solution can be obtained by setting @xmath563 and @xmath564 , and taking @xmath107 and @xmath162 to be real .
observe that following the standard treatment of the kerr solution , we do not only continue time but also the ` boost parameter ' @xmath565 , which is analogous to the angular momentum parameter of the kerr solution in boyer - lindquist coordinates .
the euclidean section of the boosted black brane in is then @xmath566 we now explore the near horizon geometry by adapting a similar calculation used to examine the kerr - newman solution in @xcite . introducing the new radial variable @xmath567 by @xmath568 , the function @xmath229 has the expansion @xmath569 around the horizon . expanding up to order @xmath506 ,
the metric takes the form @xmath570 where we have replaced the coordinates @xmath107 and @xmath219 by the new coordinates @xmath571 we remark that , in contradistinction to the kerr - newman solution discussed in @xcite , ( i ) the coordinate @xmath572 is linear rather than angular , i.e. we do not need to impose an identification on it ; and ( ii ) the coordinate @xmath573 is well defined , since @xmath345 and @xmath162 are constant , so that @xmath574 is exact .
the horizon is at @xmath575 .
the coordinates @xmath576 parametrize a three - dimensional plane with a metric which is flat up to corrections of order @xmath506 .
this part of the metric is clearly regular for @xmath577 .
the variables @xmath567 and @xmath573 parametrize a surface with metric @xmath578 dr^2 + 4 r^2 \frac{r_+^2}{l^4 } d\chi^2 \right),\ ] ] which is , up to a subleading term of order @xmath506 , the metric of a cone with apex at @xmath575 .
thus @xmath573 is an angular variable and the surface parametrized by @xmath567 and @xmath573 is topologically a disk . imposing the absence of a conical singularity at @xmath575 fixes the periodicity of @xmath573 to be @xmath579 since the coordinate @xmath572 is linear ( has no identifications ) we can determine the periodicities of @xmath107 and @xmath219 from @xmath580 with @xmath581 the hawking temperature @xmath4 is read off from the periodicity of @xmath107 by @xmath582 , so that @xmath583 which agrees with the result found by computing the surface gravity . to interpret the periodicity of @xmath219 ,
remember that the boost velocity at the horizon is @xmath584 thus @xmath585 so that the identifications take the form @xmath586 which is analogous to the identification for the euclidean kerr solution , see for example @xcite .
in this appendix we shall construct the frame fields describing the ppon associated to an observer freely falling towards the five - dimensional extremal black brane in .
the frame - dragging effects associated to the brane s boost in the @xmath219 direction mean that an observer who starts falling radially inward from infinity will acquire a velocity in the @xmath219 direction .
we want to pick our first frame field to be the vector field generating the geodesic motion of the observer . to do this
, we follow the procedure of @xcite and introduce the frame field @xmath587 where @xmath107 is the proper time of our observer or , equivalently , the affine parameter for the geodesic motion , and a dot denotes differentiation with respect to @xmath107 .
note that for simplicity we consider an observer who is not moving in the @xmath588 and @xmath589 directions .
it is clear that to obtain @xmath590 , we must first obtain @xmath591 and @xmath592 . to do this
, we recall that associated to each of the killing vector fields @xmath593 of there is an integral of motion .
these conserved quantities are the energy and momenta , @xmath594 defining the quantities @xmath595 we can simultaneously solve and to find @xmath596 notice that both of these velocities diverge as we approach the horizon at @xmath359 .
this divergence tells us that this particular coordinate system is not valid beyond the horizon .
however , for our current purposes , this is not a problem as we are only interested in tidal forces close to , but outside , the horizon . in order to write down @xmath590 , we still need to obtain @xmath592 . for this
we use that @xmath597 for a timelike observer , which is equivalent to @xmath598 where we ve taken the negative root to represent a radially infalling observer and @xmath599 are the roots of @xmath600 .
notice that had we instead picked the positive root in , describing an outgoing timelike geodesic , @xmath592 will become complex for sufficiently large @xmath233 ; this indicates that geodesic can not reach the boundary but in fact hits a turning point and returns to the bulk @xcite .
for this reason , we will only be interested in near horizon tidal forces .
we can now substitute the above expressions for @xmath601 into to obtain the following expression for the first frame field @xmath602 whilst the frame field @xmath590 correctly describes the parallel propagation , it is not correctly normalised . to form an orthonormal basis of frame fields we can apply gram - schmidt procedure to the set of linearly independent frame fields @xmath603 where @xmath604 .
this was done using maple and returns a basis of frame fields that we shall denote @xmath605 without the hat .
these still correctly characterise the parallel propagation but at the same time are fully orthonormal in the sense that they satisfy @xmath606 .
the full expressions for the individual frame fields @xmath605 are quite complicated and not especially illuminating so we omit them here .
however , we can then use the frame fields as transformation matrices to obtain the components of the riemann tensor as measured in the ppon via @xmath607 the non - zero components of the ppon riemann tensor are again rather complicated and so rather than provide full expressions , we instead list their scaling behaviour in the near horizon regime in table [ riemanntable ] . ' '' '' @xmath608 & const + ' '' '' @xmath609 & @xmath610 + ' '' '' @xmath611 & @xmath610 + ' '' '' @xmath612 & @xmath610 + ' '' '' @xmath613 & @xmath610 + ' '' '' @xmath614 & @xmath615 + ' '' '' @xmath616 & @xmath615 + ' '' '' @xmath617 & @xmath618 + ' '' '' @xmath619 & @xmath610 + ' '' '' @xmath620 & @xmath618 + ' '' '' @xmath621 & @xmath618 + ' '' '' @xmath622 & @xmath618 + ' '' '' @xmath623 & @xmath624 +
to investigate the tidal forces present for the four - dimensional extremal nernst brane solutions of @xcite we must treat the cases with finite and infinite four - dimensional chemical potential separately as we have done throughout the paper .
these have metrics given in and respectively .
we shall proceed in a similar fashion to appendix [ app:5dtidalforces ] except for the assumption that the infalling observer is now moving only in the radial direction and has no transverse momentum in either the @xmath588 or @xmath589 directions .
this is slightly different to the analysis of @xcite and means the tangent vector for the timelike geodesic on which our radially infalling observer is travelling is given by @xmath625 where dot denotes differentiation with respect to the observer s proper time , @xmath107 .
the extremal version of is given by @xmath626 the energy is again an integral of motion : @xmath627 for a timelike geodesic we have @xmath628 where we pick the negative square root to represent an observer falling radially inwards .
we could equally well pick the positive root and consider an outgoing geodesic but @xmath592 will become complex for large @xmath233 , meaning the geodesic encounters a turning point and is reflected back into the bulk .
this is reminiscent of the situation in appendix [ app:5dtidalforces ] and in fact , this inability of timelike geodesics to reach the boundary is an example of a property that hyperscaling - violating lifshitz spacetimes can inherit from their parent anti de - sitter spacetimes .
all of this means that we need only focus on the ingoing observer and near horizon tidal forces .
another similarity with appendix [ app:5dtidalforces ] is the divergence of @xmath629 and @xmath592 as @xmath361 ; again this indicates the coordinates are only valid up the horizon which is absolutely fine for the analysis of tidal forces .
next we align the frame field @xmath630 with the vector field @xmath631 responsible for generating the integral curve along which the observer is moving : @xmath632 the observer is moving in the @xmath633 directions and so there are two frame fields associated to this : @xmath630 and @xmath634 .
since the observer is nt moving in any of the @xmath635 ( @xmath636 ) directions , the frames @xmath637 for @xmath636 are just given by the square roots of the inverse metric components i.e. @xmath638 it remains to find the frame @xmath634 such that the @xmath605 form a ppon .
we have picked @xmath630 to describe the parallel propagation and so we just need a second frame field , @xmath634 , that is orthonormal to both @xmath630 and @xmath639 .
it follows from simple linear algebra that @xmath640 it is interesting to note that in the case of the static four - dimensional metric , the frame fields are already orthonormal whereas in appendix [ app:5dtidalforces ] , where the five - dimensional metric is non - static , this is not the case and we had to perform an additional gram - schmidt procedure at this point .
we next use maple to find the components of the riemann tensor in a coordinate basis with lowered indices , @xmath641 , and then multiply by frame fields to obtain the local tidal forces felt by the observer as in .
we again omit the full expressions and instead list in table [ 4daneq0tidalforces ] the scaling behaviour of the non - zero components in the near horizon regime ' '' '' @xmath608 & @xmath233 + ' '' '' @xmath614 & @xmath642 + ' '' '' @xmath616 & @xmath642 + ' '' '' @xmath620 & @xmath642 + ' '' '' @xmath623 & @xmath643 + here we repeat the same procedure as above for the @xmath201 extremal metric .
the extremal version of is given by @xmath644 the resulting nonzero components of the riemann tensor as measured in the ppon are given in table [ 4da=0tidalforces ] . ' '' '' @xmath608 & @xmath233 + ' '' '' @xmath614 & @xmath642 + ' '' '' @xmath616 & @xmath642 + ' '' '' @xmath620 & @xmath642 + ' '' '' @xmath623 & @xmath645 + the near horizon scaling behaviours of the ppon riemann tensor components in tables [ 4daneq0tidalforces ] and [ 4da=0tidalforces ] agree .
this is consistent with the fact that the parameter @xmath128 only affects the asymptotic geometry , which is why the metrics and both take the same form in the small @xmath233 limit : specifically , a hyperscaling - violating lifshitz metric with parameters @xmath646 as observed in @xcite .
it is worthwhile to check the consistency of the results of this appendix with the complete classification of hyperscaling - violating lifshitz singularities obtained in @xcite .
it can be shown that our @xmath646 geometry is equivalent to a @xmath647 geometry in their notation .
this would place our near horizon metric into class iv of the analysis in @xcite , making it both consistent with the null energy condition and indicative of a null curvature singularity ( infinite tidal forces ) at @xmath242 .
for the four - dimensional chemical potential @xmath648 , to be uniquely defined , it is crucial that the vector potential is normalized such that @xmath439 .
while this is widely used and the reason well known , see for example @xcite , we would like to review the full argument here for completeness .
assume that we are given a static space - time which has a killing horizon with killing vector field @xmath547 .
if the norm of @xmath547 has a simple zero at the horizon , in other words , if the solution is non - extremal , then the space - time can be continued analytically to a space - time which contains a bifurcate horizon @xcite .
this means that the horizon has a spatial section @xmath649 where the killing vector field @xmath547 vanishes . if @xmath128 is a well - defined one - form on this space - time , then @xmath650 on @xmath649 .
since the horizon is generated by the flow of the killing vector field @xmath547 , and if assuming that the one - form @xmath128 is invariant under @xmath547 , @xmath651 ( where @xmath652 denotes the lie derivative ) , it follows that @xmath650 on the whole horizon . outside the horizon
we can define a time coordinate @xmath280 , such that @xmath653 .
then the horizon limit of the component @xmath438 of the one - form is @xmath654 . in our application
, we have non - extremal solutions with killing horizons , generated by @xmath547 , given by @xmath550 outside the horizon .
moreover not only the metric but also the vector field is assumed static ( invariant under @xmath280 ) , and therefore @xmath438 has to vanish on the horizon . by continuity
this continues to hold in the extremal limit .
x. dong , s. harrison , s. kachru , g. torroba , and h. wang , _ aspects of holography for theories with hyperscaling violation _ , _ jhep _ * 06 * ( 2012 ) 041 , [ http://xxx.lanl.gov/abs/1201.1905 [ arxiv:1201.1905 ] ] .
m. cvetic , h. lu , and c. n. pope , _ geometry of the embedding of supergravity scalar manifolds in d = 11 and d = 10 _ , _ nucl .
phys . _ * b584 * ( 2000 ) 149170 , [ http://xxx.lanl.gov/abs/hep-th/0002099 [ hep - th/0002099 ] ] .
a. azizi , h. godazgar , m. godazgar , and c. n. pope , _ the embedding of gauged stu supergravity in eleven dimensions _ , _ phys .
* d94 * ( 2016 ) , no . 6 066003 , [ http://xxx.lanl.gov/abs/1606.0695 [ arxiv:1606.0695 ] ] .
s. barisch - dick , g. lopes cardoso , m. haack , and s. nampuri , _ extremal black brane solutions in five - dimensional gauged supergravity _
, _ jhep _ * 1302 * ( 2013 ) 103 , [ http://xxx.lanl.gov/abs/1211.0832 [ arxiv:1211.0832 ] ] .
m. cvetic , g. w. gibbons , h. lu , and c. n. pope , `` rotating black holes in gauged supergravities : thermodynamics , supersymmetric limits , topological solitons and time machines . '' e - print hept - h/0504080 , 2005 .
j. podolsky , _ interpretation of the siklos solutions as exact gravitational waves in the anti - de sitter universe _
, _ class .
* 15 * ( 1998 ) 719733 , [ http://xxx.lanl.gov/abs/gr-qc/9801052[gr-qc/9801052 ] ] .
v. cortes and t. mohaupt , _ special geometry of euclidean supersymmetry iii : the local r - map , instantons and black holes _
, _ jhep _ * 0907 * ( 2009 ) 066 , [ http://xxx.lanl.gov/abs/0905.2844[arxiv:0905.2844 ] ] .
e. bergshoeff , s. cucu , t. de wit , j. gheerardyn , r. halbersma , s. vandoren , and a. van proeyen , _ superconformal n=2 , d = 5 matter with and without actions _ , _ jhep _ * 10 * ( 2002 ) 045 , [ http://xxx.lanl.gov/abs/hep-th/0205230[hep-th/0205230 ] ] .
e. bergshoeff , s. cucu , t. de wit , j. gheerardyn , s. vandoren , and a. van proeyen , _
n = 2 supergravity in five - dimensions revisited _ , _ class .
* 21 * ( 2004 ) 30153042 , [ http://xxx.lanl.gov/abs/hep-th/0403045[hep-th/0403045 ] ] .
grav.23,7149(2006 ) ] . d. v. alekseevsky and v. corts , _ geometric construction of the r - map : from affine special real to special khler manifolds _ , _ commun
_ * 291 * ( 2009 ) 579590 , [ http://xxx.lanl.gov/abs/0811.1658[arxiv:0811.1658 ] ] .
s. cremonini , k. hanaki , j. t. liu , and p. szepietowski , _ black holes in five - dimensional gauged supergravity with higher derivatives _ , _ jhep _ * 12 * ( 2009 ) 045 , [ http://xxx.lanl.gov/abs/0812.3572 [ arxiv:0812.3572 ] ] .
p. dempster and t. mohaupt , _ non - extremal and non - bps extremal five - dimensional black strings from generalized special real geometry _ ,
_ class.quant.grav .
_ * 31 * ( 2014 ) 045019 , [ http://xxx.lanl.gov/abs/1310.5056[arxiv:1310.5056 ] ] .
i. racz and r. m. wald , _ global extensions of space - times describing asymptotic final states of black holes _ , _ class .
grav . _ * 13 * ( 1996 ) 539553 , [ http://xxx.lanl.gov/abs/gr-qc/9507055 [ gr - qc/9507055 ] ] .
s. kobayashi , d. mateos , s. matsuura , r. c. myers , and r. m. thomson , _ holographic phase transitions at finite baryon density _
, _ jhep _ * 02 * ( 2007 ) 016 , [ http://xxx.lanl.gov/abs/hep-th/0611099 [ hep - th/0611099 ] ] .
m. m. caldarelli , g. cognola , and d. klemm , _ thermodynamics of kerr - newman - ads black holes and conformal field theories _ , _ class.quant.grav . _ * 17 * ( 2000 ) 399420 , [ http://xxx.lanl.gov/abs/hep-th/9908022[hep-th/9908022 ] ] .
g. gibbons , m. perry , and c. pope , _ the first law of thermodynamics for kerr - anti - de sitter black holes _ , _ class.quant.grav .
_ * 22 * ( 2005 ) 15031526 , [ http://xxx.lanl.gov/abs/hep-th/0408217 [ hep - th/0408217 ] ] .
j. maldacena , d. martelli , and y. tachikawa , _
comments on string theory backgrounds with non - relativistic conformal symmetry _ , _ jhep _ * 0810 * ( 2008 ) 072 , [ http://xxx.lanl.gov/abs/0807.1100 [ arxiv:0807.1100 ] ] .
s. barisch - dick , g. lopes cardoso , m. haack , and .
vliz - osorio , _ quantum corrections to extremal black brane solutions _ , _ jhep _ * 02 * ( 2014 ) 105 , [ http://xxx.lanl.gov/abs/1311.3136 [ arxiv:1311.3136 ] ] .
k. hristov , s. katmadas , and i. lodato , _ higher derivative corrections to bps black hole attractors in 4d gauged supergravity _ , _ jhep _ * 05 * ( 2016 ) 173 , [ http://xxx.lanl.gov/abs/1603.0003 [ arxiv:1603.0003 ] ] . | we construct nernst brane solutions , that is black branes with zero entropy density in the extremal limit , of fi - gauged minimal five - dimensional supergravity coupled to an arbitrary number of vector multiplets .
while the scalars take specific constant values and dynamically determine the value of the cosmological constant in terms of the fi - parameters , the metric takes the form of a boosted ads schwarzschild black brane .
this metric can be brought to the carter - novotn - horsk form that has previously been observed to occur in certain limits of boosted d3-branes . by dimensional reduction to four dimensions
we recover the four - dimensional nernst branes of arxiv:1501.07863 and show how the five - dimensional lift resolves all their uv singularities .
the dynamics of the compactification circle , which expands both in the uv and in the ir , plays a crucial role . at asymptotic infinity ,
the curvature singularity of the four - dimensional metric and the run - away behaviour of the four - dimensional scalar combine in such a way that the lifted solution becomes asymptotic to ads@xmath0 .
moreover , the existence of a finite chemical potential in four dimensions is related to fact that the compactification circle has a finite minimal value .
while it is not clear immediately how to embed our solutions into string theory , we argue that the same type of dictionary as proposed for boosted d3-branes should apply , although with a lower amount of supersymmetry .
1.5 true cm five - dimensional nernst branes from special geometry + 1.0 true cm p. dempster,@xmath1 d. errington,@xmath2 j. gutowski@xmath3 and t. mohaupt@xmath1 + @xmath2department of mathematical sciences + university of liverpool + peach street + liverpool l69 7zl , uk + [email protected] , [email protected] , [email protected] + @xmath3department of mathematics + university of surrey + guildford , gu2 7xh , uk + [email protected] + september 16 , 2016 . revised : october 11 , 2016 .
final version : november 15 , 2016 .
1.0 true cm |
a satisfactory connection between quasar absorption lines and galaxies close to the qso line of sight has been provided so far only by relatively low redshift studies ( @xmath12 ) , as galaxies at high redshifts are much harder to observe . for @xmath13 ,
large samples of quasar absorption lines have been provided because at these redshifts the many strong uv lines are redshifted to the optical range , where very sensitive ground based observations can be obtained .
however a complete picture in the whole observable redshift range is missing because of the lack of information at intermediate redshifts ( @xmath14 ) . with the advent of the high resolution uv spectrograph space telescope imaging spectrograph ( stis ) , a detailed study of the intermediate redshift
forest and the connection to galaxy properties can now be obtained .
indeed , this was one of the motivations for the hubble deep field south ( hdfs ) .
much of the ground - based and hst hdfs campaign ( williams et al .
1999 ) has been devoted to the observations of the quasar j2233606 ( @xmath0 ) that lies in the stis field ( about 5 and 8 arcmin from the wfpc2 and nicmos fields ) .
the study of the quasar incorporates both spectroscopy ( sealy et al .
1998 ; savaglio 1998 ; outram et al .
1998 ; ferguson et al . 1999 ) and imaging ( gardner et al . 1999 ) . here
we present the interesting features shown by the distribution of the clouds obtained by the fitting of the absorption lines in the redshift range @xmath15 in the high and medium resolution spectra taken with stis / hst ( ferguson et al .
1999 ) and ucles / aat ( outram et al .
the spectral resolution ranges from fwhm = 10 in the interval @xmath16 , to 50 in @xmath17 , and to 8.5 in @xmath18 .
the line parameters have been obtained using the midas package fitlyman ( fontana & ballester 1995 ) in the spectral range 26703900 through @xmath19 minimization of voigt profiles . at shorter wavelengths ,
the lyman limit of the system at @xmath20 absorbs a large fraction of the qso flux , preventing the fit of absorption lines .
the optical ucles / aat and stis / hst spectra have been combined to simultaneously fit the lyman series of the ly@xmath1 clouds and that gives more robust results than using the ly@xmath1 line alone .
only the parameters of clouds at @xmath21 have been obtained using the absorption line alone .
the redshift distribution of the doppler parameters and of the column densities are shown in fig . 1 , together with the instrumental doppler width and column density 4@xmath22 detection limit for the lines
. the sample does not include hi column densities associated with known metal systems , and lines close ( @xmath23 ) to the quasar redshift that are considered to be affected by the ionization of the quasar .
the number of identified intervening metal systems is 5 .
three of them have a complex structure ( for a total of 10 components ) with hi column density in each system @xmath24 .
these clouds are presumably associated with galaxies and we consider them a different population of clouds with respect to lower hi column density clouds of the intergalactic medium , although many of them are presumably polluted of metals by nearby galaxies ( cowie & songaila , 1998 ) .
metal systems have been included in the sample only to compare the number density redshift evolution with previous studies at lower and higher redshifts ( see section 3 ) .
the total number of clouds excluding metal systems is then 210 .
the reported 4@xmath22 detection limit is determined by considering the error array of the spectra and does not take into account the line blending effect ; at @xmath25 it is not very important because the line density at intermediate redshifts is relatively low , but it becomes particularly strong at lower redshifts , due to the presence of lyman series lines of the higher redshift clouds that prevents detection of weak lines ( @xmath26 2 ) .
we therefore consider the completeness limit in the whole redshift range of @xmath27 2 .
1em -1.1em 1em some interesting features can be noticed in the column density redshift distribution .
there is a peak of high hi clouds at @xmath28 ( @xmath29 @xmath8 mpc comoving ) .
a metal system with a @xmath30 lyman limit was first detected from the test stis observations of the quasar .
therefore this peak might indicate the presence of a high density of galaxies at those redshifts . at lower redshifts ( in the interval @xmath31 ) , we see a region with a low density of lines .
the corresponding line density is @xmath32 , compared with an observed mean over the whole range of @xmath33 lines .
there are no lines with hi column density larger than the completeness limit of 10@xmath34 atoms/2 , while from the mean observed in the whole redshift range we would expect 4 lines .
although this is not statistically very significant , it is suggestive of the presence of a `` void '' of comoving size of 94 @xmath35 mpc .
the redshift measurements of galaxies from multi color ground observations and from the very deep stis images of the field will probably confirm if this is real .
the lack of lines can also be caused by a strong ionization of the clouds by the local uv radiation field , for instance in a quasar environment .
it is worth noting the decrement of lines going from the qso redshift to @xmath36 , shown by the histogram of the entire sample ( fig . 2 ) .
this effect disappears when selecting strong lines , and so it may be due to the detection limit variation along the spectrum . for smaller redshifts ,
the number of lines increases again .
we notice that at @xmath37 , a civ system has been tentatively identified ( ferguson et al .
1999 ) and a quasar at a distance of about 44.5 " from the line of sight ( @xmath38 @xmath35 kpc ) has been found from the ground by emmi / ntt observations ( tresse et al .
this might also be an indication of an overdensity of objects at that redshift .
1em -1em 0.1 in
the number density evolution of the forest , described by a power law of the type @xmath40 , has been studied at high redshift ( @xmath41 ) using samples of high resolution data by giallongo et al .
( 1996 ) and kim et al .
( 1997 ) . for lines with @xmath42
, @xmath43 shows a fast evolution , with @xmath44 . the analysis at low redshifts ( weymann et al .
1998 ) , based on low resolution spectroscopy of the quasar absorption line survey ( qals ) , gives in the range @xmath45 a much flatter @xmath43 , consistent with weak negative evolution ( for @xmath9 ) @xmath46 .
we note that in the two redshift regimes a variation of @xmath47 with column density threshold has been found ( larger for increasing threshold ) .
this would go the opposite direction than what is shown in fig . 2 where @xmath47 is smaller for higher threshold .
the discrepancy can be explained if a large number of lines with @xmath48 is present , but not detected at @xmath49 . -1em
-1em -0em 1em the problem in comparing high and low redshift data stems from the different techniques used to measure column densities . at high redshifts ,
high resolution observations allow line fitting of voigt profiles , and those give more robust results than the curve of growth technique used to estimate hi from equivalent width measurements of low resolution data at low redshifts . on the other hand ,
the strong evolution of line density makes the line deblending of complex structures rather difficult at high redshifts .
the forest of j2233606 lies between the two redshift regimes and so it is extremely useful to evaluate the correct connection between the two .
the results for the two samples of clouds with and without metal lines are shown in fig .
3 for lines with @xmath42 .
the data from j2233606 have been binned into three redshift ranges @xmath50 and @xmath51 corresponding roughly to the stis high resolution and medium resolution intervals and to the ucles high resolution interval .
the last two bins give a number density of lines that would better match the high and low redshift observations if metal systems are excluded .
the first bin gives a larger number of lines , @xmath52 including metal systems , compared with @xmath53 lines given the qals best fit ( weymann et al .
1998 ) . that deviation is not particularly significant , and
could indicate a peculiarity along the line of sight to j2233606 .
however , a similar excess at @xmath54 ( @xmath55 ) is also shown by the quasar um18 , which is part of the fos sample of the qals , but not included in the fitting of the number density evolution .
the excess found in j2233606 can be due to the presence of a cluster of lines at @xmath56 , as suggested in the previous section .
other possible explanations would be the blending effect by which nearby lines with column density below the threshold would artificially appear in the limited signal to noise spectrum as stronger single lines with hi above the threshold , or the presence of a small number ( @xmath57 would be enough ) of unidentified metal lines interpreted as lines .
this latter possibility looks very unlikely because even in clusters , metal lines typically appear very narrow at high resolution . -1em
-1em -0.65 in
a positive signal in the clustering properties of the forest has been found thanks to the advent of high resolution spectroscopy .
the investigation has been mainly based on the analysis of the two point correlation function ( cristiani et al .
1997 ; kim et al . 1997 ) but also on that of the power spectrum of mass density fluctuations ( hui et al . 1997 ; croft 1998 ; amendola & savaglio 1998 ) .
results have shown that the clustering is present up to @xmath58 and that the amplitude increases with increasing column density threshold of the lines .
most interesting for models of large scale structure formation is the redshift evolution found , showing larger values at smaller redshifts .
although the analysis of the two point correlation function is uncertain when the sample is too small , we report the result on the forest of j2233606 in fig .
4 for two column density thresholds .
the signal is significant in the first bin ( @xmath59 @xmath8 mpc ) and it is higher at the higher hi threshold at about the 6 and 4 @xmath22 level for @xmath60 and 13.8 respectively . in a previous study by cristiani et al.(1997 ) , the two point correlation function for @xmath61 at @xmath62 has shown a redshift evolution , being @xmath63 at @xmath64 ( consistent with no clustering ) , @xmath65 at @xmath66 , and @xmath67 at @xmath68 .
the line of sight of j2233606 shows @xmath69 at @xmath5 and confirms this evolution .
thanks to the high resolution observations of j2233606 , the analysis of the distribution of the doppler parameter of the forest can be performed for the first time at these low redshifts . at higher redshifts kim et al .
( 1997 ) have found that the median value of the doppler parameter is @xmath70 at a redshift of @xmath71 , and appeared to increase to @xmath72 at a redshift of @xmath73 for lines with @xmath74 , which has been interpreted as an increase of the temperature or kinematic broadening of the clouds .
the doppler distribution in j2233606 is shown in fig .
5 for the whole sample and for the subsample of lines ( 60% of the total ) with smaller uncertainties ( @xmath75 and @xmath76 ) .
the fit of the two gaussians ( in the interval @xmath77 ) gives in the two cases a mean value of 27 and 26 respectively ( @xmath19 is 2.4 and 1.5 ) , and these are slightly different than the median value ( 29 and 25 ) in the whole @xmath78 range .
the median values are larger if lines with @xmath74 are selected , being 31 in the two cases .
thus , the median value we find over the redshift range @xmath15 is considerably below the value of @xmath79 expected on the basis of the redshift evolution suggested by kim et al .
in contrast to the empty field used for the hdf north program , the hdfs has been selected close to a high redshift quasar , j2233606 ( @xmath0 ) .
the qso is centered in the stis field , located approximately 5 and 8 arcmin from the wfpc2 and nicmos fields respectively .
the forest of j2233606 has been observed at high and medium resolution ( fwhm = 10 , 50 and 8.5 in the wavelength intervals @xmath16 , @xmath17 and @xmath18 ) using stis / hst ( ferguson et al .
1999 ) and ucles / aat ( outram et al .
1998 ) , giving the opportunity to analyze the properties of the intergalactic medium in the large redshift interval @xmath39 .
this is the first time that the forest is studied in detail for @xmath80 .
more than 200 cloud redshifts have been found and 63 have hi column density in excess of @xmath81 atoms/2 , 11 of which have associated metal absorption .
the redshift distribution of the lines has shown a deficiency of lines ( 5 are found over a mean expected of 16 ) in the interval @xmath82 , corresponding to a comoving size of 94 @xmath8 mpc ( @xmath9 ) .
the region around the metal system at @xmath83 , also responsible of the lyman limit observed for @xmath84 , shows an excess of strong lines in the interval @xmath85 , indicating a clustering of clouds .
an excess of lines is also found for @xmath86 .
the evolution of the number density in the total sample shows a rapid decrease of lines with redshift .
this is partially due to the blending effect of low hi column density lines , which is stronger at lower redshifts .
indeed it tends to disappear when selecting only strong lines ( @xmath87 2 ) . on average
the number of strong clouds per unit redshift is higher than the extrapolation from studies at lower and higher redshifts : @xmath3 at @xmath88 including metal systems are found to be compared with 41 and 27 expected from the low and the high redshift predictions .
a larger sample of lines will confirm if this line density is typical at that redshift or if it is peculiar to the j2233606 line of sight .
the two point correlation function has shown a positive signal up to a scale of the order of 3 @xmath8 mpc .
the amplitude found seems to confirm the evolution with redshift , being larger at lower @xmath89 .
the mean doppler parameter of the forest is around @xmath90 , corresponding to a temperature , in the case of thermal broadening , of @xmath91 k. the median doppler parameter evolution reported by kim et al .
( 1997 ) is not confirmed by this forest .
amendola l. , savaglio s. , 1998 , mnras , submitted cristiani s. , dodorico , s. , dodorico , v. , fontana , a. , giallongo e. , savaglio s. , 1997 , mnras , 285 , 209 cowie l. l. , songaila a. , 1998 , nat , 394 , 44 croft r. a. c. , weinberg d. h. pettini m. , hrtnquist l. , katz n. , 1998 , apj , submitted , astro
ph/9809401 ferguson h. c. et al .
, 1999 , in preparation gardner j. p. , 1999 , in preparation giallongo e. , cristiani , s. , dodorico s. , fontana a. , savaglio s. , 1996 , apj , 466 , 46 fontana a. , ballester p. , 1995
, the eso messenger , 80 , 37 hui l. gnedin n. y. , zhang y. , 1997 , apj , 486 , 599 kim t s , hu e. m. , cowie l. l. , songaila a. , 1997 , aj , 114 , 1 outram p. j. , boyle b. j. , carswell r. f. , hewett p. c. , williams r. e. , norris r. p. , 1998 ,
mnras , submitted , astro
ph/9809404 savaglio s. , 1998 , aj , 116 , 1055 sealey k. m. , drinkwater m. j. , webb j. k. , 1998 , apjl , 499 , l135 tresse l. , dennefeld m. , petitjean p. et al .
, 1998 , a&a , submitted , astro ph/9812246 weymann r. j. , jannuzi b. t. , lu l. , et al . , 1998 , apj . , 506 , 1 williams r. e. et al . , 1999 , in preparation | the quasar in the hubble deep field south ( hdfs ) , j2233606 ( @xmath0 ) has been exhaustively observed by ground based telescopes and by the stis spectrograph on board the hubble space telescope ( hst ) at low , medium and high resolution in the spectral interval from 1120 to 10000 .
the combined data give continuous coverage of the lyman@xmath1 forest from redshift 0.9 to 2.24 .
this very large base
line represents a unique opportunity to study in detail the distribution of clouds associated with emitting structures in the field of the quasar and in nearby fields already observed as part of the hdfs campaign . here
we report the main properties obtained from the large spectroscopic dataset available for the clouds in the intermediate redshift range @xmath2 , where our present knowledge has been complicated by the difficulty in producing good data .
the number density is shown to be higher than what is expected by extrapolating the results from both lower and higher redshifts : @xmath3 lines with @xmath4 are found ( including metal systems ) at @xmath5 , to be compared with @xmath6 lines predicted by extrapolating from previous studies .
the redshift distribution of the lyman@xmath1 clouds shows a region spanning @xmath7 ( comoving size of 94 @xmath8 mpc , @xmath9 ) with a low density of absorption lines ; we detect 5 lines in this region , compared with the 16 expected from an average density along the line of sight .
the two point correlation function shows a positive signal up to scales of about 3 @xmath8 mpc and an amplitude that is larger for larger hi column densities .
the average doppler parameter is about 27 , comparable to the mean value found at @xmath10 , thus casting doubts on the temperature evolution of the clouds . 2@xmath11 |
a systematic approach to the study of the relation between configurational entropy , spatial correlations and local ordering in a fluid of elongated particles was recently outlined in @xcite , where the authors used the theoretical framework that had been originally introduced by giaquinta and giunta in order to identify the `` hidden '' signature of freezing in the structural properties of a homogeneous and isotropic fluid of hard spheres @xcite .
their approach was based on the well known multi - particle correlation expansion of the statistical entropy , first established for a finite system by h. s. green @xcite and later generalized to an infinite open system by r. e. nettleton and m. s. green @xcite : @xmath4 where @xmath5 is the excess entropy per particle in units of the boltzmann constant and the `` @xmath6-body entropies '' ( @xmath7 ) are obtained from the integrated contribution of spatial correlations between @xmath6-tuples of particles .
note that the one - body term coincides with the entropy of the corresponding ideal gas .
the pair entropy of a homogeneous and isotropic fluid of _ nonspherical _ molecules can be written as @xcite : + @xmath8 - g(r,\omega^2 ) + 1\ } { \rm d } { \bf r } { \rm d } \omega^2 \,,\ ] ] where @xmath9 is the pair distribution function ( pdf ) which depends on the relative separation @xmath10 between a pair of molecules and on the set of euler angles @xmath11 $ ] that specify the absolute orientations of the two particles in the laboratory reference frame .
the quantity @xmath12 represents the integral over the euler angles of one molecule ( @xmath13 for linear molecules ) , while @xmath14 is the particle number density .
the approach originally set up by giaquinta and giunta focused on the properties of the so - called residual multi - particle entropy ( rmpe ) : @xmath15 the quantity @xmath16 includes , by definition , the re - summed contributions of all correlations involving more than two particles @xcite .
this quantity , despite its minor quantitative relevance in the overall entropy balance , is far from being irrelevant in that it conceals significant indications on the statistical thermodynamics of the fluid that are intimately related to the role played by high - order density correlations in the system .
the noticeable feature exhibited by the rmpe , as compared to the pair entropy ( a negative - definite quantity ) , is its non - monotonic behavior .
in fact , @xmath17 starts decreasing for increasing densities until the function attains a minimum beyond which it undergoes a sharp increase up to positive values .
this behavior may come out , at first glance , somewhat unexpected in that one would probably surmise that , in analogy with the monotonic behavior of the two - body term @xmath18 , the growth of higher - order density correlations that is induced by a further compression of the system should throughout contribute to a contraction of the configurational states that are accessible to the system for smaller and smaller volumes , as is actually the case from low to intermediate densities .
instead , one observes a `` slowing down '' , for increasing pressures , in the reduction rate of the average number of states with respect to the non - interacting system in an equivalent thermodynamic condition : in fact , the positive cumulative contribution of multiparticle correlations to the entropic balance moderately damps the systematic drop that is caused by the dominant pair term .
this effect is suggestive of the emergence of a new structural condition in the system .
more specifically , it indicates that multiparticle correlations have started to build up some type of ordering , on a local scale , as a result of compelling entropic constraints .
the associated signature is the vanishing of the rmpe : _ a posteriori _ , it is found to herald in a very sensitive way the occurrence - in a proximate range of densities and/or temperatures - of a phase transition of the fully disordered fluid into a more structured phase . by now
, the reliability of such a one - phase ordering criterion has been extensively documented for a variety of model fluids @xcite , also in two dimensions @xcite . with specific reference to a fluid of hard spherocylinders , i.e. , cylinders of length @xmath0 and diameter @xmath1 capped at both ends with two hemispheres , preliminary evidence was reported in @xcite on the rmpe indication concerning the isotropic - nematic ( i - n ) transition . in the case of simple fluids composed of spherical particles a number of one - phase criteria have been introduced in the past with the aim of locating the transition from the liquid to the solid phase without resorting to the knowledge of the free energy of the two coexisting phases @xcite . however , no such simple criteria are available for nematic fluids . in this respect ,
the zero - rmpe criterion stands as , perhaps , a unique exception in that it can be extended in a straightforward way to fluids composed of elongated particles to deal with the transition from the homogeneous and isotropic fluid to a liquid - crystalline phase .
since the seminal work of onsager @xcite , the model of hard spherocylinders has been diffusely investigated , notwithstanding its simplicity , in order to study excluded - volume effects in real liquid crystals .
in fact , hard spherocylinders exhibit depending on the value assumed by the aspect ratio @xmath19 a surprising variety of mesophases @xcite . in @xcite constant - pressure monte carlo simulations
were carried out for @xmath20 only , a case where the disordered fluid undergoes , under compression , a first - order transition to a partially ordered nematic phase . in this paper
we have extended the above calculations to lower values of the aspect ratio in the range @xmath21 , thus encompassing a region where the homogeneous and isotropic fluid undergoes a transition to a fully ordered solid ( s ) phase as well as to the smectic - a ( sma ) phase , a layered orientationally ordered phase characterized by a one - dimensional density modulation along the direction @xmath22 of the molecular alignment @xcite .
we have carried out simulations also in the limit of very elongated particles , specifically for @xmath23 , in order to analyze the behavior of the rmpe in a regime where the i - n transition point can be determined , asymptotically , in an exact way @xcite .
this is also the range of aspect ratios relevant for the modelling of monodisperse aqueous suspensions of rodlike particles like the _ tobacco mosaic virus _ ,
a plant virus which shows an isotropic - nematic transition as well as a smectic phase at higher concentrations @xcite .
the paper is organized as follows : the theoretical framework and the numerical simulation technique , as applied to the case of cylindrical molecules with head - tail symmetry , are described in sects . 2 and 3 , respectively .
the results are presented in sect .
4 while sect . 5
is devoted to concluding remarks .
in order to evaluate the rmpe from eq .
( [ s : sex ] ) , one needs to calculate the excess and pair entropies of the fluid .
the former quantity can be obtained by integrating the equation of state ( eos ) along a thermodynamic path : @xmath24 \frac{{\rm d}{\rho'}}{\rho ' } \,.\ ] ] in eq .
( [ s : free_ex ] ) @xmath25 is the pressure , @xmath26 is the inverse temperature in units of the boltzmann constant , and @xmath27 is the density of a suitably chosen reference state .
the eos of the fluid was obtained through a monte carlo sampling of the system at different pressures , while the quantity @xmath28 was calculated using the widom test - particle insertion method @xcite at low enough densities . as for the pair entropy ,
this quantity was obtained , according to eq .
( [ eq : s2mol ] ) , after integration of the pdf which , in turn , was decoupled in the form @xcite : @xmath29 where @xmath30 is the radial distribution function of the molecular centers of mass that is obtained after integrating the full pdf over the set of euler angles , while @xmath31 represents the conditional distribution function , normalized to @xmath32 , for the relative orientation of a pair of molecules whose centers of mass lie at a distance @xmath10 . using eq . ( [ gtot ] ) ,
the pair entropy was separated into the sum of a translational ( @xmath33 ) and an orientational part ( @xmath34 ) , after sorting out an additional excluded - volume contribution which arises in eq .
( [ eq : s2mol ] ) from the integration over space regions where @xmath35 : @xmath36 in eq .
( [ s : s2 ] ) , @xmath37 is the second virial coefficient of hard spherocylinders : @xmath38 the translational contribution is formally analogous to the pair entropy of a simple fluid with no rotational degrees of freedom : @xmath39 { \rm d}{\bf r } \,,\ ] ] where the integration is carried out over non - overlapping regions . as for the orientational term ,
one finds : @xmath40 where @xmath41 { \rm d } \omega^2 \;.\ ] ] even for uniaxial molecules , the computation of the full orientational correlation function @xmath31 and its subsequent integration is a formidable task . in this specific case
there are only four independent degrees of freedom which can be fixed by providing , e.g. , the relative intermolecular distance @xmath10 and three angles @xcite . without loss of generality , we can hold the center of mass of , say , particle @xmath42 fixed at the origin of the laboratory coordinate system and let its molecular axis @xmath43 lie along the @xmath44-axis . then
, the position and orientation of particle @xmath45 can be unambiguously determined by specifying , for instance , two polar coordinates of its center of mass , i.e. , the distance @xmath10 and the angle @xmath46 formed by the intermolecular axis with the @xmath44-axis , and the euler angles @xmath47 and @xmath48 of the @xmath49 axis ( the `` twist '' angle , @xmath50 , is redundant for uniaxial molecules ) . in order to simplify the whole scheme , we neglected the dependence of @xmath31 on @xmath46 and @xmath48 , thus focusing on the information associated with the angle representing the relative orientation of the two molecular axes , @xmath51 .
indeed , this is the critical parameter that should be monitored in order to assess the onset of either nematic or smectic order in the fluid .
the contracted orientational distribution function @xmath52 is normalized in such a way that : @xmath53 we do not expect that such a simplification may heavily bias the estimate of the weight associated with orientational correlations in the _ homogeneous _ and _ isotropic _ phase .
indeed , the quantity we are ultimately interested in is just an integral measure that can help to assess the relative balance between the positional and angular contributions that intervene in the configurational entropy .
we performed monte carlo ( mc ) simulations of hard spherocylinders at constant pressure .
this choice is rather common for systems composed of non - spherical hard - core particles for which it may be difficult to calculate the equation of state through the contact value of the pdf , as is usually done in a constant - volume simulation .
furthermore , density fluctuations are more naturally accommodated in a constant - pressure simulation @xcite . the number of particles , that were enclosed in a cubic box with periodic boundary conditions , typically ranged between @xmath54 ( @xmath55 ) and @xmath56 ( @xmath57 ) , with an intermediate value of @xmath58 particles for @xmath59 . a larger sample composed of @xmath60 particles
was used for more elongated particles with @xmath23 .
a perfectly ordered , low - density lattice arrangement was the starting configuration .
then , the system was allowed to equilibrate , initially along a constant - volume mc path . a sequence of states
was then generated at constant pressure , starting from an equilibrated isotropic configuration .
we carried out rather extended simulations , as compared with those currently reported in the literature , especially in proximity of the transition threshold .
in fact , after an equilibration period of typically @xmath61 to @xmath62 mc cycles , simulation data were obtained by generating chains consisting of @xmath63 to @xmath64 mc cycles , depending on the pressure .
a mc cycle consisted of an attempt of changing , sequentially , both the position and the orientation of each molecule , followed by an attempt of changing the volume of the sample .
during such a procedure , each molecule was first displaced from its original position , by moving at random its center of mass within a cube .
the molecule was then re - oriented within a cone centered about the molecular axis .
the overall acceptance probability for the combined move was adjusted to about 50% .
the isotropic change of the simulation box was obtained by varying at random the box side @xmath0 .
finally , the maximum displacement allowed was set so as to achieve an average acceptance probability for volume changes of @xmath65% .
for hard - core molecules , the usual metropolis acceptance criterion reduces to a check of particle overlaps after each monte carlo move . in this regard
, a simple and efficient algorithm can be used for spherocylinders @xcite . during
the cumulation runs , we monitored the average density and built histograms for the translational and orientational distribution functions , respectively ; equilibrium averages and standard deviations were computed by dividing such runs into 5 to 10 independent blocks . in the following , we shall refer to the volume packing fraction @xmath66 , where @xmath67 is the volume occupied by one spherocylinder , and to the reduced pressure @xmath68 .
we sampled the radial distribution function @xmath30 at intervals of @xmath69 , while the orientational distribution @xmath52 was calculated with a grid spacing of @xmath70 .
we investigated the phase diagram of the model up to the threshold of either the nematic or the smectic phase .
the onset of liquid - crystalline order was monitored through the nematic order parameter : @xmath71 where @xmath72 is the angle formed by the molecular axis of the @xmath73-th molecule @xmath74 ( the sum running over all particles in the system ) .
the director @xmath75 defines the alignment direction attained , on average , by the molecules .
since this direction is not known _ a priori _ , the order parameter is usually determined by diagonalizing the traceless , symmetric , second - rank tensor @xmath76 that is defined in terms of the molecular axes as : @xmath77 where @xmath78 is the unit tensor .
the nematic order parameter is then obtained as the largest eigenvalue , and the director as the corresponding eigenvector @xcite .
the value attained by @xmath79 is close to zero in the isotropic phase and tends to one in a highly ordered phase . across the ordering transition , both the density and the order parameter
show a finite jump
. however , such discontinuities , that are associated with a first - order phase transition , are in certain cases so small that a closer investigation of the correlation functions or a direct observation of the arrangement of the molecules in the sample may be necessary on the thermodynamic edge of both the nematic and smectic phase . as already anticipated , we used widom s insertion method @xcite to obtain independent numerical estimates of the excess entropy at low densities , @xmath80 , to be used in eq .
( [ s : free_ex ] ) .
the widom technique provides a simple scheme for calculating the excess chemical potential , @xmath81 , of not too dense a liquid via the calculation of the average boltzmann factor associated with the ( never accepted ) random insertion of an additional particle in an @xmath82-particle system .
the excess entropy can then be obtained through the thermodynamic equation : @xmath83 depending on the density and the elongation of the spherocylinders , 50 to 200 trial insertions per mc step were typically needed in order to obtain a reliable estimate of @xmath81 .
the use of this technique at small but finite densities avoids the otherwise necessary extrapolation to the ideal gas limit , as a reference state for the integration contemplated in eq .
( [ s : free_ex ] ) .
this latter procedure was adopted in @xcite , where some spurious fluctuations of the orientational distribution function as well as some difficulties associated with the simulation in the highly dilute regime were registered . as sketched in fig .
[ f1 ] , the values obtained for @xmath5 using widom s insertion method are quite stable in all the cases we have examined : the thermodynamic potential relaxes almost immediately , with small fluctuations about the average value .
no significant drifts were observed during simulations as long as @xmath84 mc steps .
the simulation parameters and the results for the relevant properties of the fluid at low densities were collected for various elongations in table [ t1 ] .
in this subsection we shall focus on the phase behavior of the model in a range of parameters where , for increasing values of the aspect ratio , the disordered fluid undergoes a transition to either a solid , a smectic or a nematic phase .
we investigated the model for shape anisotropies @xmath85 and @xmath86 . for @xmath57 , the fluid freezes into a crystalline phase .
an i - sma phase transition is observed over a range of larger aspect ratios bounded by an isotropic - smectic - solid ( i - sma - s ) triple point ( @xmath87 ) and by an i - n - sma triple point ( @xmath88 ) , respectively . for even larger anisotropies , the disordered fluid forms , under compression , a nematic phase @xcite .
the reliability of the present results for the eos and for the nematic order parameter can be appraised through fig .
[ f2 ] where the above quantities were plotted as a function of the packing fraction for @xmath89 and @xmath86 .
the comparison with the simulation data reported by mcgrother and co - workers @xcite shows that a system with @xmath54 particles is fairly adequate for evaluating the basic thermodynamic properties of the fluid .
the agreement turned out to be equally good along the disordered branch also for @xmath57 and @xmath90 , up to the transition point .
we now turn to a discussion of the properties of the model for @xmath20 , a case where the isotropic fluid undergoes a rather weak first - order transition to a thermodynamically stable nematic phase at about @xmath91 of the close - packing density , a value corresponding to an absolute packing fraction of @xmath92 @xcite .
no specific feature indicating the transition edge is apparent either in the equation of state or in the order parameter ( see fig . [ f2 ] ) .
a rather modest degree of positional order shows up in the spatial profile of the radial distribution function that was plotted in fig .
[ f3 ] for increasing pressures , all the way up to the transition threshold . actually , for moderately small packing fractions , the local density around one generic spherocylinder
is significantly reduced , with respect to the homogeneous value taken at large distances , for interparticle separations lower than @xmath93 molecular diameters , a distance which corresponds to the minimum contact separation between two orthogonal spherocylinders .
retrospectively , this local `` depletion '' effect is more easily understood if one looks at the orientational distribution function ( see fig . [ f4 ] ) : the almost flat profile of @xmath52 for @xmath94 , when plotted as a function of @xmath47 , implies that orthogonal geometries in the relative spatial arrangement of two spherocylinders are equally probable as parallel configurations are .
the steric hindrance produced , on average , by such orthogonal ( or almost orthogonal ) particles sensitively reduces the probability of finding other molecules at shorter distances from the central one .
upon further compressing the fluid , the value of @xmath52 for aligned - particles geometries ( @xmath95 ) steadily increases while different geometries become , at the same time , rarer and rarer .
this is apparent from the progressive and systematic reduction of @xmath52 , for short and intermediate molecular separations , over a wide range of angles centered about @xmath96 .
correspondingly , the radial distribution function becomes more and more structured at short distances as is distinctly shown in fig .
[ f3 ] by the building up of a first coordination shell formed by particles lying almost at close contact with the central one .
the excess entropy of the fluid was resolved in fig .
[ f5 ] in terms of the contributions associated with pair and more - than - two - particle correlations , respectively ( see eq . ( [ s : sex ] ) ) .
the behavior of the rmpe is similar to that exhibited by the same quantity in a variety of other model systems , composed of spherical particles , across an ordering transition such as the freezing of a fluid @xcite or the phase separation of a binary mixture @xcite . in particular
, this quantity vanishes for a value of the packing fraction ( @xmath97 ) which falls precisely on top of the currently estimated i - n phase transition point ( see table [ t2 ] where we summarized the predictions of the entropy - based criterion for the presently investigated values of the aspect ratio ) .
we note that the current rmpe estimate sensitively improves on that formerly given in @xcite where the values reported for the pair entropy turned out to be affected by a systematic computational error . however , no significant morphological change resulted , _ a posteriori _ , in both the pair entropy and in the rmpe .
we also recall that the authors had shown in @xcite that the pair entropy changes with the size of the sample , for a given density , in a tiny but systematic way , the zero - rmpe density shifting towards moderately larger values .
the rmpe inverts its trend ( from decreasing to increasing ) in the density region ( @xmath98 ) corresponding to a more extended structuring of the fluid at short distances ( see fig .
[ f3 ] ) . to be specific
, we refer to the growing of a local nucleus of ( mostly ) aligned particles that is signalled by the emergence of a secondary maximum in the radial distribution function at a relative distance twice as large as that corresponding to the first coordination shell .
figure [ f5 ] also shows the pair entropy of the fluid resolved , according to eq .
( [ s : s2 ] ) , into the sum of a ( spatially independent ) excluded - volume term ( @xmath99 ) plus two more contributions arising from translational and orientational correlations , respectively . as expected , it is @xmath34 that drives the rather sharp increase of @xmath100 just beyond the minimum . _
vice versa _ , the contribution associated with positional correlations , @xmath33 , is very small all over the fluid range .
the function which , upon integration ( see eq .
( [ s : s2or ] ) ) , yields the quantity @xmath34 is shown , for increasing densities , in the right part of fig .
the value of the corresponding integral is negative and keeps close to zero at low and intermediate densities ( see fig . [ f5 ] ) . for packing fractions larger than @xmath101
, @xmath34 starts decreasing more and more rapidly and eventually overcomes ( in absolute value ) the excluded - volume term . as seen from fig .
[ f5 ] , the fingerprint of the ordering of the fluid is manifest in the persistence of angular correlations up to intermediate and long distances .
considerations similar to those developed above for @xmath20 apply also to the @xmath89 case , even if it is a smectic phase that is now formed by the isotropic fluid .
however , the basic underlying mechanism which drives the first - order i - sma transition is analogous to that exploited in the nematic case .
obviously , the denser isotropic fluid looks more structured than for @xmath20 ( see fig . [ f6 ] ) .
furthermore , at the transition point angular correlations show a very slow spatial decay .
we also note that in this case the lower molecular asymmetry modifies the scale of interparticle distances in the average density profile .
thus , the effect associated with the structuring of the second coordination shell , which we already commented upon for @xmath20 , is less resolved since the corresponding distance falls very close to the position of the lone weak maximum , located at a distance @xmath102 , that is present in the radial distribution function at low densities
. a shoulder , rather than a peak , emerges at @xmath103 , again in a density range corresponding to the position of the minimum ( @xmath104 ) attained by the rmpe ( see fig . [ f7 ] ) . in general , the ordering thresholds detected through the vanishing of the rmpe
correlate very well with the corresponding transition point ( see table [ t2 ] ) , whatever the nature of the higher - density phase coexisting with the isotropic fluid .
the evidence discussed so far justifies the claim that the rmpe of spherocylinders monitors the leading microscopic process which drives the phase transition in all such cases , i.e. , the aligning of spherocylinders along a common direction .
as such , the vanishing of this quantity presumably portends the density threshold beyond which the nematic ordering may possibly emerge in the system , regardless of other concurrent structural phenomena like those which may ultimately lead to the formation of a more ordered macroscopic state , be it smectic or crystalline , instead of a pure nematic phase .
an indication consistent with this thesis that is , perhaps , not accidental ( yet to be considered with due caution because of the numerical uncertainty arising from the computational error which affects , at least , the third significant figure ) is offered by the apparently greater accuracy of the predictions given by the rmpe - based criterion for @xmath59 and @xmath86 ( see table [ t2 ] ) , viz . , in those cases where the stable coexisting phase is a truly nematic fluid . in the other two cases investigated , i.e. , when either a solid or a smectic phase is formed , the criterion just overshoots the target , locating the internal threshold for the ordering of the fluid at a slightly higher density .
as the aspect ratio increases , the i - n transition point moves to lower and lower densities . in the onsager limit of infinitely long rods ( @xmath105 ) ,
a fluid of hard spherocylinders undergoes a first - order phase transition to a nematic phase for @xmath106 @xcite .
however , the phase behavior characteristic of the onsager regime qualitatively sets in for shape anisotropies even less than @xmath23 , even if for such values of the aspect ratio the basic assumption made by onsager that all virial coefficients of order higher than two can be neglected is not yet satisfied @xcite .
in fact , the contribution of the third virial term to the free energy , evaluated for @xmath23 at the transition density , turns out to be still comparable to the contribution of the second virial term @xcite .
we carried out simulations for @xmath107 . in passing ,
we recall that this value mimics the typical size of the tobacco mosaic virus , whose particles have a practically rigid rod shape of @xmath108 nm diameter and @xmath109 nm length @xcite . for such large values of the aspect ratio , the simulation of a fluid of hard spherocylinders dictates more severe constraints on the size of the simulation box which should be large enough to accommodate at least two rod lengths , so as to avoid multiple overlaps between a molecule and the periodic images of another particle @xcite .
the number of particles chosen ( @xmath110 ) is sufficient to ensure that the condition mentioned above is fulfilled in the ( relatively low ) density range where the disordered phase is thermodynamically stable . the reduced pressure and the nematic order parameter are shown in fig .
the isotropic liquid phase is mechanically stable up to a reduced pressure @xmath111 : in this regime the system shows , for simulations as long as @xmath112 mc cycles , no significant fluctuations about the average values of the packing fraction ( @xmath113 ) and of the nematic order parameter ( @xmath114 ) , respectively .
upon further compressing the system up to @xmath115 , we observed , after about @xmath116 mc cycles , a spontaneous and rapid transformation into a nematic phase which , at equilibrium , settled down at a packing fraction @xmath117 and was characterized by an order parameter @xmath118 .
the resulting `` boundaries '' are slightly shifted toward higher densities if compared with the direct estimates reported in @xcite , i.e. , @xmath119 or @xmath120 , and @xmath121 or 0.172 , according to whether the gibbs - duhem integration or a gibbs - ensemble technique was used .
exploratory runs performed by decompressing the system from the ordered phase distinctly show an hysteresis , the fluid remaining almost nematic for pressures as low as @xmath122 .
we calculated the pressure also using the `` decoupling approximation '' between the translational and orientational degrees of freedom originally introduced by parsons @xcite .
in this approximation a system of rods is mapped onto a system of spheres interacting with an effective orientation - dependent potential .
as already observed by mcgrother and co - workers for the shorter elongation regime @xmath123 that was analyzed in @xcite , the resulting eos turns out to be fairly accurate in the isotropic phase up to the transition point , and definitely better than the prediction based on the scaled - particle - theory approach exploited by boublik @xcite .
however , it should be noted that the latter approach is based on a one - parameter approximation for the third and fourth virial coefficients which can not be expected to yield accurate results for large elongations @xcite . a better agreement with the simulation data is obtained with a closely related description that was proposed by nezbeda @xcite :
this approximation is based on an empirical fit of the virial coefficients , especially tailored on the available results for hard spherocylinders .
all the above approximations for the eos were reported , for a comparison , in fig .
[ f8 ] : note that the simulation data are closely bracketed by the parsons and nezbeda approximations , respectively .
radial correlations are shown in fig .
[ f9 ] together with the function whose space integral yields @xmath124 . all over the liquid density range ,
the radial distribution function is practically structureless apart from the correlation hole at short distances ( see fig . [ f9 ] ) . as a result
, the translational contribution to the pair entropy keeps close to zero all over the explored density range ( see fig .
[ f10 ] ) .
the same is also true for the orientational contribution for packing fractions less than about @xmath125 .
it thus follows that the pair entropy of the isotropic fluid is substantially equal to the excluded - volume term @xmath99 , as is actually assumed in the original onsager theory . at variance with the positional term ,
the angular contribution is responsible , together with the excluded - volume term , of the vanishing of the rmpe . just beyond the transition point
, @xmath126 blows up and definitely overcomes the excluded - volume term ( see fig . [ f10 ] ) .
the resulting rmpe shows the typical behavior already observed for smaller aspect ratios .
also in this case , the zero - rmpe threshold is in very good agreement with the i - n transition point estimates given in @xcite ( see table [ t2 ] ) .
we also reported in fig .
[ f10 ] the parsons and nezbeda analytical estimates for the excess entropy .
as already pointed out for the eos , both approximations turn out to predict fairly well , almost with the same degree of accuracy , also the excess entropy of the fluid . as seen from figs .
[ f5 ] , [ f7 ] and [ f10 ] , the overall importance of the cumulative contribution of higher - order density correlations to the excess entropy decreases with increasing shape anisotropies .
this is clearly a consequence of the progressive confinement of the isotropic phase to lower and lower densities . in this regard
, we just observe that the value attained by the rmpe at its minimum drops from @xmath127 of the excess entropy for @xmath89 to @xmath128 for @xmath23 .
this behavior is consistent with the numerical findings on the virial expansion cited above and with the asymptotic trend predicted by onsager as @xmath105 .
note , however , that at low densities the ratio @xmath129 moderately increases with @xmath19 for a given packing fraction .
actually , longer and thinner particles give rise to more extended multiparticle correlations which trigger the ordering of the fluid at lower and lower densities . beyond the zero of the rmpe , where the nematic phase is thermodynamically stable , the present data monitor _ de facto _ a system in a metastable , macroscopically isotropic condition whose proper description lies beyond the standard onsager theory .
in this paper we have analyzed the contributions to the `` pair entropy '' @xmath18 of hard spherocylinders of length @xmath0 and diameter @xmath1 that are associated with excluded - volume , translational and angular correlations , with particular emphasis on the thermodynamic behavior exhibited by the above quantities on approaching the phase transition from the low - density fully disordered phase to a partially ordered phase at higher densities .
an elucidating synthesis of the results obtained for @xmath18 through an intensive monte carlo sampling of the model is offered by fig . [ f11 ] where we plotted the three contributions cited above , each of them being calculated in the state corresponding to the vanishing of the residual multi - particle entropy ( rmpe ) for an assigned value of @xmath19 .
the rmpe is a quantity that gauges the cumulative weight of spatial correlations involving _ more than two particles _ in the configurational entropy of the fluid .
figure [ f11 ] neatly indicates the relative importance of the three contributions in the overall entropic balance : for very small asymmetries , the translational term , @xmath33 , prevails ( in absolute value ) while being comparable with the excluded - volume contribution , @xmath99 ( note that the data for @xmath130 do obviously refer to the hard - sphere case @xcite ) . for increasing values of the aspect ratio the angular term , @xmath126 , acquires more and more importance : in fact , an interpolation of the available data shows that the three contributions reach the same magnitude for @xmath131 . actually , it is in this range of values ( more precisely , for @xmath132 @xcite ) that the `` rotator phase '' ( i.e. , the plastic crystal ) disappears in the phase diagram of hard spherocylinders , and is replaced at high densities by an orientationally ordered , crystal phase . as seen from fig . [ f11 ] , in the range @xmath133 the ordering of the fluid is mainly controlled by angular correlations .
in fact , the translational contribution rapidly tends to smaller and smaller values and practically vanishes for @xmath134 .
also @xmath34 asymptotically tends to zero in this range of increasing asymmetries but , as seen from fig .
[ f11 ] , the decay of this quantity is much slower than that shown by its translational counterpart .
correspondingly , the excluded - volume term becomes the dominant contribution to the configurational entropy of the fluid .
note that its value for @xmath23 ( @xmath135 ) is already very close to the onsager limit ( @xmath136 ) .
as far as the quality of the `` predictions '' offered by the zero - rmpe criterion are concerned , the agreement with the properly defined thermodynamic thresholds is impressive for all the asymmetries that have been investigated .
indeed , the vanishing of the rmpe monitors , in a very sensitive and reliable way , the ordering of the homogeneous and isotropic fluid into a more ordered phase be it nematic , smectic or solid
. however , we remark that such a close quantitative correspondence should not be necessarily and systematically expected _ a priori_. in fact , the zero of @xmath16 needs not coincide with the thermodynamic edge of the disordered phase that is properly located through the comparison of the free energies of the two coexisting phases .
actually , the rmpe conveys a type of information that is intermediate between the fully macroscopic level and the underlying microscopic description of the system .
the interpretation of such a simply accessible information ( and , more specifically , of its change of sign ) as an intrinsic signature of the incipient ordering of the fluid appears to be quite firmly established on the basis of an ample and coherent evidence that has , by now , emerged in a variety of continuum and lattice model fluids @xcite .
the present analysis of the `` fine structure '' of the configurational entropy of hard spherocylinders further corroborates the sensitivity of the rmpe to local ordering phenomena which may even turn out into a rather weak first - order phase transition .
we believe that this one - phase entropic criterion can be conveniently generalized to account for more than one phase transition , leading the system to increasing levels of self - organization .
an attempt in this direction has been recently done by cuetos and coworkers whose results seem to indicate the existence of a jump in the translational part of the pair entropy as well as in its rate of change across the nematic - smectic transition @xcite . however , this evidence was not considered conclusive because of the limited size of the simulation box .
we are also currently investigating this point with promising preliminary results .
we would like to thank professor bruno martinez - haya for useful discussions and for pointing a numerical inconsistency out in a previously published article .
d. c. gratefully acknowledges the hospitality and advice of professor m. p. allen during a visit to the university of bristol .
99 d. costa , f. saija , and p. v. giaquinta , chem .
. lett . * 283 * , 86 ( 1998 ) ; _ erratum _ : * 299 * , 252 ( 1999 ) .
p. v. giaquinta and g. giunta , physica a * 187 * , 145 ( 1992 )
. h. s. green , _ the molecular theory of fluids _ ( north - holland , amsterdam , 1952 ) .
r. e. nettleton and m. s. green , j. chem .
* 29 * , 1365 ( 1958 ) .
t. lazaridis and m. e. paulaitis , j. phys .
chem . * 96 * , 3847 ( 1992 ) ; * 98 * , 635 ( 1994 ) .
f. saija , s. prestipino and p. v. giaquinta , j. chem .
* 115 * , 7586 ( 2001 ) , and references contained therein .
f. saija and p. v. giaquinta , j. phys .
b * 106 * , 2035 ( 2002 ) . f. saija , s. prestipino and p. v. giaquinta , j. chem . phys . * 113 * , 2806 ( 2000 ) .
p. a. monson and d. a. kofke , adv .
phys . * 115 * , 113 ( 2000 ) .
l. onsager , ann .
sci . * 51 * , 627 ( 1949 ) .
d. frenkel , j. phys .
chem . * 92 * , 3280 ( 1988 ) .
d. frenkel , h. n. w. lekkerkerker , and a. stroobants , nature * 332 * , 822 ( 1988 ) .
j. a. c. veerman and d. frenkel , phys .
a * 41 * , 3237 ( 1989 ) .
s. c. mcgrother , d. c. williamson and g. jackson , j. chem .
phys . * 104 * , 6755 ( 1996 ) .
p. bolhuis and d. frenkel , j. chem .
phys . * 106 * , 666 ( 1997 ) .
h. n. w. lekkerkerker , ph .
coulon , r. van der haegen , and r. deblieck , j. chem .
phys . * 80 * , 3427 ( 1984 ) .
d. c. williamson and g. jackson , molec .
* 83 * , 603 ( 1994 ) .
j. p. straley , mol .
* 22 * , 333 ( 1973 ) ; * 24 * , 7 ( 1973 ) .
d. frenkel , j. phys .
chem . * 91 * , 4912 ( 1987 ) ; * 92 * , 5314 ( e ) ( 1988 ) for a review see s. fraden in proceed . of the nato - asi _ observation , prediction and simulation of phase transitions in complex fluids _
, m. baus , l. f. rull and j. p. ryckaert eds .
( kluwer , dordrecht , 1994 ) ; n. urakami , m. imai , y. sano , and m. takasu , j. chem . phys . * 111 * , 2322 ( 1999 ) ; h. graf and h. l@xmath137wen , phys . rev .
e * 59 * , 1932 ( 1999 ) .
b. widom , j. chem . phys . * 39 * , 2808 ( 1963 ) .
d. frenkel and b. smit , _ understanding molecular simulation _ ( academic press , london , 1996 ) .
t. lazaridis and m. karplus , j. chem . phys . *
105 * , 4294 ( 1996 ) .
see , for example : a. k. soper , c. andreani , and m. nardone , phys .
e * 47 * , 2598 ( 1993 ) .
m. p. allen , g. t. evans , d. frenkel , and b. m. mulder , adv .
* 86 * , 1 ( 1993 ) . c. vega and s. lago , comp .
chem . * 18 * , 55 ( 1994 ) .
j. d. parsons , phys .
a * 19 * , 1225 ( 1979 ) .
t. boublik , mol . phys . * 42 * , 209 ( 1981 ) . for a review see : t. boublik and i. nezbeda , collect .
comm . * 51 * , 2301 ( 1986 ) .
i. nezbeda , chem .
. lett . * 41 * , 55 ( 1976 ) .
j. s. van duijneveldt and m. p. allen , molec .
* 92 * , 855 ( 1997 ) .
a. cuetos , b. martinez - haya , l. f. rull , and s. lago , j. chem . phys . * 117 * , 2934 ( 2002 ) .
lllllllcl + @xmath19 & @xmath138 & @xmath139 & @xmath140 & @xmath141 & @xmath142 & mc cycles & trials & ratio ( % ) + + + 3 & 0.30 & 0.131(1 ) & 2.33(1 ) & -1.039(13 ) & -0.959(2 ) & @xmath143 & @xmath144 & 9.7 + 3.2 & 0.30 & 0.130(1 ) & 2.37(1 ) & -1.054(17 ) & -0.971(3 ) & @xmath145 & @xmath146 & 9.4 + 4 & 0.30 & 0.125(1 ) & 2.55(1 ) & -1.149(6 ) & -0.996(1 ) & @xmath143 & @xmath147 & 7.8 + 5 & 0.181 & 0.089(1 ) & 1.90(1 ) & -0.878(6 ) & -0.834(1 ) & @xmath148 & 20 & 14.9 + & 0.379 & 0.136(1 ) & 3.25(1 ) & -1.457(24 ) & -1.237(4 ) & @xmath143 & 200 & 3.8 + 20 & 0.10 & 0.044(1 ) & 2.39(1 ) & -1.137(13 ) & -1.066(1 ) & @xmath143 & 50 & 9.2 + & 0.20 & 0.066(1 ) & 3.79(1 ) & -1.785(26 ) & -1.588(5 ) & @xmath149 & 200 & 2.2 + + lcccc + & & & + + @xmath19 & phase transition & @xmath97 & ref .
@xcite & ref .
@xcite + + + 3 & i - s & 0.521 & 0.490 & 0.512 + 3.2 & i - sma & 0.510 & 0.513 & 0.503 + 4 &
i - n & 0.463 & 0.472 & 0.466 + 5 &
i - n & 0.399 & 0.407 & 0.398 + 20 &
i - n & 0.148 & & 0.139 + + | hard spherocylinders ( cylinders of length @xmath0 and diameter @xmath1 capped at both ends with two hemispheres ) provide a suitable model for investigating entropy - driven , mesophase formations in real colloidal fluids that are composed of rigid rodlike molecules .
we performed extensive monte carlo simulations of this model fluid for elongations in the range @xmath2 and up to @xmath3 , in order to investigate the relative importance of translational and orientational correlations allowing for the emergence of nematic or smectic order in the framework of the so - called residual multi - particle entropy ( rmpe ) .
the vanishing of this quantity , which includes the re - summed contributions of all spatial correlations involving more than two particles , signals the structural changes which take place , at increasing densities , in the isotropic fluid .
we found that the ordering thresholds detected through the zero - rmpe condition systematically correlate with the corresponding phase - transition points , whatever the nature of the higher - density phase coexisting with the isotropic fluid .
pacs 64.60.-i , 64.70.md , 61.20.ja , 65.40.gr |
antimatter search in cosmic rays ( crs ) are especially important for astrophysics and dark matter signal study . in recent years , there were great progresses in the measurements of cr antimatter particles .
in particular , the alpha magnetic spectrometer ( ams-02 ) launched in 2011 has provided unprecedentedly precise measurement of crs .
based on the results from ams-02 and previous experiments , such as pamela and fermi - lat , the properties of cr antimatter particles have been extensively studied in a quantitative way in the literature . recently , the ams-02 collaboration has reported their measurement of the @xmath0 ratio up to @xmath1 gev with a high precision @xcite .
the previous pamela @xmath0 ratio data @xcite seemed to be compatible with the conventional cr prediction in the energy range of @xmath4 gev , and can be used to set limits on the dm annihilation cross section or lifetime @xcite .
however , it is very interesting to note that the ams-02 @xmath0 ratio becomes almost flat from @xmath5 gev to several hundred gev .
this behavior is not expected from the cr secondary production since the measured b / c ratio decreases with energy as a power - law spectrum without any structure .
therefore it is necessary to perform a careful study on the secondary production of antiprotons taking all the uncertainties into account .
the major challenge to interpret the ams-02 data is how to give an accurate prediction of the secondary antiprotons .
although the ams-02 collaboration has provided precise measurements of the proton and helium spectra , there are still large uncertainties from cr propagation processes
. the complicated cr propagation in the galaxy may involve diffusion , energy loss , convection , and reacceleration effects .
the propagation parameters describing these effects can be determined by fitting the secondary - to - primary ratios , such as b / c and ( sc+ti+v)/fe , and unstable - to - stable ratios of secondary particles , such as @xmath6 and @xmath7 @xcite .
however , current measurements of these ratios are still not sufficient to distinguish different propagation models .
another important uncertainty comes from the solar modulation , which significantly affects low energy cr spectra , but it is not easy to quantify the effects . in this work , we give first a careful analysis of the secondary antiproton production .
the galprop package is used to solve the cr propagation equation @xcite .
we adopt two kinds of propagation models , namely the diffusion - reacceleration ( dr ) model and diffusion - convection ( dc ) model .
propagation parameters are determined by fitting the available b / c data with a markov chain monte carlo ( mcmc , @xcite ) algorithm @xcite , which is powerful for surveying high dimensional parameter space .
secondary antiprotons from cr interactions with the ism are calculated and the uncertainties from propagation effects are obtained .
taking all the uncertainties into account , the secondary antiprotons seem marginally consistent with ams-02 results , but show clearly different slope with energy
. then we study the implications of the ams-02 @xmath8 data on dm annihilation signals .
since the ams-02 data do not unambiguously indicate an excess , we derive upper limits on the dm annihilation rate and signal regions where the background plus dm signals give a good fit to the ams-02 data . in order to derive upper limits or signals regions ,
we carry out a global fitting to the ams-02 data with the mcmc method . based on the ams-02 @xmath0 data , upper limits on the dm annihilation rate and favored dm annihilation parameters
have also been given in refs .
@xcite .
this paper is organized as follows . in sec .
ii , we describe the cr propagation processes . in sec .
iii , we first determine the propagation parameters from the ams-02 b / c ratio data and then calculate the secondary antiproton flux . in sec .
iv , we study implication on dm annihilation by the ams-02 @xmath0 ratio data .
finally , we give the discussions and conclusions in sec . v.
after accelerated in sources , galactic crs are injected and diffuse in the interstellar space and suffer from several propagation effects before they arrive at the earth .
a conventional assumption is that crs propagate in a cylindrical halo with a half height @xmath9 , beyond which crs escape freely .
the propagation equation can be expressed as @xcite @xmath10 \nonumber\\ & & - \frac{\partial}{\partial p}\left [ \dot{p}\psi - \frac{p}{3}(\nabla\cdot\mathbf{v}_c)\psi \right ] - \frac{\psi}{\tau_f } - \frac{\psi}{\tau_r } , \label{propagation_equation}\end{aligned}\ ] ] where @xmath11 is the cr density per momentum interval , @xmath12 is the source term , @xmath13 is the momentum loss rate , and the time scales @xmath14 and @xmath15 characterize fragmentation processes and radioactive decays , respectively . [ cols="^,^,^",options="header " , ] fig . [
fig : dm_sv ] also shows the favored regions at 68% and 95% c.l . in the @xmath16-@xmath17 plane .
the favored values of @xmath17 are highly correlated with the values of @xmath16 , since the dm source term is proportional to @xmath18 , as indicated in eq . .
since the dm annihilations into quarks or gauge bosons also produce high energy photons , @xmath19-ray observations , such as those from fermi - lat , would set strong constraints on the dm annihilation cross section .
also shown in fig .
[ fig : dm_sv ] are the upper limits from the observations of spheroidal dwarf galaxies with six - year fermi - lat gamma - ray data @xcite .
it can be found that the constraint from ams-02 is similar to that derived from fermi - lat dwarf galaxy observation .
in this work we investigate the new ams-02 results of @xmath0 ratio . in order to understand the data ,
the crucial step is to perform a reliable study on the secondary antiproton flux generated by cr interactions with the ism .
we consider three propagation models , namely the dr , dr-2 and dc models defined in tab .
[ tab : propmodel ] .
the prorogation parameters are constrained by the b / c data using a mcmc algorithm .
we find that all the models can provide very good fit to the b / c data . within @xmath20 ranges of the fitted propagation parameters ,
we give the predictions of the secondary antiprotons in the three propagation models and compared them with the ams-02 @xmath0 data .
it is found that the dr model underproduces antiprotons at low energies , which has already been noticed for a long time
. the dr-2 and dc models can give a quite good fit to the ams-02 data below @xmath21 gev .
however , it should be emphasized that the predicted secondary @xmath0 ratio above @xmath21 gev has a different slop from the ams-02 data .
the data keep flat up to @xmath1 gev , but the predictions of all the three models decrease with the energy .
the key factor is that the diffusion coefficient is in a form of single power law , namely @xmath22 , which is determined by the b / c data .
if the diffusion coefficient @xmath23 permits a flatter behavior at high energies within the allowance of the b / c data , a flatter behavior of @xmath0 could also be expected @xcite .
if we consider an extra source of antiprotons from the galactic dm annihilation , the fitting to the ams-02 data can be significantly improved .
the data can be well explained by a dm particle with a mass of @xmath24 and an annihilation cross section of @xmath25 into the @xmath26 or @xmath27 final states .
if we assume that there is no obvious excess in the @xmath0 ratio , the ams-02 data can be used to set upper limits on the dm annihilation cross section into @xmath26 or @xmath27 .
we find that the constraints given by the dr-2 model are slightly stricter than those given by the dc model . for small @xmath16 ,
the constraints from ams-02 are even stronger than the constraints from fermi - lat .
we appreciate q. yuan for helpful discussions .
this work is supported by the national natural science foundation of china under grants no .
11475189 , 11475191 , 11135009 , 11175251 , and the 973 program of china under grant no .
2013cb837000 and by the strategic priority research program `` the emergence of cosmological structures '' of the chinese academy of sciences , under grant no .
xdb09000000 .
o. adriani , g. c. barbarino , g. a. bazilevskaya , r. bellotti , m. boezio , e. a. bogomolov , l. bonechi and m. bongi _ et al .
_ , phys .
lett . * 102 * , 051101 ( 2009 ) [ arxiv:0810.4994 [ astro - ph ] ] .
o. adriani _ et al . _
[ pamela collaboration ] , phys .
* 105 * , 121101 ( 2010 ) [ arxiv:1007.0821 [ astro-ph.he ] ] .
f. donato , d. maurin , p. brun , t. delahaye and p. salati , phys .
lett . * 102 * , 071301 ( 2009 ) [ arxiv:0810.5292 [ astro - ph ] ] .
m. cirelli and g. giesen , jcap * 1304 * , 015 ( 2013 ) [ arxiv:1301.7079 [ hep - ph ] ] .
t. bringmann , m. vollmann and c. weniger , phys .
d * 90 * , no . 12 , 123001 ( 2014 ) [ arxiv:1406.6027 [ astro-ph.he ] ] .
m. cirelli , d. gaggero , g. giesen , m. taoso and a. urbano , jcap * 1412 * , no .
12 , 045 ( 2014 ) [ arxiv:1407.2173 [ hep - ph ] ] .
d. hooper , t. linden and p. mertsch , jcap * 1503 * , no .
03 , 021 ( 2015 ) [ arxiv:1410.1527 [ astro-ph.he ] ] . h. b. jin , y. l. wu and y. f. zhou , arxiv:1410.0171 [ hep - ph ] .
m. boudaud , m. cirelli , g. giesen and p. salati , arxiv:1412.5696 [ astro-ph.he ] .
a. w. strong and i. v. moskalenko , astrophys . j. * 509 * , 212 ( 1998 ) [ astro - ph/9807150 ] .
g. di bernardo , c. evoli , d. gaggero , d. grasso and l. maccione , astropart .
* 34 * , 274 ( 2010 ) [ arxiv:0909.4548 [ astro-ph.he ] ] .
d. maurin , f. donato , r. taillet and p. salati , astrophys .
j. * 555 * , 585 ( 2001 ) [ astro - ph/0101231 ] .
i. v. moskalenko and a. w. strong , astrophys .
j. * 493 * , 694 ( 1998 ) [ astro - ph/9710124 ] .
a. lewis and s. bridle , phys .
d * 66 * , 103511 ( 2002 ) [ astro - ph/0205436 ]
. j. liu , q. yuan , x. bi , h. li and x. zhang , phys .
d * 81 * , 023516 ( 2010 ) [ arxiv:0906.3858 [ astro-ph.co ] ] .
j. liu , q. yuan , x. j. bi , h. li and x. zhang , phys .
d * 85 * ( 2012 ) 043507 [ arxiv:1106.3882 [ astro-ph.co ] ] .
q. yuan , x. j. bi , g. m. chen , y. q. guo , s. j. lin and x. zhang , astropart .
phys . * 60 * , 1 ( 2015 ) [ arxiv:1304.1482 [ astro-ph.he ] ] .
s. j. lin , q. yuan and x. j. bi , phys .
d * 91 * , no .
6 , 063508 ( 2015 ) [ arxiv:1409.6248 [ astro-ph.he ] ] .
m. ibe , s. matsumoto , s. shirai and t. t. yanagida , arxiv:1504.05554 [ hep - ph ] . k. hamaguchi , t. moroi and k. nakayama , arxiv:1504.05937 [ hep - ph ] .
a. w. strong , i. v. moskalenko and v. s. ptuskin , ann .
sci . * 57 * , 285 ( 2007 ) [ astro - ph/0701517 ] .
d. maurin , a. putze and l. derome , astron .
astrophys . * 516 * , a67 ( 2010 ) [ arxiv:1001.0553 [ astro-ph.he ] ] .
r. trotta , g. johannesson , i. v. moskalenko , t. a. porter , r. r. de austri and a. w. strong , astrophys .
j. * 729 * , 106 ( 2011 ) [ arxiv:1011.0037 [ astro-ph.he ] ] .
ams-02 collaboration , international cosmicray conference , rio de janeiro , brazil , 2013 , http://www.ams02.org/2013/07/newresults- from - ams - presented - at - icrc-2013/[http://www.ams02.org/2013/07/newresults- from - ams - presented - at - icrc-2013/ ] .
r. a. mewaldt , m. miller , j. r. jokipii , m. a. lee , t. h. zurbuchen and e. mobius , `` acceleration and transport of energetic particles observed in the heliosphere .
proceedings , symposium , ace 2000 , indian wells , usa , january 5 - 8 , 2000 , '' vol .
421 - 424 .
n. e. yanasak , m. e. wiedenbeck and r. a. mewaldt _ et al .
_ , astrophys .
j. * 563 * , 768 ( 2001 ) .
t. hams , l. m. barbier , m. bremerich , e. r. christian , g. a. de nolfo , s. geier , h. gobel and s. k. gupta _ et al .
_ , astrophys .
j. * 611 * , 892 ( 2004 ) . j. j. connell , astrophys .
j. lett . , * 501 * , l59 ( 1998 ) .
h. s. ahn , p. allison , m. g. bagliesi , j. j. beatty , g. bigongiari , j. t. childers , n. b. conklin and s. coutu _ et al .
_ , astrophys .
j. * 714 * , l89 ( 2010 ) [ arxiv:1004.1123 [ astro-ph.he ] ] . s. orito _ et al . _
[ bess collaboration ] , phys .
lett . * 84 * , 1078 ( 2000 ) [ astro - ph/9906426 ] . i. v. moskalenko , a. w. strong , j. f. ormes and m. s. potgieter , astrophys . j. * 565 * , 280 ( 2002 ) [ astro - ph/0106567 ] .
m. cirelli , g. corcella , a. hektor , g. hutsi , m. kadastik , p. panci , m. raidal and f. sala _ et al .
_ , jcap * 1103 * , 051 ( 2011 ) [ jcap * 1210 * , e01 ( 2012 ) ] [ arxiv:1012.4515 [ hep - ph ] ] .
j. f. navarro , c. s. frenk and s. d. m. white , astrophys .
j. * 490 * , 493 ( 1997 ) [ astro - ph/9611107 ] . m. di mauro , f. donato , a. goudelis and p. d. serpico , phys . rev .
d * 90 * , no .
8 , 085017 ( 2014 ) [ arxiv:1408.0288 [ hep - ph ] ] .
r. kappl and m. w. winkler , jcap * 1409 * , 051 ( 2014 ) [ arxiv:1408.0299 [ hep - ph ] ] .
m. ackermann _ et al .
_ [ fermi - lat collaboration ] , arxiv:1503.02641 [ astro-ph.he ] .
a. e. vladimirov , g. e. johannesson , i. v. moskalenko and t. a. porter , astrophys .
j. * 752 * , 68 ( 2012 ) [ arxiv:1108.1023 [ astro-ph.he ] ] . | the ams-02 collaboration has just released the cosmic antiproton to proton ratio @xmath0 with a high precision up to @xmath1 gev . in this work ,
we calculate the secondary antiprotons generated by cosmic ray interactions with the interstellar medium taking into account the uncertainties from the cosmic ray propagation .
the @xmath0 ratio predicted by these processes shows some tension with the ams-02 data in some regions of propagation parameters , but the excess is not significant .
we then try to derive upper bounds on the dark matter annihilation cross section from the @xmath0 data or signal regions favored by the data .
it is shown that the constraint derived by the ams-02 data is similar to that from fermi - lat observations of dwarf galaxies .
the signal region for dark matter is usually required @xmath2 tev and @xmath3 . |
our starting point is the anisotropic version of the pseudo - qed theory , described by the lagrangian @xmath97 where @xmath98 is the usual field - intensity tensor of the u(1 ) gauge field @xmath99 , which intermediates the electromagnetic interaction in 2d ( pseudo electromagnetic field ) , @xmath100 , with @xmath101 a four - component dirac spinor representing electrons in sublattices @xmath102 and @xmath43 , with different spin orientations , @xmath103 are rank-4 dirac matrices , the parameter @xmath104 is the gauge fixing , and @xmath105 is an infrared regularization parameter .
we will adopt that @xmath106 . defining the feynman s rules of the model ( see fig .
[ feynmanrules ] ) , we obtain the fermion propagator @xmath107 , @xmath108 where for dirac matrices @xmath109 , @xmath110 and @xmath111 . the photon propagator reads @xmath112}},\ ] ] where @xmath113 is the four - momentum given by @xmath114 , @xmath115 and @xmath116 is the electric permitivity .
the interaction vertex is given by @xmath117 the pole of the fermion propagator provides the energy dispersion relation @xmath118 , where @xmath119 is the energy gap .
when @xmath120 , we reproduce the tight - binding result for monolayer graphene . the non - local term present in the maxwell lagrangian in eq .
( [ anisotropic ] ) renders the canonical dimension of the gauge field equal to one , in units of mass . the same holds for the dirac field .
therefore , the coupling constant @xmath11 is dimensionless in the 2 + 1 space - time , and the theory is renormalizable , analogously to qed4 , and the primitive divergences at one - loop order are shown in fig . [ divergences ] . we have calculated these diagrams using the dimensional regularization procedure as a way to obtain finite feynman amplitudes , which do not depend on the regulator @xcite . by adopting the feynman s gauge ( @xmath121 ) and denoting @xmath122 the momentum that runs through the loop , we first perform the integration over @xmath123 , then the one over @xmath124 using standard formulas of dimensional regularization and working in spatial dimension @xmath125 .
the vertex functions , which have been made finite by the subtraction of the pole terms in the dimensionally regularized amplitudes , satisfy a t hooft - weinberg renormalization group equation @xcite @xmath126 where we have considered the massless case . here
, @xmath127 denotes the renormalized vertex function of @xmath128 fermion fields and @xmath129 gauge fields , and @xmath130 denotes the external momenta .
we have introduced the renormalization scale @xmath131 through the substitution @xmath132 , where @xmath133 .
the @xmath134 s for @xmath135 are the beta functions which describe how the electric charge , fermi velocity and the speed of light , respectively , change with the scale parameter @xmath131 and are usually defined as @xmath136 , @xmath137 , and @xmath138 .
the functions @xmath139 and @xmath140 are the anomalous dimensions of the fields @xmath141 and @xmath142 , which are given by @xmath143 and @xmath144 , where @xmath145 and @xmath146 are the wave function renormalization of the fields @xmath141 and @xmath142 , respectively . by performing the calculations up to one - loop and and using the minimal subtraction scheme , we find @xmath147 , \nonumber \\
\gamma^{\mu(2,1 ) } & = & \gamma^{\mu}_{0 } + \gamma^{\mu } , \nonumber \end{aligned}\ ] ] where @xmath148 , which leads to @xmath149\gamma^{0 } , \nonumber \\
\gamma^{i(2,1 ) } & = & -ie \beta\gamma^i+e^3\left[\mbox{finite}^{i(2,1)}+\frac{3}{2}\ln\mu\,\mbox{res}^{i(2,1)}\right]\gamma^{i } , \nonumber\end{aligned}\ ] ] where @xmath150 . in the above expression
, @xmath151 means the residue part of the diagram that is given by the coefficient of the term @xmath152 , and @xmath153 is the finite part of the amplitudes .
thus , we can write @xmath154 where @xmath155 , @xmath156 , @xmath157 and @xmath158 are constants . by using dimensional regularization
, we can anticipate that @xmath159 is finite at one - loop ( order @xmath160 ) .
hence , the anomalous dimension of the gauge field is zero in this order , i.e. , @xmath161 . after writing the rg functions perturbatively , e.g. , @xmath162
, we find that @xmath163 hence , there is neither renormalization of the speed of light , nor of the charge .
this should be contrasted to the results found by isobe and nagaosa in anisotropic qed@xmath164 , where the charge and the speed of light are renormalized @xcite . according to the feynman s rules , the electron self - energy at one - loop order represented in fig .
[ eselfenergy ] is given by @xmath165 where @xmath166 , @xmath167 , @xmath168 , @xmath169 , @xmath170 and @xmath171 .
we have used that @xmath172 and @xmath173 .
if @xmath174 , then @xmath175 , and lorentz symmetry is recovered .
after some algebra , we obtain that the divergent part of the electron self - energy is @xmath176 where the parametric integrals @xmath177 , @xmath178 and @xmath179 are @xmath180 } = - \frac{2}{\eta(\beta)}+\frac{2\beta}{\eta(\beta)^{3/2}}\tan^{-1}\left(\frac{\sqrt{\eta(\beta)}}{\beta}\right ) , \nonumber \\ c^2i_{2 } & = & \int_{0}^{1 } dx \frac{(1-x)^{-\frac{1}{2}}}{[x\,\eta(\beta)-1 ] } = -\frac{2}{\beta\sqrt{\eta(\beta)}}\tan^{-1}\left(\frac{\sqrt{\eta(\beta)}}{\beta}\right ) \ , \nonumber \\
c^2i_{3 } & = & \int_{0}^{1 } dx \frac{(1-x)^{-\frac{1}{2}}{\ensuremath{\left(1+\frac{\beta^{2}x}{x\,\eta(\beta)-1}\right)}}}{[x\,\eta(\beta)-1 ] } = \frac{1}{\eta(\beta)}-\frac{\tan^{-1}\left(\sqrt{\eta(\beta)}/\beta\right)}{\beta \,\eta(\beta)^{3/2}}. \nonumber\\end{aligned}\ ] ] finally , we may identify @xmath181 and obtain @xmath182}}e^2 , \label{betavf}\\ \gamma_{\psi } & = & \frac{c}{16\varepsilon\pi^2}{\ensuremath{\left(1 - 2\beta^2\right)}}i_1 e^2 . \label{gamapsi}\end{aligned}\ ] ] using the parametric expression for @xmath183 and @xmath184 in eq .
( [ betavf ] ) , we have @xmath185\sqrt{1-x}}{[x \eta(\beta)-1]^{2}}. \label{beta1}\ ] ] after evaluating the parametric integral in eq .
( [ beta1 ] ) , we obtain @xmath186 where @xmath187\cot^{-1}[\beta/\sqrt{\eta(\beta)}]}{\eta(\beta)^{3/2}}. \nonumber\ ] ] for @xmath174 , i.e. @xmath188 , it follows that @xmath189 , thus implying that @xmath190 .
this means that the speed of light is the infrared fixed point . for simplicity , we artificially consider @xmath191and @xmath192 to plot eq .
( [ beta2 ] ) in fig .
[ supfig5 ] .
[ betafig ] * * -beta function * * @xmath193*. the fixed point @xmath194 ( @xmath174 ) is a stable infrared fixed point .
[ supfig5],title="fig:",scaledwidth=45.0% ] eq .
( [ beta2 ] ) is a complicated function of @xmath195 , but one may explore its non - relativistic and relativistic limits , i.e , @xmath196 and @xmath197 , respectively . for @xmath197 ,
one finds @xmath198,\ ] ] with @xmath199 and @xmath200 , whereas for @xmath201 @xmath202.\ ] ] we have determined the effective ( or running ) fermi velocity @xmath203 for pseudo - qed in both limits . from eq .
( [ beta2 ] ) , the flow of the effective fermi velocity can be written as @xcite @xmath204 where @xmath205 , and we have introduced a logarithmic scale @xmath206 , with @xmath207 a reference scale where the parameter @xmath208 has been defined . for @xmath209 , the natural limit in the lowest order is @xmath210 .
therefore , by integrating eq .
( [ defvrenor ] ) , we find @xmath211 , \label{vren1}\ ] ] where we defined the coupling constant for graphene @xmath212 , using that @xmath213 , where @xmath20 is the vacuum permittivity and @xmath214 is the dielectric function that describes screening in graphene @xcite .
furthermore , we recovered the physical units where @xmath215 in order to compare with experimental data . this result agrees with the value derived earlier in ref .
@xcite . in the vicinity of the fixed point , @xmath216
, we can use that @xmath217 thus , by integrating eq .
( [ defvrenor ] ) , we obtain @xmath218\left(\frac{\mu}{\mu_0}\right)^{2\gamma}\right\}}},\label{vren2}\ ] ] where @xmath219 $ ] , with @xmath220 , the usual fine - structure constant .
we may rewrite eqs .
( [ vren1 ] ) and ( [ vren2 ] ) in terms of the electron density @xmath9 to obtain @xmath221 } } , \label{vf1}\end{aligned}\ ] ] for @xmath13 , and @xmath222}}{\ensuremath{\left(\frac{n}{n_0}\right)}}^{\gamma}\right\ } } } , \label{vf2}\end{aligned}\ ] ] for @xmath223 .
( [ vf1 ] ) and ( [ vf2 ] ) are the most important results of this section .
however , notice that the experimental data on the renormalized @xmath5 may only be compared to eq .
( [ vf1 ] ) .
although we have shown that there are two possible asymptotic regions ( @xmath13 and @xmath223 ) for a generic 2d dirac system , when comparing to graphene one would need an `` infinite '' sample to observe the relativistic effects .
indeed , by integrating eq .
( [ defvrenor ] ) for @xmath195 , and using fixed values of the reference point @xmath224 and @xmath195 , we obtain @xmath225 for a value of @xmath226 @xcite , as we increase @xmath195 to approach to 1 , the ratio between the energies @xmath227 gets smaller , even with little changes in the velocity , e.g. , @xmath228 .
hence , to reach values of @xmath229 or even @xmath230 , where the relativistic effects could be appreciable , we would be considering resolutions smaller than @xmath231mev .
this result tell us that despite the fact that we have found theoretically a relativistic limit , this regime can not be reached in graphene .
our aim in this section is to compute the electron s gyromagnetic factor @xmath0 using the anisotropic pseudo - qed . for this , it is only necessary to calculate the finite part of the spatial component of the vertex represented in fig . [ vertexgraph ] . according to feynman s rules ,
the vertex diagram is given by @xmath232 with @xmath233 .
using the anticommutation of the gamma matrices and the dirac equations in momentum space @xmath234 and @xmath235 , we can write @xmath236u(\bar{p})}{{\ensuremath{\left(k_{0}^{2 } - \delta\right)}}^{5/2}}. \end{aligned}\ ] ] in the equation above , we used the parametric integral @xmath237}}{\ensuremath{\left[(\bar{k}+\bar{p})^{2}-m^{2}\right ] } } \sqrt{k^{2}}\right\}}}^{-1}=\frac{3}{4}\int_{0}^{1}dx\int_{0}^{1-x}dy \frac{(1-x - y)^{-1/2}}{{\ensuremath{\left[(k_{0}+\omega_{0})^{2 } - \delta\right]}}^{5/2 } } , \nonumber\end{aligned}\ ] ] where @xmath238 with @xmath239}}$ ] , @xmath240 and @xmath241 .
we performed also the displacement @xmath242 , such that the terms in the numerator of eq .
( [ esp ] ) become @xmath243\gamma^{i } + 2\gamma^{i}\gamma_{0}p'^{0}v_{f}\boldsymbol{\gamma}\cdot\textbf{k}+ 2p_{0}v_{f}\gamma^{0}\gamma^{l}k_{l}\gamma^{i } - 2v_{f}k^{i } \omega_{0}\gamma^{0 } + \nonumber \\ & - & 2v_{f}^{2}k^{i}k_{l}\gamma^{l } + ( 1/2)v_{f}^{2}\textbf{k}^{2}\gamma^{i } \nonumber\end{aligned}\ ] ] and @xmath244 where we eliminated the odd terms in @xmath245 . as a next step ,
we try to simplify the lengthy expressions .
since we are interested in obtaining the gyromagnetic factor , we will disregard the terms proportional to @xmath246 . after solving the integrals over @xmath123
, we find @xmath247u(\bar{p})}{a^2{\ensuremath{\left[(\textbf{k}^{2}-v_f^2 a^{-1}\boldsymbol{\omega})^2 - \tilde{\delta}\right]}}^{5/2}},\ ] ] where @xmath248 and @xmath249 displacing @xmath250 , we find , after solving the integrals over @xmath124 , @xmath251 where we considered @xmath252 . therefore ,
working on mass - shell , we can use that @xmath253 . changing the variables @xmath254 and @xmath255 to @xmath256 and @xmath257 , in such a way that @xmath258 and @xmath259
, we can write @xmath260 and @xmath261 now , we can use the gordon identity @xmath262 and write @xmath263 where @xmath264 and @xmath265 are now the normalized solutions of the free dirac equation @xcite .
hence , the form factor @xmath266 is identified as @xmath267 } } } { av_{f}^{2}(\textbf{p'}x+\textbf{p}y)^{2}}\right\}}}. \label{f}\ ] ] by rewriting the denominator of eq .
( [ f ] ) as @xmath268 and using that @xmath269 , we obtain @xmath270 with @xmath271 } - \int_{0}^{1}dx\int_{0}^{1-x}dy \frac{(1-x - y)^{-1/2}}{[\beta^{2}(x+y)+(1-x - y)]^{2}}. \label{parametricg}\ ] ] in the limit of @xmath272 , we find @xmath273 which is exactly what is obtained in the isotropic model . on the other hand , if we solve the integrals in eq .
( [ parametricg ] ) , we find @xmath274 } } } { \beta^{3}{\ensuremath{\left(-1+\beta^{2}\right)}}^{3/2}}.\end{aligned}\ ] ] then , for @xmath201 we have @xmath275 and the correction for the @xmath0-factor reads @xmath276 whereas for @xmath197 ( isotropic limit ) the correction is given by @xmath277
in the case of the valley splitting the procedure is slightly different , since we are not able to generate this gap using perturbation theory .
the splitting found experimentally @xcite is much larger than the one for the spin , and clearly can not be conceived as a correction to the bare spin .
therefore , a non - perturbative method is required . in a recent work @xcite , using schwinger - dyson s equations , some of us have shown that a mass gap is generated in graphene solely due to interactions , even in the absence of a magnetic field , @xmath278 where @xmath63 , and @xmath64 is the critical coupling @xcite . however , the experiment in ref .
@xcite is performed in the presence of an external magnetic field , and the valley splitting exhibits a linear dependence on the field for @xmath279 tesla .
hence , we now consider the fermionic propagator in the presence of an external magnetic field , accounting for the mass that was dynamically generated due to interactions , and perform a semi - classical calculation . in the presence of an external magnetic field @xmath280 , where @xmath281 is the corresponding vector potential , the fermionic lagrangian is promptly obtained by using the minimal coupling @xmath282}}$ ] , @xmath283}}\bar{\psi}.\ ] ] within a semi - classical approximation , this external vector potential is given by its mean value , i.e. , @xmath284 , and we can rewrite the fermion propagator as @xmath285 where we considered @xmath286 , but @xmath287 . in the case of @xmath288
, we obtain the energy from the poles of the fermionic propagator @xmath289 hence , in the presence of weak magnetic fields , we may write the valley splitting @xmath77 where the only free adjustable parameter is @xmath78 , which for typical magnetic fields is of order of 1 nm .
measurements of the spin gyromagnetic factor were performed by kurganova et al . for
graphene grown on the sio@xmath56@xcite . in order to compare our result for the @xmath0-factor given by eqs .
( [ deltag1 ] ) and ( [ deltag2 ] ) with these experiments , we begin with the equation @xmath290 which relates the cyclotron mass @xmath29 to the charge carrier concentration @xmath9 , and @xmath26 is the renormalized fermi velocity . the parameter @xmath291 in eq .
( [ function1 ] ) is the effective @xmath0-factor , which , in the experimental work , is taken to be a constant and fixed parameter that best fits the experimental points @xcite , and @xmath28 denotes the electron mass . recalling that the bare @xmath0-factor in graphene is @xmath292 , after replacing eqs .
( [ deltag1 ] ) and ( [ vren1 ] ) into eq .
( [ function1 ] ) , we obtain the corrected @xmath0-factor @xmath293 , @xmath294 } } } , \nonumber \\ & = & { \ensuremath{\left\{2+\frac{e^{2}}{16\pi\hbar\varepsilon_{0}\varepsilon_g v_{f}^{0}{\ensuremath{\left[1+\frac{\alpha_0}{8\varepsilon_g}\ln{\ensuremath{\left(\frac{n_0}{n}\right)}}\right ] } } } \right\ } } } \frac{\hbar\sqrt{\pi n } } { m_e v_f^0 } \frac{1 } { { \ensuremath{\left[1+\frac{\alpha_0}{8\varepsilon_g}\ln{\ensuremath{\left(\frac{n_0}{n}\right)}}\right ] } } } , \nonumber \\ & = & { \ensuremath{\left\{\frac{2 } { { \ensuremath{\left[1+\frac{\alpha_0}{8\varepsilon_g}\ln{\ensuremath{\left(\frac{n_0}{n}\right)}}\right]}}}+\frac{e^{2}}{16\pi\hbar\varepsilon_{0 } \varepsilon_g v_{f}^{0}{\ensuremath{\left[1+\frac{\alpha_0}{8\varepsilon_g}\ln{\ensuremath{\left(\frac{n_0}{n}\right)}}\right]}}^{2 } } \right\ } } } \frac{\hbar\sqrt{\pi n } } { m_e v_f^0 } , \nonumber \end{aligned}\ ] ] where we used @xmath295 , with @xmath22 the renormalized fermi velocity given by eq .
( [ vren1 ] ) , and @xmath296 .
note that screening is taken into account in @xmath46 and in @xmath22 .
choosing the reference value of @xmath37 around the values of @xmath9 that we want to describe , and neglecting corrections of order @xmath297 , we may write @xmath298}}^{2 } & \approx & 1 + \frac{2\alpha_0}{8\varepsilon_g}\ln{\ensuremath{\left(\frac{n_0}{n}\right ) } } \nonumber\end{aligned}\ ] ] to obtain @xmath299}}}+\frac{\alpha_0}{4 \varepsilon_g { \ensuremath{\left[1 + \frac{2\alpha_0}{8\varepsilon_g}\ln{\ensuremath{\left(\frac{n_0}{n}\right)}}\right ] } } } \right\ } } } \frac{\hbar\sqrt{\pi n } } { m_e v_f^0 } .
\label{y3n } \end{aligned}\ ] ] this equation is plotted in figs .
[ fig1 ] * a * and * b * of the main text , for different values of @xmath25 and @xmath300 . measurements of the spin - splitting and , moreover , valley - splitting were performed also in graphene on sic @xcite , and we now compare these data to our results obtained within the pseudo - qed . in this experiment , the zeeman splitting , was measured , which is given by @xmath301 inserting the value found for @xmath302 in sec .
iii , @xmath303 with @xmath58 given by eqs .
( [ deltag1 ] ) and ( [ vren1 ] ) , we have @xcite @xmath304}}}\right\ } } } \mu_{b } b. \label{spinsp}\end{aligned}\ ] ] we can observe in ref .
@xcite that the experimental data , i.e. , the spin - splitting , does not change much when increasing the magnetic field from 11 to 14 t. eq .
( [ spinsp ] ) for the spin - splitting is plotted in fig .
[ gv ] * b * in the main text using @xmath25 as a fitting parameter .
we find that @xmath80 for this sample .
once we have derived the valley splitting in the presence of a weak magnetic field in eq .
( [ valleysp ] ) , we can associate this energy with the zeeman energy for the separation of the valleys , and isolate the g@xmath305-factor , @xmath306 now , replacing eq .
( [ valleysp ] ) into ( [ zeemanvalley ] ) , we find @xmath307 it is important to notice the domain of validity of eq .
( [ gv ] ) .
it only holds for b from 2 to 8 t , approximately , which is the region in the experiment @xcite where the data exhibit a linear behavior .
therefore , a possibly inferred behavior for b around zero , that would be divergent , is wrong . in this case
, one should return to eq .
( [ valleysp ] ) , which shows that the zero magnetic field mass is finite , @xmath308 .
( [ valleysp ] ) is plotted in the inset of fig .
[ gv ] * b * of the main text using the value of the gap @xmath81 found in ref .
@xcite ( for a @xmath309 ev @xcite ) , and @xmath310 in this case is simply the bare value @xmath311 m / s . here
, we also plot eq .
( [ gv ] ) in fig .
[ gvalley ] . for both plots we used the value of @xmath312 obtained from the fit of the measurement of @xmath47 and @xmath74 .
an excellent agreement is found between theory and experiment .
an experiment performed by the group of e. andrei for graphene deposited on graphite @xcite showed that for very clean samples , there is a gap being generated , even in the absence of a magnetic field , that could possibly be explained by the effect of interactions .
now , we attempt to compare our theoretical predictions to this experimental result . using the expression for the gap predicted within pqed @xcite @xmath313 where @xmath314 and substituting the values of @xmath315 @xcite , @xmath316 @xcite , @xmath89 @xcite , @xmath317 and @xmath318
, we obtain a gap @xmath319 , in excellent agreement with the measured value .
gross , d. j. applications of the renormalization group to high - energy physics , proceedings of les houches , session xxviii , 1975 , edited by roger balian and jean zinn - justin , north - holland publishing co , 1976 .
elias , d. c. , gorbachev , r. v. , mayorov , a. s. , morozov , s. v. , zhukov , a. a. , blake , p. , ponomarenko , l. a. , grigorieva , i. v. , novoselov , k. s. , guinea , f. & geim , a. k. _ nat .
_ * 7 * , 701 - 704 ( 2011 ) . | the spin - gyromagnetic factor @xmath0 is an important physical quantity relating the magnetic dipole moment of a particle to its spin .
the electron @xmath0-factor _ in vacuo _ is one of the best model - based theoretical predictions ever made , showing agreement with the measured value up to ten parts per trillion @xcite .
however , for electrons in a material the @xmath0-factor is modified with respect to its value _ in vacuo _ because of environment interactions . here ,
we report a detailed calculation of the @xmath0-factor in graphene , taking into account the full electromagnetic interaction described by pseudo - qed @xcite .
because the electrons in graphene have spin and valley internal degrees of freedom , one may associate a @xmath1-factor to both flavors .
experiments indicate that the spin @xmath1-factor @xmath2 , depending on the graphene substrate @xcite , whereas the valley @xmath1-factor is colossal , @xmath3 .
we determine the radiative corrections to the spin @xmath1-factor both in the high- and low - doping regimes , and find a very good agreement between theory and experiments @xcite .
in addition , we show that the gigantic value measured recently for the valley @xmath1-factor @xcite can also be well explained within the pseudo - qed formalism .
the energy dispersion of the electrons in solid - state materials strongly depends on the lattice geometry . in the case of graphene
, the honeycomb lattice leads to a zero - mass relativistic - like dispersion @xmath4 .
this characteristic allows us to relate the electrons in graphene to free dirac massless particles in ( 2 + 1 ) dimensions ( d ) @xcite .
however , the fact that these electrons propagate with a velocity @xmath5 three hundred times smaller than the speed of light @xmath6 @xcite has important implications upon the physical properties of the system . for long
, graphene was believed to be an effectively noninteracting system .
the recent measurement of the fractional quantum hall effect @xcite , which is a typical feature of strongly correlated systems , has changed this paradigm .
the importance of interactions in graphene was further attested by the experimental observation of the renormalization of the fermi velocity @xcite , as had been predicted by theories that accounted for _ static _ interactions @xcite .
however , dynamical effects have proven to be relevant in some cases , by generating novel quantum topological states that would not emerge in the static approximation @xcite .
although the electrons in graphene are confined to move on a plane , the electromagnetic field through which they interact extends throughout the 3d space .
integrating away out - of - the - plane photons , one obtains an effective interaction that is non - local in space and time . in spite of being fully 2d
, it conveys all properties of the genuine 3d electromagnetic interaction .
this effective theory has been called pseudo - qed because it contains pseudo - differential operators .
it has been shown to respect causality @xcite , scale invariance , the huygens principle , and unitarity @xcite , apart from exhibiting an @xmath7 static coulomb potential .
furthermore , the propagator in pseudo - qed in coordinate space coincides with the one of qed3 in momentum space @xcite , and these two theories are dual to each other @xcite .
inspired by the relevance of electron - electron interactions in graphene , we investigate in this paper the _ anisotropic _ pseudo - qed by considering a term that breaks lorentz invariance in our quantum - field - theory formalism .
we can easily include the lorentz violating term in the matter field , since the nonlocal gauge field produces the full electromagnetic interaction , independently on whether the matter is relativistic or not @xcite .
we then determine the spin gyromagnetic factor ( @xmath8-factor ) in two regimes , at small and large carrier density @xmath9 , and show that it compares very well to the experimental data available in the literature @xcite .
moreover , we compare our predictions for the valley gyromagnetic factor ( @xmath10-factor ) with experimental measurements obtained for epitaxial graphene grown on sic @xcite and graphene flakes on graphite @xcite , finding in both cases an excellent agreement .
our results exploit the relevance of interactions in estimating the g - factor in graphene in particular , and 2d relativistic condensed - matter systems in general , besides establishing a solid theoretical ground that allows one to understand puzzling and disparate experimental results observed so far .
our starting point is the anisotropic version of the pseudo - qed theory ( see sup .
info for details ) . within this effective theory both the gauge and
the dirac fields have the canonical dimension equal to one , which leads to a dimensionless coupling constant @xmath11 .
hence , pseudo - qed is renormalizable in the same way as qed4 .
we calculate the relevant feynman diagrams using the dimensional regularization procedure @xcite , and thereby determine the equivalent @xmath12-function .
afterwards , we solve the associated renormalization group equations in two different limits , namely @xmath13 and @xmath14 , to obtain the renormalized fermi velocity .
we proceed then to the evaluation of the correction to the bare spin @xmath1-factor , by calculating the one - loop radiative corrections to the electron - photon vertex diagram @xcite ( see sup .
info for details ) .
a thorough perturbative calculation up to three - loops was recently performed in ref .
@xcite , yet using the static coulomb interaction regime . an essential ingredient of our calculation is the insertion of the _ full _ _ dynamical _ interaction , which is described by pseudo - qed . in the limit @xmath13
, we obtain the enhanced spin @xmath1-factor @xmath15 @xmath16 where @xmath17 is the one - loop correction to the bare value @xmath18 , @xmath19 , @xmath20 , @xmath21 is the dielectric constant , which depends heavily on the substrate and on the doping @xmath9 , and @xmath22 is the renormalized fermi velocity . however , in the isotropic case , where @xmath23 , we find @xmath24 -factor corrected by interactions : influence of the dielectric constant @xmath25*. at high densities , the theoretical curves are given by eq .
( [ deltag1 ] ) , by using the renormalized value of @xmath26 given by eq .
( [ log ] ) , and the reference value @xmath27 m / s ( see sup . info . for details ) . here , @xmath28 denotes the electron mass and @xmath29 is the cyclotron mass .
* a * the black and pink solid curves correspond to different values of the dielectric constant , chosen _
ad hoc _ to be @xmath30 and 5 , respectively .
the light - blue solid curve denotes the bare @xmath31 factor @xcite . * b * the red solid line is the theoretical curve obtained for @xmath32 ( i.e. @xmath33 ) , which is the bare fine structure constant for graphene on sio@xmath34 @xcite .
[ fig1],title="fig : " ] -factor corrected by interactions : influence of the dielectric constant @xmath25*. at high densities , the theoretical curves are given by eq .
( [ deltag1 ] ) , by using the renormalized value of @xmath26 given by eq .
( [ log ] ) , and the reference value @xmath27 m / s ( see sup .
info . for details ) .
here , @xmath28 denotes the electron mass and @xmath29 is the cyclotron mass . *
a * the black and pink solid curves correspond to different values of the dielectric constant , chosen _
ad hoc _ to be @xmath30 and 5 , respectively .
the light - blue solid curve denotes the bare @xmath31 factor @xcite . *
b * the red solid line is the theoretical curve obtained for @xmath32 ( i.e. @xmath33 ) , which is the bare fine structure constant for graphene on sio@xmath34 @xcite .
[ fig1],title="fig : " ] these are the first major outcomes of this paper .
now , since the fermi velocity varies strongly as a function of doping , special care has to be taken when comparing the theoretical predictions with the experiments , i.e. , this functional dependence must be considered .
the fermi velocity renormalization due to interactions has been investigated earlier using the renormalization group technique , and it was found that there is one fixed point in the problem : @xmath23 @xcite . however , the system will display a different behavior when varying the doping values @xcite . at high enough doping
, the fermi velocity exhibits a logarithmic behavior @xcite , @xmath35 , \label{log}\ ] ] where @xmath36 is the bare value of the graphene fine structure constant and @xmath37 is a value of reference for doping . on the other hand , near the neutrality point @xmath38 the system flows towards @xmath23 and should become fully relativistic @xcite , where a power - law behavior is found instead , @xmath39\left(\frac{n}{n_0}\right)^{\gamma}\right\}}},\ ] ] ( see sup .
info for the full details of this calculation ) . here , @xmath40 , with @xmath41 denoting the bare fermi velocity ( reference point ) and @xmath42 .
although undoubtedly important for the complete description of the theory , the limit @xmath23 can not be reached in graphene samples .
it remains to be seen whether it could be realized or at least approached in artificial graphene samples @xcite , or other engineered systems @xcite .
the gyromagnetic factor is an intrinsic property of the free electron in a certain medium , which is seen to be corrected due to electron - electron interactions .
experimentally , the @xmath1-factor is estimated by applying a magnetic field @xmath43 perpendicularly to a sample and measuring the zeeman gap @xmath44 , where @xmath45 is the boht magneton . by inserting the renormalized @xmath22 given by eq .
( [ log ] ) inside the definition of the fine structure constant @xmath46 and replacing it into eq .
( [ deltag1 ] ) , we obtain the value of the @xmath8-factor corrected by interactions , i.e. , @xmath47 , in the physical regime @xmath13 , see fig .
[ fig1 ] * a * and ref .
@xcite . here , we plot eq .
( [ deltag1 ] ) using eq .
( [ log ] ) for _ ad hoc _ values of the dielectric constant @xmath48 and 5 ( black and pink curves , respectively ) . as shown in ref .
@xcite , the dielectric constant depends on the density concentration , and for the doping values used in ref .
@xcite , we empirically expect @xmath25 to be in this range @xcite .
the light - blue curve , corresponding to the bare value of the @xmath1-factor @xmath49 clearly can not describe the observed data , thus confirming the relevance of interactions in the description of the spin @xmath1-factor . in fig .
[ fig1 ] * b * , we still plot eq .
( [ deltag1 ] ) using eq .
( [ log ] ) , however , here we use the value of @xmath32 , as given in ref .
@xcite for graphene on sio@xmath34 .
the theoretical curve exhibits an excellent agreement with the experimental data . until now , we discussed the enhancement of the bare spin @xmath1-factor in graphene due to interaction effects .
however , besides the spin , the dirac electrons in graphene have also a valley degree of freedom .
measurements of the integer quantum hall effect indicate the four - fold degeneracy of the landau levels , due to both spin ( up and down ) and valley ( @xmath50 and @xmath51 ) . at ultra - low temperatures , in extremely clean samples ,
these degeneracies may be lifted .
this was indeed the case in ref .
@xcite , where the top layer of multilayer epitaxial graphene grown on sic was investigated by high - resolution scanning tunneling spectroscopy .
the experiments were performed at temperatures as low as 10 mk , in magnetic fields up to 15 t. interactions are expected to dominate in this epitaxial graphene sample , where quantum hall effect observations suggest that the sample is in the weak - disorder limit ( the width of the landau levels is about 0.5 mev ) .
an astonishing large valley @xmath1-factor was experimentally observed , @xmath52 , whereas the spin @xmath1-factor behaved as expected , with interactions providing a small correction @xmath53 to the bare spin @xmath1-value @xmath49 .
indeed , the values @xmath54 and @xmath55 ( there is a small difference in the value measured for each of the valleys ) are comparable to the one obtained by kurganova et al .
@xcite for graphene grown on sio@xmath56 , @xmath57 .
these experimental observations had to await for a while , before being understood .
the valley @xmath1-factor is not an intrinsic characteristic of the free dirac electron , as is the spin @xmath1-factor , but it is associated to the existence of a lattice geometry , that in the case of graphene , is the honeycomb lattice .
therefore , it can not be calculated by first principles as a correction to some bare value , as we did before for the spin . however , we show now that this baffling result can be understood within the same pqed formalism used to evaluate @xmath58 . recently , some of us studied transport properties in undoped graphene within pqed by using the kubo formula @xcite .
we found that above a critical value of the interaction strength the parity anomaly sets in , and a _ universal _ , _ exact _ , and _ quantized _ valley hall effect arises @xcite . due to interactions , time - reversal symmetry is spontaneously broken for each valley , although the system remains time - reversal invariant and obeys the vafa - witten theorem when both valleys are considered together . as an outcome , the electrons carrying the valley @xmath50 and @xmath51 pseudospin degree of freedom counterpropagate at the edges of an otherwise insulating sample .
the magnitude of the conductance quantization for each valley is precisely the same as obtained in the presence of a perpendicular magnetic field , but now it is dynamically driven by interactions in the _ absence _ of a magnetic field .
this result was further confirmed by the calculation of the energy levels as a non - perturbative solution of the schwinger - dyson equation , which leads to @xmath59 , where @xmath60 is an ultraviolet cutoff ( given by the inverse lattice constant in the case of graphene ) , @xmath61 /2 $ ] , with @xmath62 integer and @xmath63 , where @xmath64 is the critical coupling for the occurrence of the anomaly @xcite .
since undoped graphene is genuinely quantum critical , due to the zero density of states , a critical interaction strength is required for the anomaly to set in .
the existence of a gap generated by interactions has been also predicted earlier using non - perturbative techniques within qed in ( 2 + 1)d @xcite .
-factor in graphene grown on sic*. * a * sketch of the energy levels , indicating the lifting of the valley degeneracy calculated within the pqed model , in the absence of an applied magnetic field .
this energy splitting @xmath65 is identified with the valley - zeeman splitting @xmath66 .
the usual spin - zeeman splitting @xmath67 is also shown .
* b * comparison between theory and experiments for both @xmath1-factors .
the main figure refers to the spin @xmath1-factor . in the experiments
, there is an asymmetry between the valleys , indicated by the red and blue points .
they lead to a spin @xmath1-factor of @xmath68 and @xmath69 , respectively @xcite . the black - solid line , which provides a good agreement with the experimental data , is obtained by using eq .
( [ deltag1 ] ) and the fitting parameter @xmath70 , since the precise value of the dielectric constant is unknown .
the reference value for the magnetic field in the rg equations for the renormalized fermi velocity used here is @xmath71 t. inset : valley @xmath1-factor .
the green dots are the experimental results obtained in ref .
@xcite , whereas the black - solid line is our theoretical result , obtained from eq .
( [ valleysp ] ) by using the value @xmath72 determined from @xmath73 and @xmath74 . [ gv ] , title="fig : " ] -factor in graphene grown on sic*. * a * sketch of the energy levels , indicating the lifting of the valley degeneracy calculated within the pqed model , in the absence of an applied magnetic field .
this energy splitting @xmath65 is identified with the valley - zeeman splitting @xmath66 .
the usual spin - zeeman splitting @xmath67 is also shown .
* b * comparison between theory and experiments for both @xmath1-factors .
the main figure refers to the spin @xmath1-factor . in the experiments
, there is an asymmetry between the valleys , indicated by the red and blue points .
they lead to a spin @xmath1-factor of @xmath68 and @xmath69 , respectively @xcite . the black - solid line , which provides a good agreement with the experimental data , is obtained by using eq .
( [ deltag1 ] ) and the fitting parameter @xmath70 , since the precise value of the dielectric constant is unknown .
the reference value for the magnetic field in the rg equations for the renormalized fermi velocity used here is @xmath71 t. inset : valley @xmath1-factor .
the green dots are the experimental results obtained in ref .
@xcite , whereas the black - solid line is our theoretical result , obtained from eq .
( [ valleysp ] ) by using the value @xmath72 determined from @xmath73 and @xmath74 .
[ gv ] , title="fig : " ] the dynamically generated masses are the zero - momentum poles of the corrected electron propagator , @xmath75 for @xmath76 . in the presence of a magnetic field
@xmath43 , this mass is corrected , and by using a semi - classical approximation we obtain @xmath77 where @xmath78 is a adjustable parameter of order of nanometer ( see sup .
info . for details ) . here , we propose that the lifting of the valley degeneracy observed in ref .
@xcite in ultraclean graphene probed in the millikelvin regime is actually an observation of this dynamically generated mass @xmath79 . in fig .
[ gv ] * a * , we sketch the energy levels , as well as the corresponding gaps , and in fig .
[ gv ] * b * we plot our theoretical predictions , represented by the solid black lines , jointly with the experimental data @xcite .
first , we use eqs .
( [ deltag1 ] ) and ( [ log ] ) in the main figure and consider @xmath25 as a fitting parameter .
the theoretical black - solid curve corresponds to @xmath80 .
now , by using @xmath81 @xcite ( for @xmath82 ) , @xmath83 , the bare value of @xmath5 , and @xmath78 of order of nanometer in eq .
( [ valleysp ] ) , we obtain the black - solid line in the inset of fig .
[ gv ] * b*. a very good agreement between experiment and theory is obtained . moreover , eq .
( [ valleysp ] ) indicates that this gap will survive even for @xmath84 , when it becomes identical to the value predicted in ref .
@xcite .
next , we argue that this zero-@xmath43 result has possibly been observed in ref .
@xcite .
high - resolution ( @xmath85 ) landau - level ( ll ) spectroscopy performed at 4 kelvin on single - layer graphene deposited on a graphite substrate has revealed a gap of approximately @xmath86 in the lowest ll , which remained even in the absence of an applied magnetic field @xcite .
several mechanisms were conjectured to be behind the observed behavior @xcite . here , we claim that this gap could very well be due to valley splitting , which occurs due to an spontaneous symmetry breaking triggered by interactions . using that @xmath87 and @xmath88 @xcite , the ultraviolet cutoff as @xmath89 @xcite , @xmath90 and @xmath91 , together with a reduced value of the fermi velocity @xmath92 @xcite , we obtain a gap of @xmath93 , in excellent agreement with the experiment . as argued in ref .
@xcite , the fact that the lls in the experiment were better fitted by a lorentzian than by a gaussian curve indicates that the sample is very clean and dominated by many - body effects , rather than disorder .
indeed , stm topography shows no signature of structural disorder in the region where the tunneling spectra data was collected .
the agreement between the theoretical predictions and this third type of independent experiment corroborates further the validity and relevance of the theoretical model .
the implications of our findings are broad .
the possibility to solve the old graphene @xmath1-factor conundrum in an elegant way overcomes the disadvantages of having to deal with intricate calculations involving propagators that are non - local in space and time , in a model that breaks lorentz symmetry .
although the pseudo - qed formalism has been originally proposed more than a decade ago @xcite , it has remained nearly forgotten for years .
nevertheless , the discovery of graphene and related materials , like silicene , phosphorene , and metal dicalchogenides has opened a vast field of important applications in material systems .
the first attempts to use the pseudo - qed formalism in the context of two - dimensional dirac materials is revealing a giant potential to solve complex longstanding problems , as we show in the present work .
besides that , it has produced unexpected predictions of novel topological effects , such as the recently proposed interaction driven quantum valley hall effect @xcite . combined
, these results show the power of a theory that accounts for dynamical electron interactions and appropriately captures the dimensional mismatch between the electrons , which are constraint to move in a two - dimensional sheet , but interact via photons living in the three - dimensional space . the pseudo - qed theory , dual to maxwell qed in 2 + 1 dimensions , holds promises of becoming the unavoidable tool to describe the elusive effect of interactions in the new class of two - dimensional dirac systems .
this work was supported by cnpq ( brazil ) , capes ( brazil ) , faperj ( brazil ) , nwo - vici ( netherlands ) and by the brazilian government project science without borders .
we are grateful to d. haldane , s. sachdev , d. v. khveshchenko , a. h. castro neto , l. fritz , m. goerbig , and v. juricic for fruitful discussions .
the work by c.m.s . is part of the d - itp consortium , a program of the netherlands organization for scientific research ( nwo ) that is funded by the dutch ministry of education , culture and science ( ocw ) .
v.s.a . acknowledges nwo and the institute for theoretical physics of utrecht university for the kind hospitality . 100 schwinger , j. on quantum - electrodynamics and the magnetic moment of the electron .
_ phys .
rev . _ * 73 * , 416 - 417 ( 1948 ) .
van dyck , r. s. , jr . , schwinberg , p. b. & dehmelt , h. g. new high - precision comparison of electron and positron g factors .
_ phys .
rev .
lett . _ * 59 * , 26 - 29 ( 1987 ) .
hanneke , d. , fogwell , s. & gabrielse , g. new measurement of the electron magnetic moment and the fine structure constant .
_ phys .
rev .
lett . _ * 100 * , 120801 - 1/4 ( 2008 ) .
aoyama , t. , hayakawa , m. , kinoshita , t. & nio , m. tenth - order qed contribution to the electron g-2 and an improved value of the fine structure constant .
_ phys .
rev .
lett . _ * 109 * , 111807 - 1/4 ( 2012 ) .
kovner , a. , & rosenstein , b. kosterlitz - thouless mechanism of two - dimensional superconductivity .
_ phys .
rev .
b _ * 42 * , 4748 - 4751 ( 1990 ) ; dorey , n. & mavromatos , n. e. qed@xmath94 and two - dimensional superconductivity without parity violation . _ nucl .
phys .
b _ * 386 * , 614 - 680 ( 1992 ) ; teber , s. electromagnetic current correlations in reduced quantum electrodynamics , phys . rev .
d * 86 * , 025005 ( 2012 ) ; teber , s. two - loop fermion self - energy and propagator in reduced * * qed**@xmath95 .
phys .
rev .
d * 89 * , 067702 ( 2014 ) .
marino , e. c. quantum electrodynamics of particles on a plane and the chern - simons theory .
_ nucl .
phys .
b _ * 408 * , 551 - 564 ( 1993 ) .
kurganova , e. v. , elferen , h. j. , mccollam , a. , ponomarenko , l. a. , novoselov , k. s. , veligura , a. , wees , b. j. , maan , j. c. & zeitler , u. spin splitting in graphene studied by means of tilted magnetic - field experiments .
_ phys .
rev .
b _ * 84 * , 121407 - 1/4 ( r ) ( 2011 )
. song , y. j. , otte , a. f. , kuk , y. , hu , y. , torrance , d. b. , first , p. n. , de heer , a. w. , min , h. , adam , s. , stiles , m. d. , macdonald , a. h. & stroscio , j. a. high - resolution tunnelling spectroscopy of a graphene quartet .
_ nature _ * 467 * , 185 - 189 ( 2010 ) .
novoselov , k. s. , geim , a. k. , morozov , s. m. , jiang , d. , katsnelson , m. i. , grigorieva , i. v. , dubonos , s. v. & firsov , a. a. two - dimensional gas of massless dirac fermions in graphene .
_ nature _ * 438 * , 197 - 200 ( 2005 ) .
castro neto , a. h. , guinea , f. , peres , n. m. r. , novoselov , k. s. , geim , a. k. the electronic properties of graphene .
_ rev .
mod .
phys . _ * 81 * , 109 - 162 ( 2009 ) .
du , x. , skachko , i. , duerr , f. , luican , a. & andrei , e. y. fractional quantum hall effect and insulating phase of dirac electrons in graphene .
_ nature _ * 462 * , 192 - 195 ( 2009 ) .
bolotin , k. i. , ghahari , f. , shulman , m. d. , stormer , h. l. & kim , p. observation of the fractional quantum hall effect in graphene .
_ nature _ * 462 * , 196 - 199 ( 2009 ) .
dean , c. r. , young , a. f. , cadden - zimansky , p. , wang , l. , ren , h. , watanabe , k. , taniguchi , t. , kim , p. , hone , j. & shepard , k. l. multicomponent fractional quantum hall effect in graphene .
_ nat .
phys . _ * 7 * , 693 - 696 ( 2011 ) .
luican , a. , li , g. & andrei , e. y. quantized landau level spectrum and its density dependence in graphene .
_ phys .
rev .
b _ * 83 * , 041405(r ) ( 2011 ) .
elias , d. c. , gorbachev , r. v. , mayorov , a. s. , morozov , s. v. , zhukov , a. a. , blake , p. , ponomarenko , l. a. , grigorieva , i. v. , novoselov , k. s. , guinea , f. & geim , a. k. dirac cones reshaped by interaction effects in suspended graphene .
_ nat .
phys . _ * 7 * , 701 - 704 ( 2011 ) .
chae , j. , jung , s. , young , a. f. , dean , c. r. , wang , l. , gao , y. , watanabe , k. , taniguchi , t. , hone , j. , shepard , k. l. , kim , p. , zhitenev , n. b. & stroscio , j. a. renormalization of the graphene dispersion velocity determined from scanning tunneling spectroscopy .
_ phys .
rev .
lett . _
* 109 * , 116802 ( 2012 ) .
gonzlez , j. , guinea , f. & vozmediano , m. a. h. non - fermi liquid behavior of electrons in the half - filled honeycomb lattice ( a renormalization group approach ) .
_ nucl .
phys .
b _ * 424 * , 595 - 618 ( 1994 ) ; gonzlez , j. , guinea , f. & vozmediano , m. a. h. marginal - fermi - liquid behavior from two - dimensional coulomb interaction . _
phys .
rev .
b _ * 59 * , 2474 - 2477 ( 1999 ) .
foster , m. s. & aleiner , i. l. graphene via large _
n _ : a renormalization group study .
_ phys .
rev .
b _ * 77 * , 195413 - 1/24 ( 2008 ) .
gonzlez , j. , guinea , f. & vozmediano , m. a. h. unconventional quasiparticle lifetime in graphite .
_ phys .
rev .
lett . _ * 77 * , 3589 - 3592 ( 1996 ) .
juan , f. , grushin , a. g. & vozmediano , m. a. h. renormalization of coulomb interaction in graphene : determining observable quantities .
_ phys .
rev .
b _ * 82 * , 125409 - 1/7 ( 2010 ) . vozmediano , m. a. h & guinea , f. effect of coulomb interactions on the physical observables of graphene .
_ phys , scr . _ * t146 * , 014015 ( 2012 ) .
marino , e. c. , nascimento , l. o. , alves , v. s. & morais smith , c. interaction induced quantum valley hall effect in graphene . _
phys .
rev .
x _ * 5 * , 011040 - 1/13 ( 2015 ) .
amaral , r. l. p. g. & marino , e. c. canonical quantization of theories containing fractional powers of the dalembertian operator .
_ j. phys .
a _ * 25 * , 5183 - 5200 ( 1992 ) .
marino , e. c. , nascimento , l. o. , alves , v. s. & smith , c. m. unitarity of theories containing fractional powers of the dalembertian operator .
_ phys .
rev .
d _ * 90 * , 105003 - 1/6 ( 2014 ) .
li , g. , luican , a. & andrei , e. y. scanning tunneling spectroscopy of graphene on graphite . _ phys . rev
. lett . _ * 102 * , 176804 - 1/4 ( 2009 ) .
bollini , c. g. & giambiagi , j. j. lowest order `` divergent '' graphs in v - dimensional space .
_ phys .
lett .
b _ * 40 * , 566 - 568 ( 1972 ) ; t hooft , g. & veltman , m. regularization and renormalization of gauge fields . _
nucl .
phys.b _ * 44 * , 189 - 213 ( 1972 ) .
weinberg , s. new approach to the renormalization group .
_ phys .
rev .
d _
* 8 * , 3497 - 3509 ( 1973 ) ; t hooft , g. dimensional regularization and the renormalization group .
_ nucl .
phys .
b _ * 61 * , 455 - 468 ( 1973 ) .
weinberg , s. the quantum theory of fields , cambridge university press ( 1995 ) .
barnes , e. , hwang , e. h. , throckmorton , r. e. & das sarma , s. effective field theory , three - loop perturbative expansion , and their experimental implications in graphene many - body effects .
_ phys .
rev .
b _ * 89 * , 235431 - 1/57 ( 2014 ) . giuliani , a. , mastropietro , v. & porta , m. lattice quantum electrodynamics for graphene . _
ann .
phys .
_ ( ny ) * 327 * , 461 ( 2012 ) .
das sarma , s. , hwang , e. h. & tse , w .- k .
many - body interaction effects in doped and undoped graphene : fermi liquid versus non - fermi liquid .
_ phys .
rev .
b _ * 75 * ,
121406 ( 2007 ) . goerbig , m. electronic properties of graphene in a strong magnetic field .
_ rev .
mod .
phys . _ * 83 * , 1193 - 1243 ( 2011 ) .
das sarma , s. & hwang , e. h. velocity renormalization and anomalous quasiparticle dispersion in extrinsic graphene .
_ phys .
rev .
b _ * 87 * , 045425 ( 2013 ) .
kalesaki , e. , delerue , c. , smith , c. m. , beugeling , w. , allan , g. & d. vanmaekelbergh , d. dirac cones , topological edge states , and nontrivial flat bands in two - dimensional semiconductors with a honeycomb nanogeometry .
_ phys .
rev .
x _ * 4 * , 011010 ( 2014 ) .
boneschanscher , m. p. , evers , w. h. , geuchies , j. j. , altantzis , t. , goris , b. , rabouw , f. t. , van rossum , s. a. p. , van der zant , h. s. j. , siebbeles , l. d. a. , van tendeloo , g. , swart , i. , hilhorst , j. , petukhov , a. v. , bals , s. , vanmaekelbergh , d. long - range orientation and atomic attachment of nanocrystals in 2d honeycomb superlattices .
_ science _ * 344 * , 1377 - 1380 ( 2014 ) .
beugeling , w. , kalesaki , e. , delerue , c. , niquet , y .- m . ,
vanmaekelbergh , d. & smith , c. m. topological states in multi - orbital hgte honeycomb lattices .
_ nat .
comm . _ * 6 * , 6316 ( 2015 ) .
khveshchenko , d.v . & leal , h. excitonic instability in layered degenerate semimetals .
_ nucl .
phys .
b _ * 687 * , 323 ( 2004 ) ; khveshchenko , d. v. coulomb - interacting dirac fermions in disordered graphene . _ phys .
rev .
b _ * 74 * , 161402 ( 2006 ) ; khveshchenko , d. v. massive dirac fermions in single - layer graphene .
_ j. phys . : condens
. matter _ * 21 * , 075303 ( 2009 ) . according to refs .
@xcite , @xmath96 .
there is a large imprecision , and the actual value may depend on doping , on substrate , etc . *
supplementary information + valley- and spin - splitting due to interactions in graphene * + natlia menezes et al . |
one of the most persistent problems in solar physics has been the unambiguous identification of the mechanisms that heat the sun s corona and accelerate the solar wind .
many processes have been proposed for converting some fraction of the mechanical energy in subphotospheric convective motions to heat , but it has proved very difficult to make distinguishing comparisons between the predictions of these competing ideas and specific observations .
we are entering an era , though , where both the models and the measurements are improving to the point of soon being able to eliminate many of the candidate theories .
for example , it seems increasingly clear that closed loops in the low corona are heated by small - scale , intermittent magnetic reconnection that is driven by the continual stressing of their magnetic footpoints ( see recent reviews by longcope 2004 ; gudiksen 2005 ; aschwanden 2006 ; klimchuk 2006 ) . in this paper
we model the coronal heating along open field lines that reach into interplanetary space .
we construct a self - consistent model of the photosphere , chromosphere , corona , and solar wind that is driven mainly by magnetohydrodynamic ( mhd ) turbulence and is free of arbitrary `` heating functions '' that have been used in the past .
this work continues earlier studies ( cranmer & van ballegooijen 2005 ; cranmer 2005a ) which used prescribed empirical descriptions of the density and flow velocity along an open flux tube in order to compute the rates of turbulent heating and acceleration . the main goal of this paper is to provide a possible explanation for the origin and properties of fast and slow solar wind streams .
another goal , though , is to illustrate how the particular description of mhd turbulence can be applied to more advanced modeling efforts specifically those that attempt to reproduce the full three - dimensional and time - dependent nature of the corona and heliosphere ( e.g. , riley et al . 2001 ;
roussev et al . 2003 ; tth et al .
2005 ; usmanov & goldstein 2006 ) .
coronal heating and solar wind acceleration have been known to be closely linked since the initial contributions of parker ( 1958 ) .
however , nearly all subsequent theoretical attempts to model both processes together have made limiting assumptions about either the heating ( i.e. , ad hoc energy input rates ) or the acceleration ( i.e. , a prescribed mass flux ) .
it has been realized over the past few decades that it is key to resolve the full chromosphere - corona transition region in order to produce a model with an internally consistent radial dependence of pressure , density , and flow speed ( see hammer 1982 ; hansteen & leer 1995 ; hansteen et al .
1997 ; lie - svendsen & esser 2005 ) . to our knowledge , the only solar wind models
other than the ones presented in this paper that contain _ both _ a first - principles approach to coronal heating and a self - consistent chromosphere , transition region , and mass flux are those of suzuki & inutsuka ( 2005 , 2006 ) . and @xmath0 k ) .
see 2 below for a comparison of the relevant physical assumptions and numerical approaches . ]
the study of mhd turbulence as a potential source of heating for the solar wind goes back to coleman ( 1968 ) and jokipii & davis ( 1969 ) .
hollweg ( 1986 ) extended these ideas down into the corona and laid the foundations for the cascade - driven heating rates used by isenberg ( 1990 ) , li et al .
( 1999 ) , matthaeus et al .
( 1999 ) , dmitruk et al .
( 2001 , 2002 ) , and this paper .
these ideas are highly complementary to theoretical models of wave dissipation via ion cyclotron resonance ( e.g. , hollweg & isenberg 2002 ; cranmer 2000 , 2001 , 2002 ) that have been invoked to explain the observed preferential heating of heavy ions in the extended corona ( kohl et al .
1997 , 1998 , 2006 ) . in the cascade paradigm
, the energy in large - scale alfvnic fluctuations must eventually be dissipated in small - scale ( short wavelength or high frequency ) collisionless kinetic modes .
the results from this paper can thus be used as initial conditions for detailed models of the anisotropic turbulent cascade and minor ion heating .
the remainder of this paper is organized as follows . in
2 we present an overview of the general principles involved in our modeling effort ( i.e. , why certain physical processes are included or excluded ) . 3 gives the conservation equations that are solved and the adopted prescriptions for basic ingredients such as radiative heating / cooling and conduction .
the detailed treatment of waves is described in the following three sections : acoustic waves and shock heating in 4 , alfvn waves and turbulent cascade in 5 , and the ponderomotive wave pressure acceleration due to both types of fluctuations in
the numerical methods used to solve the equations and the newly developed computer code , called zephyr , are described in 7 .
we then present a comprehensive set of solutions for the solar wind emerging from coronal holes , helmet streamers , and active regions ( 8) .
a summary of the major results of this paper , together with a discussion of the implications for understanding the winds of other stars , is given in
we consider the _ one - dimensional _ variation of plasma parameters along a radially pointed magnetic flux tube rooted in the solar photosphere and open to interplanetary space .
some effects of the multi - dimensional magnetic field in the photosphere and chromosphere are included as explicit superradial expansion of the flux tube ( see also cranmer & van ballegooijen 2005 ) , but otherwise we ignore the effects of neighboring closed flux tubes .
we assume that most of the plasma that eventually becomes the time - steady solar wind originates in small ( 100 km sized ) intergranular flux concentrations that are concentrated most densely in the supergranular network .
the models we construct are _ time - independent _ solutions to the hydrodynamic conservation equations .
this may be a restrictive simplification since all observed layers of the solar atmosphere are intrinsically dynamic over a wide range of time scales .
it is evident , for example , that the transition region is `` strongly nonuniform and magnetically structured '' ( marsch et al . 2006 ) to the extent that one - dimensional layered models may miss some important physical effects at those heights .
however , the corona above the transition region seems to exist in its hot ( @xmath1 k ) state all the time .
also , in situ measurements of the solar wind plasma often can be interpreted clearly as a superposition of a quasi - steady supersonic flow ( which depends mainly on the coronal source region of the streams being measured or on stream - stream interactions ) and a rapidly varying turbulent or wavelike component .
we thus separate these two scales and model the fluctuations using energy conservation equations derived from linear wave theory .
it is especially beneficial to treat waves and turbulent motions statistically rather than follow the oscillations explicitly in time when the dynamically important ranges of wavelength and period span many orders of magnitude .
alfvn waves in the solar wind are measured to have periods from seconds to days ( e.g. , goldstein et al . 1995 ; tu & marsch 1995 ) and there is evidence that the waves that dominate the dissipation in the collisionless extended corona have periods even below @xmath2 s ( mckenzie et al . 1995 ; cranmer et al .
if this full range of scales had to be resolved in order to track the motions in a time - dependent numerical model , it would be extremely difficult to simulate the macroscopic plasma properties simultaneously from the photosphere to the heliosphere .
there are , however , some clear disadvantages in modeling fluctuations as statistical `` wave energy fluxes '' rather than as explicit variations .
any nonlinear processes , such as shock steepening , turbulent cascade , mode conversion , or ponderomotive forces , must be inserted beforehand as source or sink terms in the time - steady conservation equations .
time - dependent mhd models have the benefit of producing such effects naturally ( e.g. , ofman & davila 1998 ; ofman 2005 ; bogdan et al . 2002 , 2003 ; suzuki & inutsuka 2005 , 2006 ) .
however , by resolving only certain temporal and spatial scales , these models are limited in ways that a statistical treatment is not .
for example , the relative importance of shock dissipation in the time - dependent models of suzuki & inutsuka ( 2005 , 2006 ) may be an artifact of either the inability to resolve small enough scales ( which could drive processes such as perpendicular cascade , landau damping , and collisionless particle energization at shocks ) or the waveguide - like trapping of fluctuations along the model flux tube .
it is important to note , though , that neither extreme i.e .
, neither coarse time - dependent simulations nor statistical ( and more approximate ) time - independent models is ideal .
there may be value in seeking some kind of hybrid methodology between these two approaches . as indicated above
, the only source of heating for the chromosphere and corona that we consider is the dissipation of _ waves and turbulent motions . _ for open flux tubes , a substantial fraction of the energy appears to be deposited at large heights in the wind s acceleration region
i.e . , at spatial scales much larger than the sizes of low - lying closed loops in the quiet sun and active regions .
this demands the energy be propagated for some distance , presumably by waves or turbulent eddies , before it is dissipated .
however , the general phenomenology of turbulence is probably not limited to the open - field regions .
concepts from turbulence theory have been applied to the full range of time scales for closed - field coronal energy input as well , from the most rapid ( ac ) wavelike oscillations to the slowest ( dc ) quasistatic stresses on magnetic footpoints ( see van ballegooijen 1986 ; hendrix & van hoven 1996 ; milano et al .
1997 ; gmez et al .
2000 ; chae et al .
indeed , some recent simulations of intermittent turbulent heating in closed loops have been interpreted using similar cascade rate expressions as we use in this paper ( rappazzo et al .
2007 ) .
the models presented below include the effects of both alfvn waves and acoustic waves on the mean flow .
we assume implicitly that all waves propagate parallel to the radially oriented flux tube , but this is not an essential feature .
in the magnetically dominated corona ( where @xmath3 ; @xmath4 being the ratio of gas pressure to magnetic pressure ) the modeled alfvn waves can be considered essentially to be the sum of true alfvnic wave power and fast - mode mhd wave power . the acoustic waves can be considered equivalent to slow - mode mhd waves . as a starting point
, we neglect all nonlinear couplings between the alfvn and acoustic wave modes . by ignoring the enhanced reflection and dissipation that may arise because of these couplings ( e.g. , suzuki & inutsuka 2005 , 2006 ) we essentially _ underestimate _ the heat deposited in the extended corona and heliosphere .
the fact that sufficient energy nonetheless exists to heat the corona and accelerate the wind ( mainly from the alfvn waves ) appears to deemphasize the need for such mode couplings .
it should be noted that the attention given to modeling the steepening and shock dissipation of the acoustic waves ( 4 ) does not seem to have much of a `` payoff '' in the resulting properties of the corona and solar wind .
the fully ionized outer atmosphere is extremely insensitive to the magnitude ( or sometimes even the presence ) of acoustic wave power that comes up from below the photosphere ( 8.2 ) . accordingly
, many prior studies of the coupled chromosphere and corona used a much simpler prescription for maintaining the chromosphere .
we believe it is important , though , to treat both the chromospheric and coronal heating at a comparable level of first - principles modeling .
we also anticipate that the zephyr code developed here will be applied to the simulations of winds of other late - type stars that may have acoustically heated coronae ( e.g. , mullan & cheng 1993 ; schrijver 1995 ) , and thus the acoustic waves should be treated as realistically as possible .
one final limitation of the models presented below is that the energy conservation is treated in a _
one - fluid _ manner .
the protons , electrons , and heavy ions are modeled as having a common flow speed @xmath5 and temperature @xmath6 , with microscopic velocity distributions that are simple isotropic maxwellians .
this is an extreme simplification of reality , since it has been known since the 1960s that the in situ solar wind exhibits significant departures from a thermalized equilibrium ( see reviews by hundhausen 1972 ; feldman & marsch 1997 ) .
many of these effects persist down into the extended corona as well ( e.g. , kohl et al . 2006 ) .
therefore , theoretical models have long included a range of attempts to deal with these features .
fluid - based models have been extended to solve separate conservation equations for each particle species , and they have been reconstructed in various ways based on different parameterizations for the anisotropic velocity distributions .
several purely kinetic models have also been attempted ( see above - cited reviews , also cranmer 2002 ; hollweg & isenberg 2002 ; marsch 2005 ) . despite the potentially clumsy `` averaging '' over real kinetic effects
, we believe the one - fluid approach is the _ most consistent with our present state of knowledge _ about the primary source of energy deposition : mhd turbulent cascade .
there are a number of competing ideas in the literature regarding how the cascade proceeds to its smallest kinetic scales and how either linear or nonlinear damping transfers wave energy to the particles ( for a recent summary see
5.2.4 of kohl et al . 2006 ) .
we thus do not yet know , from first principles , how to partition the cascaded wave energy between particle species ( @xmath7 ) and between various directions in microscopic velocity space ( @xmath8 ) .
performing such partitioning at the present time would essentially add new free parameters into the model .
our adopted rate of alfvnic coronal heating ( eq . [ [ eq : dmit ] ] ) thus deals only with the total energy flux that cascades from large to small scales and not the specific means of dissipation once the energy reaches the small scales .
the equations governing the expansion of a time - steady stellar wind are derivable by taking successive velocity moments of the boltzmann equation , together with some assumption about the shape of the velocity distribution function in order to close the otherwise infinite chain of moment equations ( see , e.g. , braginskii 1965 ; collins 1989 ; marsch 2005 ) . in this section
we describe the conservation equations used in the models ( 3.1 ) and the adopted prescriptions for heat transport due to radiation ( 3.2 ) and conduction ( 3.3 ) .
we consider the flow of a pure hydrogen plasma along a radially oriented magnetic field .
the goal is to solve for the time - steady radial dependence of the mass density @xmath9 , the bulk flow speed @xmath5 , and the maxwellian temperature @xmath6 .
the distance along the magnetic flux tube is denoted either as @xmath10 , measured from sun - center , or @xmath11 , measured from the lower boundary of the model in the solar atmosphere ( essentially the photosphere ) . for completeness
, the equations below contain the dependence on time @xmath12 , though these terms are set to zero in the time - independent solutions that we describe below .
the equation of mass conservation is @xmath13 where @xmath14 is the cross - sectional area of the one - dimensional flux tube along which the wind flows .
magnetic flux conservation demands that the product @xmath15 is constant along the flux tube , where @xmath16 is the field strength that we specify explicitly ( see 8) .
a time - steady one - dimensional flow thus constrains the product @xmath17 to be proportional to @xmath18 .
the equation of momentum conservation is @xmath19 where @xmath20 is the gas pressure , @xmath21 is the newtonian gravitation constant , and @xmath22 is the mass of the star .
the mass of the plasma in the modeled stellar wind region is assumed to be negligible from a gravitational standpoint . also , @xmath23 is the bulk acceleration on the plasma due to _ wave pressure ; _ i.e. , the nondissipative net ponderomotive force due to the propagation of waves through an inhomogeneous medium ( 6 ) .
the equation of internal energy conservation is @xmath24 where @xmath25 is the internal energy density and the terms on the right - hand side are volumetric heating / cooling rates due to radiation , conduction , alfvn wave damping , and acoustic ( sound ) wave damping .
the terms on the left - hand side that depend on @xmath5 are responsible for enthalpy transport and adiabatic cooling in the accelerating wind .
the coupling of the above equations requires additional constitutive relations to be specified : @xmath26 @xmath27 @xmath28 where @xmath29 is boltzmann s constant and a monatomic ratio of specific heats @xmath30 is used .
the total particle number density @xmath31 is given by the sum of the hydrogen number density @xmath32 and the electron density @xmath33 , and the number densities of neutral hydrogen and protons are denoted @xmath34 and @xmath35 , respectively . for the pure hydrogen plasma assumed in this paper , @xmath36 .
the mass density is given by @xmath37 , although the second term is safely considered to be negligible .
note that equation ( [ eq : edef ] ) becomes the ideal gas relation @xmath38 in the purely neutral limit , where @xmath39 is the specific heat at constant volume .
ionization is taken into account by the internal energy s dependence on @xmath40 , the ionization potential of hydrogen from the ground level ( 13.6 ev ) .
we use the same convention in the definition of @xmath25 as ulmschneider & muchmore ( 1986 ) and mullan & cheng ( 1993 ) .
an alternate definition of the internal energy is possible , though , which is essentially given by @xmath41 .
this version reduces to the above ideal - gas relation in the limit of a fully ionized plasma ( e.g. , mcclymont & canfield 1983 ; fontenla et al .
either definition is consistent with the combined equations of mass and energy conservation given above .
we adopted a relatively simple prescription to compute the ionization fraction @xmath42 .
this quantity is parameterized as a tabulated function of @xmath6 only , where we have used the ionization balance from a recent semi - empirical model of the solar photosphere , chromosphere , and transition region ( e. avrett 2005 , private communication ; see also fontenla et al . 1993 , 2002 , 2006 ) .
the tabulated model we use is a modified version of the quiet - sun model c from the above papers , and it is the same one used by cranmer & van ballegooijen ( 2005 ) to model the magnetic structure of the chromospheric network . for convenience
we call this model fal - c@xmath43 .
figure 1 shows the adopted ionization fraction @xmath44 , as well as the corresponding neutral fraction @xmath45 ( in order to see small departures from total ionization at high temperatures ) and the radiative cooling function that was derived in part from the fal - c@xmath43 model ( see 3.2 ) . for temperatures below the minimum tabulated value of about 4500 k , we prevent @xmath44 from decreasing below a minimum value of @xmath46 , because photoionization is expected to always keep some small fraction of metals ionized at and above the photosphere .
we tested the zephyr code with a more self - consistent model of the hydrogen ionization balance .
the results were extremely similar , though , to those using the tabulated fraction shown in figure 1 .
this model consisted of a three - level hydrogen atom , where the @xmath47 and @xmath48 levels were assumed to remain in relative local thermodynamic equilibrium ( lte ) and the full rate equation between @xmath48 and the continuum was solved iteratively with collisional and radiative terms from hartmann & macgregor ( 1980 ) , vriens & smeets ( 1980 ) , ferland et al .
( 1992 ) , and ferguson & ferland ( 1997 ) . because the tabulated version was much faster in terms of computation time ,
though , we decided to use it in the models shown below .
it will be important to include this kind of self - consistent ionization balance when adapting this method to the winds of other stars ( e.g. , natta et al .
1988 ) .
a complete treatment of the non - lte transfer of energy between radiation and matter in a partially ionized plasma is beyond the scope of this paper .
detailed computational efforts to model chromospheric and coronal radiative transfer effects ( e.g. , avrett & loeser 1992 ; carlsson & stein 1997 , 2002 ; rammacher et al . 2005 )
are important for reproducing the spectrum or studying time - dependent or multidimensional dynamics , but for our purposes a simpler treatment is warranted .
we use a similar general `` bridging '' approach as mullan & cheng ( 1993 ) to combine different limiting cases in the optically thick lower atmosphere and the optically thin upper atmosphere .
the adopted radiative heating / cooling rate is given by @xmath49 where @xmath50 is the rosseland mean optical depth and @xmath51 is a constant that defines where the rate is dominated by the optically thick or thin limits ( see also ludwig et al .
for a spherical stellar atmosphere , we use a definition for the rosseland optical depth , @xmath52 that contains the geometrical correction factor suggested by lucy ( 1971 , 1976 ) ; it is unimportant in the case of the sun where @xmath53 in regions where @xmath54but we include it for later use in the extended atmospheres of evolved stars .
the rosseland mean opacity @xmath55 ( in @xmath56 g@xmath57 ) is interpolated as a function of temperature and pressure from the table of kurucz ( 1992 ) . for a given one - dimensional model
, we integrate downward from the upper boundary far out in the supersonic wind ( which has an assumed optical depth of zero ) to compute @xmath58 . in the optically thick photosphere and lower chromosphere
we assume that the heating and cooling is dominated by continuum photons emitted and absorbed in lte , with @xmath59 ( e.g. , mihalas 1978 ; vgler et al .
2004 ) . above
, @xmath60 is the frequency - integrated source function , assumed here in lte to equal the local integrated planck function , @xmath61 , with @xmath62 being the stefan - boltzmann constant .
the frequency - integrated mean intensity @xmath63 is given by the gray atmosphere dependence on optical depth @xmath50 , @xmath64 \,\ , , \label{eq : jtau}\ ] ] where @xmath65 k is the solar effective temperature and @xmath66 is the hopf ( 1930 , 1932 ) function , for which we use the following fit ( accurate to better than 0.2% ) , @xmath67 also , @xmath68/2 $ ] is the spherical dilution factor suggested for use in this context by chandrasekhar ( 1934 ) and lucy ( 1971 , 1976 ) .
@xmath69 is positive giving net heating when @xmath6 is less than the radiative equilibrium temperature ( @xmath70 ) given by @xmath71 , and it is negative with net cooling when @xmath72 . in the optically thin upper chromosphere and corona
, we use a modified temperature - dependent radiative cooling function that has been computed for photon losses due to a large collection of spectral lines and continuum processes , @xmath73 ( see also cox & tucker 1969 ; anderson & athay 1989b ; schmutzler & tscharnuter 1993 ) . the radiative loss function @xmath74 is shown in figure 1 , and it has been assembled from three sources .
( 1 ) for hydrogen and helium above 10@xmath75 k , we used the output from the pandora radiative transfer code which produced the fal - c@xmath43 semi - empirical model discussed above .
( 2 ) for other elements above 10@xmath75 k , we used a tabulated radiative loss function from version 4.2 of the chianti atomic database ( young et al . 2003 ) using a traditional solar abundance mixture ( grevesse & sauval 1998 ) and collisional ionization balance ( mazzotta et al .
( 3 ) for temperatures below 10@xmath75 k , we used the fitting function for partially ionized hydrogen given by scholz & waters ( 1991 ) .
for these low temperatures we also added a constant lower - limit value of @xmath76 erg @xmath77 s@xmath57 to @xmath78 to account for photoionized metals , molecules , and dust , which may be important contributors to radiative cooling at very low temperatures in late - type stellar atmospheres ( e.g. , schirrmacher et al .
2003 ) . in order to ensure that the optically thin parts of the atmosphere would smoothly approach radiative equilibrium ( in the absence of nonradiative heating ) in the same way as in the optically thick parts of the atmosphere , we multiplied the standard cooling function by the term in parentheses in equation ( [ eq : qradthin ] ) . for temperatures above @xmath79
this correction factor rapidly approaches unity .
note from figure 1 that there is no isolated `` lyman alpha peak '' at temperatures of 12 @xmath80 k. for solar atmosphere models , this peak seems to be a spurious feature that appears in optically thin radiative loss curves computed without a radiation field .
the fal - c@xmath43 model used here contained a full non - lte treatment of hydrogen as well as ambipolar diffusion that also provides additional smearing of discrete features related to the strong h i ly@xmath81 transition ( see also kuin & poland 1991 ; fontenla et al .
2002 ) . with regard to the extension to other stars ,
the computation of the radiative cooling rate is in a similar situation as the ionization fraction discussed in
our use of tabulated quantities from fal - c@xmath43 seems to be reasonable for modeling the solar atmosphere , but it will eventually need to be replaced with a more self - consistent procedure . specifically , evolved low - gravity stars with high mass loss rates are expected to have more optically thick chromospheres .
hartmann & macgregor ( 1980 ) , canfield & ricchiazzi ( 1980 ) , and mullan & cheng ( 1993 ) computed approximate optically thick radiation losses by taking account of reduced escape probabilities in the damping wings of strong lines ; it is possible that these methods can be extended in a robust way to future time - steady stellar wind models .
for the radially oriented flux tubes that we consider , the conductive energy exchange rate is dominated by the divergence of a parallel heat flux density , i.e. , @xmath82 because the solar atmosphere undergoes a transition from being strongly collisionally coupled ( at low heights and high densities ) to being nearly collisionless ( at large heights and low densities ) , it was realized long ago that the classical spitzer - hrm ( sh ) prescription for thermal conductivity must break down somewhere in the corona and solar wind .
we thus follow wang ( 1993 ) , and others , by using a semi - empirical bridging law between the sh heat flux @xmath83 and a completely collisionless `` free - streaming '' heat flux @xmath84 , @xmath85 where @xmath86 is the local wind expansion rate and @xmath87 is the electron - electron coulomb collision frequency , @xmath88 ( braginskii 1965 ; olsen & leer 1996 ) .
the electron coulomb logarithm is approximated by @xmath89 a more accurate version of equation ( [ eq : qpara ] ) was derived by cuperman & dryer ( 1985 ) . under a number of simplifying approximations ,
though ( such as nearly maxwellian distributions ) , their expression reduces to something very close to the above bridging formula . the radial distance where the heat flux undergoes the transition from collisional to collisionless
can be estimated by locating the point at which @xmath90 .
for the models of low - density high - speed solar wind streams presented below , this occurs at about @xmath91 solar radii ( @xmath92 ) ; for the models of high - density low - speed solar wind , this occurs at @xmath9380 @xmath92 . in the collisionally dominated limit , we assume sh conduction , @xmath94 where the thermal conductivity is assumed to be dominated by contributions from free electrons and neutral hydrogen , @xmath95 ( e.g. , nowak & ulmschneider 1977 ; mcclymont & canfield 1983 ) .
the effects of protons and heavy ions are neglected because they tend to be overwhelmed by the electron conductivity in regions of appreciable ionization ( see also ulmschneider 1970 ; hansteen & leer 1995 ) .
the electron conductivity is given by @xmath96 ( spitzer 1962 ; braginskii 1965 ) .
partial ionization effects ( i.e. , @xmath97 and @xmath98 ) are parameterized using the results of shmeleva & syrovatskii ( 1973 ) and mcclymont & canfield ( 1983 ) , with @xmath99 where @xmath100 is the ratio of the electron - electron collision time @xmath101 to the electron - neutral hydrogen time @xmath102 , and @xmath103 similarly , the neutral hydrogen conductivity is given by @xmath104 the above partial ionization effects are generally unimportant in the solar atmosphere , but we include them for some attempt at completeness , and with anticipation that the cooler and more optically thick chromospheres of other stars may depend more sensitively on them .
in the collisionless limit we use the free - streaming heat flux derived for escaping electrons by hollweg ( 1974 , 1976 ) , which was based on empirical constraints from forslund ( 1970 ) and perkins ( 1973 ) on the electron velocity distribution , @xmath105 where @xmath106 is an order - unity correction factor that depends on how the wings of the electron velocity distribution depart from a maxwellian shape . we take a constant value of @xmath107 as has been often used in solar wind modeling ( see , e.g. , leer et al . 1982 ; scudder & olbert 1983 ; withbroe 1988 ; canullo et al . 1996 ; landi & pantellini 2003 ) .
tests using hollweg s ( 1974 ) more detailed prescription for how @xmath106 should depend on the solar wind speed and electron temperature produced no more than a 10% change in @xmath108 from the @xmath107 model ( and no more than a 2% change in the mass loss rate or terminal wind speed ) .
the simpler constant value was used in all subsequent zephyr models .
note that equations ( [ eq : qpara ] ) and ( [ eq : qfs ] ) are in some ways similar to the heat flux relations used by smith & auer ( 1980 ) and others for flare plasmas , where the classical heat flux is assumed to _ saturate _ at a value no larger than a threshold flux proportional to @xmath109 ( where @xmath110 is the electron thermal speed ) .
the free - streaming heat flux given above is dimensionally similar to the saturated heat flux , but the former is proportional to @xmath111 rather than @xmath109 .
these two limits are related , though , because saturation occurs when the electron - electron collisional mean free path becomes larger than the local temperature scale height .
when heat can no longer be carried diffusively by classical conduction , it may be advected directly at some characteristic velocity @xmath112 , and the heat flux then becomes proportional to @xmath113 . in the low - density supersonic solar wind ,
the characteristic speed for , e.g. , the total enthalpy flux is @xmath114 . to the extent
that @xmath106 is an order - unity correction factor ( that depends weakly on @xmath110 ) , this applies also to the heat flux carried in the non - maxwellian tail of the electron distribution . for a high - density laboratory plasma ,
though , mannheimer & klein ( 1975 ) showed that @xmath112 scales directly with @xmath110 ( see also smith & lilliequist 1979 ; craig & davys 1984 ) .
we include the time - averaged effects of acoustic waves that propagate parallel to the magnetic field .
the only source of these waves that we consider is the convective motion below the photosphere which channels compressive wave energy into the magnetic flux tubes .
deep in the atmosphere these fluctuations take the form of longitudinal , or sausage - mode tube waves ( e.g. , spruit 1982 ; roberts 2000 ) , but because the individual flux tubes appear to merge together somewhere in the low chromosphere into a region filled with magnetic field ( cranmer & van ballegooijen 2005 ) we treat these waves as standard acoustic oscillations and do not consider thin - tube dispersive effects .
a separate source of compressive waves , which we do not model , may be the gradual parametric decay of nonlinear mhd waves in the outflowing stellar wind ( see sagdeev & galeev 1969 ; goldstein 1978 ; jayanti & hollweg 1993 ) .
it is usually assumed that these waves have very low frequencies and thus are likely to have small rates of damping and steepening , and thus a minimal impact on chromospheric or coronal heating .
below we describe how zephyr models the propagation , steepening , and dissipation of individual monochromatic `` packets '' of acoustic wave energy ( 4.1 ) and how the complete power spectrum is specified ( 4.2 ) .
an arbitrarily steepened acoustic wave / shock train travels along the field with constant frequency @xmath115 and a radially varying wavenumber @xmath116 determined by the dispersion relation @xmath117 .
the sound speed is given by @xmath118 .
the energy density @xmath119 of linear acoustic fluctuations obeys an equation of wave action conservation , @xmath120 \ , = \ , - \frac{q_s}{\omega ' } \label{eq : j77ac}\ ] ] ( e.g. , jacques 1977 ; koninx 1992 ) .
the doppler shifted frequency in the frame of the accelerating wind is given by @xmath121 and the wave energy density is given by @xmath122 where @xmath123 is a dimensionless shape factor determined by the spatial profile of the waves ( see below ) .
for a small - amplitude sinusoid , @xmath124 , and for a fully steepened sawtooth or n - wave , @xmath125 .
the parallel velocity variance , or squared wave amplitude , is specified as @xmath126 .
below we give the wave energy flux @xmath127 as a lower boundary condition for the zephyr code ; this quantity is converted into wave energy density using @xmath128 u_{s } \,\ , .
\label{eq : fsdef}\ ] ] note that in the limit of a static plane - parallel atmosphere ( @xmath129 and @xmath130 ) equations ( [ eq : j77ac])([eq : fsdef ] ) simplify into the standard flux conservation quantities implicit in wave - heated models of the solar atmosphere since the initial work of schwarzschild ( 1948 ) , biermann ( 1948 ) , and osterbrock ( 1961 ) .
the right - hand side of equation ( [ eq : j77ac ] ) couples the dissipation of acoustic wave energy to the @xmath131 heating term in equation ( [ eq : dedt ] ) .
we include two sources of wave damping : heat conduction and entropy gain at shock discontinuities . for the present models we ignore radiation damping of the waves ( which damps high frequencies at low heights , but does not technically provide heat to the plasma ) , viscosity , and ion - neutral friction .
the acoustic heating rate is given by @xmath132 where @xmath133 is the linear damping rate due to heat conduction and @xmath134 is the net entropy jump across a shock ( which is nonzero only above the height where the wave train has steepened into shocks )
. the damping rate due to heat conduction is given for adiabatic waves ( @xmath30 ) by @xmath135 ( e.g. , landau & lifshitz 1959 ; hung & barnes 1973 ; whang 1997 ) .
we use the same thermal conductivity @xmath136 as is used in the classical sh heat flux ( eq . [ [ eq : qsh ] ] ) . for simplicity , the electron and neutral hydrogen conductivities
are assumed to dominate proton and heavy ion heat conduction as well as other classical transport processes .
note that @xmath133 is the damping rate for the wave amplitude ; the damping rate for wave energy is twice that value ( see the factor of 2 in eq . [ [ eq : qs ] ] ) .
the major contribution to the acoustic heating rate comes from shock steepening and dissipation .
the gain in internal energy across an ideal inviscid shock is given by @xmath137 \label{eq : tds}\ ] ] where subscripts 1 and 2 denote quantities measured on the upstream ( supersonic ) and downstream ( subsonic ) sides of the shock ( landau & lifshitz 1959 ) .
the above expression does not contain the internal energy component from @xmath138 work but only the energy that goes into dissipation .
equation ( [ eq : qs ] ) uses the approximation from so - called `` weak - shock theory '' that the volumetric heating rate is given by the internal energy dissipated at one shock divided by the mean time between shock passages in a periodic train .
this assumption breaks down for very strong shocks in the chromosphere , which dissipate their energy in a relatively narrow zone behind the shock ( e.g. , carlsson & stein 1992 , 1997 ) , but the models presented below do not develop such strong shocks . to evaluate equation ( [ eq : tds ] )
we used the classical rankine - hugoniot relations for a monatomic ( @xmath30 ) gas .
these relations also are valid for a plasma with a constant ionization state across the shock , and carlsson & stein ( 2002 ) found that shock trains in the solar chromosphere often approach this nearly steady - state condition .
the density and temperature jump relations can be written in terms of the upstream mach number @xmath139 , @xmath140
@xmath141 [ ( \gamma - 1 ) m_{1}^{2 } + 2]}{(\gamma + 1)^{2 } m_{1}^{2 } } \label{eq : t2t1}\ ] ] note , though , that for shocks of arbitrary strength , equation ( [ eq : tds ] ) requires the absolute upstream and downstream temperatures @xmath142 and @xmath143 to be computed , not just their ratio .
( in the weak - shock limit , @xmath144 , the latter being the `` mean '' model atmospheric temperature . ) in many astrophysical models of shocks , @xmath142 is often assumed to be the undisturbed equilibrium state and @xmath143 is computed from equation ( [ eq : t2t1 ] ) .
however , in time - dependent simulations of chromospheric shocks , @xmath142 is often seen to fall _ below _ the time - averaged mean temperature @xmath6 and often also below the radiative equilibrium value @xmath145 k. this is believed to arise from adiabatic expansion behind the shock . to compute @xmath142 and @xmath143
, we first determine the upstream and downstream densities @xmath146 and @xmath147 relative to the known background model density @xmath9 .
with respect to the propagating shock train , the background density can be defined as that which occurs when the shock passes through zero velocity in the reference frame of the undisturbed atmosphere .
for an ideal sawtooth - shaped n - wave , this occurs for a given height at a time halfway between shock passages . using this definition ,
the analytic results of bertschinger & chevalier ( 1985 ) can be used to estimate the ratio of the minimum ( preshock ) density to the background value to be @xmath148 which then allows both @xmath146 and @xmath147 to be computed . to convert densities into temperatures , some knowledge of the thermodynamic cycle of the shock must be incorporated . in other words , after the gas is heated and compressed , we need to know what `` path '' it takes as it cools and expands back to the preshock values , in order to be heated and compressed again as the next shock in the train goes by .
nearly all time - dependent models of periodic shocks in stellar atmospheres have found that shocks first undergo rapid radiative cooling at a roughly constant density , followed by nearly adiabatic expansion back to the preshock density and temperature ( e.g. , weymann 1960 ; osterbrock 1961 ; ulmschneider et al .
1978 ; bowen 1988 ) .
this second phase seems to dominate the time between shock passages usually encompassing the halfway point used above as the definition for the undisturbed density
so we assume that the cooling from the mean temperature @xmath6 to the preshock , or upstream temperature @xmath142 is adiabatic , and @xmath149 this , in combination with equations ( [ eq : d2d1])([eq : d1d0 ] ) , completes the specification of @xmath142 and @xmath143 needed to compute @xmath150 . in the limit of low - amplitude shocks ( i.e. , @xmath151 , where @xmath152 ) , equation ( [ eq : tds ] ) reduces to the standard weak - shock limit @xmath153 ( e.g. , ulmschneider 1970 ; stein & schwartz 1972 , 1973 ; mihalas & mihalas 1984 ) . in the strong - shock limit ( @xmath154 and @xmath155 ) ,
the ratio @xmath156 approaches a constant value of @xmath157 and @xmath143 grows without bound .
the first term of equation ( [ eq : tds ] ) dominates the second and @xmath158 the weak and strong limiting expressions are valid to within about 25% of the exact result for @xmath159 and @xmath160 , respectively . because the peak mach numbers of the shock trains in the zephyr models shown below are typically outside these ranges ( i.e. , @xmath161 ) we use the full procedure described by equations ( [ eq : tds])([eq : t1t0 ] ) , which is valid for shocks of arbitrary strength . in order to incorporate the shock dissipation into the acoustic wave transport equation ,
the wave amplitude @xmath162 must be converted , where appropriate , into the mach number @xmath163 .
we must take account of how the initially sinusoidal wave profile steepens into a sawtooth shock train .
often shock heating is applied only above an estimated shock formation height .
however , there exists a finite range of heights between the first formation of the shock ( i.e. , where the acoustic wave train first `` breaks '' ) and the height where the shock train has evolved into a complete sawtooth shape . between these heights
the velocity amplitude of the shock may only be a fraction of the crest - to - trough velocity amplitude of the wave . to solve for the wave evolution and dissipation in this region , we have computed numerical profiles of both the shape factor @xmath123 and a steepening efficiency factor @xmath164 .
these quantities are computed as a function of a dimensionless steepening parameter @xmath165 , which measures by how much a wave crest is approaching ( or has overtaken ) the zero - velocity node immediately ahead of it .. ] the efficiency @xmath164 is defined as the ratio of the shock velocity amplitude to the full velocity amplitude of the ( arbitrarily steepened ) wave profile , and we define @xmath166 where the wave is still a sinusoid , @xmath167 and @xmath168 , which results in no heating .
as the wave steepens , @xmath164 grows from 0 to 1 . at each radial grid zone
, we compute the crest - to - node distance factor @xmath169 , where @xmath170 is the local parallel wavelength and @xmath171 where the integration is taken from the lower boundary ( at which @xmath172 ) up to an arbitrary height . by computing this quantity point - by - point along the grid ,
the atmospheric stratification is taken into account accurately . at the lower boundary
, the wave profile is assumed to be a perfect sinusoid , and @xmath173 .
this quantity decreases steadily as the wave train propagates upward and steepens .
when @xmath165 reaches zero , the shock amplitude has grown to be equal to the wave amplitude and the profile is assumed to remain a sawtooth as it propagates upward .
we continue to follow the ever - decreasing @xmath165 to values below zero , though , because the profile only reaches the exact sawtooth `` n - wave '' shape in the limit of @xmath174 .
figure 2 shows the normalized phase function @xmath175 of a gradually steepening acoustic wave , where the phase @xmath176 is defined at a given ( constant ) height as @xmath177 ; it varies from @xmath178 to @xmath179 .
figure 2 also plots the efficiency @xmath164 and the shape factor @xmath123 as a function of @xmath165 .
these have been derived from numerical simulations of this steepening sinusoidal wave profile , and we give parameterized fits ( used by the zephyr code ) below .
the fit to the efficiency is given by @xmath180 the constant @xmath181 denotes the critical breakpoint at which the wave train first steepens to infinite slope at the node ahead of the crest ; it is defined as @xmath182 .
note that as @xmath165 decreases from 0.25 to @xmath183 , the efficiency @xmath164 remains zero because no shock transition has yet formed .
only when @xmath165 decreases below @xmath183 does there exist a shock with a finite strength .
the shape factor @xmath123 is used to convert between wave amplitude and energy density ( eq . [ [ eq : usdef ] ] ) , and it is defined as the inverse of the phase - averaged square of the normalized phase function , i.e. , @xmath184^{2}\ ] ] ( see also koninx 1992 ; suzuki 2004 ) . as an example , for a sinusoid the average of @xmath185 over a full period is 1/2 , and thus @xmath186 .
we found the following fit from a series of simulated wave profiles undergoing gradual steepening , @xmath187 ( see figure 2 for a comparison between the numerically determined values and the fit ) . for most of the pre - break steepening ( i.e. , from @xmath188 down to @xmath183 )
@xmath123 remains close to 2 .
note , though , that when the shock grows to full amplitude at @xmath189 the shape factor has increased only to about 2.4 .
the additional increase up to 3 occurs for negative values of @xmath165 .
the convection zone generates a continuous spectrum of acoustic power , and we need to specify this distribution of energy as a function of frequency at the photospheric base of the model .
the power spectrum at larger heights is determined implicitly as a result of solving the monochromatic wave action conservation equations ( eq . [
[ eq : j77ac ] ] ) for a range of frequencies . the continuous power spectrum @xmath190 is defined at the base in units of the photospheric acoustic flux , with @xmath191 for waves that escape into the upper atmosphere
, there should be negligible power at frequencies below the acoustic cutoff , @xmath192 because waves having @xmath193 are evanescent in an ideal hydrostatic atmosphere ( e.g. , mihalas & mihalas 1984 ; see , however , wang et al . 1995 ;
schmitz & fleck 1998 ) .
we used a constant value of @xmath194 rad s@xmath57 ( i.e. , @xmath195 mhz , or a period of 3.5 minutes ) , which corresponds to temperature - minimum conditions in the upper photosphere . from a theoretical standpoint , this value is slightly on the high side , since the photospheric @xmath196 of 5800 k gives a cutoff frequency about 13% lower . our higher value for @xmath197 is an attempt to ensure that the modeled frequencies are those that should be propagating _ everywhere _ in the model i.e .
, since we do not treat evanescent energy loss , we want to include only waves that can make it through both the photosphere and the temperature minimum region without becoming evanescent . note , though , that from the standpoint of observational helioseismology , our value is slightly on the low side ( e.g. , jimnez 2006 ) , so it seems to be a satisfactory median value .
the shape of the power spectrum above the cutoff is a subject of some ongoing controversy , which we do not attempt to address fully .
the presence of substantial power at high frequencies ( @xmath198 mhz ) is predicted by traditional theories of sound generation from turbulent convection ( e.g. , lighthill 1952 ; stein 1967 ; musielak et al .
1994 ; ulmschneider et al .
1996 ) and also by observational inferences of time - steady chromospheric heating ( kalkofen et al . 1999 ; ulmschneider et al . 2005 ; cuntz et al .
however , evidence also exists that there may be an extremely steep decline in the acoustic power spectrum before frequencies of order 20 mhz are reached , and thus that high frequencies would not be important to atmospheric heating ( e.g. , judge et al .
2003 ; fossum & carlsson 2005 , 2006 ) .
recent advances in detecting high - frequency acoustic fluctuations have been made by wunnenberg et al .
( 2002 ) , deforest ( 2004 ) , muglach ( 2006 ) , and van noort & rouppe van der voort ( 2006 ) , but no firm conclusions yet exist regarding their impact on chromospheric heating . future observations with higher spatial and temporal resolution are definitely needed . provisionally , we model @xmath199 with a high - frequency tail reminiscent of the turbulent convection theories cited above .
the following parameterization @xmath200 & \omega \geq \omega_{\rm ac } \\ 0 & \omega
< \omega_{\rm ac } \end{array } \right .
\label{eq : pomegas}\ ] ] has finite power at the cutoff , a peak value at @xmath201 , and a declining tail with @xmath202 at high frequencies .
typically , @xmath203 and @xmath204 is a factor of 2 to 5 larger than @xmath197 .
the normalization is given by specifying a known value of @xmath127 and using equation ( [ eq : pints ] ) .
figure 3 illustrates the shape of the acoustic power spectrum that is used in all of the solar models described in this paper .
we use constant values of @xmath205 and @xmath206 .
the adopted value of @xmath207 is the most sensitive to the controversy over high - frequency acoustic waves , and it deserves further discussion . the standard lagrangian treatment of the lighthill - stein sound generation mechanism tends to give values of @xmath207 between 3 and 3.5 in the high - frequency limit for both acoustic and longitudinal flux - tube waves ( ulmschneider et al .
1996 ; musielak et al .
2000 ) . however , an alternate eulerian treatment of the turbulent correlations ( rubinstein & zhou 2002 ) yielded a much shallower decline with increasing frequency ; @xmath2082.4 .
we also note that the value @xmath209 is a special case that corresponds to frequency - independent shock steepening for the discrete bins that we use ( see below ) ; i.e. , when @xmath209 the increase in frequency from one bin to the next is balanced exactly by the power decrease , such that the wave train in each bin steepens into a shock train at the same height .
we find that a realistic chromospheric temperature rise occurs only when higher frequency waves steepen at successively lower heights ( which requires @xmath210 ) .
our value of 2.5 satisfies this requirement while being only slightly lower than the range predicted by traditional lighthill - stein theory . in the zephyr code ,
the continuous spectrum @xmath199 is modeled as a series of discrete frequency bins , each of which is treated independently as described in 4.1 .
we define the bins as `` octaves '' in frequency space ; i.e. , the first bin encompasses @xmath197 to @xmath211 , the second encompasses @xmath211 to @xmath212 , and so on . ideally , it would have been preferable to use even narrower bins in order to more accurately represent the shape of the spectrum . however ,
if the frequency bins were too narrow , each would contain a vanishingly small amount of wave energy .
the calculation of nonlinear steepening ( eq . [ [ eq : delz ] ] ) depends on the total amplitude that remains reasonably coherent at a given frequency .
narrow bins treated independently as described above would essentially destroy this coherence and produce an unphysical delay in the onset of steepening due to the low amplitudes in each bin .
the use of octaves is an attempt to balance the needs of frequency resolution and realistic coherence for steepening . in the models computed for this paper ,
a series of 14 bins is used .
the power missed by ignoring frequencies above the maximum value of @xmath213 is only @xmath214 of the total flux .
the monochromatic frequencies used in the wave action conservation equation are computed as mean , or first - moment frequencies over the power spectrum of each bin .
the first one ( @xmath197 to @xmath211 ) , for example , has a mean of @xmath215 ( i.e. , @xmath216 mhz ) .
we model transverse mhd fluctuations in the radially oriented magnetic flux tube as ideal incompressible alfvn waves .
the equations and assumptions we use generally follow cranmer & van ballegooijen ( 2005 ) , who computed various properties of alfvn waves from the photosphere to the interplanetary medium assuming a known background plasma state ( @xmath5 , @xmath9 , @xmath18 ) . here
we compute the wave properties and background plasma properties together in a self - consistent manner .
the alfvn waves are believed to arise from transverse jostling of magnetic flux tubes in the intergranular photosphere by convective motions on granular ( @xmath2171000 km ) horizontal scales .
note that , similar to the acoustic waves , the alfvn waves are expected to have modified dispersive properties in the photosphere and low chromosphere ( e.g. , they become _ kink - mode _ tube waves in regions where the flux tubes are surrounded by field - free regions ; see spruit 1981 , 1982 ) , but we ignore these effects because they occur over a small range of heights .
our treatment of alfvn wave reflection and turbulent dissipation is complementary to the recent work of verdini et al .
( 2005 , 2006 ) . as in the previous section ,
we first discuss the overall wave energy transport and dissipation mechanisms ( 5.1 ) and then describe the power spectrum of the fluctuations ( 5.2 ) .
the way that the frequency dependence is assimilated into the model is slightly different from the acoustic - wave case , though . for transverse incompressible waves , we solve a single wave action conservation equation with a damping term that contains information about the power spectrum and wave reflection .
we believe that a monochromatic treatment of the wave action conservation ( i.e. , to treat each frequency bin independently from the rest ) is inappropriate to use when the dominant dissipation mechanism is mhd turbulence .
a turbulent cascade inherently contains strong nonlocal `` mixing '' in frequency and wavenumber space .
thus the turbulent damping rate is computed only using quantities that have been integrated over the power spectrum ( see below ) .
we assume the alfvn waves are dispersionless with a phase speed given by @xmath218 in the stationary reference frame of the sun .
the alfvn speed is defined as @xmath219 .
although we solve the full non - wkb transport equation to obtain the _ relative _ distribution of outward and inward propagating waves , we use the following simplified version of the wave action conservation equation to compute the radial evolution of the total ( frequency - integrated ) wave energy density @xmath220 : @xmath221 \ , = \ , - \frac{q_a}{\omega ' } \label{eq : j77alf}\ ] ] ( see , e.g. , jacques 1977 ; isenberg & hollweg 1982 ; tu & marsch 1995 ) .
the presence of the doppler shifted frequency @xmath222 ( measured in the solar wind frame ) in the above equation is deceptive , since the wave action conservation is essentially _ frequency - independent .
_ note that @xmath223 , and thus a constant factor of @xmath115 can be factored out of each term in equation ( [ eq : j77alf ] ) .
the total alfvnic energy density @xmath220 is the sum of the energy densities in kinetic and magnetic fluctuations , @xmath224 where @xmath225 and @xmath226 are the transverse velocity and magnetic field oscillation amplitudes .
we assume equipartition between the two components ( @xmath227 ) , such that @xmath228 .
the use of equation ( [ eq : j77alf ] ) contains the assumption that the energy balance is dominated by outward - propagating waves , whose energy density @xmath229 exceeds the energy density of inward - propagating waves @xmath230 . by definition , @xmath231 , and @xmath232 where
the elsasser ( 1950 ) amplitudes are defined here as @xmath233 . for future reference
, we give the expression for the net outward alfvn wave flux , @xmath234 ( e.g. , heinemann & olbert 1980 ) .
the two assumptions of kinetic / magnetic equipartition and outward - wave dominance in equation ( [ eq : j77alf ] ) tend to break down in the lower atmosphere when there is strong non - wkb wave reflection , but in the _ corona_where the dominant alfvnic heating occurs
these assumptions seem to be appropriate ( see also cranmer & van ballegooijen 2005 ) .
the only physical source of alfvn wave damping that we include is the turbulent cascade , which ultimately must terminate in an irreversible conversion of wave energy into heat .
this is , in some sense , a controlled experiment to evaluate to what degree turbulence may be the _
dominant _ cause of coronal heating and solar wind acceleration , but other damping mechanisms for alfvn waves have been suggested .
collisional damping mechanisms for mhd waves have been studied extensively in the context of the dense coronal base ( e.g. , alfvn 1947 ; kuperus et al .
1981 ; narain & ulmschneider 1990 , 1996 ; porter et al . 1994 ; roberts 2000 ) ,
but the open field lines that feed the solar wind have lower densities and are thus less collisionally dominated than closed loops in the low corona .
cranmer & van ballegooijen ( 2005 ) investigated linear viscous dissipation of alfvn waves along an open magnetic flux tube and found that viscosity should be negligible for waves having periods longer than about 1 minute .
we adopted the following phenomenological form for the mhd turbulence damping rate , @xmath235 ( hossain et al .
1995 ; zhou & matthaeus 1990 ; matthaeus et al . 1999 ; dmitruk et al .
2001 , 2002 ) .
the transverse length scale @xmath236 represents an effective perpendicular correlation length of the turbulence for the largest `` driving '' eddies .
we used the standard assumption that @xmath237 scales with the transverse width of the open flux tube ; i.e. , that it remains proportional to @xmath238 ( see also hollweg 1986 ) .
ideally , the evolution of @xmath237 should be coupled to the radial variation of the fluctuation energy ( see , e.g. , eq .
[ 3 ] of matthaeus et al .
1999 ) as well as to the non - varying field strength .
future work should include both this effect and additional tests of whether the above phenomenological form is an adequate representation of the true anisotropic cascade .
the normalization of @xmath237 is one of the few free parameters of our model .
we note that cranmer & van ballegooijen ( 2005 ) found that @xmath237 should be about 1100 km at the height in the low chromosphere where thin flux tubes merge with one another . mapping this value down to the photosphere
would yield a lower boundary value @xmath239 km .
this spatial scale is intermediate between the probable horizontal size an individual flux tube in the photosphere ( 50100 km ) and the size of a convective granule ( 1000 km ) .
this value is also similar to the width of an intergranular lane and also the mean separation between photospheric flux tubes in the quiet - sun supergranular network ( 350700 km ; see cranmer & van ballegooijen 2005 ) .
the factor @xmath240 in equation ( [ eq : dmit ] ) is a turbulent efficiency that accounts for regions where the turbulence may not have time to develop before the waves or the wind carry away the energy ( see dmitruk & matthaeus 2003 ) .
we estimated this efficiency factor to be @xmath241 where the two time scales in this expression are @xmath242 , a nonlinear outer - scale eddy cascade time , and @xmath243 , a timescale for macroscopic alfvn wave reflection .
in most of the models presented below we take @xmath244 , based on analytic and numerical models of dobrowolny et al .
( 1980 ) , matthaeus & zhou ( 1989 ) , and oughton et al .
dmitruk & matthaeus ( 2003 ) found that the turbulent cascade has sufficient time to develop and heat the plasma only when @xmath245 .
thus , our efficiency factor above quenches the turbulent heating when @xmath246 , i.e. , when the alfvn waves want to propagate away much faster than the cascade can proceed `` locally . ''
the reflection time is defined simply as @xmath247 , and the eddy cascade time is given by @xmath248 where the alfvnic mach number @xmath249 and the numerical factor of @xmath250 comes from the normalization of an assumed shape of the turbulence spectrum ( see appendix c of cranmer & van ballegooijen 2005 ; see also higdon 1984 ; shebalin et al . 1983 ; goldreich & sridhar 1995 , 1997 ; bhattacharjee & ng 2001 ; cho et al .
when @xmath47 , the efficiency factor provides an approximate bridging between a kolmogorov ( 1941 ) scaling , when @xmath245 , and an ik - like ( iroshnikov 1963 ; kraichnan 1965 ) scaling , when @xmath246 .
there is still some controversy , though , over which type of cascade rate is appropriate for mhd turbulence in the solar wind .
we separated the alfvn wave energy into outward ( @xmath251 ) and inward ( @xmath252 ) components by solving a modified form of the non - wkb transport equations of heinemann & olbert ( 1980 ) , barkhudarov ( 1991 ) , velli ( 1993 ) , and orlando et al .
eigenfunctions are computed numerically without the use of a wavenumber .
] these frequency - dependent equations are discussed below in 5.2 and their solution provides a spectrum - averaged value of the effective local reflection coefficient @xmath253 .
knowing @xmath254 and @xmath220 ( from eq .
[ [ eq : j77alf ] ] ) allows the elsasser amplitudes to be computed at all heights , @xmath255 and the turbulent damping rate in equation ( [ eq : dmit ] ) is then specified . for a series of alfvn wave frequencies @xmath115
, we solved the non - wkb wave transport equations in the dimensionless form given by equations ( 24 ) and ( 33 ) of barkhudarov ( 1991 ) , @xmath256 @xmath257 where @xmath258 is related to the frequency - dependent reflection coefficient @xmath259 via @xmath260 @xmath261 is the angular phase shift between the @xmath251 and @xmath252 wave trains , and @xmath262 is the ( signed ) scale height for the alfvn speed , or @xmath263 .
although the models of barkhudarov ( 1991 ) were limited to spherical geometry ( @xmath264 ) the above relations are valid for any @xmath265 ( see also cranmer & van ballegooijen 2005 ) .
we followed the general solution procedure outlined by barkhudarov ( 1991 ) for integrating across the alfvnic singular point @xmath266 , where @xmath267 and thus @xmath268 .
the reflection coefficient remains finite at this point , and it can be written exactly as @xmath269 where @xmath270 and @xmath271 are radial derivatives of the outflow speed and alfvn speed taken at @xmath272 .
the zephyr code utilizes fourth - order runge - kutta numerical integration to solve for @xmath273 above and below @xmath266 .
we ensured that @xmath274 at all heights and for all frequencies .
the dissipation of alfvn waves is explicitly not included in the equations given above for the reflection coefficient .
although it is possible to include nonlinear damping consistent with equation ( [ eq : dmit ] ) in these transport equations ( see , e.g. , verdini et al .
2005 , 2006 ) , we remain cautious about the combination of strong turbulent damping and `` monochromatic '' wave quantities .
additional simulations may be required in order to better guide the use of phenomenological nonlinear terms when following the development of a spectrum of alfvnic fluctuations .
the full frequency - averaged reflection coefficient is computed by weighting @xmath259 by the alfvn wave power spectrum @xmath275 , with @xmath276 where the square of @xmath254 is used because the power spectrum is an energy density quantity and @xmath254 is a ratio of amplitudes . the spectrum is used essentially as a weighting function , and we assume it has a constant shape as a function of height ( see , e.g. , figure 8 of cranmer & van ballegooijen 2005 ) .
figure 3 shows the shape of the alfvn wave power spectrum that we use in the zephyr models discussed below . specifically , this is the power spectrum of total ( kinetic plus magnetic ) energy from the model of cranmer & van ballegooijen ( 2005 ) taken at the height of the transition region .
it was computed from an empirically constrained _ photospheric _ spectrum of the dynamics of thin flux - tubes , which is described as a linear superposition of two types of motion .
first , isolated flux tubes undertake `` random walks '' in response to convective granulation ; they have a power spectrum that has been constrained by observed g - band bright point motions ( e.g. , van ballegooijen et al .
1998 ; nisenson et al . 2003 ) .
second , flux tubes exhibit sporadic rapid horizontal `` jumps '' that probably represent merging , fragmenting , or reconnecting with surrounding magnetic fields ( e.g. , choudhuri et al .
1993 ; berger & title 1996 ; berger et al . 1998 ; hasan et al .
2000 ) which we modeled as a series of periodic impulsive motions .
high - resolution photospheric observations thus provided the kinetic energy spectrum , and the partitioning between kinetic and magnetic energy components ( needed to compute the total energy spectrum ) was determined using the analytic linear theory of kink - mode waves in a stratified atmosphere .
finally , propagation effects between the photosphere and the transition region ( @xmath277 in the model of cranmer & van ballegooijen 2005 ) were taken into account by solving the combined non - wkb kink - mode and alfvn - mode transport equations for an empirically constrained ( fal - c@xmath43 ) background plasma state .
there are several noteworthy features of the spectrum @xmath275 in figure 3 .
there is power at the lowest frequencies , corresponding to periods of hours to days , because of the assumed random - walk component of the photospheric flux tube motion ( with @xmath278 , where @xmath279 s ) .
however , the relative amount of power at periods longer than 1 hour is negligible compared to the total ; this stands in contrast with in situ measurements that show the majority of power to be at these long periods ( e.g. , goldstein et al .
1995 ; tu & marsch 1995 ) .
there is growing evidence that the low - frequency fluctuations seen in interplanetary space may be the result of the passage of multiple uncorrelated flux tubes past the spacecraft and not intrinsic turbulence within any one flux tube ( see , e.g. , mccracken & ness 1966 ; jokipii & parker 1969 ; bruno et al .
2001 ; giacalone & jokipii 2004 ; giacalone et al .
2006 ; borovsky 2006 ) . the damping of evanescent kink - mode waves can be seen in figure 3 from the mild discontinuity at the cutoff frequency of 1.4 mhz . only about 40% of the energy of evanescent waves
is lost , though , because the flux tubes remain thin ( and thus gravitationally buoyant ) only below a low - chromosphere `` merging height '' of 600 km . above this height
the magnetic fields expand horizontally to fill the volume above supergranular network lanes and the transverse waves both below and above the cutoff can propagate freely as alfvn waves .
the multiple peaks in the spectrum between about 3 and 20 mhz are a result of propagation effects between the photosphere and the merging height . in the model of cranmer & van ballegooijen ( 2005 )
the atmosphere was modeled with the nonisothermal fal - c@xmath43 temperature structure and a radially varying filling factor for the flux tubes ; both of these factors resulted in a complicated frequency dependence for the non - wkb transmission of kink - mode waves between @xmath280 and 600 km . coincidentally , the frequencies of the maxima in both @xmath281 and @xmath282 are each about a factor three higher than their respective cutoff frequencies .
it is important to note that the alfvn wave amplitude @xmath225 ( obtained from the wkb - like eq .
[ [ eq : j77alf ] ] ) diverges from the actual transverse velocity of oscillating magnetic flux tubes in the lower atmosphere .
the effects of evanescence and non - wkb wave reflection are not _ directly _ included in equation ( [ eq : j77alf ] ) , but they end up being of minimal importance in the corona and solar wind .
we thus can estimate the true velocity amplitude @xmath283 after the final iterated solar atmosphere parameters have been determined , as @xmath284 , & \omega < \omega_{\rm kc } \\
\end{array } \right . \
, , \label{eq : wperp}\ ] ] where @xmath285 is the kink - mode cutoff frequency . above the mid - chromosphere height where thin flux tubes merge with one another
, @xmath285 effectively goes to zero and there is no evanescence .
the @xmath259 factor above in parentheses corrects for the approximation that equation ( [ eq : j77alf ] ) follows only outward - going waves . for the zephyr models presented below
, the spectrum - averaged value of @xmath283 at the photosphere is typically a factor of 2 to 5 times larger than @xmath225 at this lower boundary . above the transition region , though , @xmath286 ( see also figure 11 of cranmer & van ballegooijen 2005 ) .
despite the fact that @xmath225 underestimates the actual velocity amplitude , we believe it is more appropriate to use this as an imposed lower boundary condition rather than to use @xmath283 .
the latter quantity depends on quantities that are computed self - consistently along with the other plasma parameters and are not known a priori .
just as electromagnetic waves carry momentum and exert pressure on matter , propagating acoustic and mhd waves can also do work on the mean fluid via a similar kind of radiation stress .
( bretherton & garrett 1968 ; dewar 1970 ; belcher 1971 ; alazraki and couturier 1971 ) . for parallel - propagating acoustic and alfvn waves ,
the time - averaged radial wave pressure acceleration was derived in detail by jacques ( 1977 ) to be @xmath287 as above , we ignore departures from kinetic - magnetic energy equipartition for the alfvn waves ; this ends up being a good approximation for solar wind conditions ( see , e.g. , heinemann & olbert 1980 ; cranmer & van ballegooijen 2005 ) . to implement the above expression in the zephyr code
, we used the wave action conservation equations ( [ eq : j77ac ] ) and ( [ eq : j77alf ] ) to reformulate the momentum conservation equation into a time - independent parker ( 1958 , 1963 ) critical point equation with additional terms that affect the definition of the critical point . assuming @xmath30 throughout
, the modified momentum equation becomes @xmath288 @xmath289 \label{eq : critmod}\ ] ] where @xmath290 is the isothermal sound speed @xmath291 and the modified critical speed is given by @xmath292 with the bulk - flow mach numbers @xmath249 and @xmath293 . in the presence of acoustic waves ,
the additional terms above are given by @xmath294 @xmath295 the modified critical radius @xmath296 is found by locating the point where the right - hand side of equation ( [ eq : critmod ] ) is zero , and thus @xmath297 .
for rapidly expanding superradial flux tubes , kopp & holzer ( 1976 ) pointed out that there may be more than one possible location for the critical point . defining the right - hand side of equation ( [ eq : critmod ] ) as the radial derivative of a known function @xmath298
, they essentially found that the _ global minimum _ of @xmath298 specifies the critical point location that allows for a stable and continuous solution for @xmath299 from the sun to interplanetary space ( see also vsquez et al .
2003 ; cranmer 2005a ) .
the kopp & holzer ( 1976 ) result , though , was derived for an ideally polytropic corona without wave pressure acceleration .
it is not clear if this global minimum condition is applicable to the models described in this paper .
for all of the zephyr models discussed below in 8 , though , the global minimum of @xmath298 coincides with the _ largest _ possible value of @xmath296 . for expediency , then , when there are multiple possibilities we choose the largest value for the critical radius that also exhibits a minimum in @xmath300 ( i.e. , @xmath301 ) .
note that equation ( [ eq : critmod ] ) includes the wave pressure acceleration , but not the explicit derivatives of @xmath119 and @xmath220 that appear in equation ( [ eq : ddef ] ) .
the modified momentum equation thus alleviates the need to perform noisy numerical differentiation ( see also jacques 1977 ; hartmann & macgregor 1980 ; decampli 1981 ; holzer et al .
1983 ; wang & sheeley 1991 ) .
the damping rates @xmath131 and @xmath302 appear explicitly in the momentum equation , which provides additional nonlinear coupling between the momentum and energy equations . wave pressure acceleration has been invoked by laming ( 2004 ) as a potential explanation for the relative enhancement of low fip ( first ionization potential ) elements relative to high fip elements in the corona and solar wind .
if alfvn waves exert an appreciable force on ions in the chromosphere and transition region ( where there are still neutrals that do not feel this force ) significant `` fractionation '' may occur .
we apply laming s ( 2004 ) idea to the models presented below by evaluating his integrated momentum equation for atoms and ions of element @xmath123 undergoing fractionation .
a slightly simplified version of this equation ( which assumes that there is no explicit dependence of @xmath23 on the density gradient ) is @xmath303 \label{eq : lamfip}\ ] ] where @xmath304 and @xmath305 are the element s mass density and effective parallel turbulent speed .
the ionization fraction of the element essentially one minus the neutral fraction is given by @xmath306 , and the collision rates between ions and neutrals of element @xmath123 and the ambient gas are given by @xmath307 and @xmath308 ( see laming 2004 for detailed expressions ) .
subscripts @xmath309 and @xmath5 denote quantities computed at heights @xmath310 ( a lower boundary we take as the photosphere ) and @xmath311 ( an upper boundary that can be anywhere in the corona or solar wind ) , respectively .
the degree of fip enhancement between these heights is obtained by solving equation ( [ eq : lamfip ] ) twice : once for the ratio @xmath312 for a low fip element , and once for a high fip element .
dividing one ratio by the other cancels out the overall density stratification between @xmath310 and @xmath311 and leaves only the fractionated abundance difference .
we compute time - independent solutions to the conservation equations given above by applying a new hybrid method of iteration and relaxation .
reasons for not following time - dependent fluctuations explicitly were summarized in
it has become common , though , to use time - dependent hydrodynamics codes to find stable time - steady solar wind solutions ( see recent work by , e.g. , li et al .
2004 ; lionello et al . 2005 ; lie - svendsen & esser 2005 , and references therein ) .
we break from this tradition for several reasons .
first , in the one - fluid case , the time - independent mass and momentum conservation equations are relatively easy to solve , and only the energy equation requires special treatment .
second , when modeling the photosphere , chromosphere , corona , and wind all in one grid , the energy equation changes its basic character at different locations depending on which terms are dominant . in the corona ,
the spitzer - hrm conduction makes it a second - order parabolic differential equation . in the outer solar wind the gradual transition to collisionless heat conduction reduces it to a first - order differential equation . in the photosphere and
chromosphere it is essentially a zeroth - order differential equation i.e .
, an algebraic balance between heating and cooling .
these changes , combined with the huge dynamic range in quantities such as density , temperature , and wave phase speeds , make it difficult to implement and optimize a _
robust _ time - dependent numerical scheme .
even implicit methods , which are not necessarily limited by propagation across the smallest grid zones , appear to be prohibitively difficult to set up properly when the source terms ( i.e. , the waves ) have such a complex nonlinear dependence on the primary plasma parameters .
the code developed for this work is called zephyr .
after setting up an initial trial guess for the plasma parameters @xmath9 , @xmath5 , and @xmath6 , the code iterates a fixed number of times ( typically 100 to 200 ) alternately between solutions of the energy equation ( eq . [ [ eq : dedt ] ] ) and the other constitutive and conservation equations ( eqs . [ [ eq : masscon ] ] , [ [ eq : momcon ] ] , [ [ eq : pdef ] ] , [ [ eq : edef ] ] , [ [ eq : ntot ] ] , [ [ eq : j77ac ] ] , [ [ eq : j77alf ] ] , [ [ eq : dpsi ] ] , [ [ eq : dgam ] ] ) to find a steady - state accelerating wind solution . to stabilize the iteration process and avoid pathological solutions , most of the plasma parameters are `` undercorrected '' using a scheme described below
, rather than replacing the old solution with the new one , a fractional step _ toward _ the new solution is taken .
the spatial grid used by zephyr has variable zone spacing that depends on the height above the lower boundary .
the total number of grid zones @xmath313 is divided into two subsets : 45% of the zones are allocated to a fine mesh with constant spacing in the lower atmosphere ( i.e. , between @xmath280 and a fixed zone - midpoint height @xmath314 ) and 55% of the zones occupy the rest of the grid with spacing that increases by 1% per zone from the value at @xmath315 up to the top of the grid . all models described in this paper have @xmath316 and a fixed grid spacing of @xmath317 km in the lower atmosphere . at the top of the grid
( @xmath318 au ) the grid spacing has increased to 24 @xmath92 .
the relative spacing @xmath319 thus increases from about @xmath320 in the lower atmosphere up to 0.02 at the top of the grid .
the initial condition for the temperature is a gray radiative equilibrium atmosphere ( i.e. , @xmath321 as defined in
3.2 ) that is perturbed by transitioning ( using a hyperbolic tangent function ) to a corona / wind value above a specified transition region height of @xmath322 .
the temperature above this transition obeys a power law in @xmath10 , with a basal value of @xmath323 k and a very slow radial decline proportional to @xmath324 . the lower boundary ( @xmath280 ) is defined as @xmath325 , which corresponds to a base density @xmath326 g @xmath327 .
this choice for the lower boundary condition is used in order to include the traditional photosphere ( either @xmath328 or 2/3 ) in the interior of the grid .
the temperature at the lower boundary is defined by the gray atmosphere condition ( eq . [ [ eq : jtau ] ] ) .
the initial density distribution is assumed to be hydrostatic ( i.e. , @xmath329 ) .
refined initial guesses for the density and outflow speed are evaluated as the first steps in the iteration process described below .
tests have shown that finding the proper solution is insensitive to the details of the initial guess , but the iteration method converges faster when the initial guess has properties closer to a realistic solar atmosphere .
the main `` outer '' iteration loop consists of two interior modules , each of which undergoes a number of `` inner '' iterations .
the first module solves for the dynamics ( @xmath9 , @xmath5 ) and for the various source terms in the momentum and energy equations .
the five steps taken in one inner iteration of this module are as follows . 1 . the hydrogen ionization fraction @xmath44 , the gas pressure @xmath20 , and the internal energy density @xmath25 are computed as described in 3.1 .
the optical depth scale @xmath50 is integrated and the radiative cooling / heating rate @xmath330 is computed ( 3.2 ) . the heat conduction rate @xmath331 ( 3.3 ) is also determined using the four - point finite differencing scheme discussed below for the radial derivatives @xmath332 and @xmath333 . the heat conduction @xmath331 is artificially suppressed in the outermost 10 grid zones in order to produce a more well - behaved upper boundary condition that is dominated by the outward advection of all characteristics .
( these 10 zones are not considered to be part of the actual solar wind solution . ) 2 .
the modified parker critical point equation ( [ eq : critmod ] ) is solved for @xmath299 by integrating up and down from @xmath296 . to step from the critical point ( which generally falls between grid zones ) to the grid zones immediately above and below , an analytic derivative @xmath334 computed from lhpital s rule
is used in order to avoid the well - known instability at an x - type singular point .
at all other grid zones , fourth - order runge - kutta integration is used ( e.g. , press et al .
the undercorrection scheme described below is used for @xmath299 .
the time - steady mass conservation equation ( [ eq : masscon ] ) is solved for @xmath335 in a straightforward way .
the base density @xmath336 described above is kept fixed and the prior step s solution for the basal outflow speed @xmath337 is used to determine the constant mass flux @xmath338 .
the density is then computed exactly at each grid zone , but the undercorrection scheme is also used for @xmath335 in order to prevent too rapid a change .
the properties of the acoustic wave spectrum are computed at each grid point and the spectrum - integrated values of @xmath119 and @xmath131 are determined ( 4 ) .
the steepening of each monochromatic wave train is computed by integrating equation ( [ eq : delz ] ) up from the lower boundary simultaneously with the wave action equation ( [ eq : j77ac ] ) .
simple first - order euler steps are used to perform both integrations .
the non - wkb alfvn wave transport equations ( 5.2 ) are solved for a number of frequency dependent reflection coefficients @xmath273 which are summed over the power spectrum to obtain @xmath339 .
the adaptive fourth - order runge - kutta scheme developed by cranmer & van ballegooijen ( 2005 ) has been included in the zephyr code . once the @xmath340 elsasser variables have been determined , the heating rate @xmath302 is computed and the alfvn wave action conservation equation is integrated up from the lower boundary to obtain @xmath220 ( 5.1 ) .
these steps are repeated for a fixed number of inner iterations ( typically 50 to 100 ) in order to reach internal consistency . however , because the non - wkb wave reflection equations ( [ eq : dpsi][eq : dgam ] ) dominate the computation time , the reflection coefficient @xmath254 is recomputed only for the first two of these iterations .
the second main module of zephyr solves the energy conservation equation and obtains a time - independent temperature distribution @xmath108 .
there are three main steps that are repeated until either a certain degree of convergence has been attained ( @xmath341 ) or a maximum number of iterations ( typically 1000 ) is exceeded .
the convergence quantity @xmath342 is defined to be an average over the relative convergence at each grid zone . for the radial grid with @xmath313 discrete zones , @xmath343 where @xmath344 denotes quantities computed at each grid zone .
the numerator is obtained by solving equation ( [ eq : dedt ] ) for the explicit time dependent term @xmath345 at each grid zone and taking the absolute magnitude .
this residual - like quantity is nonzero for solutions that have not yet converged to a steady state .
the denominator is the absolute magnitude of the largest single term in equation ( [ eq : dedt ] ) ; this includes the two advection terms on the left - hand side , and it also includes the separation of the optically thick @xmath63 and @xmath60 terms in @xmath69 ( eq . [ [ eq : qradthick ] ] ) which balance one another in the photosphere .
the solution that satisfies exact time - independence ( for the energy equation ) would have @xmath346 .
the following three steps repeated by the second module of zephyr are designed to hone in on this solution . 1 .
the core procedure in solving the energy equation is _ relaxation _ using @xmath347 at each grid zone .
the sign of @xmath347 is used to determine whether the current solution for @xmath6 should be increased or decreased , and the magnitude of the change is computed from a positive - definite correction factor @xmath348 .
is _ not _ used is the main reason that this technique is called `` relaxation '' and is not really a variety of time - dependent hydrodynamic evolution . ]
this factor is initialized to a constant value of 0.16 at the start of the inner iteration loop . during each relaxation step it
is either kept constant , if @xmath347 has kept the same sign from the last step to the current step , or it is reduced in magnitude by 7% , if @xmath347 has switched signs from the last step to the current step .
the reduction in @xmath349 that occurs when @xmath347 oscillates in sign is a kind of `` annealing '' that allows the relaxation method to focus in on the time - independent solution ( i.e. , the solution with @xmath350 everywhere ) . the updated temperature at each grid zone is thus determined by multiplying the old value by a factor of @xmath351 .
the optimized numerical values given above were determined from a large number of tests with a range of values .
2 . because the relaxation method above can lead to discontinuous jumps in @xmath108
, we perform trial piecewise smoothing in order to reduce unphysical fluctuations .
the radial grid with @xmath352 is broken up into 130 pieces each having 10 zones .
( at the start of each iteration , the offset point that defines the start of the first 10-zone piece is shifted by one grid zone . )
if a piece contains more than one change of sign in the slope @xmath332 , it is smoothed using a gaussian filter ( with a two zone half - width ) until only 0 or 1 changes of sign remain .
the piecewise nature of this smoothing is necessary so that the entire grid does not `` suffer '' when just a small part of it contains numerical noise .
the convergence parameter @xmath342 for the current inner iteration is compared to the best solution ( i.e. , the lowest value of @xmath342 ) that has been found during this outer iteration loop .
if the solution has improved , the best solution is updated .
if the solution has gotten worse , we discard it and revert to the saved best case .
note , though , that this comparison is not performed at every inner iteration step .
( doing it every time could lead to an infinite loop with no changes ever made to @xmath6 . )
tests showed that it is best to allow the solutions to evolve for a while and only perform this comparison every 15 to 20 iteration steps ( we use 17 ) .
the converged value of @xmath6 that emerges from this module is undercorrected before starting the next outer iteration .
the undercorrection scheme that is used at various points in the zephyr code was motivated by globally convergent backtracking methods for finding roots of nonlinear equations ( e.g. , dennis & schnabel 1983 ) . rather than taking the full suggested iteration step , which may propel the solution away from the desired region of convergence , it is sometimes best to take only a partial step . for a scalar quantity @xmath353 at radial grid zone @xmath344 and iteration step @xmath354 , we specify this partial step as @xmath355 where @xmath356 is the next suggested iteration that was obtained by solving one of the conservation equations . the exponent @xmath357 describes the degree of undercorrection .
when the solutions are nearly converged , the full iteration step should be taken ( @xmath358 ) .
when the solutions are far from convergence , though , we require substantial undercorrection ( @xmath359 ) . to obtain this exponent
, we use @xmath360 where the minimum is taken over the entire radial grid ( thus the absence of @xmath344 subscripts ) so that the worst agreement between the current iteration and the next suggested iteration is highlighted .
because @xmath361 may be larger or smaller than @xmath362 , both the ratio and its reciprocal are used above , and the largest value that the minimum can take is 1 .
tests have shown @xmath363 to be a robust value ; it represents a practical lower limit to @xmath357 that alleviates making infinitesimally small corrections .
when radial derivatives need to be taken in the zephyr code , we use the following four - point finite differencing scheme , @xmath364 for discretized quantities @xmath365 on the radial grid that has unequally spaced heights @xmath366 .
the constant @xmath367 determines the weighting between two pairs of centered differences .
the limit of @xmath368 corresponds to standard two - point finite differencing , which is generally accurate to second order .
further taylor - series expansion , assuming a constant - spacing grid and a reasonably smooth function @xmath369 , would give a result that is accurate to fourth order for @xmath370 .
note , though , that quantities in the zephyr code are tabulated on a grid with variable spacing and they often exhibit numerical noise . in this case , errors may be reduced by essentially averaging between the @xmath371 difference and the @xmath372 difference ( i.e. , using @xmath373 ) .
tests with the zephyr code found that @xmath374 provides the most accurate differentiation and noise reduction , and this value is used in the models presented below . for the four zones at the bottom ( @xmath375 ) and top ( @xmath376 ) of the grid we use a combination of the standard two - point finite difference expression and linear extrapolation to evaluate the derivatives .
figure 4 illustrates the convergence of the outer iteration process for the main polar coronal hole model described in 8.3 .
the displayed plasma parameters reach reasonably steady final values in about 70 outer iteration steps .
the energy convergence parameter @xmath342 decreases to its minimum range of variation ( 0.010.02 ) in about 100 iterations .
note , though , that @xmath342 is averaged over the entire radial grid ; in many parts of the grid the convergence is much better than the average value indicates . below the chromosphere - corona transition region
the convergence is excellent ( i.e. , @xmath342 is always less than @xmath377 ) . in the narrow transition region itself , though , the piecewise smoothing discussed above smears out the temperature distribution so that the ideal `` conduction dominated '' solution can not be sustained exactly .
thus , the average value of the convergence parameter for @xmath378 k is about 0.20.3 . in the corona and solar wind acceleration region above the transition region
the average value of the convergence parameter is about 0.01 , similar to the global average .
if the 30 or so grid zones of the transition region are excluded , the global average @xmath342 is reduced by about a factor of two to @xmath2170.005 .
the full zephyr code comprises approximately 2000 lines of fortran . with the typical iteration parameters given above
, the code runs in 2 to 4 hours of cpu time on various sun microsystems ( ultra and sunfire ) computers .
in this section we present a series of solar atmosphere models computed by the zephyr code .
a series of tests was first performed to evaluate the sensitivity of the dominant plasma properties to various input parameters ( 8.18.2 ) .
we used the grids of test models to determine the most likely input parameters for a detailed model of a flux tube emerging from a polar coronal hole at solar minimum ( 8.3 ) .
then , using the same input parameters and varying only the radial dependence of the magnetic field @xmath379 , we modeled the pole - to - equator variation of solar wind conditions at solar minimum ( 8.4 ) and explored the properties of slow wind streams that are connected to flux tubes emerging from active regions ( 8.5 ) .
in addition to the global magnetic field strength ( our primary `` control knob '' ) there are three key parameters that we varied in the course of exploring the physics of atmospheric heating and solar wind acceleration : 1 . the _ photospheric acoustic flux _
@xmath380 injected at the lower boundary mainly affects the chromospheric heating .
probable values for @xmath380 seem to range between @xmath381 and @xmath382 erg s@xmath57 @xmath383 ( see , e.g. , musielak et al .
1994 ; ulmschneider et al .
1996 , 2001 ; carlsson & stein 1997 ; fawzy et al .
2002 ) . for standard photospheric densities and sound speeds ,
this gives a range for the photospheric acoustic velocity amplitude @xmath384 of about 0.1 to 1 km s@xmath57 .
these values are a bit smaller than traditional `` laminar '' granulation velocities of 1 to 2 km s@xmath57 .
the sources of propagating waves are believed to be concentrated in the dark intergranular lanes and thus are able to extract only a fraction of the total kinetic energy of granulation ( rimmele et al .
1995 ; nesis et al .
1997 , 1999 ; cadavid et al .
2 . the _ photospheric alfvn wave amplitude _
@xmath385 is specified instead of the basal flux @xmath386 or non - wkb velocity amplitude @xmath283 , since the latter quantities depend on the cancellation between upward and downward propagating waves that is determined as a part of the self - consistent solution .
observational determinations of @xmath283 from the footpoint motions of g - band bright points yield mean values around 1 km s@xmath57 , with transient speeds up to 5 km s@xmath57 ( e.g. , berger & title 1996 ) .
larger - scale analyses of the horizontal diffusion of magnetic flux elements give smaller speeds of order 0.1 to 0.3 km s@xmath57 ( schrijver et al .
recall that @xmath225 is likely to be factors of 2 to 5 smaller than @xmath283 , so the range of possible values for @xmath385 may extend from below 0.1 up to 1 or 2 km s@xmath57 .
the _ photospheric alfvn wave correlation length _ @xmath387 sets the scale of the turbulent heating rate @xmath302 ( eq .
[ [ eq : dmit ] ] ) .
once this parameter is set , the value of @xmath237 at all larger heights is determined by the adopted proportionality with @xmath238 .
a practical lower limit for @xmath387 seems to be about 10 km ; i.e. , the spatial scale over which radiative diffusion may inhibit the collapse of strong fields into thin flux tubes ( e.g. , venkatakrishnan 1986 ; snchez almeida 2001 ; cameron & galloway 2005 ) . the upper limit may be of the order of the size of photospheric granules ( @xmath2171000 km ) .
possible intermediate length scales include the photospheric radius of a of thin flux tube ( 50100 km ) , the width of an intergranular lane ( about 300 km ) , and the mean separation between the flux tubes in the supergranular network ( 350700 km ; see cranmer & van ballegooijen 2005 ) .
all other parameters have been fixed with the values given in earlier sections .
the test models discussed in this section all used the polar coronal hole magnetic field model that was derived by cranmer & van ballegooijen ( 2005 ) . in the photosphere , chromosphere , and low corona ( i.e. , from @xmath280 to 12 mm ) , this model was obtained by tracing the radial magnetic field strength from the central axis of a two - dimensional numerical model of the supergranular network .
this model contains thin intergranular flux tubes between the photosphere and a `` merging height '' of 0.6 mm where the flux tubes have expanded laterally to the extent that the surrounding field - free plasma disappears . above this height
the merged network element undergoes further funnel - like horizontal expansion to fill a supergranular canopy .
the field is directed completely vertically by the height of 12 mm . above this height
we applied a slightly modified version of the empirically derived solar - minimum magnetic model of banaszkiewicz et al .
the radial dependence of @xmath16 is shown in figure 5 .
figure 6 shows the result of producing a two - dimensional grid of zephyr models by varying @xmath388 and @xmath389 . the acoustic flux @xmath380 was kept fixed at a median value of @xmath390 erg s@xmath57 @xmath383 ( see above ) .
the other two quantities were varied between the limits of @xmath391 km s@xmath57 and @xmath392 km , with 9 points per quantity spread logarithmically between those limits .
figure 6a displays contours of the `` terminal '' outflow speed at the upper edge of the spatial grid , which we call @xmath393 .
this quantity is slightly larger than the outflow speed at 1 au , but never by more than 5% .
figure 6b shows the spherical mass loss rate , which is defined as @xmath394 with the quantities on the right - hand side evaluated at the top of the grid ( where @xmath264 ) .
figures 6c and 6d show the maximum coronal temperature and the heliocentric radius of the wave - modified critical point ( see 6 ) , respectively .
several general trends are evident in figure 6 .
the mass loss rate is primarily dependent on @xmath388 , which represents the total amount of alfvnic wave energy available to be dissipated in the corona .
the outflow speed , though , seems to depend mainly on @xmath389 ; this parameter tells us _ where _ the wave energy is damped . for large values of @xmath389
the damping occurs over a large range of heights , with increasingly more heating and acceleration taking place above the critical point .
it has been known for some time that the relative heights of the critical point and the dominant energy deposition are key in determining the nature of a pressure - driven wind ( leer & holzer 1980 ; pneuman 1980 ; leer et al . 1982 ) .
heat that is deposited _ above _ the critical point is converted nearly completely into kinetic energy of the wind ( and a higher value of @xmath393 ) .
on the other hand , low values of @xmath389 give a more concentrated heat deposition mostly _ below _ the critical point .
this energy raises the temperature in the subsonic part of the corona , increases its scale height , and provides more downward heat conduction into the upper transition region .
less energy is thus available to accelerate the wind and @xmath393 is lower .
the mass loss rate @xmath395 is generally believed to be set in the transition region by a balance between conduction , radiative losses , and an upward enthalpy flux ( e.g. , hammer 1982 ; leer et al .
1998 ) .
interestingly , for a large range of parameters in the lower - right of figure 6b ( @xmath396 km and @xmath397 km / s ) , a combined power - law fit to the parameter dependence of the mass loss rate yields good agreement with the zephyr models for @xmath398 .
this resembles the classical kolmogorov ( 1941 ) energy flux due to isotropic hydrodynamic turbulence : @xmath399 .
it seems to corroborate the general idea that the mass flux of a thermally driven wind is proportional to the deposited coronal energy flux . the large variation in coronal heating , though , does _ not _ seem to produce a large variation in the maximum coronal temperature .
figure 6c shows that this temperature varies only by about a factor of two over most of the parameter space .
owocki ( 2004 ) summarized how the energy losses due to both conduction and the solar wind act as an effective `` thermostat '' to keep the coronal temperature from varying too widely .
figure 7 illustrates how the coronal heating rate @xmath302 changes as @xmath388 is varied and @xmath389 is kept fixed .
we plot the heating rate per unit mass @xmath400 in order to more easily show which heights receive the most heating on a particle - by - particle basis .
note that larger values of @xmath388 produce larger values of both @xmath400 and @xmath6 and move their local maxima down to lower heights .
the presence of damping , though , leads to differences in how the heating rate depends on @xmath388 in various regions . in the lower corona ( @xmath401 ) , before substantial alfvn - wave damping has had time to occur , the three models show a power - law dependence of @xmath402 that is similar to the mass loss rate dependence of @xmath403 for these models .
in the extended corona , though , the _ peak value _ of @xmath400 ( above @xmath404 ) varies more weakly as @xmath405 .
the models with higher alfvn wave amplitudes have undergone relatively more damping in the extended corona than the models with lower @xmath225 .
the above arguments do not explain all of the features of the contours shown in figure 6 .
for example , although there is a general trend ( in the center and lower - right of figure 6b ) for a decreasing @xmath389 to give a larger value of @xmath395 , the globally largest mass loss rate occurs in the upper - right corner of the plot i.e .
, for the largest values of _ both _ @xmath388 and @xmath389 . in this region of the plot ,
added insight can be found in the `` cold '' wave - driven wind model of holzer et al .
for undamped alfvn waves that dominate the critical velocity ( eq . [ [ eq : ucrit ] ] ) the terminal wind speed and mass loss rate can be computed analytically . isolating the proportionality with @xmath388 in this
model yields @xmath406 where the constant term is typically negligible , thus giving @xmath407 ( see eqs . [ 39 ] and [ 41 ] of holzer et al . 1983 ) .
in the upper - right corners of figures 6a and 6b , these relations come the closest to being satisfied . specifically , for @xmath408 km s@xmath57 and @xmath409 km
, the best power - law fits to the model results are @xmath410 and @xmath411 .
these parameters indeed correspond to a wave - dominated critical velocity ( i.e. , large @xmath388 ) without much damping ( large @xmath389 ) . for the entire grid of values shown in figure 6 ,
a power - law is a poor fit to the overall variations of @xmath395 and @xmath393 as a function of the two input parameters . for completeness , though , we report the best - fitting exponents over the full grid : @xmath412 and @xmath413 .
we produced a second grid of exploratory zephyr models that kept the coronal heating parameters ( @xmath388 , @xmath389 ) fixed and varied only the basal acoustic wave flux @xmath380 . as described further in 8.3 , we chose optimal values of @xmath414 km s@xmath57 and @xmath415 km in order to model the fast wind that emerges from a polar coronal hole ( see also the stars in figure 6 ) . for the median value of @xmath416 erg s@xmath57 @xmath383 used above , the transition region occurs at a relatively large height of @xmath417 .
a traditional view of chromospheric energy balance is that when the acoustic heating is increased , the height of the transition region should decrease ( since the temperature would rise more rapidly as a function of @xmath11 ) .
figure 8 shows that the _ opposite _ actually occurs when we vary @xmath380 and keep everything else fixed .
we produced a series of models with @xmath418 , @xmath419 , @xmath381 , @xmath390 , @xmath382 , and @xmath420 erg s@xmath57 @xmath383 , as well as a model with @xmath421 ( i.e. , a model with alfvn wave heating only ) .
the models with @xmath421 and @xmath422 are virtually identical and the latter is not shown . before discussing the chromospheric heating
, we first note that the coronal and solar wind parameters for these models are remarkably constant .
this shows that varying the acoustic heating has relatively little impact above the transition region . as @xmath380
is increased from zero to its maximum value , the mass loss rate decreases by only 7% , the terminal speed increases by less than 1% , and the peak coronal temperature decreases by 5% . why does the transition region height increase as additional chromospheric heating is imposed ?
the models having larger values of @xmath380 have _ larger scale heights _ because they have both larger chromospheric temperatures and higher acoustic wave pressures .
the larger scale heights lead to a more shallow density decrease in the approximately hydrostatic chromosphere . the transition region tends to occur at a critical density at which radiative cooling can no longer keep pace with the imposed acoustic heating ( i.e. , because @xmath78 has reached its peak and can increase no further ) .
the `` shallower '' models with more acoustic heating reach this critical density at a larger height .
the modeled transition region heights @xmath423 range from 3800 km ( for @xmath421 ) to 11500 km ( for @xmath424 erg s@xmath57 @xmath383 ) .
for consistency , we define @xmath423 as the height where @xmath6 first reaches @xmath425 k , which is the peak of the radiative cooling curve ; see figure 1
. these heights are conspicuously larger than the standard values of , e.g. , 1700 to 2300 km from `` fal '' semi - empirical solar atmosphere models ( fontenla et al .
1993 ) . although there are some solar limb observations that suggest large values similar to our modeled @xmath423 ( zhang et al .
1998 ) , these may be affected by the dynamics of spicules and mass flows along closed magnetic loops , and thus may not be comparable directly to the present models ( see also filippov & koutchmy 2000 ) .
the relative stretching of the chromospheres shown in figure 8 can be understood further by examining the density dependence of the total heating rate : @xmath426 .
the exponent @xmath427 was computed for all of the models by utilizing the radial dependence of both @xmath9 and the total heating rate , with @xmath428 because the heating is balanced by radiative cooling ( which usually has @xmath429 ) , we find that values of @xmath427 closer to 2 give rise to a more extended `` matching '' between heating and cooling at chromospheric temperatures and thus a larger overall extent of the chromosphere .
thus , since @xmath430 for all of the models shown below , we can understand how larger [ smaller ] values of @xmath427 correspond to a higher [ lower ] transition region height .
for the shock heating described in 4.1 , the possible values of @xmath427 range between 0.5 and @xmath4311
. the lower limit would occur for an undamped weak shock train , for which @xmath432 ( eq . [ [ eq : wst ] ] ) and @xmath433 .
the upper limit occurs when the shock train strengthens to the point where @xmath162 saturates to a nearly constant value and thus @xmath434 .
( the peak mach number @xmath163 tends to saturate at values between 2 and 3 for the majority of the @xmath380 models . ) the zephyr models shown in figure 8 exhibit intermediate values of the density exponent @xmath427 , with the value of the exponent varying slightly as a function of height in the chromosphere .
the lowest value of @xmath423 corresponds to the @xmath421 model and @xmath4350.5 .
the highest value of @xmath423 corresponds to the @xmath424 model and @xmath4361 .
it is clear that , in these models , an even lower ( fal - like ) transition region height would require @xmath437 , and thus would require the heating rate to increase with height .
such heating may occur for the following reasons . 1 .
as described above , if the acoustic wave energy were to act more like a weak undamped shock train , @xmath427 would approach 0.5 .
the emerging picture of the chromosphere as a confluence of multiple acoustic sources ( each expanding quasi - spherically from small regions at or below the photosphere ) may be consistent with the need for more low - amplitude acoustic wave energy ( see , e.g. , ulmschneider et al .
a collection of weak and incoherent acoustic wave packets may also exhibit effectively larger frequencies because of their random phases and constructive interference .
2 . there could be additional important sources of linear wave dissipation in the low chromosphere , similar in form to @xmath133 ( which gives @xmath438 to 0.5 depending on the magnetic geometry ) .
for example , goodman ( 2000 , 2004 ) suggested that the resistive dissipation of currents driven by mhd waves may provide enough energy to heat the chromosphere .
similarly , it has been argued on the basis of multiple lines of observational evidence that both solar and stellar chromospheres may be dominated by magnetic and not acoustic heating ( e.g. , judge & carpenter 1998 ; judge et al .
2003 ; bercik et al .
if these mechanisms grow stronger as one moves from the lower to the upper chromosphere ( the latter being more strongly ionized and magnetized ) , this could produce the radial increase of the total heating rate that would be needed to push down the transition region .
however , we should note that anderson & athay ( 1989a ) determined empirically that the chromospheric heating rate seems to be roughly proportional to @xmath9 ( and thus @xmath439 ) throughout most of the chromosphere .
the mechanisms included in the zephyr models may thus be reasonable without the need for additional physics .
we experimented with several other variations on the acoustic heating in order to test the assumptions described in 4 .
a test model was created with the power spectrum @xmath190 divided into a finer mesh of discrete frequency bins : separated by factors of @xmath440 rather than 2 .
this model , which had @xmath416 erg s@xmath57 @xmath383 , produced a nearly identical chromosphere to the standard model with the same value of @xmath380 .
the fine - mesh model s transition region height @xmath423 was about 10% higher than that of the standard model because slightly more power was given to the lower frequencies ( which form shocks at larger heights ) .
presumably , using a coarser frequency spectrum would lead to a lower value of @xmath423 , but this would start to become unfaithful to the modeled spectral shape ( eq . [ [ eq : pomegas ] ] ) .
another model was run with just a single frequency bin ; i.e. , a monochromatic low - frequency wave train with @xmath441 .
this model was designed to explore what would happen if the high - frequency tail in @xmath190 were absent ( e.g. , fossum & carlsson 2005 , 2006 ) .
this model exhibited @xmath442 km , about a factor of two larger than the standard model , and would certainly require some additional kind of chromospheric heating to produce a realistically low value of @xmath423 .
note that for all of the above models the minimum temperature in the upper photosphere never dropped below the radiative equilibrium value of @xmath145 k. there is observational evidence , though , for a lower minimum temperature that could extend intermittently down to values between 3000 to 4000 k ( see model a of fontenla et al .
1993 ; carlsson & stein 1997 ; as well as recent work by ayres et al . 2006 ; fontenla et al .
additional sources of cooling that could be included in zephyr include molecular opacity and dust formation ( for cooler stars ) as well as adiabatic expansion effects due to waves and shocks that may _ not _ be confined to the modeled flux tube .
the exploratory models discussed above led to a choice for the optimal set of parameters which would reproduce the observed properties of high - speed solar wind streams that emerge from polar coronal holes ( mainly at the minimum of the solar cycle ) .
these parameters are @xmath414 km s@xmath57 , @xmath415 km , and @xmath416 erg s@xmath57 @xmath383
. the terminal speed @xmath393 and mass loss rate @xmath395 computed for this model are 753.5 km s@xmath57 and @xmath443 , respectively .
these values give a total hydrogen number density of 2.9 @xmath327 at @xmath444 au .
the semi - empirical alfvnic turbulence model of cranmer & van ballegooijen ( 2005 ) determined a value for @xmath237 at the mid - chromosphere merging height of about 1100 km , which corresponds to a photospheric value @xmath445 km .
the cranmer & van ballegooijen ( 2005 ) model also predicted a photospheric non - wkb amplitude @xmath446 of about 3.1 km s@xmath57 , which translates into @xmath447 km s@xmath57 .
these values are both slightly higher than the ones chosen on the basis of the self - consistent zephyr models , though they certainly fall within the range of plausibility .
the smaller value of the turbulence correlation length ( 75 km ) seems to be consistent with the observed horizontal size of a thin flux tube in the photosphere ( 50100 km ) .
this seems nicely consistent with being an outer `` stirring '' scale for the turbulence , since the horizontal shaking and distortion of the flux tube can be expected to take place mainly on the spatial scale of its own size .
( earlier justifications for the larger [ @xmath217 300 km ] correlation length were based on this being either the mean separation between flux tubes or the intergranular lane width .
these scales may be excited by convective driving as well , but it makes sense for the primary _ response _ of each tube to be on the smaller scale of its radius or diameter . ) for simplicity we continue to use the fiducial value for @xmath380 that was used in the two - dimensional grid shown in figure 6 . this energy flux density ( @xmath448 erg s@xmath57 @xmath383 ) corresponds to a photospheric acoustic wave amplitude @xmath384 of about 0.29 km s@xmath57 , and it is close to that computed by musielak et al .
( 2000 ) for linear longitudinal flux - tube waves .
the flux computed in the lighthill - stein sound generation models depends sensitively on the partitioning of gas and magnetic pressure in the photosphere ; these models have @xmath449 , where @xmath450 is the field strength consistent with equipartition between gas and magnetic pressures , and @xmath451 takes on large values typically between 6 and 9 .
our polar coronal hole model has a photospheric ratio @xmath452 .
interpolating from table 1 of musielak et al .
( 2000 ) would yield a flux for this model of about @xmath453 erg s@xmath57 @xmath383 , which is extremely close to what we use .
figure 9 displays the radial dependence of several key plasma parameters for the polar coronal hole model .
the critical point is denoted by a symbol , with @xmath454 and @xmath455 km s@xmath57 .
the wave - modified critical point is only a small distance above the classical sonic point , which occurs at 1.74 @xmath92 ( where @xmath456 km s@xmath57 ) .
this is somewhat surprising because the fast solar wind is typically assumed to be strongly `` wave dominated , '' and thus one might have expected the modified critical point to be far from the unmodified sonic point .
both of these points are also near the so - called _ turbopause _ radius at which the bulk solar wind speed begins to exceed the turbulent fluctuation amplitude @xmath225 ( @xmath457 ; see veselovsky 2001 ) .
in contrast , the alfvnic singular point ( where @xmath267 ) occurs at a substantially larger radius of @xmath458 and @xmath459 km s@xmath57 . note that figure 9a displays a frequency - integrated acoustic wave amplitude that we define for convenience as @xmath460 ( see eq .
[ [ eq : usdef ] ] ) , where we summed over the discrete frequency bins discussed in 4.2 and assumed @xmath186 everywhere .
this acoustic wave amplitude peaks at about 17 km s@xmath57 in the upper chromosphere ( and is mildly nonlinear , with @xmath461 peaking there at 0.96 ) , then is damped strongly in the low corona .
the temperature and density curves shown in figure 9b appear to be in reasonable agreement with prior expectations for the fast solar wind ( e.g. , kohl et al .
when @xmath9 is converted into electron number density @xmath33 , the modeled values agree with the canonical polar coronal hole measurements of sittler & guhathakurta ( 1999 ) to within @xmath462 20% .
these measurements come mainly from white - light polarization brightness ( @xmath463 ) observations between @xmath464 and 4 @xmath92 . considering that the absolute @xmath463 calibration uncertainties are typically about 10% and that the measured density depends on the filling factor of polar plumes along the line of sight ( which introduces @xmath463 variations up to a factor of two in magnitude ; see fisher & guhathakurta 1995 ) , this agreement is good .
the alfvn wave amplitudes ( @xmath225 in figure 9a and @xmath226 in figure 5 ) compare favorably with the semi - empirical models of cranmer & van ballegooijen ( 2005 ) , who compared those models with various remote - sensing and in situ measurements .
the spectrum - averaged reflection coefficient @xmath254 shown in figure 9b is also similar to that of cranmer & van ballegooijen ( 2005 ) ; note that it need not be monotonically decreasing with increasing distance .
the photospheric value of @xmath254 in this model is 0.883 , which is smaller than the value of 0.974 from cranmer & van ballegooijen ( 2005 ) .
the latter model utilized the fal - c@xmath43 temperature structure which had a sharper transition region than that produced by the zephyr code .
a more abrupt transition region has a larger @xmath465 gradient and thus experiences stronger reflection . defining the transition region thickness @xmath466 as the distance between temperatures of @xmath467 and @xmath468 k
, the fal - c@xmath43 model has @xmath469 km and the zephyr polar coronal hole model has a value of 560 km .
although some of the computed thickness may be due to numerical smoothing ( see 7 ) not all of it is .
the zephyr code did produce a sharper transition region ( @xmath470 km ) for the @xmath421 model shown in figure 8which was also the model with the lowest transition region height @xmath423 .
the radial dependence of temperature in the extended corona and heliosphere can be compared with various analytic limiting cases ( e.g. , hundhausen 1972 ) . at large distances where the wind speed is approximately constant and @xmath471
, there are several cases that give a power - law dependence @xmath472 .
the polar coronal hole model shown in figure 9 exhibits @xmath473 at 1 au . if the advection terms on the left - hand side of equation ( [ eq : dedt ] ) are dominant , one obtains the purely adiabatic exponent @xmath474 . if classical heat conduction were to dominate the energy equation out to 1 au , @xmath475 .
however , we find that conduction and advection tend to _ balance _ one another in the heliosphere and yield intermediate values . for the case that the advection terms are balanced by classical spitzer - hrm conduction , there is no exact power - law solution for @xmath108 . for advection
being balanced by collisionless conduction ( eq . [
[ eq : qfs ] ] ) , one obtains @xmath476 , which for the adopted value @xmath107 gives @xmath477 . in the zephyr models , the hybrid heat conduction ( eq .
[ [ eq : qpara ] ] ) is mostly collisionless at 1 au , but it still contains some classical heat conduction .
the computed values of @xmath4 thus tend to fall between the two above cases of 2/7 and 4/9 .
figure 9c shows the terms that dominate the energy conservation equation as a function of distance .
the areas plotted here were computed by normalizing the absolute values of the individual terms in equation ( [ eq : dedt ] ) by the maximum value at each height ( as in the denominator of eq . [ [ eq : dele ] ] ) and then `` stacking '' them so that together they fill the region between 0 and 1 .
the photosphere is essentially definable as the region where radiative heating balances radiative cooling ( both denoted as @xmath330 ) and the chromosphere is the region where acoustic heating ( @xmath131 ) balances @xmath330 .
the transition region and the very low corona ( @xmath478 mk ) exhibit a complicated balance of radiation , conduction , advection ( i.e. , enthalpy flux ) , and some alfvnic and acoustic heating .
the extended corona is mainly a balance between the alfvn wave heating ( @xmath302 ) and the advection terms on the left - hand side of equation ( [ eq : dedt ] ) , although conduction remains nonnegligible . in the heliosphere above 2030 @xmath92
the direct heating becomes less important and advection balances conduction . in order to produce further comparisons with in situ solar wind measurements
, we computed the nonequilibrium ionization balance of oxygen as a function of distance in the zephyr models .
specifically , the ratio of number densities of o@xmath479 to o@xmath480 is often used to aid in the identification of fast and slow wind streams ( e.g. , zurbuchen et al .
2002 ) . because of the steep decline in electron density with increasing height , solar wind ions above a certain `` freezing - in radius '' encounter virtually no electrons , and thus are not sensitive to ionization and recombination processes in interplanetary space ( hundhausen et al . 1968
; owocki et al . 1983 ) .
interplanetary charge states thus carry information about the plasma properties in the corona .
we adopted the nonequilibrium ionization code of gaetz et al .
( 1988 ) , esser et al .
( 1998 ) , and esser & edgar ( 2000 , 2001 ) to the zephyr models at temperatures above @xmath481 k. ionization and recombination are most sensitive to the electron velocity distribution in the corona , and we relied on our basic one - fluid assumption of a maxwellian distribution with @xmath482 .
we also assumed that all oxygen ions flow with the bulk wind speed @xmath5 , independent of their charge .
figure 10 shows the result of computing the oxygen ionization state for the polar coronal hole model .
the mean charge state @xmath483 is computed as an average of the net charge of the ions ( in units of @xmath484 ) weighted by the computed number density fractions of each stage of ionization .
the equilibrium solution assumes a local `` coronal '' balance between collisional ionization , radiative recombination , and dielectronic recombination , and is a strict function of temperature .
the full nonequilibrium solution freezes in at a relatively low height in the corona ( @xmath401 ) and is roughly constant at all larger heights .
note , though , that the equilibrium and nonequilibrium solutions are also different from one another in the upper transition region ( @xmath4850.02 @xmath92 ) where the wind flow time over a scale height is beginning to approach the relevant ionization and recombination times .
the dominant ionization state at 1 au is o@xmath480 ( i.e. , @xmath486 ) and the ratio of o@xmath479 to o@xmath480 is discussed further below .
there is a definite empirical relationship between the solar wind speed measured in situ and the inferred lateral expansion of magnetic flux tubes near the sun ( levine et al .
1977 ; wang & sheeley 1990 ; arge & pizzo 2000 ) .
specifically , flux tubes that expand more rapidly between the coronal base and @xmath487 tend to have lower wind speeds at 1 au .
there have been many theoretical attempts to explain this observed anticorrelation .
some models have treated the flux expansion as the fundamental cause of the difference between fast and slow streams ( e.g. , kovalenko 1978 , 1981 ; wang & sheeley 1991 , 2006 ; wang 1993 ; cranmer 2005a ) , some have treated it as a by - product of varying boundary conditions in the low corona ( e.g. , fisk 2003 ; schwadron & mccomas 2003 ) , and some have considered a combination of the two ( bravo & stewart 1997 ; suzuki 2006 ) . our goal is to test these ideas by varying _ only _ the flux expansion rate and keeping the lower boundary parameters ( @xmath388 , @xmath389 , @xmath380 ) fixed . the first way that we modify
the flux expansion is to utilize the full two - dimensional axisymmetric magnetic field model of banaszkiewicz et al .
this model utilizes a sum of dipole , quadrupole , and current - sheet ( effective monopole ) terms to reproduce various observed properties of the solar - minimum field .
note that other assumptions lead to slightly different solar - minimum field configurations ( e.g. , sittler & guhathakurta 1999 , 2002 ; vsquez et al .
2003 ) but the general trends of the polar and equatorial expansion factors are not substantially different from the banaszkiewicz et al .
( 1998 ) case .
figure 11 illustrates a selection of the open field lines in the banaszkiewicz et al .
( 1998 ) model .
figure 5 also shows the magnetic field @xmath488 traced along a subset of these field lines .
we define the surface colatitude @xmath489 as the main identifier of each flux tube .
the polar flux tube that was used above in
8.3 has @xmath490 .
we take the `` last '' open field line ( at the outermost edge of the open - field region ) as @xmath491 . because the latter field line eventually stretches to a colatitude of nearly 90 we also call this the `` equatorial '' flux tube .
we assume that the differences in @xmath16 between flux tubes occur only above @xmath492 and that in the photosphere and chromosphere all flux tubes are identical ( see , e.g. , aiouaz & rast 2006 ) .
let us define an effective wang & sheeley ( 1990 ) superradial expansion factor for each flux tube as @xmath493 where the coronal base is fixed at @xmath494 ( i.e. , a height equivalent to the size of one supergranule , well above any `` canopy '' structure ) and we use the traditional source surface radius of @xmath495 ( hoeksema & scherrer 1986 ) .
it would not be proper to use the photosphere as the basal height because the low - resolution magnetograms that typically are used to determine @xmath496 resolve neither the individual photospheric flux tubes nor the contrast between supergranular network and cell interior .
for the banaszkiewicz et al .
( 1998 ) field configuration , @xmath496 ranges from 4.5 for the polar flux tube to 9.1 for the equatorial flux tube .
note that this model does not produce the larger values of @xmath49740 that have been found during more active phases of the solar cycle ( see 8.5 ) .
a series of 20 zephyr models was computed with the same photospheric boundary conditions as the polar coronal hole model discussed in 8.3 , but varying @xmath498 from 0 to 29.7 .
going from pole to equator , the asymptotic solar wind speed @xmath393 was found to decrease and the mass loss rate @xmath395 was found to increase .
figure 11 shows how the radial dependence of outflow speed changes as @xmath498 increases from 0 to 29.7 .
there is an abrupt bifurcation between two classes of solar wind solution : ( 1 ) polar solutions that have fast wind speeds , low densities , and wave - modified critical points @xmath499 , and ( 2 ) equatorial solutions that have slow wind speeds , high densities , and @xmath500 .
the colatitude at which this bifurcation occurs is @xmath501 . as discussed by vsquez et al .
( 2003 ) and in 6 above , when there is more than one possible location for the critical point , the most stable time - steady solution tends to be the one with the largest value of @xmath296 .
for the banaszkiewicz et al .
( 1998 ) field geometry , the second ( large-@xmath296 ) solution appears only for @xmath502 .
the field line that divides the two classes of solutions extends out to a heliospheric _ latitude _ ( measured from the ecliptic plane at @xmath503 ) of 16 .
figure 11 shows that the solar wind solutions all exhibit substantial wind speeds just above the transition region ( @xmath504 km s@xmath57 ) .
esser et al .
( 2005 ) found that such rapid flows at the coronal base are a consequence of the rapid superradial divergence in supergranular funnels , and that these flows seem to be necessary to explain observed h i ly@xmath81 emission from the solar disk .
note also that our modeled flux tubes nearest to the equator exhibit a pronounced local minimum in the wind speed at the cusp - radius of the current sheet in the banaszkiewicz et al .
( 1998 ) model .
observational evidence for this kind of `` stagnation point '' in equatorial streamers was reported by strachan et al .
( 2002 ) and was also modeled in a similar way as in this paper by chen et al .
( 2004 ) and li et al .
( 2004 ) .
it is also evident from figure 11 that the polar model has a larger alfvn wave amplitude than the equatorial model .
some of this difference can be attributed directly to the varying flux - tube area factors @xmath265 ( i.e. , different manifestations of wave action conservation ) .
however , the equatorial flux tube also exhibits relatively more damping than the polar model . in the corona
, the alfvn waves essentially `` spend more time '' at low heights in the equatorial flux tube because of its lower phase speed ( @xmath505 ) , and this allows a given damping rate @xmath302 to have more of an impact on diminishing the wave energy .
thus , the equatorial models exhibit more heating at low heights ( due to the increased damping and the higher densities at the coronal base ) and less heating at large heights ( due to the lower wave amplitudes @xmath225 and larger turbulent correlation lengths @xmath237 in the extended corona ) .
the peak temperature for the polar model , @xmath506 mk , occurs at @xmath507 , and the peak temperature for the equatorial model , @xmath508 mk , has a similar magnitude but occurs at a lower height of @xmath509 ( see also figure 17 below ) .
this shift in the range of heights over which coronal heating occurs seems to be an important factor in producing a fast solar wind for the polar model ( which is heated in the supersonic region ) and a slow wind for the equatorial model ( which is heated in the subsonic region ) ; see 8.1 .
the differences in temperature persist out to 1 au , where the polar temperature ( 0.41 mk ) exceeds the equatorial temperature ( 0.16 mk ) by a larger relative amount than in the extended corona . the radial temperature exponent @xmath4 ( i.e. , @xmath510 ) grows steeper from pole ( @xmath473 ) to equator ( @xmath511 ) .
although the photospheric alfvn wave amplitude parameter @xmath388 was kept fixed in these models , the degree of non - wkb reflection decreased slightly from the polar to the equatorial flux tube .
this resulted in a decrease in the computed value of @xmath446 ( eq . [ [ eq : wperp ] ] ) from its polar value of 0.72 km s@xmath57 to an equatorial value of 0.59 km s@xmath57 .
interestingly , if we could have forced @xmath446 to remain fixed as a function of latitude , the resulting equatorial value of @xmath388 would have been about 20% _ larger _ than the polar value . figure 6
indicates that the effect would have been an even larger equatorial mass loss rate and a slightly lower @xmath393 .
thus , the current set of models is in some sense `` robust '' because these effects arise even without such an extra fine - tuning in @xmath446 .
the detailed latitudinal dependence of the zephyr models ( for an observer in interplanetary space ) is shown in figure 12 .
we compare the model predictions with _
ulysses _ swoops ( solar wind observations over the poles of the sun ; bame et al .
1992 ) measurements during its first fast latitude scan in 1994 and 1995 ( see goldstein et al .
the measured values for flow speed , proton number density , and proton temperature were obtained at heliocentric distances between 1.3 and 2.3 au .
the densities were scaled to a common distance of 1 au by using an assumed @xmath512 dependence . for comparison
we plot @xmath393 , which does not vary substantially between these distances , and the proton number density ( @xmath513 ) at 1 au , scaled for an assumed helium - to - hydrogen number density ratio of 0.05 .
the model temperature @xmath108 has been interpolated to varying heights as a function of latitude to match the elliptical orbit of the _ ulysses _ probe during its polar pass .
also , we used the banaszkiewicz et al .
( 1998 ) model to map the surface colatitude @xmath489 of each model flux tube to its corresponding heliospheric latitude at @xmath514 .
the latter is the abscissa coordinate used in figure 12 . to within about 10% accuracy
, one can estimate this latitude as @xmath515 .
there are two sets of pole - to - equator zephyr models shown in figure 12 .
we compare the standard model discussed above with a preliminary model presented by cranmer ( 2006 ) , which we hereafter call the `` durham '' model .
the models differ in two specific ways .
( 1 ) the standard model used the alfvn wave frequency spectrum shown in figure 3 , but the durham model assumed a single frequency for the alfvn waves corresponding to a period of 6 minutes .
( 2 ) the standard model used an exponent @xmath47 in equation ( [ eq : eturb ] ) , but the durham model used @xmath48 .
the latter was an early guess for the functional form of @xmath240 that we used prior to the incorporation of the insights of oughton et al .
the larger exponent resulted in more efficient quenching of the mhd turbulence in the lower atmosphere , which thus required a larger amount of alfvn wave energy to produce plasma conditions appropriate for a polar coronal hole . specifically , the durham model used @xmath516 km s@xmath57 , @xmath517 km , and the same value of @xmath380 as was used in the standard model .
the durham model s @xmath393 and @xmath395 for the polar flux tube were nearly identical to those of the standard model .
the peak coronal temperature , though , was substantially larger ( 2.06 mk ) and the temperature at 1 au was slightly smaller ( 0.31 mk ) in comparison to the standard polar model .
there is reasonably good agreement between the modeled and observed plasma properties shown in figure 12 .
we of course do not reproduce any of the measured north - south asymmetry in the data , except for that in @xmath518 that arose because the path of _ ulysses _ was not perfectly north - south symmetric .
we also do not attempt to model the rapid near - ecliptic variability that seems to result from longitudinal variations in @xmath16 . in some sense , the durham model does better at achieving the low equatorial wind speeds ( down to 330 km s@xmath57 ) and high number densities ( up to about 10 @xmath327 ) that are often observed in the plane of the ecliptic , but that model exhibits a narrower range of slow - wind ( large-@xmath296 ) latitudes than both the standard model and the _ ulysses _ data .
it is expected , though , that some fraction of the observed @xmath519 latitudinal extent of the slow - wind region may be due to a slight tilt in the main dipole component of the solar magnetic field ( i.e. , a `` ballerina skirt '' phenomenon ) because the first _ ulysses _ polar pass did not occur exactly at the minimum of the activity cycle .
note from figure 12c that the zephyr model temperatures between 1.3 and 2.3 au are systematically higher than the measured values of @xmath518 .
the trend as a function of the coupled distance - latitude coordinate , though , is similar to the measurements .
our heliospheric temperatures are sensitive to the adopted method of bridging between collisional and collisionless heat conduction ( eq . [ [ eq : qpara ] ] ) and the present method is admittedly somewhat ad hoc . however , some of the discrepancy in figure 12c may come from the fact that @xmath520 in interplanetary space and that we are essentially modeling their average . in situ measurements
have shown that @xmath521 in the fast solar wind and @xmath522 in the slow wind ( see , e.g. , feldman & marsch 1997 ) .
the discrepancy is worst in the polar , fast - wind regions where our modeled average temperature would exceed _ both _ the observed @xmath518 and @xmath523 .
however , the zephyr code does not yet contain the dissipative physical processes that would lead to temperature equilibration between protons and electrons ; these may alter subtly both the net heating and the form of the conduction flux @xmath524 .
figure 13 plots the terminal speeds of the standard and durham zephyr models as a function of the wang - sheeley superradial expansion factor @xmath496 .
the abrupt changes in @xmath393 that occur because of the emergence of the outer critical point are evident as vertical discontinuities at the critical values of @xmath489 . because of the relatively narrow range of @xmath496 values subtended by the pole - to - equator ( banaszkiewicz et al .
1998 ) flux tubes , we see that the standard model actually follows the the empirical wang & sheeley ( 1990 ) relationship better than the durham model , despite the fact that the standard model has a relatively high equatorial terminal speed of 500 km s@xmath57 . figure 14 compares modeled and measured alfvn wave amplitudes in the heliosphere as a function of the in situ wind speed .
it is difficult to process the richly nonlinear turbulent fluctuations ( as measured in situ ) into a single `` amplitude , '' so we made an attempt to convert the zephyr values of @xmath225 into something analogous to the frequency - averaged elsasser spectrum quantities reported by tu et al .
those data were obtained by the two _ helios _ spacecraft between 0.29 and 0.52 au in the years 19791980 .
( measurements taken closest to the sun are optimal if the goal is to obtain the properties of the `` pristine '' alfvn waves . )
individual time series of @xmath525 ( outward - propagating kinetic energy ) magnitudes for one - day periods were transformed into power spectra , denoted @xmath526 .
the values shown in figure 14 were averaged over frequencies @xmath527 between @xmath377 and @xmath528 hz .
the plotted data points also correspond to time periods when the solar wind exhibited its highest range of total energy flux ( @xmath529/2 $ ] ) , which we verified to correspond most closely with the zephyr model results .
we extracted the elsasser amplitudes @xmath251 from the models at a mean distance of @xmath530 au and converted them into the average value of @xmath531 as follows .
first , representative spectral shapes for these fluctuations were used in order to determine how much of the total energy is contributed by the specified frequency band ( 1@xmath528 hz ) .
we used the power - law exponents and breakpoints reported by tu et al .
( 1989 ) , using the same _ helios _ data , to determine that fast and slow wind streams contain approximately 11% and 21% , respectively , of their power in this band .
we used a mean value of 16% for all zephyr models to avoid introducing any possibly spurious correlations with wind speed . to convert the ( implicitly frequency - integrated ) energy fraction into an `` average '' spectrum quantity we divided by an effective frequency @xmath532 that was defined to perform this conversion exactly for the idealized spectra of tu et al .
( 1989 ) ; it turned out that @xmath533 hz for both fast and slow wind spectra and we used just @xmath377 hz . thus we plot the derived quantity @xmath534 in figure 14 in order to compare with the measured average spectra from tu et al .
( 1992 ) . the modeled and measured values and trends with wind speed
appear to agree quite nicely . in the remainder of this section
we discuss two supplementary calculations of in situ quantities that are often used to diagnose the properties of fast and slow wind streams : the ionization fraction ( figure 15 ) and the fip effect ( figure 16 ) .
the _ nonequilibrium oxygen ionization state _ at 1 au has been computed for each of the standard and durham model flux tubes .
figure 15 shows the ratio of o@xmath479 to o@xmath480 number densities as a function of the modeled terminal speed .
these are compared to statistical summaries of _ ulysses _ swics ( solar wind ion composition spectrometer ; gloeckler et al .
1992 ) measurements of this ratio taken during two time periods : the first polar pass discussed above ( september 1994 to august 1995 ) and the initial phase of the mission that sampled the peak of the solar cycle ( december 1990 to september 1994 ) .
the data points , obtained from the online database of t. zurbuchen and r. von steiger ( see also von steiger et al . 2000 ) , were grouped in sequential bins of wind speed of size 25 km s@xmath57 , and the mean and standard deviation of the ionization ratio in each bin were computed .
the regions plotted in figure 15 show the values within @xmath535 standard deviations of the mean for each bin in the two data sets .
the general trend of the pole - to - equator zephyr models is for lower wind speeds to have higher ratios of o@xmath479 to o@xmath480 at 1 au , in qualitative agreement with the observations .
the absolute values of the modeled ionization ratios , though , are up to an order of magnitude lower than the _
ulysses _ measurements , with slightly better agreement for the hotter durham models .
this discrepancy may be the result of our simple assumption of maxwellian electron distributions and equal flow speeds for all ion charge states ( see , e.g. , esser & edgar 2000 , 2001 ) .
a ubiquitous weak `` halo '' of suprathermal electrons in the low corona could be responsible for the larger overall degree of ionization that is observed . as one moves from the polar flux tube to the middle latitudes ( i.e. , from @xmath536 km s@xmath57 down to about 650 km s@xmath57 ) the modeled ionization ratio decreases , but then as one passes the bifurcation point between the low and high critical radii , the ionization ratio increases for the rest of the way from the mid - latitude region to the equator .
the overall trend for slow wind to have a high ionization ratio is expected because of the higher temperatures in the narrow zone just above the transition region .
in fact , the temperature at a constant height of @xmath537 has the same quantitative trend as the modeled o@xmath479 to o@xmath480 ratios : there is an initial decrease from the polar model ( 0.37 mk ) to just before the bifurcation latitude ( 0.28 mk ) , then an abrupt jump back to nearly the polar value ( 0.34 mk ) followed by a steady increase toward the equatorial flux tube ( 0.46 mk ) .
the freezing in occurs at about this height , so it makes sense that the ionization state tracks these variations .
notice that this ionization ratio is sensitive only to the temperature trends at low heights ( and that the o@xmath479 to o@xmath480 ratio freezes in just above the transition region ) and is _ not _ sensitive to the temperatures at larger heights in most of the extended corona .
some interpretations of the in situ charge states make the implicit assumption that high electron temperatures in the slow wind must persist all the way through the extended corona into interplanetary space , and our models show that this need not be the case .
a specific diagnostic of _ fip fractionation _ has been computed for each of the standard and durham model flux tubes .
figure 16 shows a commonly measured ratio of the abundances of a low - fip element ( fe , 7.9 ev ) and a moderately high - fip element ( o , 13.6 ev ) normalized to their photospheric abundance ratio .
we solved equation ( [ eq : lamfip ] ) as described in 6 , and by laming ( 2004 ) , for the various models and compared the ratios to the _ ulysses _ swics observations of the fe / o ratio .
the swics ratio was normalized to an assumed photospheric iron - to - oxygen abundance ratio of 0.0468 ( grevesse & sauval 1998 ) and the modeled quantity is naturally in units of the photospheric ratio . because the fip fractionation takes place in the uppermost part of the chromosphere ( which is largely optically thin ) and the very low transition region , we assumed `` coronal '' ionization equilibrium to hold for the modeled fe and o ions .
we used the code developed by cranmer ( 2000 ) to compute the fe and o ionization fractions @xmath306 that are used in equation ( [ eq : lamfip ] ) ; the rates were taken largely from mazzotta et al .
( 1998 ) . figure 16 shows that our standard pole - to - equator zephyr model exhibits a roughly constant fe / o ratio of about 2 for outflow speeds between 500 and 750 km s@xmath57 .
this is in reasonable agreement with the _ ulysses _ data , which have a roughly constant mean value of about 1.5 over these speeds .
the durham model varies nonmonotonically between ratios of about 3 and 5.5 over its wider range of outflow speeds and is clearly inconsistent with most of the in situ data .
the equatorial durham model , with @xmath538 km s@xmath57 , does seem to fall into agreement with the measured data , but the overall trend as a function of wind speed is not the same . the most promising models for matching the observed trend for a steep enhancement of the fe / o ratio ( as the wind speed decreases from 500 to 300 km s@xmath57 ) appear to be the active - region flux tubes discussed below in 8.5 .
we should note that an alternate theoretical explanation for fip fractionation may also be consistent with the zephyr models .
wang ( 1996 ) proposed that flux tubes with a stronger degree of downward heat conduction in the low corona could give rise to a larger amount of transient evaporative outflow of protons relative to neutral hydrogen atoms .
these protons could collisionally dredge up singly ionized species ( which are preferentially low - fip ) from the transition region and upper chromosphere . in our models ,
the maximum value of @xmath539 occurs within the transition region ( at @xmath540 k ) , and in the equatorial models this peak value is consistently _ larger _ than in the polar models .
the ratio of equatorial to polar maximum conduction rates is 1.7 for the standard model and 2.8 for the durham model . whether this is enough to produce a fip effect with wang s ( 1996 ) mechanism
is not known , but the variations are at least in the correct qualitative direction to produce more of an effect in the slow wind than in the fast wind .
note , though , that schwadron et al .
( 1999 ) disagree with some of the basic assumptions of the wang ( 1996 ) mechanism , and a definitive explanation for the fip effect remains elusive .
there are additional observational diagnostics that appear to be correlated with fast and slow wind speeds .
kojima et al . ( 2004 ) and suzuki ( 2006 ) found that the ratio of basal magnetic field strength to the superradial flux - tube divergence factor ( i.e. , @xmath541 ) may be a better predictor of solar wind speed than just using @xmath496 itself ( see , however , wang & sheeley 2006 ) .
mcintosh & leamon ( 2005 ) found that the difference in formation heights between the emission seen in the _ trace _ 1600 and 1700 filter bandpasses could be a useful diagnostic of the eventual wind speed in flux tubes that are associated with specific regions on the solar surface .
although there have been numerous observational inferences of blueshifts ( i.e. , outflow ) in various spectral lines , it is unclear to what degree these may be utilized as empirical `` lower boundary conditions '' for solar wind flux tubes ( see reviews by peter et al .
2004 ; jones 2005 ) . finally , there are definite fast / slow wind correlations in the wealth of existing turbulence measurements both in situ and remote sensing that could be key tests of models that use waves and turbulence for energy and momentum deposition .
future development of the zephyr models will encompass such additional comparisons . at times other than solar minimum
, the coronal magnetic field is substantially more complex than the axisymmetric configuration modeled in the previous section . during the approach to solar maximum
the large polar coronal holes disappear as magnetic flux of the opposite polarity is brought toward the poles .
active regions with strong fields emerge at all latitudes and give rise to an interconnected web of closed loops and intermittent open flux tubes . in the extended corona
, active regions often have the appearance of cusp - like `` active streamers '' that seem like smaller , more intense versions of the quiescent equatorial streamers that dominate at solar minimum ( newkirk 1967 ; liewer et al .
2001 ; ko et al .
there is increasing evidence that much of the slow solar wind is connected somehow to open flux tubes that are rooted in or near active regions ( e.g. , neugebauer et al .
2002 ; liewer et al .
2004 ) .
we investigated several different ways to model active - region flux tubes with the zephyr code .
first , we explored just changing the radial location of the cusp in the equatorial banaszkiewicz et al .
( 1998 ) flux tube .
observationally , streamers exhibit a wide range of cusp heights which gives rise to a large variation in their rates of superradial expansion .
the cusp height was varied by first computing the ratio of the equatorial ( @xmath491 ) to the polar ( @xmath490 ) magnetic fields @xmath542 , as shown in figure 5 , then shifting the height coordinate @xmath11 either upward or downward by a constant multiplier . the `` new '' equatorial field
was obtained by taking the product of this shifted ratio and the polar field strength . changing the cusp height by up to
a factor of two in either direction did not have much of an effect on the resulting solar wind properties .
the mass loss rates varied by about 18% for these models ( lower cusps having lower @xmath395 ) , whereas the terminal speeds @xmath393 varied by less than 3% .
this near constancy in the solar wind speed occurred despite the fact that @xmath496 was made to increase and decrease by factors of @xmath2172 about the equatorial value of 9.1 .
the above models can not be considered true active - streamer models because they were not given the significantly _ stronger _ mean magnetic field strengths that active regions are observed to have in the lower atmosphere . on the smallest scales in the photosphere , plage and active regions
exhibit a much higher filling factor of strong - field flux concentrations than quiet regions and coronal holes . in the most tightly - packed active regions
these flux concentrations do not appear as isolated flux tubes but instead as fluted ribbons , buckled flux sheets , and flower - like patterns ( e.g. , berger et al .
2004 ; rouppe van der voort et al .
we are concerned here only with the subset of magnetic flux that extends out into interplanetary space , so we provisionally retain the `` flux tube '' picture .
we thus modeled @xmath488 in an active - region flux tube by enhancing the field strength in the chromosphere and low corona .
this is done in general agreement with the larger observed magnetic fields , but also to account for the smaller area @xmath265 subtended by open flux tubes that are `` crowded '' by neighboring closed loops in active regions .
we added an active - field component to the standard equatorial banaszkiewicz et al .
( 1998 ) magnetic field by taking @xmath543 \label{eq : bactive}\ ] ] where @xmath544 is the equatorial ( @xmath491 ) model and @xmath545 is a constant that we set at 50 g. we produced a grid of models that varied the scale height @xmath546 between 0 and 0.07 @xmath92 .
the model with @xmath547 is equivalent to the standard equatorial model with no active - field enhancement .
equation ( [ eq : bactive ] ) takes the greater of the two magnetic field quantities , which tends to retain the original @xmath548 in the photosphere , low chromosphere , and outer corona and wind , but produces the active - region enhancement in the upper chromosphere and low corona ( see figure 5 ) .
the smallest values of the scale height ( @xmath549 ) do not produce any enhancement in @xmath18 because the quantity @xmath550 is rapidly decreasing and it never exceeds @xmath548 . for larger values of @xmath546 , the magnetic field strength at the coronal base
is enhanced and thus the wang - sheeley expansion factor @xmath496 is enhanced as well .
the maximum value of @xmath551 ( i.e. , 49 mm ) corresponds to @xmath552 .
even though we used the exact values of @xmath496 computed for each model in the plots below , we found the following approximate fit to be a useful illustration of how @xmath496 depends on @xmath546 : @xmath553 where @xmath546 is in units of @xmath92 .
the threshold value of 0.02 @xmath92 corresponds to the point at which @xmath550 first produces a significant enhancement at @xmath554 .
zephyr models were computed for active region field enhancements with @xmath555 , 0.02 , 0.024 , 0.03 , 0.035 , 0.04 , 0.045 , 0.05 , 0.06 , and 0.07 @xmath92 .
the active region models exhibited lower terminal speeds and higher mass loss rates than the standard equatorial model . as @xmath546 was increased from 0 to 0.07 @xmath92 , @xmath393 decreased from 500 to about 390 km / s .
the mass loss rate increased from @xmath556 to @xmath557 @xmath558 , thus increasing the proton number density at 1 au from 6.4 @xmath327 to about 10 @xmath327 .
figure 13 shows how @xmath393 decreases as @xmath496 increases for these models .
there is reasonable agreement with the empirical wang & sheeley ( 1990 ) relationship , and it is important to reiterate that all models in figure 13 were produced by fixing the basal parameters ( @xmath388 , @xmath389 , @xmath380 ) at values appropriate for the polar coronal hole and varying only the magnetic field above the photosphere in order to produce the fast - to - slow wind transitions . figures 15 and 16 illustrate the computed o@xmath479 to o@xmath480 freezing - in ratios and fe / o ( fip fractionation ) ratios for the active region models .
these models mirror the observed trends for there to be larger ratios for both quantities at lower wind speeds .
the strengthening of the fip ratio , from @xmath2171.5 to 4 as the wind speed decreases from 500 to 400 km s@xmath57 , is especially comparable to the observed increase toward the lowest speeds .
figure 17 compares the radial dependence of temperature for a selection of active models to that for the standard polar and equatorial models . as @xmath546 increases from 0 to 0.07 @xmath92 , the peak coronal temperature increases from the equatorial model s value of 1.29 mk up to 1.67 mk .
the heliocentric radius of this peak decreases from 1.26 to 1.11 @xmath92 .
these trends appear to be a continuation of the variations in coronal heating from the polar to the equatorial flux tubes .
as one progresses from pole to equator ( @xmath559 ) to active region ( @xmath560 ) , the models show a successively stronger decrease in the alfvn speed with height , more reflection ( and damping ) in the low corona , and thus less alfvn wave power available to heights at and above the critical point .
the iterative convergence for the largest-@xmath546 active region zephyr models was slower and less robust than for the models discussed in
8.18.4 .
the models having @xmath561 were run for 400 outer iterations in order to reach acceptable convergence parameter values : @xmath562 to 0.02 .
( trial models with even larger values of @xmath563 did not converge below @xmath564 after 400 iterations and were not used . ) unlike in figure 4 , where acceptable values of 0.010.02 were retained in nearly every iteration ( after 100 or so outer iterations had passed ) , the active models exhibited these values only sporadically .
thus , the iteration with the lowest value of @xmath342 did not necessarily represent the `` best '' single solution . in order to evaluate the most likely time - steady plasma parameters
, we considered a collection of the lowest-@xmath342 solutions sampled from the iteration stream for each model run .
this was done by taking the distribution of values for a given parameter , such as @xmath393 , for all 400 outer iterations and weighting it by a factor proportional to @xmath565 in order to deemphasize the early iterations that are farthest from convergence .
figure 13 shows both the mean and @xmath535 standard deviation values for this weighted distribution of well - converged terminal speed solutions .
as @xmath546 is increased , the standard deviation first decreases from about 20 to 6 km s@xmath57 , then increases again to 35 km s@xmath57 .
it is unclear whether the nonconvergence of the large-@xmath546 active region models represents a failing of the numerical technique or an intrinsic lack of a steady - state solution .
the coronal sources of slow solar wind are traditionally viewed as more variable and filamentary than the source regions of fast wind .
high - resolution coronagraphic observations of streamers reveal an almost continual release of low - contrast density inhomogeneities ( sheeley et al .
1997 ; wang et al .
multidimensional simulations of streamers have also exhibited time - variable solutions ( e.g. , suess et al .
1996 ; endeve et al . 2003 , 2004 ) .
for flux tubes in the vicinity of the boundary between open and closed magnetic fields , the complex interplay between heat conduction and transverse pressure balance can lead to various instabilities that drive variability .
certain open magnetic field configurations ( possibly like the one adopted in eq .
[ [ eq : bactive ] ] ) may not be stable when considered in the context of its neighbors in a full - sun model .
other techniques for constructing active - region solar wind models should be explored .
the specific functional form of the @xmath16 enhancement in the low corona can be improved in order to agree with , e.g. , force - free magnetic field reconstructions of active regions ( schrijver et al .
in addition , the photospheric field strength may also be substantially higher in active regions than elsewhere .
it is unclear at first glance how the photospheric wave fluxes would be modified in these regions , though . for strong enough basal fields ( i.e. , in sunspots )
convection is quenched and the wave fluxes should be lower . however , active regions exhibit a larger degree of magnetic flux emergence ( e.g. , abramenko et al . 2006 ) which could increase the frequency of impulsive horizontal `` jumps '' exhibited by flux - tube footpoints . as summarized above , flux tubes are more more highly concentrated in active - region intergranular lanes than elsewhere .
the close presence of additional flux tubes may either increase the basal wave amplitude in a given tube ( via `` sympathetic '' motions from the neighbors ) or decrease it ( because large horizontal displacements are blocked by the neighbors ) .
similarly , the correlation length @xmath389 in these regions may be either lower ( because of the smaller domain of motion when crowded ) or higher ( because the lane - filled `` ribbons '' may move together ) .
comparisons between high - resolution observations of g - band bright points in active regions ( mstl et al .
2006 ) and three - dimensional magnetoconvection simulations ( e.g. , bushby & houghton 2005 ) are needed in order to clarify these competing effects .
the primary aim of this paper has been to construct self - consistent models of chromospheric and coronal heating , waves and turbulence , and wind acceleration in an open magnetic flux tube rooted in the solar photosphere .
a key aspect of the zephyr models presented above is that the only true free parameters are : ( 1 ) the properties of the waves injected at the base , and ( 2 ) the background geometry and magnetic field strength along the modeled flux tubes .
everything else ( e.g. , the radial dependence of the rates of chromospheric and coronal heating , the resulting temperature structure of the atmosphere , and the solar wind speed and mass flux ) is an emergent property of the model and not an ad hoc input . in the above sense
, the zephyr models are similar to those of suzuki & inutsuka ( 2005 , 2006 ) ; the differences between the two approaches involve mainly the specific physical processes that are assumed to dissipate the alfvn waves and heat the corona . as discussed in 2 ,
the zephyr models presented here are limited by being one - dimensional , time - independent , and one - fluid solutions to the hydrodynamic conservation equations , with heating that is derived only from a subset of all possible wave and turbulent dissipation processes .
even so , the results given in
8 show that a realistic variation of asymptotic solar wind conditions can be produced by varying only the background magnetic field geometry , as predicted by wang & sheeley ( 1990 , 1991 , 2003 , 2006 ) .
specifically , our models show general agreement with some well known empirical correlations : i.e. , a larger @xmath496 ( expansion factor ) gives rise to a slower and denser wind , less intense alfvnic fluctuations at 1 au , and larger values of both the o@xmath479/o@xmath480 charge state ratio and the fip - sensitive fe / o abundance ratio . satisfying these kinds of scalings are necessary but not sufficient conditions for validating the idea that the wind is driven by a combination of mhd turbulence and non - wkb alfvn wave reflection .
future work must involve more physical realism for the models , expanded comparisons with existing observations , and predictions of as - yet unobserved quantities that may be key discriminators between competing theoretical models .
an important component of all three efforts will be to model the divergent temperatures and flow speeds of protons , electrons , and various heavy ion species in the extended corona and heliosphere .
these mainly _ collisionless _ regions allow the kinetic physics of wave dissipation to be probed to a level of detail not possible in a collisionally coupled ( i.e. , essentially one - fluid ) plasma . even in a perfectly collisional plasma
, there can be _ macroscopic _ dynamical consequences depending on how the energy is deposited into protons , electrons , and possibly heavy ions as well .
for example , if all of the heat goes into electrons , there can be substantially more downward conduction than in a proton - heated model , which would affect the coronal temperature distribution ( hansteen & leer 1995 ) and the stability of helmet streamers ( endeve et al .
two - fluid effects also may affect the phenomenological form of the mhd turbulent cascade assumed in the above models ( e.g. , galtier 2006 ) . the original motivation for this work was to understand the preferential ion heating and acceleration in the corona revealed by the uvcs ( ultraviolet coronagraph spectrometer ) instrument on _ soho _
( kohl et al . 1995 , 1997 , 2006 ) .
if mhd turbulence gives rise to cyclotron - resonant alfvn waves which are likely to be responsible for the observed ion properties then we must first have a large - scale description of the energy flux injected into the turbulent cascade in order to further model the end - products of that cascade .
we hope to link the macroscopic properties modeled by zephyr with the microscopic kinetic processes studied by , e.g. , leamon et al .
( 1999 , 2000 ) , cranmer & van ballegooijen ( 2003 ) , voitenko & goossens ( 2003 , 2004 ) , gary & nishimura ( 2004 ) , dmitruk et al .
( 2004 ) , chandran ( 2005 ) , gary et al . ( 2006 ) , and markovskii et al .
( 2006 ) .
in addition to increasing the realism ( and complexity ) of the models , attempts should be made to extract the dominant physical processes of these models and create simpler `` scaling laws '' that can be used to estimate the wind conditions for varying magnetic field geometries and photospheric wave properties ( see also leer et al .
1982 ; schwadron & mccomas 2003 ; suzuki 2006 ) .
an obstacle to accomplishing this for the present zephyr models is that the alfvn wave reflection is nontrivially coupled to the plasma state via the alfvn speed @xmath465 and solar wind speed @xmath5 ( and their gradients ) .
the radial dependence of these quantities can not be derived a priori from just the background magnetic field @xmath488 .
approximations for @xmath251 and @xmath252 such as those given by dmitruk et al .
( 2002 ) should be tested further to evaluate ways of estimating the amount of reflection without having to solve the full non - wkb equations .
we also intend to use the methodology developed here to model the winds of other kinds of stars on the cool side of the hertzsprung - russell ( h - r ) diagram .
similar ideas are beginning to be applied to younger and older stars with solar - type winds ( e.g. , airapetian et al .
2000 ; falceta - gonalves et al . 2006
; suzuki 2007 ; holzwarth & jardine 2007 ) . a benefit of computing the alfvn wave evolution with a physically motivated damping rate ( like eq .
[ [ eq : dmit ] ] ) is that the artificial `` damping lengths '' used in the past are no longer needed .
many cool stars that exhibit observational evidence for winds have lower surface gravities than the sun , stronger x - ray emission ( i.e. , more coronal heating ) , and much larger mass loss rates .
these properties are likely to be causally linked , but no comprehensive and predictive models yet exist . in order to model cool - star winds with zephyr
we must have values for the input parameters that drive the atmospheric heating and wind acceleration .
for example , we must be able to predict the properties of waves at the photospheric lower boundary for a given star .
amplitudes can be estimated from existing models of flux - tube wave generation from turbulent convection .
figure 18 shows an example of how transverse kink - mode wave amplitudes in stellar photospheres ( for which we retain the symbol @xmath388 ) can be estimated from the peak value of the subphotospheric convective velocity @xmath566 .
musielak et al .
( 2000 ) computed a series of stellar interior models with a range of effective temperatures ( 2000 to 10000 k ) and surface gravities ( @xmath567 , 4 , and 5 , with @xmath568 in units of cm s@xmath569 ) and presented the maximum values of the ratio @xmath570 in their figure 1 . also , musielak & ulmschneider ( 2002 ) computed the photospheric transverse wave fluxes for the same grid of stellar models .
we rescaled both quantities into velocity units using a sound speed consistent with @xmath196 , a magnetic field strength consistent with the degree of pressure equipartition assumed by musielak et al . , and a photospheric density computed from the condition @xmath571 , where @xmath572 is the photospheric scale height ( proportional to @xmath573 ) and @xmath55 is the rosseland mean opacity interpolated from the same tables used in other parts of the zephyr code ( see also cranmer 2005b ) .
note that although the two velocities span many orders of magnitude they `` collapse '' into something close to a one - to - one power - law relation ( with @xmath574 ) , and that the solar value used in 8 coincides reasonably closely with the model curves .
another key input to cool - star wind models is an adequate description of the large - scale magnetic field .
observational determinations of the field geometries of rapidly rotating young stars have been made possible via zeeman - doppler imaging ( e.g. , donati et al . 1990 , 2003 ) , time - resolved x - ray spectroscopy ( hussain et al . 2005 ) , and combinations of these and related techniques .
however , not enough is known about how the magnetic flux tubes break up and penetrate the photosphere on spatial scales relevant to convective granulation and supergranulation .
the relationship between the star - averaged mean field strength and the stronger fields inferred for thin flux tubes is not yet well understood ( e.g. , saar 2001 ) .
ongoing improvements in the observations make possible more comprehensive and predictive models .
we anticipate that phenomenological approaches like those described in this paper can contribute to substantial progress in our understanding of magnetic activity , coronal heating , and wind mass loss in a wide range of stars .
the authors would like to thank wolfgang kalkofen , jonathan slavin , eugene avrett , rudy loeser , and takeru suzuki for valuable discussions .
this work was supported by the national aeronautics and space administration ( nasa ) under grants nng04ge77 g , nnx06ag95 g , and nag5 - 11913 to the smithsonian astrophysical observatory .
this research made extensive use of nasa s astrophysics data system .
peter , h. , gudiksen , b. v. , & nordlund , .
2004 , in cool stars , stellar systems , and the sun , 13th cambridge workshop , ed .
f. favata , g. hussain , & b. battrick , ( noordwijk , the netherlands : esa ) , esa sp-560 , 59 | we present a series of models for the plasma properties along open magnetic flux tubes rooted in solar coronal holes , streamers , and active regions .
these models represent the first self - consistent solutions that combine : ( 1 ) chromospheric heating driven by an empirically guided acoustic wave spectrum , ( 2 ) coronal heating from alfvn waves that have been partially reflected , then damped by anisotropic turbulent cascade , and ( 3 ) solar wind acceleration from gradients of gas pressure , acoustic wave pressure , and alfvn wave pressure .
the only input parameters are the photospheric lower boundary conditions for the waves and the radial dependence of the background magnetic field along the flux tube .
we have not included multifluid or collisionless effects ( e.g. , preferential ion heating ) which are not yet fully understood . for a single choice for the photospheric wave properties ,
our models produce a realistic range of slow and fast solar wind conditions by varying only the coronal magnetic field .
specifically , a two - dimensional model of coronal holes and streamers at solar minimum reproduces the latitudinal bifurcation of slow and fast streams seen by _ ulysses . _
the radial gradient of the alfvn speed affects where the waves are reflected and damped , and thus whether energy is deposited below or above the parker critical point . as predicted by earlier studies , a larger coronal `` expansion factor ''
gives rise to a slower and denser wind , higher temperature at the coronal base , less intense alfvn waves at 1 au , and correlative trends for commonly measured ratios of ion charge states and fip - sensitive abundances that are in general agreement with observations .
these models offer supporting evidence for the idea that coronal heating and solar wind acceleration ( in open magnetic flux tubes ) can occur as a result of wave dissipation and turbulent cascade . |
the search for exotic quantum states in two - dimensional spin systems with frustrated interactions has been of great recent interest.@xcite among various materials studied recently , @xmath10 layered copper oxyhalides ( cux)a@xmath11b@xmath12o@xmath13 offer an interesting family of frustrated magnets with rich variety of behaviors . in this family , each magnetic cux layer ( with x = cl ,
br ) is sandwiched by nonmagnetic layers , forming an ideal two - dimensional structure . in each layer , the cu@xmath14 ions form a square lattice of @xmath10 spins , and a competition between ferromagnetic and antiferromagnetic interactions is anticipated from the small curie - weiss temperatures ( relative to other characteristic energy scales ) commonly observed in this family .
extensive experimental investigations have uncovered a collective singlet ground state with a spin gap in ( cucl)lanb@xmath0o@xmath1 ( refs . ) , a collinear stripe magnetic order in ( cubr)lanb@xmath0o@xmath1 ( ref . ) , and a magnetization plateau at @xmath15 of the saturated moment in ( cubr)sr@xmath0nb@xmath160@xmath17 ( ref . ) .
furthermore , chemical substitution was used to observe quantum phase transitions between magnetic and non - magnetic ground states.@xcite it would be interesting to determine what kind of spin models can capture the wide variety of physics in this family . a frustrated @xmath18-@xmath19 model on the square lattice with ferromagnetic @xmath18 and antiferromagnetic @xmath19 , as initially postulated for this family , does not seem to exhibit a spin gap or a magnetization plateau.@xcite a recent study by tassel _ et al._@xcite has proposed a remarkable microscopic structure in ( cucl)lanb@xmath0o@xmath1 . the x - ray and neutron diffraction studies have identified a considerable distortion of the cu - cl bonds ; consequently , the unit cell is doubled along @xmath20 and @xmath21 axes , as shown in fig .
[ fig : lattice](a ) . based on the obtained crystal structure
, a density functional calculation was carried out to construct the microscopic spin model .
the dominant interaction was then found to be the fourth - neighbor antiferromagnetic exchange coupling @xmath22 , which pairs spins into dimer singlets shown as thick lines in fig .
[ fig : lattice](a ) .
the calculation also suggested that the next leading couplings are ferromagnetic couplings ( denoted by @xmath4 and @xmath5 in this paper ) between these dimers , which are shown by solid and broken lines in fig .
[ fig : lattice](a ) . remarkably , the resultant interaction network has the structure of the distorted shastry - sutherland lattice@xcite as shown in fig .
[ fig : lattice](b).@xcite in this paper , we study a spin-@xmath23 model of coupled dimers on the distorted shastry - sutherland lattice , as the simplest starting point for understanding the interplay of competing interactions in ( cucl)lanb@xmath0o@xmath1 and related compounds .
we consider the heisenberg model in a magnetic field , @xmath24 where @xmath25 runs over all the bonds in fig .
[ fig : lattice](b ) , and @xmath26 , @xmath4 , and @xmath5 for diagonal , horizontal and vertical bonds , respectively .
we are primarily concerned with the case of antiferromagnetic @xmath2 and ferromagnetic @xmath3 ; however , results for other signs of @xmath4 and @xmath5 are presented alongside for comparison .
a common viewpoint for dealing with coupled dimer systems is to regard the triplet excitation at each dimer as a particle ( called a `` triplon '' ) and to describe the system as a bose gas of such particles.@xcite a notable feature of the shastry - sutherland lattice is the strong suppression of the triplon hopping due to frustration.@xcite in the antiferromagnetic case with @xmath2 and @xmath27 , the localized nature of the triplons gives rise to various fractional plateaux in the magnetization process,@xcite which are experimentally observed@xcite in srcu@xmath0(bo@xmath16)@xmath0 . frustration also exist in the case of our interest , with ferromagnetic @xmath3 .
the consequence of frustration in the ferromagnetic case has not been addressed in previous studies . o@xmath1 .
thick solid ( @xmath22 ) , thin solid ( @xmath4 ) , and broken ( @xmath5 ) lines indicate the three major couplings revealed in the electronic structure calculation of ref . .
other interactions @xmath28 and @xmath29 shown by dotted lines will also be considered in sec .
[ sec : compound ] .
( b ) topologically equivalent shastry - sutherland picture of the three major couplings @xmath22 , @xmath4 , and @xmath5 . a dotted green plaquette indicates a unit cell .
the four sublattices are labeled as @xmath30 , @xmath31 , @xmath32 , and @xmath33 .
, title="fig:",scaledwidth=40.0% ] + o@xmath1 .
thick solid ( @xmath22 ) , thin solid ( @xmath4 ) , and broken ( @xmath5 ) lines indicate the three major couplings revealed in the electronic structure calculation of ref . .
other interactions @xmath28 and @xmath29 shown by dotted lines will also be considered in sec .
[ sec : compound ] .
( b ) topologically equivalent shastry - sutherland picture of the three major couplings @xmath22 , @xmath4 , and @xmath5 . a dotted green plaquette indicates a unit cell .
the four sublattices are labeled as @xmath30 , @xmath31 , @xmath32 , and @xmath33 .
, title="fig:",scaledwidth=37.0% ] our main results for the ferromagnetic case @xmath3 are as follows .
when @xmath6 , the system stays in a dimer singlet phase with a finite spin gap .
this state is adiabatically connected to the decoupled dimer limit @xmath7 .
the magnetization process of this phase depends crucially on the spatial anisotropy of the inter - dimer couplings : the magnetization shows a jump and a smooth increase for weak and strong anisotropy , respectively , after the spin gap closes at a certain magnetic field . when @xmath8 or @xmath9 , the spin gap of the dimer singlet phase closes , and quantum phase transitions to various magnetically ordered phases ( ferromagnetic , collinear stripe , and spiral ) occur .
it is demonstrated that quantum fluctuations split the classical degeneracy of different spiral ground states .
the detailed phase diagrams of the classical and quantum models are constructed .
these results are based on a strong - coupling expansion for weak @xmath4 and @xmath5 , exact diagonalization for small clusters , and a schwinger boson mean field theory .
using the existing data on the magnetization process and the triplon bandwidth in ( cucl)lanb@xmath0o@xmath1 , and comparing with the corresponding theoretical results , we provide a consistency check for the appropriateness of the proposed spin model .
the rest of the paper is organized as follows . in sec .
[ sec : phase ] , we present the classical and quantum phase diagrams of the model and summarize the main results of the paper . in sec .
[ sec : classical ] , we discuss the details of the computations for the classical model and the resulting ground states . in sec . [
sec : sc ] , we perform the strong - coupling expansion of the quantum model to derive an effective low - energy hamiltonian . this effective model will be very useful in determining the global phase diagram of the quantum model and in estimating relevant physical quantities . in sec .
[ sec : ed ] , the numerical analysis of the quantum model based on the exact diagonalization is presented .
some parts of the phase boundaries are determined quite accurately within this approach . in sec .
[ sec : schwingermft ] , the schwinger boson mean field theory analysis of the model is reported .
these three techniques are collectively used to understand the quantum phase diagram . in sec .
[ sec : compound ] , we compare these theoretical results on the quantum model with experimental data on ( cucl)lanb@xmath0o@xmath1 .
we conclude the paper in sec .
[ sec : discussion ] . appendix [ app : exactgs ] explains an exact solution available in the isotropic case @xmath34 . in appendix [ app :
hartree ] , the hartree variational approach is discussed . in appendix
[ app : saturation ] , the saturation field in the magnetization process is determined exactly by considering a single - magnon excitation from the fully polarized state .
green dashed lines at @xmath35 , @xmath36 , @xmath34 are guides for eyes . , scaledwidth=40.0% ] in this section we present the classical and quantum phase diagrams of the model in figs .
[ fig : phase_cl ] , [ fig : phase_ed ] , and [ fig : phase_sb ] , and summarize the main results of the paper .
we assume @xmath2 and @xmath37 throughout the paper . with the assumption of @xmath2 , frustration at the classical level
occurs only when @xmath38 .
when we discuss the isotropic case @xmath34 , we denote these two parameters by @xmath39 .
the classical spin model is defined by replacing all the spin operators in eq . by classical @xmath40 vectors of length @xmath41 .
assuming @xmath42 for simplicity , the ground state phase diagram in fig .
[ fig : phase_cl ] is determined exactly for different signs of @xmath4 and @xmath5 .
the details of our analysis will be presented in sec .
[ sec : classical ] .
below we summarize the main characteristics of each phase . [ [ collinear - stripe - phases . ] ] collinear stripe phases .
+ + + + + + + + + + + + + + + + + + + + + + + + these phases appear naturally in the unfrustrated regions @xmath43 and penetrate into some parts of the frustrated regions @xmath38 .
the spin configuration has a propagation vector of @xmath44 or @xmath45 , where @xmath46 and @xmath47 are defined in fig .
[ fig : lattice](b ) . in the ground state with @xmath48 ,
for example , the system forms stripes of up or down spins running along the @xmath49 direction , while up and down alternate in the @xmath50 direction .
[ [ nel - phase . ] ] nel phase .
+ + + + + + + + + + + this phase appears for large positive @xmath51 and @xmath52 , where there is a minimal effect from the @xmath22 coupling .
the spin configuration is collinear and has a propagation vector of @xmath53 , as usual for the square lattice antiferromagnet .
[ [ ferromagnetic - fm - phase . ] ] ferromagnetic ( fm ) phase
. + + + + + + + + + + + + + + + + + + + + + + + + + this phase appears for large negative @xmath51 and @xmath52 , where ferromagnetic @xmath4 and @xmath5 dominate over antiferromagnetic @xmath22 .
the spins are all aligned in the same direction and have a propagation vector of @xmath54 .
[ [ spiral - phases . ] ] spiral phases .
+ + + + + + + + + + + + + + these phases appear when the magnitudes of @xmath22 , @xmath4 , and @xmath5 are comparable .
this is the case where the effect of the frustration is most prominent .
the ground state is a coplanar spiral state with an incommensurate propagation vector , as originally found in the isotropic case @xmath34 in ref . .
a notable feature of the spiral phase is that two kinds of spiral ground states are degenerate , apart from the trivial degeneracy associated with the global @xmath40 rotation of spins .
one ground state is the `` @xmath49-spiral '' in fig .
[ fig : spiral_qx](c ) , where spins rotate by a uniform angle @xmath55 in the @xmath49 direction and alternating angles @xmath56 in the @xmath50 direction .
the other is the `` @xmath50-spiral '' , which is defined by interchanging the roles of @xmath49 and @xmath50 directions .
remarkably , in the anisotropic case @xmath57 , the @xmath49- and @xmath50-spirals are not related to each other by any symmetry operation .
their degeneracy therefore comes from the particular geometry of the lattice .
we note that the classical phase boundaries in fig .
[ fig : phase_cl ] are symmetric with respect to the sign flips @xmath58 .
this is because the signs of @xmath4 and @xmath5 can both be flipped by reversing the spins on the @xmath30 and @xmath33 sites of all units cells , as seen in fig.[fig : lattice](b ) .
this transformation is _ not _ allowed in the quantum case , since the simultaneous reversal of all the spin components @xmath59 changes the commutation relations of spin operators .
the quantum phase diagram presented next therefore depends on the signs of @xmath4 and @xmath5 .
the quantum model with @xmath10 has been studied intensively in the isotropic antiferromagnetic case @xmath60 .
a remarkable feature of this model is that for @xmath42 and @xmath61 , the ground state of the model is exactly given by the product of dimer singlets@xcite @xmath62 where @xmath63 labels a dimer and @xmath64 is the singlet state on the dimer .
this solution can be extended to the ferromagnetic region @xmath65 , as described in appendix [ app : exactgs ] .
a notable difference between the ferromagnetic- and antiferromagnetic-@xmath39 cases occurs when a magnetic field @xmath66 is applied . in the antiferromagnetic-@xmath39 case ,
the magnetization process shows various fractional plateaux,@xcite which can be viewed as density wave formations of triplon excitations . in the ferromagnetic-@xmath39 case
, the magnetization process shows a jump from the dimer singlet state to the fully saturated state , as in fig .
[ fig : mag_h](a ) . as will be discussed in sec .
[ subsec : sc2 ] , this can be viewed as a consequence of the phase separation of triplons .
case , determined by the strong coupling expansion ( sec .
[ sec : sc ] ) and exact diagonalization ( sec .
[ sec : ed ] ) analyses .
green dashed lines at @xmath35 , @xmath36 , @xmath34 are guides for the eye .
the dimer singlet ( ds ) phase is divided into three regions , i , ii , and iii , which are characterized by the magnetization processes in fig . [ fig : mag_h ] .
red broken lines around the origin indicate the region boundaries determined from the the first - order effective hamiltonian ( see fig .
[ fig : phase_sc1 ] ) .
the square and circular symbols are based on exact diagonalization for the number of spins , @xmath67 .
the three different symbol sizes are in the order of @xmath68 .
the classical ferromagnetic phase boundary is superposed on the square symbols , showing good agreement .
the diamond symbols are the boundaries between the ds - ii regions and the stripe phases , determined by the fidelity susceptibility analysis in fig .
[ fig : fid ] .
narrow spiral phases may appear between the ds - i region and the stripe phases for large @xmath69 or @xmath70 .
for this reason , the ds - stripe phase boundaries ( diamond symbols ) are not calculated beyond @xmath51 or @xmath71 . ,
scaledwidth=40.0% ] of the saturated moment @xmath72 .
, scaledwidth=50.0% ] our major interest is in the ferromagnetic case of @xmath3 with spatially anisotropic couplings . for comparison
, we will also include some results for other signs of @xmath4 and @xmath5 .
the ground state phase diagrams are presented in figs .
[ fig : phase_ed ] and [ fig : phase_sb ] .
figure [ fig : phase_ed ] is based on a strong - coupling expansion for weak @xmath4 and @xmath5 ( sec .
[ sec : sc ] ) and exact diagonalization for finite clusters ( sec .
[ sec : ed ] ) .
when @xmath6 , the system stays in a dimer singlet phase with a finite spin gap , which is adiabatically connected to the decoupled - dimer limit @xmath7 .
this phase is split into three regions ( i , ii , and iii ) , in terms of the behaviors of the magnetization processes , as shown in fig .
[ fig : mag_h ] . in the ds - i region
, the magnetization @xmath73 shows a jump as in fig .
[ fig : mag_h](a ) . in the ds - ii region
, the magnetization @xmath73 smoothly increases towards saturation after the spin gap closes at a certain magnetic field , as in fig .
[ fig : mag_h](b ) . finally , in the ds - iii region , the magnetization process shows a plateau at @xmath74 of the saturated moment @xmath72 ( and more plateaux can appear near the isotropic case@xcite @xmath34 ) . when @xmath69 or @xmath75 , the spin gap of the dimer singlet phase closes , leading to various magnetically ordered states .
the characteristics of the ferromagnetic and collinear stripe phases are rather similar to those in the classical model .
the transition lines from the dimer singlet phase to these two phases are determined by exact diagonalization with relatively good accuracy ( square and diamond symbols in fig .
[ fig : phase_ed ] ) . ) for @xmath76 .
solid lines indicate second - order transitions , while broken lines indicate first - order transitions .
the isolated dimer singlet ( ds ) state is found for small @xmath77 , while the ferromagnetic state stabilizes for larger @xmath8 and @xmath78 .
the stripe state expands from the classical case to fill @xmath8 or @xmath79 .
the spiral state exists in the middle .
the @xmath49-spiral appearing for @xmath80 has an incommensurate long - range order in the @xmath49 direction as in fig .
[ fig : spiral_qx ] .
the @xmath50-spiral similarly appears for @xmath81 .
the dotted line separates the two spiral phases . ] the exact diagonalization for small systems , however , is not adequate to find incommensurate spiral phases that may intervene between the stripe , dimer singlet , and ferromagnetic phases . in order to get insight into this question
, we performed a schwinger boson mean - field analysis ( sec .
[ sec : schwingermft ] ) . in this approach ,
the spin magnitude @xmath41 can be varied freely , and one can discuss how the classical phase diagram in the limit @xmath82 ( fig .
[ fig : phase_cl ] ) changes as quantum fluctuations are gradually taken into account .
the phase diagram for @xmath76 is shown in fig . [
fig : phase_sb ] . in this result , we find that the spiral phases do appear in some parts of the phase diagram , while the presence of various magnetically ordered phases are somewhat exaggerated , as expected in mean - field methods .
providing an interesting difference from the classical case , quantum fluctuations split the classical degeneracy of the two spiral states , favoring the @xmath49- or @xmath50-spirals for @xmath80 and @xmath81 , respectively .
although the mean - field method is not useful in discussing the precise locations of the phase boundaries , it is natural to expect that these spirals appear in narrow regions between the ds - i region and the stripe phases in fig .
[ fig : phase_ed ] , particularly for large @xmath69 or @xmath70 . in experiments on ( cucl)lanb@xmath0o@xmath1 ,
the magnetization increases smoothly beyond a critical magnetic field until it reaches the saturated value.@xcite this suggests that , at least within the @xmath22-@xmath4-@xmath5 model , the ground state should belong to the ds - ii region as described in fig .
[ fig : phase_ed ] and there should be substantial anisotropy in @xmath4 and @xmath5 . considering the magnetization data@xcite and the measured energy range @xmath83 of the triplon excitations,@xcite and comparing them with theoretical results , we find that within the @xmath22-@xmath4-@xmath5 model , one of @xmath4 and @xmath5 may be ferromagnetic while the other may be antiferromagnetic .
however , depending on how to interpret the existing experimental data , the energy range of the triplon excitations can be larger ; if this is the case , @xmath4 and @xmath5 can be both ferromagnetic .
we plot the triplon dispersions for the cases of ( a ) ferromagnetic @xmath84 and antiferromagnetic @xmath85 and ( b ) both ferromagnetic @xmath86 ( fig .
[ fig : band_jjxjy ] ) . these plots can be used to test the @xmath22-@xmath4-@xmath5 model further and to determine the signs of @xmath4 and @xmath5 , when more detailed information of the triplon excitations is provided from experiments .
we also discuss the effects of other exchange couplings @xmath28 and @xmath29 shown in fig .
[ fig : lattice](a ) , which are also contained in the model of ref . .
these couplings induce oscillating behaviors in the triplon dispersions , which can be used as fingerprints of their existence .
the details of these analyses are presented in sec .
[ sec : compound ] .
in the @xmath49 direction ( `` @xmath49-spiral '' ) . the number at each site indicates the angle ( in the @xmath87 plane ) of the spin , where @xmath55 and @xmath88 are given by eq . .
, scaledwidth=50.0% ] in this section we describe the exact solution for the classical ground state .
the phase diagram is presented in fig .
[ fig : phase_cl ] , with the phases summarized in sec .
[ subsec : classical ] . for simplicity
, we restrict ourselves to the case of @xmath42 and @xmath89 .
the case of @xmath2 and @xmath3 can be treated in parallel by applying the spin reversal transformation on the @xmath30 and @xmath33 sites , as explained in the last paragraph of sec .
[ subsec : classical ] . in the case of @xmath43 ,
the system is not frustrated and the ground states are naturally determined as collinear stripe states . to find the ground state of the classical model with @xmath90
, we decompose the hamiltonian into triangular parts and solve a single - triangle problem shown in fig .
[ fig : spiral_qx](a ) . here
the diagonal coupling @xmath22 is divided by two , since it is shared by two neighboring triangles . by rewriting the three - spin energy @xmath91 as @xmath92 we find that the ground state of @xmath91 is obtained by minimizing the length of the vector @xmath93 .
when the vector lengths @xmath94 , @xmath95 , and @xmath96 have comparable magnitudes , the three vectors can form a triangle so that the above sum vanishes , as shown in fig .
[ fig : spiral_qx](b ) .
this leads to a coplanar spin configuration as shown in fig .
[ fig : spiral_qx](a ) , where the spins rotate counterclockwise by @xmath55 and then by @xmath88 when moving between sites @xmath97 .
another configuration where the spins instead rotate clockwise by the same angles also gives a ground state . here
@xmath55 and @xmath88 satisfy the relations @xmath98 which are solved as [ eq : qxqy ] @xmath99 using a ground state of @xmath91 locally on every triangle , one can construct a ground state of the whole lattice as in fig . [
fig : spiral_qx](c ) . in this state ,
the spins rotate by a uniform angle @xmath55 in the @xmath49 direction , and by alternating angles @xmath56 in the @xmath50 direction .
we call this state the `` @xmath49-spiral . ''
similarly , one can construct the `` @xmath50-spiral '' , where the spins rotate by a uniform angle @xmath88 in the @xmath50 direction and by alternating angles @xmath100 in the @xmath49 direction . as explained in sec .
[ sec : phase ] , the degeneracy of these states does not result from symmetry , and thus it should be regarded as a consequence of the particular geometry of the lattice .
when @xmath101 ( with @xmath90 ) , the three spins of @xmath91 can no longer form a triangle as in fig .
[ fig : spiral_qx](b ) . instead , @xmath102 and @xmath103
align antiparallel to @xmath104 , forming a collinear configuration with @xmath105 .
this leads to a nel ground state of the whole lattice .
similarly , for @xmath106 and @xmath107 , one obtains collinear stripe ground states with propagation vectors @xmath48 and @xmath45 , respectively .
the phase diagram obtained above for the antiferromagnetic quadrant @xmath108 can be mapped onto the ferromagnetic quadrant @xmath3 by applying the spin reversal mentioned in the last paragraph of sec .
[ subsec : classical ] .
the nel ground state maps onto the ferromagnetic state .
the @xmath48 stripe state maps onto the @xmath45 stripe state . a spiral state with @xmath109 maps to a spiral state with @xmath110 .
these arguments complete the phase diagram for all possible signs of @xmath4 and @xmath5 , as shown in fig .
[ fig : phase_cl ] .
in this section we analyze the spin-@xmath23 model by means of a strong coupling expansion.@xcite . in this approach , we start from the limit @xmath7 , where dimers are decoupled from each other .
we then perturbatively include the effects of @xmath4 and @xmath5 and derive an effective hamiltonian of the model . in sec .
[ subsec : heff ] .
we obtain the effective hamiltonian up to second order in @xmath4 and @xmath5 .
then , in secs . [ subsec : sc1 ] and [ subsec : sc2 ] , we analyze the first- and second - order effective hamiltonians , respectively , to deduce the physical properties of the original hamiltonian @xmath111 .
the first - order hamiltonian is exactly equivalent to the xxz model on the square lattice . using the known results on the xxz model , we map out a qualitative phase diagram of the model .
the second - order hamiltonian provides more accurate estimations of physical quantities and phase boundaries than the first - order one .
the obtained phase boundaries show a remarkable agreement with the exact diagonalization result of sec .
[ sec : ed ] .
the spin gap and the triplon band width derived in sec .
[ subsec : sc2 ] will be used to fit the experimental data of ( cucl)lanb@xmath0o@xmath1 in sec .
[ sec : compound ] . and ( b ) @xmath112 , and a three - body interaction term ( c ) @xmath113 . in each process , @xmath114 acts first on @xmath115 and @xmath116 , as indicated by the dotted lines , to yield the excited state .
it then acts on @xmath115 and @xmath117 to yield the final state of the process .
, scaledwidth=50.0% ] we start from the limit @xmath7 , where dimers are decoupled from each other .
we derive an effective hamiltonian of the model by perturbatively including the effects of @xmath4 and @xmath5 .
we label each dimer by its center position @xmath63 as in fig .
[ fig : lattice](b ) ; the two spins on the dimer are then denoted by @xmath118 and @xmath119 . when @xmath42 , the eigenstates of each dimer consist of a singlet @xmath120 and a triplet @xmath121 , where @xmath122 represents the @xmath123 component of the total spin , @xmath124 .
when a magnetic field @xmath125 is applied , the degeneracy of the triplet is split , with @xmath126 having the lowest energy of the three .
then it is useful to focus on the low - energy sector of the hilbert space which consists of @xmath120 and @xmath127 .
we apply degenerate perturbation theory to derive an effective hamiltonian in this restricted hilbert space . specifically , we adjust the magnetic field to @xmath128 so that the singlet @xmath120 and the up - polarized state @xmath127 become degenerate .
the unperturbed hamiltonian is @xmath129.\ ] ] the remaining part of the hamiltonian is @xmath130,\ ] ] where @xmath131 runs over all the sites .
we treat @xmath114 as a perturbation , assuming that the coefficient of each term is sufficiently smaller than @xmath22 .
note that the coefficient of the zeeman term in @xmath114 is @xmath132 ( _ not _
@xmath66 ) since the magnetic field of magnitude @xmath22 is included in @xmath133 ; therefore , the perturbation theory is most accurate around the intermediate magnetic field @xmath128 .
the ground state of @xmath133 is macroscopically degenerate with a degeneracy of @xmath134 , where @xmath135 is the number of unit cells in the system .
let @xmath136 be the subspace of the hilbert space spanned by these states .
regarding @xmath127 and @xmath120 on each dimer @xmath63 as the presence and the vacancy of a particle ( `` triplon '' ) , we treat the system as a hard - core boson gas on the square lattice of dimer centers . on each dimer @xmath63
, we define a creation operator @xmath137 and a number operator @xmath138 .
degenerate perturbation theory up to second order yields an effective hamiltonian @xmath139 where @xmath140 is the ground state energy of @xmath133 , and @xmath141 and @xmath142 are the first- and second - order contributions , respectively .
the first - order contribution @xmath141 is calculated by projecting @xmath114 onto the ground - state manifold @xmath136 of @xmath133 : @xmath143 \notag\\ & - ( h - j ) \sum_{{\bm{r}}}n_{{\bm{r}}},\end{aligned}\ ] ] where @xmath144 runs over all the nearest - neighbor pairs of dimers , and @xmath145 the @xmath4 and @xmath5 couplings produce the triplon hopping and interaction terms between nearest - neighbor dimers .
the two couplings contribute with opposite signs to the hopping term , so the coefficient is proportional to @xmath146 .
this leads to a suppression of the boson hopping near the isotropic case @xmath34 , which is a remarkable feature of the shastry - sutherland lattice.@xcite the interaction term is repulsive or attractive depending on the sign of @xmath147 .
the zeeman term @xmath132 in @xmath114 plays the role of the chemical potential of the hard - core boson gas .
since this term is diagonal in the particle number basis @xmath148 , it contributes only to the first - order hamiltonian @xmath141 . to calculate the second - order contribution @xmath142 , we take into account various processes where a state in @xmath136 is virtually promoted to an excited state of @xmath133 by the operation of @xmath114 , and then comes back to @xmath136 by @xmath114 again .
such virtual processes give rise to correlated hopping terms as well as three - body interaction terms , as illustrated in fig .
[ fig : virtual ] .
the resulting form of @xmath142 is here , the vectors @xmath150 and @xmath151 are defined as in fig . [ fig : lattice ] , and @xmath152 are defined via the relations @xmath153 and @xmath154 .
@xmath155 and @xmath156 refer to the sets of dimers oriented in the directions of @xmath150 and @xmath151 , respectively . by setting @xmath34 ,
the effective hamiltonian presented here coincides with the one derived by momoi and totsuka.@xcite by identifying the hard - core boson operators with the spin-@xmath23 operators via @xmath157 and @xmath158 , the first - order effective hamiltonian @xmath141 is equivalent to the xxz model on the square lattice , @xmath159 - \tilde{h } \sum_{{\bm{r}}}s^z_{{\bm{r}}},\ ] ] with @xmath160 note that this model has a higher symmetry than the original model ; the unit cell has been reduced from two dimers to a single dimer .
the phase diagram of this model has been studied in detail in the literature.@xcite by changing @xmath161 from large negative to large positive values , the ground state of @xmath141 changes between the fully down - polarized ( @xmath162 ) and fully up - polarized ( @xmath163 ) states . in the original model
, this corresponds to a change between the dimer singlet state with magnetization @xmath164 and the fully polarized state with @xmath165 .
the process of this change can be smooth or sudden ; the details depend on the value of the xxz anisotropy @xmath166 of @xmath141 , and are classified into three cases ( i)-(iii ) , described below .
we note that the sign of @xmath167 is not crucial in the present argument , since it can be flipped by applying a gauge transformation to @xmath141 or by exchanging the roles of @xmath4 and @xmath5 in @xmath111 .
\(i ) ferromagnetic ising case @xmath168 . at @xmath169 ,
the fully up- and down - polarized states are degenerate in the ground state of @xmath141 .
therefore , the magnetization of the effective xxz model shows a jump from @xmath162 to @xmath170 as @xmath161 changes its sign from negative to positive .
this corresponds , in the original model , to a jump of the magnetization @xmath73 at an intermediate magnetic field @xmath171 as in fig .
[ fig : mag_h](a ) .
\(ii ) xy case @xmath172 .
the ground state of @xmath141 at @xmath169 is given by a ferromagnetic or antiferromagnetic state in the @xmath87 plane for @xmath173 and @xmath174 , respectively .
this xy ordered state changes to a canted state by a magnetic field @xmath175 and the magnetization @xmath176 varies smoothly between @xmath177 and @xmath170 in the range @xmath178 .
therefore , in the original model , the magnetization @xmath73 increases smoothly from zero to the saturation in the field range between @xmath179 and @xmath180 , as in fig .
[ fig : mag_h](b ) .
the canted ferromagnetic and antiferromagnetic orders in the effective model give rise to transverse magnetizations in the original model .
specifically , in the field range @xmath181 , the transverse component shows a stripe order with a propagation vector of @xmath182 and @xmath45 , for @xmath183 and @xmath184 , respectively .
the connection between the magnetic orders of the effective and original models is understood most easily from a variational ground state approach , which is presented in appendix [ app : hartree ] .
\(iii ) antiferromagnetic ising case @xmath185 .
in contrast to the xy case above , the ground state of @xmath141 at @xmath169 is given by an antiferromagnetic state polarized along the @xmath123 axis , featuring gapped excitations .
when @xmath161 is increased above zero , the magnetization @xmath176 stays zero up to certain critical @xmath186 and then shows a jump in the transition to the canted antiferromagnetic state ( known as the spin - flopping process).@xcite a similar thing occurs when decreasing @xmath161 below zero .
therefore , in the original model , the magnetization process shows a plateau at @xmath74 of the saturation , accompanied by jumps at the edges , as in fig .
[ fig : mag_h](c ) .
the expressions of @xmath187 and @xmath188 are the same as the xy case ( ii ) above . model , derived from the first - order effective hamiltonian @xmath141 of the strong - coupling expansion . in the ds - i and ds - ii regions , the magnetization shows a jump and a smooth increase as in figs . [
fig : mag_h](a ) and ( b ) , respectively . in the ds - iii region , at least one magnetization plateau is expected to appear at @xmath74 of the saturated moment as in fig .
[ fig : mag_h](c ) .
more plateaux may be expected near the isotropic case.@xcite , scaledwidth=50.0% ] the present argument allows us to map out a qualitative phase diagram as in fig .
[ fig : phase_sc1 ] .
the dimer singlet phase appearing for @xmath189 are divided into three regions ( i , ii , and iii ) in terms of the behaviors of the magnetization processes shown in fig .
[ fig : mag_h ] .
the boundaries between the regions are given by @xmath190 from the above argument . in principle
, the effective hamiltonian @xmath141 can give reliable results for @xmath111 only when @xmath4 and @xmath5 are sufficiently smaller than @xmath22 . beyond this region ,
we nonetheless use @xmath141 to investigate possible instabilities of the dimer singlet state . by increasing @xmath8 and @xmath78 in the ds - i region
, the critical field @xmath171 goes to zero , indicating a transition to the ferromagnetic phase .
similarly , by increasing @xmath8 or @xmath78 in the ds - ii regions , @xmath179 goes to zero , leading to the collinear stripe phases .
the boundaries between the ferromagnetic and stripe phases are determined by the condition @xmath191 . the obtained phase diagram in fig .
[ fig : phase_sc1 ] qualitatively agrees with fig .
[ fig : phase_ed ] obtained by exact diagonalization .
however , some aspects of the phase diagram are not captured in this approach .
for example , the spiral and nel phases ( expected from the classical analysis of sec . [ sec : classical ] and the schwinger boson analyses of refs . and
[ sec : schwingermft ] ) do not appear in this approach .
these phases are beyond the scope of the present perturbative analysis .
in passing , we note that the phase diagram in fig .
[ fig : phase_sc1 ] can also be obtained using a hartree variational state approach , which is presented in appendix [ app : hartree ] .
we now move to the analysis of the second - order effective hamiltonian @xmath192 .
on one hand , this allows us to make more accurate estimations of physical quantities , such as the spin gap , compared to the first - order case above . on the other hand
, the new terms appearing in @xmath192 may open up possibilities of new phenomena . in the isotropic antiferromagnetic case @xmath27 , momoi and totsuka@xcite
have shown that correlated hopping processes @xmath193 induce the formation of bound states of two triplons . in the ferromagnetic case of our main interest ,
however , we will argue that the formation of bound states is rather unlikely , and that the second - order terms in @xmath192 do not change the essential physical properties of the system .
instead , the improved accuracy in the estimation of physical quantities allows us to make more quantitative comparison with the exact diagonalization results of sec .
[ sec : ed ] . and
@xmath194 in the ds - ii regions , where the magnetization process shows a smooth increase as in fig .
[ fig : mag_h](b ) . here
@xmath195 is the width between the starting and ending of the magnetization process , and @xmath196 is the width of the triplon dispersion .
the contours are shown only inside the pink dotted lines , where eq .
the lines of @xmath197 and @xmath198 correspond to transition lines to the ds - i region and the stripe phases , respectively .
the transition lines obtained from exact diagonalization analysis of sec .
[ sec : ed ] are also shown ( by the same symbols as in fig .
[ fig : phase_ed ] ) .
fit with the experimental data of ( cucl)lanb@xmath0o@xmath1 yields estimates of @xmath51 and @xmath52 , as indicated by the asterisks ( see sec .
[ sec : compound ] ) .
, scaledwidth=45.0% ] we first calculate the energy of a single - triplon excitation . by restricting to the single - triplon sector , we can neglect all the terms of the forms , @xmath199 , @xmath200 and @xmath193 .
the resulting hamiltonian contains only the chemical potential @xmath201 and the nearest - neighbor , second - neighbor , and third - neighbor hoppings , whose amplitudes are given respectively by @xmath202 , @xmath203 , and @xmath204 .
these parameters are given by @xmath205 the unit cell is again reduced to a single dimer ; however , this occurs only in a single - triplon sector .
this hamiltonian leads to a triplon dispersion , @xmath206 \\ & - 2 t ' ( \cos 2k_1 + \cos 2k_2 ) , \end{split}\ ] ] with @xmath207 . using @xmath208 and @xmath209 ,
this is rewritten as @xmath210 this dispersion has the minimum energy @xmath211 and the band width @xmath212 ( if @xmath213 ) .
the triplon condensation field @xmath187 in the ds - ii regions is determined by the condition @xmath214 , leading to @xmath215 the saturation field @xmath188 is calculated exactly in appendix [ app : saturation ] .
using @xmath216 in eq . ,
which is valid in the collinear stripe regions of the classical model , the width @xmath217 between @xmath187 and @xmath188 is determined as @xmath218 in fig .
[ fig : hc1 ] , we display the contour maps of @xmath219 and @xmath194 calculated in this manner .
since @xmath220 in the decoupled - dimer limit @xmath7 , the two parameters indicate to what extent the system is separated from this limit . however , the two parameters show slightly different behaviors .
the contours of @xmath219 and @xmath194 are approximately parallel to the lines @xmath221 ( or @xmath222 ) and @xmath34 , respectively .
the lines of @xmath197 corresponds to the transition between the ds - i and ds - ii regions .
the lines of @xmath223 corresponds to the quantum phase transitions into magnetically ordered states associated with the triplon condensation .
these boundaries agree well with the exact diagonalization results plotted together in fig .
[ fig : hc1 ] .
this means that our perturbative calculation likely gives a good estimation of @xmath187 up to rather large values of @xmath69 and @xmath70 .
next , we discuss the role of the correlated hopping processes @xmath193 , which appear in the second - order effective hamiltonian . in the isotropic antiferromagnetic case @xmath27 , momoi and totsuka@xcite
have shown that these processes induce the formation of bound states of two triplons .
namely , two triplons pair together and achieve a lower energy than two independent triplons . in this case , as the magnetic field is increased , the bound states condense before the triplons do , leading to a bond - nematic order.@xcite it is interesting to look for this possibility in the ferromagnetic case of our interest .
we performed exact diagonalization analysis of the original spin model on various points in the ferromagnetic case @xmath3 , and searched for the existence of bound states .
however , we found no indication of the formation of bound states .
this result can be interpreted from the study of schmidt _
et al_ .. @xcite these authors have analyzed a simple hard - core boson model on the square lattice containing a single - particle hopping @xmath202 , a correlated hopping @xmath224 , and the nearest - neighbor repulsion @xmath225 .
they have demonstrated numerically that the correlated hopping process enhances the tendency towards phase separation as well as boson pairing ; a jump of the magnetization [ fig .
[ fig : mag_h](c ) ] in the ds - i region indeed occurs as a result of the phase separation.@xcite it was shown@xcite that the correlated hopping process induces the boson paring only when a modest magnitude of @xmath226 is introduced . in the ferromagnetic region @xmath3 of the model ,
the nearest - neighbor interaction is attractive ( @xmath227 ) , and hence phase separation and a standard superfluid state are more likely to occur than the boson pair - condensed state .
, and @xmath228 used for exact diagonalization analyses . , scaledwidth=50.0% ] as a function of the magnetization @xmath73 for ( a ) @xmath229 and @xmath230 and ( b ) @xmath229 and @xmath231 . a concave curve in ( a ) leads to a jump in the magnetization as in fig .
[ fig : mag_h](a ) , while a convex curve in ( b ) leads to a smooth increase as in fig .
[ fig : mag_h](b ) .
broken lines indicate the variational estimates using the ansatz .
, title="fig:",scaledwidth=40.0% ] as a function of the magnetization @xmath73 for ( a ) @xmath229 and @xmath230 and ( b ) @xmath229 and @xmath231 . a concave curve in ( a ) leads to a jump in the magnetization as in fig . [ fig : mag_h](a ) , while a convex curve in ( b ) leads to a smooth increase as in fig .
[ fig : mag_h](b ) .
broken lines indicate the variational estimates using the ansatz .
, title="fig:",scaledwidth=40.0% ] versus @xmath232 for @xmath233 .
( b ) extrapolation of the peak position of @xmath234 into @xmath235 using eq . .
, title="fig:",scaledwidth=40.0% ] versus @xmath232 for @xmath233 .
( b ) extrapolation of the peak position of @xmath234 into @xmath235 using eq . .
, title="fig:",scaledwidth=40.0% ] the strong coupling expansion in the previous section has provided controlled analyses for small @xmath69 and @xmath70 . to investigate the ground - state phase diagram of the model in a wide parameter space ,
we have performed exact ( lanczs ) diagonalization calculations in finite - size clusters .
the shapes of the clusters are shown in fig .
[ fig : clusters ] , where @xmath68 is the total number of spins in the system . in every cluster ,
periodic boundary conditions are imposed at the edges .
the clusters with @xmath236 and @xmath237 are invariant under simultaneous operations of @xmath238 lattice rotation and interchange of @xmath4 and @xmath5 .
therefore , these clusters treat @xmath4 and @xmath5 equally , and are less biased in studying the ground - state properties than the clusters with @xmath239 , which do not have this invariance .
however , to investigate the size dependence of the calculated quantities , we analyze the @xmath239 clusters as well .
we first investigate the magnetization process in the dimer singlet phase .
the magnetization process is determined by calculating the ground - state energy @xmath240 as a function of the total magnetization @xmath241 as in fig .
[ fig : ene_m ] . to determine the magnetization @xmath73 for a given field @xmath66 from this figure , we first find the line of slope @xmath242 that touches the curve in such a way as to minimize the vertical axis intercept .
the magnetization is then found at the touching point of the line and the curve . a concave curve , as in fig .
[ fig : ene_m](a ) , leads to a jump of @xmath73 as in fig .
[ fig : mag_h](a ) , with @xmath243 a convex curve , as in fig .
[ fig : ene_m](b ) , leads instead to a smooth increase of @xmath73 as in fig .
[ fig : mag_h](b ) with @xmath244 a convenient way to judge whether the curve is concave or convex is to examine which of @xmath245 and @xmath187 defined above is smaller . by plotting @xmath245 and @xmath187 as functions of model parameters ( @xmath51 or @xmath52 ) and finding the crossing point , one can determine the boundary between the ds - i and ds - ii regions , as indicated by circular symbols in fig . [
fig : phase_ed ] . here , the data for @xmath236 , @xmath246 , and @xmath228 are shown by symbols of different sizes in order of @xmath68 .
the obtained boundaries show little dependence on the system size @xmath68 . in fig .
[ fig : ene_m ] , we also plot the estimate of the energy using the hartree variational ansatz .
we find that this ansatz can capture the qualitative shapes of the curves . as @xmath4 or @xmath5 increases ,
the spin gap of the system becomes smaller and eventually vanishes , leading to magnetically ordered states .
the transition points to the ferromagnetic state can be determined accurately by finding the level crossing between the singlet ground state and the fully polarized state .
these transition points are plotted by square symbols in fig .
[ fig : phase_ed ] .
the obtained boundary agrees very well with the classical ferromagnetic phase boundary.@xcite the transitions to the collinear stripe states can not be detected by a level crossing , and we instead calculate the fidelity susceptibility , as presented below .
the fidelity susceptibility@xcite shows a peak at the transition point , which we use to detect the phase boundary . since the phase boundaries are parallel to the line @xmath34 in the qualitative phase diagram of fig .
[ fig : phase_sc1 ] , it is efficient to probe the parameter space in the direction perpendicular to this line .
therefore we fix @xmath247 and change the parameter @xmath232 , where @xmath248 as defined in eq . .
the fidelity between two parameter points @xmath249 and @xmath250 is defined as an overlap between the ground states , @xmath251 when the two points are close enough , this quantity has an expansion @xmath252 which allows us to define the fidelity susceptibility@xcite @xmath253 the result for @xmath233 is presented in fig .
[ fig : fid](a ) .
we observe a peak in @xmath254 , which indicates a phase transition ; this peak shifts gradually and grows sharper with increasing @xmath68 .
the peak position @xmath255 in a finite - size system is expected to approach the transition point @xmath256 with increasing @xmath68 . in fig .
[ fig : fid](b ) , we extrapolate the data of @xmath255 into @xmath235 , assuming the scaling form@xcite @xmath257 \sim n_s^{-1/(2\nu)},\ ] ] with @xmath258 for an @xmath40 quantum phase transition , corresponding to triplon condensation .
the obtained stripe phase transition points are plotted by diamond symbols in fig .
[ fig : phase_ed ] .
the phase boundary crosses the vertical axis at @xmath259 .
when @xmath35 and @xmath260 , the present model is equivalent to a heisenberg antiferromagnet on the honeycomb lattice with spatially anisotropic exchange couplings .
quantum monte carlo analysis of such a model@xcite has given a precise estimate @xmath261 for the transition point .
our result roughly agrees with this , although it slightly overestimates the range of the dimer singlet phase .
the classical phase diagram in fig .
[ fig : phase_cl ] suggests that there can be a narrow spiral phase between the ferromagnetic and collinear stripe phases .
since incommensurate spin correlations existing in the spiral phase are difficult to treat in the small system sizes available to exact diagonalization , we do not analyze the existence of this phase within exact diagonalization .
instead , we analyze this phase using the schwinger - boson mean field theory in the next section .
= v = = = in this section , we analyze the heisenberg model introduced in eq .
using a schwinger boson mean - field theory.@xcite the mean - field method can cover a wide range of circumstances where our other methods may be limited . unlike the strong - coupling expansion , this method is not limited to small @xmath51 and @xmath52 .
the incommensurate spiral state can be easily found by this mean - field theory , in contrast with exact diagonalization .
furthermore , the theory can determine the splitting of the @xmath49- and @xmath50-spiral state degeneracy due to quantum fluctuations .
finally , one can control the strength of quantum fluctuations by varying the spin magnitude " @xmath41 , connecting the @xmath262 limit of semi - classical ordering to the low - spin limit of @xmath76 .
we consider the case @xmath42 for simplicity .
the schwinger boson representation of the spin on site @xmath131 is given by @xcite @xmath263 here , the local constraint of the form @xmath264 is imposed so that the spin quantum number @xmath41 is given by the number of bosons per site .
the @xmath265 spin labels run over @xmath266 , and @xmath267 with @xmath268 are the pauli matrices .
heisenberg terms are quartic in the schwinger bosons , so we apply a mean - field decoupling to obtain a quadratic hamiltonian .
the heisenberg term is given by @xmath269 to decouple this , we use the following pauli matrix product identities to decouple in an attractive channel for each possible sign of the interaction : @xmath270 here , @xmath271 is the totally antisymmetric tensor defined with @xmath272 .
this gives the relations for ferromagnetic and antiferromagnetic interactions , @xmath273 where we have defined the operators @xmath274 we enforce the boson number constraint by introducing a lagrange multiplier term @xmath275 we treat this on average by taking @xmath276 constant on each of the four sublattices .
we perform a mean - field decoupling of the terms given by ( [ eqn : sintermsofab ] ) .
for instance , @xmath277 with @xmath278 . assuming the translational invariance , there are eight @xmath31 mean fields , four for the nearest - neighbour interactions in the @xmath46 direction , four for the @xmath47 direction , and two @xmath30 mean fields for the diagonal antiferromagnetic interactions . following this
, we perform a fourier transform defined by @xmath279 @xmath280 labels a position in a unit cell as in fig . [
fig : lattice](b ) , so we rewrite the label @xmath281 .
@xmath135 is the number of unit cells in the system .
we have used the @xmath30 site position @xmath282 to label a unit cell , and @xmath283 to represent the position of each site relative to the @xmath30 site , @xmath284 at this point , we assume the chemical equivalence of the four sites in the unit cell , by taking @xmath285 the same for all sites in the lattice . in conjunction , we use an ansatz for the mean fields consistent with the choice of chemical potential .
namely , we take the @xmath46-direction @xmath31 to be equal @xmath286 , and do the same for the @xmath47-direction @xmath31 ( @xmath287 ) .
the two @xmath30 mean fields are between sites @xmath30 and @xmath33 ( @xmath288 , and @xmath31 and @xmath32 ( @xmath289 ) .
furthermore , we take all mean fields to be real .
we take the two @xmath30 mean fields to have equal magnitude , so that @xmath290 and @xmath291 gives two possible choices for the relative signs of a. with these assumptions , the mean field theory can still describe all the relevant phases expected to appear in the model .
we are left , in the case of a gapped dispersion , with the mean - field hamiltonian @xmath292 where @xmath293 the matrix @xmath294 is given by @xmath295 @xmath296 @xmath297 with @xmath298 , @xmath299 , and @xmath300 as defined above . the quadratic hamiltonian in ( [ eqn : hamilmf ] )
is diagonalized by the bogoliubov transformation @xmath301 , with @xmath302 defined in the same manner as @xmath303 : @xmath304 this is a canonical transformation , where the operators in @xmath302 preserve the bosonic commutation relations of the operators in @xmath303 .
these commutation relations are given by @xmath305 = \delta_{{{\ensuremath{\mathbf{k}}}}{{\ensuremath{\mathbf{k } } } } ' } \eta , \end{aligned}\ ] ] where the @xmath306 matrix @xmath307 can be written in @xmath308 blocks : @xmath309 the proper commutation relation are obtained by taking the columns of @xmath310 to be the eigenvectors @xmath311 ( with eigenvalues @xmath312 ) of @xmath313.@xcite here , @xmath314 labels the eight eigenvectors .
when the hamiltonian @xmath294 is positive definite , the corresponding spinon dispersions @xmath315 are gapped .
this leads to a disordered ground state . in the case of a gapless dispersion ,
@xmath316 for some @xmath317 , so that both @xmath318 and @xmath319 have a zero eigenvalue at the dispersion minimum @xmath320 .
condensation of such zero - energy bosons gives rise to a magnetically ordered state . to describe the condensation
, we replace the operators @xmath321 with macroscopic constant values , @xmath322 where @xmath323 . with the condensate contribution ,
the diagonalized hamiltonian is @xmath324 spin ordering is found at twice the spinon minimum wavevector , @xmath325 , governing relative spin orientation between unit cells .
given the diagonalized hamiltonian , we must solve for the mean - field values of @xmath326 and @xmath249 , as well as any condensate vectors @xmath327 and associated minimum wavevectors @xmath320 .
it turns out that the solution depends only on the total condensate density @xmath328 .
we have the mean - field equations @xmath329 which we solve self - consistently for a given set of parameters @xmath51 , @xmath52 , and @xmath41 , and the associated values of @xmath317 and @xmath320 . in the semi - classical limit @xmath82
, we recover the classical phase diagram shown earlier in fig . [
fig : phase_cl ] .
we present the phase diagram as a function of @xmath51 and @xmath52 for the cases of @xmath76 ( fig .
[ fig : phase_sb ] ) and @xmath330 ( fig .
[ fig : total0.15 ] ) . since a mean - field theory
is expected to underestimate quantum fluctuations , it can be instructive to look at spin values smaller than the actual case .
the schwinger boson mean - field theory finds several magnetically ordered phase , and one disordered phase .
the magnetically ordered phases are the ferromagnetic , spiral and stripe phases seen in the classical limit , while the disordered phase features isolated dimers .
below we summarize how each phase is described in the schwinger boson formalism . .
solid lines indicate second - order transitions , while dashed lines indicate first - order transitions .
the dotted line separates the two spiral phases .
the disordered dimer state ( ds ) expands significantly compared to the @xmath76 case in fig .
[ fig : phase_sb ] .
the stripe state expands moderately , pushing the ferromagnetic ( fm ) state to moderately larger @xmath4 and @xmath5 . in between these phases ,
the spiral state is found in a significantly reduced area . with an even smaller @xmath331 ,
the spiral state will have disappeared entirely . ] [ [ ferromagnetic - phase . ] ] ferromagnetic phase .
+ + + + + + + + + + + + + + + + + + + + in this state , the antiferromagnetic order parameter @xmath332 , while both ferromagnetic order parameters take their maximum values , @xmath333 .
the spinon minimum wavevector @xmath334 , as is the spin ordering wavevector , so that the spins are fully polarized throughout the entire lattice .
[ [ stripe - phase . ] ] stripe phase .
+ + + + + + + + + + + + + in this state , the ferromagnetic order parameter associated with the smaller of @xmath4 and @xmath5 is zero .
the ordering wavevector is @xmath335 , and a stripe magnetization pattern is found in the direction of the larger ferromagnetic interaction .
depending on the relative signs of the @xmath30 parameters , the spinon minimum wavevector is either @xmath48 or @xmath335 , for @xmath184 , both leading to the same ordered state of same energy .
[ [ spiral - phase . ] ] spiral phase .
+ + + + + + + + + + + + + this is an incommensurate magnetically ordered state with a nonzero ordering wavevector in the direction of smallest ferromagnetic interaction : @xmath336 for @xmath337 .
all of the mean - field parameters are nonzero .
there are two inequivalent spinon minimum wavevectors @xmath338 found in the spiral state .
they both lead to the same type of spiral order , described in sec .
[ sec : classical ] .
the relative sign of the two @xmath30 mean fields yields either the @xmath49-spiral or @xmath50-spiral ordering . we find that the classical degeneracy between @xmath49- and @xmath50-spiral ordering is broken by quantum fluctuations , determining the ordering direction . for @xmath184 , the @xmath50-spiral state has a lower energy , while the @xmath49-spiral state has lower energy for @xmath339 .
[ [ disordered - phase . ] ] disordered phase .
+ + + + + + + + + + + + + + + + + the disordered phase is gapped , with no boson condensate , and has no ferromagnetic correlations : @xmath340 .
the antiferromagnetic dimer order parameter @xmath341 , leaving a decoupled dimer state .
this is a mean field description corresponding to the dimer singlet state .
we consider the effect on the phase diagram of lowering spin from the large-@xmath41 semiclassical limit , by comparing the classical , @xmath76 , and @xmath330 phase diagrams shown in figs .
[ fig : phase_cl ] , [ fig : phase_sb ] and [ fig : total0.15 ] , respectively .
as @xmath41 decreases , the disordered phase appears and expands around @xmath7 , pushing out the magnetically ordered phases to larger @xmath4 and @xmath5 .
of particular notes is the shift of the ferromagnetic phase boundary to larger @xmath4 and @xmath5 in contradiction with the exact boundary for @xmath342 seen in appendix [ app : exactgs ] .
similarly , the stripe state boundary shifts to larger @xmath5 for @xmath184 , and larger @xmath4 for @xmath183 . between these aforementioned phase boundaries ,
we find the spiral state , which consequently gets pushed out to larger @xmath4 and @xmath5 .
this state shrinks as @xmath41 decreases , having almost disappeared at @xmath330 , as seen in fig .
[ fig : total0.15 ] .
in contrast with the case of the purely antiferromagnetic shastry - sutherland lattice,@xcite we see no other disordered states than the dimer one down to @xmath331 . in particular , there are no short - range ordered analogues of the ferromagnetic , stripe , or spiral ordered states .
in this section we compare our theoretical results with the experimental data on ( cucl)lanb@xmath0o@xmath1 .
the magnetization process of the compound shows a smooth increase between @xmath344 t and @xmath345 t,@xcite as in fig .
[ fig : mag_h](b ) .
therefore , the compound should be located inside the ds - ii regions in fig .
[ fig : phase_ed ] within the @xmath22-@xmath4-@xmath5 model . assuming the land factor @xmath346 , the above magnetic fields are translated into energy units as @xmath347 according to a recent inelastic neutron scattering experiment,@xcite the triplet excitations at @xmath348 range from approximately 1.2 to 3.0 mev .
good agreement of the lower bound of the excitations with @xmath187 suggests that the emergence of the magnetization is associated with the condensation of single - triplon excitations .
indeed , a bose - einstein condensation of some sort of bosonic excitations have been observed in the field range between @xmath349 and @xmath350.@xcite here we do not address the possibility of a condensation of bound magnons proposed in refs .
, since it does not seem to occur within the @xmath22-@xmath4-@xmath5 model ( except for the case @xmath108 ) as discussed in sec .
[ subsec : sc2 ] . below
we first discuss the experimental data within the @xmath22-@xmath4-@xmath5 model , using the second - order perturbative results of sec .
[ subsec : sc2 ]
. then we discuss possible effects of other exchange couplings @xmath28 and @xmath29 shown in fig .
[ fig : lattice](a ) , within the first - order perturbation theory . -@xmath4-@xmath5 model in the cases of ( a ) ferromagnetic @xmath351 and antiferromagnetic @xmath260 in eq . and ( b ) both ferromagnetic @xmath3 in eq . .
black solid and broken lines correspond to @xmath352 and @xmath353 , respectively .
these are plotted along a triangle connecting @xmath354 , @xmath355 , and @xmath356 in the brillouin zone .
blue horizontal dotted lines indicate the energy range of triplet excitations ( from @xmath357 to @xmath358 ) observed in ref . .
to compare with the magnetic layer of ( cucl)lanb@xmath0o@xmath1 in fig .
[ fig : lattice](a ) , we have replaced @xmath359 , where @xmath360 is the distance between neighboring cu@xmath14 ions .
, title="fig:",scaledwidth=50.0% ] -@xmath4-@xmath5 model in the cases of ( a ) ferromagnetic @xmath351 and antiferromagnetic @xmath260 in eq . and ( b ) both ferromagnetic @xmath3 in eq . .
black solid and broken lines correspond to @xmath352 and @xmath353 , respectively .
these are plotted along a triangle connecting @xmath354 , @xmath355 , and @xmath356 in the brillouin zone .
blue horizontal dotted lines indicate the energy range of triplet excitations ( from @xmath357 to @xmath358 ) observed in ref . .
to compare with the magnetic layer of ( cucl)lanb@xmath0o@xmath1 in fig .
[ fig : lattice](a ) , we have replaced @xmath359 , where @xmath360 is the distance between neighboring cu@xmath14 ions .
, title="fig:",scaledwidth=50.0% ] using the above data of the magnetization process and the triplet excitations , the two ratios calculated in fig .
[ fig : hc1 ] are given by @xmath361 these can be used to fix the values of @xmath362 , as shown by asterisks in fig .
[ fig : hc1 ] . adjusting @xmath22 to give @xmath188 in eq .
, we obtain the estimates @xmath363 another parameter set where the values of @xmath4 and @xmath5 are interchanged is also possible . in these naive estimates within the @xmath22-@xmath4-@xmath5 model , one of @xmath4 and @xmath5
is ferromagnetic while the other is antiferromagnetic .
this is in contrast to both ferromagnetic values @xmath3 predicted in the electronic structure calculation.@xcite we note , however , that the estimation depends on the interpretation of the experimental data .
in addition to the excitations ranging between @xmath364 and @xmath365 mev above , an earlier inelastic neutron scattering experiment@xcite also detected excitations centered around @xmath366 mev ( @xmath367 ) with small scattering intensities . in ref .
, these excitations are interpreted as bound states of triplets .
if we regard that these also originate from triplet excitations , the range @xmath83 is much larger . then in fig .
[ fig : hc1 ] , the estimates of @xmath362 can change , keeping the relation @xmath368 , to the case of ferromagnetic @xmath3 . for example , in the parameter set @xmath369 the triplon excitation energy in eq
. ranges up to @xmath370 .
it is expected that by increasing @xmath8 further , the energy upper bound would reach @xmath371 observed in experiment .
we do not discuss a larger-@xmath8 region , since our perturbative result would be less reliable there . to further discuss the consistency between the present model and ( cucl)lanb@xmath0o@xmath1
, it would be useful to look at the detailed shapes of the triplon dispersions .
here we plot the triplon dispersions using the perturbative result in eq . .
when more detailed information of the triplet excitations is provided from experiments , these plots can be used to test the @xmath22-@xmath4-@xmath5 model and to determine the signs of @xmath4 and @xmath5 .
equation contained only a single band , since the single - triplon hopping problem had the same periodicity as the square lattice ( of dimer centers ) . to discuss the physical excitations
, however , one needs to take into account the fact that there are two dimers in a unit cell of the original lattice , and to fold the brillouin zone in such a way as to identify @xmath372 , @xmath373 , and @xmath374 .
consequently , the first brillouin zone is given by a square ranging over @xmath375 .
after the folding , there are two dispersions expressed as @xmath376 these dispersions are plotted in fig .
[ fig : band_jjxjy ] for the parameter sets in eqs . and .
while fig .
[ fig : band_jjxjy](a ) shows a nearly symmetric shape centered around @xmath377 , fig .
[ fig : band_jjxjy](b ) shows a highly asymmetric shape .
-@xmath4-@xmath5-@xmath28-@xmath29 model .
compared to fig .
[ fig : band_jjxjy ] , the path @xmath378 is added , since the triplon dispersions are anisotropic in the @xmath379 and @xmath380 directions in the presence of nonzero @xmath28 or @xmath29 . ,
scaledwidth=50.0% ] in addition to @xmath22 , @xmath4 , and @xmath5 , the model of ref . also contains some other exchange couplings such as @xmath28 and @xmath29 , shown in fig .
[ fig : lattice](a ) . here
we discuss their effects on the triplon dispersions . in the first - order perturbation theory ,
the additional couplings above introduce triplon hopping and interaction terms between the dimers they connect , similar to eq . .
using the effective hamiltonian , triplon dispersions at @xmath42 are calculated as @xmath381 with @xmath382 around @xmath383 , these dispersions are expanded as @xmath384 k_x^2\\ & \mp \left ( \frac{|\delta j|}2 \mp j_4 ' \right ) k_y^2 + { { \mathcal o}}(k^4 ) , \end{split}\ ] ] with @xmath385 .
this dispersion implies that the minimum of the triplon excitation energy is given by @xmath386 if the following inequalities are satisfied : @xmath387 equation can be used to determine the critical fields as@xcite @xmath388 the electronic structure calculation of ref . produced the estimates @xmath389 these parameter values satisfy the condition and give @xmath390 and @xmath391 .
we find @xmath392 , which means that the system is in fact inside the ds - i region , and the magnetization jumps as in fig . [ fig : mag_h](a ) .
we confirmed the occurrence of a jump by directly performing an exact diagonalization for the @xmath22-@xmath4-@xmath5-@xmath28-@xmath29 model with the values of eq . .
in order to be consistent with the smooth increase of the magnetization observed in experiment,@xcite the estimates in eq . must be modified .
given the difficulty in uniquely determining the exchange couplings on the basis of the current experimental data , we here discuss possible effects of @xmath28 and @xmath29 on the triplon dispersions , which could be tested in future experiments .
we add small values of @xmath28 and @xmath29 to the parameter set in eq .
, and plot the triplon dispersions eq . in fig . .
compared to fig .
[ fig : band_jjxjy](b ) , the dispersions show oscillating behaviors coming from the long - range triplon hopping induced by @xmath28 and @xmath29 .
these oscillating behaviors can be used as fingerprints for the existence of these couplings .
motivated by recent experiments on ( cucl)lanb@xmath0o@xmath1 , we have studied a spin-@xmath23 model of ferromagnetically coupled dimers on the distorted shastry - sutherland lattice . using the three different approaches ( the strong - coupling expansion , exact diagonalization , and schwinger boson mean field theory ) , we have determined the ground state phase diagram of this model , as shown in figs . [ fig : phase_ed ] and [ fig : phase_sb ] .
we have shown that in the dimer singlet phase appearing for @xmath189 , the magnetization process depends crucially on the spatial anisotropy of @xmath4 and @xmath5 . in the ds - i region with weak anisotropy , the magnetization shows a jump , as in fig .
[ fig : mag_h](a ) . in the ds - ii regions with strong anisotropy ,
the magnetization smoothly increases after the spin gap closes at a certain magnetic field , as in fig .
[ fig : mag_h](b ) . when @xmath8 or @xmath9 , quantum phase transitions to various magnetically ordered phases ( ferromagnetic , collinear stripe , and spiral ) occur .
these magnetic phases also appear in the classical limit of the model as shown in fig .
[ fig : phase_cl ] . providing a notable difference from the classical case
, the schwinger boson analysis has demonstrated that quantum fluctuations split the classical degeneracy of two kinds of spirals ( @xmath49- and @xmath50-spirals ) .
we have compared our theoretical results with the existing experimental data on ( cucl)lanb@xmath0o@xmath1 .
a smooth magnetization process observed in the compound@xcite suggests that , at least within the @xmath22-@xmath4-@xmath5 model , the compound should be located in the ds - ii region , and that there should be substantial anisotropy in @xmath4 and @xmath5 . comparing the second - order perturbative results with the experimental data for the magnetization process@xcite and the triplon excitations,@xcite
we found that one of @xmath4 and @xmath5 may be ferromagnetic while the other may be antiferromagnetic .
this is in contrast to both the ferromagnetic estimates in the previous electronic structure calculation.@xcite however , we leave open the possibility of both ferromagnetic @xmath4 and @xmath5 , given the possible existence of higher - energy triplet excitations .
we plotted the triplon dispersions for different signs of @xmath4 and @xmath5 .
these plots can be used to test the @xmath22-@xmath4-@xmath5 model further and to determine the signs of @xmath4 and @xmath5 , when more detailed information of the triplon excitations is provided from experiments .
we have also discussed the effects of other exchange couplings @xmath28 and @xmath29 [ shown in fig .
[ fig : lattice](a ) ] on the triplon dispersions .
our analysis of the distorted shastry - sutherland model may provide a useful starting point for understanding the wide range of physics in the layered copper oxyhalide family . a quantum phase transition from a singlet ground state to a collinear stripe state observed by chemical substitution@xcite
may be interpreted as the transition from the ds - ii region with @xmath183 to the stripe state with a propagation vector @xmath393 in fig .
[ fig : phase_ed ] .
here we have excluded the case of @xmath184 , since the stripe state with @xmath394 appearing in this region is in fact a nel state with @xmath395 in the original square lattice of the compound in fig .
[ fig : lattice](a ) .
the ion - exchange method used to synthesize the layered copper oxyhalide family has a remarkable flexibility of changing the constituent atoms in various ways .
we expect that the jump of the magnetization in the ds - i region and a spiral order with a unique structure as in fig .
[ fig : spiral_qx ] will also be observed in future experiments on related compounds .
we are grateful to eric lee , seung - hun lee , and tsutomu momoi for stimulating discussions .
this work was supported by the nserc of canada , the canada research chair program , and the canadian institute for advanced research .
numerical diagonalization calculations were performed using titpack ver
. 2.@xcite
here we discuss the exact ground states of the hamiltonian in the isotropic case @xmath60 .
the discussion of this section is based on the argument of shastry and sutherland@xcite and its extension to the case of ferromagnetic @xmath39 .
we decompose the hamiltonian @xmath111 into triangular parts and solve a 3-spin problem as in fig .
[ fig : spiral_qx](a ) .
the 3-spin hamiltonian is given by @xmath396 note that here the zeeman coupling with the magnetic field is introduced only for the first and third spins and not for the second spin .
the zeeman term for the second spin is included in the hamiltonians of different triangles . in general , the decomposition of @xmath111 to triangular parts is not unique . however , definition of @xmath397 used in eq . , introduced by shastry and sutherland,@xcite
will be useful in the following discussion . when @xmath42 , @xmath397 is @xmath398-symmetric , and its eigenstates consist of a quadruplet @xmath399 ( with @xmath400 labeling @xmath401 ) and two doublets . because @xmath397 is also symmetric under the permutation of the sites @xmath198 and @xmath402 ,
the doublets are classified into ones symmetric and antisymmetric with respect to this operation , @xmath403 and @xmath404 respectively . when @xmath405 , the @xmath398 symmetry of @xmath397 is reduced to @xmath406 .
then @xmath407 and @xmath403 are mixed to form new eigenstates , @xmath408 and @xmath409 .
the eigenstates are summarized as follows : [ eq:3spin_state ] @xmath410 and the other eigenstates with @xmath411 or @xmath412 are obtained by reversing all the spins and replacing @xmath413 with @xmath414 .
the parameters @xmath415 are given by @xmath416 the corresponding eigenenergies are given by @xmath417 crucially , the two states @xmath404 have the same eigenenergies and are written as product states of a singlet on the sites @xmath198 and @xmath402 , and an isolated spin state on the site @xmath418 . the exact dimer singlet state @xmath419 in eq .
can be written in terms of @xmath404 on a given triangle .
the state @xmath419 is therefore an eigenstate of the triangular hamiltonian @xmath397 .
if @xmath404 are the ground states of the triangular hamiltonian , the dimer singlet state @xmath419 will also minimize the energy of the triangular hamiltonian .
consequently , @xmath419 becomes an exact ground state of @xmath111 , since it minimizes the local energy of every triangular hamiltonian .
we note that the degeneracy of @xmath404 is not split by a magnetic field @xmath66 , which comes from the specific form of the zeeman coupling in eq . .
the solid lines indicate the boundaries on which the 3-spin ground state changes .
the exact ground states of @xmath111 derived from the 3-spin solution are indicated in the parentheses . for @xmath420 , the exact dimer singlet state @xmath419 continue to be the ground state of @xmath111 beyond the solid line ( see the text ) . to indicate the range where @xmath419 actually remains the ground state , we show broken and dotted line .
these lines respectively indicate the magnetic fields on which two- and one - triplon excitations condense according to the perturbative calculation of ref . .
, scaledwidth=45.0% ] assuming @xmath2 and @xmath37 , the ground state of @xmath397 is summarized in fig .
[ fig : phase_exact ] . for @xmath42 , @xmath404 are the ground states of @xmath397 when @xmath421 these states remain the ground states of @xmath397 up to a certain magnetic field , indicated by solid lines in fig . [
fig : phase_exact ] . in this regime
, the dimer singlet state @xmath419 is an exact ground state of @xmath111 with eigenenergy @xmath422 . for @xmath423 , the ground state of @xmath397 changes to @xmath424 as we pass the solid line in fig .
[ fig : phase_exact ] . in this regime
, the ground state of @xmath111 is given by a fully polarized ferromagnetic state since it contains the state @xmath424 locally on every triangle .
combined with the above result , the present exact argument can complete the phase diagram of @xmath111 in the case of ferromagnetic @xmath39 .
specifically , the phase boundary is given by @xmath425 .
for @xmath420 , the ground state of @xmath397 is replaced by @xmath426 beyond the solid line in fig .
[ fig : phase_exact ] .
the solid line is give by @xmath427 outside of this region , we can not use the 3-spin solution to predict the ground state of @xmath111 , since @xmath426 is not factorizable .
however , one can show that @xmath419 is always an exact eigenstate of @xmath111 , regardless of the values of @xmath22 , @xmath39 and @xmath66 .
therefore , beyond the line , @xmath419 should still remain the exact ground state of @xmath111 until it encounters a level crossing with another eigenstate . for @xmath42 ,
this crossing has been found to occur at @xmath428 in an exact diagonalization study of finite clusters.@xcite for @xmath125 , the level crossing is associated with a condensation of triplon excitations . as shown in ref .
, a bound state of two triplons condense before one - triplon state does . in fig .
[ fig : phase_exact ] , we draw the two lines , derived from the strong coupling expansion to third order in ref . , where one- and two - triplon condensations occur .
here we discuss an alternative derivation of the phase diagram in fig .
[ fig : phase_sc1 ] , using the hartree variational state.@xcite the discussion goes in parallel with sec .
[ sec : sc ] , and some notations are common with this section .
the hartree variational state is given by a product of local states on dimers : @xmath429 this state smoothly connects between the dimer singlet state ( @xmath430 ) at @xmath42 and the fully polarized state ( @xmath431 ) at a high field . for @xmath432 ,
this state describes a superfluid state of triplons , which breaks the @xmath406 symmetry .
we minimize the energy @xmath433 with respect to the variational parameters @xmath434 to find a variational ground state . using the obtained values of @xmath434
, one can determine the magnetic structure via the relations [ eq : hartree_s ] @xmath435 since @xmath436 , we find @xmath437 where @xmath438 is the projection operator onto @xmath136 .
since eq . can also be viewed as a coherent state of @xmath439 spins , minimizing @xmath440 is precisely equivalent to finding the classical ground state of the xxz model @xmath141 .
the solution to the classical problem@xcite is essentially similar to the quantum solution ( except for the detailed shape of the magnetization process ) , and leads to the same phase diagram of the original model @xmath111 as in fig .
[ fig : phase_sc1 ] . in the entire parameter space
, the obtained variational ground state has the same periodicity as the original model ; the values of @xmath441 and @xmath442 are uniform in each subset of dimers ( @xmath155 or @xmath156 ) .
therefore , the ground states are characterized by four parameters @xmath443 and @xmath444 defined for @xmath445 . in the ds - ii region with @xmath183 , the classical ground state of @xmath141 is given by a canted antiferromagnetic state with @xmath446 and @xmath447 .
using eq .
, this leads , in the original model , to a canted stripe state with a propagation vector @xmath48 in the transverse component .
similarly , in the ds - ii region with @xmath184 , the classical solution of @xmath141 is given by @xmath448 and @xmath447 , leading to a canted stripe state with a propagation vector @xmath45 .
the magnetization curve in the ds - ii region is determined as in fig .
[ fig : mag_h](b ) , but the curves are all straight in the current variational approach . as we increase @xmath8 or @xmath78 , @xmath187 goes to zero , signaling a phase transition to a collinear stripe phase .
we note that even after the transition , the variational ground state at @xmath42 remains the dimer singlet state @xmath449 with @xmath450 .
this is an artifact of the ansatz . in the ds - iii region ,
the variational solution for @xmath451 is given by @xmath452 and @xmath45 , indicating th existence of the plateau .
here we consider a single - magnon excitation from the fully polarized state @xmath453 and determine the saturation field @xmath188 .
we introduce the fourier - transformed basis @xmath454 where @xmath280 labels a position in a unit cell as in fig .
[ fig : lattice](b ) .
we have used the @xmath30 site position @xmath455 to label a unit cell , and @xmath283 represents the position of each site relative to the @xmath30 site , as in eq . . using the basis ,
we define a @xmath308 matrix @xmath456 by @xmath457 here , the energy of the fully polarized state is subtracted . to represent the matrix @xmath456 in a compact form
, we identify @xmath280 with fictitious two - spin-@xmath23 states @xmath458 and introduce two sets of pauli matrices , @xmath459 and @xmath460 .
the obtained expression is @xmath461 where @xmath462 with @xmath298 and @xmath299 . by diagonalizing @xmath463 and obtaining the eigenvalues @xmath464 , one obtains four magnon dispersions @xmath465 . assuming that the leading instability from the fully polarized state is a condensation of single - magnon excitations , the saturation field @xmath188 is found when the minimum of the lowest dispersion @xmath466 touches zero .
this leads to an expression @xmath467 , where @xmath468 is the minimum of @xmath469 over @xmath470 .
we now determine @xmath468 .
one can show , through the formulation of luttinger and tisza,@xcite that this problem is precisely equivalent to the classical ground - state problem .
therefore , the analysis can be done in parallel with sec .
[ sec : classical ] . in the collinear stripe phases of fig .
[ fig : phase_cl ] , the minimum of the dispersion is found at @xmath471 . at this point , @xmath472 is expressed using only mutually commuting matrices , @xmath473 and @xmath474 , and thus is diagonalized easily .
the saturation field is @xmath475 in the classically spiral regime , the minima of the dispersion are found at @xmath476 and @xmath477 , where @xmath55 and @xmath88 are given in eq . .
when one of @xmath379 and @xmath380 is set to zero , the diagonalization of @xmath463 is simple , because one of @xmath473 and @xmath474 becomes a good quantum number leaving a single - spin problem in a magnetic field .
the saturation field is calculated as @xmath478 we comment that in the isotropic case @xmath34 , the minimum of the dispersion is obtained along a circle in @xmath470 space , leading to an interesting pseudo - one - dimensional density of states of magnons.@xcite in closing , we stress that our argument for determining @xmath188 in this section is based on the assumption that the leading instability from the fully polarized state is a condensation of single - magnon excitations .
the present argument therefore does not apply to the ds - i region , where a direct transition occurs from the fully polarized state to a singlet state as one lowers the magnetic field .
h. kageyama , t. kitano , n. oba , m. nishi , s. nagai , k. hirota , l. viciu , j. b. wiley , j. yasuda , y. baba , y. ajiro , and k. yoshimura , j. phys .
jpn . * 74 * , http://jpsj.ipap.jp/link?jpsj/74/1702/[1702 ] ( 2005 ) .
h. kageyama , j. yasuda , t. kitano , k. totsuka , y. narumi , m. hagiwara , k. kindo , y. baba , n. oba , y. ajiro , and k. yoshimura , j. phys .
* 74 * , http://jpsj.ipap.jp/link?jpsj/74/3155/[3155 ] ( 2005 ) .
a. kitada , z. hiroi , y. tsujimoto , t. kitano , h. kageyama , y. ajiro , and k. yoshimura , j. phys .
. jpn . * 76 * , http://jpsj.ipap.jp/link?jpsj/76/093706/[093706 ] ( 2007 ) .
m. yoshida , n. ogata , m. takigawa , j. yamaura , m. ichihara , t. kitano , h. kageyama , y. ajiro , and k. yoshimura , j. phys .
. jpn . * 76 * , http://jpsj.ipap.jp/link?jpsj/76/104703/[104703 ] ( 2007 ) .
n. oba , h. kageyama , t. kitano , j. yasuda , y. baba , m. nishi , k. hirota , y. narumi , m. hagiwara , k. kindo , t. saito , y. ajiro and k. yoshimura , j. phys .
. jpn . * 75 * , http://jpsj.ipap.jp/link?jpsj/75/113601/[113601 ] ( 2006 ) .
y. tsujimoto , y. baba , n. oba , h. kageyama , t. fukui , y. narumi , k. kindo , t. saito , m , takano , y. ajiro , and k. yoshimura , j. phys .
. jpn . * 76 * , http://jpsj.ipap.jp/link?jpsj/76/063711/[063711 ] ( 2007 ) .
y. j. uemura , a. a. aczel , y. ajiro , j. p. carlo , t. goko , d. a. goldfeld , a. kitada , g. m. luke , g. j. macdougall , i. g. mihailescu , j.a .
rodriguez , p.l .
russo , y. tsujimoto , c.r .
wiebe , t.j .
williams , t. yamamoto , k. yoshimura , and h. kageyama , phys . rev .
b * 80 * , http://link.aps.org/abstract/prb/v80/e174408[174408 ] ( 2009 ) .
a. kitada , y. tsujimoto , h. kageyama , y. ajiro , m. nishi , y. narumi , k. kindo , m. ichihara , y. ueda , y. j. uemura , and k. yoshimura , phys . rev .
b * 80 * , http://link.aps.org/abstract/prb/v80/e174409[174409 ] ( 2009 ) .
y. tsujimoto , a. kitada , h. kageyama , m. nishi , y. narumi , k. kindo , y. kiuchi , y. ueda , y.j .
uemura , y. ajiro , and k. yoshimura , j. phys .
. jpn . * 79 * , http://jpsj.ipap.jp/link?jpsj/79/014709/[014709 ] ( 2010 ) .
n. shannon , t. momoi , and p. sindzingre , phys .
96 * , http://link.aps.org/abstract/prl/v96/e027213[027213 ] ( 2006 ) . c. tassel , j. kang , c. lee , o. hernandez , y. qiu , w. paulus , e. collet , b. lake , t. guidi , m .- h .
whangbo , c. ritter , h. kageyama , and s .- h .
lee , phys .
* 105 * , http://link.aps.org/abstract/prl/v105/e167205[167205 ] ( 2010 ) .
b.s . shastry and b. sutherland , physica * 108b * , 1069 ( 1981 ) . another coupled - dimer model for ( cucl)lanb@xmath0o@xmath1 has been proposed in a.a .
tsirlin and h. rosner , phys .
b * 82 * , http://link.aps.org/abstract/prb/v82/e060409[060409 ] ( r ) ( 2010 ) . in this model ,
the arrangement of dimers are similar to ref . , and the inter - dimer couplings are non - frustrating .
we note that the two different models were based on density functional calculations for different crystal structures .
t. giamarchi , ch .
regg , and o. tchernyshyov , nat . phys . * 4 * , 198 ( 2008 ) .
t.m . rice , science * 298 * , 760 ( 2002 ) .
s. miyahara and k. ueda , phys .
* 82 * , http://link.aps.org/abstract/prl/v82/p3701[3701 ] ( 1999 ) .
s. miyahara and k. ueda , j. phys .
matter * 15 * , r327 ( 2003 ) .
s. miyahara and k. ueda , phys . rev .
b * 61 * , http://link.aps.org/abstract/prb/v61/p3417[3417 ] ( 2000 ) .
t. momoi and k. totsuka , phys .
b * 61 * , http://link.aps.org/abstract/prb/v61/p3231[3231 ] ( 2000 ) ; phys . rev .
b * 62 * , http://link.aps.org/abstract/prb/v62/p15067[15067 ] ( 2000 ) .
g. misguich , t. jolicoeur , s.m .
girvin , phys .
lett . * 87 * , http://link.aps.org/abstract/prl/v87/p097203[097203 ] ( 2001 ) .
h. kageyama , k. yoshimura , r. stern , n.v .
mushnikov , k. onizuka , m. kato , k. kosuge , c.p .
slichter , t. goto , and y. ueda , phys .
82 * , http://link.aps.org/abstract/prl/v82/p3168[3168 ] ( 1999 ) ; h. kageyama , k. onizuka , t. yamauchi , y. ueda , s. hane , h. mitamura , t. goto , k. yoshimura , and k. kosuge , j. phys .
. jpn . * 68 * , http://jpsj.ipap.jp/link?jpsj/68/1821/[1821 ] ( 1999 ) ; k. onizuka , h. kageyama , y. narumi , k. kindo , y. ueda , and t. goto , j. phys .
soc . jpn .
* 69 * , http://jpsj.ipap.jp/link?jpsj/69/1016/[1016 ] ( 2000 ) .
k. totsuka , phys .
b * 57 * , http://link.aps.org/abstract/prb/v57/p3454[3454 ] ( 1998 ) .
kolezhuk , phys .
b * 59 * , http://link.aps.org/abstract/prb/v59/p4181[4181 ] ( 1999 ) . for an introductory review of the method , see f. mila and k.p . schmidt , chapter 20 in ref . ; arxiv : http://arxiv.org / abs/1005.2495[1005.2495 ] .
h. matsuda and t. tsuneto , prog .
. suppl . * 46 * , 411 ( 1970 ) .
m. kohno and m. takahashi , phys .
b * 56 * , http://link.aps.org/abstract/prb/v56/p3212[3212 ] ( 1997 ) . s. yunoki , phys . rev .
b * 65 * , http://link.aps.org/abstract/prb/v65/p092402[092402 ] ( 2002 ) .
h.t . ueda and k. totsuka ,
b * 76 * , http://link.aps.org/abstract/prb/v76/e214428[214428 ] ( 2007 ) .
schmidt , j. dorier , a. luchli , and f. mila , phys .
b * 74 * , http://link.aps.org/abstract/prb/v74/e174508[174508 ] ( 2006 ) . a phase - separated state consist of fully occupied ( @xmath479 ) and empty ( @xmath480 ) domains of hard - core bosons .
if the total particle number is increased gradually , one observes gradual expansion and shrink of the occupied and empty domains , respectively .
if the chemical potential is changed , one instead observes a first - order transition between fully empty and occupied state which are homogeneous in the entire system .
this leads to a jump in the magnetization in the spin hamiltonian .
the classical ferromagnetic phase boundary corresponds , in the quantum case , to the line where one - magnon excitations from the fully polarized state condense ( namely , @xmath188 in eq
. becomes zero ) . in the transition from the ferromagnetic phase to the ds - i region in the spin-@xmath23 case ,
the ground state directly change from the fully polarized state to a singlet state before the one - magnon condensation occurs .
a good agreement of the quantum and classical phase boundaries means that the @xmath481-@xmath73 curve as in fig .
[ fig : ene_m](a ) becomes almost straight around the phase boundary .
you , y .- w .
li , and s .- j .
gu , phys .
e * 76 * , http://link.aps.org/abstract/pre/v76/e022101[022101 ] ( 2007 ) .
albuquerque , f. alet , c. sire , and s. capponi , phys .
b * 81 * , http://link.aps.org/abstract/prb/v81/e064418[064418 ] ( 2010 ) .
jiang and u. gerber , j. stat .
( 2009 ) p09016 .
a. auerbach and d.p .
arovas , phys .
lett . * 61 * , http://link.aps.org/abstract/prl/v61/p617[617 ] ( 1988 ) .
s. sarker , c. jayaprakash , h.r .
krishnamurthy , and m. ma , phys .
b * 40 * , http://link.aps.org/abstract/prb/v40/p5028[5028 ] ( 1989 ) .
n. read and s. sachdev , phys .
lett . * 66 * , http://link.aps.org/abstract/prl/v66/e1773[1773 ] ( 1991 ) ; s. sachdev and n. read , int .
phys . * 5 * , 219 ( 1991 ) .
a. auerbach , _ interacting electrons and quantum magnetism _ ( springer , new york , 1998 ) . in principle , the minimum of the dispersion at @xmath42 in eq . can be used only for determining @xmath187 . to calculate @xmath188
, one needs to consider a single - magnon excitation from the fully polarized state and to find the minimum of the dispersion . within the first - order perturbation theory ,
however , the two problems can be solved in parallel , since the effective model is equivalent to an xxz model and has the particle - hole symmetry .
h. nishimori , url : http://www.stat.phys.titech.ac.jp/ + nishimori / titpack2/index - e.html | a recent study [ tassel _ et al . _ , phys .
rev . lett . * 105 * , 167205 ( 2010 ) ] has proposed a remarkable spin model for ( cucl)lanb@xmath0o@xmath1 , in which dimers are ferromagnetically coupled to each other on the distorted shastry - sutherland lattice . in this model ,
the intra - dimer exchange coupling @xmath2 is antiferromagnetic , while the inter - dimer exchange couplings are ferromagnetic and take different values , @xmath3 , in the two bond directions .
anticipating that the highly frustrated character of this model may lead to a wide range of behaviors in ( cucl)lanb@xmath0o@xmath1 and related compounds , we theoretically investigate the ground state phase diagram of this model in detail using the following three approaches : a strong - coupling expansion for small @xmath4 and @xmath5 , exact diagonalization for finite clusters , and a schwinger boson mean field theory .
when @xmath6 , the system stays in a dimer singlet phase with a finite spin gap .
this state is adiabatically connected to the decoupled - dimer limit @xmath7 .
we show that the magnetization process of this phase depends crucially on the spatial anisotropy of the inter - dimer couplings .
the magnetization shows a jump or a smooth increase for weak and strong anisotropy , respectively , after the spin gap closes at a certain magnetic field .
when @xmath8 or @xmath9 , quantum phase transitions to various magnetically ordered phases ( ferromagnetic , collinear stripe , and spiral ) occur .
the schwinger boson analysis demonstrates that quantum fluctuations split the classical degeneracy of different spiral ground states .
implications for ( cucl)lanb@xmath0o@xmath1 and related compounds are discussed in light of our theoretical results and existing experimental data . |
let @xmath9 be an inhomogeneous poisson process of rate @xmath2 such that @xmath10 for all @xmath11 .
the epochs of @xmath12 , in increasing order , are denoted by @xmath13 , @xmath14 , so that the gaps are given by @xmath15 with @xmath16 .
the objects of study of the present paper are the longest gap , @xmath0 , before time @xmath1 and its right - end position , @xmath17 : @xmath18 note that the definition does not include the gap straddling time @xmath1 , but this is in fact unimportant for our asymptotic results , see remark [ rem : count - the - gap ] . in the homogeneous case ,
the discrete time analogue of the longest gap is the longest run , @xmath19 , of ones before time @xmath20 in a bernoulli@xmath21 sequence .
the study of the longest run has a long history going back to , among others , @xcite ; a recent survey is in @xcite .
a main result is that @xmath19 is of order @xmath22 . in the homogeneous poisson case ,
@xmath23 , there is a neat analogue of this ( which despite its simplicity we have not seen stated in the literature ) : @xmath24 where @xmath25 is gumbel with cumulative distribution function ( cdf ) @xmath26 @xmath27 .
the proof is equally neat : with @xmath28 @xmath29 for @xmath30 one has @xmath31 for large @xmath1 with high probability .
further , by standard extreme value theory , the random variables @xmath32 have gumbel limits as @xmath33 , so one can just let first @xmath1 tend to infinity and next @xmath34 tend to @xmath35 .
note that fails in the bernoulli setting because the extreme value behaviour of geometric random variables is more complicated than the one of exponential random variables , cf .
@xcite . as mentioned already ,
our interest is in time inhomogeneity .
this may occur in at least two ways .
firstly , one may consider fluctuations around a long - term average which is conveniently modelled in a hidden markov setting , see @xcite .
secondly , the rates @xmath2 may exhibit a systematic deterministic trend .
the only reference here seems to be @xcite ( though cf . also @xcite ) , continuing a study of @xcite related to problems from computer reliability .
the results in @xcite are of large deviations type , giving asymptotic estimates of @xmath36 in the rare - event setting where @xmath37 with @xmath38 fixed .
our concern here is the typical behaviour , that is , analogues of . as in @xcite ,
the quantitative form of @xmath2 is crucial both for the form of the results and the difficulty of the analysis .
first , we concentrate on what is maybe the simplest form , a power function @xmath39 , and then provide extensions to regularly varying functions .
the power function is a rather natural choice with which to start the analysis , and already this case presents substantial challenges .
the case @xmath40 is settled by and the behaviour when @xmath41 or @xmath42 is easily resolved , see remark [ rem:18.11a ] below .
thus what is left for analysis is the case @xmath43 , and here our result is the following : [ thm : main ] let @xmath44 be an inhomogeneous poisson process of rate @xmath45 with @xmath46 and @xmath47 .
let @xmath0 and @xmath17 be as in and , respectively .
it holds that @xmath48 where @xmath49 and @xmath50 are independent random variables : @xmath25 is gumbel and @xmath51 is exponential of rate @xmath52 . in fact , we prove a much more general result establishing weak convergence of a sequence of point processes , from which theorem [ thm : main ] easily follows . here and as usual , convergence in distribution of point processes is with respect to the vague topology in the space of radon measures on @xmath53 ^ 2 $ ] .
[ thm : point_proc ] under the assumptions of theorem [ thm : main ] consider the point process @xmath54 on @xmath53 ^ 2 $ ] consisting of the points @xmath55 then @xmath56 as @xmath57 , where @xmath58 is a poisson point process with intensity measure @xmath59 importantly , in theorem [ thm : main ] we consider the compactified euclidean plane @xmath53 ^ 2 $ ] so that the set @xmath60\times [ -z,\infty]$ ] is compact .
the points of @xmath54 in this set correspond to @xmath61 and @xmath62 .
hence our result concerns all large enough gaps of @xmath12 up to the time @xmath63 . furthermore , since vague convergence of point measures implies convergence of the respective points in any compact set ( * ? ? ?
3.13 ) , we conclude that the map @xmath64 is continuous apart from possible discontinuities at point measures with @xmath65 or @xmath66 for some @xmath67 . since @xmath58 is not of such form a.s .
, the continuous mapping theorem gives that @xmath68 , where @xmath69 has the distribution arising from the application of the above map to @xmath58 .
a standard calculation reveals that for @xmath70 we have @xmath71 proving theorem [ thm : main ] ; see also the light - gray region in figure [ fig : points ] .
[ rem : count - the - gap]in order to give a feeling for some further results we define @xmath72 to be the position of the first gap exceeding @xmath0 and its value , respectively . from theorem
[ thm : point_proc ] and the continuous mapping theorem applied to the appropriate map , we find that @xmath73 where the conditional distribution of @xmath74 is easily identified to be @xmath75 for @xmath76 and @xmath77 ; see the dark - grey region in figure [ fig : points ] . and
@xmath78 and the associated empty regions @xmath79 in ( light - gray ) and @xmath80 in ( dark - gray ) , respectively.,scaledwidth=40.0% ] in particular , we find after some algebra that @xmath81
. one may proceed even further and obtain convergence of extremal processes ( on the skorokhod space of two - sided paths ) identifying the record gaps and their times , see ( * ? ? ?
4.20 ) for the classical setting .
finally , note that @xmath82 and so @xmath83 , showing that the corresponding gap does not straddle time @xmath1 in the limit . when trying to adapt the above proof of , with scale constant @xmath84 say , one quite easily gets @xmath85 , which gives a rough estimate of @xmath0 in terms of @xmath86 .
the difficulty is that these interarrival times @xmath87 are no longer independent nor exponentially distributed .
nevertheless , the @xmath87 are not too far from exponential random variables with rates @xmath88 , because @xmath89 for large @xmath90 .
hence our first step is to consider extreme value theory for sequences of i.i.d .
random variables equipped with weights .
some references in that direction are @xcite and , of particular relevance for us , ( * ? ? ?
4.1 ) , from which the following result can be extracted : [ prop : wn ] [ prop : max ] let @xmath91 be independent unit exponential random variables and let @xmath92 .
then with @xmath93 it holds that @xmath94 where @xmath95 .
our analysis supplements this result by identifying the location of the maximum and providing the analogue of theorem [ thm : point_proc ] .
this location is trivially uniform for i.i.d .
sequences or homogeneous poisson processes , but has an interesting limiting distribution in the nonhomogeneous case .
we also give an extension to weights in proposition [ prop : wn ] and rates in theorem [ thm : main ] which are regularly varying rather than of simple power form .
such an extension is of course expected , but the proof is surprisingly complicated , and in fact , we need some regularity conditions on the slowly varying function .
[ rem:18.11a]when @xmath41 , @xcite gives that the increasing process @xmath0 has a proper limiting distribution , of @xmath96 , say .
that is , from ( 7 ) in @xcite it follows that @xmath97 as @xmath98 .
the case @xmath99 is trivial since then @xmath100 and so @xmath101 for all large @xmath90 ; in this case the indexing of gaps should be modified since there are infinitely many epochs in the neighbourhood of 0 .
the boundary case @xmath102 is also easy : if @xmath103 for some scale constant @xmath46 ,
then @xmath104 indeed , fix @xmath105 and define the time - changed process @xmath106 by @xmath107 for @xmath108 . its intensity measure , @xmath109 , satisfies @xmath110 for any @xmath111 .
it follows that @xmath112 has the same distribution as @xmath113 , and it is then clear that @xmath114 has the same distribution as @xmath115 .
finally , observe that @xmath116 is of order @xmath1 when @xmath102 or @xmath117 , the two boundary cases in theorem [ thm : main ] . in contrast
, theorem [ thm : main ] gives the smaller order @xmath118 when @xmath47 .
therefore , it is intuitive that the limiting random variable @xmath51 must increase to @xmath119 as @xmath120 approaches 0 or 1 .
this is indeed the case .
as in proposition [ prop : wn ] , we consider a sequence @xmath91 of independent , unit exponential random variables .
we fix @xmath121 and let @xmath122 denote the partial maximum of the weighted sequence @xmath123 and the location of that maximum , respectively . in the i.i.d .
case , @xmath124 , the random variable @xmath125 is uniformly distributed on @xmath126 . since the weights @xmath127 increase to infinity
, one would expect that @xmath128 as @xmath129 .
the following proposition makes this precise .
[ prop : loc ] for @xmath130 and @xmath131 as in , it holds that @xmath132 with @xmath133 and where @xmath25 and @xmath134 are independent random variables , @xmath25 is gumbel and @xmath134 is exponential with rate @xmath135 .
we start by proving a lemma which is basic for the proof of proposition [ prop : loc ] and the associated point process result given in proposition [ prop : point_discrete ] .
[ lem : mz1z2 ] for every @xmath136 , we have @xmath137 where @xmath25 is a gumbel random variable . letting @xmath138 we find from proposition [ prop : max ] that @xmath139 further , @xmath140 an elementary calculation yields @xmath141 the result follows by slutsky s lemma and the fact that @xmath142 , where @xmath143 means that @xmath144 as @xmath129 .
the following result establishing convergence of the underlying point processes is close in spirit to , e.g. , ( * ? ? ?
* thm . 1 ) , and it serves as the basis for theorem [ thm : point_proc ] .
[ prop : point_discrete ] the point process @xmath145 on @xmath53 ^ 2 $ ] consisting of the points @xmath146 converges in distribution as @xmath147 to the poisson point process @xmath148 with mean measure @xmath149 let @xmath150 . according to the result of grigelionis , see e.g. ( * ? ? ?
16.18 ) , applied to a null array of single points , it is only required to show that @xmath151 for any finite union @xmath152 of rectangles in @xmath153 ^ 2 $ ] . in our
setting it is sufficient to check the above limits for @xmath154\times [ z,\infty]$ ] .
the first limit result follows from monotonicity of @xmath155 and @xmath156 using this and lemma [ lem : mz1z2 ] we also find that @xmath157\times [ z,\infty]),\end{aligned}\ ] ] as required .
it follows by the continuous mapping theorem applied to proposition [ prop : point_discrete ] in the same way as theorem [ thm : main ] follows from theorem [ thm : point_proc ] .
alternatively , one may proceed directly by identifying the limit distribution : @xmath158 and then expressing the distribution of interest using @xmath159 and the above quantity .
let @xmath160 be the points of a poisson process @xmath161 with rate @xmath162 for some @xmath163 and with cumulative rate function @xmath164 .
note that we assume that @xmath165 ; the case @xmath166 for general @xmath46 follows by the time change argument , but see also section [ sec : reg_var ] . recall that @xmath167 for integer @xmath168 , where @xmath169 .
define @xmath170 for integer @xmath171 , so that @xmath172 are the points of a unit - rate homogeneous poisson process @xmath173 .
let @xmath174 be its interarrival times , for integer @xmath168 .
the random variables @xmath91 are independent unit exponentials .
put @xmath175 the following result provides the basic approximation .
[ lem : rbyx ] it holds that @xmath176 since @xmath177 is the partial sum process of a sequence of independent unit exponentials , the law of the iterated logarithm states that @xmath178 which further implies @xmath179 and so @xmath180 a.s . as required . concerning the second part , we write using the mean - value theorem @xmath181 with @xmath182 .
hence it is left to show that @xmath183 a.s . , which again follows from .
in the following we relate the points of the point process @xmath54 in theorem [ thm : point_proc ] to the corresponding points of the process @xmath184 in proposition [ prop : point_discrete ] with rescaled second component .
[ lem : points_close ] let @xmath185\times[z_1,z_2]$ ] and put @xmath186 with @xmath187 .
then @xmath188 as @xmath189 with the convention that @xmath190 .
letting @xmath191 we see that @xmath192 and hence also @xmath193 uniformly in @xmath194 as @xmath37 .
now according to lemma [ lem : rbyx ] , for all @xmath194 it holds that @xmath195 where @xmath196 as @xmath189 a.s .
so we have @xmath197 where in the last line we used the facts : @xmath198 and @xmath199 .
this and the analogous lower bound imply that @xmath200 as @xmath37 a.s . , because @xmath201 is bounded for the indices of interest . upon recalling that @xmath202 uniformly in @xmath194 , for all such @xmath90 we find that @xmath203 this and the analogous lower bound yield @xmath204 as @xmath37 a.s . , because now @xmath205 is bounded for the indices of interest .
next , consider the set of indices @xmath206 . in this case
we use the fact that @xmath207 uniformly in @xmath208 .
furthermore , with probability 1 as @xmath189 the corresponding indices @xmath90 converge to @xmath119 too , and since @xmath209 we must have that @xmath193 uniformly in @xmath208
. thus holds true and hence also @xmath210 the corresponding upper bound , as well as the bounds on @xmath211 stemming from , complete the proof , because @xmath212 and @xmath213 are bounded for all @xmath208 .
[ rem : rescaling ] the point process @xmath214 with @xmath215 defined in lemma [ lem : points_close ] is a rescaled version of @xmath216 in proposition [ prop : point_discrete ] , and the proof of the latter easily yields that @xmath214 converges in distribution to a poisson point process with intensity measure for the set @xmath60\times[z,\infty]$ ] given by @xmath217\times[{\alpha}^2 z,\infty ] ) = { \mathrm{e}}^{-x-{\alpha}(1-{\alpha})z } = \mu([x,\infty]\times[z,\infty]).\ ] ] that is , the corresponding limit is @xmath58 .
the following lemma shows that compact sets of the form @xmath60\times[z,\infty]$ ] can be truncated to finite rectangles .
[ lem : bounding ] for any @xmath218 and @xmath219 there exist @xmath220 and @xmath221 such that @xmath222\times[z,\infty ] ) \setminus ( [ x , x']\times[z , z']))>0)<\epsilon.\ ] ] put @xmath223 and observe using that @xmath224 where in the last equality we used the law of large numbers applied to @xmath225 and the fact that @xmath226 is asymptotically gumbel .
but then @xmath227 in probability .
thus it is sufficient to restrict our attention to the indices @xmath228 , in which case we have for all such @xmath90 a.s .
observe that for @xmath228 the bound is still true .
letting @xmath229 be the set of indices @xmath90 such that @xmath230 $ ] or @xmath231 $ ] , we see from the proof of lemma [ lem : points_close ] that @xmath202 uniformly in @xmath232 , and also that @xmath233 hence for any fixed @xmath234 with arbitrarily high probability the following is true for large enough @xmath1 : if for some @xmath228 it holds that @xmath235 then @xmath236 because for @xmath237 monotonicity of @xmath13 implies @xmath238 . thus it is left to apply proposition [ prop : point_discrete ] and to note that @xmath239 with @xmath240\times[{\alpha}^2(z'-\delta),\infty],\quad b_2=[x'-\delta,\infty]\times[{\alpha}^2(z-\delta),\infty]\ ] ] as @xmath241 , which implies that @xmath242 . according to ( * ? ? ?
1 ) it is sufficient to show that @xmath243 where @xmath244 is a compact rectangle in @xmath53 ^ 2 $ ] and @xmath152 is a finite union of such rectangles .
according to lemma [ lem : bounding ] we may choose @xmath245 such that @xmath246\times[z_1,z_2]))>0 \bigr ) < \epsilon\end{aligned}\ ] ] for @xmath247\times [ z_1,z_2])$ ] and all @xmath1 large enough .
furthermore , we may additionally ensure that @xmath248 .
a similar observation holds true with respect to @xmath249 and the corresponding difference for the process @xmath58 .
hence it is sufficient to prove and for any finite rectangle @xmath244 and a finite union @xmath152 of such rectangles .
fix @xmath234 and define the @xmath250-enlarged set @xmath251 and @xmath250-narrowed set @xmath252 . according to lemma [ lem : points_close ] with the respective rectangle chosen to cover @xmath152 , it holds that @xmath253 for all @xmath1 large . noting that @xmath254 we obtain from remark [ rem : rescaling ] based on proposition [ prop : point_discrete ] . in a similar way we also find that @xmath255 .
the proof is complete .
let @xmath256 denote the set of measurable functions @xmath257 which are regularly varying at @xmath119 with index @xmath258 , i.e. , satisfying @xmath259 as @xmath260 for all @xmath261 .
any such @xmath262 can be represented as @xmath263 with @xmath264 a slowly varying function .
regularly varying functions are thus a generalization of the power functions considered above .
let us also recall the basic theorem concerning regularly varying functions @xmath265 , the uniform convergence theorem ( * ? ? ?
1.5.2 ) : @xmath266 on intervals @xmath267 $ ] with @xmath268 for @xmath269 , and on intervals @xmath270 $ ] for @xmath271 if @xmath262 is locally bounded .
assume that the rate function @xmath272 is in @xmath273 for some @xmath274 , so that @xmath275 .
let @xmath276 be the inverse function of @xmath277 and let @xmath278 be its derivative
. then @xmath279 and @xmath280 with @xmath281 .
the point process @xmath282 is a unit - rate homogenous poisson process with epochs @xmath283 .
we generalize our main results imposing just one condition on the slowly varying function @xmath38 associated to @xmath284 , i.e. , @xmath285 ; see condition [ cond : loc ] below
. the basis of our analysis will be the approximation @xmath286 inspired by the mean - value theorem applied to @xmath287 and the strong law of large numbers applied to the partial sum sequence @xmath225 .
therefore , we first study the behaviour of the maximum of the weighted exponentials @xmath288 .
let @xmath91 be a sequence of i.i.d .
unit exponentials and let @xmath289 with @xmath121 .
write @xmath285 with @xmath290 .
we consider the maximum of the weighted exponentials @xmath288 and the location of that maximum : @xmath291 [ lem : tauto1 ] we have @xmath292 as @xmath129 .
let @xmath293 .
first , we prove that @xmath294 as @xmath129 , or equivalently , that @xmath295 on the one hand , we have @xmath296 but @xmath297 can be assumed to be locally bounded [ otherwise redefine @xmath284 by @xmath298 , and so by it follows that @xmath299 } u^\gamma = h^\gamma.\ ] ] since @xmath91 are i.i.d .
unit exponentials , we have @xmath300 and thus @xmath301 on the other hand , let @xmath302 . by a similar argument as in the previous paragraph , we find @xmath303 since @xmath304 , we obtain , as required .
concerning the first statement , observe from above and proposition [ prop : loc ] that with arbitrarily high probability @xmath305 for large enough @xmath20 .
hence it is sufficient to show that @xmath306 and the same for @xmath307 , which again follows from . in order to generalize proposition [ prop : max ] we need a stronger statement than the readily available , and so we assume the following additional condition on the slowly varying function @xmath38 . [
cond : loc ] whenever @xmath308 as @xmath33 , we have @xmath309t)}{\ell(t ) } - 1 \right ) \to 0.\ ] ] in section [ sec : condition ] we provide a simple sufficient criterion under which condition [ cond : loc ] holds .
it is important to realize that condition [ cond : loc ] is equivalent to a seemingly stronger condition stated in the following lemma .
[ lem : cond_equiv ] condition [ cond : loc ] is equivalent to @xmath310 for any @xmath311 . given in appendix [ appendix ] .
recall @xmath312 .
[ lem : mnstarz1z2 ] assuming condition [ cond : loc ] we have @xmath313 . since @xmath314 as @xmath129 by lemma [
lem : tauto1 ] and proposition [ prop : loc ] , we can find @xmath315 such that @xmath316 and @xmath317 \to 1 $ ] as @xmath129 .
hence with arbitrarily high probability we have @xmath318 lemma [ lem : cond_equiv ] and lemma [ lem : mz1z2 ] show that @xmath319 completing the proof .
let @xmath289 for some @xmath121 and put @xmath320 .
if @xmath38 satisfies condition [ cond : loc ] then proposition [ prop : loc ] and proposition [ prop : point_discrete ] hold with @xmath321 in place of @xmath322 .
one may follow the same steps as in the original proofs .
in addition , for the analogue of we use lemma [ lem : rv ] , whereas the extension of proposition [ prop : point_discrete ] requires showing that @xmath323 which follows from applied to the function @xmath324 .
let @xmath325 be an inhomogenous poisson process as in the beginning of this section .
as a consequence of lemma [ lem : cond_equiv ] , we have that@xmath326 because @xmath327 .
let us now provide a generalization of lemma [ lem : rbyx ] .
[ lem : approxrbyx ] if @xmath38 satisfies condition [ cond : loc ] , then @xmath328 as @xmath329 . using monotonicity of @xmath287 and we find that a.s .
@xmath330 which is @xmath331 by lemma [ lem : rv ] .
a similar bound from below completes the proof of the first part . for the second part we write @xmath332
so that @xmath333 but then @xmath334 from and we find that the last term is @xmath331 a.s . as required .
[ thm : rv ] if the rate function @xmath335 , where @xmath47 , is such that @xmath38 satisfies condition [ cond : loc ] , then theorem [ thm : main ] and theorem [ thm : point_proc ] hold with such @xmath2 and @xmath336 in this more general setting we use @xmath337 and so @xmath338 . concerning the generalization of lemma [ lem : points_close ] we only need to show that and hold when adapted according to lemma [ lem : approxrbyx ] .
that is , @xmath339 this hinges on the following : ( i ) @xmath340 , ( ii ) @xmath341 , and ( iii ) @xmath342 uniformly in @xmath194 .
identity ( i ) holds , because by @xmath343 where , indeed , @xmath344 . identity ( ii )
is rather obvious , whereas concerning ( iii ) we have @xmath345 but @xmath346 by a slight extension of lemma [ lem : rv ] upon noting that @xmath347 uniformly in @xmath90 concerned .
it is left to show that lemma [ lem : bounding ] still holds , and the only non - trivial step is to show that @xmath348 in probability and hence a.s . , which we obtain in the following .
observe from that @xmath349 where @xmath350 and concerning the latter term we have @xmath351 by , provided that @xmath284 is locally bounded .
this shows that @xmath352 and hence holds in view of . in general , however , we only have that @xmath284 is bounded on @xmath267 $ ] for some @xmath353 and all @xmath354 . with arbitrarily high probability
we may choose an index @xmath355 such that @xmath356 , and then the above steps can be repeated for @xmath357 , whereas obviously @xmath358 a.s .
let us note that virtually all standard examples of slowly varying functions , e.g. @xmath359 and @xmath360 , satisfy condition [ cond : loc ] .
this can be easily checked using the following result .
[ lem : condition ] condition holds true if @xmath264 is eventually differentiable and @xmath361 using the mean value theorem we have @xmath362t)-\ell(t)|\leq t\epsilon(t)\sup_{t \le s \le [ 1+\epsilon(t)]t } |\ell'(s)|.\ ] ] moreover , @xmath363 t } |\ell'(s)| \leq \sup_{t \le s \le [ 1+\epsilon(t)]t } \left|\frac{s \ , \ell'(s)\log s}{\ell(s)}\right| \sup_{t \le s \le [ 1+\epsilon(t)]t } \left|\frac{\ell(s)}{s\log s}\right|,\ ] ] where the first term on the right hand side is @xmath364 according to .
hence condition holds if @xmath363t}\frac{t\log(t)\ell(s)}{s\log(s)\ell(t)}\ ] ] is bounded for large @xmath1 , but this term tends to 1 by applied to the regularly varying function @xmath365 .
concerning theorem [ thm : rv ] it is more useful to express the sufficient condition of lemma [ lem : condition ] using the slowly varying function associated to the rate function @xmath2 instead of that associated to @xmath366 , which is the contents of the next result .
[ prop : cond ] let @xmath367 with @xmath47 and @xmath368 . if @xmath369 is eventually continuously differentiable and if @xmath370 then condition [ cond : loc ] is satisfied and the result of theorem [ thm : rv ] holds true .
first , we show that is equivalent to @xmath371 using @xmath372 and @xmath373 we find that @xmath374 plugging in @xmath375 and noting that @xmath376 we confirm the equivalence . thus it is sufficient to establish that @xmath377 the left statement is a result of a simple calculation , and so we concentrate on the right statement . using integration by parts we find @xmath378 for all @xmath379 and some level @xmath380 ( to be fixed high enough ) .
hence it is left to show that @xmath381 form our assumption we see that @xmath382 for large enough @xmath383 . finally , by karamata s theorem ( * ? ? ?
1.5.8 ) we have @xmath384 because @xmath385 , and so follows .
as mentioned above , virtually all standard examples of slowly varying functions satisfy the assumption of proposition [ prop : cond ] . in particular , so does @xmath386 for @xmath387 .
hence @xmath388 is an example of a rate function for which the asymptotic results derived in this work are true . for a simple example that does not satisfy the assumption of proposition [ prop : cond ] ,
consider @xmath389 . indeed , this is a slowly varying function for which @xmath390 is unbounded .
[ [ appendix ] ] it is clearly sufficient to show that condition [ cond : loc ] implies .
firstly , from condition [ cond : loc ] we have that @xmath391t)}{\ell(t ) } - 1\right)=-\log t\left(\frac{\ell([1+\hat\epsilon(\hat t)]\hat t)}{\ell(\hat t ) } - 1\right)\frac{\ell([1-\epsilon(t)]t)}{\ell(t)}\to 0,\ ] ] where @xmath392 and @xmath393 . next , for any @xmath218 and any large @xmath1 we can choose @xmath394 $ ] such that @xmath395t)}{\ell(t ) } - 1 \right|+\epsilon/\log t>\sup_{-\epsilon(t)\leq x\leq \epsilon(t)}\left| \frac{\ell([1+x]t)}{\ell(t ) } - 1 \right|.\ ] ] but the term on the left when multiplied by @xmath396 must converge to @xmath34 , because @xmath397 .
the limit result in follows since @xmath218 is arbitrary .
the following technical result concerning regularly varying functions may well exist in the literature .
[ lem : rv ] let @xmath262 be an increasing , absolutely continuous function such that its radon nikodym derivative @xmath398 is in @xmath399 for some @xmath400 .
if @xmath401 as @xmath33 , then @xmath402 we have @xmath403 and thus @xmath404 but then @xmath405 the term in curly brackets converges to zero by applied to @xmath398 , the fact that @xmath406 , and the direct half of karamata s theorem , see e.g. theorem 1.5.11 in @xcite .
j. ivanovs acknowledges support by t.n . thiele center at aarhus university
. j. segers gratefully acknowledges funding by contract `` projet dactions de recherche concertes '' no .
12/17 - 045 of the `` communaut franaise de belgique '' and by iap research network grant p7/06 of the belgian government ( belgian science policy ) .
s. asmussen , p. fiorini , l. lipsky , t. rolski & r. sheahan ( 2008 ) on the distribution of total task times for tasks that must restart from the beginning if failure occurs .
res . _ * 33 * , 932944 .
s. asmussen , l. lipsky & s. thompson ( 2016 ) markov renewal methods in restart problems in complex systems , in : _ the fascination of probability , statistics and their applications . in honour of ole e. barndorff - nielsen _ ,
. 501527 . | let @xmath0 be the longest gap before time @xmath1 in an inhomogeneous poisson process with rate function @xmath2 proportional to @xmath3 for some @xmath4 .
it is shown that @xmath5 has a limiting gumbel distribution for suitable constants @xmath6 and that the distance of this longest gap from @xmath1 is asymptotically of the form @xmath7 for an exponential random variable @xmath8 .
the analysis is performed via weak convergence of related point processes .
subject to a weak technical condition , the results are extended to include a slowly varying term in @xmath2 . |
the leo a dwarf galaxy was discovered by @xcite and is one of the most remote galaxies in the local group .
leo a is gas rich , with an h i velocity dispersion of 3.5 to 9 km s@xmath1 and with no observed rotation @xcite .
leo a is also extremely metal poor , with an abundance of @xmath6 to 7.4 measured from h ii regions @xcite .
photometric studies of leo a reveal both a red and blue plume of stars in its color - magnitude diagram indicating recent star formation @xcite .
_ hubble space telescope _
observations have resolved the stellar population of leo a , which shows evidence for numerous epochs of star formation spanning billions of years @xcite as well as an old stellar `` halo '' @xcite .
rr lyrae variables confirm the presence of an @xmath7 gyr old population , and place leo a at a distance of @xmath8 kpc @xcite .
recently , @xcite reported the first spectroscopy of stars in leo a : two b supergiants stars observed serendipitously as part of their hypervelocity star survey .
the b supergiants provide spectroscopic proof of star formation as recently as @xmath530 myr ago in leo a. inspired by the b supergiant observations , we have obtained spectroscopy for ten additional blue - plume objects in leo a. there is no a - priori reason to expect that leo a s steller and h i gas velocity dispersions are identical .
detailed h i maps show velocity structure , which suggests that the gas may be affected by cooling or may not yet be relaxed @xcite .
our observations allow us to measure the stellar velocity dispersion , and thus estimate the mass of leo a s dark matter halo . in
2 we discuss our target selection , observations , and stellar radial velocity determinations . in 3 we present the resulting velocity dispersion and mass - to - light ratio of leo a. we conclude in 4 .
we use sloan digital sky survey ( sdss , * ? ? ?
* ) photometry to select candidate leo a blue plume stars by color .
we illustrate our target selection in figure [ fig : ugr ] , a color - color diagram of every star in sdss data release 5 with @xmath9 and within @xmath10 of leo a ( see also figure [ fig : radec ] ) .
we compute de - reddened colors using extinction values obtained from @xcite ; the adopted extinction values are @xmath11 and @xmath12 .
objects with @xmath13 and @xmath14 are objects in the blue plume
. the blue plume can contain massive main sequence stars , blue supergiant stars , and blue - loop stars ( e.g. * ? ? ? * ) .
we target the 12 blue plume objects with @xmath9 ( solid squares and triangles )
. objects with @xmath15 in figure [ fig : ugr ] have colors consistent with foreground stars , ranging from f - type stars at the main sequence turn - off @xmath16 to late m dwarfs @xmath17 . stars with @xmath18 may include some asymptotic giant branch stars in leo a. figure [ fig : radec ] plots the position of every star in figure [ fig : ugr ] . for reference ,
the ellipses follow leo a s observed h i profile , with center @xmath19 @xmath20 ( j2000 ) , semiminor to semimajor axis ratio 0.6 , and position angle @xmath21 @xcite .
the solid ellipse marks leo a s holmberg radius @xmath22 @xcite , and the dotted ellipse with @xmath23 marks the extent of leo a s h i gas @xcite and stellar `` halo '' @xcite .
all twelve blue plume candidates are located within @xmath3 of the center of leo a ; probable foreground objects are distributed more uniformly across the field .
we obtained spectroscopy of the twelve blue plume objects with the 6.5 m mmt telescope and the blue channel spectrograph .
observations occurred during the course of our hypervelocity star survey program on the nights of 2005 dec 5 - 6 , 2006 may 24 - 25 , 2006 june 20 , 2006 dec 27 , and 2007 mar 18 .
we operated the blue channel spectrograph with the 832 line mm@xmath1 grating in 2nd order and with a 1.25@xmath24 slit .
these settings provided a wavelength coverage of 3650 to 4500 and a spectral resolution of 1.2 .
one object ( an h ii region ) was re - observed with the 300 line mm@xmath1 grating and a 1 @xmath24 slit , providing wavelength coverage from 3400 to 8600 with a spectral resolution of 6.2 .
exposure times were 30 minutes .
we obtained comparison lamp exposures after every exposure .
the wavelength solutions are determined from 44 lines with typical root - mean - square residuals of @xmath25 , or @xmath26 km s@xmath1 .
we note that the single slit spectrograph is a compact instrument with minimal flexure : wavelength solutions shift by less than 1 pixel ( 0.355 ) during a night , easily measured from individual comparison lamp exposures .
ten blue plume objects are stars of b spectral type and two are h ii regions .
figure [ fig : spectra ] plots the spectra of the ten stars and the two h ii regions , summed and shifted to the rest frame .
the signal - to - noise ratios ( @xmath27 ) of the individual spectra range from @xmath28 to 15 per pixel at 4000 , and depend on target s apparent magnitude and the seeing conditions of the observation .
the ten b - type stars have visibly narrower balmer lines and thus lower surface gravity than the other b - type stars in the @xcite hypervelocity star survey .
cross - correlation with mk spectral standards @xcite indicates that the stars are probably luminosity class i or ii b supergiants , consistent with the stars inferred luminosities . at the distance modulus of leo a @xmath29 @xcite , the ten b - type stars have absolute magnitudes ranging from @xmath30 to @xmath31 . for comparison ,
@xcite give absolute magnitudes @xmath32 for a b9 ib star and @xmath33 for a b9 ii star .
we conclude the ten stars are likely b supergiants in leo a. such b supergiants have ages ranging from @xmath530 myr for the most luminous stars to @xmath5200 myr for the least luminous stars @xcite .
we measure radial velocities with the cross - correlation package rvsao @xcite .
we begin by observing the b9 ii star @xmath34 lyr by quickly scanning the star across the spectrograph slit .
this procedure provides us with a very high signal - to - noise ratio cross - correlation template with a known velocity @xcite .
the accuracy of the velocity zero - point comes from the error on the mean of the 44 comparison lamp lines used to determine the template s wavelength solution , @xmath35 km s@xmath1 .
is is important that we maximize velocity precision for our velocity dispersion measurement , and we achieve the best precision by cross - correlating the stars with themselves .
thus , after measuring the stars velocities with the @xmath34 lyr template , we shift the spectra to the rest frame and sum them together to create a second template ( shown in figure [ fig : spectra ] ) .
we then cross - correlate the ten stars with this second template of themselves . table [ tab : obs ] lists the resulting heliocentric radial velocities and errors .
the mean cross - correlation precision is @xmath36 km s@xmath1 .
we also measure the radial velocities of the h ii regions with rvsao , but this time using gaussian fits to the emission lines .
the final velocity of sdss j095927.532 + 304457.75 comes from a weighted mean of the 3727 [ oii ] doublet ( resolved in our spectra ) , h@xmath37 , and h@xmath34 emission lines .
a low - dispersion spectrum of sdss j095933.320 + 304439.21 provides additional line measurements from h@xmath38 , [ oiii ] , and h@xmath39 for that object .
the velocity of sdss j095933.320 + 304439.21 is the weighted average of all of its observed lines .
the mean emission - line velocity error is @xmath40 km s@xmath1 .
the average velocity of our twelve leo a objects is @xmath41 km s@xmath1 ( see figure [ fig : vel ] ) , statistically identical with the @xmath42 km s@xmath1 systemic h i velocity measured by @xcite and the 23.2 - 24.0 km s@xmath1 systemic h i velocities measured by @xcite .
thus the velocities of our twelve objects are all consistent with membership in leo a. the root - mean - square velocity dispersion of our twelve objects is 10.0 km s@xmath1 .
we derive the intrinsic velocity dispersion by subtracting in quadrature the average 3.8 km s@xmath1 uncertainty of the observations .
thus we measure an intrinsic stellar velocity dispersion of @xmath0 km s@xmath1 .
we estimate the robustness of the velocity dispersion measurement by comparing the cumulative distribution of velocities to a gaussian distribution ( see figure [ fig : vel ] ) .
a kolmogorov - smirnov test finds a 0.5 likelihood of drawing the twelve objects from a gaussian distribution with the observed velocity dispersion .
greater number statistics are always desirable , but it appears that the twelve blue plume objects provide a statistically sound measurement of leo a s stellar velocity dispersion .
our stellar velocity dispersion measurement is identical to the h i gas velocity dispersion measured by @xcite : @xmath43 km s@xmath1 .
@xcite also observe an h i component with @xmath44 km s@xmath1 dispersion localized in high column - density regions .
if we remove the two h ii regions from our own analysis , the b - type stars have a mean velocity of @xmath45 km s@xmath1 and an intrinsic velocity dispersion of @xmath46 km s@xmath1 .
this dispersion is statistically identical to our original value
. there is no evidence for rotation of the stellar component of leo a ; the high- and low - velocity blue plume objects appear inter - mixed on the sky .
this result is consistent with absence of rotation seen in the h i gas @xcite . given that detailed h i maps show velocity structure in leo a @xcite , it is possible that additional observations may reveal structure in the stellar radial velocity distribution .
lrcccccl 9:59:15.124 & 30:44:10.40 & b & @xmath47 & 19.896 & 0.761 & -0.279 + 9:59:16.940 & 30:43:48.22 & b & @xmath48 & 19.050 & -0.021 & -0.353 + 9:59:20.223 & 30:43:52.71 & b & @xmath49 & 19.435 & 0.458 & -0.375 + 9:59:23.220 & 30:45:06.23 & b & @xmath50 & 20.026 & 0.584 & -0.347 + 9:59:24.909 & 30:44:36.69 & b & @xmath51 & 19.797 & 1.004 & -0.257 + 9:59:25.980 & 30:46:10.44 & b & @xmath52 & 20.964 & 0.652 & -0.397 + 9:59:26.351 & 30:45:26.09 & b & @xmath53 & 19.131 & 0.412 & -0.300 + 9:59:27.058 & 30:45:38.79 & b & @xmath54 & 20.267 & 0.982 & -0.388 + 9:59:27.326 & 30:45:44.69 & b & @xmath55 & 20.661 & 1.082 & -0.420 + 9:59:27.532 & 30:44:57.75 & hii & @xmath56 & 19.984 & -0.238 & -0.607 + 9:59:32.129 & 30:43:48.55 & b & @xmath57 & 20.471 & 0.383 & -0.360 + 9:59:33.320 & 30:44:39.21 & hii & @xmath58 & 19.520 & -0.288 & -0.252 + we now estimate the kinematic mass of leo a. because there is no evidence for rotation , we assume that the galaxy is in pressure equilibrium and apply two simple mass estimators : the virial theorem , and the projected mass estimator of @xcite .
the virial mass is given by @xmath59 , where @xmath60 is the effective radius in pc and @xmath61 is the one - dimensional velocity dispersion in km s@xmath1 .
our objects are located inside a radius of @xmath62 pc , while leo a s observed stellar distribution extends to a radius of @xmath63 pc .
if we choose @xmath64 pc , leo a s virial mass is @xmath65 . the virial theorem , however , is both more biased and less stable for small numbers of test particles than is the projected mass estimator @xcite
. thus we use the @xcite projected mass estimator to obtain a more accurate estimate of leo a s mass : @xmath66 where @xmath67 is the gravitational constant , @xmath68 is the number of stars , @xmath39 is an empirical correction to the center of mass ( heisler et al .
use @xmath69 ) , @xmath70 is the velocity relative to the mean , @xmath71 is the projected separation from the center of the galaxy , and @xmath72 is a constant that depends on the eccentricity of the stellar orbits . for purely isotropic orbits @xmath73 , while for purely radial orbits@xmath74 . using the velocities and positions in table [ tab : obs ] , leo a has a kinematic mass of @xmath75 for purely isotropic orbits and @xmath76 for purely radial orbits .
heisler et al .
prefer using the smaller mass derived from isotropic orbits , but for purposes of discussion , we will assume that leo a s mass is the average of the two projected mass estimates : @xmath77 .
this mass is derived from objects inside a radius of @xmath62 pc . by comparison ,
@xcite estimate that leo a s stellar mass is @xmath78 m@xmath79 , consistent with the galaxy s optical luminosity .
more recently , @xcite use _ spitzer _ 4.5 imaging to estimate that leo a s total stellar mass is @xmath80 m@xmath79 with an uncertainty of 0.5 dex .
these stellar mass estimates are factors of 20 - 100 times smaller than our kinematic mass estimate .
leo a s total @xmath81-band luminosity is @xmath82 , which comes from the apparent magnitude @xmath83 @xcite and the distance modulus @xmath29 @xcite . assuming the sun has @xmath84 , leo a s total luminosity in solar units is @xmath85 .
thus the mass - to - light ratio of leo a is @xmath86 for a mass of @xmath87 . because our spectroscopic targets do not sample the full extent of leo a , this mass - to - light ratio is a lower limit to leo a s true mass - to - light ratio .
a mass - to - light ratio of 20 suggests that leo a is dominated by dark matter .
@xcite reach the opposite conclusion from their h i velocity dispersion , but their result is explained by the revision of leo a s distance from 2.2 mpc to 800 kpc .
leo a s total h i mass within @xmath88 is @xmath89 @xcite at a distance of 800 kpc .
the gas mass includes the 10% correction for helium gas .
thus baryonic matter stars plus gas accounts for at most @xmath520% of leo a s total mass .
we have obtained spectroscopy for twelve blue plume objects in the central @xmath3 of leo a. ten of these objects are young b supergiants .
we measure a stellar velocity dispersion of @xmath0 km s@xmath1 , identical to leo a s h i gas dispersion @xcite . from this
we estimate a projected mass of @xmath90 , which implies that leo a s mass is at least @xmath580% dark matter .
dwarf galaxies are thought to be the smallest bodies containing dynamically significant amounts of dark matter , and so it is interesting to place leo a in the context of cosmological simulations .
@xcite show that the velocity dispersion of dark matter halos follow a tight correlation with total mass , @xmath91 , where @xmath92 is the mass within a sphere with mean interior density 200 times the critical density .
leo a s mass , @xmath93 , would fill such a sphere with a radius of @xmath94 kpc .
the halo virial relation is derived from @xmath95 dark matter halos , but evrard et al .
show it is valid down to @xmath96 halos .
if we simply equate leo a s mass to @xmath92 , the halo virial relation predicts @xmath97 km s@xmath1 for @xmath98 .
this prediction is less than half of the observed velocity dispersion .
one possible explanation for the discrepancy is that leo a has not reached dynamical equilibrium , and thus its velocity dispersion is inflated @xcite . or , perhaps the discrepancy suggests that dwarfs like leo a experience a different evolutionary path than a purely hierarchical growth of dark matter halos .
remarkably , leo a s stellar velocity dispersion is very similar to that of local group dwarf spheroidals ( dsphs ) , which have central velocity dispersions of 8 to 10 km s@xmath1 @xcite .
one explanation for the common central velocity dispersion is that all local group dwarfs are enclosed in dark matter halos of similar total mass @xcite .
galaxies with smaller velocity dispersions ( total mass @xmath99 ) are possibly re - ionized and thus never form stars ( e.g. * ? ? ?
if this picture is correct , then the total mass to light ratio of a dwarf is a function of its luminosity @xmath100 , where @xmath101 is the fixed dark matter halo mass , @xmath102 is the total @xmath81-band luminosity , and @xmath103 is the stellar mass to light ratio . in figure
[ fig : ml ] we plot the @xmath104 versus @xmath81-band luminosity for local group dsphs with central velocity dispersion measurements .
we note that leo a has a central velocity dispersion and no observed rotation , thus its dynamical mass is directly comparable with dsphs .
dwarf irregulars have masses determined from rotation and are not directly comparable .
we base figure [ fig : ml ] on the @xcite version of @xcite s plot .
the solid squares are and ii @xcite , and ix @xcite , botes @xcite , carina and sextans @xcite , draco and ursa minor @xcite , fornax @xcite , leo i @xcite , leo ii and sculptor @xcite , and ursa major @xcite .
the dashed line shows the @xmath104 relation for a fixed dark matter halo mass @xmath105 and stellar @xmath106 @xcite .
leo a , plotted as a star , falls very near the fixed halo mass relation for dsphs .
however , leo a s stellar population is quite different from that of the dsphs .
integrated colors provide a quantitative measure of the difference : leo a has @xmath107 , systematically bluer than the average dsph with @xmath108 @xcite . as its stellar population ages ,
leo a s luminosity will decrease and its mass to light ratio will increase .
we estimate this change using starburst99 @xcite with @xmath109 padova tracks .
we find that in a couple of gyr , assuming leo a has no further star formation , it will reach @xmath110 and will have faded @xmath111 magnitudes in @xmath112 .
we indicate this evolution with the arrow in figure [ fig : ml ] .
leo a still falls well within the observed scatter around the @xmath104 relation .
comparing leo a with `` transition dwarfs '' may be more fair than comparing with dsphs .
transition dwarfs have old stellar populations like dsphs , but also contain gas and young stars like leo a. a central velocity dispersion is available for the phoenix transition dwarf @xcite ( the open square in figure [ fig : ml ] ) , which places it squarely on the @xmath104 relation .
thus , despite their different star formation histories , leo a , phoenix , and the dsphs appear to share remarkably similar kinematics and dark matter halo mass .
if transition dwarfs represent the stage between gas - rich dwarf irregulars and gas - poor dsphs , this evolution must involve some amount of galaxy interaction .
most dsphs in the local group are located near the major spirals , so the dsphs lack of gas and young stars likely results from repeated gravitational and/or hydrodynamic interactions with the spirals . in a comprehensive study of minor galaxy interactions in the sdss ,
@xcite find that the lowest luminosity galaxies in close pairs experience the largest fractional boosts in their specific star formation rates .
perhaps leo a s episodic star formation history is a history of its interactions with objects in the local group .
one clue to the evolution of transition dwarfs in the local group may come from comparison of the rotation velocity and central velocity dispersion .
a wide variety of studies demonstrate the relation between these kinematic measures and the formation history of galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
multi - slit spectrographs can now provide radial velocities for hundreds of stars in nearby dwarfs , making such studies possible for the first time .
we thank k. rines for helpful discussions and thank the referee for comments that improved this paper .
we thank m. alegria , j. mcafee , and a. milone for their assistance with observations obtained at the mmt observatory , a joint facility of the smithsonian institution and the university of arizona .
this project made use of data products from the sloan digital sky survey , which is managed by the astrophysical research consortium for the participating institutions .
this research made use of the smithsonian / nasa astrophysics data system bibliographic services .
this work was supported in part by w. brown s clay fellowship and by the smithsonian institution .
, m. i. , kleyna , j. t. , wyn evans , n. , gilmore , g. f. , read , j. i. , koch , a. , grebel , e. k. , & irwin , m. j. 2006 , in eas publications series , ed .
g. a. mamon , f. combes , c. deffayet , & b. fort , 105112 | we measure the first stellar velocity dispersion of the leo a dwarf galaxy , @xmath0 km s@xmath1 .
we derive the velocity dispersion from the radial velocities of ten young b supergiants and two hii regions in the central region of leo a. we estimate a projected mass of @xmath2 within a radius of @xmath3 , and a mass to light ratio of at least @xmath4 .
these results imply leo a is at least @xmath580% dark matter by mass . |
btev is an experiment designed to explore heavy flavor phenomenology thoroughly @xcite , with particular emphasis on mixing , @xmath2 violation and rare and forbidden decays .
the pixel detector system is crucial to the experiment .
we have engaged in a thorough study of several different factors influencing the sensor performance , with particular emphasis on spatial resolution and detector occupancy . as the radiation dose is very intense at small distances from the beam axis ,
our main emphasis is on @xmath3 sensors @xcite , where the collected charge carriers are electrons .
we report the most important results on the expected performance of the btev baseline front end - sensor systems , as well as a comparison between our predictions and results recently obtained in a test beam run that took place at fermilab , where we took extensive data using different prototype sensors and readout electronics @xcite .
we have investigated several factors affecting the spatial resolution , in particular charge diffusion , magnetic field , electronic noise , discriminator threshold and digitization resolution .
we have modeled the signal induced by minimum ionizing tracks traversing silicon using a charge straggling distribution function supported by experimental data and a detailed theoretical model of the interactions responsible for the energy loss in silicon @xcite .
the detector has been conceptually divided into @xmath4 30 @xmath0 m thick slices to model fluctuations in the energy deposition along the charged particle track path .
the most probable charge signal produced by a minimum ionizing particle in the 280 @xmath0 m p - stop sensors used in the test beam data is found to be 24000 @xmath5 with a full width at half maximum of 10000 @xmath5 , in good agreement with the corresponding parameters of the measured distribution @xcite .
electrons and holes produced by the energy deposited by a traversing charged particle drift along the electric field lines ( @xmath6 ) in the detector .
the equations describing this drift motion are : @xmath7 where @xmath8 is the magnitude of the electron charge , @xmath9 and @xmath10 are the mobility of electrons and holes respectively , @xmath11 is the number of free electrons per unit volume , and @xmath12 is number of free holes per unit volume . the charge cloud spreads laterally due to diffusion .
the parameters characterizing the drift in the electric field ( @xmath10 , @xmath9 ) are related to the parameter describing the diffusion of the charge cloud ( @xmath13,@xmath14 ) by the einstein equation : @xmath15 where @xmath13 and @xmath14 are the diffusion coefficients and @xmath16 is the product of the boltzmann constant and the absolute temperature of the silicon .
the average square deviation with respect to the trajectory of the collected charge without diffusion is @xmath17 . in our study
we have used @xmath18 @xcite .
a magnetic field perpendicular to @xmath6 produces a distortion of the collected charge distribution parameterized in terms of the hall mobility @xmath19 , proportional to the drift mobility .
finally , the charge - cloud is collected on more pixels when the incident track crosses the detector at an angle , as the generation points of the electron - hole pairs spread out along the track path .
these various effects are illustrated in fig .
[ chdist ] as a function of the incident angle @xmath20 .
the discriminator threshold also influences the spatial resolution . in this study
we have assumed that only the pixels having a signal above threshold are read out , and we have varied the threshold for analog and binary readout .
[ resth ] ( top ) shows the effect of increasing the threshold for an incident angle @xmath21 mr , for analog and digital readout respectively .
the bottom plot shows the fraction of events having @xmath22 pixels hit for a given threshold .
for instance , for a threshold of 1000 electrons , about 60% of the events have 3 pixels hit , and about 40% have 2 pixels hit .
the resolution achievable for binary readout shows a characteristic oscillatory behavior as we change the threshold . as the digital clustering algorithm exploits the information provided by the number of pixels in a cluster , its accuracy is best when there is an almost equal population in two different cluster sizes : the smaller cluster corresponding to a track incident in a pixel center and the bigger cluster corresponding to incidence close to the boundary between two pixels . in the analog readout case ,
the accuracy of any position reconstruction algorithm is degraded as the threshold increases .
a pixel detector has the potential for being a very low noise system , since the capacitive load at the input of the charge sensitive preamplifier is negligible .
we have achieved @xcite 100 @xmath5 noise or below on the test bench . in order to take full advantage of the low noise ,
the minimum threshold spread among channels needs to be small .
the measured spread in the devices used in the test beam run is about 380 @xmath5 @xcite .
prototypes of the more advanced version of this design show that a threshold spread of 180 @xmath5 or better is achievable .
therefore noise and threshold spread figures are not a limiting factor in the detector performance .
an analog readout is a preferred solution for several reasons , including more effective monitoring of the stability of the detector properties and improvement in the spatial resolution .
we need to know how many bits are required because the analogue circuit must fit within the small pixel cell boundary .
furthermore , we need to extract the digitized information very quickly .
we have thus investigated several different combinations of adc transfer functions , varying both the dynamic range and the digitization accuracy .
[ fpix2 ] shows an example where the expected performance of our analog front end electronics combined with a logarithmic 3 bit flash adc is shown . as a comparison , the resolution expected from a 8 bit linear adc
is included .
note that the difference is not too dramatic .
this supports our conclusion that the digitization resolution of 3 bits , chosen for our front end electronics , fpix2 , is optimal for our application . in order for this simulation tool to be effective
, its accuracy must be checked with experimental data .
we have done a systematic study of the performance expected from various sensor and readout electronics combinations used in the recently completed test beam run .
details on the data taken and the analysis procedure are given elsewhere in these proceedings @xcite .
the comparison between predicted and measured resolution is shown in fig .
[ resmix ] for two different digitization accuracies ( 8 bit and 2 bit adcs ) as well as binary readout .
the data for the two adcs tested were taken with different thresholds .
[ resmix ] includes a simulation of the 2 bit adc case with the lower threshold , showing that the higher threshold used with the coarser adc resolution is a dominant effect , in particular at small incidence angles .
the agreement with the data is very good , especially if we take into account that factors like imperfect alignment , track projection errors and angular resolution have not yet been included .
the monte carlo simulation algorithm discussed in this paper has been a key element in our optimization of the baseline design of the btev pixel detector system .
the good agreement with recent test beam data gives us confidence that the most important factors affecting the pixel performance are modeled accurately and that this baseline design will be an excellent tool to achieve our physics goals .
while the results presented here are specific to the btev experiment , the algorithm developed is quite general and can be applied to other detector systems .
we thank d. christian , s. kwan , r. mountain and s. stone for interesting discussions .
ma thanks the organizers for a very informative workshop , in a beautiful setting .
this work was supported by the u.s
. national science foundation .
999 the btev collaboration , _ the btev proposal _ ( 2000 ) ; http://www - btev.fnal.gov / public_documents / btev_proposal/. r. horisberger , _ nucl .
_ * a384 * ( 1996 ) 185 .
g. chiodini , _ these proceedings .
_ h. bichsel _ rev .
phys . _ * 60 * ( 1988 ) 663 .
scharfetter ( rd19 ) , _ active pixel detectors for large hadron colliders _ , cern thesis , ( 1997 ) .
a. mekkaoui , _ these proceedings . | a monte carlo simulation has been developed to predict the spatial resolution of silicon pixel detectors .
the results discussed in this paper focus on the unit cell geometry of 50 @xmath0 m @xmath1 400 @xmath0 m , as chosen for btev .
effects taken into account include energy deposition fluctuations along the charged particle path , diffusion , magnetic field , and response of the front end electronics .
we compare our predictions with measurements from a recent test beam study performed at fermilab .
pixel ; silicon ; |
the eigenvalue statistics of large random hermitian matrices with independent entries are known to exhibit universal behaviour .
wigner proved @xcite that the eigenvalue density converges ( on the macroscopic scale ) to the semicircle law as the dimension of the matrix tends to infinity .
he also observed that the local statistics of individual eigenvalues ( e.g. the gap statistics ) are universal , in the sense that they depend only on the symmetry class of the matrix but are otherwise independent of the distribution of the matrix entries . in the gaussian case ,
the local spectral statistics were identified by gaudin , mehta , and dyson @xcite , who proved that they are governed by the celebrated sine kernel . in this paper and its companion @xcite , we focus the universality of the eigenvalue density statistics on intermediate , so - called _ mesoscopic _ , scales , which lie between the macroscopic and the local scales .
we study _ random band matrices _ , commonly used to model quantum transport in disordered media . unlike the mean - field wigner matrices ,
band matrices possess a nontrivial spatial structure .
apart from the obvious mathematical interest , an important motivation for this question arises from physics , namely from the theory of conductance fluctuations developed by thouless @xcite . in the next sections we explain the physical background of the problem .
thus , readers mainly interested in the mathematical aspects of our results may skip much of the introduction . according to the anderson metal - insulator transition @xcite , general disordered quantum systems are believed to fall into one of two very distinctive regimes . in the _ localized regime _ ( also called the _ insulator regime _ ) , physical quantities depending on the position , such as eigenvectors and resolvent entries , decay on a length scale @xmath0 ( called the _ localization length _ ) that is independent of the system size . the unitary time evolution generated by the hamiltonian
remains localized for all times and the local spectral statistics are poisson . in contrast , in the _ delocalized regime _
( also called the _ metallic regime _ ) , the localization length is comparable with the linear system size .
the overlap of the eigenvectors induces strong correlations in the local eigenvalue statistics , which are believed to be universal and to coincide with those of a gaussian matrix ensemble of the appropriate symmetry class .
moreover , the unitary time evolution generated by the hamiltonian is diffusive for large times .
strongly disordered systems are in the localized regime . in the weak disorder regime ,
the localization properties depend on the dimension and on the energy . despite compelling theoretical arguments and numerical evidence ,
the anderson metal - insulator transition has been rigorously proved only in a few very special cases .
the basic model is the random schrdinger operator , @xmath1 , typically defined on @xmath2 or on a graph ( e.g. on a subset of @xmath3 ) . here
@xmath4 is a random potential with short - range spatial correlations ; for instance , in the case of a graph , @xmath4 is a family of independent random variables indexed by the vertices .
the localized regime is relatively well understood since the pioneering work of frhlich and spencer @xcite , followed by an alternative approach by aizenman and molchanov @xcite .
the poissonian nature of the local spectral statistics was proved by minami @xcite . on the other hand ,
the delocalized regime has seen far less progress . with the exception of the bethe lattice @xcite ,
only partial results are available .
they indicate delocalization and quantum diffusion in certain limiting regimes @xcite , or in a somewhat different model where the static random potential is replaced with a dynamic phonon field in a thermal state at positive temperature @xcite .
another much studied family of models describing disordered quantum systems is random matrices .
delocalization is well understood for random wigner matrices @xcite , but , owing to their mean - field character , they are always in the delocalized regime , and hence no phase transition takes place .
the local eigenvalue statistics are universal .
this fundamental fact about random matrices , also known as the wigner - dyson - mehta conjecture , has been recently proved @xcite ( see also @xcite for a partially alternative argument in the hermitian case ) . in a seminal paper @xcite ,
altshuler and shklovskii computed a new physical quantity : the variance of the number @xmath5 of eigenvalues on a mesoscopic energy scale @xmath6 in @xmath7-dimensional metallic samples with disorder for @xmath8 ; here _ mesoscopic _ refers to scales @xmath6 that are much larger then the typical eigenvalue spacing @xmath9 but much smaller than the total ( macroscopic ) energy scale of the system .
their motivation was to study fluctuations of the conductance in mesoscopic metallic samples ; see also @xcite and @xcite .
the relationship between @xmath10 and the conductance is given by a fundamental result of thouless @xcite , asserting that the conductance of a sample of linear size @xmath11 is determined by the ( one - particle ) energy levels in an energy band of a specific width @xmath6 around the fermi energy .
in particular , the variance of @xmath12 directly contributes to the conductance fuctuations .
this specific value of @xmath6 is given by @xmath13 , where @xmath14 is the temperature and @xmath15 is the _ thouless energy _
@xcite . in diffusive models
the thouless energy is defined as @xmath16 , where @xmath17 is the diffusion coefficient .
( in a conductor the dynamics of the particles , i.e. the itinerant electrons , is typically diffusive . )
the thouless energy may also be interpreted as the inverse diffusion time , i.e. the time needed for the particle to diffuse through the sample . as it turns out , the mesoscopic linear statistics @xmath10 undergo a sharp transition precisely at to indicate that @xmath18 and @xmath19 have comparable size .
see the conventions at the end of section [ sec : intro ] . ]
for small energy scales , @xmath21 , altshuler and shklovskii found that the variance of @xmath10 behaves according to to denote the covariance and abbreviate @xmath22 .
see below . ] @xmath23 as predicted by the dyson - mehta statistics @xcite .
the unusually small variance is due to the strong correlations among eigenvalues ( arising from level repulsion ) . in the opposite regime , @xmath24 ,
the variance is typically much larger , and behaves according to @xmath25 the threshold @xmath15 may be understood by introducing the concept of ( an energy - dependent ) _ diffusion length _
@xmath26 , which is the typical spatial scale on which the off - diagonal matrix entries of those observables decay that live on an energy scale @xmath6 ( e.g. resolvents whose spectral parameters have imaginary part @xmath6 ) .
alternatively , @xmath26 is the linear scale of an initially localized state evolved up to time @xmath27 .
the diffusion length is related to the localization length @xmath0 through @xmath28 .
assuming that the dynamics of the quantum particle can be described by a classical diffusion process , one can show that @xmath29 and the relation @xmath30 may be written as @xmath31 .
the physical interpretation is that the sample is so small that the system is essentially mean - field from the point of view of observables on the energy scale @xmath6 , so that the spatial structure and dimensionality of the system are immaterial .
the opposite regime @xmath24 corresponds to large samples , @xmath32 , where the behaviour of the system can be approximated by a diffusion that has not reached the boundary of the sample .
these two regimes are commonly referred to as _ mean - field _ and _ diffusive _ regimes , respectively .
a similar transition occurs if one considers the correlation of the number of eigenvalues @xmath33 and @xmath34 around two distinct energies @xmath35 whose separation is much larger than the energy window @xmath6 ( i.e. @xmath36 ) . for small samples , @xmath37 , the correlation decays according to @xmath38 this decay holds for systems both with and without time reversal symmetry .
the decay is in agreement with the dyson - mehta statistics , which in the complex hermitian case ( corresponding to a system without time reversal symmetry ) predict a correlation @xmath39 for highly localized observables on the scale @xmath40 . for mesoscopic scales , @xmath41 ,
the oscillations in the numerator are averaged out and may be replaced with a positive constant to yield .
a similar formula with the same decay holds for the real symmetric case ( corresponding to a system with time reversal symmetry ) . on the other hand , for large samples , @xmath24
, we have @xmath42 for @xmath43 the correlation vanishes to leading order .
the power laws in the energies @xmath6 and @xmath44 given in and respectively are called the _ altshuler - shklovskii formulas_. they express the variance and the correlation of the density of states in the regime where the diffusion approximation is valid and the spatial extent of the diffusion , @xmath26 , is much less than the system size @xmath11 . in contrast , the mean - field formulas and describe the situation where the diffusion has reached equilibrium .
note that the behaviours and as well as and are very different from the ones obtained if the distribution of the eigenvalues were governed by poisson statistics ; in that case , for instance , and would be zero . from a mathematical point of view , the significance of these mesoscopic quantities is that their statistics are amenable to rigorous analysis even in the delocalized regime . in this paper
we demonstrate this by proving the altshuler - shklovskii formulas for random band matrices .
we consider @xmath7-dimensional random band matrices , which interpolate between random schrdinger operators and mean - field wigner matrices by varying the band width @xmath45 ; see @xcite for an overview of this idea .
these matrices represent quantum systems in a @xmath7-dimensional discrete box of side length @xmath11 , where the quantum transition probabilities are random and their range is of order @xmath46 .
we scale the matrix so that its spectrum is bounded , i.e. the macroscopic energy scale is of order 1 , and hence the eigenvalue spacing is of order @xmath47 .
band matrices exhibit diffusion in all dimensions @xmath7 , with a diffusion coefficient @xmath48 ; see @xcite for a physical argument in the general case and @xcite for a proof up to certain large time scales . in @xcite
it was showed that the resolvent entries with spectral parameter @xmath49 decay exponentially on a scale @xmath50 , as long as this scale is smaller than the system size , @xmath51 .
( for technical reasons the proof is valid only if @xmath11 is not too large , @xmath52 . )
the resolvent entries do not decay if @xmath53 , in which case the system is in the mean - field regime for observables living on energy scales of order @xmath37 .
notice that the crossover at @xmath54 corresponds exactly to the crossover at @xmath20 mentioned above .
our main result is the proof of the formulas ( with @xmath55 ) and for @xmath7-dimensional band matrices for @xmath56 ; we also obtain similar results for @xmath57 , where the powers of @xmath6 and @xmath44 are replaced with a logarithm .
this rigorously justifies the asymptotics of silvestrov ( * ? ? ?
* equation ( 40 ) ) , which in turn reproduced the earlier result of @xcite . for technical reasons
, we have to restrict ourselves to the regime @xmath58 .
for convenience , we also assume that @xmath59 , which guarantees that @xmath60 ( or , equivalently , @xmath24 ) .
hence we work in the diffusive regime .
however , our method may be easily extended to the case @xmath31 as well ( see remark [ rem : diffusion ] and section [ sec : wigner ] below ) .
we also show that for @xmath61 the universality of the formulas and breaks down , and the variance and the correlation functions of @xmath5 depend on the detailed structure of the band matrix .
we also compute the leading correction to the density - density correlation . in summary
, we find that for @xmath62 the leading and subleading terms are universal , for @xmath63 only the leading terms are universal , and for @xmath64 the density - density correlation is not universal . the case @xmath43 is special , since the coefficient of the leading term in vanishes owing to an algebraic cancellation .
the first non - vanishing term was predicted in @xcite .
we rigorously identify this term in the regime @xmath65 , and find a substantial discrepancy between it and the prediction of @xcite .
for an outline of our proof , and the relation between this paper and its companion @xcite , see section [ sec : struc ] .
our analysis is valid in the mesoscopic regime , i.e. when @xmath66 , and concerns only density fluctuations . for completeness
, we mention what was previously known in this and other regimes . in the macroscopic regime , @xmath67 , the quantity @xmath10 should fluctuate on the scale @xmath68 according to . for the wigner case , @xmath69
, it has been proved that a smoothed version of @xmath10 , the linear statistics of eigenvalues @xmath70 , is asymptotically gaussian . the first result in this direction for analytic @xmath71
was given in @xcite , and this was later extended by several authors to more general test functions ; see @xcite for the latest result .
the first central limit theorem for matrices with a nontrivial spatial structure and for polynomial test functions was proved in @xcite .
very recently , it was proved in @xcite for one - dimensional band matrices that , provided that @xmath72 , the quantity @xmath73 is asymptotically gaussian with variance of order @xmath68 . for a complete list of references in this direction ,
see @xcite .
the asymptotics in the completely mean - field case , corresponding to wigner matrices ( i.e. @xmath74 so that @xmath75 ) , was proved in @xcite ; see the remarks following theorem [ thm : wigner ] for more details about this work .
we note that the formula for random band matrices with @xmath76 was derived in @xcite , using an unphysical double limit procedure , in which the limit @xmath77 was first computed for a fixed @xmath6 , and subsequently the limit of small @xmath6 was taken .
note that the mesoscopic correlations can not be recovered after the limit @xmath77 .
hence the result of @xcite describes only the macroscopic , and not the mesoscopic , correlations .
much less is known about the local spectral statistics of random band matrices , even for @xmath76 .
the tracy - widom law at the spectral edge was proved in @xcite .
based on a computation of the localization length , the metal - insulator transition is predicted to occur at @xmath78 ; see @xcite for a non - rigorous argument and @xcite for the best currently known lower and upper bounds .
hence , the local spectral statistics are expected to be governed by the sine kernel from random matrix theory in the regime @xmath79 .
very recently , the sine kernel was proved @xcite for a special hermitian gaussian random band matrix with band width @xmath45 comparable with @xmath11 .
universality for a more general class of band matrices but with an additional tiny mean - field component was proved in @xcite .
we also mention that the local correlations of determinants of a special hermitian gaussian random band matrix have been shown to follow the sine kernel @xcite , up to the expected threshold @xmath80 .
the diagrammatic calculation of @xcite uses the diffusion approximation , and formulas are supposed to be valid in the delocalized regime .
nevertheless , our results also hold in the localized regime , in particular even @xmath76 and for @xmath81 , in which case the eigenvectors are known to be localized @xcite . in this regime , and for @xmath82 , we also prove and .
both formulas show that poisson statistics do not hold on large mesoscopic scales , despite the system being in the localized regime . indeed ,
if @xmath10 were poisson - distributed , then we would have @xmath83 . on the other hand ,
gives @xmath84 .
we conclude that the prediction of for the magnitude of @xmath85 is much smaller than that predicted by poisson statistics provided that @xmath86 .
the fact that the poisson statistics breaks down on mesoscopic scales is not surprising .
indeed , the basic intuition behind the emergence of poisson statistics is that eigenvectors belonging to different eigenvalues are exponentially localized on a scale @xmath87 , typically at different spatial locations .
hence the associated eigenvalues are independent . for larger @xmath6 , however , the observables depend on many eigenvalues , which exhibit nontrivial correlations since the supports of their eigenvectors overlap .
a simple counting argument shows that such overlaps become significant if @xmath88 , at which point correlations are expected to develop . in other words ,
we expect a transition to / from poisson statistics at @xmath89 . in the previous paragraph
, we noted that predicts a transition in the behaviour of @xmath10 to / from poisson statistics at @xmath90 . combining these observations
, we therefore expect a transition to / from poisson statistics for @xmath87 .
this argument predicts the correct localization length @xmath87 without resorting to grassmann integration .
it remains on a heuristic level , however , since our results do not cover the full range @xmath91 .
we note that this argument may also be applied to @xmath92 , in which case it predicts the absence of a transition provided that @xmath93 .
the main conclusion of our results is that the local eigenvalue statistics , characterized by either poisson or sine kernel statistics , do not in general extend to mesoscopic scales .
on mesoscopic scales , a different kind of universality emerges , which is expressed by the altshuler - shklovskii formulas .
we use @xmath94 to denote a generic large positive constant , which may depend on some fixed parameters and whose value may change from one expression to the next . similarly
, we use @xmath95 to denote a generic small positive constant .
we use @xmath96 to mean @xmath97 for some constants @xmath98 .
also , for any finite set @xmath99 we use @xmath100 to denote the cardinality of @xmath99 .
if the implicit constants in the usual notation @xmath101 depend on some parameters @xmath102 , we sometimes indicate this explicitly by writing @xmath103 .
we are very grateful to alexander altland for detailed discussions on the physics of the problem and for providing references .
fix @xmath104 , the physical dimension of the configuration space . for @xmath105
we define the discrete torus of size @xmath11 @xmath106 and abbreviate @xmath107 let @xmath108 denote the band width , and define the deterministic matrix @xmath109 through @xmath110 where @xmath111 denotes the periodic euclidean norm on @xmath112 , i.e. @xmath113 . note that @xmath114 the fundamental parameters of our model are the linear dimension of the torus , @xmath11 , and the band width , @xmath45 .
the quantities @xmath115 and @xmath116 are introduced for notational convenience , since most of our estimates depend naturally on @xmath115 and @xmath116 rather than @xmath11 and @xmath45 .
we regard @xmath11 as the independent parameter , and @xmath117 as a function of @xmath11 . next , let @xmath118 be a hermitian random matrix whose upper - triangular on the torus @xmath112 . ]
entries @xmath119 are independent random variables with zero expectation .
we consider two cases . *
_ the real symmetric case ( @xmath120 ) _ , where @xmath121 satisfies @xmath122 . * _ the complex hermitian case ( @xmath123 ) _ , where @xmath121 is uniformly distributed on the unit circle @xmath124 . here
the index @xmath125 is the customary symmetry index of random matrix theory .
we define the _ random band matrix _ @xmath126 through @xmath127 note that @xmath128 is hermitian and @xmath129 , i.e. @xmath130 is deterministic .
moreover , we have for all @xmath131 @xmath132 with this normalization , as @xmath133 the bulk of the spectrum of @xmath134 lies in @xmath135 $ ] and the eigenvalue density is given by the wigner semicircle law with density @xmath136\,.\ ] ] let @xmath71 be a smooth , integrable , real - valued function on @xmath137 satisfying @xmath138 .
we call such functions @xmath71 _ test functions_. we also require that our test functions @xmath71 satisfy one of the two following conditions .
* @xmath71 is the cauchy kernel @xmath139 * for every @xmath140 there exists a constant @xmath141 such that @xmath142 a typical example of a test function @xmath71 satisfying * ( c2 ) * is the gaussian @xmath143 .
we introduce the rescaled test function @xmath144 .
we shall be interested in correlations of observables depending on @xmath145 of the form @xmath146 where @xmath147 denote the eigenvalues of @xmath134 .
( the factor @xmath148 is a mere convenience , chosen because , as noted above , the asymptotic spectrum of @xmath134 is the interval @xmath135 $ ] . )
the quantity @xmath149 is the smoothed local density of states around the energy @xmath150 on the scale @xmath6 .
we always choose @xmath151 for some fixed @xmath152 , and we frequently drop the index @xmath6 from our notation .
the strongest results are for large @xmath153 , so that one should think of @xmath153 being slightly less than @xmath154 .
we are interested in the correlation function of the local densities of states , @xmath155 and @xmath156 , around two energies @xmath157 .
we shall investigate two regimes : @xmath158 and @xmath159 . in the former regime , we prove that the correlation decay in the energy difference @xmath160 is universal ( in particular , independent of @xmath6 , @xmath161 , and @xmath162 ) , and we compute the correlation function explicitly . in the latter regime , we prove that the variance has a universal dependence on @xmath6 , and depends on @xmath161 and @xmath162 via their inner product in a homogeneous sobolev space .
the case * ( c2 ) * for our test functions is the more important one , since we are typically interested in the statistics of eigenvalues contained in an interval of size @xmath6 .
the cauchy kernel from the case * ( c1 ) * has a heavy tail , which introduces unwanted correlations arising from the overlap of the test functions and not from the long - distance correlations that we are interested in .
nevertheless , we give our results also for the special case * ( c1)*. we do this for two reasons .
first , the case * ( c1 ) * is pedagogically useful , since it results in a considerably simpler computation of the main term ( see ( * ? ? ?
* section 3 ) for more details ) .
second , the case * ( c1 ) * is often the only one considered in the physics literature ( essentially because it corresponds to the imaginary part of the resolvent of @xmath128 ) .
hence , our results in particular decouple the correlation effects arising from the heavy tails of the test functions from those arising from genuine mesoscopic correlations .
as proved in theorem [ thm : theta 1 ] below , the effect of the heavy tail is only visible in the leading nonzero corrections for @xmath163 . for simplicity , throughout the following we assume that both of our test functions satisfy * ( c1 ) * or both satisfy * ( c2)*. since the covariance is bilinear , one may also consider more general test functions that are linear combinations of the cases * ( c1 ) * and * ( c2)*. throughout the following we use the quantities @xmath164 and @xmath165 interchangeably . without loss of generality we always assume that @xmath166 .
for the following we choose and fix a positive constant @xmath167 .
we always assume that @xmath168 \
, , \qquad \omega \;{\leqslant}\ ; c_*\ ] ] for some small enough positive constant @xmath169 depending on @xmath167 .
these restrictions are required since the nature of the correlations changes near the spectral edges @xmath170 . throughout the following we regard the constants @xmath167 and @xmath169 as fixed and do not track the dependence of our estimates on them .
we now state our results on the density - density correlation for band matrices in the diffusive regime ( section [ sec : band_results ] ) .
the proofs are given in the current paper and its companion @xcite .
as a reference , we also state similar results for wigner matrices , corresponding to the mean - field regime ( section [ sec : wigner ] ) .
our first theorem gives the leading behaviour of the density - density correlation function in terms of a function @xmath171 , which is explicit but has a complicated form .
in the two subsequent theorems we determine the asymptotics of this function in two physically relevant regimes , where its form simplifies substantially .
we use the abbreviations @xmath172 [ thm : main result ] fix @xmath173 and @xmath104 , and set @xmath174 .
suppose that the test functions @xmath161 and @xmath162 satisfy either both * ( c1 ) * or both * ( c2)*. suppose moreover that @xmath175 for some constant @xmath94 .
then there exist a constant @xmath176 and a function @xmath177 which is given explicitly in and below , and whose asymptotic behaviour is derived in theorems [ thm : theta 2 ] and [ thm : theta 1 ] below such that , for any @xmath178 satisfying for small enough @xmath179 , the local density - density correlation satisfies @xmath180 where we defined @xmath181 moreover , if @xmath161 and @xmath162 are analytic in a strip containing the real axis ( e.g. as in the case * ( c1 ) * ) , we may replace the upper bound @xmath182 in @xmath183 for some small constant @xmath184 .
we shall prove that the error term in is smaller than the main term @xmath185 for all @xmath186 .
the main term @xmath185 has a simple , and universal , explicit form only for @xmath187 .
why @xmath57 is the critical dimension for the universality of the correlation decay is explained in section [ sec : heuristic leading term ] below .
the two following theorems give the leading behaviour of the function @xmath185 for @xmath187 in the two regimes @xmath188 and @xmath189 .
in fact , one may also compute the subleading corrections to @xmath185 .
these corrections turn out to be universal for @xmath190 but not for @xmath191 ; see theorem [ thm : theta 1 ] and the remarks following it . in order to describe the leading behaviour of the variance , i.e. the case @xmath188
, we introduce the fourier transform @xmath192 for @xmath187 we define the quadratic form @xmath193 through @xmath194 note that @xmath195 is real since both @xmath161 and @xmath162 are . in the case * ( c1 ) * we have the explicit values @xmath196 for the following statements of results , we recall the density @xmath197 of the semicircle law from , and remind the reader of the index @xmath125 describing the symmetry class of @xmath128 .
[ thm : theta 2 ] suppose that the assumptions in the first paragraph of theorem [ thm : main result ] hold , and let @xmath198 be the function from theorem [ thm : main result ] .
suppose in addition that @xmath188 .
then there exists a constant @xmath199 such that the following holds for @xmath200 satisfying . 1 .
for @xmath201 we have @xmath202 2 .
for @xmath203 we have @xmath204 in order to describe the behaviour of @xmath185 in the regime @xmath189 , for @xmath205 we introduce the constants @xmath206 explicitly , @xmath207 [ thm : theta
1 ] suppose that the assumptions in the first paragraph of theorem [ thm : main result ] hold , and let @xmath198 be the function from theorem [ thm : main result ] .
suppose in addition that @xmath208 for some arbitrary but fixed @xmath209 .
then there exists a constant @xmath199 such that the following holds for @xmath210 satisfying for small enough @xmath179 . 1 .
for @xmath211 we have @xmath212 2 . for @xmath163 does not identify the leading term since @xmath213 .
the leading nonzero correction to the vanishing leading term is @xmath214 in the case * ( c1 ) * and @xmath215 in the case * ( c2)*. 3 .
for @xmath203 we have @xmath216 note that the leading non - zero terms in the expressions , , are much larger than the additive error term in .
hence , theorems [ thm : main result ] and [ thm : theta 2 ] give a proof of the first altshuler - shklovskii formula , .
similarly , theorems [ thm : main result ] and [ thm : theta 1 ] give a proof of the second altshuler - shklovskii formula , . the additional term in
as compared to originates from the heavy cauchy tail in the test functions @xmath217 at large distances . in theorem [ thm :
theta 1 ] ( ii ) , we give the leading correction , of order @xmath218 , to the vanishing main term for @xmath163 .
similarly , for @xmath219 one can also derive the leading correction to the nonzero main term ( which is of order @xmath220 ) .
this correction turns out to be of order @xmath221 ; we omit the details .
the leading term in originates from the so - called one - loop diagrams in the terminology of physics .
the next - order term after the vanishing leading term for @xmath43 ( recall that to @xmath213 ) was first computed by kravtsov and lerner ( * ? ? ?
* equation ( 13 ) ) .
they found that for @xmath120 it is of order @xmath222 and for @xmath123 even smaller , of order @xmath223 .
part ( ii ) of theorem [ thm : theta 1 ] shows that , at least in the regime @xmath224 , the true behaviour is much larger .
the origin of this term is a more precise computation of the one - loop diagrams , in contrast to @xcite where the authors attribute the next - order term to the two - loop diagrams .
( see ( * ? ? ?
* section 3 ) for more details . ) if the distribution of the eigenvalues @xmath225 of @xmath134 were governed by poisson statistics , the behaviour of the covariance would be very different .
indeed , suppose that @xmath226 is a stationary poisson point process with intensity @xmath115 .
then , setting @xmath227 and supposing that @xmath228 , we find @xmath229 where @xmath230 .
this is in stark contrast to , , , and . in particular , in the case @xmath189 the behaviour of the covariance on @xmath231 depends on the tails of @xmath161 and @xmath162 , unlike in and .
hence , if @xmath161 and @xmath162 are compactly supported then the covariance for the poisson process is zero , while for the eigenvalue process of a band matrix it has a power law decay in @xmath231 .
[ rem1 ] we emphasize that theorem [ thm : main result ] is true under the sole restrictions on @xmath231 and @xmath210 .
however , the leading term @xmath185 only has a simple and universal form in the two ( physically relevant ) regimes @xmath188 and @xmath232 of theorems [ thm : theta 2 ] and [ thm : theta 1 ] . if neither of these conditions holds , the expression for @xmath185 is still explicit but much more cumbersome and opaque .
it is given by the sum of the values of eight ( sixteen for @xmath120 ) skeleton graphs , after a ladder resummation ; these skeleton graphs are referred to as the `` dumbbell skeletons '' @xmath233 in section [ sec : main argument ] below , and are depicted in figure [ fig : dumbbell ] below .
( they are the analogues of the _ diffusion _ and _ cooperon _ feynman diagrams in the physics literature . )
[ rem : diffusion ] the upper bound in the assumption is technical and can be relaxed .
the lower bound in , however , is a natural restriction , and is related to the _ quantum diffusion _ generated by the band matrix @xmath128 . in @xcite , it was proved that the propagator @xmath234 behaves diffusively for @xmath235 , whereby the spatial extent of the diffusion is @xmath236 .
similarly , in @xcite , it was proved that the resolvent @xmath237 has a nontrivial profile on the scale @xmath238 .
( note that @xmath6 is the conjugate variable to @xmath239 , i.e. the time evolution up to time @xmath239 describes the same regime as the resolvent with a spectral parameter @xmath240 whose imaginary part is @xmath241 . ) since in theorem [ thm : main result ] we assume that @xmath242 , the condition simply states that the diffusion profile associated with the spectral resolution @xmath6 does not reach the edge of the torus @xmath112 .
thus , the lower bound in imposes a regime in which boundary effects are irrelevant .
hence we are in the diffusive regime a basic assumption of the altshuler - shklovskii formulas and .
our method can easily be applied to the case where the lower bound in is not satisfied . in this case , however , the leading behaviour @xmath243 is modified by boundary effects . to illustrate this phenomenon , we state the analogue of theorems [ thm : main result][thm : theta 1 ] for the case of @xmath244 . in this case
, the physical dimension @xmath7 in section [ sec : setup1 ] is irrelevant .
the off - diagonal entries of @xmath128 are all identically distributed , i.e. @xmath128 is a standard wigner matrix ( neglecting the irrelevant diagonal entries ) , and we have @xmath245 , and @xmath246 where @xmath247 . in particular , @xmath128 is a mean - field model in which the geometry of @xmath112 plays no role ; the effective dimension is @xmath248 . in this case
remains valid , and we get the following result .
[ thm : wigner ] suppose that @xmath249 .
fix @xmath173 and set @xmath250 .
suppose that the test functions @xmath161 and @xmath162 satisfy either both * ( c1 ) * or both * ( c2)*. then there exists a constant @xmath176 and a function @xmath251 such that for any @xmath210 satisfying for small enough @xmath179 the following holds . 1 .
the local density - density correlations satisfy @xmath252 2 . if holds then @xmath253 3 . if @xmath188 then @xmath254 where @xmath255 was defined in .
the proof of theorem [ thm : wigner ] proceeds along the same lines as that of theorems [ thm : main result][thm : theta 1 ] .
in fact , the simple form of @xmath256 results in a much easier proof ; we omit the details .
we remark that a result analogous to parts ( i ) and ( ii ) of theorem [ thm : wigner ] , in the case where @xmath257 are given by , was derived in @xcite .
more precisely , in @xcite , the authors assume that @xmath128 is a goe matrix and derive ( i ) and ( ii ) of theorem [ thm : wigner ] for any @xmath258 ; in @xcite , they extend these results to arbitrary wigner matrices under the additional constraint that @xmath259 .
moreover , results analogous to ( iii ) for the gaussian circular ensembles were proved in @xcite .
more precisely , in @xcite it is proved that in gaussian circular ensembles the appropriately scaled mesoscopic linear statistics @xmath149 with @xmath260 are asymptotically gaussian with variance proportional to @xmath261 .
we remark that for random band matrices the mesoscopic linear statistics also satisfy a central limit theorem ; see ( * ? ? ?
* corollary 2.6 ) .
the starting point of the proof is to use the fourier transform to rewrite @xmath262 , the spectral density on scale @xmath6 , in terms of @xmath263 up to times @xmath264 .
the large-@xmath239 behaviour of this unitary group has been extensively analysed in @xcite by developing a graphical expansion method which we also use in this paper .
the main difficulty is to control highly oscillating sums . without any resummation ,
the sum of the absolute values of the summands diverges exponentially in @xmath11 , although their actual sum remains bounded .
the leading divergence in this expansion is removed using a resummation that is implemented by expanding @xmath263 in terms of chebyshev polynomials of @xmath128 instead of powers of @xmath128 .
this step , motivated by @xcite , is algebraic and requires the deterministic condition @xmath265 .
( the removal of this condition is possible , but requires substantial technical efforts that mask the essence of the argument ; see section [ sec : outlook ] ) . in the jargon of diagrammatic perturbation theory
, this resummation step corresponds to the self - energy renormalization .
the goal of @xcite was to show that the unitary propagator @xmath263 can be described by a diffusive equation on large space and time scales .
this analysis identified only the leading behaviour of @xmath263 , which was sufficient to prove quantum diffusion emerging from the unitary time evolution . the quantity studied in the current paper
the local density - density correlation is considerably more difficult to analyse because it arises from higher - order terms of @xmath263 than the quantum diffusion .
hence , not only does the leading term have to be computed more precisely , but the error estimates also require a much more delicate analysis .
in fact , we have to perform a second algebraic resummation procedure , where oscillatory sums corresponding to families of specific subgraphs , the so - called ladder subdiagrams , are bundled together and computed with high precision . estimating individual ladder graphs in absolute value
is not affordable : a term - by - term estimate is possible only after this second renormalization step .
although the expansion in nonbacktracking powers of @xmath128 is the same as in @xcite , our proof in fact has little in common with that of @xcite ; the only similarity is the basic graphical language .
in contrast to @xcite , almost all of the work in this paper involves controlling oscillatory sums , both in the error estimates and in the computation of the main term .
the complete proof is given in the current paper and its companion @xcite . in order to highlight the key ideas , the current paper contains the proof assuming three important simplifications , given precisely in * ( s1)**(s3 ) * in section [ sec_41 ] below .
they concern certain specific terms in the multiple summations arising from our diagrammatic expansion .
roughly , these simplifications amount to only dealing with typical summation label configurations ( hence ignoring exceptional label coincidences ) and restricting the summation over all partitions to a summation over pairings . as explained in @xcite ,
dealing with exceptional label configurations and non - pairings requires significant additional efforts , which are however largely unrelated to the essence of the argument presented in the current paper . how to remove these simplifications , and hence complete the proofs ,
is explained in @xcite .
in addition , the precise calculation of the leading term is also given in @xcite ; in the current paper we give a sketch of the calculation ( see section [ sec : heuristic leading term ] below ) .
we close this subsection by noting that the restriction @xmath266 for the exponent of @xmath267 is technical and stems from a fundamental combinatorial fact that underlies our proof the so - called _
2/3-rule_. the 2/3-rule was introduced in @xcite and is stated in the current context in lemma [ lem:2/3_rule ] below . in ( * ? ? ?
* section 11 ) , it was shown that the 2/3-rule is sharp , and is in fact saturated for a large family of graphs . for more details on the 2/3-rule and
how it leads the the restriction on @xmath153 , we refer to the end section [ sec : large_sigma ] below .
we conclude this section by summarizing some extensions of our results from the companion paper @xcite .
first , our results easily extend from the two - point correlation functions of to arbitrary @xmath268-point correlation functions of the form @xmath269 in ( * ? ? ?
* theorem 2.5 ) , we prove that the joint law of the smoothed densities @xmath270 is asymptotically gaussian with covariance matrix @xmath271 , given by the altshuler - shklovskii formulas .
this result may be regarded as a wick theorem for the mesoscopic densities , i.e. a central limit theorem for the mesoscopic linear statistics of eigenvalues . in particular , if @xmath272 , the finite - dimensional marginals of the process @xmath273 converge ( after an appropriate affine transformation ) to those of a gaussian process with covariance @xmath274 .
second , in ( * ? ? ?
* section 2.4 ) we introduce a general family of band matrices , where we allow the second moments @xmath275 and @xmath276 to be arbitrary translation - invariant matrices living on the scale @xmath45 . in particular , we generalize the sharp step profile from and relax the deterministic condition @xmath277 .
note that we allow @xmath278 to be arbitrary up to the trivial constraint @xmath279 , thus embedding the real symmetric matrices and the complex hermitian matrices into a single large family of band matrices .
in particular , this generalization allows us to probe the transition from @xmath120 to @xmath123 by rotating the entries of @xmath128 or by scaling @xmath278 .
note that @xmath280 corresponds to the real symmetric case , while @xmath281 corresponds to a complex hermitian case where the real and imaginary parts of the matrix elements are uncorrelated and have the same variance .
we can combine this rotation and scaling into a two - parameter family of models ; roughly , we consider @xmath282 where @xmath283 $ ] are real parameters .
we show that the mesoscopic statistics described by theorems [ thm : theta 2 ] and [ thm : theta 1 ] take on a more complicated form in the case of the general band matrix model ; they depend on the additional parameter @xmath284 , which also characterizes the transition from @xmath120 ( small @xmath285 ) to @xmath123 ( large @xmath285 ) .
we refer to ( * ? ? ?
* section 2.4 ) for the details .
since the left - hand side of is invariant under the scaling @xmath286 for @xmath287 , we assume without loss of generality that @xmath288 for @xmath289
. we shall make this assumption throughout the proof without further mention .
we expand @xmath290 in nonbacktracking powers @xmath291 of @xmath128 , defined through @xmath292 from @xcite , section 5 , we find that @xmath293 where @xmath294 is the @xmath295-th chebyshev polynomial of the second kind , defined through @xmath296 the identity first appeared in @xcite .
note that it requires the deterministic condition @xmath277 on the entries of @xmath128 .
however , our basic approach still works even if this condition is not satisfied ; in that case the proof is more complicated due to the presence of a variety of error terms in . see ( * ? ? ?
* section 5.3 ) for more details . from @xcite
, lemmas 5.3 and 7.9 , we recall the expansion in nonbacktracking powers of @xmath128 .
[ lemma : uexp ] for @xmath297 we have @xmath298 where @xmath299 and @xmath300 denotes the @xmath301-th bessel function of the first kind . throughout the following we denote by @xmath302 the analytic branch of @xmath302 extended to the real axis by continuity from the upper half - plane .
the following coefficients will play a key role in the expansion . for @xmath303 and @xmath304 define @xmath305 we have @xmath306 using we find @xmath307 where in the second step we used the identity @xmath308 which is an easy consequence of ( * ? ? ?
* formulas 6.693.16.693.2 ) and analytic continuation .
this concludes the proof .
define @xmath309 where we used the notation .
note that the left - hand side of may be written as @xmath310 the expectations in the denominator are easy to compute using the local semicircle law for band matrices ; see lemma [ lem : ey ] below .
our main goal is to compute @xmath311 . throughout
the following we use the abbreviation @xmath312 and define @xmath313 , @xmath314 , and @xmath315 similarly similarly in terms of @xmath316 , @xmath317 , and @xmath318 .
we also use the notation @xmath319 to denote convolution . the normalizing factor @xmath320
is chosen so that @xmath321 .
observe that @xmath322 we note that in the case where @xmath323 , we have @xmath324 . hence implies @xmath325 this may also be interpreted using the identity @xmath326 we now return to the case of a general real @xmath71 . since @xmath71 is real , we have @xmath327 .
we may therefore use lemma [ lemma : uexp ] and fourier transformation to get @xmath328 where @xmath329 denotes the hermitian part of a matrix , i.e. @xmath330 , and in the last step we used and the fact that @xmath291 is hermitian .
we conclude that @xmath331 because the combinatorial estimates of section [ sec : main argument ] deteriorate rapidly for @xmath332 , it is essential to cut off the terms @xmath333 in the expansion , where @xmath334 . thus , we choose a cutoff exponent @xmath335 satisfying @xmath334 .
all of the estimates in this paper depend on @xmath336 , and @xmath71 ; we do not track this dependence .
the following result gives the truncated version of , whereby the truncation is done in @xmath337 and in the support of @xmath338 .
[ prop : expansion with trunction ] choose @xmath339 and @xmath340 satisfying @xmath341 . define @xmath342 and @xmath343 let @xmath140 be arbitrary .
then for any @xmath344 and recalling we have the estimates @xmath345 and @xmath346 moreover , for all @xmath140 we have @xmath347 if @xmath161 and @xmath162 are analytic in a strip containing the real axis , the factors @xmath348 on the right - hand sides of and may be replaced with @xmath349 for some @xmath184 , and the factor @xmath350 on the right - hand side of by @xmath351 .
the proof of proposition [ prop : expansion with trunction ] is given in ( * ? ? ?
* appendix a ) . at this point
we make a short digression to outline how we compute the leading term of @xmath352 .
the precise calculation is given in the companion paper @xcite . in section [ sec : main argument ] below , we express the right - hand side of as a sum of terms indexed by graphs , reminiscent of feynman graphs in perturbation theory .
we prove that the leading contribution is given by a certain set of relatively simple graphs , which we call the _
dumbbell skeletons_. their value @xmath353 may be explicitly computed and is essentially given by @xmath354 ( see below for the precise statement ) .
the summations represent `` ladder subdiagram resummations '' in the terminology of graphs .
proving that the contribution of all other graphs is negligible , and hence that gives the leading behaviour of , represents the main work , and is done in sections [ sec : main argument ] . assuming that this approximation is valid , we compute as follows .
we use @xmath355 on the right - hand side of , and only consider the first resulting term ; the second one will turn out to be subleading in the regime @xmath356 , owing to a phase cancellation . recalling the definition of @xmath357 from , we find that the summations over @xmath358 are simply geometric series , so that @xmath359 where we abbreviated @xmath360 , and wrote , by a slight abuse of notation , @xmath361 . in order to understand the behaviour of this expression
, we make some basic observations about the spectrum of @xmath256 .
since @xmath256 is translation invariant , i.e. @xmath362 , it may be diagonalized by fourier transformation , @xmath363 where @xmath364 is the normalized indicator function of the unit ball in @xmath2 ; in the last step we used the definition of @xmath256 and a riemann sum approximation .
( note that , since @xmath256 lives on the scale @xmath45 , it is natural to rescale the argument @xmath365 of the fourier transform by @xmath366 . ) from this representation it is not hard to see that @xmath367 for some constant @xmath184 . moreover , for small @xmath365 we may expand @xmath368 to get @xmath369 , where we defined the covariance matrix differs from the @xmath17 used in the introduction by a factor of order @xmath370 .
] @xmath371 of @xmath256 through @xmath372 we deduce that @xmath256 has a simple eigenvalue at @xmath373 , with associated eigenvector @xmath374 , and all remaining eigenvalues lie in the interval @xmath375 $ ] for some small constant @xmath184 . therefore the resolvent on the right - hand side of is near - singular ( hence yielding a large contribution ) for @xmath376 .
this implies that the leading behaviour of is governed by small values of @xmath365 in fourier space .
we now outline the computation of in more detail .
let us first focus on the regime @xmath189 , i.e. the regime from theorem [ thm : theta 1 ] .
thus , the function @xmath315 may be approximated by @xmath377 times a delta function , so that the convolutions may be dropped .
what therefore remains is the calculation of the trace .
we write @xmath378 in the regime @xmath379 .
we use the fourier representation of @xmath256 and only consider the contribution of small values of @xmath365 . after some elementary computations we get , for @xmath8 , @xmath380 a similar calculation
may be performed for @xmath203 , which results in a logarithmic behaviour in @xmath231 .
this yields the right - hand sides of and .
for @xmath187 the main contribution arises from the regime @xmath381 and is therefore universal . if @xmath64 the leading contribution to arises from all values of @xmath365 .
while may still be computed for @xmath64 , it loses its universal character and depends on the whole function @xmath368 .
in the regime @xmath188 , i.e. the regime from theorem [ thm : theta 2 ] , we introduce @xmath382 ( recall ) and write ( for simplicity setting @xmath383 and @xmath8 ) @xmath384 where in the third step we used an elementary identity of fourier transforms . from this expression
it will be easy to conclude , and an analogous calculation for @xmath203 yields .
in this section we prove theorems [ thm : main result][thm : theta 1 ] by computing the limiting behaviour of @xmath385 . for simplicity , throughout this section we assume that we are in the complex hermitian case , @xmath123 .
the real symmetric case , @xmath120 , can be handled by a simple extension of the arguments of this section , and is presented in section [ sec : sym ] . due to the independence of the matrix entries ( up to the hermitian symmetry ) ,
the expectation of a product of matrix entries in can be computed simply by counting how many times a matrix entry ( or its conjugate ) appears .
we therefore group these factors according to the equivalence relation @xmath386 if @xmath387 as ( unordered ) sets . since @xmath388 , every block of the associated partition must contain at least two elements ; otherwise the corresponding term is zero . if @xmath128 were gaussian , then by wick s theorem only partitions with blocks of size exactly two ( i.e. _ pairings _ ) would contribute .
since @xmath128 is not gaussian , we have to do deal with blocks of arbitrary size ; nevertheless , the pairings yield the main contribution . in order to streamline the presentation and focus on the main ideas of the proof , in the current paper we do not deal with certain errors resulting from partitions that contain a block of size greater than two ( i.e. that are not pairings ) , and from some exceptional coincidences among summation indices . ignoring these issues results in three simplifications , denoted by * ( s1)**(s3 )
* below , to the argument . throughout the proof we use the letter @xmath389 to denote any error term arising from these simplifications . in the companion paper @xcite
, we show that the error terms @xmath389 are indeed negligible ; see proposition [ prop : cale ] below . the proof of proposition [ prop : cale ] is presented in a separate paper , as it requires a different argument from the one in the current paper . in order clarify our main argument
, it is actually helpful to generalize the assumptions on the matrix of variances @xmath256 .
this more general setup is also used in the generalized band matrix model analysed in @xcite and outlined in section [ sec : outlook ] . instead of , we set @xmath390_l}{w}}\biggr)}\ , , \qquad m \;{\mathrel{\vcenter{\baselineskip0.65ex \lineskiplimit0pt \hbox{.}\hbox{.}}}=}\ ; \sum_{x \in { \mathbb}t } f { \biggl({\frac{x}{w}}\biggr)}\,,\ ] ] where @xmath391_l$ ] denotes the representative of @xmath392 in @xmath112 , and @xmath393 is an even , bounded , nonnegative
, piecewise is _
piecewise @xmath394 _ if there exists a finite collection of disjoint open sets @xmath395 with piecewise @xmath394 boundaries , whose closures cover @xmath2 , such that @xmath364 is @xmath394 on each @xmath396 . ] @xmath394 function , such that @xmath364 and @xmath397 are integrable .
we also assume that @xmath398 for some @xmath184 .
note that @xmath116 and @xmath45 satisfy .
we introduce the covariance matrices of @xmath256 ( see also ) and @xmath364 , defined through @xmath399 it is easy to see that @xmath400 .
we always assume that @xmath401 in the sense of quadratic forms , for some positive constants @xmath95 and @xmath94 .
note that , since holds for @xmath402 , it also holds for @xmath17 for large enough @xmath45 . in the case
we have the explicit diagonal form @xmath403 in addition , for @xmath163 we introduce the quantities @xmath404 which also depend on the fourth moments of @xmath256 and @xmath364 respectively .
( here @xmath111 denotes the euclidean norm on @xmath405 . ) as above , it is easy to see that @xmath406 . in the case
we have the explicit form @xmath407 the main result of this section is summarized in the following proposition [ prop : main ] , which establishes the leading asymptotics of @xmath408 , defined in , for small @xmath409 .
once proposition [ prop : main ] is proved , our main results , theorems [ thm : main result][thm : theta 1 ] will follow easily ( see section [ sec : conclusion ] ) .
recall the definition of @xmath410 from .
[ prop : main ] suppose that the assumptions of the first paragraph of theorem [ thm : main result ] as well as the assumptions on @xmath128 made in hold . then there is a constant @xmath176 such that , for any @xmath210 satisfying for small enough @xmath179 , we have @xmath411 where the leading contribution @xmath412 satisfies the following estimates . 1 .
suppose that holds .
then for @xmath211 we have @xmath413 where @xmath414 was defined in .
moreover , for @xmath203 we have @xmath415 2 .
suppose that holds and that @xmath163 .
if @xmath161 and @xmath162 satisfy * ( c1 ) * then @xmath416 and if @xmath161 and @xmath162 satisfy * ( c2 ) * then @xmath417 3 .
suppose that @xmath188 .
then the exponent @xmath335 from proposition [ prop : expansion with trunction ] may be chosen so that there exists an exponent @xmath199 such that for @xmath201 we have @xmath418 and for @xmath203 we have @xmath419 the rest of this section is devoted to the proof of proposition [ prop : main ] . in order to express the nonbacktracking powers of @xmath128 in terms of the entries of @xmath128 , it is convenient to index the two multiple summations arising from when plugged into using a graph .
we note that a similar graphical language was developed in @xcite , and many of basic definitions from sections [ sec_41 ] and [ sec_42 ] ( such as bridges , ladders , and skeletons ) are similar to those from @xcite .
we introduce a directed graph @xmath420 defined as the disjoint union of a directed chain @xmath421 with @xmath422 edges and a directed chain @xmath423 with @xmath424 edges . throughout the following , to simplify notation
we often omit the arguments @xmath422 and @xmath424 from the graphs @xmath425 , @xmath426 , and @xmath427 . for an edge @xmath428
, we denote by @xmath429 and @xmath430 the initial and final vertices of @xmath431 .
similarly , we denote by @xmath432 and @xmath433 the initial and final vertices of the chain @xmath434 .
we call vertices of degree two _ black _ and vertices of degree one _ white_. see figure [ fig : cal c ] for an illustration of @xmath425 and for the convention of the orientation .
we assign a label @xmath435 to each vertex @xmath436 , and write @xmath437 . for an edge @xmath428 define the associated pairs of ordered and unordered labels
@xmath438 \;{\mathrel{\vcenter{\baselineskip0.65ex \lineskiplimit0pt \hbox{.}\hbox{.}}}=}\ ; \{x_{a(e ) } , x_{b(e)}\}\,.\ ] ] using the graph @xmath439 we may now write the covariance @xmath440 } - { \mathbb{e}}{\bigl({\operatorname{tr}h^{(n_1)}}\bigr ) } { \mathbb{e}}{\bigl({\operatorname{tr}h^{(n_2)}}\bigr ) } \;=\ ; \sum_{{\boldsymbol{\mathrm{x } } } \in { { \mathbb}t}^{v({\mathcal}c ) } } i({\boldsymbol{\mathrm{x } } } ) a({\boldsymbol{\mathrm{x}}})\,,\ ] ] where we introduced @xmath441 and the indicator function @xmath442 the indicator function @xmath443 implements the fact that the final and initial vertices of each chain have the same label , while @xmath444 in addition implements the nonbacktracking condition .
when drawing @xmath425 as in figure [ fig : cal c ] , we draw vertices of @xmath425 with degree two using black dots , and vertices of @xmath425 with degree one using white dots .
the use of two different colours also reminds us that each black vertex @xmath445 gives rise to a nonbacktracking condition in @xmath444 , constraining the labels of the two neighbours of @xmath445 to be distinct . in order to compute the expectation in
, we decompose the label configurations @xmath446 according to partitions of @xmath447 .
[ def : partitions ] we denote by @xmath448 for the set of partitions of a set @xmath449 and by @xmath450 the set of pairings ( or matchings ) of @xmath449 .
( in the applications below the set @xmath449 will be either @xmath447 or @xmath451 . )
we call blocks of a pairing _
bridges_. moreover , for a label configuration @xmath452 we define the partition @xmath453 as the partition of @xmath447 generated by the equivalence relation @xmath454 if and only if @xmath455 = [ x_{e'}]$ ] .
hence we may write @xmath456 at this stage we introduce our first simplification . 1 .
we only keep the pairings @xmath457 in the summation . using simplification * ( s1 )
* , we write @xmath458 here , as explained at the beginning of this section , we use the symbol @xmath389 to denote an error term that arises from any simplification that we make .
all such error terms are in fact negligible , as recorded in proposition [ prop : cale ] below , and proved in the companion paper @xcite .
we use the symbol @xmath389 without further comment throughout the following to denote such error terms arising from any of our simplifications * ( s1)**(s3)*. fix @xmath457 . in order to analyse the term resulting from the first term of
, we write @xmath459 where we used that @xmath460 and @xmath461 are independent if @xmath455 \neq [ x_{e'}]$ ] as well as @xmath462 and @xmath463 .
note that the indicator function @xmath464 imposes precisely two things : first , if @xmath431 and @xmath465 belong to the same bridge of @xmath466 then @xmath455 = [ x_{e'}]$ ] and , second , if @xmath431 and @xmath465 belong to different bridges of @xmath466 then @xmath455 \neq [ x_{e'}]$ ] .
the second simplification that we make neglects the second restriction , hence eliminating interactions between the labels associated with different bridges . 1 .
after taking the expectation , we replace the indicator function @xmath464 with the larger indicator function @xmath467 = [ x_{e'}])}$ ] .
thus we have @xmath468 = [ x_{e ' } ] ) } { { \boldsymbol{\mathrm{1 } } } ( x_e \neq x_{e ' } ) } s_{x_e}}\biggr ) } + { \mathcal}e\,.\ ] ] a similar analysis may be used for the term resulting from the second term of to get @xmath469 = [ x_{e ' } ] ) } { { \boldsymbol{\mathrm{1 } } } ( x_e \neq x_{e ' } ) } s_{x_e}}\biggr ) } + { \mathcal}e\,,\end{gathered}\ ] ] where we used that if any bridge @xmath470 intersects both @xmath471 and @xmath472 then the left - hand side vanishes since @xmath473 .
next , we note that @xmath474 = [ x_{e ' } ] ) } { { \boldsymbol{\mathrm{1 } } } ( x_e \neq x_{e ' } ) } \;=\ ; { { \boldsymbol{\mathrm{1 } } } ( x_{a(e ) } = x_{b(e ' ) } ) } { { \boldsymbol{\mathrm{1 } } } ( x_{a(e ' ) } = x_{b(e ) } ) } \;{=\mathrel{\vcenter{\baselineskip0.65ex \lineskiplimit0pt \hbox{.}\hbox{.}}}}\ ; j_{\{e , e'\}}({\boldsymbol{\mathrm{x}}})\,.\ ] ] plugging and back into therefore yields @xmath475 where we introduced the _ subset of connected pairings _ of @xmath447 @xmath476 the formula provides the desired expansion in terms of pairings .
a pairing @xmath466 may be conveniently represented graphically by drawing a line ( or _ bridge _ ) joining the edges @xmath431 and @xmath465 whenever @xmath477 .
see figure [ fig : pairing ] for an example . ]
the following notations will prove helpful .
we introduce the set of all connected pairings , @xmath478 [ def : c_gamma ] with each pairing @xmath479 we associate its underlying graph @xmath480 , and regard @xmath422 and @xmath424 as functions on @xmath481 in self - explanatory notation .
we also frequently abbreviate @xmath482 , and refer to @xmath483 as the vertices of @xmath466 .
next , we observe that the indicator function @xmath484 in constrains some labels of @xmath446 to coincide .
we introduce a corresponding partition @xmath485 of the vertices of @xmath466 , whereby @xmath445 and @xmath486 are in the same block of @xmath487 if and only if @xmath488 and @xmath489 are constrained to be equal by .
equivalently , we define @xmath487 as the finest partition of @xmath483 with the following properties . 1 .
@xmath429 and @xmath490 belong to the same block of @xmath487 whenever @xmath477 .
( note that , by symmetry , @xmath491 and @xmath430 also belong to the same block . ) 2 .
@xmath492 and @xmath493 belong to the same block of @xmath487 .
@xmath494 and @xmath495 belong to the same block of @xmath487 .
graphically , the first condition means that the two vertices on either side of a bridge are constrained to have the same label
. see figure [ fig : q ] for an illustration of @xmath487 .
we emphasize that we constantly have to deal with two different partitions .
the expectation originally introduced a partition on the edges , which , after simplification * ( s1 ) * , is in fact a pairing .
this pairing , in turn , induces constraints on the labels that are assigned to vertices ; more precisely , it forces the labels of certain vertices to coincide .
together with the coincidence of the first and last labels on @xmath426 and @xmath427 , imposed by taking the trace , this defines a partition on the vertices . depending on @xmath446
it may happen that more labels coincide than required by @xmath487 ; the partition @xmath487 encodes the minimal set of constraints .
we therefore call @xmath487 the _ minimal vertex partition induced by @xmath466_. notice that , by construction , @xmath487 does not depend on @xmath446 . next
, suppose that there is a block @xmath496 that contains two vertices @xmath497 such that @xmath498 .
we conclude that the contribution of @xmath466 to the right - hand side of vanishes , since the indicator function @xmath444 vanishes by the nonbacktracking condition @xmath499 .
hence we may restrict the summation over @xmath466 in to the subset of pairings @xmath500 [ lem : q_geq_2 ] for any @xmath501 , all blocks of @xmath487 have size at least two .
if @xmath502 then , by definition of @xmath487 , the degree of @xmath445 is two and both edges incident to @xmath445 belong to the same bridge .
this implies that the two vertices adjacent to @xmath445 belong to the same block of @xmath487 , which is impossible by definition of @xmath503 . at this point
we introduce our final simplification . 1 .
after restriction the summation over @xmath466 to the set @xmath503 in , we neglect the indicator function @xmath444 .
note that the main purpose of @xmath444 was to restrict the summation over pairings @xmath466 to the set @xmath503 , which is still taken into account if one assumes * ( s3)*. the presence of @xmath444 in simply results in some additional error terms @xmath389 that are ultimately negligible .
note that @xmath444 also restricts the summation to labels satisfying @xmath504 ; this condition is still imposed in the definition of @xmath503 .
hence we get @xmath505 where we introduced a set of independent summation labels @xmath506 , indexed by the blocks of @xmath487 .
the summation in is highly oscillatory , which requires a careful resummation of graphs of different order .
we perform a local resummation procedure of the so - called _ ladder _ subdiagrams , which are subdiagrams with a pairing structure that consists only of parallel bridges .
this is the second resummation procedure mentioned in section [ sec : struc ] .
concretely , we regroup pairings @xmath466 into families that have a similar structure , differing only in the number of parallel bridges per ladder subdiagram . their common structure is represented by the simplest element of the family , the _ skeleton _ , whose ladders consist of a single bridge .
we now introduce these concepts precisely .
the skeleton of a pairing @xmath479 is generated from @xmath466 by collapsing parallel bridges . by definition , the bridges @xmath507 and @xmath508 are _ parallel _ if @xmath509 and @xmath510 . with each @xmath479
we associate a couple @xmath511 , where @xmath512 has no parallel bridges , and @xmath513 .
the pairing @xmath514 is obtained from @xmath466 by successively collapsing parallel bridges until no parallel bridges remain .
the integer @xmath515 denotes the number of parallel bridges of @xmath466 that were collapsed into the bridge @xmath285 .
conversely , for any given couple @xmath516 , where @xmath512 has no parallel bridges and @xmath517 , we define @xmath518 as the pairing obtained from @xmath514 by replacing , for each @xmath519 , the bridge @xmath285 with @xmath515 parallel bridges .
thus we have a one - to - one correspondence between pairings @xmath466 and couples @xmath516 .
the map @xmath520 corresponds to the collapsing of parallel bridges of @xmath466 , and the map @xmath521 to the `` expanding '' of bridges of @xmath514 according to the multiplicities @xmath522 . instead of burdening the reader with formal definitions of the operations @xmath520 and @xmath521 , we refer to figures [ fig : pairing ] and [ fig : skeleton ] for an illustration .
when no confusion is possible , in order to streamline notation we shall omit @xmath520 and @xmath521 and identify @xmath466 with @xmath516 .
in particular , the minimal vertex partition @xmath487 induced by @xmath518 is denoted by @xmath523 , and is not to be confused with @xmath524 , the minimal vertex partition on the skeleton @xmath514 .
[ def : ladders ] fix @xmath512 and @xmath525 .
as above , abbreviate @xmath526
. 1 . for @xmath519
we introduce the _ ladder encoded by @xmath285 _ , denoted by @xmath527 and defined as the set of bridges @xmath470 that are collapsed into the skeleton bridge @xmath285 by the operation @xmath520 .
note that @xmath528 consists of @xmath529 parallel bridges .
we say that a vertex @xmath530 _ touches _ the bridge @xmath477 if @xmath445 is incident to @xmath431 or @xmath465 .
we call a vertex @xmath445 a _ ladder vertex _ of @xmath531 if it touches two bridges of @xmath531 .
note that a ladder consisting of @xmath19 parallel bridges gives rise to @xmath532 ladder vertices .
we say that @xmath530 is a _ ladder vertex _ of @xmath466 if it is a ladder vertex of @xmath531 for some @xmath519 .
we decompose the vertices @xmath533 , where @xmath534 denotes the set of ladder vertices of @xmath466 .
see figure [ fig : ladder vertices ] for an illustration .
, drawn in grey , of the pairing @xmath466 from figure [ fig : pairing ] .
the vertices @xmath535 are drawn in black or white . in this example
, there are @xmath536 ladders .
[ fig : ladder vertices ] ] due to the nonbacktracking condition and the requirement that parallel bridges are collapsed , not every pairing can be a skeleton , and not every family of multiplicities is admissible ; however , the few exceptions are easy to describe .
the following lemma characterizes the explicit set @xmath537 of allowed skeletons @xmath514 and the set of allowed multiplicities , @xmath538 , which may arise from some graph @xmath539 .
[ lem : skeletons ] for any @xmath501 with @xmath540 we have @xmath541 , where @xmath542 moreover , defining @xmath543 for any @xmath541 , we have that @xmath544 is finite .
roughly , this lemma states two things .
first , if a skeleton bridge @xmath519 touches two adjacent vertices of @xmath514 that belong to the same block of @xmath524 , then we have @xmath545 .
second , if @xmath524 yields the label structure @xmath546 for three consecutive vertices of @xmath514 , then @xmath547 where @xmath285 and @xmath548 are the two bridges touching the innermost of these three vertices ( in such a situation @xmath549 is impossible by nonbacktracking condition implemented by @xmath503 ) .
see figure [ fig : d8 ] below for an illustration of this latter restriction .
both of these restrictions are consequences of the nonbacktracking condition implemented in the definition of @xmath503 .
for example , the skeleton @xmath550 , defined in figure [ fig : dumbbell ] below , may arise as a skeleton of some @xmath466 , so that @xmath551 . using @xmath552 , @xmath553 , and @xmath554 to denote the multiplicities of the top , bottom , and middle bridges respectively
, we have @xmath555 .
indeed , it is easy to check that the condition on the right - hand side of is satisfied if and
only if @xmath556 . let @xmath501 and @xmath540 .
clearly , @xmath514 has no parallel bridges .
moreover , if @xmath524 has a block of size one then @xmath514 must have a bridge that connects two adjacent edges .
hence @xmath466 also has a bridge that connects two adjacent edges . by definition of @xmath503 , this is impossible .
this proves the first claim . in order to prove the second claim
, we simply observe that if @xmath541 and @xmath557 for all @xmath519 , then @xmath558 .
this follows easily from the definition of @xmath503 and the fact that the two vertices located between two parallel bridges of @xmath466 always form a block of size two in @xmath487 .
lemma [ lem : skeletons ] proves that there is a one - to - one correspondence , given by the maps @xmath520 and @xmath521 , between pairings @xmath501 and couples @xmath516 with @xmath541 and @xmath559 . throughout the following ,
we often make use of this correspondence and tacitly identify @xmath466 with @xmath516 .
we now use skeletons to rewrite @xmath560 : from we get @xmath561 next , we observe that we have a splitting @xmath562 so that the indicator function in factors into an indicator function involving only labels @xmath563 and @xmath488 with @xmath564 and @xmath565 , and another indicator function involving only labels @xmath563 and @xmath488 with @xmath566 and @xmath567 .
summing over the latter ( `` ladder '' ) labels yields @xmath568 where we defined the _ value _ of the skeleton @xmath541 as @xmath569 here we recall definition [ def : c_gamma ] for the meaning of the vertex set @xmath570 .
the entry @xmath571 arises from summing out the @xmath572 independent labels associated with the ladder vertices of @xmath573 , according to @xmath574 the labels @xmath575 in are not free ; we use them for notational convenience .
they are a function @xmath576 of the independent labels @xmath577 .
the function @xmath578 is defined by the indicator function in the second parentheses on the second line of , i.e. @xmath579 where @xmath580 . in summary
, we have proved that @xmath560 can be written as a sum of contributions of skeleton graphs ( up to errors @xmath389 that will prove to be negligible ) .
the value of each skeleton is computed by assigning a positive power @xmath581 of @xmath256 to each bridge of @xmath514 , and summing up all powers @xmath581 and all labels that are compatible with @xmath514 ( in the sense that the vertices touching a bridge , on the same side of the bridge , must have identical labels ) .
we now compute the leading contribution to .
as it turns out , it arises from a family of eight skeleton pairings , which we call _ dumbbell _ skeletons .
they are defined in figure [ fig : dumbbell ] .
we denote by @xmath582 the @xmath445-th dumbbell skeleton , where @xmath583 . at this point in the argument , it is not apparent why precisely these eight skeletons yield the leading contribution .
in fact , our analysis will reveal the graph - theoretic properties that single them out as the leading skeletons ; see section [ sec : small_skeletons ] below for the details .
we now _ define _ @xmath353 as the contribution of the dumbbell skeletons : @xmath584 [ prop : leading term ] under the assumptions of proposition [ prop : main ] , the contribution of the dumbbell skeletons defined in satisfies ( i ) , and ( ii ) , and ( iii ) of proposition [ prop : main ] .
see ( * ? ? ?
* propositions 3.4 and 3.7 ) . while the proof of proposition [ prop : leading term ]
is given in the companion paper @xcite , here we explain how to obtain the ( approximate ) expression from the definition .
the main work , performed in ( * ? ? ? * sections 3.3 and 3.4 ) , is the asymptotic analysis of the right - hand side of , which was outlined in section [ sec : heuristic leading term ] .
we first focus on the most important skeleton , @xmath585 .
see figure [ fig : d8 ] for our choice of labelling the vertex labels and the multiplicities of the bridges of @xmath585 .
in particular , @xmath586 consists of four blocks , which are assigned the independent summation vertices @xmath587 .
from we get @xmath588 here we used that @xmath589 , as may be easily checked from the definition of @xmath503 .
similarly , we may compute @xmath590 for @xmath591 ; it is not hard to see that all of them arise from the expression for @xmath592 by setting @xmath552 , @xmath553 , or @xmath593 to be zero ; setting a multiplicity @xmath594 to be zero amounts to removing the corresponding bridge from the skeleton . since the skeleton has to be connected , @xmath554 and @xmath593 can not both be zero and we choose to assign @xmath554 to the bridge with nonzero multiplicity .
the eight combinations generated by @xmath595 or @xmath596 , @xmath597 or @xmath598 , @xmath599 or @xmath600 correspond precisely to the eight graphs @xmath233 ( the case @xmath601 , @xmath602 corresponds to the first four graphs , @xmath603 , while @xmath604 corresponds to @xmath605 ) .
moreover , recalling , we can perform the sum over @xmath606 : @xmath607 where we defined @xmath608 the choice of the symbol @xmath609 suggests that for most purposes @xmath609 should be thought of as @xmath373 . putting everything together
, we find @xmath610 where @xmath611 note that here we exclude the two cases where @xmath612 , since in those cases it may be easily checked that @xmath613 violates the defining condition of @xmath503 . in all other cases , this condition is satisfied .
next , we use to decouple the upper bound in the summations over @xmath552 , @xmath553 , @xmath554 , and @xmath593 . using we easily find @xmath614 similarly , using to replace @xmath615 by @xmath616 ,
we get @xmath617 this is the precise version of . for the asymptotic analysis of the right - hand side of , see ( * ? ? ?
* sections 3.3 and 3.4 ) .
we now focus on the essence of the proof of proposition [ prop : main ] : the estimate of the non - dumbbell skeletons .
we have to estimate the contribution to the right - hand side of of all skeletons @xmath514 in the set @xmath618 it turns out that when estimating @xmath619 we are faced with two independent difficulties .
first , strong oscillations in the @xmath522-summations in the definition of @xmath619 give rise to cancellations which have to be exploited carefully .
second , due to the combinatorial complexity of the skeletons , the size of @xmath620 grows exponentially with @xmath116 , which means that we have to deal with combinatorial estimates .
it turns out that these two difficulties may be effectively decoupled : if @xmath621 is small then only the first difficulty matters , and if @xmath621 is large then only the second one matters .
the sets of small and large skeletons are defined as @xmath622 where @xmath623 is a cutoff , independent of @xmath115 , to be fixed later . in this subsection
, we deal with large @xmath621 , i.e. we estimate @xmath624 .
the only input on @xmath625 that the argument of this subsection requires is the estimate . in particular , in this subsection we deal with both cases * ( c1 ) * and * ( c2 ) * simultaneously .
[ prop : large_sigma ] for large enough @xmath626 , depending on @xmath335 , we have @xmath627 recall that , according to proposition [ prop : main ] , the value of the main terms ( the dumbbell skeletons ) is larger than @xmath628 .
the rest of this subsection is devoted to the proof of proposition [ prop : large_sigma ] .
we begin by introducing the following construction , which we shall make use of throughout the remainder of the paper .
see figure [ fig : z_sigma ] for an illustration .
[ def : z_sigma ] let @xmath541 be a skeleton pairing .
we define a graph @xmath629 on the vertex set @xmath630 as follows .
each bridge @xmath631 gives rise to the edge @xmath632 of @xmath629 , where @xmath365 and @xmath633 are defined as the blocks of @xmath524 that contain @xmath429 and @xmath430 respectively .
( note that , by definition of @xmath634 , we also have @xmath635 and @xmath636 ) .
we call @xmath629 the graph associated with @xmath514
. recall definition [ def : c_gamma ] for the meaning of @xmath637 .
then @xmath629 is simply obtained as a minor of @xmath637 after contracting ( identifying ) vertices that belong to the same blocks of @xmath524 and replacing every pair of edges of @xmath637 forming a bridge with a single edge . in particular , the skeleton bridges of @xmath514 become the edges of @xmath629 , i.e. @xmath514 and @xmath638 may be canonically identified .
similarly , @xmath524 is canonically identified with @xmath639 , the vertex set of the associated graph .
[ lem : z_sigma_connected ] for any @xmath541 the associated graph @xmath629 is connected .
this follows immediately from the definition of @xmath629 and the fact that @xmath512 .
next , let @xmath541 be fixed .
starting from the definition , we use to get @xmath640 for future reference we note that the right - hand side of may also be written without the partition @xmath524 as @xmath641 where @xmath642 was defined in and @xmath643 in .
recall that the free variables in are @xmath506 . using @xmath629 , we may rewrite in the form @xmath644 where we recall the convention @xmath645 .
let @xmath646 it is easy to see that @xmath647 .
next , we state the fundamental counting rule behind our estimates ; its analogue in @xcite was called the 2/3-rule .
it says that each block of @xmath524 contains at least three vertices , with the possible exception of blocks consisting exclusively of white vertices .
[ lem:2/3_rule ] let @xmath541 . for all
@xmath648 we have @xmath649 .
moreover , @xmath650 by definition of @xmath537 , we have that @xmath651 .
now suppose that @xmath652 .
let @xmath653 be a black vertex of @xmath570 .
since @xmath652 , we conclude that the two bridges of @xmath514 touching @xmath445 ( see definition [ def : ladders ] ( ii ) ) are parallel .
this is in contradiction with the definition of @xmath537 .
finally , follows directly from @xmath649 , since @xmath654 since @xmath655 , we get from that @xmath656 next , using lemma [ lem : z_sigma_connected ] we choose some ( immaterial ) spanning tree @xmath657 of @xmath629 . clearly , @xmath658 and @xmath659 , so that yields @xmath660 we now sum over @xmath506 in , using the estimates , valid for any @xmath661 , @xmath662 which are easy consequences of @xmath663 and @xmath664 . in the product on the right - hand side of
, we estimate each factor associated with @xmath665 by @xmath666 , using the second estimate of .
we then sum out all of the @xmath506-labels , starting from the leaves of @xmath657 ( after some immaterial choice of root ) , at each summation using the first estimate of .
this yields @xmath667 where in the last step we used .
the factor @xmath115 results from the summation over the label associated with the root of @xmath657 .
thus we find from @xmath668 - 1}{{\lvert \sigma \rvert } - 1 } \;{\leqslant}\ ;
n { \biggl({\frac{c}{m}}\biggr)}^{{\lvert \sigma \rvert}/3 - 5/3 } \frac{m^{\mu { \lvert \sigma \rvert}}}{({\lvert \sigma \rvert } - 1 ) ! } \,.\end{gathered}\ ] ] next , for any @xmath669 , a simple combinatorial argument shows that the number of skeleton pairings @xmath541 satisfying @xmath670 is bounded by @xmath671 here the factor @xmath672 is the number of pairings of @xmath673 edges , and the factor @xmath674 is the number of graphs @xmath425 with @xmath673 edges .
we therefore conclude that @xmath675 choosing @xmath626 large enough completes the proof of proposition [ prop : large_sigma ] .
we conclude this subsection by summarizing the origin of the restriction @xmath339 ( and hence @xmath676 ) , as it appears in the preceding proof of proposition [ prop : large_sigma ] .
the total contribution of a skeleton is determined by a competition between its _ size _ ( given by the number of bridges ) and its _ entropy factor _ ( given by the number of independent summation labels @xmath506 ) .
each bridge yields , after resummation , a factor @xmath677 , so that the size of the graph is @xmath678 where @xmath679 is the number of ladders .
the entropy factor is @xmath680 where @xmath681 is the number independent summation labels .
the 2/3-rule from lemma [ lem:2/3_rule ] states roughly that @xmath682 .
the sum of the contributions of all skeletons is convergent if @xmath683 , which , by the 2/3-rule , holds provided that @xmath684 .
we now focus on the estimate of the small skeletons , i.e. we estimate @xmath619 for @xmath685 ( recall the splitting ) .
the details of the following estimates will be somewhat different for the two cases * ( c1 ) * and * ( c2 ) * ; for definiteness , we focus on the ( harder ) case * ( c2 ) * , i.e. we assume that @xmath161 and @xmath162 both satisfy . the analogue of the following result in the case * ( c1 ) * is given in proposition [ prop : small_sigma_c1 ] at the end of this subsection .
[ prop : small_sigma ] suppose that @xmath161 and @xmath162 satisfy .
suppose moreover that holds for some small enough @xmath179 .
then for any fixed @xmath623 and small enough @xmath340 in proposition [ prop : expansion with trunction ] there exists a constant @xmath176 such that @xmath686 where we recall the definition of @xmath410 from .
note that , by proposition [ prop : leading term ] , the size of the dumbbell skeletons is @xmath687 unless @xmath163 and @xmath189 , in which case we have @xmath688 we conclude that the right - hand side of is much smaller than the contribution of the dumbbell skeletons . in particular , the proof of proposition [ prop : small_sigma ] reveals why precisely the dumbbell skeletons provide the leading contributions . in this section
we give a sketch of the proof of proposition [ prop : small_sigma ] , followed by the actual proof in the next section .
as explained at the beginning of section [ sec : large_sigma ] , the combinatorics of the summation over @xmath514 are now trivial , since the cardinality of the set @xmath689 depends only on @xmath626 , which is fixed . however , the brutal estimate of , which neglects the oscillations present in the coefficients @xmath616 , is not good enough . for small skeletons , it is essential to exploit these oscillations .
first we undo the truncation in the definition of @xmath690 and use to replace @xmath690 with @xmath691 in the definition of @xmath619 .
then we rewrite the real parts in using this gives rise to two terms , and we focus on the first one , which we call @xmath692 .
( the other one may be estimate in exactly the same way and is in fact smaller . )
the summation over @xmath522 in can now be performed explicitly using geometric series .
the result is that each skeleton bridge @xmath519 encodes an entry of the quantity @xmath693 , which is roughly a resolvent of @xmath256 multiplied by a phase @xmath102 , i.e. @xmath694 .
it turns out that these phases @xmath102 depend strongly on the type of bridge they belong to .
we split the set of skeleton bridges @xmath695 into the `` domestic bridges '' which join edges within the same component of @xmath425 and `` connecting bridges '' which join edges in different components of @xmath425 ; see definition [ def_split_sigma ] below for more details .
the critical regime is when @xmath696 , which yields a singular resolvent @xmath697 ( see the discussion on the spectrum of @xmath256 in section [ sec : heuristic leading term ] ) .
the phase @xmath102 associated with a domestic bridge is separated away from @xmath373 , which yields a regular resolvent .
( this may also be interpreted as strong oscillations in the geometric series of the resolvent expansion . )
the phase @xmath102 associated with a connecting bridge is close to @xmath373 and the associated resolvent is therefore much more singular . more precisely ( see lemma [ lem : bounds on s ] below ) , we find that these resolvents @xmath693 satisfy the bounds @xmath698 for domestic bridges @xmath699 and @xmath700 for connecting bridges @xmath701 .
( recall the definition of @xmath410 from . ) using the bounds and we get a simple bound on @xmath692 .
the rest of the argument is purely combinatorics and power counting : we have to make sure that for any @xmath685 this bound is small enough , i.e. @xmath702 . without loss of generality
we may assume that @xmath514 does not contain a bridge that touches ( see definition [ def : ladders ] ) the two white vertices of the same component of @xmath425 . indeed , if @xmath514 contains such a bridge , we can sum up the ( coinciding ) labels of the two white vertices using the second bound of , which effectively removes such a bridge , as depicted in figure [ fig : ssbar ] below . in particular , we have @xmath703 .
( recall the definitions of @xmath524 after and of @xmath704 from ) .
we perform the summation over the labels @xmath446 as in section [ sec : large_sigma ] : by choosing a spanning tree on the graph @xmath629 .
recall that there is a canonical bijection between the edges of @xmath629 and the bridges of @xmath514 .
denote by @xmath705 the bridges associated with the spanning tree of @xmath629 .
the combinatorics rely on the following quantities : @xmath706 note that the total number of bridges is @xmath707 .
moreover , @xmath708 since @xmath629 is connected and @xmath709 since @xmath710 is part of a spanning tree . from the 2/3-rule in
we conclude that @xmath649 for all @xmath711 and @xmath712 using the bounds and , we sum over the labels @xmath446 associated with the vertices of @xmath514 , and find the estimate @xmath713 indeed , the root of the spanning tree gives rise to a factor @xmath115 ; each one of the @xmath714 bridges not associated with the spanning tree gives rise to a factor @xmath715 ; each one of the @xmath716 connecting loop bridges gives rise to an additional factor @xmath410 ; and each one of the @xmath717 connecting tree bridges gives rise to a factor @xmath718 .
it is instructive to compare the upper bound for @xmath514 being a dumbbell to the true size of the dumbbell skeletons from .
since we exclude pairings with bridges touching the two white vertices of the same component of @xmath425 , we may take @xmath514 to be @xmath719 or @xmath720 ( see figure [ fig : dumbbell ] ) . of these two
, @xmath720 saturates the 2/3-rule and is of leading order .
for @xmath721 we have @xmath722 , @xmath723 , @xmath724 , @xmath725 .
hence the bound reads @xmath726 this is in general much larger than the true size ; they become comparable for @xmath727 ( i.e. on very small scales ) , which is ruled out by our assumptions on @xmath231 and @xmath6 .
now we explain how the estimate on @xmath692 can be improved if @xmath514 _ is not a dumbbell skeleton_. we rely on two simple but fundamental observations . first , if @xmath514 does not saturate the 2/3-rule then the right - hand side of contains an extra power of @xmath728 as compared to the leading term .
second , if @xmath514 saturates the 2/3-rule and is not a dumbbell skeleton then @xmath514 must contain a domestic bridge ( joining edges within the same component of @xmath425 ) .
having a domestic bridge implies that @xmath729 instead of the trivial bound @xmath730 .
this implies that the power of one of the large factors @xmath410 or @xmath718 on the right - hand side of will be reduced by one ; as it turns out , this is sufficient to make the right - hand side of subleading .
note that the absence of such domestic bridges in @xmath514 is the key feature that singles out the dumbbells among all skeletons that saturate the @xmath731-rule .
this explains why the leading contribution in comes from the dumbbell skeletons .
we now explain these two scenarios more precisely . for the rest of this subsection
we suppose that @xmath514 is not a dumbbell skeleton . hence @xmath732 the first two estimates are immediate , and the last one follows from the first combined with and the fact that @xmath733 .
suppose first that @xmath514 saturates the 2/3-rule .
then @xmath734 for all @xmath735 , and it is not too hard to see that @xmath514 must contain a domestic bridge , i.e. @xmath736 .
roughly , this follows from the observation that in order to get a block of size three , the bridges touching the vertices of this block must be as in figure [ fig : three - lump ] below . plugging into yields @xmath737 where the second step follows from @xmath738 and the third step from @xmath729 ( since @xmath736 ) .
we conclude that @xmath739 where we used and @xmath740 .
next , consider the case where @xmath514 does not saturate the 2/3-rule . in this case
it may well be that @xmath741 .
however , if is not saturated , then there must exist a @xmath735 satisfying @xmath742 .
thus improves to @xmath743 thus we find that @xmath744 note that we have @xmath745 and @xmath746 . using @xmath738 we therefore get @xmath747 from we therefore get @xmath748
this concludes the sketch of the proof of proposition [ prop : small_sigma ] .
we begin the proof by rewriting in a form where the oscillations in the summation over @xmath522 may be effectively exploited .
this consists of three steps , each of which results in negligible errors of order @xmath749 for any @xmath140 . in the first step ,
we decouple the @xmath522-summations by replacing the indicator function @xmath750 with the product @xmath751 , using the estimate . in the second step ,
we replace the factors @xmath752 with @xmath753 , using the estimate .
these two steps are analogous to the steps from to . in the third step , we truncate in the tails of the functions @xmath314 on the scale @xmath754 , where @xmath340 is the constant from proposition [ prop : expansion with trunction ] . to that end , we choose a smooth , nonnegative , symmetric function @xmath755 satisfying @xmath756 for @xmath757 and @xmath758 for @xmath759 .
we split @xmath760 , where @xmath761 this yields the splitting @xmath762 of the rescaled test function @xmath763 .
this splitting is done on the scale @xmath764 , and we have @xmath765\,.\ ] ] moreover , recalling and using the trivial bound @xmath766 we find @xmath767 for any @xmath140 .
the truncation of the third step is the replacement of @xmath768 with @xmath769 , using .
applying these three steps to the definition yields @xmath770 the errors arising from each of the three steps are estimated using , , and respectively .
the summations over @xmath522 and @xmath506 in the error terms are performed brutally , exactly as in the proof of proposition [ prop : large_sigma ] ( in fact here we only need that @xmath629 be connected ) ; we omit the details .
next , we use to write @xmath771 , where @xmath772 and @xmath773 is defined similarly but without the complex conjugation on @xmath774 .
we shall gives the details of the estimate for the larger error term , @xmath692 .
the term @xmath773 may be estimated using an almost identical argument ; we sketch the minor differences below . in order to estimate the right - hand side of , we shall have to classify the bridges of @xmath514 into three classes according to the following definition .
[ def_split_sigma ] for @xmath289 we define @xmath775 the set of bridges consisting only of edges of @xmath434 .
we abbreviate @xmath776 ( the set of `` domestic bridges '' ) .
we also define @xmath777 , the set of bridges connecting the two components of @xmath425 .
moreover , for @xmath778 we introduce the set @xmath779 defined as the subset of @xmath638 encoded by @xmath780 under the canonical identification @xmath781 , according to definition [ def : z_sigma ] .
since each @xmath701 contains one edge of @xmath426 and one edge of @xmath427 , and each @xmath782 contains two edges of @xmath434 , we find that the number of edges in the @xmath445-th chain @xmath783 of the graph @xmath784 with pairing @xmath785 is @xmath786 here we identify @xmath466 with @xmath516 , as remarked after lemma [ lem : skeletons ] .
we may now plug into the explicit expression for @xmath357 from , at which point it is convenient to introduce the abbreviations @xmath787 thus we get from @xmath788 here , by a slight abuse of notation , we write @xmath361 . using the associated graph @xmath629 from definition [ def : z_sigma ]
, we may rewrite this as @xmath789\}^{e({\mathcal}y(\sigma ) ) } } { { \boldsymbol{\mathrm{1 } } } ( { \boldsymbol{\mathrm{b } } } \in b(\sigma ) ) } \\
\times { \biggl({t(e_1 ) { \overline{t(e_2 ) } \!\ , } \ , { \mathrm{e}}^{{\mathrm{i}}(a_1 - a_2 ) } \ ,
\prod_{e \in e_1({\mathcal}y(\sigma ) ) } ( -{\mathrm{e}}^{2 { \mathrm{i}}a_1})^{b_e } \prod_{e \in e_2({\mathcal}y(\sigma ) ) } ( -{\mathrm{e}}^{-2 { \mathrm{i}}a_2})^{b_e } \prod_{e \in e_c({\mathcal}y(\sigma ) ) } { \mathrm{e}}^{{\mathrm{i}}(a_1 - a_2 ) b_e}}\biggr ) } \\ * \psi_1^{{\leqslant } , \eta}(e_1 ) * \psi_2^{{\leqslant } , \eta}(e_2 ) \ , \sum_{{\boldsymbol{\mathrm{y } } } \in { \mathbb}t^{v({\mathcal}y(\sigma ) ) } } { \biggl({\prod_{e \in e({\mathcal}y(\sigma ) ) } ( s^{b_e})_{y_e}}\biggr)}\,,\end{gathered}\ ] ] where we used the canonical identification between @xmath514 and @xmath638 to rewrite the set @xmath790 from lemma [ lem : skeletons ] as a subset @xmath791 ( by a slight abuse of notation ) . also , to avoid confusion , we emphasize that the expressions @xmath792 and @xmath793 have nothing to do with each other .
next , we split @xmath794 using the splitting @xmath795 in .
we first focus on main term , @xmath796 . in the definition of @xmath797
, we may sum the geometric series associated with each summation variable @xmath581 to get @xmath798 where we abbreviated @xmath799 } x^b \;=\ ; \frac{x ( 1 - x^{[m^\mu]})}{1 - x}\,\ ] ] for any quantity @xmath131 , which may be a number or a matrix .
the explicit summation over @xmath522 exploits the cancellations associated with the highly oscillating summands . from now on , we shall freely estimate the summation over @xmath506 by taking the absolute value inside the sum . on the right - hand side of , each edge @xmath800
encodes a symmetric matrix of the form @xmath801 , where @xmath802 . in order to estimate the right - hand side of
, we therefore need appropriate resolvent bounds on the entries of @xmath801 . to that end , we improve the second bound of using the following local decay bound . recall the definition of @xmath609 from .
[ lem : lct ] for all @xmath803 we have @xmath804 for some constant @xmath94 depending only on @xmath364 .
this follows from a standard local central limit theorem ; see for instance the proof in ( * ? ? ?
* section 3 ) . in particular , for @xmath805 we have @xmath806 recalling and
, we find from that for @xmath807 we have @xmath808 the bound is sharp if @xmath809 , i.e. if the sum in is not oscillating . if oscillations are present , we get better bounds which we record in the following lemma .
it is a special case of ( * ? ? ?
* proposition 3.5 ) .
[ lem : bounds on s ] let @xmath256 be as in and @xmath810 satisfy @xmath807 and @xmath811 . there exists a constant @xmath812 , depending only on @xmath7 and the profile function @xmath364 , such that @xmath813 under the same assumptions we have @xmath814 where the constant @xmath94 depends only on @xmath7 and @xmath364 . from , , and lemma [ lem : bounds on s ]
we get @xmath815 we apply to estimating via the following key estimate .
let @xmath816 and denote by @xmath817 the value of @xmath818 in the convolution integral . for small enough @xmath340 and @xmath819 ( i.e. @xmath820 and @xmath821 in the support of the convolution integral ) we have @xmath822 and @xmath823 to prove , we set @xmath824 , in which case an elementary estimate yields @xmath825 .
similarly , we have @xmath826 , which yields @xmath827 . now follows from and .
to prove , we set @xmath828 . in order to estimate @xmath829 , we distinguish two cases according to whether @xmath830 .
suppose first that @xmath830 .
then we have @xmath831 .
we therefore find from the first inequality of that @xmath832 where in the second step we used that @xmath833 and that @xmath410 is monotone decreasing for small enough arguments ( see its definition in ) . on the other hand ,
if @xmath834 then we get from that @xmath835 where in the second step we used that @xmath836 , and in the last step that @xmath837 . putting both cases together we get .
next , we plug the estimates and into and sum over @xmath506 ; we use for @xmath838 with @xmath289 , and for @xmath839 .
we perform the summation over @xmath506 as in section [ sec : large_sigma ] : by choosing an arbitrary spanning tree @xmath657 of @xmath629 along with an arbitrary root of @xmath657 . in the summation over @xmath506 on the right - hand side of ,
each edge @xmath800 encodes a matrix entry that we estimate as follows .
for @xmath840 we use the first estimate of , for @xmath841 the second estimate of , for @xmath842 the first estimate of , and for @xmath843 the second estimate of .
the result is @xmath844 where we used that @xmath659 and @xmath658 .
as before , the factor @xmath115 arises from the summation over the label of @xmath506 associated with the root of @xmath657 .
next , we remark that the above proof may be repeated verbatim for the other error term , @xmath845 .
this case is in fact easier : since @xmath846 is a finite set ( see lemma [ lem : skeletons ] ) , we do not have to exploit the cancellations from the summation over @xmath522 . repeating the above argument for @xmath845 , with the right - hand sides of the corresponding estimates from and replaced with @xmath847 , @xmath94 , @xmath847 , and @xmath94 respectively , we find @xmath848 in order to show that @xmath849 is small enough , we shall use a graph - theoretic argument to derive appropriate bounds on the exponents .
it relies on the following further partition of the set @xmath850 according to whether a bridge touches both endpoints ( white vertices ) of a chain .
[ def_s^1 ] we partition @xmath851 , where @xmath852 we also use @xmath853 and @xmath854 to denote the corresponding disjoint subsets of @xmath855 .
note that @xmath856 may contain at most two bridges : one only touching the white vertices of @xmath426 and one only touching the white vertices of @xmath427 .
see figure [ fig : bridges ] for an illustration of these three types of bridges .
( left ) , @xmath856 ( centre ) , and @xmath857 ( right ) . [
fig : bridges ] ] for the following counting arguments , for definiteness it will be convenient to assume that @xmath858 .
hence , we first show that skeleton pairings with @xmath859 can be easily estimated by those with @xmath858 , at the expense of an unimportant factor .
the following lemma states this fact precisely .
let @xmath860 [ lem : two_e_edges ] for each @xmath685 there exists a @xmath861 such that and @xmath862 .
the operation @xmath863 amounts to simply removing all bridges of @xmath856 from @xmath514 . instead of a formal definition
, we refer to figure [ fig : ssbar ] for a graphical depiction of this operation . by definition of @xmath864
, we find that the operation @xmath863 amounts to removing any of the two vertices @xmath865 and @xmath866 that belongs to @xmath524 .
this results in a removal of the corresponding number of leaves from the spanning tree @xmath657 .
( the removed bridges always correspond to leaves in @xmath657 .
in particular , @xmath867 and @xmath868 are remain unchanged by this removal . )
note that if @xmath869 then @xmath861 , since by construction if @xmath870 then @xmath871 . the claim now follows easily from the bound with argument @xmath514 , as well as the observations that @xmath872 , that @xmath873 , that @xmath874 , and that the two last exponents on the right - hand side of are the same for @xmath514 and @xmath875 . by lemma [ lem : two_e_edges ]
, it suffices to estimate @xmath849 for @xmath876 .
for @xmath876 we have @xmath877 .
moreover , if there is a bridge touching @xmath492 and @xmath494 as well as a bridge touching @xmath493 and @xmath495 , we find that all four white vertices constitute a single block of @xmath524 .
otherwise , since @xmath878 , every block of @xmath524 contains a black vertex , so that @xmath703 , where @xmath704 was defined in . either way ,
recalling lemma [ lem:2/3_rule ] , we conclude for @xmath876 and @xmath735 that @xmath879 in order to complete the estimate of , and hence the proof of proposition [ prop : small_sigma ] , we shall have to distinguish between the case where @xmath734 for all @xmath735 and the case where there exists a @xmath735 with @xmath880 .
[ lem : saturated2/3 ] suppose that @xmath876 and @xmath734 for all @xmath735 .
then @xmath881 [ lem : non_saturated2/3 ] suppose that @xmath876 and there exists a @xmath882 with @xmath883 .
then @xmath884 we first claim that there is at least one domestic bridge , i.e. that @xmath885 .
clearly , @xmath886 is even .
recall that @xmath887 . since each block of @xmath524 has size @xmath888 ,
we conclude that @xmath886 is multiple of @xmath888 , and hence of @xmath889 .
a simple exhaustion of all possible pairings @xmath876 that saturate the first inequality in shows that there is no such @xmath514 satisfying @xmath890 .
( in fact , any connected pairing with at most six vertices is a dumbbell pairing , which are excluded by the definition of @xmath620 . ) hence we find that @xmath891 , so that @xmath892 . next ,
note that @xmath524 contains at most two blocks @xmath365 that contain white vertices of @xmath514 , since @xmath514 contains four white vertices , and , for each @xmath893 , those of @xmath434 are in the same block of @xmath524 .
( recall items ( ii ) and ( iii ) after ) . since @xmath892
, we find that there is a block @xmath735 that contains only black vertices .
let @xmath894 be the set of bridges of @xmath514 touching a vertex of @xmath365 ; see figure [ fig : three - lump ] . by definition of @xmath524 , we have @xmath895 .
now if all vertices of @xmath365 belong to the same connected component of @xmath425 , then @xmath896 . otherwise , let @xmath897 with @xmath486 and @xmath268 belonging to the same connected component of @xmath425 .
then both bridges touching @xmath445 are in @xmath857 ; the remaining bridge of @xmath894 must touch both @xmath486 and @xmath268 , and is therefore in @xmath850 .
either way , we find that @xmath885 , as claimed above . for the following , abbreviate @xmath898 and @xmath899 .
from the saturated inequality we find @xmath900 plugging this into yields @xmath901 recall that @xmath902 and @xmath903 .
moreover , @xmath904 . since @xmath905 , we conclude @xmath906 thus we get @xmath907 where in the last step we used to get @xmath908 , as well as @xmath892 and @xmath909 . here
we chose @xmath340 in proposition [ prop : expansion with trunction ] small enough that @xmath910 .
we also used that @xmath911 .
this concludes the proof . since @xmath912 and all other blocks of @xmath524 have size at least 3 by , we find that @xmath913 where @xmath914 is the number of vertices of @xmath514 not in @xmath915 .
note the improvement of over . using the notation of the proof of lemma [ lem : saturated2/3 ]
, we get from , in analogy to , @xmath916 now we proceed as in the proof of lemma [ lem : saturated2/3 ] , using @xmath917 , @xmath903 , and @xmath918 .
we get @xmath919 in the last step we used that @xmath920 , which follows from and from @xmath921 for @xmath922 , and that @xmath923 .
( in fact , since @xmath870 one may easily check that @xmath924 . )
this concludes the proof . from lemmas [ lem : saturated2/3 ] , [ lem : non_saturated2/3 ] , and [ lem : two_e_edges ]
, we conclude that for all @xmath685 we have @xmath925 in order to conclude the proof of proposition [ prop : small_sigma ] , we need an analogous estimate of @xmath773 .
this may be obtained by repeating the above argument almost verbatim ; the only nontrivial difference is that , on the right - hand side of , the factor @xmath926 associated with the edge @xmath839 is replaced with @xmath927 . since @xmath928 on the support of the convolution integral , we replace with @xmath929 thus we find , for any @xmath685 , that @xmath930 hence proposition [ prop : small_sigma ] follows from , , and the observation that @xmath931 is a finite set that is independent of @xmath115 .
this concludes the proof of proposition [ prop : small_sigma ] .
we conclude this subsection with an analogue of proposition [ prop : small_sigma ] in the case * ( c1)*. its proof follows along the same lines as that of proposition [ prop : small_sigma ] , and is omitted .
[ prop : small_sigma_c1 ] suppose that @xmath161 and @xmath162 satisfy .
suppose moreover that holds for some small enough @xmath179 .
then for any fixed @xmath623 we have . for future reference
we emphasize that the only information about the matrix entries of @xmath932 that is required for the estimate to hold is and .
thus , the conclusion of the above argument may be formulated in the following more general form .
let @xmath933 , and suppose that we have a family of matrices @xmath934 parametrized by @xmath519 satisfying @xmath935 for @xmath936 and @xmath937 for @xmath701 .
then for small enough @xmath9 there exists a @xmath176 such that @xmath938 we may now conclude the proof of proposition [ prop : main ] .
as indicated before , the error terms @xmath389 resulting from the simplifications * ( s1)**(s3 ) * are small ; the precise statement is the following proposition that is proved in @xcite .
[ prop : cale ] the error term @xmath389 in arising from the simplifications * ( s1)**(s3 ) * satisfies @xmath939 for some constant @xmath176 .
this is an immediate consequence of propositions 4.5 , 4.6 , and 4.15 in @xcite . combining propositions [ prop : large_sigma ] , [ prop : small_sigma ] , [ prop : small_sigma_c1 ] , and [ prop : cale ] yields @xmath940 together with proposition [ prop : leading term ] ,
this concludes the proof of proposition [ prop : main ] . using , , and proposition [ prop : main ] we therefore get , for @xmath128 as in section [ sec : setup ] , @xmath941 in order to compute the left - hand side of , and hence conclude the proof of theorems [ thm : main result][thm : theta 1 ] , we need to control the denominator of using the following result .
[ lem : ey ] for @xmath942 $ ] we have @xmath943 in the case * ( c1 ) * we have @xmath944 for any @xmath184 . here in the last step we introduced the stieltjes transform of the semicircle law , @xmath945 , and invoked ( * ? ? ?
* theorem 2.3 and equation ( 7.6 ) ) .
the claim then follows from the estimate @xmath946 , which itself follows from ( * ? ? ?
* equations ( 3.3 ) and ( 3.5 ) ) . in the case * ( c2 ) * , we first split @xmath947 as in .
the contribution of @xmath948 is small by the strong decay of @xmath71 .
the error in the main term , @xmath949 may be estimated using ( * ? ? ?
* theorem 2.3 ) and helffer - sjstrand functional calculus , as in e.g. ( * ? ? ?
* section 7.1 ) ; we omit the details .
then the claim follows from @xmath950 .
now we define @xmath951 then theorems [ thm : theta 2 ] and [ thm : theta 1 ] follow from lemma [ lem : ey ] and , recalling and
. moreover , theorem [ thm : main result ] follows from and lemma [ lem : ey ] .
this concludes the proof of theorems [ thm : main result][thm : theta 1 ] under the simplifications * ( s1)**(s3)*.
in this section we explain the changes needed to the arguments of section [ sec : main argument ] to prove theorems [ thm : main result ] , [ thm : theta 2 ] , and [ thm : theta 1 ] for @xmath120 instead of @xmath123 .
the difference is that for @xmath120 we have @xmath952 , while for @xmath123 we have @xmath462 ( in addition to @xmath953 , which is valid in both cases ) .
this leads to additional terms for @xmath120 , which may be conveniently tracked in our graphical notation by introducing _
twisted bridges _ , in analogy to section 9 of @xcite .
as it turns out , allowing twisted bridges results in eight new dumbbell skeletons , called @xmath954 below , each of which has the same value @xmath955 as its counterpart without a tilde .
hence , for @xmath120 the leading term is simply twice the leading term of @xmath123 , which accounts for the trivial prefactor @xmath956 in the final formulas .
any other skeleton may be estimated by a trivial modification of the argument from sections [ sec : large_sigma][proof : small_sigma ] . as in section
[ sec : main argument ] , we make the simplifications * ( s1)**(s3 ) * , and do not deal with the errors terms @xmath389 resulting from them .
they are handled in @xcite .
we now give a more precise account of the proof for @xmath120 .
we start from , which remains unchanged .
since @xmath957 , holds for @xmath120 without the indicator function @xmath958 that was present for @xmath123 .
hence also holds without the indicator function @xmath958 .
we now write @xmath959 = [ x_{e ' } ] ) } \;=\ ; { { \boldsymbol{\mathrm{1 } } } ( [ x_e ] = [ x_{e ' } ] ) } { { \boldsymbol{\mathrm{1 } } } ( x_e \neq x_{e ' } ) } + { { \boldsymbol{\mathrm{1 } } } ( x_e = x_{e ' } ) } \;{=\mathrel{\vcenter{\baselineskip0.65ex \lineskiplimit0pt \hbox{.}\hbox{.}}}}\ ; j_{\{e , e'\}}({\boldsymbol{\mathrm{x } } } ) + { \widetilde}j_{\{e , e'\}}({\boldsymbol{\mathrm{x}}})\,,\ ] ] in self - explanatory notation ( recall that @xmath960 was already defined in ) . thus becomes @xmath961 multiplying out the parentheses in yields @xmath962 terms ,
each of which is characterized by the set of bridges of @xmath466 associated with a factor @xmath963 ; the other bridges are associated with a factor @xmath964 .
we call the former _ straight bridges _ and the latter _ twisted bridges_. this terminology originates from the fact that a twisted bridge forces the labels of the adjacent vertices to coincide on opposite sides of the bridge ; see figure [ fig : twisted bridges ] for an illustration .
next , we extend the definition of skeletons from section [ sec_42 ] to pairings containing twisted bridges . recall that the key observation behind the definition of a skeleton was that parallel straight bridges yield a large contribution but a small combinatorial complexity .
now _ antiparallel _ twisted bridges behave analogously , whereby two bridges @xmath507 and @xmath508 are _ antiparallel _ if @xmath509 and @xmath965 .
( recall that they are parallel if @xmath509 and @xmath510 . )
see figure [ fig : twisted bridges ] for an illustration .
antiladder _ is a sequence of bridges such that two consecutive bridges are antiparallel .
all of section [ sec_41 ] , in particular the partition @xmath487 , may now be taken over with trivial modifications .
as in section [ sec_42 ] , to each tagged pairing @xmath466 we assign a tagged skeleton @xmath514 with associated multiplicities @xmath522 . the skeleton @xmath514 is obtained from @xmath466 by successively collapsing _ parallel straight bridges _ and _ antiparallel twisted bridges _ until none remains .
parallel twisted bridges and antiparallel straight bridges remain unaltered .
the skeleton @xmath514 inherits the tagging of its bridges in the natural way : two parallel straight bridges are collapsed into a single straight bridge , and two antiparallel twisted bridges are collapsed into a single twisted bridge .
see figure [ fig : twisted skeleton ] for an illustration . allowing twisted bridges leads to a further eight skeleton graphs , which we denote by @xmath954 , whose contribution is of leading order .
they are the same graphs as @xmath233 from figure [ fig : dumbbell ] , except that the ( one or two ) vertical antiparallel straight bridges ( depicted by solid lines ) are replaced with the same number of vertical parallel twisted bridges ( depicted by dashed lines ) .
we use the notations @xmath966 we record the following simple result , whose proof is immediate .
what remains is the estimate of @xmath619 for @xmath969 .
we proceed exactly as in sections [ sec : large_sigma][proof : small_sigma ] .
the key observation is that the 2/3-rule from lemma [ lem:2/3_rule ] remains true thanks to the definition of skeletons .
when estimating the large skeletons ( without making use of oscillations ) in section [ sec : large_sigma ] , we get an extra factor @xmath970 to the left - hand side of arising from the sum over all possible taggings of a skeleton ; this factor is clearly immaterial .
finally , the argument of sections [ sec : small_skeletons ] and [ proof : small_sigma ] may be taken over with merely cosmetic changes .
the set @xmath856 from definition [ def_s^1 ] remains unchanged , and in particular only contains straight bridges .
note that the basic graph - theoretic argument from lemmas [ lem : saturated2/3 ] and [ lem : non_saturated2/3 ] remains unchanged .
in particular , exactly as in the proof of lemma [ lem : saturated2/3 ] , if all blocks of @xmath524 have size three and @xmath514 is not a dumbbell skeleton , then @xmath514 contains a domestic bridge ( which may be straight or twisted ) .
this concludes the proof of theorems [ thm : main result][thm : theta 1 ] for the case @xmath120 under the simplifications * ( s1)**(s3)*. r. froese , d. hasler , and w. spitzer , _ transfer matrices , hyperbolic geometry and absolutely continuous spectrum for some discrete schrdinger operators on graphs _ ,
* 230 * ( 2006 ) , 184221 . | we consider the spectral statistics of large random band matrices on mesoscopic energy scales .
we show that the correlation function of the local eigenvalue density exhibits a universal power law behaviour that differs from the wigner - dyson - mehta statistics .
this law had been predicted in the physics literature by altshuler and shklovskii @xcite ; it describes the correlations of the eigenvalue density in general metallic samples with weak disorder .
our result rigorously establishes the altshuler - shklovskii formulas for band matrices . in two dimensions , where the leading term vanishes owing to an algebraic cancellation
, we identify the first non - vanishing term and show that it differs substantially from the prediction of kravtsov and lerner @xcite . the proof is given in the current paper and its companion @xcite . |
most scientists publish their findings to disseminate their research results widely and hope that their research has some impact in the scientists community and society in general . for many years
, citation counts have been used for this goal .
this research evaluation approach has produced interesting results identifying the complex nature of physics impact in the research community ( for instance see refs .
@xcite , @xcite , @xcite ) .
this identification of diverse research impacts is important to research managers/ sponsors/ evaluators , and of course performers .
they are interested in the types of people and organizations citing the research outputs , and whether the citing audience is the target audience .
also , they are interested in whether the development categories and technical disciplines impacted by the research outputs are the desired targets .
since fundamental research can evolve along myriad paths , tracking diverse impacts becomes very complex .
recently , scientists have addressed the problem of citation in scientific research from different perspectives : looking for topological description of citation @xcite , or for power laws in citation networks @xcite , or obtaining power laws in number of cites received by journals according to their number of published papers @xcite , or though two - step competition model relating the number of publications and number of cites .
these different approaches use power laws trying to obtain simple results from complex interactions , assuming that the precise details of the interactions among the parts of the system play no role in determining the overall behavior of the system@xcite .
other approaches try to find some kind of universality in the behavior of research institutions @xcite,@xcite .
however , in order to obtain a detailed representation of the system it is important to know the details of the interactions .
the creation of roadmaps for science and technology illuminates these interactions , and allows the progression of research to be portrayed from both retrospective and prospective perspectives . and with these @xcite , @xcite .
the analysis of the cites and authorship from a social network perspective gives other useful information to characterize scientific disciplines @xcite , @xcite .
however , these approaches : scaling , networks and roadmaps , have limitations due to the fact that they explore only partially the data available from the citation system .
the detailed analysis of all the available data of the citing community is required to obtain more information and knowledge .
until now there has been no comprehensive systematic methodology to deal with the information available through cites of the scientific article . to overcome the above mentioned limitations of these techniques , we have developed a phenomenological approach to deal with all the citation information available , and obtain a more detailed description of this complex system .
the aim of this paper is to show how we can obtain a more complete profile of the citing papers , and thereby get a more complete representation of the impact of science .
the application of this kind of phenomenological detailed description is useful to obtain a different and illustrative view of the complex systems .
the enhanced coverage of the research literature by the web version of the science citation index ( sci-5300 leading research journals ) allows a broad variety of bibliometric analyses of r&d units ( papers , researchers , journals , institutions , countries , technical areas ) to be performed .
aggregation of citation number counts is characteristic of almost all published citation studies@xcite,@xcite , @xcite ; this approach identifies r&d units that have had ( and have not had ) gross impact on the user community .
however , as we have already mentioned , this abscence of fine - structure represents a limited perspective ; therefore , we require an approach that utilizes all the available citing data and could help answer questions such as : what types of people and organizations are citing the research outputs ; is this the desired target audience ?
what development categories are citing the research outputs ?
what technical disciplines are citing the research outputs ?
what are the relationships between the citing technical disciplines and the cited technical disciplines ?
the aim of the present study is to show the power and capability of this new phenomenological approach to citer profiling .
it is necessary to stress that the aim is not to assess the productivity and magnitude of impact of any individual researcher , research group , laboratory , institution , or country . to perform such an assessment , the authors would need a charter and statistically representative data based on the unit of assessment , i.e. to make a portrait of the research impact according with the current available scientific databases .
the organization of this paper is the following : in the second section , we summarize the methodology , and describe citation mining . in the third section , we present the results of analyzing four sets of papers . in the following
we will refer to these papers as cited papers , in order to distinguish them from the citing papers that are the papers in which the cited papers are referenced .
we profile the four selected paper sets through the analysis of the characteristics of the citing authors , journals where the citing papers appeared , references in the citing papers , and also through the analysis of linguistic correlations in the abstract of the cited and citing papers as well as the titles and the keywords , and other registers available in the sci database .
finally , we conclude with some remarks on the present study .
in this section we describe a simple procedure to incorporate much of the sci information in the analysis of the impact of scientific papers .
first , we identify the types of data contained in the sci ( circa early 2000 ) , and the types of analyses that will be performed on this information ( see table 1 ) .
table 1 shows a record from the sci , without the field tags . the actual paper that it represents is referred in the following description as the full paper. starting from the top , the individual fields are : [ cols="<,<",options="header " , ] the citing papers to usf paper representing different categories of development and different disciplines from those of the cited paper
are portrayed graphically in figure 12 , the axes are category , alignment and papers .
the category represents the level of development characterized by the citing paper ( 1=basic research ; 2=applied research ; 3=advanced development/ applications ) , and the alignment represents the degree of similarity between the main themes of the citing and cited papers ( 1=strong alignment ; 2=partial alignment ; 3=little alignment ) .
there are three interesting features on figure 12 .
first , the tail of total annual citation counts is very long , and shows little sign of abating , this is one characteristic feature of a seminal paper .
second , the fraction of extra - discipline basic research citing papers to total citing papers ranges from about 20 - 40% annually , with no latency period evident .
this instant extra - disciplinary diffusion may have been due to the combination of intrinsic broad - based applicability of the subject matter and publication of the paper in a high - circulation science journal with very broad - based readership .
third , there was a four - year latency period before the higher development category citing papers began to emerge .
one can see that black dots ( earlier cites ) are completely in the category .
this correlates with the results from the bibliometrics component .
the latency could have been due to the information remaining in the basic research journals , and not reaching the applications community , or the time that an application needs to be developed is of the order of four years .
thus , the basic science publication feature that may have contributed heavily to extra - discipline citations may also have limited higher development category citations for the latency period .
the present phenomenological approach of identifying impact themes through text mining allows a much more detailed and informative picture of the impact of research to be obtained compared to semi - automated journal classification comparison approaches @xcite .
it represents the difference between stating that a physics paper impacted geology research and a paper focused on sand - pile avalanches for surface smoothing impacted analyses of steep hill - slope landslides . in the final data analysis
, a taxonomy of the usf citing papers was generated using phrase clustering .
the abstracts of all the usf citing papers were converted to phrases and their frequencies of occurrence with use of a natural language processor contained in the techoasis software package .
the 153 highest frequency technical content phrases ( expert - selected ) were exported to a statistical clustering software package ( winstat ) . based on the relations among phrases generated by this package , a taxonomy was generated by the authors .
a particularly helpful output for each clustering run was the dendogram , a tree - like diagram showing the structural branches that define the clusters .
figure 13 is one dendogram based on the 48 highest frequency phrases ( for illustration purposes only ) .
the abscissa contains the phrases that are clustered .
the ordinate is a distance metric
. the smaller the distance at which phrases , or phrase groups , are clustered , the closer is the connection between the phrases .
thus , samples of the phrases combined , near the right hand end of the graph include dissipation , collisions and energy , and mixtures/ alternating layers/ small grains/ large grains/ stratification . in the middle part of the graph vibration and amplitude can be found . at some later time , the vibration - amplitude combination is grouped with the gravity - granular media combination to form the next hierarchical level grouping , and so on .
many statistical agglomeration techniques for clustering were tested ; the average neighbor method appeared to provide reasonably consistent good results .
analyses were performed of the numerous cluster options that were produced .
the following is one of the top - level cluster descriptions that represented the results of the phrase and word lists clustering best , as well as the factor matrix clustering from the techoasis results ( more clusters can be found in kostoff et al .
@xcite ) . the highest level categorization based on the highest frequency 153 phrases produced three distinct clusters @xcite : structure/ properties , flow - based experiments , modeling and simulation . in the description of the structure and properties cluster ( right part of the dendogram figure 13 ) that follows
, phrases that appeared within the clusters will be capitalized .
this cluster contained mixtures of large grains and small grains , with stratification along alternating layers based on size segregation and grain shape and geometrical profile .
the mixture forms a pile with an angle of repose . when the angle of repose is larger than a critical angle , dynamical processes produce avalanches , resulting in surface flow within thin layers .
in this paper , we have presented a phenomenological technique to analyze some aspects of a complex system like the multi - path non - monotonic impact of scientific research .
the result of using citation mining ( bibliometrics and text mining ) to analyze the impact of science , through the use of the available information from the web of science isi , allows the profiling of the citing papers of a given paper , research group , scientific organization , etc .
we illustrated citation mining through the analysis of four research groups .
this analysis provided multiple facets and perspectives of the myriad impacts of research .
citation mining offers insights that would not emerge if only separate citing paper counts were used independently , as is the prevalent use of citation analysis today .
moreover , by removing the need to actually read thousands of abstracts through the use of text mining , comprehensive assessments of research impact become feasible .
one important result from the basic research citation mining @xcite was that impacts are possible in myriad fields and applications not envisioned by the researchers .
this reference also questioned whether fundamental sand - pile research would receive funding from tokamak , air traffic control , or materials programs , even though sand - pile research could impact these or many other types of applications , as shown in the paper .
the reference concluded that sponsorship of some unfettered research must be protected , for the strategic long - term benefits on global technology and applications !
99 amaral , lan , gopikrishnan , p. , matia , k. , plerou , v. and stanley e.h .
application of statistical physics methods and concepts to the study of science & technology systems , scientometrics , 51 , 3 ( 2001 ) .
katz , j.s .
the self - similar science systems , research policy 28 , 501 - 517 ( 1999 ) ; katz , j.s . scale - independent indicators and research evaluation , electronic working paper series spru , university of sussex , ( 2000 ) http://www.sussex.ac.uk / spru/. kostoff , r.n . ,
del ro , j.a .
, humenik .
j.a . , garca , e.o . and ramrez , a.m. citation mining : integrating text mining and bibliometrics for research user profiling j. am .
52 1148 - 1156 ( 2001 ) .
newman , m.e.j .
scientific collaboration networks i , phys .
e 64 , 016131 ( 2001 ) ; newman , m.e.j . scientific collaboration networks ii , phys .
e 64 , 016132 ( 2001 ) ; newman , m.e.j . the structure of scientific collaboration networks , p. natl .
usa 98 , 404 - 409 ( 2001 ) .
plerou , v. amaral , l.a.n . ,
goplkrishnan , p. , meyer , m. and stanley h.e .
similaties between the growth dynamics of university research and of competitive economic activities , nature 400 , 433437 ( 1999 ) .
stanley , h.e . , amaral , lan , goplkrishnan , p. , ivanov p.ch .
keitt t.h . and plerou , v. scale invariance and universality : organizing principles in complex systems .
physica a 281 , 60 - 68 ( 2000 ) .
time profile of citing papers for usf .
here it is important to stress a four year delay in the appearence of applied citing papers , but no delay in the appearence of extradiscipline fundamental papers .
figure 6 . citing country development phase . in this graph
we are plotting the relative number of citing countries according to their development phase .
each axis indicates the ratio between developed or developing citing countries and the total number of citing countries to each group .
lines are drawn in order to guide the eye and group the different origins of the citing papers . in the fundamental groups ,
the difference between the number of papers produced in developed and developing countries is clear .
usf , usa and brif receive more cites from developed countries than from developing ones
. however , note that again the topic of low cost technology is more interesting for developing countries . | in this paper we present a phenomenological approach to describe a complex system : scientific research impact through citation mining .
the novel concept of citation mining , a combination of citation bibliometrics and text mining , is used for the phenomenological description .
citation mining starts with a group of core papers whose impact is to be examined , retrieves the papers that cite these core papers , and then analyzes the technical infrestructure ( authors , jorunals , institutions ) of the citing papers as well as their thematic characteristics .
the science citation index is used as the source database for the core and citing papers , since its citation - based structure enables the capability to perform citation studies easily .
this paper presents illustrative examples in photovoltaics ( applied research ) and sandpile dynamics ( basic research ) to show the types of output products possible .
bibliometric profiling is used to generate the technical infrastructure , and is performed over a number of the citing papers record fields to offer different perspectives on the citing ( user ) community .
text mining is performed on the aggregate citing papers , to identify aggregate citing community themes , and to identify extra - discipline and applications themes .
the photovoltaics applied research papers had on the order of hundreds of citations in aggregate .
all of the citing papers ranged from applied research to applications , and their main themes were fully aligned with those of the aggregate cited papers .
this seems to be the typical case with applied research .
the sandpile dynamics basic research papers had hundreds of citations in aggregate .
most of the citing papers were also basic research whose main themes were aligned with those of the cited paper .
this seems to be the typical case with basic research . however , about twenty percent of the citing papers were research or development in other disciplines , or development within the same discipline .
there was no - time lag between publication and citation by the extra - discipline research papers , but there was a four year lag time between publication and citation by the development papers . |
ideal hydrodynamics is very successful in describing experimental data at the relativistic heavy ion collider .
ideal hydrodynamics requires local thermal equilibrium .
the ideal hydrodynamics equations are valid for systems with at least local isotropy .
it is very interesting to study the equilibration process to get a more detailed understanding of when and how much thermal equilibration can be achieved in relativistic heavy ion collisions @xcite .
the isotropization process has been studied with the ampt model @xcite .
in particular , the pressure anisotropy , i.e. , the longitudinal to transverse pressure ratio , is used to measure isotropization in the central cell in central heavy ion collisions .
the string melting model is seen to give faster isotropization than the default model due to more partons in the initial stage . as the parton - parton cross section increases , isotropization increases .
the pressure anisotropy crosses one at late times when transverse expansion sets in .
the current ampt has fixed partonic cross sections and only two to two partonic processes . in order to go beyond the current ampt description ,
we introduce the medium dependent cross sections and particle number changing processes .
the @xmath4 ( @xmath5 ) cross section is taken to be the pqcd cross section regulated by a debye - screening mass .
it is inversely proportional to the screening mass squared .
the screening mass depends on the medium locally and is related to the density of particles . as the density increases , the screening mass increases , and the cross section decreases .
the @xmath6 ( @xmath7 ) cross section is taken to be one - half of the @xmath5 cross section , consistent with a more sophisticated calculation from xu and greiner .
the @xmath8 reaction integral is determined from detailed balance to ensure correct chemical equilibration .
isotropic cross sections will be used and should be reasonable for the dense central region of a collision . with this radiative transport , chemical and kinetic equilibrations can be studied .
the evolutions of collision rates are shown in fig .
[ fig : box_rate ] beginning with @xmath9 particles in a box of @xmath10 @xmath11 with a temperature of @xmath1 gev and coupling constant @xmath12 .
the initial @xmath7 rate is large than @xmath8 rate and there will be particle production .
as time goes on , these two rates converge toward the rate in chemical equilibrium .
it is interesting to see whether kinetic equilibrium is maintained throughout the chemical equilibration process .
[ fig : box_ene ] shows that the energy distribution deviates from an exponential distribution at early times . in particular , at @xmath13 fm/@xmath3 ,
low energy part is closer to the final equilibrium value .
the system relaxes toward the final thermodynamics equilibrium and at time @xmath14 fm/@xmath3 , it is very close to the expected distribution with chemical and kinetic equilibrium .
kinetic equilibration during the early stage of a heavy ion collision can be studied with the radiative transport model @xcite .
we will start with either initial isotropic or initial transverse distribution .
the initial particle distributions are all exponential .
two initial energy densities @xmath15 gev/@xmath11 will be used .
the former corresponds to an initial temperature of @xmath16 gev for the isotropic initial conditions , while the latter corresponds to @xmath1 gev .
two values are used for the strong interaction coupling constant @xmath17 .
[ fig : bj_plopt ] shows the pressure anisotropy ratio evolutions . with isotropic initial conditions , @xmath18 first decreases due to
the influence of longitudinal expansion then increases as thermalization wins over expansion .
in contrast to the case with only @xmath5 collisions , instead of the @xmath19 scaling , there is an approximate @xmath0 scaling .
furthermore , the transverse initial condition case has the same asymptotic evolution as the isotropic initial condition case . in other words , the memory of the initial pressure anisotropy
is lost after some relaxation time .
as the initial energy density increases or as the coupling constant increases , the pressure anisotropy goes closer to one and there is more thermalization and more memory loss of initial conditions .
the energy density evolutions are shown in fig .
[ fig : bj_ene ] .
this is for the high initial energy density case , similar results are obtained for the low initial energy density case .
[ fig : bj_ene ] shows that the energy density evolution at very early times is determined by the initial pressure anisotropy and the late time evolution is determined by final state interactions .
in particular , we observe that the evolutions are bounded by @xmath20 for the free streaming case and @xmath21 when @xmath22 .
this falls a little short of the ideal hydro @xmath23 behavior and shows that viscosity is important in the longitudinal expansion stage .
9 z. xu and c. greiner , phys .
c 76 ( 2007 ) 024911
. b. zhang , l. w. chen and c. m. ko , j. phys .
g 35 ( 2008 ) 065103 .
p. huovinen and d. molnar , phys .
c 79 ( 2009 ) 014906
. b. zhang , eur .
j. c 62 ( 2009 ) 25 .
b. zhang , arxiv:0903.2055 [ nucl - th ] . | a radiative transport model is used to study kinetic equilibration during the early stage of a relativistic heavy ion collision .
the parton system is found to be able to overcome expansion and move toward thermalization via parton collisions .
scaling behaviors show up in both the pressure anisotropy and the energy density evolutions .
in particular , the pressure anisotropy evolution shows an approximate @xmath0 scaling when radiative processes are included .
it approaches an asymptotic time evolution on a time scale of @xmath1 to @xmath2 fm/@xmath3 .
the energy density evolution shows an asymptotic time evolution that decreases slower than the ideal hydro evolution .
these observations indicate that partial thermalization can be achieved and viscosity is important for the evolution during the early longitudinal expansion stage of a relativistic heavy ion collision . |
deep x - ray surveys have provided the most effective method of identifying reliable and fairly complete samples of active galactic nuclei ( agns ) out to @xmath18 ( e.g. , see @xcite for a review ) .
the observed agn sky density reaches @xmath19 deg@xmath20 in the deepest x - ray surveys , the _ chandra _ deep fields ( e.g. , * ? ? ?
* ; * ? ? ?
these point sources are largely responsible for the observed cosmic x - ray background ( xrb ) .
a significant portion ( @xmath21 ) of the xrb in the kev range has been resolved into discrete sources by _ chandra _ and _ xmm - newton _ ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the majority of which are agns with moderate x - ray obscuration ( @xmath22 @xmath1 ; e.g. , @xcite ) . however , the resolved fraction of the xrb decreases toward higher energies , being @xmath23 in the 68 kev band and @xmath24 in the 812 kev band @xcite .
population - synthesis models suggest that at @xmath2530 kev , where the xrb reaches its peak flux ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , the unobscured or moderately obscured agns discovered at lower energies can not account for the xrb flux entirely , and an additional population of heavily obscured ( @xmath0 @xmath1 ) , or even compton - thick ( @xmath7 @xmath1 , hereafter ct ) , agns at mainly @xmath261.5 is required .
the number density of these heavily obscured agns is estimated to be of the same order as that of moderately obscured agns ( e.g. , * ? ? ?
however , as the x - ray emission ( below at least 10 kev ) of such sources is significantly suppressed , only a few distant heavily obscured agns have been clearly identified even in the deepest x - ray surveys ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and thus a significant fraction of the agn population probably remains undetected .
infrared ( ir ) selection is a powerful tool for detecting heavily obscured agns that can not be identified in the x - ray band .
x - ray and uv emission absorbed by the obscuring material is reprocessed and reemitted mainly at mid - ir wavelengths ( e.g. , * ? ? ? * ; * ? ? ?
the mid - ir emission is less affected by dust extinction than at optical or near - ir wavelengths , rendering it more suitable for identifying obscured agns . with the _ spitzer _ mission @xcite ,
deep mid - ir data in multiple bands can be obtained by the infrared array camera ( irac ) and multiband imaging photometer ( mips ) instruments .
various ir selection methods for obscured agns have been proposed . _
spitzer _ irac power - law selection chooses sources whose irac spectral energy distributions ( seds ) follow a power law with a slope of @xmath27 ( @xmath28 ; e.g. , @xcite ) .
this criterion was determined based on the average optical - to - ir spectral properties of x - ray selected qsos ( e.g. , * ? ? ?
it can select a relatively pure sample of agns with little contamination from star - forming galaxies but is generally limited to the most - luminous sources ( e.g. , * ? ? ?
irac color - color selection applies color cuts in the irac color space to identify agns ( e.g. , * ? ? ?
* ; * ? ? ?
this method is based on the irac color properties of optically selected qsos and agns , and it tends to have significant contamination from star - forming galaxies when applied to deep mid - ir data ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
ir - excess selection uses a combined uv
ir color cut to identify agns ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this technique is also contaminated by star - forming galaxies ( e.g. , * ? ? ?
compared to the above methods , which utilize observational data directly and determine the selection criteria empirically , @xcite adopted a generally different approach to find obscured agns .
they selected candidates with significantly higher ir - based ( also including the component corresponding to the transmitted uv emission ; see 2.1 below for details ) star formation rates ( sfrs ) than their uv - based ( after correcting for dust extinction ) sfrs .
the amount of excess in the ir - based sfr is a measurement of the excess in the ir emission , which is likely caused by reprocessed emission from the dusty torus of a heavily obscured agn .
this relative ir sfr excess ( isx ) selection method requires and utilizes redshift information to measure the ir excess effectively , and it appears to isolate a different ir sample from the methods above @xcite .
a few of the candidates in @xcite have been reliably identified as ct agns @xcite .
the isx sample is also contaminated by star - forming galaxies due to uncertainties in the calculation of sfrs ; the heavily obscured agn fraction in the @xcite sample is estimated to be @xmath16 @xcite , and the rest of the objects are likely star - forming galaxies . in this paper ,
we improve the isx selection method and utilize it to search for undetected , heavily obscured agns at @xmath21 in the _ chandra _ deep field - south (; @xcite ) and extended _ chandra _ deep ( e - cdf - s ; @xcite ) . we choose to apply the isx method on the cdf - s and e - cdf - s galaxy sample in this redshift range for the following reasons : * _ ir - selection of heavily obscured agns at @xmath29 is a poorly explored territory .
_ most of the previous ir selections were focused on samples at @xmath30 , including the @xcite work . since agns at @xmath21 contribute significantly to the resolved xrb ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and heavily obscured agns in this redshift range are expected to play a similar role in the missing fraction of the xrb ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , it is of interest to study the heavily obscured population at these lower redshifts .
@xcite selected a sample of ir - excess sources in the cosmos field , which included some objects at @xmath31 .
however , detailed studies ( e.g. , the agn fraction ) were not focused on this low - redshift bin .
moreover , the x - ray and ir ( 24 @xmath32 m ) flux limits in the cdf - s field are @xmath33 times more sensitive than those in the cosmos field @xcite , allowing identification of the population of heavily obscured agns with lower intrinsic luminosities ( which are likely more numerous ) .
recently , @xcite selected a sample of 19 ir - excess sources at @xmath34 in the aegis and goods - n fields , but they argued that most of those x - ray undetected candidates are not heavily obscured agns based on ir sed modeling . * _ the isx method is best applicable at @xmath29 . _
the isx method proposed by @xcite is a promising technique because of its simple physical motivation and its success in selecting identified ct agns at @xmath30 .
however , the isx method has not been applied to any other studies , partially due to its known limitations .
there are two essential quantities in the isx method , the ir - based sfr ( @xmath35 ) and the uv - based sfr ( @xmath36 ) .
the ir - based sfr is derived mainly from the observed _ spitzer _ mips 24 @xmath32 m flux using the @xcite galaxy sed templates .
it has been noted that at @xmath37 , the ir luminosity ( and thus @xmath35 ) derived this way is overestimated by an average factor of @xmath14 @xcite , m and the ir seds at @xmath37 ; they show weaker pah emission than that in high - redshift galaxies by an average factor of @xmath14 . ]
leading to additional contamination from galaxies in the isx sample .
the uv - based sfr is calculated using the dust - extinction corrected uv luminosity .
@xcite used an empirical color - extinction relation to estimate the dust extinction .
however , this correlation can not be applied to relatively old galaxies that are intrinsically red ( dominated by old stellar populations instead of being reddened ) , and all these galaxies were removed from the @xcite sample.median(ssfr)/3 , where ssfr is the specific sfr , defined as @xmath38 with @xmath39 being the stellar mass . about 15% of the sources were excluded this way , the median stellar mass of which is @xmath14 times larger than that of the remaining sources . ] as agns tend to be hosted by massive , red galaxies ( e.g. , * ? ? ?
* ) , this limitation might significantly affect the completeness of the resulting isx sample .
both these problems can be substantially alleviated if the isx method is applied at lower redshifts .
has the advantage of being able to get a better contrast between star - formation and agn emission at the observed 24 @xmath32 m wavelength than at lower redshifts . ]
the 24 @xmath32 m deduced ir luminosity is robust at @xmath40 ( e.g. , * ? ? ?
* ) , while the dust extinction can be derived via sed fitting ( e.g. , * ? ?
* ; * ? ? ?
the technique is reliable with the high - quality multiwavelength photometric data achievable at @xmath40 . * _ the cdf - s and e - cdf - s are excellent fields in which to perform such a study .
_ they have been covered by extensive multiwavelength photometric and spectroscopic surveys , including very deep x - ray and 24 @xmath32 m exposures .
the combo-17 @xcite and subaru @xcite surveys covering the entire cdf - s and e - cdf - s are particularly valuable for the determination of dust extinction via sed fitting .
the 17-filter coverage of combo-17 between @xmath41 and @xmath42 and 18 medium - band coverage of subaru between @xmath43 and @xmath44 are useful for distinguishing between a red dust - free galaxy and a blue dusty galaxy .
reliable photometric redshifts ( photo-@xmath45 s ) can also be obtained with the high - quality multiwavelength data available . *
_ we can assess the possibility of directly detecting the missing population of heavily obscured agns with future hard x - ray observatories .
_ one of the science goals of several future hard x - ray missions , such as the _ nuclear spectroscopic telescope array _ ( _ nustar _ ;
@xcite ) and _
astro - h _
@xcite , is to resolve better the xrb at @xmath4630 kev via detecting heavily obscured and ct agns directly in the distant universe .
while ir - selected heavily obscured agn candidates at @xmath30 are likely below their sensitivity limits , @xmath1 , the 1030 kev flux of the intrinsically luminous ( kev luminosity of @xmath47 erg s@xmath10 ) ct agn hdf - omd49 at @xmath48 @xcite will still be below the detection limit of _ nustar _ with a 1 ms exposure ( the _ nustar _ sensitivity limit is from http://www.nustar.caltech.edu/ ) .
the average luminosity of the ct agn candidates in @xcite is about an order of magnitude fainter than that of hdf - omd49 .
] candidates at @xmath21 should be more easily detectable , and we will critically assess such possibilities .
this paper is organized as follows .
we describe the multiwavelength data and isx sample selection in 2 . in
3 we present an x - ray stacking analysis of the isx sample and estimate the heavily obscured agn fraction among the sample via simulations . in 4 , we calculate the space density of the selected heavily obscured agns and their contribution to the xrb .
we also discuss the feasibility of detecting these isx sources with future hard x - ray observatories .
we summarize in 5 . throughout this paper , we adopt the latest cosmology with @xmath49 km s@xmath10 mpc@xmath10 , @xmath50 , and @xmath51 @xcite .
all given magnitudes are in the ab system ( e.g. , * ? ? ?
* ) unless otherwise stated .
the cdf - s is the deepest x - ray survey ever performed , having a total _
chandra _
exposure of @xmath52 ms and covering a solid angle of @xmath53 arcmin@xmath54 @xcite .
x - ray images in three standard bands , ( full band ; fb ) , ( soft band ; sb ) , and ( hard band ; hb ) , along with other relevant products such as the main x - ray source catalog ( 740 sources ) and sensitivity maps , are used in the sample creation and x - ray stacking analysis , which are discussed below and in 3 .
the cdf - s is flanked by the e - cdf - s , which consists of four contiguous @xmath55 ks _ chandra _ observations with a total solid angle of @xmath56 arcmin@xmath54 @xcite .
the e - cdf - s main x - ray source catalog ( 762 sources ) is used only in the sample creation .
mid - ir - to - optical multiwavelength data are required to calculate the ir - based and sfrs in the isx method .
the cdf - s and e - cdf - s have been covered by extensive photometric and spectroscopic surveys , and we constructed a sample of mid - ir and optically selected sources in the cdf - s and e - cdf - s region . for the mid - ir data , we used the _ spitzer _
mips 24 @xmath32 m source catalog from the _ spitzer _ far infrared deep extragalactic legacy survey ( fidel ; m. dickinson et al .
2011 , in preparation ) .
the fidel survey covers the entire cdf - s and , and it has a 5 @xmath17 limiting flux of @xmath25 @xmath32jy . for the data , we used the master source catalog compiled by @xcite , which consists of @xmath57 optically selected galaxies covering the entire cdf - s and .
the master source catalog was created based on the musyc catalog @xcite , the combo-17 catalog @xcite , and the goods - s music catalog @xcite , and it was also cross - matched to several other photometric catalogs such as the simple irac catalog @xcite and the _ galex _ uv catalog ( e.g. , * ? ? ?
* ) to include up to 42 bands of ir - to - uv data
. additionally , we included the 18 medium - band subaru photometric data that have become available recently @xcite ; about 80% of the sources in the master catalog have subaru data .
for our purpose of computing reliably the uv - based sfrs ( requiring robust sed fitting ) , we chose sources with relatively bright @xmath58-band magnitudes ( @xmath59 ) ; there are @xmath60 such objects .
these sources were matched to the 24 @xmath32 m sources using the likelihood - ratio matching technique described in @xcite , resulting in 5237 matches and a false - match probability of @xmath61 . for sources in the x - ray stacking samples discussed in 3 ,
we have visually examined their ir and optical images and removed sources that are probably affected by source blending , and thus the false - match probability is negligible for those samples .
reliable spectroscopic redshifts ( spec-@xmath45 s ; 979/5237 sources ) for sources in the cdf - s and e - cdf - s were collected from the following catalogs : @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
if a spec-@xmath45 is not available for a given source , we calculated its photo-@xmath45 using the zurich extragalactic bayesian redshift analyzer ( zebra ; @xcite ) .
zebra utilizes a maximum - likelihood approach to find the best - fit sed template ; more details about the zebra sed - fitting procedure are discussed in 3.2 of @xcite . for the purpose of our study here , we used zebra to derive photo-@xmath45 s and dust extinction ( in terms of the extinction in the @xmath62 band , @xmath63 ) simultaneously using up to 60 photometric bands .
we selected 99 typical pegase galaxy templates from @xcite that span a wide range of star - formation history , and we applied intrinsic extinction in the range @xmath644 with an increment of 0.1 to these templates employing the @xcite extinction law .
the resulting sed templates were used to fit the sed data ( excluding the 24 @xmath32 m data ) of the sources . for sources with spec-@xmath45 s ,
the redshifts were fixed at the spec-@xmath45 values during the sed fitting , and the values of @xmath63 were obtained from the best - fit templates . for the other sources ,
both the photo-@xmath45 and @xmath63 values were determined based on the best - fit templates . to check the quality of the photo-@xmath45 s
, we performed another zebra run of the spec-@xmath45 sources , setting the redshift as a free parameter , and then compared the resulting photo-@xmath45 s with the spec-@xmath45 s .
the photo-@xmath45 accuracy was estimated using the outlier fraction and the normalized median absolute deviation ( @xmath65 ; e.g. , @xcite ) parameters .
outliers are defined as sources having , where @xmath66 , and @xmath65 is defined as @xmath67 for all the spec-@xmath45 sources , we found an outlier fraction of 6% and @xmath68 .
for sources in the redshift range of 0.51 , which are of primary interest for this study , the outlier fraction is 3% and @xmath69 .
these photo-@xmath45 s have comparably high quality to those recently obtained for galaxies in the cdf - s region @xcite .
we calculated the ir - based and uv - based sfrs following @xcite , @xmath70 and @xmath71 where @xmath72 erg s@xmath10 and @xmath73 , @xmath74 , and @xmath75 are the ir , uv , and corrected uv luminosities , respectively .
the @xcite initial mass function was adopted here .
the ir luminosity was estimated by finding an ir sed template that produces the observed 24 @xmath32 m luminosity via interpolation of a library of 105 ir seds and then calculating the integrated 81000 @xmath32 m luminosity for this template @xcite .
it has been demonstrated that the ir luminosity determined this way is similar to that derived with additional longer wavelength photometric data and mid - ir spectroscopic for sources with @xmath76 and @xmath77 @xcite .
the uv luminosity was computed following @xmath78 ( see 3.2 of @xcite ) , where the rest - frame 2800 monochromatic luminosity @xmath79 was interpolated from the multiwavelength data .
we calculated the correction for the uv luminosity employing the @xcite extinction law , @xmath80 .
as agns tend to reside in massive galaxies ( e.g. , * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , we estimated stellar masses for the 5237 galaxies and used these masses to filter out low - mass objects . the stellar mass is calculated following @xcite : @xmath81 where @xmath82 is the monochromatic @xmath83-band luminosity of the sun ( @xmath84 erg s@xmath10 hz@xmath10 ) , the coefficients @xmath85 and @xmath86 are from @xcite , and the monochromatic luminosity @xmath87 and color @xmath88 in the vega system are from the results .
all the sources have irac detections , and thus they all have @xmath83-band coverage in the sed fitting .
the normalization has been adjusted by @xmath89 dex to account for our adopted @xcite initial mass function .
the @xmath83-band luminosity was used because it is @xmath1410 times less sensitive to dust and stellar - population effects than optical luminosities ( e.g. , * ? ? ?
we define our parent sample from the 5237 optically and 24 @xmath32 m selected galaxies above with the following criteria : ( 1 ) the redshift of the source is between 0.5 and 1 ; there are 2037 such sources .
( 2 ) the source must have combo-17 or subaru detections ( in @xmath90 photometric bands ) to ensure robust sed fitting ; there are 24 sources removed by this criterion .
( 3 ) the source has a stellar mass @xmath91 , as agns tend to be hosted by massive galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) ; this criterion removes @xmath16 of the sources .
( 4 ) the source is not x - ray detected or overlapping with the 90% encircled - energy aperture ( see 3.2 of @xcite ) of any known x - ray source ; the latter criterion is to avoid any contamination from nearby x - ray sources in our x - ray stacking analysis .
we used the 4 ms cdf - s and 250 ks e - cdf - s x - ray source catalogs for this purpose @xcite .
about 200 sources are removed by this criterion .
we note that the basic source properties , such as the dust extinction , stellar mass , and uv - based sfr , were calculated based on the assumption that the optical sed is dominated by the host galaxy ; this assumption is valid after we exclude the x - ray sources from the sample ( e.g. , see 4.6.3 of @xcite for further discussion ) .
the parent sample defined above consists of 1313 sources , about 20% of which have .
the spectroscopic completeness does not have any significant dependence on the ir or optical magnitude , and it is largely limited by the spectroscopic coverage in the area ; for example , @xmath23 of the sources in the goods - s region have spec-@xmath45 s .
we show in figure [ plotav ] the dust extinction ( @xmath63 ) against the ratio of the ir - based ( @xmath35 ) and uv - uncorrected [ @xmath92 sfrs , which is also a measure of extinction .
the @xmath63 values derived from sed fitting can well account for the extinction in general . in figure 2
, we plot the logarithmic ratios of the ir - based and uv - based sfrs , defined as @xmath93 .
the distribution of @xmath94 displays an excess on the positive side ( @xmath95 ) similar to that observed in @xcite for @xmath30 galaxies .
various statistical errors are present in the calculations of @xmath35 and @xmath36 , such as the uncertainties in @xmath73 , @xmath74 , and @xmath63 .
these tend to make the @xmath94 distribution follow approximately a gaussian function .
we mirror the negative half of the histogram to the positive side , and the combined distribution ( blue histogram in fig .
2 ) can be approximated by a gaussian function with @xmath96 .
we thus consider that the excess in the distribution of @xmath94 is caused by galaxies hosting heavily obscured or even ct agns .
we define isx sources using the same criterion as in @xcite : @xmath97 . for comparison , we define ir sfr normal ( isn ) sources as those having @xmath98
. there are 242 isx and 736 isn sources in the parent sample . assuming that the intrinsic dispersion of @xmath94 is gaussian ( blue histogram in fig .
2 ) , and that the excess ir emission is powered by a heavily obscured agn , about 73% of the isx sources ( 176 objects out of the 242 isx sources ; shaded region in fig .
2 ) host such agns . for simple comparison , we selected x - ray detected sources in the same way from the initial 5237 sources ( revising the fourth criterion above to require detection in the cdf - s or e - cdf - s source catalogs ) ; there are 198 x - ray sources selected . following the agn classification scheme in 4.4 of @xcite , including criteria for intrinsic x - ray luminosity , effective power - law photon index , and x - ray - to - optical flux ratio
, we identified 155 agns from the x - ray sources , mostly unobscured or moderately obscured .
this number is comparable to that of expected heavily obscured agns ( 176 objects ) in the isx sample , consistent with predictions from population - synthesis models ( e.g. , * ? ? ?
* ) .
the isx and isn sources have different physical properties . the median stellar mass , ir luminosity , and ir - based sfr of the isx sample are 2.2 , 1.4 , and 1.2 times those for the isn sample , respectively , while the median uv luminosity , uv - based sfr , and @xmath63 of the isx sample are 50% , 15% , and 30% those for the isn sample .
if the relative excess in the ir luminosities is contributed by underlying heavily obscured agns , then such objects selected by the isx method at @xmath21 appear to be hosted by massive galaxies with little dust . since agns are preferentially hosted by massive galaxies ( e.g. , * ?
* ; * ? ? ?
* ; * ? ? ?
* ) , it is natural to find isx sources in such galaxies .
the dust extinction ( in terms of @xmath63 ) applies to the entire galaxy , and a small value of @xmath63 does not necessarily conflict with a high absorbing column density for a heavily obscured agn in the nucleus ( e.g. , * ? ? ?
* ; * ? ? ?
the finding of smaller @xmath63 values in the isx sources may be a selection effect .
heavily obscured agns at are numerically dominated by moderate - luminosity objects ( intrinsic x - ray luminosity according to population - synthesis models ( e.g. , * ? ? ?
* ) , and sources hosted by galaxies with less dust ( less ir emission from the host galaxy ) will have more prominent ir excesses and are more likely to be identified by ir selection methods . as a test of whether isx sources are preferential hosts of obscured agns , we also included x - ray detected cdf - s agns @xcite in figure 2 .
there are 74 cdf - s agns selected from the initial 5237 sources ( revising the fourth criterion above to require detection in the cdf - s source catalog ) .
we further removed 6 luminous agns with @xmath99 erg s@xmath10 as the optical seds for such sources may be affected by agn contamination ( e.g. ,
4.6.3 of * ? ? ?
the level of intrinsic absorption was estimated by assuming an underlying x - ray power - law photon index of @xmath100 and using xspec ( version 12.5.1 ; @xcite ) to derive the appropriate @xmath101 value that produces the observed x - ray band ratio ( defined as the ratio of count rates between the hb and sb ) .
we consider a source to be obscured if @xmath102 @xmath1 ; otherwise , it is unobscured or only weakly obscured
. figure 2 shows clearly that agns in the isx ( @xmath97 ) region are mostly obscured ( 17 cases out 19 ) , while only half of the agns in the isn region are obscured .
we caution that the nature of the isx population as a whole is likely different from these x - ray detected agns , given the apparent difference in the ir - to - x - ray flux ratios , and thus we perform detailed x - ray studies of the isx sample in the following section .
x - ray stacking analysis is a useful technique to obtain the average x - ray properties and probe the nature of a sample of x - ray undetected objects , and it has been used extensively in previous studies of ir - selected agn candidates ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we performed x - ray stacking of the isx and isn samples in the three standard x - ray bands of the 4 ms cdf - s , fb , sb , and hb . we did not stack the isx and isn sources in the 250 ks e - cdf - s because the total x - ray exposure is low and no significant stacked signal can be achieved .
sources located further than 6 from the average aim point are not included in the stacking due to sensitivity degradation at large off - axis angles ( e.g. , * ? ? ?
* ; * ? ? ?
we visually examined the ir and optical images of the remaining isx and isn sources , and excluded additional 10 sources ( two isx and eight isn sources ) whose 24 @xmath32 m data appear to be affected by source blending . the final isx and isn samples used in the stacking contain 23 and 58 sources , respectively .
arcmin@xmath54 , and thus the number of isx sources within the inner 6-radius area is only @xmath103 of the total . ] about 70% of these sources have spec-@xmath45 s .
we also visually checked the x - ray images and verified that these isx and isn sources are not close to any bright x - ray sources which could affect the stacking results .
basic properties of the isx and isn sources used in the x - ray stacking analysis are listed in table 1 .
we followed a stacking procedure similar to that discussed in @xcite . for each source in each of the three standard bands , we calculated the total ( source plus background ) counts within a 3-diameter circular aperture centered on its optical position .
this extraction radius was found to produce the best signal - to - noise ratio ( s / n ) ( e.g. , * ? ? ?
* ) compared to other choices .
the corresponding background counts within this aperture were determined with a monte carlo approach which randomly ( avoiding known sources ) places 1000 apertures within a 1-radius circle of the source position to measure the mean background ( e.g. , * ? ? ? * ) .
the total counts ( @xmath104 ) and background counts ( @xmath105 ) for the stacked sample were derived by summing the counts for individual sources .
the net source counts are then given by @xmath106 , and the s / n is @xmath107 ; we note that the numbers of source and background counts are large ( @xmath108 ) , and thus the s / n can be calculated using gaussian statistics .
the 3-diameter aperture does not encircle the full point - spread function ( psf ) .
thus we determined an encircled - energy fraction ( eef ) for each source in each band by interpolating an eef table given the aperture radius , angle , and photon energy .
the eef table was derived from the psf images of the sources in the 4 ms cdf - s generated by acis extract ( ae ; @xcite ) that uses the marx ray - trace simulator ; psf images at five different photon energies ( @xmath1098.5 kev ) for each source in each _ chandra _ observation were provided by ae ( see 3.2 of @xcite ) .
we created the eef table by calculating the eefs at different aperture radii , off - axis angles , and photon energies , averaged over all the observations weighted by exposure time .
the eefs calculated this way are the best representative of the real cdf - s data .
the eef in a 3-diameter aperture has a strong off - axis angle dependence , being @xmath5 for on - axis sources , and @xmath24 for sources at a 6 off - axis angle .
the aperture correction for the stacked counts was calculated using the exposure - weighted average of the eefs for all the sources .
the effective power - law photon indices and fluxes were calculated based on the band ratios and aperture - corrected count rates using the cxc s portable , interactive , multi - mission simulator ( pimms ; see details in 3.3.1 of @xcite ) .
the stacking results are listed in table 2 , which show statistically significant differences between the isx and isn samples . for the isn sample , there is a strong detection ( @xmath110 ) in the sb , while the stacked signal is weak ( @xmath111 ) . treating this kind of weak signal as a detection does not affect our later analyses . ] in the hb .
the corresponding band ratio is @xmath112 ( 1 @xmath17 errors ) , which indicates an effective power - law photon index of @xmath113 , consistent with x - ray emission from starburst galaxies ( e.g. , * ? ? ?
the isx sample has a @xmath114 detection in the sb and a @xmath115 detection in the hb ( corresponding to rest - frame @xmath11614 kev ) , with a band ratio of @xmath117 , corresponding to @xmath3 .
x - ray sources with very flat spectral slopes ( @xmath118 ) are almost exclusively identified as heavily obscured agns ( e.g. , * ? ? ?
* ) , and thus a significant contribution from obscured agns is required to produce this kind of hard x - ray signal for the isx sample .
we note that the isn sample may still contain a fraction of low - luminosity agns with soft spectra comparable to those of star - forming galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
these agns are generally unobscured and are not of primary interest to this study . in figure 3 , we show the adaptively smoothed stacked images for the isx and isn samples .
it is clear that the isx sample has a much harder x - ray signal than the isn sample .
the difference between the isx and isn samples is also apparent via comparison of their fluxes ; the isx sb flux is @xmath119 smaller than the isn sb flux , while the isx hb flux is @xmath14 times higher than the isn hb flux .
the stacked x - ray signal ( @xmath120 ) for the isx sample indicates the existence of embedded agns and is consistent with x - ray emission from heavily obscured or ct agns modeled with a reflection spectrum ( e.g. , * ? ? ?
* ; * ? ? ?
* ) or a spectrum from a toroidal reprocessor @xcite .
however , we can not derive the average agn properties using the stacked signal directly , as the isx sources have a range of redshifts and the sample is likely contaminated by star - forming galaxies ( see , e.g. , fig .
[ f01 ] ) .
we thus performed monte carlo simulations to assess the fraction of heavily obscured agns / ct agns in the sample , using a procedure refined from that discussed in @xcite . for a given agn fraction , we ran 10000 simulations to estimate the expected x - ray emission from both the star - forming and agn activities of the sample , and we then compared the average simulated results to the observed stacked signal .
we first used the isn sample ( 58 star - forming galaxies ) to test our simulation method .
for galaxies , a number of studies have found an approximately linear correlation between the sfr and rest - frame 210 kev x - ray luminosity @xmath121 ( e.g. , * ? ? ? * ; * ? ? ? * ) .
however , later studies have pointed out that the x - ray
sfr correlation is likely to have significant scatter due to the fact that sfr only relates to the population of relatively young high - mass x - ray binaries ( hmxbs ) , and the older population of x - ray binaries ( lmxbs ) likely relates to the galaxy stellar mass ( e.g. , * ? ? ?
* ; * ? ? ?
therefore , we adopted the most recent relation presented in @xcite , which considers the correlation of @xmath121 with both the sfr and stellar mass , @xmath122 where @xmath123 erg s@xmath10 @xmath124 and @xmath125 erg s@xmath10 ( @xmath126 yr@xmath10)@xmath10 ; here @xmath127 has been increased by a factor of 2.6 to account for the average difference between stellar masses calculated in this work and @xcite .
ideally , the two sfrs for a given isn galaxy , @xmath35 and @xmath36 , should be equal to each other .
we consider that the gaussian spread in @xmath94 for isn galaxies shown in figure 2 is caused by uncertainties in the two sfrs , and the intrinsic sfr is estimated as @xmath128/2 $ ] .
the corresponding 210 kev x - ray luminosity was then derived from equation [ sfrx ] , also including a random gaussian dispersion of 0.34 dex @xcite to account for the scatter of that relation .
we assumed an x - ray power - law photon index of @xmath129 ( given the stacking results ) and no intrinsic absorption for the star - forming galaxies , and then converted the predicted 210 kev x - ray luminosity to the observed sb and hb fluxes using the individual redshifts of the sources .
as none of the isn sources is individually x - ray detected , we required that the simulated fluxes in the sb and hb of each source do not exceed the sensitivity limits at the source position .
the x - ray sensitivity limits are derived following @xcite , by calculating the minimum flux at each pixel ( converted from the minimum number of counts ) required for detection under the catalog source - detection criteria .
we added a small gaussian dispersion of 0.1 dex to the sensitivity limits to account for statistical fluctuations ( estimated using the few cdf - s x - ray sources with fluxes below the nominal sensitivity limits ) .
if the simulated x - ray fluxes are greater than the sensitivity limits , we made another random generation of the source properties ( in this case , only eq .
[ sfrx ] and the sensitivity limits have random scatters ) .
the expected sb and hb counts were calculated given the exposure time , ( using pimms ; see @xcite ) , and eef of the source . for each simulation
, net counts of every source were summed in the sb and hb as the output counts .
we performed 10000 simulations , and the average numbers of output counts are @xmath130 for the sb and @xmath131 for the hb ( the errors are the standard errors of the mean ; the 1 @xmath17 dispersions for the average numbers of counts are both @xmath132 ) , matching well with the observed stacked counts ( @xmath133 and @xmath134 )
. therefore we can successfully reproduce the x - ray emission from star - forming activity with this simulation approach .
we then performed simulations for the 23 sources in the isx sample adopting a similar approach , calculating x - ray emission from both the star - forming and agn activities of the sample . assuming an agn fraction in the range of @xmath4690% , the x - ray emission from the star - forming galaxies in the sample was derived following the same procedure as above . for a source selected as an agn , its x - ray emission and ir emission consist of both a star - formation component ( host galaxy ) and an agn component
for the star - formation component , we estimated its ir - based sfr to be @xmath135 , where the factor @xmath136 is randomly drawn from a gaussian function with @xmath96 ( blue histogram in fig . [ f01 ] ) to account for the scatter of the calculated sfr .
the intrinsic sfr of the star - formation component was then estimated as @xmath137/2 $ ] , and the corresponding x - ray emission was calculated . for the agn component , the intrinsic 210 kev luminosity , @xmath138 , can be estimated from the rest - frame 6 @xmath32 m luminosity of the agn , @xmath139 ( e.g. , * ? ? ?
the relation in @xcite was derived using a sample of local agns with the 6 @xmath32 m agn continua decomposed from the ir spectra .
here we used the luminosity - dependent version of the relation which takes into account the possible luminosity dependence @xcite , @xmath140 a gaussian dispersion of 0.4 dex was added to account for the scatter of the relation , estimated based on the range of intrinsic x - ray - to - mid - ir luminosity ratios for local agns @xcite .
we note that this relation is consistent with the x - ray - to - mid - ir luminosity ratios of a sample of x - ray selected obscured quasars ( sturm et al .
2006 ) , suggesting that it is applicable for deriving the intrinsic luminosity of heavily obscured or ct agns . to derive the rest - frame 6 @xmath32 m monochromatic luminosity coming from the agn , we first computed the agn 24 @xmath32 m flux by subtracting the star - formation contribution from the observed 24 @xmath32 m flux ; the star - formation contribution to the 24 @xmath32 m flux was estimated using the ir - based sfr ( @xmath141 ) above and the @xcite ir sed templates .
the residual 24 @xmath32 m flux was then converted to the rest - frame 6 @xmath32 m flux using the ir sed of the local ct agn ngc 1068 ( sed data from @xcite and @xcite ) .
the observed 24 @xmath32 m wavelength corresponds to rest - frame 1216 @xmath32 m for our sources , and the conversion factor to the rest - frame 6 @xmath32 m flux is @xmath1420.3 .
we adopted an intrinsic photon index of @xmath100 for the agn and an @xmath101 value randomly drawn from the @xmath101 distribution shown in figure 7 of @xcite , in which about half of the agns are ct .
the adopted @xmath101 distribution has only a small impact on the final @xmath101 distribution , as will be discussed later .
given the intrinsic 210 kev luminosity , absorption , and photon index , we estimated the observed absorbed sb and hb fluxes using the mytorus model @xcite implemented in xspec .
the mytorus model is a recently developed spectral - fitting suite for modeling the transmitted and scattered x - ray spectra from a toroidal - shaped reprocessor , and it is especially designed to treat the spectra of ct agns . compared to the commonly used disk - reflection model for ct agns ( e.g. , the _ pexrav _ model in xspec ; see @xcite ) , the mytorus model is more physically motivated and takes into account absorption and scattering simultaneously .
we only considered the continuum output from mytorus for simplicity , neglecting the fluorescent iron emission lines which contribute less than 8% of the hb continuum flux .
two basic parameters describing the geometry of the model are the angle of the torus and the inclination angle ( 0 corresponds to a face - on observing angle ) .
the angle was set at 37 , corresponding to a scenario where the number of obscured agns is four times the number of unobscured agns ; such an obscured agn fraction ( @xmath6 ) for moderate - luminosity agns at @xmath34 has been reported in several studies ( e.g. , * ? ? ?
* ; * ? ? ?
we randomly chose an inclination angle between 37 and 90 ( probability weighted by solid angle ) for each source .
other parameters , such as the relative cross - normalizations of different components , were set at the default values ( see 8.2 of the mytorus manual ) .
the @xmath101 value in the mytorus model has an upper limit of @xmath143 @xmath1 . for column densities beyond this limit ,
the observed spectrum is highly geometry dependent , and the scattered flux is likely dominated by scattering in any optically thin parts of the torus , even if the effective solid angle of the optically thin gas is tiny ( t. yaqoob and k. d. murphy 2011 , private communication ) .
the above parameterization of the model can not reproduce the exact environment of real obscured agns , which could be much more complicated ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , but we consider it to be the best available approximation of the average properties .
we then combined the emission emerging from this obscured agn component with that from the star - formation component to derive the expected observed emission .
as for the simulations of star - forming galaxies , we required that the observed sb and hb fluxes do not exceed the sensitivity limits at the source position .
we would continue randomly regenerating the source properties ( e.g. , @xmath101 value ) until this requirement is satisfied .
this is a strong constraint on the simulated source properties , and it generally requires the agn to be heavily obscured or ct . finally , we converted the fluxes into counts in the extraction aperture , and added the sb and hb counts from every source as the output of a single simulation .
we again performed 10000 simulations , and the average output counts were compared to the stacking results in table 2 .
the simulation results are shown in figure 4 , with the simulated sb and hb counts at different assumed agn fractions .
as the agn fraction increases , the band ratio also increases and the simulated spectrum becomes harder .
we note that the simulated sb photons almost exclusively originate from star - forming galaxies and agn hosts ( their star - formation components ) , while the hb photons are mainly from the agns ; the increase of band ratio with agn fraction is largely due to the increase in the hb counts .
however , we did not include a soft - scattering component in the spectra of heavily obscured agns , which has been observed to provide significant sb emission in most of the local ct agns ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
the normalization of this component to the intrinsic power - law continuum is around 1% but has large object - to - object variations and uncertainties ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and the resulting contribution to the sb emission relative to the host - galaxy contribution depends on both this normalization and the intrinsic x - ray luminosity . therefore the sb simulated counts should be considered lower limits , and we use only the hb results to determine the agn fraction . as shown in figure 4 , an agn fraction of @xmath4 can produce the observed hb counts ( see table 3 ) .
star - forming galaxies can still host low - luminosity agns , as long as the agn x - ray luminosities are not comparable to their host galaxies ; this kind of agns is not the focus of this study . ]
the error on the agn fraction was determined given the error on the observed hb counts .
the @xmath144 agn fraction is in good agreement with the fraction of extra sources ( 73% , shaded area ) in the @xmath94 distribution plot in figure 2 . for this best - fit agn fraction
, we also computed the stacked and simulated signals in several sub - bands , and the comparisons are shown in table 3 .
the simulations reproduced well the stacked counts in general .
the cumulative @xmath101 distribution for the 75% of agns in the 10000 simulations is displayed as an inset of figure 4 , where @xmath5 of the agns are heavily obscured , and @xmath6 are ct .
the ct agn fraction in the entire isx sample is thus @xmath23 .
the isx agns have moderate x - ray luminosities . the median intrinsic 210 kev luminosity derived from the simulations is @xmath145 erg s@xmath10 ; this value would be @xmath24 higher if the intrinsic luminosity was calculated ( equation 6 ) using the observed ir flux ( without subtracting the star - formation contribution ) .
the interquartile range of the x - ray luminosities is ( 0.94)@xmath146 erg s@xmath10 [ ( 15)@xmath146 erg s@xmath10 ] for the heavily obscured ( ct ) agns .
we explored how the agn fraction could change by varying some of the assumptions in the simulations .
first , we tried the commonly used disk - reflection model instead of the mytorus model to compute obscured agn emission .
an absorbed power - law ( _ zwabs*zpow _ in xspec ) plus reflection ( _ pexrav _ in xspec ) model was used .
the absorbed power - law component typically dominates in the compton - thin ( @xmath148 @xmath1 ) regime , and the reflection component dominates in the ct regime .
we note that the _ pexrav _ model has non - negligible uncertainties on some of its parameters , such as the reflection scaling factor . for simplicity
, we adopted the same parameters used in @xcite to model the xrb ; we assumed a cut - off energy of 200 kev , a reflection scaling factor of 0.37 , and an inclination angle of 60 for the reflecting material .
the hard x - ray flux produced with this model is about three times higher than that from the mytorus model @xcite , and thus the required agn fraction in the isx sample is smaller .
the fraction drops from @xmath4 to @xmath149 , still indicating a substantial agn contribution .
again , almost all ( @xmath5 ) of the agns are heavily obscured .
simply increasing the @xmath138-to-@xmath139 ratio in equation [ irlxagn ] also reduces the agn fraction .
for example , if we adopt the relation used in @xcite , which gives an x - ray luminosity that is @xmath142 dex more luminous than ours , the resulting agn fraction is @xmath150
. the initial @xmath101 distribution input into the simulation will affect the agn fraction slightly .
we adopted an alternative initial @xmath101 distribution that has only @xmath151 ct agns , which is similar to the @xmath101 distribution for high excitation - line galaxies in @xcite .
the derived agn fraction is @xmath152 with @xmath153 being ct .
the ct agn fraction in the entire isx sample ( @xmath154 ) is thus comparable to that ( @xmath23 ) for the best - fit result above . even in an extreme case where the initial @xmath101 distribution has only 10% ct agns , an agn fraction of @xmath23 and a ct agn fraction of @xmath154
are still required to produce the stacked x - ray emission and satisfy the requirement that the simulated sources can not be individually detected in the sb or hb .
we note that the adopted photon index for star - forming galaxies ( @xmath129 ) has little effect on the estimated agn fraction as these galaxies produce mainly sb counts and we do not consider sb counts to be constraining . in general , the requirement that the simulated sources can not be individually detected places a strong constraint on the final results , and an agn fraction of @xmath155 ( mostly heavily obscured ) is always expected .
we also caution that the simulation results were taken to be the average of 10000 tests , and thus they ( as well as the following analyses based on these results ) only represent the most probable scenario of the isx source properties and might deviate from the real source properties .
we consider that this is the best available approach to probe the nature of isx sources given the x - ray and multiwavelength data available .
given the simulation results for the isx sample , we can estimate the space density of heavily obscured agns at @xmath21 and calculate their contribution to the xrb at high energies ( 1030 kev ) .
there are @xmath156 ( @xmath157 ) heavily obscured ( ct ) agns expected in the stacking isx sample .
these objects are located within the central 6-radius region of the cdf - s ; the corresponding area is 0.029 deg@xmath158 after excluding the regions masked by x - ray sources .
we further correct the number of sources by 10% to account for the fraction removed due to blended 24 @xmath32 m photometry .
therefore the sky density of heavily obscured ( ct ) agns in our isx sample is @xmath159 deg@xmath20 ( @xmath160 deg@xmath20);1 x - ray agns from the original 5237 galaxies in the same region for the 4 ms cdf - s is @xmath161 deg@xmath20 . ] the corresponding space density is @xmath162 mpc@xmath12 [ @xmath11 mpc@xmath12 ] given the comoving volume between redshift 0.5 and 1 .
note that there are several factors affecting the estimated space density above .
the expected number of agns in the isx sample may be up to @xmath164 lower if different assumptions are made in the simulations ( 3.2 ) , and thus the space density will be reduced by the same fraction . on the other hand ,
there is probably a small fraction of sources excluded from the parent sample with stellar mass @xmath165 ( 2.2 ) that also host heavily obscured agns .
also , the stacking isx sample was selected in a relatively small area , and thus we have missed a few rare but intrinsically luminous agns , which is also suggested by the moderate ir luminosities of all the isx sources in table 1 . moreover ,
the x - ray spectra of heavily ct ( @xmath166 @xmath1 ) agns can not be modeled by mytorus and an upper limit of @xmath143 @xmath1 was used for the column density ; if these heavily ct agns are present in the isx sample , we could have overestimated their x - ray emission and thus underestimated the agn fraction and space density .
most importantly , we define the isx sample conservatively ( @xmath97 ) to avoid significant contamination from star - forming galaxies in the sample . there will be additional heavily obscured agns among those sources that are not in the isx or isn sample ( @xmath167 ; see [ appsamples ] below for x - ray stacking results ) ; the expected number is nearly comparable to that in the isx sample given the fraction of excess sources in the distribution of @xmath94 in figure 2 .
taking into account the uncertainties and incompletenesses , we expect that the true space density of heavily obscured agns at @xmath21 is @xmath1682 times that estimated above . in the following analyses
, we still use our conservative estimation as it is difficult to determine the exact correction factor .
we show in figure 5 the space densities of ct agns selected in this work and some previous studies ( mainly ir based ) , as well as those from model predictions by @xcite . the ct agn space density for the @xcite cdf - s sample with and @xmath169 erg s@xmath10 is estimated to be @xmath170 mpc@xmath12 ;
the actual value is expected to be lower ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and was recently revised downward to @xmath171 mpc@xmath12 @xcite . at lower redshift ,
the space density for the @xcite cosmos ct agn sample with @xmath1721.2 is @xmath173 mpc@xmath12 .
we are probing a different population of ct agns from the @xcite cosmos sample in terms of intrinsic x - ray luminosities .
the interquartile range of the intrinsic 210 kev luminosities is ( 15)@xmath146 erg s@xmath10for the ct agns in our sample , while it is ( 310)@xmath174 erg s@xmath10 for the @xcite sample .
we thus selected mainly moderate - luminosity and more typical ct agns ; this is largely attributed to the much higher sensitivities of the mid - ir and x - ray observations in the cdf - s . on the other hand
, we expect that we have missed some rare , unrepresentative objects due to the smaller area of the cdf - s .
we note that the data points in various studies and model predictions have been derived with different assumptions ( see , e.g. , 3.2.2 ) and are thus not strictly comparable .
we estimated the expected xrb flux in the 1030 kev band provided by the heavily obscured agns in the isx sample .
we used the mytorus model to calculate the emergent flux from these agns , given their redshifts , assumed power - law photon index ( @xmath100 ) , intrinsic x - ray luminosities , and column densities ( the latter two quantities are from the simulation results ) . on average ,
these agns produce a flux of @xmath175 kev @xmath1 s@xmath10 sr@xmath10 in the kev band , @xmath176 of which is from ct agns .
the total xrb flux in this band is about 44 kev @xmath1 s@xmath10 sr@xmath10 ( e.g. , @xcite and references therein ) .
therefore , this population of @xmath21 heavily obscured ( ct ) agns selected using the isx method is expected to contribute @xmath177 ( @xmath178 ) of the total xrb in the 1030 kev band .
observations have not directly resolved the xrb at energies @xmath90 kev ; however , theoretical predictions have been made by population - synthesis models about the contributions from unobscured , moderately obscured , and heavily obscured agns .
for example , @xcite concluded that @xmath179 of the xrb in the 1030 kev band ( the missing xrb ) is produced by ct agns , while @xcite considered the total contribution from ct agns to be @xmath180 .
the predicted contribution from the ct agns selected here ( @xmath178 ) is thus @xmath181 @xcite or @xmath182 @xcite of the missing xrb in the 1030 kev band .
given the expected properties ( luminosity , redshift , column density ) of the isx sources , the @xcite model predicts that @xmath183 of the missing xrb comes from such objects , agreeing well with our results .
the remaining missing xrb is largely attributed to luminous ( @xmath184@xmath185 erg s@xmath10 ) ct agns at @xmath186 @xcite ; the stacking isx sample selected here appears to have missed some of the high - luminosity objects due to the limited volume surveyed .
heavily obscured agns are expected also to contribute to the xrb at energies @xmath187 kev , though with a smaller contributed fraction than that for the 1030 kev band ( see , e.g. , fig . 15b of @xcite ) .
the xrb in the 18 kev band , unlike that in the 1030 kev band , has been largely resolved into discrete sources ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . with the deepest _ chandra _ data available , the 4 ms , we expect to resolve the xrb further and improve our understanding of the nature of the x - ray source populations at low flux levels , including the heavily obscured population .
we performed x - ray stacking analyses on the x - ray sources and optical galaxies in the cdf - s to explore the resolved xrb fraction and hidden agn contribution to the unresolved fraction .
measurements of the normalization of the xrb spectrum have non - negligible uncertainties and field - to - field variations ; the combined uncertainty on the normalization is @xmath4620% ( e.g. , * ? ? ?
* ; * ? ? ?
here we adopted the normalization from @xcite , which was derived from the _ chandra _ deep fields data including the 2 ms _ chandra _ deep field - north ( cdf - n ; @xcite ) and 1 ms cdf - s @xcite .
the xrb has a power - law spectral slope with @xmath188 and a normalization of 10.9 photons s@xmath10 kev@xmath10 sr@xmath10 at 1 kev .
x - ray stacking analyses were performed in the following energy bands : 0.51 , 12 , 24 , 46 , and 68 kev . for x - ray sources
, we used the 740 sources in the 4 ms cdf - s main x - ray source catalog @xcite .
the stacking procedure was the same as that described in 3.1 ; to maximize the s / n ratio , a 3-diameter circular aperture was used and the stacking was performed for the 389 sources within the inner 6-radius area .
to account properly for bright x - ray sources that have a rare occurrence in the narrow cdf - s region , we adopted the bright - end correction in @xcite : sources brighter than @xmath189 erg @xmath1 s@xmath10 in the sb or @xmath190 erg @xmath1 s@xmath10 in the hb are removed from the stacking , and the background intensity produced by such bright sources was calculated using the number counts of x - ray sources @xcite .
four sources in the sb and two in the hb are removed this way ; the final stacking samples include 385 sources in the 0.51 and 12 kev bands , and 387 sources in the three @xmath191 kev bands .
for the optical galaxies , we chose @xmath45-band sources in the goods - s _ hst _ version r2.0z catalog @xcite with a 5 @xmath17 limiting ab magnitude of @xmath192 .
we stacked the x - ray undetected optical galaxies in the central 6-radius region .
galaxies within twice the 90% encircled - energy aperture radius of any known x - ray source are also removed from the stacking to avoid x - ray source contamination .
there are 18272 optical sources included in the stacking .
note that this galaxy sample contains all of the 23 sources in the x - ray stacking isx sample .
the stacking results are shown in figure [ xrbfrac ] .
the 1 @xmath17 errors on the stacked fluxes were calculated following @xcite including measurement errors and a 3% _ chandra _ flux - calibration error . for the stacking of the x - ray sources
, there is an additional poisson error due to the limited number of sources below the bright - end flux cut @xcite .
taking into account both the x - ray source contribution and bright - end correction , the resolved xrb fractions are @xmath14480% in all the energy bands , indicating that the average photon index of the x - ray sources is @xmath193 .
this is consistent with the fact that the majority of the x - ray sources in the cdf - s are obscured agns ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . for the x - ray undetected optical galaxies ,
significant detections are found in all the bands except the 46 kev band , where there is only a 1 @xmath17 signal ( corresponding to a @xmath194 chance that the signal was created by poisson noise ) .
we thus calculated 3 @xmath17 upper limits on the stacked counts and resolved xrb fraction in this band . in the 68 kev band ,
the optical galaxies produced a 2.5 @xmath17 stacked signal ( @xmath13 chance of being generated by poisson noise ) , responsible for @xmath195 of the xrb .
significance ) . ] with all of the x - ray sources and galaxies considered , it appears that the xrb in the 68 kev band can be fully explained , though the uncertainties in the normalization and stacked signals are large ( the slightly higher stacked flux than the xrb flux in this band could also be caused by cosmic variance in the cdf - s ) . at 16 kev
, x - ray sources and galaxies can account for @xmath196@xmath197 of the xrb .
the remaining unresolved fraction is probably due to cosmic variance in the narrow cdf - s region ( e.g. , * ? ? ?
* ; * ? ? ?
it could also be partially contributed by extended x - ray sources ( e.g. , galaxy groups or clusters ) in the cdf - s . the 2.5 @xmath17 stacked signal from galaxies in the 68 kev band ( @xmath195 of the xrb ; we refer to this as the galaxy ct xrb ) , combined with the weak signal in the 46 kev band ( @xmath198 of the xrb ) , suggest an underlying population of heavily obscured agns among the x - ray undetected galaxies ; emission in the 46 kev band is more heavily absorbed than that in the 68 kev band . as a subsample of the optical galaxies ,
the 23 isx sources produce a stacked signal that is @xmath199 of the total xrb flux in the 46 kev band ; the stacked signal in the 68 kev band is weak ( 1.4 @xmath17 ) , and we estimated a 3 @xmath17 upper limit that is @xmath200 of the total xrb . therefore ,
in the 68 kev band , the isx sources contribute @xmath15 of the galaxy ct xrb .
it is expected that there are some additional heavily obscured agns not selected with the isx method among the optical galaxies ( see 4.1 and 4.3 ) .
the number density of such objects is of about the same order of magnitude as the isx sources .
based on the model in @xcite , we derived that heavily obscured agns below the 4 ms sensitivity limit are responsible for @xmath201 of the xrb in the 68 kev band ( we refer to this model prediction as the model ct xrb ) .
this model ct xrb is only @xmath164 of the galaxy ct xrb , due to possible cosmic variance in the and uncertainties in the model assumptions ( e.g. , assumptions about the spectral shape , ct agn number density , and obscured fraction ; see 9 of @xcite for discussion ) .
we also note that the @xcite model does not include low - luminosity ( @xmath202 erg s@xmath10 ) agns .
ir studies suggest that the agn fraction in low - mass galaxies may be significantly higher than previously reported using optical spectroscopy ( e.g. , * ? ? ?
the obscured agn fraction also increases as luminosity decreases ( e.g. , * ? ? ?
it is thus probable that low - luminosity heavily obscured agns have a significant stacked contribution to the galaxy ct xrb .
these agns are faint in the x - ray band and challenging to identify even in deep _
chandra _ observations ( e.g. , * ? ? ?
their ir luminosities are more likely to be dominated by host - galaxy emission and thus are difficult to detect with the isx method ( or any ir - based method ) .
we extracted three additional x - ray stacking samples with @xmath167 to test whether there are agns among these sources .
we further break the isx sample into two subsamples to explore how the stacking results depend on the @xmath94 threshold value .
the threshold cuts in @xmath94 for these samples ( samples a1a5 ) are listed in table 2 , and they were chosen so that each sample has a similar number ( @xmath46 ) of sources .
we performed x - ray stacking analyses on the two subsamples and three additional samples following the same procedure as that for the isx and isn samples .
the results are presented in table 2 . due to limited sample sizes ,
these samples have less - significant detections in the hb than the isx sample .
however , the band ratios and effective photon indices for samples all suggest heavily obscured agn contributions to the stacked x - ray signals .
a plot of the effective photon index as a function of the average @xmath94 is shown in figure 7 ; in general , the larger the @xmath94 threshold value ( more ir excess ) , the harder the stacked signal .
similar behavior has also been observed by @xcite , and it is likely due to less contamination from star - forming galaxies at larger threshold values ( see fig . 2 ) .
given the stacking results for these additional samples , it is not likely that the hard x - ray stacked signal of the isx sample was produced by coincidence . a potentially straightforward way to detect distant heavily obscured or ct
agns is via hard observations at @xmath46100 kev .
several future hard x - ray missions , such as _ nustar _ ( planned launch year 2012 ; @xcite ) and _ astro - h _ ( planned launch year 2014 ; @xcite ) , have as one science goal to detect hard x - ray emission from distant heavily obscured agns .
for the three typical hard x - ray bands of _ nustar _ , , 1030 kev , and 3060 kev , the expected sensitivity limits are @xmath203 , @xmath204 , and @xmath205 erg @xmath1 s@xmath10 for a 1 ms exposure , respectively .
these flux limits are about two orders of magnitude more sensitive than those of previous missions .
the sensitivities are expected to be comparable to those for _ nustar_. the angular resolutions ( half - power diameters ) of _ nustar _ and are expected to be @xmath206 and @xmath207 , respectively .
the _ nustar _
positional accuracy is expected to be @xmath208 for strong sources . to assess
if the agns in our isx samples at @xmath21 will be detectable , we utilized the properties for sources in the x - ray stacking isx sample from the best - fit simulations ( for the 74% agn fraction ) and the mytorus model to calculate their expected fluxes in the _ nustar _ bands . only a tiny fraction ( @xmath209 ) of the simulated sources in our 10000 simulations
have fluxes above the _ nustar _ sensitivity limit in its most sensitive band ( ) ; this corresponds to no detected sources expected within the @xmath156 heavily obscured agns in the stacking isx sample ( or one detectable source among all the 242 isx objects ) . the median simulated flux in the kev band ( @xmath210 erg @xmath1 s@xmath10 ) is about two orders of magnitude below the sensitivity limit .
this result is a natural consequence of the requirement that the sources are not individually detected in the 28 kev band of the 4 ms _ chandra _ exposure ; the extraordinarily high sensitivity of the 4 ms cdf - s places a tight constraint on the intrinsic luminosities of the isx sources and prevents them from being detected by _
nustar_. therefore , it is not likely that _ nustar _ or _ astro - h _ will detect any of the isx sources presented here .
however , they will probably detect some of the x - ray selected ct agn candidates ( e.g. , * ? ? ?
* ; * ? ? ?
* ) in hard x - rays ; such detections will be useful for a clear determination of the intrinsic spectral shape and power of these sources .
one other approach to identify heavily obscured or ct agns is via x - ray spectroscopy at relatively low energies ( @xmath187 kev ) complemented by multiwavelength data ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the x - ray emission of these objects is characterized by a flat continuum and often a strong rest - frame 6.4 kev iron k@xmath127 fluorescent line ( e.g. , * ? ? ?
* ; * ? ? ?
for the isx sources presented here , spectroscopic analyses are probably not feasible due to the small number of counts expected
. however , it is worth performing a further study of the selected ct agn candidates in @xcite , which were previously studied using only the 1 ms cdf - s data . with the current 4 ms cdf - s data and the 3 ms observations on the cdf - s
, we will be able to constrain better their nature ( e.g. , * ? ? ?
in the case of the cdf - s receiving further _ chandra _ exposure ( e.g. , 10 ms total ) , some ( @xmath182 ) of the heavily obscured agns in the isx sample could be detected in the hb , given the simulated properties and expected 10 ms sensitivity . given the above , the majority of the heavily obscured agns in the isx sample will remain undetected in the x - ray , and the few percent of the xrb at @xmath46100 kev produced by these agns can not be directly resolved in the near future .
we note that x - ray absorption variability appears common among local seyfert 2 galaxies ( e.g. , * ? ? ?
* ) , and a few sources have even exhibited ct to compton - thin transitions ( e.g. , * ? ? ?
* ; * ? ? ?
therefore , a fraction of the isx agns may be detectable in the future if they become less obscured due to absorption variability .
also , optical / mid - ir spectroscopy has been shown to be a promising technique for identifying distant heavily obscured agns , particularly when combined with sensitive x - ray constraints ( e.g. , * ? ? ?
* ; * ? ? ?
future optical / ir instruments such as the _ james webb space telescope _
( _ jwst _ ) , extremely large telescopes ( elts ) , and atacama large millimeter array ( alma ) will provide great opportunities for detecting the remaining sources .
we have identified a population of heavily obscured / ct agns at @xmath21 in the cdf - s and e - cdf - s utilizing the isx selection method .
the key points from this work are listed below : 1 .
we have improved the isx method from @xcite by deriving the dust extinction via sed fitting and studying sources at lower redshifts with no bias in their ir - luminosity calculation .
the isx method can be well applied to the x - ray undetected galaxies in the cdf - s region at @xmath21 , given the superb multiwavelength data available .
the parent sample of 1313 galaxies were defined from an initial sample of 5237 optically and 24 @xmath32 m selected sources in the e - cdf - s , with the requirements of selected redshift interval , excellent multiwavelength coverage , high stellar mass , and no x - ray detection ( 2 ) .
2 . we have identified 242 isx sources in the cdf - s and e - cdf - s at @xmath21 ; these sources tend to be hosted by evolved galaxies with high stellar mass and little dust ( 2.2 ) .
an x - ray stacking analysis of 23 of the objects in the central cdf - s region resulted in a very hard x - ray signal with an effective photon index of @xmath3 , indicating a significant contribution from obscured agns ( 3.1 ) .
we have performed monte carlo simulations to estimate the agn fraction in the isx sample and assess the intrinsic properties of the obscured agns .
we modeled the observed x - ray flux considering both the star - formation and agn contributions , and we utilized the mytorus model to treat the spectra of heavily obscured agns .
the requirement that the sources are not individually detected in the 4 ms cdf - s sets strong constraints on their intrinsic properties .
we infer that @xmath4 of the isx sources are obscured agns , within which @xmath5 are heavily obscured and @xmath6 are ct ( 3.2 ) .
the heavily obscured ( ct ) agns discovered in our isx sample have moderate intrinsic x - ray luminosities ; the interquartile range of the intrinsic 210 kev luminosities is erg s@xmath10 [ ( 15)@xmath146 erg s@xmath10 ] .
the space density of the heavily obscured ( ct ) agns is @xmath162 mpc@xmath12 [ @xmath11 mpc@xmath12 ]
. these heavily obscured ( ct ) agns are expected to contribute @xmath211 ( @xmath178 ) of the total xrb flux in the 1030 kev band . in the 68 kev band ,
the 23 isx sources provide @xmath212 of the xrb flux .
the x - ray undetected optical galaxies in the cdf - s produce a 2.5 @xmath17 stacked signal in the 68 kev band , accounting for @xmath195 of the xrb flux , which is about the entire unresolved xrb fraction .
the space density of the isx agns and their contribution to the xrb could be increased by a factor of @xmath2132 due to our conservative isx definition ( 4 ) .
these heavily obscured or ct agns will probably not be detected by hard x - ray observatories under development such as _ nustar _ or _ astro - h _ due to their moderate intrinsic x - ray luminosities and significant obscuration .
most of the hard x - ray sources that will be detected by these facilities are likely already detected in the 4 ms cdf - s given the extremely high sensitivity of _ chandra _ to point sources ( 4.4 ) .
the isx selection method presented in this paper can be applied to other survey fields with good optical - to - ir coverage , and it can be expanded to higher redshifts if the ir luminosity can be estimated reliably , e.g. , combining mips 24 @xmath32 m data with data from deep _ herschel _ surveys at 100 and 160 @xmath32 m ( e.g. , @xcite ) .
we will explore these possibilities in future work .
+ we acknowledge financial support from cxc grant sp1 - 12007a ( bl , wnb , yqx ) , cxc grant g09 - 0134a ( bl , wnb , yqx ) , nasa adp grant nnx10ac99 g ( wnb ) , the science and technology facilities council ( dma ) , and asi / inaf grant i/009/10/0 ( ac , cv , rg ) .
we are grateful to t. yaqoob and k. d. murphy for providing support for the mytorus model and making available the data for our desired half - opening angle .
we thank a. t. steffen for helpful discussions .
we also thank the referee for carefully reviewing the manuscript and providing helpful comments that improved this work .
+ & & & & & & & + 03 32 49.61 & @xmath21427 49 00.1 & 0.98 & 0.97 & 1.02 & 0.2 & 58.7 & 10.74 & 6.2 & 1.1 & 0.76 + 03 32 45.58 & @xmath21427 49 36.4 & 0.680 & @xmath2141.00 & @xmath2141.00 & 0.9 & 472.3 & 11.46 & 29.5 & 5.0 & 0.77 + & & & & & & & + + & & & & & & & + 03 32 50.38 &
@xmath21427 47 07.1 & 0.537 & @xmath2141.00 & @xmath2141.00&1.9 & 166.5 & 10.86 & 7.7 & 14.1&@xmath215 + 03 32 48.58 & @xmath21427 45 04.9 & 0.89 & 0.89&0.92 & 1.2 & 121.8 & 11.03 & 12.2 & 12.2&@xmath216 + & & & & & & & & & & + isx ( @xmath97 ) & 23 & 0.69 & 76.7 & @xmath217 & @xmath218 & @xmath219 & 6.0 & 4.9 & 3.9 & @xmath220&@xmath221&4.9 & 36.6&8.3 + isn ( @xmath98 ) & 58 & 0.77 & 193.0 & @xmath222 & @xmath133 & @xmath134 & 6.9 & 10.6 & 1.6 & @xmath112&@xmath223&7.7&7.3 & 4.2 + & & & & & & & & & & + & & & & & & & & & & + a1 ( @xmath224 ) & 14&0.68&46.9&@xmath225 & @xmath226 & @xmath227&4.0 & 3.5&2.6 & @xmath228 & @xmath229&4.2&28.4&6.3 + a2 ( @xmath230 ) & 9&0.71&29.8&@xmath231 & @xmath232 & @xmath233&4.5 & 3.5&3.2 & @xmath234 & @xmath235&5.9&50.7&12.2 + a3 ( @xmath236 ) & 10&0.76&33.3&@xmath237 & @xmath238 & @xmath239&3.1 & 3.2&1.8 & @xmath240 & @xmath241&5.6&22.7&7.1 + a4 ( @xmath242 ) & 11&0.69&37.7&@xmath243 & @xmath244 & @xmath245&5.1 & 5.9&2.4 & @xmath246 & @xmath247&8.8&28.1&7.4 + a5 ( @xmath248 ) & 13&0.74&43.7&@xmath249 & @xmath250 & @xmath251&4.0 & 5.6&1.3 & @xmath252 & @xmath253&8.1&12.7&5.0 + 0.52.0 & @xmath218 & @xmath254 + 2.08.0 & @xmath219 & @xmath255 + 0.51.0 & @xmath256 & @xmath257 + 1.02.0 & @xmath258 & @xmath259 + 2.04.0 & @xmath260 & @xmath261 + 4.08.0 & @xmath262 & @xmath263 + | heavily obscured ( @xmath0 @xmath1 ) active galactic nuclei ( agns ) not detected in even the deepest x - ray surveys are often considered to be comparably numerous to the unobscured and moderately obscured agns .
such sources are required to fit the cosmic x - ray background ( xrb ) emission in the 1030 kev band .
we identify a numerically significant population of heavily obscured agns at @xmath21 in the _ chandra _ deep ( cdf - s ) and extended _ chandra _ deep field - south by selecting 242 x - ray undetected objects with infrared - based star formation rates ( sfrs ) substantially higher ( a factor of 3.2 or more ) than their sfrs determined from the uv after correcting for dust extinction .
an stacking analysis of 23 candidates in the central cdf - s region using the 4 ms _ chandra _ data reveals a hard x - ray signal with an effective power - law photon index of @xmath3 , indicating a significant contribution from obscured agns .
based on monte carlo simulations , we conclude that @xmath4 of the selected galaxies host obscured agns , within which @xmath5 are heavily obscured and @xmath6 are compton - thick ( ct ; @xmath7 @xmath1 ) .
the heavily obscured objects in our sample are of moderate intrinsic x - ray luminosity [ @xmath8@xmath9 erg s@xmath10 in the 210 kev band ] .
the space density of the ct agns is @xmath11 mpc@xmath12 .
the @xmath21 ct objects studied here are expected to contribute @xmath13 of the total xrb flux in the 1030 kev band , and they account for @xmath1415% of the emission in this energy band expected from all ct agns according to population - synthesis models . in the 68 kev band ,
the stacked signal of the 23 heavily obscured candidates accounts for @xmath15 of the unresolved xrb flux , while the unresolved @xmath16 of the xrb in this band can probably be explained by a stacking analysis of the x - ray undetected optical galaxies in the cdf - s ( a 2.5 @xmath17 stacked signal ) .
we discuss prospects to identify such heavily obscured objects using future hard x - ray observatories . |
the observation of a non - vanishing reactor angle @xmath1 represents an important achievement towards a more detailed understanding of the lepton flavour sector of the standard model .
results on @xmath1 from reactor and accelerator experiments t2k @xcite , minos @xcite , double chooz @xcite , daya bay @xcite and reno @xcite have been considered in global fits to the pontecorvo - maki - nakawaga - sakata ( pmns ) matrix @xcite .
interestingly , there is at the moment a controversial indication for deviations from the maximal value of the atmospheric angle @xmath2 , that is advocated in refs .
@xcite , but not confirmed in refs .
@xcite .
the discovery of the non - vanishing reactor angle and the indication of a non - maximal atmospheric angle have a deep impact on flavour model building .
indeed , models based on discrete symmetries , that dominated the flavour scenario in the past years for their ability to describe in first approximation specific mixing patters with @xmath3 and @xmath4 @xcite , need now some adjustment .
a few strategies have been followed : introduction of additional parameters in preexisting minimal models ; implementation of features that allow next order corrections only in specific directions in the flavour space ; search for alternative mixing patterns or flavour symmetries that lead already in first approximation to @xmath5 and @xmath6 ( see for example the reviews in refs .
@xcite and references therein ) . in other words ,
the latest neutrino data can indeed be described in the context of discrete symmetries , but at the prize of fine - tunings and/or eccentric mechanisms .
this suggests to investigate approaches alternative to discrete flavour symmetries : from models based on continuous symmetries such as @xmath7 ( i.e. refs .
@xcite ) or the simplest @xmath0 ( e.g. , refs .
@xcite ) , to models where no symmetry at all is acting on the neutrino sector ( e.g. , refs .
the latter go under the name of _ anarchical _ models , which have been characterised as models in which the coupling constants and mass matrices are `` random '' numbers drawn from an invariant probability distribution .
it has been claimed that such matrices generically prefer large mixings @xcite and that the observed sizable deviation from a zero reactor angle seems to favour anarchical models when compared to other more symmetric constructions @xcite .
however , as discussed in ref .
@xcite for the case of neutrino mass matrices , how much a large value of a parameter is preferred can depend strongly on the definition of `` preferred '' and of `` large '' .
it has been suggested in ref .
@xcite that the performances of anarchical models in reproducing the 2012 neutrino data are worse than those of models constructed upon the @xmath0 flavour symmetry . the analysis in ref
@xcite is based on the fact that anarchy can be formulated in a @xmath0 context , giving no charges to the left - handed fields , but non - zero @xmath0 charges to the right - handed ones in order to describe the charged lepton mass hierarchy .
this indeed allows a consistent comparison between anarchical and hierarchical models .
the constructions with built - in hierarchies among the matrix elements considered in ref .
@xcite , that resulted to be favoured with respect to anarchical models , have been chosen due to particular phenomenological features that lead to a good description of the data . on the other hand ,
these models have not been shown to be necessarily the best ones available and other @xmath0 models could provide an even better description of the data .
the method of analysis used in ref . @xcite treated all mass matrix elements as random complex numbers with modulus of order one and the models were judged according to the _ success probability _ ,
i.e. , the fraction of the generated points which satisfied some experimentally motivated cuts .
this kind of analysis allows to naively estimate the relative success of one model with respect to the others , but it is not based on , nor motivated by , any established method of statistical analysis , and hence the results can not be given any proper statistical interpretation .
however , some aspects of that method are rather good approximations of some procedures having well - defined meanings within bayesian inference . in comparison ,
a standard @xmath8-analysis of these models is not possible since they can all fit the data equally well for any values of the interesting parameters .
the main objective of this work is to make a statistically principled analysis of the models considered in ref .
@xcite , to make a systematic search for even better models , and to make the appropriate generalisations , allowing for a meaningful comparison between anarchy and hierarchy in the form of @xmath0 models .
furthermore , data on the masses of the charged leptons will be included in the analysis .
this is important since this data gives constraints on the model parameters constraints that must be consistent with neutrino oscillation data which will results in a large impact on the final results . in the following ,
basics concepts of bayesian inference are briefly summarized in sec .
[ sect : bayesian ] .
section [ u1mod ] is devoted to the presentation and analysis of specific flavour models , where the lepton @xmath0 charges are chosen a priori due to particular features of the yukawa matrices . in sec .
[ sec : generalised_models ] the lepton charges are instead treated as free parameters in the bayesian analysis . we discuss future prospects to further distinguish among the considered models in sec .
[ sect : future ] and present the conclusions in sec . [ conclusions ] .
in the bayesian interpretation , probability is associated with the _ plausibility _ or _ credibility _ of a proposition .
perhaps the main goal of science is to infer which model or hypothesis best describes a certain set of collected data .
also , these models should preferably be `` simple '' or `` economical '' in some sense . if one accepts the bayesian interpretation of probability , a very powerful arsenal of inference tools become available . in a nutshell ,
the idea is to use the laws of probability to calculate the probabilities of different hypotheses or models , when conditioned on some known ( or presumed ) information .
if the collected data is denoted by @xmath9 and the set of considered hypotheses or models is @xmath10 , bayes theorem gives the plausibilities of each of the hypotheses after considering the data , the _ posterior probabilities _
, @xmath11 here , the _ evidence _ @xmath12 is the probability of the data , assuming the model @xmath13 to be true , while @xmath14 is the _ prior probability _ of @xmath13 , which is how plausible @xmath13 is before considering the data .
@xmath15 is the probability of the data without assuming any particular model .
one can then perform _ model comparison _ by calculating ratios of posterior probabilities , the _ posterior odds _
, of two models as @xmath16 in words , the posterior odds is given by the _
prior odds _
@xmath17 multiplied by the _ bayes factor _ @xmath18 , which quantifies how much better @xmath13 describes that data than @xmath19 .
the prior odds quantifies how much more plausible one model is than the other a priori , i.e. , without considering the data .
if there is no reason to favor one of the models over the other , the prior odds should be taken to equal unity ( in which case the posterior odds equals the bayes factor ) , but sometimes one must consider this point more carefully . if the model contains free parameters @xmath20 , the evidence is given by = ( |m ) & = & ( , |m)d^n + & = & ( | , m ) ( |m)d^n + & = & d^n .
[ eq : z ] here , the _ likelihood function _
@xmath21 is the probability ( density ) of the data @xmath9 , assuming parameter values @xmath20 and @xmath22 is the prior probability ( density ) , which should reflect how plausible different values of the parameters are , assuming the model to be correct .
it should always be normalized , i.e. , integrate to unity .
the assignment of priors are probably the most discussed and controversial part of bayesian inference .
this assignment is often far from trivial , but constitutes a very important part of any bayesian analysis .
one observes that the evidence is the average of the likelihood over the prior , and hence this method automatically implements a form of _ occam s razor _ , since in general a more predictive model with a smaller parameter space will have a larger evidence than a less predictive one , unless the latter can fit the data substantially better .
the probabilities of the different hypotheses give the complete posterior inference on the space of models , and these have a somewhat unique and meaningful interpretation on their own .
however , bayes factors , or rather posterior odds , are usually interpreted or `` translated '' into ordinary language using the so - called _ jeffreys scale _ , given in tab . [
tab : jeffreys ] ( `` @xmath23 '' denotes the natural logarithm ) .
this scale has been used in applications in cosmology and astro- and particle phsycis such as refs .
@xcite ( and refs .
@xcite in neutrino physics ) although slightly more aggressive scales have been used previously @xcite .
note that it is often the case that the evidence is quite dependent on the prior used , although the bayes factor will generally favour the correct model once `` enough '' data have been obtained .
.__the jeffreys scale , often used for the interpretation of bayes factors , odds , and model probabilities .
the posterior model probabilities for the preferred model are calculated by assuming only two competing hypotheses and equal prior probabilities . _ _ [ cols="^,^,^,^ " , ] from the lepton charges in tab . [ tab - models ] , the textures for the charged leptons @xmath24 and neutrino @xmath25 yukawa matrices are as follows : a : y_e&= ( ^3 & ^2 & 1 + ^3 & ^2 & 1 + ^3 & ^2 & 1 ) , y_= ( 1 & 1 & 1 + 1 & 1 & 1 + 1 & 1 & 1 ) , + a _ : y_e&= ( ^4 & ^3 & + ^3 & ^2 & 1 + ^3 & ^2 & 1 ) , y_= ( ^2 & & + & 1 & 1 + & 1 & 1 ) , + h : y_e&= ( ^7 & ^5 & ^2 + ^6 & ^4 & + ^5 & ^3 & 1 ) , y_= ( ^4 & ^3 & ^2 + ^3 & ^2 & + ^2 & & 1 ) , [ oldmodels ] a : y_e&= ( ^3 & & 1 + ^3 & & 1 + ^3 & & 1 ) , y_= ( 1 & 1 & 1 + 1 & 1 & 1 + 1 & 1 & 1 ) , + h : y_e&= ( ^10 & ^6 & ^2 + ^9 & ^5 & + ^8 & ^4 & 1 ) , y_= ( ^4 & ^3 & ^2 + ^3 & ^2 & + ^2 & & 1 ) .
[ newmodels ] in the spirit of @xmath0 models , the coefficients in front of @xmath26 are expected to be complex numbers with absolute values of @xmath27 and arbitrary phases . considering that @xmath25 is a symmetric matrix , the total number of parameters that should be consider in the analysis is @xmath28 , from the yukawa matrices , plus the unknown value of @xmath29 . in ref .
@xcite , the performances of the first three models , @xmath30 , @xmath31 and @xmath32 , were evaluated by considering the fraction of the corresponding parameter spaces which were consistent ( at a fixed confidence level ) with the experimental constraints , for fixed values of @xmath33 .
the main result of that analysis was that , using a uniform distribution for the @xmath27 coefficients in the interval @xmath34 $ ] or @xmath35 $ ] ( and phases with a uniform distribution in @xmath36 $ ] ) , @xmath32 was the best performing model for values of @xmath29 larger than about @xmath37 , while for smaller values @xmath31 had the best success rate .
at the same time , @xmath30 seemed disfavoured when compared to the previous models , for almost all the range of values of @xmath29 .
however , it is worth to mention that i ) the doublet charges of @xmath32 and @xmath31 were chosen in order to naturally reproduce the neutrino data ; ii ) these models had an intrinsic advantage over @xmath30 , as they have an extra parameter in the form of @xmath29 , while @xmath30 is insensitive to its value ; iii ) although the charges and the relevant values of @xmath29 considered were chosen using the observed charged lepton masses ( as well as quark masses and mixings as these models where formulated in a @xmath38 gut context ) , they were not subsequently used in the numerical analysis .
the charges of the remaining models , @xmath39 and @xmath40 , are chosen following the analysis in sec .
[ sec : generalised_models ] , where they are treated in general as free parameters .
in particular , the charges of @xmath39 are identified as the best ( i.e. having large posterior probabilities ) in the case of vanishing doublet charges , whereas those of @xmath40 are the best assuming all free charges .
they are rather similar to those of @xmath30 and @xmath32 , respectively . in the remaining part of this section
, we will consider the models listed in tab .
[ tab - models ] with the aim of : * analyse and compare the models using bayesian inference ; * check whether the results of ref .
@xcite remain valid when performing the full bayesian analysis ( assuming that the differences in the data sets used for our analysis and that in ref .
@xcite are irrelevant for the model comparison ) ; * determine the importance of including the charged lepton data . in order to calculate the evidence and obtain the posterior distributions ,
we need to specify priors on the @xmath41 free parameters of these models .
* it is reasonable to take @xmath29 as a priori independent of all the @xmath42 coefficients , phases , and charges , and then we use consistency of the lagrangian expansion to set an upper bound of @xmath43 .
we use a prior ( ) = , , [ priorepsilon ] which behaves uniformly in @xmath44 for @xmath45 and uniformly in @xmath33 for @xmath46 .
@xmath47 is the required normalisation factor .
we take @xmath48 as our default choice , but we will find that our results are quite insensitive to changes in @xmath49 , which is to be expected since it is a free parameter in all models we consider .
* for the @xmath27 parameters , we also make the reasonable assumption that they are a priori independent of the charges , and so their priors should be the same for the cases of vanishing and non - zero charges .
this translates into a _ a priori _ invariance under basis rotations and leads to a unique measure on the leptonic mixing matrix , the so - called haar measure .
this has been studied in some detail in refs .
@xcite , and its interpretation has often been that it describes how matrix elements or mixing angles are `` randomly distributed '' in some sense .
however , these distributions are more naturally considered in a bayesian context , and the use of haar measures to construct prior distributions has been extensively studied in the statistics literature ( see ref .
@xcite and references therein ) .
+ in addition to the mixing angles , there still remains an arbitrary measure over the neutrino mass eigenvalues . as pointed out in ref .
@xcite , under the additional assumption that the matrix elements are independent of each other , the measure on the mass matrix becomes unique ( up to a scale ) , ( m_ij ) e^-(m m^)/2 , and so the real and imaginary parts of each element are a priori independent with gaussian priors .
note that , since the mass matrix is symmetric , the off - diagonal elements are on average a factor @xmath50 smaller than the diagonal elements , but this will have a negligible impact on the results .
equivalently , the prior on the absolute value @xmath51 and phase @xmath52 of each ( off - diagonal ) element is ( q , ) = , [ eq : mag_prior ] with mode at @xmath53 . + by analogy , we take the same prior for the elements of the charged lepton mass matrix , although , since it is not symmetric , the elements have the same widths of their priors .
we consider the following relevant data : * neutrino oscillation data constrain the parameters @xmath54 , @xmath55 , @xmath56 , and @xmath57 ( using @xmath58 for @xmath59 ) .
although there are some constraints on the cp - violating phase @xmath60 , we will not take this into account as induced priors on @xmath60 are rather similar and independent of @xmath29 in all models , and this will have negligible impact on the results .
the oscillation parameters are rather well constrained and the correlations between the oscillation in the standard parameterization are rather small , and so we can approximate the oscillation likelihood as _ osc ( ) ^1(r)^2(s_12 ^ 2)^3(s_23 ^ 2)^4(s_13 ^ 2 ) .
we take the individual likelihood components as gaussian functions using the results of ref .
although this might not be a perfect approximation , this should not have a noticeable impact on any results . *
the ratio of the charged lepton masses , @xmath61 and @xmath62 have been measured with very good accuracy , and so we can approximate the associated likelihoods with dirac @xmath60-functions . to allow convergence of our numerical analysis we will in turn approximate these likelihoods using rather broad gaussians .
this will not change any conclusions as long as the widths of these gaussians are taken small enough ( so that all priors are effectively constant over these widths ) .
we take our errors to be about @xmath63 , and have checked numerically that our results are insensitive to changes in these widths . * in what follows , only the normal mass ordering for the neutrino spectrum will be considered , as this is much strongly preferred than the inverted one .
this is a well known result and already pointed out in ref .
@xcite : the inverted mass ordering is typically linked to a maximal value of the solar angle , in contrast with the observations .
we first focus on the posterior distributions in all the models .
however , since @xmath29 is really the only parameter of interest , and its marginal posteriors are at least not very far from being gaussian , we only consider the posterior means and standard deviations ( which closely matches the posterior medians and 68% credible intervals ) .
the inclusion of the charged lepton data is particularly important since , by themselves , they can give strong constraints on @xmath33 .
consequently , the same preferred values of @xmath33 must be used when fitting the neutrino data .
in general , we expect naive estimates as [ eq : naive ] ( ) ^1/(n^_1+n^_1 ) ( ) ^1/(n^_2+n^_2 ) , where @xmath64 are the @xmath42 factors entering the yukawa matrix as defined in eq . .
if each of these @xmath42 factors is identified with the absolute value of the matrix element in eq . , implying a prior uncertainty of @xmath65 , then one can obtain naive analytical estimates of @xmath44 and its uncertainty , which can be compared with the numerical results .
note that each of the two @xmath42 factors gives a contribution to the uncertainty of @xmath29 .
figure [ fig : epscons ] shows these naive estimates ( in green - circles and black - diamonds , respectively ) together with the numerical results when using only neutrino oscillation data ( in red - crosses ) and finally when also charged lepton data are considered ( in blue - squares ) .
the error bars are given by twice the naive estimates and the posterior standard deviations , respectively .
in the anarchical models there are , of course , no constraints on @xmath33 from neutrino data .
the two naive estimates are relatively consistent with each other and with the ones using neutrino data in all the models .
the largest tension can be found in the model @xmath30 between the electron and muon estimates , but it is still naively smaller than 3@xmath66 . within this model , these discrepant data can still be consistently combined , but this tension will work against the model @xmath30 when it is compared to other models . ] furthermore , the constraints using all the data is generally consistent with being the naive combination of the three other sets of data ( i.e. when treated as independent measurements ) , for both the best estimates and the size of the errors . for some of the models
the combined uncertainty is somewhat smaller than the naive expectation , and for the model @xmath32 the posterior mean is smaller than all the three partial estimates .
however , the naive combination is expected to be valid only when all the individual constraints are gaussian and there are no common nuisance parameters . in the present case , we do see some non - guassian features of the posteriors , and there are 30 common nuisance parameters whose correlations can invalidate the naive combination . finally , we note that no constraints on @xmath29 can be obtained in a standard @xmath8-analysis in which the @xmath8 is minimized ( or likelihood maximized ) over the remaining parameters .
all the models can fit the data equally well , i.e. , perfectly , for any non - zero values of @xmath29 .
however , depending on the value of @xmath29 , the fit would require more or less fine - tuning among the @xmath42 parameters .
this fine - tuning is automatically considered in the bayesian analysis and is what yields the above constraints on @xmath29 . in the different models , with the error bars being twice the standard uncertainties .
see text for the description of the naive estimates . for the numerical estimates ( red and blue ) , the points are the posterior means and the uncertainties are the posterior standard deviations.__,scaledwidth=48.0% ] all the models have a @xmath8-minimum of zero and therefore a @xmath8-analysis can never `` exclude '' any of the models , and moreover two models can not be meaningfully compared .
the different models require more or less fine - tuning among the @xmath42 parameters , and here also of @xmath29 , and this aspect is automatically considered in the bayesian analysis discussed here below .
we note that if @xmath67 denotes the model with fixed charges @xmath68 , then @xmath69 if there is no reason that a particular set of charges is a priori more plausible than any other , one sets @xmath70 and uses the jeffreys scale to interpret the strength of evidence .
the logarithms of the evidences of the models in tab .
[ tab - models ] normalised to the evidence of @xmath39 , i.e. , the bayes factor between all the models and @xmath39 , are reported in fig .
[ fig : bfactors ] , and this allows a quantitative comparison of all those models .
the numerical ( statistical ) uncertainty on the individual log - evidence estimates , as reported by , is about @xmath71 for all the models .
hence , the uncertainty on the logarithms of the bayes factors are about @xmath72 . for the models in tab .
[ tab - models ] using only neutrino data ( dark - red bars ) and all data ( light - blue bars ) .
the dashed lines correspond to the boundaries given in tab .
[ tab : jeffreys ] in the comparison with @xmath39 .
_ , scaledwidth=48.0% ] we first focus on the models @xmath30 , @xmath31 and @xmath32 that have been previously discussed in ref .
@xcite . when using only the neutrino data , we see from fig . [
fig : bfactors ] that the hierarchical models are all weakly preferred compared to the anarchical ones .
these results are in line with the ones quoted in ref .
@xcite , in the sense that the anarchical model is less appropriate to describe the data . however , this depends somewhat on the specific choice for the prior on @xmath33 : while the evidence of @xmath30 are independent on the value of @xmath73 , those of the hierarchical models are not .
taking , say , @xmath74 as small as @xmath75 would reduce the @xmath76 s of @xmath31 and @xmath32 by @xmath77 units .
this can be considered the `` punishment '' that the hierarchical models receive for their advantage of having an additional parameter , as already discussed in sec .
[ sec : specificu1 ] .
when also the charged lepton data is included , we notice that @xmath30 is strongly disfavoured compared to the other models since it predicts the charged lepton data rather badly . considering in the comparison also the @xmath40 and @xmath39 models ,
@xmath40 is the best model : it is moderately better than @xmath31 and @xmath39 , and weakly preferred over @xmath32 .
the different evidences of @xmath30 and @xmath39 when only neutrino data are taken into account , are due to statistical fluctuations of the evidence estimates , but the difference is consistent with being zero within the uncertainties .
the charged lepton data has again a deep impact on the analysis : once considered , @xmath29 is well - constrained in all models , which implies that there is basically no sensitivity to the volume of the prior any more , although a small dependence on the shape of the prior could still remain .
for example , when the prior becomes effectively uniform ( taking @xmath74 defined in eq .
( [ priorepsilon ] ) very large ) , models which subsequently prefer large values of @xmath33 are favoured by roughly the log of the ratio of the preferred values of @xmath33 s . in this case , the evidences for @xmath40 and @xmath32 get stronger than the values shown in fig .
[ fig : bfactors ] : h&b 4.5 , + h&b 2.5 .
the models @xmath30 , @xmath31 and @xmath32 discussed in the previous section were taken as in ref .
@xcite were the charges have been motivated by the data .
however , one could argue that in some sense the data are used twice - first to choose the charges and then to analyse the models . on the other hand ,
the charges can be thought of as just a set of additional unknown free parameters , which should preferably not be fixed from the beginning , but instead inferred in the bayesian analysis . in this section , this latter strategy will be followed to study parameter constraints and make a more general comparison of anarchy vs. hierarchy . beside the @xmath41 free parameters considered
when the charges are fixed , the most general model with free charges has four additional discrete parameters , which we can assign priors and subsequently calculate the posteriors of in the usual way .
we consider the following generalised models : @xmath78 : : : generalised anarchy .
the doublets have no charges and only the @xmath79 singlet charges are non - vanishing and are taken as free parameters . @xmath80 : : : generalised hierarchy .
this is the most general model where all the charges are free .
potentially , this model can give the best predictions in both the charged lepton and neutrino sectors .
on the other hand , it has a larger parameter space and hence more regions in which it could fail to predict the experimental data .
hence , this model could as a whole have small predictive power .
we include only non - vanishing doublet charges , so that @xmath78 is not contained within setup .
this will have negligible impact on our results ( except of this specific region in parameter space ) .
in other words , @xmath78 is the union of all models with anarchy in the neutrino sector , while @xmath80 is the union of all models with hierarchy . in order to perform a bayesian analysis with free charges
, one needs to assign them priors .
the doublet charges should be independent on the singlet charges , but the charges of the same field must be dependent , since it holds that @xmath81 . a general prior
can be written as ( n_1,n_2 ) = ( n_1|n_2)(n_2 ) .
the charges are integers , and a naturalness criterion can be introduce such that the preference falls on the set with smaller charges : very large charges should not be equally plausible as small ones , but at the same time it is wise to not assign a prior of exactly zero to a particular charge .
it seems reasonable to expect @xmath82 , with @xmath83 .
then the unique distribution on the non - negative reals which has maximum entropy ( `` minimum information '' , see , e.g. , refs .
@xcite ) satisfying this constraint is the geometric distribution , @xmath84 with @xmath85 .
if the same condition is imposed on @xmath86 , @xmath87 is expected to be @xmath88 larger than @xmath89 and then the prior on both charges simply becomes ( n_1,n_2 ) = ( 1-p)^n_1p^2 .
notice that this prior only depends on the value of the largest charge and the possible dependence of the results on the value of @xmath88 will be discussed in the next sections . on the other side , a uniform prior @xmath90 for charges with @xmath91
will also be considered .
it is possible now to analyse the models with free charges and compare them using the bayesian evidence .
the posteriors of the parameters of @xmath78 and @xmath80 models is shown in figs .
[ fig : modav_an ] and [ fig : modav_hi ] , respectively . in the model
@xmath92 using the full data set . in the two - dimensional plots , the color scale denotes the base-10 logarithm of the posterior probability of each region.__,scaledwidth=50.0% ] in fig . [ fig : modav_an ] the posteriors of the singlet charges and @xmath29 in the generalised anarchical model is plotted using priors with @xmath93 for the mean of the charges and @xmath94 for the parameter in the prior on @xmath29 . from the three two - dimensional posteriors of the singlet charges and @xmath29 ,
one notes the strong degeneracy between the parameters , which follows from eq .
( [ eq : naive ] ) and the fact that the neutrino data is insensitive to the values of @xmath29 and the charges .
( [ eq : naive ] ) translates into @xmath95 and @xmath96 .
the width of these regions are determined by the widths of the @xmath42 factors , and the uncertainty of @xmath33 is as in sec .
[ sec : spec_epsconstr ] .
one naively expects an exact one - dimensional degeneracy ; however , this is broken by the assignment of a prior which gives different posteriors to different points along that curve .
these are the prior on the charges and on @xmath33 , in particular its upper limit which yields upper limits also on the charges . also
the widths of the @xmath42 elements affects the viability of any point along the degeneracy differently , since points which require less fine - tuning of the @xmath42 elements are inherently favoured .
no really strong constraints on any single parameter can be extracted though , although with a high probability @xmath97 $ ] , @xmath98 $ ] and @xmath99 . in the posterior of @xmath29
there are multiple peaks , each of which is generally dominated by a single pair of charges .
the first three peaks correspond to charges @xmath100 , and @xmath101 , respectively , while the remaining ones which can be seen are dominated for pairs of charges with @xmath102 . + in fig . [
fig : modav_hi ] the posteriors of all the charges and @xmath29 in the generalised hierarchical model is plotted , also using @xmath93 and @xmath94 . once again , @xmath103 is excluded by the definition of the model .
first , the three two - dimensional posteriors of the singlet charges and @xmath29 show that the degeneracies have changed shape and position , and gotten wider with respect to the same plots in fig .
[ fig : modav_an ] .
these are of course just what is expected from eq .
( [ eq : naive ] ) when the doublet charges are allowed to be non - zero .
in particular , the extra suppression from the doublet charges requires @xmath29 to be larger in @xmath80 than @xmath78 for fixed singlet charges , while the ratio @xmath104 is smaller , around 2 .
most of the posterior probability is assigned to @xmath105 and @xmath106 .
two of these combinations were considered in the sec .
[ u1mod ] .
we note that in general @xmath29 prefers to be larger in @xmath80 than in @xmath78 , and also the appearance of two peaks in its posterior ( on top of a large `` background '' ) . from the plots of @xmath29 vs. @xmath107 and @xmath108 one sees that the peak at around @xmath109 comes mainly from @xmath110 , while the peak at around @xmath111 is associated with @xmath112 and @xmath106 .
this is consistent with the fact that @xmath110 and @xmath106 were the doublet charges of the @xmath113 and @xmath114 models in sec .
[ sec : specificu1 ] , and that these different values of @xmath33 were found to be preferred in fig .
[ fig : epscons ] .
finally , @xmath29 has a quite strict lower bound , since for even the smallest possible choice of the doublet charges , a very small @xmath29 will give a too large hierarchy in the neutrino sector .
furthermore , the charges for the singlets must typically be larger in @xmath115 than in @xmath116 .
this mainly follows from the fact that the doublet charges together with @xmath29 are essentially used to fit four observables ( the three mixing angles and the ratio @xmath117 ) , while the singlet charges are used to fit only two ( the mass hierarchies of the charged leptons ) .
then , the non - zero doublet charges require @xmath29 to not be very small so that only a mild hierarchy in the neutrino sector is achieved . therefore , in order to fit the charged lepton data , larger singlet charges are preferred in @xmath115 than in @xmath116 . similarly as in sec . [
sec : specificu1 ] , no parameter constraint can be extracted and no model comparison can be performed from a @xmath8 analysis . indeed for any given value of the charges , and indeed for any fixed tuple @xmath118 , the data can always be fitted perfectly . the comparison among @xmath80 and @xmath78
is then performed on posteriors and evidences . the evidence of a model @xmath119 with free charges @xmath120 is given by ( |m ) = _ c ( |c , m)(n|m ) = _ c _ c ( c ) and so the full evidence is the average of the evidences with fixed charges over the prior on the charges .
hence , a model with a large number of unspecified charges can have a small evidence if a large portion of those charges does not predict the data accurately , even if there is some specific combination of charges which predict the data well . for our default choice of priors with @xmath93 and @xmath94
, the comparison of @xmath80 and @xmath78 yields a bayes factor = ( _ _ g/__g ) = 1.7 , which means that the hierarchical model is weakly preferred compared to the anarchical one . the uncertainty on the value reported in the previous equation , as on all the bayes factors in this section , is about @xmath72 .
bayes factors can often depend crucially on the priors employed , in particular when the parameter in question only appears in one of the models .
since the doublet charges are free parameters only in @xmath121 , it is then important to check whether the results discussed above are stable under different assumptions on the priors .
usually , the prior on the additional parameters are taken as unrelated to the common parameters , while in the present case the singlet and doublet charges have the same priors , which could result in a smaller prior dependence . as a first possibility
, the priors on the charges have been taken as the geometric distribution ( as used in the default case ) with the expected value @xmath88 of the smallest charges varied between @xmath122 and @xmath123 .
a second possibility considered is of a uniform prior on the charges up to a maximum of @xmath124 between @xmath125 and @xmath126 : this implies that the marginal prior on the largest charge is proportional to its value ( up to @xmath124 ) , and so implicitly larger charges are a priori favoured .
for the prior on @xmath29 , we have varied the value of @xmath74 , that in the previous section was fixed at @xmath127 . since @xmath29 is a free parameter of all models , the impact on the evidence is in general expected to be small . as long as @xmath74 is smaller than the smallest value preferred by any model
, it should not effect neither the posteriors nor the bayes factors , and in fact all posterior inferences should have a unique limit as @xmath128 ( this also applies to the specific models in sec .
[ sec : specificu1 ] ) .
@xmath94 is already small enough for this limit to be well approximated ( we have also checked this numerically ) , and so we do not show any results for smaller @xmath74 . as @xmath129 , @xmath130 becomes a uniform distribution .
we find the following .
* geometric charge priors * + for small @xmath74 the evidence for @xmath78 does not really depend on @xmath88 , which is expected since in this case charges of all magnitudes are roughly equally good ( along the degeneracy ) . however , for @xmath74 with a uniform prior ( i.e. large @xmath74 ) , larger @xmath88 gives larger evidence : indeed large @xmath33 , and so larger charges , are preferred . for @xmath80 ,
the effect of varying @xmath88 is in most cases very small , regardless of the prior on @xmath29 .
if only the prior on the doublet charges is made wider , the evidence decreases ( since small charges predict the data better ) , and vice versa .
if only the prior on the singlet charges is made wider , the evidence instead increases , and vice versa . since the priors are the same on both sets of charges , these effect partially cancel , leaving only small changes in the evidence .
since large @xmath29 is preferred in @xmath80 , the uniform prior gives a larger evidence .
when the bayes factor is calculated , many of the changes in the evidences of the models tend to cancel , giving : & = 1.7 1.5 (: 1 3 , _ 0
< 10 ^ -2 ) + & = 2.3 2.0 (: 1 3 , _ 0 = ) . hence , in total the model comparison is very stable against changes in the priors with weak to `` almost moderate '' preference of @xmath80 over @xmath78 . *
uniform charge priors * + for the uniform prior on the charges , much more prior is put on large values of the charges , especially for large @xmath124 .
hence , one expects @xmath78 to be relatively unaffected by this modification , while the small doublet charges required in @xmath80 will work against it . indeed ,
as expected , we find : & = 1.3 0.8 ( n_max : 7 15 , _ 0
< 10 ^ -2 ) + & = 1.8 1.0 ( n_max : 15 , _ 0 = ) . a uniform prior in the interval @xmath131
$ ] on the @xmath27 parameters has also been considered and it turns out that the difference compared to the standard ( gaussian ) case is within the numerical uncertainties .
in this section , we comment on the ability of future low - energy experiments to distinguish between our considered models .
if a new data set @xmath132 is added to the current one , @xmath133 , the resulting bayes factor between two models becomes _ f = = , i.e. , the new bayes factor equals the present one multiplied by @xmath134 , with evidences given by ( _ f | , m ) = ( _ f | , m ) ( | , m ) d.hence , the change in the bayes factor by the addition of a new set of data is given by calculating the average of the future likelihood over the present - day posterior .
as an ideal situation , consider the case where the new data determines some combinations of parameters exactly ( for example , some set of low - energy observables ) , so that @xmath135
. then = , and therefore if a perfect measurement of a single observable is to be able to increase the evidence of , say , 5 log units , then the ratio of the current posteriors at the true value must differ by a factor of @xmath136 . regarding the data adopted here , all the functions of parameters used to constrain the models are well - constrained , so that the posteriors of those parameters follow the experimental likelihoods closely in all the considered models .
hence , improved measurements of those parameters can not further discriminate between the models .
, the lightest neutrino mass @xmath137 , the effective @xmath138-decay effective mass @xmath139 , the @xmath140-decay mass @xmath141 , and the sum of the neutrino masses @xmath142 , for @xmath39 ( black - continuous line ) and @xmath40 ( red - dashed line).__,scaledwidth=50.0% ] there are other observables which could be accurately measured in future experiments , and in principle could be used to distinguish between the models .
these are primarily the cp - phase @xmath60 and observables related to the values of neutrino masses ( @xmath139 , @xmath141 , @xmath142 ) .
we plot the posteriors of these variables as well as the lightest neutrino mass @xmath137 in fig .
[ fig : obs_1d ] for the model @xmath39 and @xmath40 .
similar posteriors are expected for the other hierarchical and anarchical models .
correlations among ( @xmath139 , @xmath141 , @xmath142 ) and @xmath137 can be read in the two - dimensional posterior plots , displayed in fig .
[ fig : obs_2d ] .
a precise measurement of @xmath60 can only give a very minor further discrimination between the models .
on the other hand , a precise measurement of the sum of neutrino masses @xmath142 at about @xmath143 could give in principle strongly increased support for anarchical models ; similar arguments hold for the other variables .
contrary , very stringent upper limits on the different observables could give support to the hierarchical models .
however , the practical feasibility of these measurements are not very good in the near future . versus @xmath139 , @xmath141 , and @xmath142 .
the contours represent the @xmath144 and @xmath145 confidence level regions for @xmath39 ( black ) and @xmath40 ( red).__,scaledwidth=50.0% ] as discussed earlier , only the normal mass ordering for the neutrino spectrum has been considered here since it is rather strongly preferred in both the anarchical and hierarchical models .
however , if definite evidence for the inverted mass ordering would emerge , all the models considered here would be disfavoured in comparison with their inverted - ordering counter parts .
after that , the comparison would be between inverted hierarchical and inverted anarchical models .
in this letter , a conclusive comparison among anarchical and more symmetric approaches to explain the flavour puzzle in the leptonic sector is presented . the method used
is based on bayesian inference and has been applied to a series of flavour models whose effective lagrangian shows an invariance under the abelian group @xmath0 .
two distinct setups have been considered : i ) in the first one , the @xmath0-charges of the sm fields and of the additional scalar field responsible for the flavour symmetry breaking mechanism are fixed to specific values , determined through phenomenological considerations or by a statistical analysis ; ii ) in the second one , the @xmath0-charges , or part of them , are kept as free parameters and the statistical procedure determined those ones , for which the model best reproduces the data , limiting fine - tunings .
case i ) corresponds of giving to the anarchical models @xmath30 and @xmath39 the same prior probability with respect to constructions with built - in hierarchies among the matrix elements , @xmath31 , @xmath32 and @xmath40 . on the other hand , in case ii ) the comparison is among the class of models embedding the anarchical ansatz @xmath116 and that of models based on a more symmetric principle @xmath115 . in both cases , models with hierarchical matrix elements
are preferred over the anarchical ones and the only difference among the two approaches is the precise values of the evidence : almost moderate in case i ) , and weak in case ii ) .
this study confirms and extends the previous results in ref .
@xcite .
the stability of the results have been checked modifying the most sensible aspect of bayesian inference , i.e. the priors on the parameters of the models .
the prior of @xmath29 has been essentially taken to be either log - uniform or uniform .
the priors of the @xmath42 parameters entering the yukawa matrices follow either a distribution constructed from the haar measure or a uniform distribution in the interval @xmath131 $ ] .
finally , the priors of the charges , when taken as free parameters , have chosen with either a geometric or a uniform distribution in a given support .
the results show a slight dependence on the choice of priors , that however do not significantly change the conclusions .
improvements on the precision of the used data or the addition of new data related to the neutrino mass spectrum and mixing matrix , such as @xmath146 , @xmath141 , @xmath142 and @xmath60 , will most likely a very minor effects on the results presented here . on the other hand , understanding the neutrino spectrum ordering could have an impact on the analysis : here the focus was only on the normal ordering , as the inverted one is typically linked to a maximal value of the solar angle , in contrast with the observations . this work shows the power of the bayesian inference in comparing models and extracting information on the model parameters .
the only constructions considered here are based on the abelian @xmath0 flavour symmetry and with neutrino masses described by the effective weinberg operator . extending the analysis to the case where a see - saw mechanism explains
the lightness of the active neutrino masses is straightforward . on the other hand , to include in the comparison models based on different symmetries than @xmath0 , such us non - abelian continuous or discrete ones , is a non trivial task and is left for future studies .
we thank mattias blennow , enrique fernandez martinez and tommy ohlsson for participating at the initial phase of this project .
j.b . and l.m .
acknowledge partial support by european union fp7 itn invisibles ( marie curie actions , pitn - ga-2011 - 289442 ) .
d.m . acknowledges miur ( italy ) for financial support under the program futuro in ricerca 2010 ( rbfr10o36o ) .
l.m . acknowledges partial support by the juan de la cierva programme ( jci-2011 - 09244 ) and by the spanish minecos `` centro de excelencia severo ochoa '' programme under grant sev-2012 - 0249 .
thanks the gran gustafsson foundation for financial support and kth royal institute of technology , the dipartimento di matematica e fisica delluniversit roma tre and the theoretical group at cern for hospitality during the development of this project . | whether the anarchical ansatz or more symmetric structures best describe the neutrino parameters is a long standing question that underwent a revival of interest after the discovery of a non - vanishing reactor angle and the indication of a non - maximal atmospheric angle . in this letter , a bayesian statistical approach is adopted in order to analyse and compare the two hypotheses within the context of @xmath0 flavour models .
we study the constraints on individual model parameters and perform model comparison : the results elect constructions with built - in hierarchies among the matrix elements as preferred over the anarchical ones , with values of the evidence that depends slightly on whether the @xmath0 charges are also considered as free parameters or not , and on the priors used . |
the excess free energy of a crystal in contact with a wall is an important quantity governing many interfacial phenomena such as wetting and heterogeneous nucleation @xcite .
theoretical approaches based on cell theory have been used to compute this quantity for model systems such as hard spheres @xcite .
atomistic simulation studies have also been carried out employing widely - used interaction potentials such as lennard - jones ( lj ) models and systems of hard spheres @xcite . in most previous studies ,
the crystal is either in contact with a flat structureless wall @xcite or a structured wall consisting of particles attached to sites corresponding to an ideal crystalline structure @xcite . in both of these situations , the structure and energy of the crystalline layers in the neighborhood of the wall is not the same as in the bulk leading to an excess free energy as compared to the bulk . an interesting situation arises when the energy and structure of the crystalline layers in contact with the wall remains the same as in the bulk .
such a scenario occurs when the wall consists of particles frozen into positions occupied by the same crystalline phase at equilibrium . in this case
, the wall is a frozen version of the same crystal .
then , it is of interest from a fundamental statistical mechanical point of view , to determine the excess free energy of the system in comparison to the bulk .
an analogous situation corresponding to a supercooled liquid in contact with a wall consisting of particles frozen into the same amorphous structure as the liquid has been widely studied in recent years in order to investigate the relaxation dynamics of glass - forming systems @xcite .
recently , an experimental investigation of such a system was carried out using colloidal particles @xcite .
such computational and experimental studies extract a growing length scale based on a slowing down of the dynamics near the wall and relate it to the slowing down of glassy dynamics in the bulk . in these investigations
, it is implicitly assumed that the thermodynamics of the system is unperturbed by the wall if an average is carried out over the thermal fluctuations and different realizations of the wall @xcite .
however , in a recent work we obtained a non - zero interfacial free energy for a glass - forming binary lennard - jones liquid @xcite in contact with a wall with the same amorphous structure as the supercooled liquid , indicating that thermodynamics of the liquid is affected by the presence of such a frozen wall @xcite . in this work
, we compute the excess free energy of a crystal in contact with a wall formed by its frozen version , using molecular dynamics simulations @xcite in conjunction with thermodynamic integration @xcite .
we obtained a non - zero interfacial free energy for particles interacting via a lennard - jones potential @xcite , which increases with increasing density of the crystal suggesting that the thermodynamics of the crystal in contact with such a wall is perturbed and not the same as in the bulk .
moreover , at low densities @xmath0 is negative , indicating an increase of entropy as compared to the bulk . to test the robustness of our results and find out
whether a negative value of @xmath0 is model - dependent or not , we carried out investigations with two other interaction potentials : a purely repulsive inverse - twelfth power soft - sphere potential and the hard sphere interaction ( using a @xmath1 potential ) @xcite .
however , @xmath0 for the two purely repulsive interactions , while non - zero , is positive at all crystal densities . in the next section
we specify the interaction potentials considered in this work . in section [
sect : system_setup ] , the system is described and then we outline the thermodynamic integration scheme adopted to compute @xmath0 . in section [
sect : simulation ] , we describe the simulation details and then present our results in section [ sect : results ] .
finally , we end with a conclusion .
in order to study the robustness of our findings , we obtained @xmath0 corresponding to three different interaction potentials .
the interaction potentials are denoted by @xmath2 [ @xmath3 , @xmath4 being the distance between the particles .
energies and lengths are expressed in units of the parameters @xmath5 and @xmath6 , respectively ( see below ) , and the masses of the particles are set to @xmath7 . in the following , all other quantities are given in terms of the latter parameters , i.e. time @xmath8 in units of @xmath9 , pressure @xmath10 in units of @xmath11 , and the interfacial free energy in units of @xmath12 .
the following interaction potentials are considered in this work : \(i ) a modified lennard - jones interaction potential ( mlj ) is chosen , defined by @xmath13 with @xmath14 , \label{eq : lj}\ ] ] for @xmath15 , and @xmath16 for @xmath17 and @xmath18 for @xmath19 .
the constants in eqs .
( [ eq : ur1 ] ) and ( [ eq : ur2 ] ) are given by @xmath20 , @xmath21 , @xmath22 , @xmath23 , and @xmath24 .
\(ii ) the second potential we consider is the purely repulsive force - shifted inverse - twelfth power potential ( fsi12 ) , @xmath25 where , @xmath26 . here ,
the potential is truncated and shifted such that @xmath27 and its derivative with respect to @xmath28 goes to zero at @xmath29 .
\(iii ) the third potential is a hard - sphere interaction modeled by a inverse - power potential ( i256 ) : @xmath30 the potential is cut off at @xmath31 .
as shown previously @xcite , the coexistence and other bulk properties of this interaction potential are very similar to those of the hard - sphere system .
( color online ) interaction potentials corresponding to a modified lj potential ( mlj ) , the inverse-256th power interaction approximating the hard sphere potential ( i256 ) , and the force - shifted inverse - twelfth power potential ( fsi12).,width=288 ] in fig .
[ fig1 ] , we show the three interaction potentials as a function of the distance @xmath4 between the particles .
our systems consist of face centered cubic ( fcc ) crystals of @xmath32 particles , with the ( 100 ) face oriented along the @xmath33 direction , enclosed in a simulation box of dimension @xmath34 .
periodic boundary conditions are maintained along the @xmath35 and @xmath36 directions . along the @xmath33 direction ,
the particles are confined by several layers of the same crystalline particles frozen into their equilibrium configuration ( see fig .
[ fig2 ] ) .
since there are two wall - crystal interfaces , the total area of the interface is @xmath37 . in order to prevent any crystalline particle from penetrating the wall ,
a short - ranged flat wall , modeled by a gaussian potential is inserted at the boundaries along the @xmath33 direction ( at @xmath38 and @xmath39 ) . as shown in our earlier works @xcite
, such a short - ranged flat wall leads to a negligible contribution to the interfacial free energy ( @xmath40 ) .
the interfacial free energy is defined as @xmath41 , where @xmath42 is the free energy of the crystal confined by the frozen walls on both sides , while @xmath43 is the free - energy of the bulk crystal with periodic boundary conditions along all the three cartesian axes .
our objective is to compute @xmath0 using molecular dynamics simulation and thermodynamic integration .
the thermodynamic integration scheme adopted in this work is based on a similar approach devised by us in previous studies to compute the interfacial free energy of a liquid / crystal in contact with structured walls @xcite and to determine the crystal - liquid interfacial free energy @xcite .
the goal is to transform a bulk crystal with periodic boundaries in all three cartesian - axes directions into a crystal in contact with frozen crystalline layers comprising the frozen walls on either side . to accomplish this
, it is neccessary that the boundary conditions along the @xmath33 direction are rearranged at some point during the transformation and that the crystalline particles do not cross the boundaries during this rearrangement .
hence , in the first step , two extremely short - ranged flat walls , modeled by a gaussian potential and denoted by @xmath44 , are gradually inserted at the boundaries of the simulation box at @xmath38 and @xmath45 with periodic boundary conditions maintained along all three axes .
since the structureless flat wall has a very short range ( @xmath46 ) , few crystalline particles interact with it . as a result ,
the contribution of this step to the total interfacial free energy is negligible and does not affect the structure of the crystal near the walls . to integrate the equations of motion in molecular dynamics in presence of such extremely short ranged walls , a very small time - step is needed .
this reduces the computational efficiency since the other forces in the simulation such as those between the crystalline particles are more comparatively long - ranged ( @xmath47 ) .
however , by the use of a multiple time - step scheme @xcite , the computational overhead is reduced significantly and our simulations are slightly less than two times slower than those carried out with a single large time - step . in the second step , interactions of particles through the periodic boundaries along the @xmath33 direction
are gradually switched off while interactions between the crystal and the frozen crystalline wall are gradually switched on , in presence of the flat walls .
the final state is a crystal with periodic boundaries along the @xmath35 and @xmath36 axes and confined by frozen crystalline walls and the flat walls along the @xmath33 direction .
since the free - energy difference per unit area corresponding to the first step is negligible the total contribution to @xmath0 comes only from the second step .
( color online ) crystal ( blue ) in contact with several layers of frozen - in crystalline particles ( red ) , forming walls on either side .
this snapshot represents an instataneous state of the system at equilibrium corresponding to the modified lennard - jones potential ( mlj ) at @xmath48 and @xmath49.,width=288 ] in the ti scheme adopted in this work , similar to our earlier works , the transformation from one state to the next is brought about by changing a switching parameter @xmath50 , which couples to the interaction potential between the walls and the crystalline particles .
the ti scheme consists of two steps . in step 1 ,
the @xmath50 dependent hamiltonian has the form @xmath51 , where @xmath52^{\rm 2})$ ] , with @xmath53 and @xmath54 . by switching @xmath50 from @xmath55 to @xmath56 ,
the system is transformed from a pure bulk crystal ( @xmath57 , no wall ) , to a crystal in contact with an extremely short - ranged gaussian flat wall ( @xmath58 ) . in step 2 ,
the @xmath50-dependent part of the hamiltonian has different forms for the different interaction potentials . for the lj and inverse - twelfth power potentials , @xmath59 , while for the hard - sphere potential , @xmath60 .
here , @xmath61 represents interactions between the crystalline particles through the periodic boundaries , while @xmath62 denotes the interaction between a crystal and a frozen wall particle and is identical to the interaction between two crystalline particles . in the second step
, @xmath57 corresponds to the crystal in contact only with a flat wall , while at @xmath58 the crystal is in contact with the frozen wall in presence of the short - ranged flat wall .
the free - energy difference in each step is computed from the relation @xmath63 where the integration over @xmath50 is computed using simpson s rule .
finally , the interfacial free - energy is directly obtained from @xmath64 and @xmath65 via @xmath66 .
( color online ) density profile of crystal in contact with a wall comprising frozen - in crystalline layers ( solid line ) and a wall consisting of particles arranged in an ideal fcc lattice structure having the same density as the crystal ( dotted line ) .
the density profiles were averaged over @xmath67 independent configurations of the crystal in contact with the same frozen wall .
the vertical dashed lines indicate the positions of the short - ranged flat walls at @xmath38 and @xmath39 .
the profiles were obtained for the mlj potential at the density @xmath48 and the temperature @xmath49.,width=288 ]
to integrate the equations of motion , the velocity form of the verlet algorithm @xcite was implemented with a multiple time - step scheme . for the mlj and fsi12 potentials a smaller time - step @xmath68
was used along with a larger time - step @xmath69 , while for the hard - sphere potential , @xmath68 and @xmath70 were chosen . to maintain constant temperature , @xmath49 ,
every 200 time steps the velocity of the particles was drawn from a maxwell - boltzmann distribution at the desired temperature .
we carried out molecular dynamics simulations in the @xmath71 ensemble at various densities of the crystal at the temperature @xmath49 .
initially , an ideal fcc crystal was put in the simulation box and it was ensured that the center of mass of the crystal along the @xmath33 direction was at @xmath72 , i.e. the two outermost crystalline layers are at the same distance from the boundaries at the two ends . at all densities , the crystal was first equilibrated at the desired temperature in order to chose the reference configurations for the pinned layers of crystalline particles forming the wall .
the frozen wall is constructed by choosing particle positions from crystalline slabs of width @xmath73 ( @xmath74 in case of the hard - sphere potential ) from the boundaries at @xmath38 and @xmath45 , at the two ends of the simulation box , from an equilibrium configuration . when equilibrating the crystal , every time the thermostat was imposed the average center of mass momentum of the crystal was set to zero .
new particles fixed at these instantaneous equilibrated positions were then juxtaposed on the right and left sides of the simulation cell from the slabs at the left and right ends of width @xmath75 or @xmath6 [ equal to the cut - off range of the interaction potential @xmath76 , by shifting their @xmath33-positions by @xmath77 and @xmath39 , respectively .
the system sizes corresponding to the mlj and fsi12 interaction potentials consisted of @xmath78 particles while for the hard - sphere potential , @xmath79 , were chosen .
simulations carried out at larger system sizes yielded values of @xmath0 in agreement with the smaller systems within the statistical errors . for each interaction potential , @xmath0 was computed at various crystal densities ranging from @xmath80 . at each density , @xmath0 was calculated for several realizations of wall configurations .
however , data corresponding to the different realizations were in close agreement with each other , within the numerical errors
. in all values of @xmath0 reported in the next section , the error bars are less than the symbol size and hence they are not specified .
( color online ) thermodynamic integrand as a function of @xmath50 for the second step of the ti scheme at various densities @xmath81 for ( a ) the mlj potential , ( b ) the fsi12 potential and ( c ) the i256 potential.,width=288 ] ( color online ) interfacial free energy vs. density for the ( 100 ) orientation of the crystal in contact with a frozen wall for the three interaction potentials .
the dashed horizontal line indicates the zero axis.,width=288 ]
figure [ fig2 ] shows the simulation set - up of the system with the crystal in contact with several frozen - in layers of crystalline particles forming walls on both sides .
it is clear from fig .
[ fig2 ] that the structure of the crystal near the wall is identical to that in the bulk .
this is also clear from the density profile shown in fig .
there is no change in the density profiles in the layers in contact with the wall and those far away from it .
we have also observed that the energy profile of the crystalline layers near the wall is the same as in the bulk indicating that the wall particles do not change the structure and energy of the crystal layers in contact with it .
if , instead , the wall is constituted by particles attached to ideal fcc lattice sites , the peaks of the density profiles near the wall are higher than those in layers away from the walls ( cf .
[ fig3 ] ) , indicating that the structure of the crystalline layers in the vicinity of such an ideal wall is different from the bulk .
the thermodynamic integrands corresponding to step 2 of the ti scheme and pertaining to the different densities are plotted as a function of @xmath50 in fig .
[ fig4 ] , for the three interaction potentials .
the contribution of the gaussian flat walls in the first step is negligible ( less than @xmath82 of @xmath0 ) and hence not reported .
a negligible value of @xmath83 indicates that the gaussian flat walls effectively do not perturb the thermodynamics of the system .
the contribution to @xmath0 comes almost entirely from the second step . as fig .
[ fig4 ] shows , the thermodynamic integrands corresponding to the two relatively long - ranged interaction potentials are similar , while that of the short - ranged hard - sphere potential is qualitatively different .
the smoothness of the integrands leads to an accurate determination of the free - energy difference from numerical integration . in fig .
[ fig5 ] , we show @xmath0 for the three interaction potentials as a function of the density . for the hard - sphere system ,
the maximum density we consider is @xmath84 since the packing fraction at this density is @xmath85 , which is very close to the maximum theoretical packing fraction @xmath86 . for the other two soft sphere potentials , we are able to consider larger densities , up to @xmath87 .
figure [ fig5 ] shows a non - zero value for @xmath0 as a function of @xmath81 for all the three interaction potentials .
the helmholtz free - energy @xmath88 is given by @xmath89 , where @xmath90 is the energy of the system , @xmath91 the temperature and @xmath92 the entropy .
since the potential energy of the system in contact with the wall is the same as in the bulk and the system is maintained at the same temperature throughout the thermodynamic transformation from the initial to the final states , a non - zero free - energy difference can only be attributed to a change in the entropy of the system .
a non - zero free energy difference indicates that the frozen - in layer of crystalline particles imposes an external field on the crystalline particle , perturbing the thermodynamics as compared to the bulk . for the purely repulsive inverse - twelfth power potential and the hard - sphere potential ,
the interfacial free energy is always positive from the lowest to the highest densities considered . as can be seen in fig .
[ fig1 ] , for @xmath93 , the inverse twelfth - power potential is more repulsive as compared to the hard sphere interaction , resulting in a larger interfacial free energy . the positive excess free energy corresponding to both the purely repulsive potentials at all densities and the modified lj potential at high densities originates due to a decrease of the entropy of the particles near the wall as they have less number of ways to arrange themselves near the wall and the thermal oscillations of the particles around their mean positions is also suppressed in the vicinity of the wall .
in case of the modified lj potential , the interfacial free energy is positive only at high densities .
as we go to lower and lower densities , @xmath0 decreases and eventually becomes negative .
we have also computed the excess free energy for other interaction potentials and observed that this negative excess free energy occurs only for potentials with attractive interactions .
a negative @xmath0 for the mlj potential at low densities indicates that the wall exerts an effective attractive pinning field on the free particles .
since all the contribution to the change in free - energy comes from the change in entropy @xmath94 , a negative @xmath0 corresponds to @xmath95 , meaning that the entropy of the crystal in contact with a wall which is its frozen version , increases at low densities , if the interaction potential has an attractive component .
as shown in a recent work @xcite , this counterintuitive finding is similar to the case of a supercooled binary lj liquid in contact with an amorphous frozen wall having the same structure as the liquid , where a negative @xmath0 was obtained at low temperatures .
in this work , we computed the excess free energy of a crystal in contact with a wall composed of several frozen - in layers of the same crystal , as a function of the density of the crystal .
even though the structure and energy of the crystal in the vicinity of such a wall is the same as in the bulk a non - zero interfacial free energy is obtained , whose origin is purely entropic .
a non - zero interfacial free energy indicates that the wall effectively imposes an external field on the bulk system and thereby the thermodynamics of the system is altered . for purely repulsive interaction potentials , the excess free energy is always positive indicating that entropy of the crystal in contact with the wall is less than that of a bulk crystal .
however , for a modified lj potential , which has an attractive component , the interfacial free energy is negative at low densities , indicating that entropy of the system in contact with the wall is larger than that of the bulk crystal .
99 p. g. de gennes , rev .
* 57 * , 827 ( 1985 ) ; p. g. de gennes , f. brochart - wyart , and d. qur , _ capillarity and wetting phenomena .
drops , bubbles , pearls , waves _ ( springer , new york , 2004 ) .
a. c. zettlemoyer ( ed ) , _ nucleation _
( new york : dekker , 1969 ) .
f. f. abraham , _ homogeneous nucleation theory _
( new york : academic , 1974 ) .
d. kashchiev , _ nucleation : basic theory with applications _
( oxford : butterworth - heinemann , 2000 ) .
a. w. adamson and a. p. gast , _ physical chemistry of surfaces _
( wiley - interscience , new york , 1997 ) . m. heni and h. lwen , phys . rev .
e * 60 * , 7057 ( 1999 ) .
r. benjamin and j. horbach , j. chem . phys . * 137 * , 044707 ( 2012 ) .
r. benjamin and j. horbach , j. chem . phys . * 139 * , 039901 ( 2013 ) .
r. benjamin and j. horbach , j. chem . phys .
* 139 * , 084705 ( 2013 ) .
a. fortini and m. dijkstra , j. phys . : condens .
matter * 18 * , l371 ( 2006 ) .
b. b. laird and r. l. davidchack , j. phys .
c * 111 * , 15952 ( 2007 ) .
v. krackoviak , phys . rep . * 476 * , 51 ( 2009 ) .
p. scheidler , w. kob , and k. binder , j. phys .
b * 108 * , 6673 ( 2004 ) .
p. scheidler , w. kob and k. binder , europhys .
lett . * 59 * , 701 ( 2002 ) .
l. berthier and w. kob , phys .
e * 85 * , 011102 ( 2012 ) .
a. cavagna , t. s. grigera , and p. verrocchio , phys .
98 * , 187801 ( 2007 ) .
g. gradenigo _ et .
al . _ , j. chem . phys . * 138 * , 12a509 ( 2013 ) .
w. kob , s. roldn - vargas , and l. berthier , nature phys .
* 8 * , 164 ( 2011 ) . c. cammarota and g. biroli , proc .
sci . * 109 * , 8850 ( 2011 ) .
k. h. nagamanasa , s. gokhale , a. k. sood and r. ganapathy , nature phys . * 11 * , 403 ( 2015 ) .
w. kob and h. c. andersen , phys .
e * 51 * , 4626 ( 1995 ) .
r. benjamin and j. horbach , phys . rev .
e * 90 * , 060101(r ) ( 2014 ) .
m. p. allen and d. j. tildesley , _ computer simulations of liquids _
( clarendon , oxford , 1987 ) .
d. frenkel and b. smit , _ understanding molecular simulation _
( academic , san diego , 2002 ) ; t. p. straatsma , m. zacharias , and j. a. maccammon , _
computer simulations of biomolecular systems _
( escom , keiden , 1993 ) .
r. benjamin and j. horbach , j. chem .
phys . * 141 * , 044715 ( 2014 ) .
r. benjamin and j. horbach , phys .
e * 91 * , 032410 ( 2015 ) . | using a thermodynamic integration scheme , we compute the free energy cost per unit area , @xmath0 , of forming an interface between a crystal and a frozen structured wall , formed by particles frozen into the same equilibrium structure as the crystal . even though the structure and potential energy of the crystalline phase in the vicinity of the wall is same as in the bulk
, @xmath0 has a non - zero value and increases with increasing density of the crystal and the wall . investigating the effect of several interaction potentials between the particles , we observe a positive @xmath0 at all crystalline densities if the potential is purely repulsive . for models with attractive interactions , such as the lennard - jones potential ,
a negative value for @xmath0 is obtained at low densities . |
the experimental study of nuclei around the doubly - magic @xmath8sn nucleus is of special interest .
such new data allow us to test ingredients of shell - model ( sm ) calculations far away from the valley of stability .
for instance , a realistic effective interaction to be used in the @xmath9 valence shell was derived from the cd - bonn nucleon - nucleon potential and has been tested in several nuclei having few proton particles and neutron holes away from the doubly - magic core : @xmath10te@xmath11 , @xmath12sn@xmath13 , @xmath14sb@xmath15 and @xmath16te@xmath13 @xcite .
the experimental energy levels of these four nuclei are well reproduced showing that such an approach for determining the two - body matrix elements could provide an accurate description of nuclear structure properties .
more recently , the experimental high - spin states of five @xmath17 isotones @xcite were described using this interaction @xcite , named sn100pn .
nevertheless because of their neutron magic number , only the proton parts of the interaction take place in such sm calculations . in order to assess more precisely the quality of this realistic effective interaction
, the results of sm calculations have to be compared to experimental data in other nuclei .
for instance all the parts of the effective interaction can be tested using nuclei having neutron holes in presence of a few proton particles , such as the @xmath18te isotopes with @xmath19 .
yrast excitations in several heavy - mass te isotopes have been already experimentally obtained using deep - inelastic reactions , the states of @xmath20te being identified up to spin ( 23/2 ) and those of @xmath21te up to spin ( 12 ) or ( 14 ) @xcite , i.e. , up to the breaking of the first @xmath7 pair .
more recently , a new level scheme of @xmath22te has been established , showing an isomeric state at 4.4 mev excitation energy @xcite .
the use of binary fission induced by heavy ions allows us to extend the yrast line to higher - spin states , particularly those coming from the breaking of several pairs .
thus , in this work , the @xmath0te isotopes have been produced as fragments of two fusion - fission reactions .
their level schemes have been built from the @xmath5 high - fold events and @xmath23 angular correlations have been analyzed in order to assign spin and parity values to most of the states .
in addition , the half - lives of isomeric states have been measured from the delayed coincidences between a fission fragment detector and the gamma array , or from the timing of the ge detectors .
several te isotopes with heavier masses were also observed in the c+u reaction .
the identification of their complementary fragments by means of the cross coincidences of their @xmath5-rays indicates that they are not produced from the same reaction channel as @xmath0te , but from fissions following transfer or incomplete fusion .
many other nuclei belonging to the @xmath24 region have also singular partners , implying that they are produced from the same mechanism .
in the last part of this paper , all the yrast states of @xmath0te are firstly discussed in comparison with the general features known in this mass region .
the predictions from sm calculations using the sn100pn effective interaction @xcite then are presented , starting from the description of the high - spin states of several sn isotopes in which the breakings of several neutron pairs occupying the @xmath7 orbit were experimentally identified @xcite . with regard to the te isotopes ,
the sm calculations are done in the full valence space in four cases , @xmath25te and the theoretical predictions are compared to the experimental results . in addition , the detailed analysis of the wave functions of the high - spin states shows the effects of the proton - pair breaking along the yrast lines of the heavy te isotopes .
the @xmath18te isotopes of interest were obtained as fission fragments in two experiments .
first , the @xmath1c + @xmath2u reaction was studied at 90 mev incident energy , with a beam provided by the legnaro xtu tandem accelerator .
second , the @xmath3o + @xmath4pb reaction was studied with a 85 mev incident energy beam provided by the vivitron accelerator of ires ( strasbourg ) .
the @xmath5 rays were detected with the euroball array @xcite .
the spectrometer contained 15 cluster germanium detectors placed in the backward hemisphere with respect to the beam , 26 clover germanium detectors located around 90@xmath26 and 30 tapered single - crystal germanium detectors located at forward angles .
each cluster detector consists of seven closely packed large - volume ge crystals @xcite and each clover detector consists of four smaller ge crystals @xcite . in order
to get rid of the doppler effect , both experiments have been performed with thick targets in order to stop the recoiling nuclei ( 47 mg/@xmath27 for @xmath2u and 100 mg/@xmath27 for @xmath4pb targets , respectively ) . the data of the c+u experiment were recorded in an event - by - event mode with the requirement that a minimum of five unsuppressed ge detectors fired in prompt coincidence .
a set of 1.9@xmath28 three- and higher - fold events was available for a subsequent analysis .
for the o+pb experiment , a lower trigger condition ( three unsuppressed ge ) allowed us to register 4@xmath29 events with a @xmath5-fold greater than or equal to 3 . the offline analysis consisted of both multigated spectra and three - dimensional cubes built and analyzed with the radware package @xcite .
more than one hundred nuclei are produced at high spin in such experiments , and this gives several thousands of @xmath5 transitions which have to be sorted out .
single - gated spectra are useless in most of the cases .
the selection of one particular nucleus needs at least two energy conditions , implying that at least two transitions have to be known .
it is worth noting that prompt @xmath5 rays emitted by couples of complementary fragments are detected in coincidence @xcite .
because the isotopes of interest are produced from two different fissioning compound nuclei in this work , the complementary fragments are different in the two reactions .
this gives a fully unambiguous assignment of transitions seen in both experiments .
the relative intensity of the lowest transitions of the even-@xmath30 te isotopes have been measured in spectra in double coincidences with two transitions emitted by one partner . then , we have used the spectra in double coincidences with one low - lying transition of the te isotope of interest and one transition of a partner .
finally , for determining the intensity of the weak transitions , we have analyzed spectra in double coincidences with two transitions of the te level schemes and normalized the obtained results by using relative intensities extracted from the spectra mentioned above .
as reported in previous papers @xcite , another experiment was performed using the saphir heavy - ion detector @xcite , here composed of 32 photovoltaic cells , in order to identify new isomeric states in the fission fragments .
placed in the target chamber of euroball , saphir was used to detect the escaping fission - fragments of the @xmath1c ( 90 mev ) + @xmath2u reaction from a thin 0.14 mg/@xmath27 uranium target .
the detection of the two fragments in coincidence providing a clean signature of fission events was used as the trigger for euroball .
the euroball time window was [ 50 ns@xmath31s ] , allowing detection of delayed @xmath5-rays emitted during the de - excitation of isomeric states .
time spectra between fragments and @xmath5-rays were analyzed in order to measure the half - life of isomeric levels , in a range of several tens to several hundreds of nanoseconds .
it is well known that the @xmath5 rays emitted by fusion - fission fragments do not show any anisotropy in their angular distributions with respect to the incident beam .
however , angular correlations of two successive transitions are meaningful . in order to determine the spin values of excited states ,
the coincidence rates of two successive @xmath5 rays are analyzed as a function of @xmath32 , the average relative angle between the two fired detectors .
the euroball spectrometer had @xmath33=28441 combinations of two crystals , out of which @xmath34 2000 involved different values of relative angle within 2@xmath26 . therefore , in order to keep reasonable numbers of counts , all these angles have been gathered around three average relative angles : 22@xmath26 , 46@xmath26 , and 75@xmath26 . the coincidence rate increases between 0@xmath26 and 90@xmath26 for the dipole - quadrupole cascades , whereas it decreases for the quadrupole - quadrupole or dipole - dipole ones .
the theoretical values of several coincidence rates for the euroball geometry have been already given in previous papers @xcite .
the method has been checked by correctly reproducing the expected angular correlations of @xmath5-transitions having well - known multipole orders and belonging to various fission fragments . when the statistics of our data are too low to perform such a measurement , the spin assignments are based upon ( i ) the already known spins of some states , ( ii ) the assumption that in yrast decays , spin values increase with the excitation energy , ( iii ) the possible existence of cross - over transitions , and ( iv ) the analogy with the level structures of the other isotopes .
the @xmath5-rays emitted by the low - lying states of @xmath35te isotopes have been observed in both fusion - fission reactions used in the present work . regarding @xmath36te and @xmath8te , their yield are so low that only the decays of their low - lying yrast states are observed and the transitions of their partners could not be identified in gated spectra . on the other hand ,
we have measured many new @xmath5-rays emitted by the high - spin states of @xmath0te .
the results are presented in the two following sections . in the third one , we discuss the particular cases of @xmath37te which have been only observed in the @xmath1c + @xmath2u reaction .
previous information of the medium - spin excited states of @xmath39te comes from results of the ( @xmath40,2n@xmath5 ) reaction @xcite .
the positive - parity yrast band was identified up to the @xmath41 state at 3152 kev and the negative - parity one up to the @xmath42 state at 3987 kev .
in addition , a @xmath43 level was proposed at 3850 kev .
we confirm the decay schemes of the 10@xmath44 and 11@xmath45 states and we have added a few new states at higher energy ( see the colored states in fig . [ schema124 ] ) .
we have gathered in table [ gammas_te124 ] the properties of all the transitions assigned to @xmath39te from this work . te deduced in the present work .
the colored levels are new .
the width of the arrows is proportional to the relative intensity of the @xmath5 rays .
, width=291 ] .properties of the transitions assigned to @xmath39te observed in this work . [ cols=">,>,^,^,^",options="header " , ] @xmath46 the number in parenthesis is the error in the least significant digit shown .
+ [ [ odd - n - isotopes-129131te . ] ] odd-@xmath47 isotopes : @xmath48te .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the negative - parity yrast line predicted in @xmath48te ( see figs . [ sm_129te ] and [ sm_131te ] ) exhibits several new high - spin states as compared to the one of the odd-@xmath47 sn isotopes , such as the 21/2@xmath45 level lying above the 19/2@xmath45 one , or the 29/2@xmath45 and 33/2@xmath45 levels lying between the 31/2@xmath45 and 35/2@xmath45 ones .
this is in good agreement with the experimental results .
the color code of the sm states is the same as that of fig .
[ sm_128te ] .
, width=340 ] te .
the color code of the sm states is the same as that of fig .
[ sm_128te ] .
, width=340 ] all the calculated levels of @xmath48te show the same sequence as the experimental ones and have an energy in the right energy range .
nevertheless it is important to notice that the differences between the experimental and calculated excitation energies can amount up to @xmath34400 kev in @xmath49te , its sm level scheme being more compressed than the experimental one .
the components of typical negative - parity states of @xmath49te are given in figs .
[ config129](a) [ config129](d ) .
all the components of the 21/2@xmath45 state predicted at 1796 kev come from the breaking of the proton pair ( @xmath50 ) , and the major component ( 63% ) has one odd neutron in the @xmath51 orbit ( @xmath52 ) .
the major component ( 55% ) of the 27/2@xmath45 state predicted at 2485 kev comes from the breaking of a neutron pair ( @xmath53 ) , the two protons being paired ( @xmath54 ) . above the 27/2@xmath45 state , two sets of states are predicted to coexist in the same energy range .
the wave function of the 31/2@xmath45 state calculated at 3189 kev displays many components with various values of both @xmath55 and @xmath56 ( similar results are obtained for the 35/2@xmath45 state at 4193 kev and the 39/2@xmath45 state at 5254 kev ) . on the other hand , the components of the 29/2@xmath45 state calculated at 3464 kev come mainly from the breaking of the proton pair ( @xmath50 ) . the same is observed for the 33/2@xmath45 state at 4034 kev and the 37/2@xmath45 state at 4855 kev .
the components of typical positive - parity states of @xmath49te are given in figs .
[ config129](e) [ config129](h ) .
the main components of the 15/2@xmath44 state calculated at 1687 kev come from the breaking of the proton pair ( @xmath50 ) , and the major component ( 64% ) has one odd neutron in the @xmath57 orbit ( @xmath58 ) .
less than 300 kev above , the 23/2@xmath44 state has a major component ( 59% ) coming from the breaking of a @xmath7 pair ( @xmath59 and @xmath54 ) .
the two sets of states predicted above the 23/2@xmath44 have different configurations .
the 27/2@xmath44 level exhibits many components with various values of both @xmath55 and @xmath56 ( similar results are obtained for the 31/2@xmath60 state at 3545 kev and the 35/2@xmath44 state at 4513 kev ) , while the components of the 29/2@xmath44 state calculated at 3422 kev come mainly from the breaking of the proton pair ( @xmath50 ) , as the ones of the 31/2@xmath61 and 33/2@xmath44 states . in summary , the calculated states of @xmath48te given in figs .
[ sm_129te ] and [ sm_131te ] can be sorted in three families drawn with the same colors as used for @xmath62te .
the breaking of the first neutron pair leads to the major component of the 19/2@xmath44 , 23/2@xmath44 , and 27/2@xmath45 states .
two sets of states are predicted to coexist above them .
the levels of one set ( drawn in blue ) have many components , resembling collective states , while all the levels of the other one ( drawn in green ) have mainly a broken proton pair .
the neutron part of the sn100pn effective interaction gives a very good description of the high - spin states of the heavy sn isotopes .
the agreement between the calculated and experimental energies of the excited states of the heavy te isotopes is slightly less good , the predicted level schemes being too compressed as compared to the experimental results .
one could suspect that the values of some proton - proton two - body matrix elements ( tbme ) are slightly too attractive , as similar features were observed previously in several @xmath17 isotones , where only such tbme are active .
it is worth recalling that the sn100pn interaction has been recently used @xcite to calculate their high - spin states which were measured previously @xcite .
the theoretical levels schemes of @xmath63cs , @xmath64ba , @xmath65la , and @xmath66ce display the same sequences as the experimental ones , but some high - spin states are predicted too low in energy ( by an amount from @xmath34 200 kev to 400 kev ) .
noteworthy is the fact that only the fully aligned states corresponding to the breaking of one neutron pair in the @xmath51 orbit ( see the cases written in bold in table [ spinmax ] ) are observed in the heavy-@xmath30 te isotopes . on the other hand ,
the main configurations of all the higher spin states contain the breaking of the proton pair .
this is at variance with the sn isotopes where the breaking of two and three neutron pairs are identified @xcite .
in this work , the @xmath0te isotopes were produced in two fusion - fission reactions , @xmath3o+@xmath4pb and @xmath1c+@xmath2u , and the emitted @xmath5 rays were detected by the euroball array .
a fragment detector was also associated to the @xmath5-ray detection , in another experiment .
all the data sets recorded in these experiments have allowed us to extend the level schemes of @xmath0te isotopes to higher spins ( @xmath67 ) and higher excitation energies ( @xmath68 mev ) .
furthermore , the performed @xmath6 angular correlations supported the former spin assignments in most cases and yielded about 30 new spin assignments .
the half - lives of three isomeric states were measured , showing that the decay of the 10@xmath44 state of @xmath69te is much faster than reported in the last compilation .
in addition , the unexpected production of heavier-@xmath30 te nuclei as well as of some isotopes of xe , ba , and ce was discussed in terms of transfer / incomplete fusion in the @xmath1c+@xmath2u reaction , thanks to the unambiguous identification of their complementary fragments .
the high - spin structures of the @xmath0te have been first discussed in comparison with the general features known in the mass region .
then shell - model calculations using the sn100pn effective interaction have been successfully compared to experimental results .
thanks to the @xmath56 and @xmath55 components of the sm wave functions , the effect of the proton - pair breaking has been identified .
the fully aligned component with @xmath50 shows up clearly by the appearance of sets of excited levels , for instance above the @xmath70 10@xmath44 state in the even-@xmath47 isotopes and the @xmath71 27/2@xmath45 state in the odd-@xmath47 ones .
the euroball project was a collaboration among france , the united kingdom , germany , italy , denmark and sweden .
the first experiment has been performed under u.e .
contract ( erb fhgect 980 110 ) at legnaro .
the second experiment has been supported in part by the eu under contract hpri - ct-1999 - 00078 ( euroviv ) .
we thank many colleagues for their active participation in the experiments , drs .
a. bogachev , a. buta , j.l .
durell , th .
ethvignot , f. khalfalla , i. piqueras , a.a . roach , a.g .
smith and b.j .
we thank the crews of the vivitron .
we are very indebted to m .- a .
saettle for preparing the pb target , p. bednarczyk , j. devin , j .-
gallone , p. mdina , and d. vintache for their help during the experiment .
brown , n.j .
stone , j.r .
stone , i.s .
towner and m. hjorth - jensen , phys .
c * 71 * , 044317 ( 2005 ) .
a. astier _ et al .
c * 85 * , 064316 ( 2012 ) .
srivastava , m.j .
ermamatov and i. o. morales , j. phys .
g : nucl . part
. phys . * 40 * , 035106 ( 2013 ) .
j. simpson , z. phys .
a * 358 * , 139 ( 1997 ) and f. a. beck prog .
a * 28 * , 443 ( 1992 ) .
j. eberth _
et al . _ ,
methods * a 369 * , 135 ( 1996 ) .
g. duchne _
methods * a 432 * , 90 ( 1999 ) .
d. radford , nucl .
methods * a 361 * , 297 and 306 ( 1995 ) .
et al . _ ,
a530 * , 111 ( 1991 ) .
et al . _ , acta phys .
polonica b * 27 * , 179 ( 1996 ) .
r. lucas _ et al .
j. a * 15 * , 315 ( 2002 ) .
j. a * 24 * , 39 ( 2005 ) .
_ , _ proceedings of the 2nd international workshop on nuclear fission and fission product spectroscopy , seyssins , france _ , april 1998 , aip conf . proc . * 447 * , 143 ( 1998 ) . a. astier _ et al .
j. a * 30 * , 541 ( 2006 ) .
et al . _ ,
c * 84 * , 054305 ( 2011 ) .
a. astier _
c * 88 * , 024321 ( 2013 ) .
et al . _ ,
nucl . phys . *
a636 * , 379 ( 1998 ) .
a. kerek , nucl . phys . *
a176 * , 466 ( 1971 ) .
ensdf database , http://www.nndc.bnl.gov / ensdf/. a.m. bruce _ et al . _ ,
c * 55 * , 620 ( 1997 ) .
xu , p.m. walker , and r. wyss , phys .
c * 59 * , 731 ( 1999 ) .
a. kerek , nucl . phys .
* a185 * , 497 ( 1972 ) . m. houry , thesis , universit paris - sud , orsay , 2000 , saclay report no
. dapnia / sphn-00 - 01-t .
m. kanbe and k. kitao , nuclear data sheets * 94 * , 227 ( 2001 ) .
a. kerek , p. carl , and j. mcdonald , nucl . phys .
* a198 * , 466 ( 1972 ) . j. genevey _
et al . _ ,
c * 63 * , 054315 ( 2001 ) .
_ , _ proceedings of the 2nd international workshop on nuclear fission and fission product spectroscopy , seyssins , france _ , april 1998 , aip conf
* 447 * , 220 ( 1998 ) .
t. kibdi et al . , nucl .
methods a * 589 * , 202 ( 2008 ) .
a. kerek , j. kownacki , a. marekius and j. pihl , nucl . phys . *
a194 * , 64 ( 1972 ) .
j. katakura , nuclear data sheets * 112 * , 1647 ( 2011 ) .
a. hazhizume , nuclear data sheets * 112 * , 495 ( 2011 ) .
h. huck , m.l .
prez and j.j .
rossi , phys . rev .
* c 26 * 621 ( 1982 ) .
stone and w.b .
walters , z. phys .
a * 328 * , 257 ( 1987 ) . c. t. zhang _ et al . _ , z. phys .
a * 353 * , 11 ( 1995 ) .
y. tendow , nuclear data sheets * 77 * , 631 ( 1996 ) .
b. fogelberg _
_ , _ proceedings of the 2nd international workshop on nuclear fission and fission product spectroscopy , seyssins , france _ , april 1998 , aip conf
. proc . * 447 * , 191 ( 1998 ) . c. t. zhang _ et al .
* 77 * , 3743 ( 1996 ) .
a. korgul _
et al . _ , eur .
j. a * 7 * , 167 ( 2000 ) .
paul , d.b .
fossan , g.j . lane , j.m .
sears , i. thorslund , and p. vaska , phys . rev . *
c 53 * 1562 ( 1996 ) .
_ , z. phys .
a * 354 * , 233 ( 1996 ) .
p. bhattacharyya _
et al . _ ,
phys . rev .
* c64 * 054312 ( 2001 ) .
j. a * 20 * , 245 ( 2004 ) . | the @xmath0te nuclei have been produced as fission fragments in two fusion reactions induced by heavy - ions ( @xmath1c + @xmath2u at 90 mev bombarding energy and @xmath3o + @xmath4pb at 85 mev ) and studied with the euroball array .
their high - spin level schemes have been extended to higher excitation energy from the triple @xmath5-ray coincidence data .
the @xmath6 angular correlations have been analyzed in order to assign spin and parity values to many observed states .
moreover the half - lives of isomeric states have been measured from the delayed coincidences between the fission - fragment detector saphir and euroball , as well as from the timing information of the ge detectors .
the behaviors of the yrast structures identified in the present work are first discussed in comparison with the general features known in the mass region , particularly the breakings of neutron pairs occupying the @xmath7 orbit identified in the neighboring sn nuclei .
the experimental level schemes are then compared to shell - model calculations performed in this work .
the analysis of the wave functions shows the effects of the proton - pair breaking along the yrast lines of the heavy te isotopes . |
chvtal @xcite proposed to call a linear order @xmath1 on the vertex set of an undirected graph @xmath2 _ perfect _ if the greedy coloring algorithm applied to each induced subgraph @xmath3 of @xmath2 gives an optimal coloring of @xmath3 : consider the vertices of @xmath3 sequentially by following the order @xmath1 and assign to each vertex @xmath4 the smallest color not used on any neighbor @xmath5 of @xmath4 , @xmath6 .
a graph is _ perfectly orderable _ if it admits a perfect order .
chvtal proved that @xmath1 is a perfect order if and only if there is no chordless path with four vertices @xmath7 and three edges @xmath8 ( written @xmath9 ) with @xmath10 and @xmath11 .
he also proved that perfectly orderable graphs are perfect .
the class of perfectly orderable graphs properly contains many important , well - known classes of perfect graphs such as chordal graphs and comparability graphs .
perfectly orderable graphs have been extensively studied in the literature ; see hong s comprehensive survey @xcite for more information .
recognizing perfectly orderable graphs is np - complete @xcite ( see also @xcite ) . also , no characterization of perfectly orderable graphs by forbidden induced subgraphs is known .
these facts have motivated researchers to study subclasses of perfectly orderable graphs ; see , e.g. , @xcite and the literature given there .
observe that a linear order @xmath1 corresponds to an acyclic orientation by directing the edge @xmath12 from @xmath13 to @xmath14 if and only if @xmath15 .
thus , a graph is perfectly orderable if and only if it admits an acyclic orientation such that the orientation of no chordless path @xmath16 is of type @xmath17 ( equivalently , the orientation of every @xmath16 is of type @xmath18 , or @xmath19 ) ; see figure [ fig:1 ] .
= [ draw , circle , inner sep=1.5pt ] ( 1 ) at ( 0,1 ) ; ( 2 ) at ( 1.5,1 ) ; ( 3 ) at ( 3,1 ) ; ( 4 ) at ( 4.5,1 ) ; \(1 ) ( 2 ) ; ( 2 ) ( 3 ) ; ( 4 ) ( 3 ) ; ( 2.25,0.2 ) node type @xmath17 ; = [ draw , circle , inner sep=1.5pt ] ( 1 ) at ( 0,1 ) ; ( 2 ) at ( 1.5,1 ) ; ( 3 ) at ( 3,1 ) ; ( 4 ) at ( 4.5,1 ) ; \(2 ) ( 1 ) ; ( 2 ) ( 3 ) ; ( 3 ) ( 4 ) ; ( 2.25,0.2 ) node type @xmath20 ; = [ draw , circle , inner sep=1.5pt ] ( 1 ) at ( 0,1 ) ; ( 2 ) at ( 1.5,1 ) ; ( 3 ) at ( 3,1 ) ; ( 4 ) at ( 4.5,1 ) ; \(1 ) ( 2 ) ; ( 2 ) ( 3 ) ; ( 3 ) ( 4 ) ; ( 2.25,0.2 ) node type @xmath21 ; = [ draw , circle , inner sep=1.5pt ] ( 1 ) at ( 0,1 ) ; ( 2 ) at ( 1.5,1 ) ; ( 3 ) at ( 3,1 ) ; ( 4 ) at ( 4.5,1 ) ; \(1 ) ( 2 ) ; ( 3 ) ( 2 ) ; ( 3 ) ( 4 ) ; ( 2.25,0.2 ) node type @xmath19 ; one of the natural subclass of perfectly orderable graphs for which the recognition complexity , as well as an induced subgraph characterization are still unknown is the following ( cf .
@xcite ) .
[ def : coali ] a graph is a _ coalition graph _ if it admits an acyclic orientation such that every induced @xmath9 has the end - edges @xmath22 and @xmath23 oriented in the same way , that is , every oriented @xmath16 is of type @xmath21 or @xmath19 .
such an orientation is called a _
coalition orientation_. equivalently , a graph is a coalition graph if it admits a linear order @xmath1 on its vertex set such that every induced @xmath9 has @xmath10 if and only if @xmath24 . in @xcite ,
coalition graphs are called one - in - one - out graphs .
examples of coalition graphs include comparability graphs , hence all bipartite graphs .
a related graph class has been introduced by olariu in @xcite : [ def : oppo ] a graph is an _ opposition graph _ if it admits an acyclic orientation such that every induced @xmath9 has the end - edges @xmath22 and @xmath23 oriented in opposition , that is , every oriented @xmath16 is of type @xmath17 or @xmath20 .
such an orientation is called an _ opposition orientation_. equivalently , a graph is an opposition graph if it admits a linear order @xmath1 on its vertex set such that every @xmath9 has @xmath10 if and only if @xmath11 .
olariu @xcite proved that opposition graphs are perfect .
he also conjectures ( @xcite ) that not all opposition graphs are perfectly orderable .
examples of opposition graphs include all split graphs .
the recognition and characterization problems for opposition graphs are still open . coalition graphs and opposition graphs have been studied in the past from the combinatorial and algorithmic point of view .
the characterization and recognition problems for these graphs have been solved for a few special graph classes so far .
a natural subclass of opposition graphs consists of those admitting an acyclic orientation in which every @xmath16 is oriented as type 1 ( equivalently , every @xmath16 is oriented as type 0 ) .
these are called _ bipolarizable graphs _ , and have been characterized by ( infinitely many ) forbidden induced subgraphs in @xcite , and have been recognized using @xmath25 adjacency matrix multiplications and thus in @xmath26 time in @xcite , and in @xmath27 time in @xcite ; @xmath28 is the vertex number and @xmath29 is the edge number of the input graph .
another subclass of opposition graphs are the so - called welsh - powell opposition graphs ; see @xcite for more information . in a recent paper @xcite
, bipartite opposition graphs have been characterized by ( infinitely many ) forbidden induced subgraphs , and have been recognized in linear time .
this paper gives also characterizations for complements of bipartite graphs that are coalition or opposition graphs .
it turns out that co - bipartite coalition graphs and co - bipartite opposition graphs coincide , and they are exactly the complements of bipartite permutation graphs , hence can be recognized in linear time .
there is also a characterization of co - bipartite coalition / opposition graphs in terms of their bi - matrices .
this characterization is similar to the one of co - bipartite perfectly orderable graphs given by chvtal in @xcite , which has a close connection to a theorem in mathematical programming .
we first address in section [ sec : genoppo ] the so - called generalized opposition graphs introduced by chvtal ; these graphs are obtained when the condition acyclic in definition [ def : oppo ] is dropped .
this concept turns out to be useful when considering opposition graphs in certain graph classes .
we give a characterization for generalized opposition graphs in terms of an auxiliary graph , which leads to a polynomial time recognition algorithm . in section
[ sec : gemhouse ] we extend the results in @xcite on bipartite opposition graphs to the larger class of @xmath30-free graphs .
it turns out that inside this graph class , opposition graphs and generalized opposition graphs coincide , hence @xmath30-free opposition graphs can be recognized in polynomial time . in section [ sec : dhoppo ] we give a forbidden subgraph characterization for distance - hereditary opposition graphs , a subclass of @xmath30-free opposition graphs
; this result includes the subgraph characterization for tree opposition graphs found in @xcite . in section [ sec : dhcoali ]
we show that @xmath31-free coalition graphs can be recognized in polynomial time by modifying the auxiliary graph for generalized opposition graphs . for the smaller class of distance - hereditary coalition graphs ,
we give a faster recognition algorithm by showing that they are indeed comparability graphs .
[ [ definitions - and - notation ] ] definitions and notation + + + + + + + + + + + + + + + + + + + + + + + + we consider only finite , simple , and undirected graphs . for a graph @xmath2
, the vertex set is denoted @xmath32 and the edge set is denoted @xmath33 . for a vertex @xmath5 of a graph @xmath2 ,
the neighborhood of @xmath5 in @xmath2 is denoted @xmath34 or simply @xmath35 if the context is clear , and the degree of @xmath5 is @xmath36 .
write @xmath37=n(u)\cup\{u\}$ ] . for a set @xmath38 of vertices of a graph @xmath2 , write @xmath39 and @xmath40=n(u)\cup u$ ] .
the subgraph of @xmath2 induced by @xmath38 is denoted @xmath41 $ ] . if @xmath5 is a vertex of a graph @xmath2 , then @xmath42 is @xmath43 $ ] . for @xmath44 ,
let @xmath45 denote a chordless path with @xmath46 vertices and @xmath47 edges , and for @xmath48 , let @xmath49 denote a chordless cycle with @xmath46 vertices and @xmath46 edges .
we write @xmath50 and @xmath51 , meaning the chordless path with vertices @xmath52 and edges @xmath53 , @xmath54 , respectively , the chordless cycle with vertices @xmath52 and edges @xmath53 , @xmath54 , and @xmath55 .
the edges @xmath56 and @xmath57 of the path @xmath45 ( @xmath58 ) are the _ end - edges _ and the other edges are the _ mid - edges _ of the path , while the vertices @xmath59 and @xmath60 are the _ end - vertices _ and the other vertices are the _ mid - vertices _ of the path . in this paper ,
all paths @xmath45 and all cycles @xmath49 will always be induced .
an _ orientation _ of an undirected graph @xmath2 is a directed graph @xmath61 obtained from @xmath2 by replacing each edge @xmath12 of @xmath2 by either @xmath62 or @xmath63 ( but not both ) .
recall that a graph is a comparability graph if it admits an acyclic orientation of its edges such that whenever we have @xmath64 and @xmath65 we also have @xmath66 .
equivalently , a graph is a comparability graph if it admits an acyclic orientation which puts the two edges of every @xmath67 in opposition .
it follows that comparability graphs are coalition graphs .
complements of bipartite / comparability graphs are _ co - bipartite / co - comparability graphs_. a cycle @xmath68 of length @xmath69 is also called a _ hole _ , and the complement of the path @xmath70 is also called a _ house_. a _ gem _ is obtained from the path @xmath16 by adding a vertex adjacent to all four vertices of the @xmath16 .
a _ domino _ is obtained from the cycle @xmath71 by adding a chord between two ( distance @xmath19)-vertices .
a _ chordal _ graph has no induced @xmath68 , @xmath72 .
a _ distance - hereditary _ graph has no gem , no house , no domino , and no hole as induced subgraphs .
chordal distance - hereditary graphs are also called _ ptolemaic
_ graphs . for more information on graph classes appearing in this paper , and for
basic graph notation and definitions not given here , see , e.g. , @xcite .
it was conjectured by chvtal @xcite and is implied by the strong perfect graph theorem , that if the term acyclic in definition [ def : oppo ] of opposition graphs is dropped , the larger class is still a class of perfect graphs .
chvtal @xcite proposed to call these graphs _
generalized opposition graphs_. [ def : genoppo ] a graph is a
_ generalized opposition graph _ if it admits an orientation such that every oriented @xmath16 is of type @xmath17 or @xmath20 .
such an orientation is called a _ generalized opposition orientation_. the co - bipartite graphs @xmath73 , @xmath74 , are examples of generalized opposition graphs that are non - opposition graphs .
the main result of this section is a good characterization of generalized opposition graphs in terms of an auxiliary graph defined below . given a graph @xmath75 , the graph @xmath76 has * @xmath77 as vertex set , * two vertices @xmath78 and @xmath79 are adjacent in @xmath76 if @xmath80 , or else @xmath81 or @xmath82 is an induced @xmath16 in @xmath2 .
notice that for each edge @xmath83 of @xmath2 that is an end - edge of an induced @xmath16 in @xmath2 there are two corresponding vertices in @xmath76 , namely @xmath78 and @xmath84 .
these two vertices are adjacent in @xmath76 and indicate the two possible orientations for the edge @xmath83
. see figure [ fig : genoppo ] for an example .
= [ draw , circle , inner sep=1pt ] ( 1 ) at ( 0,3 ) @xmath20 ; ( 2 ) at ( 3,2 ) @xmath21 ; ( 3 ) at ( 0,1 ) @xmath19 ; ( 4 ) at ( 4,3 ) @xmath0 ; ( 5 ) at ( 1,2 ) @xmath85 ; ( 6 ) at ( 4,1 ) @xmath86 ; \(1 ) ( 3 ) ( 5 )
( 1 ) ; ( 2 ) ( 4 ) ( 6 ) ( 2 ) ; ( 1 ) ( 4 ) ; ( 3 ) ( 6 ) ; ( 5 ) ( 2 ) ; ( 2,0 ) node @xmath2 ; = [ draw , rectangle , inner sep=1pt ] \(15 ) at ( 1.5,5 ) @xmath87 ; ( 26 ) at ( 4.2,5 ) @xmath88 ; ( 31 ) at ( 5.7,3 ) @xmath89 ; ( 42 ) at ( 4.2,1 ) @xmath90 ; ( 53 ) at ( 1.5,1 ) @xmath91 ; ( 64 ) at ( 0,3 ) @xmath92 ; ( 51 ) at ( 2.2,4 ) @xmath93 ; ( 62 ) at ( 3.5,4 ) @xmath94 ; ( 13 ) at ( 4.2,3 ) @xmath95 ; ( 24 ) at ( 3.5,2 ) @xmath96 ; ( 35 ) at ( 2.2,2 ) @xmath97 ; ( 46 ) at ( 1.5,3 ) @xmath98 ; \(51 ) ( 15 ) ; ( 62 ) ( 26 ) ; ( 15 ) ( 26 ) ; ( 51 ) ( 62 ) ; ( 26 ) ( 31 ) ( 42 ) ( 53 ) ( 64 ) ( 15 ) ; ( 51 ) ( 62 ) ( 13 ) ( 24 ) ( 35 ) ( 46 ) ( 51 ) ; ( 13 )
( 31 ) ; ( 24 ) ( 42 ) ; ( 35 ) ( 53 ) ; ( 46 ) ( 64 ) ; ( 2.9,0 ) node @xmath76 ; [ thm : genoppo ] a graph @xmath2 is a generalized opposition graph if and only if @xmath76 is bipartite .
write @xmath99 for the vertex set of @xmath76 .
suppose @xmath2 is an opposition graph , and consider an opposition orientation @xmath100 of @xmath2 .
partition @xmath99 into @xmath101 then , by definition of adjacencies in @xmath76 , @xmath102 and @xmath103 are independent sets in @xmath76 , i.e. , @xmath76 is bipartite .
suppose @xmath76 is bipartite , and consider a bipartition @xmath104 into independent sets @xmath102 and @xmath103 .
we first orient the end - edges @xmath83 of @xmath16 s in @xmath2 according to the bipartition : the edge @xmath83 gets the direction @xmath105 if @xmath106 , otherwise @xmath63 . since @xmath78 and @xmath84 are adjacent in @xmath76 , orienting the edge @xmath83 of @xmath2 in this way is well - defined . now , extending this partial orientation @xmath107 to the remaining edges of @xmath2 in an arbitrary way , we obtain an orientation @xmath61 of @xmath2 , which is indeed a generalized opposition orientation of @xmath2 . since , given a graph @xmath2 with @xmath29 edges , @xmath76 can be constructed in @xmath108 steps , we have : [ cor1 ] generalized opposition graphs with @xmath29 edges can be recognized in @xmath108 time .
moreover , a generalized opposition orientation of a generalized opposition graph can be obtained in the same time .
assume @xmath76 is bipartite and @xmath104 is a bipartition of @xmath76 , and let @xmath107 be the partial orientation on the end - edges of @xmath16 s in @xmath2 in the second part of the proof of theorem [ thm : genoppo ] above .
then , if @xmath107 is _ acyclic _ , we can extend it to an _ acyclic _ orientation @xmath61 of @xmath2 .
hence we have : [ pro : oppo ] a graph @xmath2 is an opposition graph if and only if * @xmath76 is bipartite , and * there exists a bipartition @xmath102 , @xmath103 of @xmath76 such that @xmath107 is acyclic . as connected bipartite graphs have only one bipartition , and checking if a directed graph is acyclic is doable in linear time , we obtain : given @xmath2 such that @xmath76 is connected , it can be decided in linear time if @xmath2 is an opposition graph .
it would be interesting to discuss opposition graphs @xmath2 in the case of _ disconnected _ @xmath76 .
notice that opposition graphs are special cases of generalized opposition graphs .
so , if @xmath2 is an opposition graph , then , by theorem [ thm : genoppo ] , @xmath76 is bipartite .
the converse is , in general , not true ; recall the example depicted in figure [ fig : genoppo ] .
however , the converse does hold true for certain special graph classes , e.g. , bipartite graphs as proved in @xcite recently . in this section
, we extend this result for @xmath30-free graphs .
[ thm : gemhouse ] the following statements are equivalent for a @xmath30-free graph @xmath2 : * @xmath2 is an opposition graph ; * @xmath2 is a generalized opposition graph ; * @xmath76 is bipartite .
the implication ( i ) @xmath109 ( ii ) is obvious , ( ii ) @xmath109 ( iii ) is implied by theorem [ thm : genoppo ] .
it remains to prove the implication ( iii ) @xmath109 ( i ) .
suppose that @xmath76 is bipartite , and @xmath102 , @xmath103 is an arbitrary bipartition of @xmath76 .
in view of proposition [ pro : oppo ] , we will show that @xmath107 is acyclic . recall that an edge @xmath83 of @xmath2 that is an end - edge of an induced @xmath16 is oriented @xmath105 in @xmath107 if @xmath106 , otherwise @xmath110 . also recall that any induced @xmath111 in @xmath2 is good : it has @xmath105 and @xmath112 in @xmath107 or else @xmath63 and @xmath113 in @xmath107 .
assume , by contradiction , that @xmath114 is a directed cycle in @xmath107 .
note that , since @xmath115 , by definition of @xmath107 , @xmath116 for all @xmath117 .
choose such a cycle @xmath118 with minimum length @xmath119 .
then , @xmath120 otherwise there would be a shorter directed cycle in @xmath107 .
let @xmath121 be an induced @xmath16 containing @xmath122 as an end - edge .
let , without loss of generality , @xmath121 be @xmath123 for some vertices @xmath124 and @xmath125 . since @xmath126 , @xmath127 , and hence @xmath128 .
we distinguish three cases ; in each case we will arrive at a contradiction .
case 1 : @xmath124 and @xmath125 are outside @xmath118 . in this case
@xmath129 , otherwise @xmath130 would induce a @xmath131 ( note that @xmath132 contains a @xmath131 ) , a gem or a house in @xmath2 .
hence @xmath133 , otherwise @xmath134 would be a bad induced @xmath16 with @xmath135 but @xmath128 .
it follows that @xmath136 and @xmath137 , otherwise @xmath138 would be a bad @xmath16 ( if @xmath139 ) or the chord @xmath140 of @xmath118 ( if @xmath141 ) would be an end - edge of the @xmath142 , contradicting ( [ eq:1 ] ) .
thus , @xmath143 is an induced @xmath16 , hence @xmath144 .
next , we show that @xmath145 .
assume not .
then @xmath146 , otherwise @xmath147 would be a bad @xmath16 ( if @xmath148 ) or @xmath149 would be a chord of @xmath118 that is an end - edge of the induced @xmath150 .
but then @xmath151 and @xmath124 induce a house ( if @xmath152 or a gem ( otherwise ) .
thus , @xmath145 .
hence @xmath153 ( otherwise @xmath154 would be a bad @xmath16 ) and @xmath155 ( otherwise @xmath156 would induce a house or a gem ) .
now , if @xmath157 , then @xmath158 is a bad @xmath16 , a contradiction .
if @xmath146 , then @xmath159 is a chord of @xmath118 that is an end - edge of the induced @xmath160 , a contradiction to ( [ eq:1 ] ) .
case 1 is settled .
case 2 : @xmath161 for some @xmath162 .
then , by ( [ eq:1 ] ) , @xmath125 is outside @xmath118 or else @xmath163 .
( note that @xmath164 , hence @xmath165 . )
first , @xmath129 , otherwise @xmath166 would induce a gem , house or a @xmath131 .
hence @xmath167 , otherwise @xmath168 would be a bad @xmath16 , and @xmath169 , otherwise @xmath154 would be a bad @xmath16 ( if @xmath170 ) or the chord @xmath171 of @xmath118 would be an end - edge of the induced @xmath172 .
but now the chord @xmath173 of @xmath118 is an end - edge of the @xmath174 , contradicting ( [ eq:1 ] ) .
case 2 is settled .
case 3 : @xmath124 is outside @xmath118 and @xmath175 for some @xmath176 .
note that @xmath177 .
first , @xmath167 ( otherwise @xmath178 would induce a @xmath131 , a house or a gem ) , @xmath133 ( otherwise @xmath179 would be a bad @xmath16 ) , and @xmath137 ( otherwise the chord @xmath173 of @xmath118 would be an end - edge of the induced @xmath180 ) .
thus , @xmath143 is an induced @xmath16 , hence @xmath144 .
but now @xmath181 is a directed cycle in @xmath107 of length @xmath182 , contradicting the choice of @xmath118 .
case 3 is settled , and the proof of theorem [ thm : gemhouse ] is complete .
since bipartite graphs are @xmath30-free , theorem [ thm : gemhouse ] implies : a bipartite graph is an opposition graph if and only if it is a generalized opposition graph .
moreover , theorem [ thm : gemhouse ] and corollary [ cor1 ] imply the following result : given a @xmath30-free graph @xmath2 with @xmath29 edges , it can be decided in @xmath108 time if @xmath2 is an opposition graph , and if so , an opposition orientation of @xmath2 can be obtained in the same time .
we were not able to find a forbidden subgraph characterization for @xmath183 , @xmath184-free opposition graphs .
this section gives such a characterization for distance - hereditary opposition graphs which form an important subclass of @xmath183 , @xmath184-free opposition graphs .
[ pro : tkag1g2 ] the graphs @xmath185 @xmath186 depicted in figure @xmath187 are minimal non - opposition graphs .
= [ draw , circle , inner sep=1.5pt ] ( 1 ) at ( 0,1 ) ; ( 2 ) at ( 1,1 ) ; ( 3 ) at ( 2,1 ) [ label = below:@xmath20 ] ; ( 4 ) at ( 3,1 ) [ label = below:@xmath21 ] ; ( 7 ) at ( 6,1 ) ; ( 8) at ( 7,1 ) [ label = below:@xmath190 ; ( k1 ) at ( 8,1 ) ; ( k2 ) at ( 9,1 ) ; ( a ) at ( 2,2 ) ; ( b ) at ( 7,2 ) ; \(x ) at ( 4,1 ) ; ( y ) at ( 5.5,1 ) ; \(1 ) ( 2 ) ( 3 ) ( 4 )
( x ) ; ( x ) ( y ) ; ( y ) ( 7 ) ( 8) ( k1 ) ( k2 ) ; ( a ) ( 3 ) ; ( b ) ( 8) ; ( name ) at ( 8,0.3 )
[ label = right:@xmath185 ] ; = [ draw , circle , inner sep=1.5pt ] ( 1 ) at ( 0,1 ) ; ( 2 ) at ( 1,1 ) ; ( 3 ) at ( 2,1 ) ; ( 4 ) at ( 3,1 ) ; ( 5 ) at ( 1,2 ) ; ( 6 ) at ( 2,2 ) ; \(1 ) ( 2 ) ( 3 ) ( 4 ) ; ( 2 ) ( 5 ) ( 6 ) ( 3 ) ; ( 1.5,0.3 ) node @xmath192 ; = [ draw , circle , inner sep=1.5pt ] ( 1 ) at ( 0,1 ) ; ( 2 ) at ( 1,1 ) ; ( 3 ) at ( 2,1 ) ; ( 4 ) at ( 3,1 ) ; ( 5 ) at ( 4,1 ) ; ( 6 ) at ( 5,1 ) ; ( 7 ) at ( 2.5,2 ) ; ( 8) at ( 2.5,3 ) ; ( 9 ) at ( 2.5,4 ) ; \(1 ) ( 2 ) ( 3 ) ( 4 ) ( 5 ) ( 6 ) ; ( 3 ) ( 7 )
( 8) ( 9 ) ; ( 4 ) ( 7 ) ; ( 2.5,0.3 ) node @xmath193 ; = [ draw , circle , inner sep=1.5pt ] ( 1 ) at ( 0,2 ) ; ( 2 ) at ( 1,2 ) ; ( 3 ) at ( 2,2 ) ; ( 4 ) at ( 3,2 ) ; ( 5 ) at ( 4,2 ) ; ( 6 ) at ( 5,2 ) ; ( 7 ) at ( 2.5,1 ) ; ( 8) at ( 2.5,3 ) ; ( 9 ) at ( 2.5,4 ) ; \(1 ) ( 2 ) ( 3 ) ( 4 ) ( 5 ) ( 6 ) ; ( 7 ) ( 3 )
( 8) ( 9 ) ; ( 7 ) ( 4 ) ( 8) ; ( 2.5,0.3 ) node @xmath194 ; [ thm : dhoppo ] the following statements are equivalent for a distance - hereditary graph @xmath2 : * @xmath2 is an opposition graph ; * @xmath2 is a generalized opposition graph ; * @xmath76 is bipartite ; * @xmath2 is @xmath185-free @xmath195 and @xmath196-free @xmath197 .
the main and technical part in the proof of theorem [ thm : dhoppo ] is the corresponding result for ptolemaic graphs ( i.e. , @xmath198-free chordal graphs ) .
[ thm : ptooppo ] the following statements are equivalent for a ptolemaic graph @xmath2 : * @xmath2 is an opposition graph ; * @xmath2 is a generalized opposition graph ; * @xmath76 is bipartite ; * @xmath2 is @xmath185-free @xmath195 and @xmath199-free @xmath197 .
theorem [ thm : dhoppo ] can be derived from theorem [ thm : ptooppo ] as follows .
two distinct vertices @xmath5 and @xmath4 of a graph @xmath2 are _ twins _ if @xmath200 . if @xmath201 are twins in @xmath2 , then , clearly , @xmath2 is an opposition graph if and only if @xmath42 is an opposition graph .
so , by induction , we may assume that the distance - hereditary graph @xmath2 in theorem [ thm : dhoppo ] has no twins .
now , if @xmath2 has no @xmath202 , then @xmath2 is chordal , and theorem [ thm : dhoppo ] follows from theorem [ thm : ptooppo ] .
if @xmath2 contains an induced @xmath202 , then , as @xmath2 has no twins , it can be shown ( see also ( * ? ? ?
* claim 3.5 ) ) that @xmath2 must contain an induced @xmath192 or an induced house , or an induced domino .
as @xmath2 is a distance - hereditary graph , none of these cases is possible , hence theorem [ thm : dhoppo ] .
* proof of theorem [ thm : ptooppo ] . * ( i ) @xmath109 ( ii ) is obvious and ( ii ) @xmath109 ( iii ) follows from theorem [ thm : genoppo ] .
the implication ( iii ) @xmath109 ( iv ) follows by noting that the graphs @xmath2 depicted in fig .
[ fig : dhoppo ] have non - bipartite @xmath76 .
it remains to prove ( iv ) @xmath109 ( i ) .
let @xmath2 be a @xmath185-free @xmath195 and @xmath199-free ptolemaic graph .
we have to show that @xmath2 is an opposition graph .
the proof is divided into three parts : @xmath2 is @xmath70-free ( lemma [ lem : p5oppo ] ) , @xmath2 is @xmath203-free but contains an induced @xmath70 ( lemma [ lem : h1oppo ] ) , and @xmath2 contains an induced @xmath203 ( lemma [ lem : final ] ) .
see figure [ fig : h1 ] for the graph @xmath203 .
= [ draw , circle , inner sep=1.5pt ] ( a ) at ( 0,1 ) [ label = above:@xmath204 ; ( b ) at ( 1,1 ) [ label = above:@xmath205 ; ( v1 ) at ( 2,1 ) [ label = above:@xmath206 ; ( v1a ) at ( 3,1 ) [ label = above:@xmath207 ; ( v1b ) at ( 4,1 ) [ label = above:@xmath208 ; ( c ) at ( 2,0 ) [ label = right:@xmath209 ; \(a ) ( b ) ( v1 )
( v1a ) ( v1b ) ; ( v1 )
( c ) ; [ lem : p5oppo ] every @xmath70-free ptolemaic graph is an opposition graph .
a short proof uses the stronger result of @xcite , which implies that every @xmath70-free ptolemaic graph is bipolarizable .
now , let @xmath2 be a ptolemaic graph , and let @xmath210 be a vertex of @xmath2 . for each integer @xmath211 , let @xmath212 be the set of all vertices at distance
exactly @xmath117 from @xmath210
. notice that @xmath213 .
we will make use of the following facts .
[ obs ] let @xmath2 be a ptolemaic graph .
let @xmath214 , and let @xmath3 be a connected component of @xmath215 $ ] .
then @xmath216 is a clique and every vertex in @xmath216 is adjacent to all vertices in @xmath3 .
we first show that @xmath217 to see ( [ eq:2 ] ) , suppose , by contradiction , @xmath218 are two non - adjacent vertices in @xmath216 , and let @xmath121 and @xmath219 be a shortest path in @xmath2 connecting @xmath210 and @xmath13 , @xmath210 and @xmath14 , respectively . consider a neighbor @xmath220 of @xmath13 and a neighbor @xmath221 of @xmath14 with smallest distance in @xmath3 , and let @xmath222 be a shortest path in @xmath3 connecting @xmath223 and @xmath224
. then @xmath225 $ ] contains an induced cycle ( containing @xmath226 and @xmath14 ) of length at least four , contradicting the chordality of @xmath2 , hence ( [ eq:2 ] ) .
next we show , by induction on @xmath117 , that @xmath227 the case @xmath228 is obvious .
let @xmath229 , and suppose that some @xmath230 is non - adjacent to some vertex in @xmath3 .
then , since @xmath3 is connected , @xmath13 is adjacent to a vertex @xmath5 and non - adjacent to a vertex @xmath4 of @xmath3 such that @xmath231 is an edge .
let @xmath232 be a neighbor of @xmath4 .
by ( [ eq:2 ] ) , @xmath12 is an edge .
hence , by induction , there is some vertex @xmath233 adjacent to all vertices of the connected component in @xmath234 $ ] containing @xmath13 and @xmath14 .
but then @xmath235 $ ] is a house or gem , a contradiction , hence ( [ eq:3 ] ) .
it follows from observation [ obs ] that , for any @xmath211 , @xmath215 $ ] is @xmath16-free , otherwise @xmath2 would contain an induced gem .
let @xmath121 be an induced @xmath16 in @xmath2 .
if @xmath121 has three vertices in @xmath212 , then , by applying observation [ obs ] to the connected components of @xmath215 $ ] containing vertices of @xmath121 , we conclude that one end - vertex of @xmath121 must belong to @xmath236 .
similarly , if @xmath121 has exactly two vertices in @xmath212 , then ( i ) the mid - edge of @xmath121 is in @xmath237 and the end - vertices of @xmath121 are in @xmath236 , or ( ii ) an end - edge of @xmath121 is in @xmath212 , one mid - vertex of @xmath121 is in @xmath236 , and one end - vertex of @xmath121 is in @xmath238 , or ( iii ) one mid - vertex of @xmath121 is in @xmath239 and one end - vertex of @xmath121 is in @xmath236 .
thus , any induced @xmath9 of @xmath2 is of exactly one of the following five types for some index @xmath211 ( up to re - labeling of the vertices ; see also figure [ fig : types ] ) : * @xmath240 , and @xmath241 .
* @xmath242 , and @xmath243 . *
@xmath244 , and @xmath245 .
* @xmath246 , and @xmath245 . * @xmath247 , and @xmath248 .
= [ draw , circle , inner sep=1.55pt ] ( -2.0cm,-1.0 cm ) grid ( 27.cm,9.0cm ) ; ( 0,8 ) node @xmath212 ; ( 0,6 ) node @xmath236 ; ( 0,4 ) node @xmath238 ; ( 0,2 ) node @xmath249 ; ( aa ) at ( 4,8 ) [ label = right:@xmath124 ] ; ( ab ) at ( 4,6 ) [ label = right:@xmath125 ] ; ( ac ) at ( 4,4 ) [ label = right:@xmath252 ; ( ad ) at ( 4,2 ) [ label = right:@xmath253 ; ( aa ) ( ab ) ( ac ) ( ad ) ; ( 4,0 ) node ( a ) ; ( ba ) at ( 7,6 ) [ label = left:@xmath124 ] ; ( bb ) at ( 7,8 ) [ label = left:@xmath125 ] ; ( bc ) at ( 9,8 ) [ label = right:@xmath252 ; ( bd ) at ( 9,6 ) [ label = right:@xmath253 ; ( ba ) ( bb ) ( bc ) ( bd ) ; ( 8,0 ) node ( b ) ; ( ca ) at ( 12,6 ) [ label = left:@xmath124 ] ; ( cb ) at ( 13,8 ) [ label = right:@xmath125 ] ; ( cc ) at ( 14,6 ) [ label = right:@xmath252 ; ( cd ) at ( 14,4 ) [ label = right:@xmath253 ; ( ca )
( cb ) ( cc ) ( cd ) ; ( 13,0 ) node ( c ) ; ( da ) at ( 17,8 ) [ label = left:@xmath124 ] ; ( db ) at ( 19,8 ) [ label = right:@xmath125 ] ; ( dc ) at ( 19,6 ) [ label = right:@xmath252 ; ( dd ) at ( 19,4 ) [ label = right:@xmath253 ; ( da )
( dd ) ; ( 18,0 ) node ( d ) ; ( ea ) at ( 22,8 ) [ label = above:@xmath124 ] ; ( eb ) at ( 24,8 ) [ label = above:@xmath125 ] ; ( ec ) at ( 26,8 ) [ label = above:@xmath252 ; ( ed ) at ( 26,6 ) [ label = right:@xmath253 ; ( ea )
( ed ) ; ( 24,0 ) node ( e ) ; we now orient @xmath2 as follows .
first , for each edge @xmath231 with @xmath254 and @xmath255 : * if @xmath256 or @xmath257 , orient the edge @xmath231 forward , i.e. , @xmath113 is the corresponding directed edge ; * otherwise , orient the edge @xmath231 backward , i.e. , @xmath112 is the corresponding directed edge . clearly , in this partial orientation of @xmath2 , any induced @xmath16 of type ( a ) or ( b ) is good , i.e. , the orientation puts the two end - edges in opposition .
next , we orient the remaining edges as follows : let @xmath231 be an edge such that @xmath258 for some @xmath214 . if @xmath231 is not an end - edge of any induced @xmath16 , orient @xmath231 arbitrarily .
if @xmath231 is an end - edge of an induced @xmath16 , say @xmath121 , then @xmath121 must be of type ( d ) or ( e ) . in this case
, @xmath231 gets the direction that is in opposition to the direction of the other end - edge of @xmath121 . in the proofs of lemmas
[ lem : h1oppo ] and [ lem : final ] below , the vertex @xmath210 will be chosen appropriately , and we will see that this orientation is well - defined , i.e. , the other induced @xmath16s of types ( d ) and ( e ) are good , too .
we will also see that any induced @xmath16 of type ( c ) is also good , provided @xmath2 is a @xmath185-free @xmath195 and @xmath199-free ptolemaic graph .
[ lem : h1oppo ] let @xmath2 be a @xmath259-free ptolemaic graph containing an induced @xmath70 . then @xmath2 is an opposition graph .
let @xmath210 be the midpoint of an induced @xmath70 , say @xmath260 .
write @xmath261 for @xmath212 .
we first show that any @xmath262 of type ( c ) is good .
let @xmath211 such that @xmath263 , and @xmath264
. then @xmath265 , otherwise there would exist vertices @xmath266 such that @xmath267 and @xmath14 induce an @xmath203 .
assume @xmath228 .
let @xmath268 .
if @xmath4 is non - adjacent to @xmath269 , then @xmath270 are also non - adjacent to @xmath269 ( as @xmath2 is chordal ) , hence @xmath271 and @xmath269 induce an @xmath203 .
thus , @xmath4 is adjacent to @xmath269 , implying @xmath272 and @xmath4 is adjacent to both @xmath273 .
then , @xmath270 are non - adjacent to @xmath273 ( as @xmath2 is gem - free chordal ) .
this implies that @xmath274 ; moreover , since @xmath2 is chordal , @xmath275 is non - adjacent to @xmath276 .
hence @xmath277 and @xmath14 induce an @xmath203 .
so , @xmath278 , and @xmath279 .
thus we have @xmath112 and @xmath105 , i.e. , the @xmath262 is good .
next , we show that the orientation for edges @xmath231 with @xmath280 for some @xmath214 is well defined .
let @xmath231 be an end - edge of an induced @xmath16 , say @xmath281 . if @xmath121 is the only @xmath16 containing @xmath231 , then the orientation of @xmath231 , implied by the one of @xmath12 , is well defined
otherwise , we have two cases below .
let us assume that @xmath12 is oriented by @xmath105 .
case 1 : @xmath121 is of type ( e ) . in this case ,
@xmath282 and @xmath283
. + by observation [ obs ] , there is a vertex @xmath284 adjacent to @xmath285 and @xmath13 .
thus , @xmath286 is an induced @xmath16 of type ( c ) which has been shown to be good .
in particular , @xmath287 is oriented @xmath288 . now consider another induced @xmath16 containing @xmath231 , say @xmath219 .
assume @xmath289 is of type ( d ) .
then @xmath290 is of type ( a ) , hence @xmath291 is oriented @xmath292 .
so @xmath121 and @xmath219 imply the same orientation of @xmath231 .
assume @xmath293 is of type ( d ) .
then @xmath294 and @xmath295 are not adjacent to @xmath296 ( as @xmath2 is gem - free chordal ) , and @xmath297 and @xmath14 induce an @xmath203 .
thus such a @xmath16 of type ( d ) can not exist . at this point
, @xmath219 is not of type ( d ) , and we may assume that @xmath219 is of type ( e ) . if @xmath219 is @xmath298 , then @xmath299 induce a @xmath202 ( as no @xmath16 has all vertices in @xmath261 ) which is impossible as @xmath2 is chordal .
if @xmath219 is @xmath300 , then @xmath121 and @xmath219 imply the same orientation of @xmath231 .
case 1 is settled .
case 2 : @xmath121 is of type ( d ) . in this case ,
@xmath301 and @xmath264 .
+ by observation [ obs ] , there is a vertex @xmath284 adjacent to @xmath5 and @xmath4 .
let @xmath219 be another induced @xmath16 containing @xmath231 . by case 1
, we may assume that @xmath219 is of type ( d ) , and in order to create a conflict let @xmath219 be @xmath298 .
then @xmath302 , otherwise there would exist vertices @xmath303 such that @xmath304 induce a @xmath193 ( note that , as @xmath2 is chordal , there are no edges between @xmath305 and @xmath296 ) .
we will see that the cases @xmath306 are also impossible .
. then @xmath307 .
at least one of @xmath273 is different from @xmath201 ; let @xmath308 , say . if @xmath5 or @xmath4 is non - adjacent to @xmath309 , then @xmath310 induce a @xmath193 or there is a gem or @xmath2 is not chordal .
so , @xmath5 and @xmath4 are adjacent to @xmath309 , which implies that @xmath311 ; then , similarly , @xmath285 are adjacent to @xmath312 .
now , @xmath13 is non - adjacent to @xmath309 and @xmath312 : if @xmath13 is adjacent to @xmath309 and @xmath312 , then @xmath313 $ ] is a @xmath202 ; if @xmath314 and @xmath315 , say , then @xmath316 $ ] is a gem .
thus , @xmath317 and @xmath312 induce an @xmath203 .
. then @xmath319 , and let without loss of generality that @xmath320 .
then , @xmath269 is adjacent to @xmath309 : if @xmath321 , @xmath275 would be non - adjacent to @xmath322 ( otherwise @xmath2 would contain an induced @xmath323 ) , and thus @xmath2 would contain an induced @xmath193 .
thus , @xmath324 , and by symmetry , @xmath269 is adjacent to @xmath312 .
now , @xmath5 is non - adjacent to @xmath309 and @xmath312 : if @xmath5 is adjacent to both @xmath325 , then @xmath326 $ ] is a @xmath202 ; if @xmath327 and @xmath328 , say , then @xmath329 $ ] is a gem . by symmetry
, @xmath4 is non - adjacent to @xmath325 .
moreover , @xmath275 is non - adjacent to @xmath269 ( otherwise @xmath330 $ ] would be a gem ) .
thus , @xmath331 induce a @xmath193 ( note that , as @xmath2 is chordal , @xmath275 is non - adjacent to @xmath332 ) .
case 2 is settled , and the proof of lemma [ lem : h1oppo ] is complete .
[ lem : final ] let @xmath2 be a @xmath185-free @xmath195 and @xmath199-free ptolemaic graph containing an induced @xmath203 . then @xmath2 is an opposition graph .
the main difficulty in this case is the choice of the root vertex @xmath210 in defining the sets @xmath212 as in the proof of lemma [ lem : h1oppo ] .
first , let @xmath333 and define the graphs @xmath334 and @xmath335 , @xmath336 , as follows .
@xmath334 ( respectively , @xmath335 ) is obtained from @xmath337 ( respectively , @xmath338 ) by taking three new vertices @xmath339 , and joining @xmath340 to all vertices in @xmath341 $ ] ( respectively , in @xmath341\setminus\{v_0\}$ ] ) , @xmath342 to @xmath340 , and @xmath343 to @xmath342 .
note that , for @xmath344 , @xmath338 is obtained from @xmath337 by deleting all edges @xmath345 .
see figure [ fig : hk ] .
note also that , for all @xmath336 , each of @xmath346 and @xmath338 is an opposition graph admitting exactly two opposition orientations : in the first one , @xmath347 is a source .
the second one is the reverse of the other ( in which , @xmath347 is a sink ) . in any opposition orientation of @xmath346 and of @xmath338 ,
each @xmath348 , @xmath349 , is neither a source nor a sink .
= [ draw , circle , inner sep=1.55pt , fill = gray!30 ] ; = [ draw , circle , inner sep=1.55pt ] ( v0 ) at ( 0,2 ) [ label = left:@xmath209 ; ( v0 ) at ( 2,2 ) [ label = below:@xmath350 ; ( v0 ) at ( 2,0 ) [ label = below:@xmath351 ; ( v1 ) at ( 4,2 ) [ label = below:@xmath206 ; ( v1 ) at ( 6,2 ) [ label = below:@xmath352 ; ( v1 ) at ( 6,0 ) [ label = below:@xmath353 ; ( v2 ) at ( 4,4.7 ) [ label = above:@xmath354 ; ( v2 ) at ( 8,2 ) [ label = below:@xmath355 ; ( v2 ) at ( 8,0 ) [ label = below:@xmath356 ; ( v2 ) ( v0 ) ; ( v2 ) ( v0 ) ; ( v2 ) ( v1 ) ; ( v2 ) ( v1 ) ; ( v2 ) ( v2 ) ; ( v0 ) ( v0 ) ( v1 ) ( v1 ) ( v1 ) ; ( v2 ) ( v2 ) ; ( v0 ) to[bend angle=20 , bend left ] ( v1 ) ; ( 6,4.3 ) node @xmath357 ; = [ draw , circle , inner sep=1.55pt , fill = gray!30 ] ; = [ draw , circle , inner sep=1.55pt ] ( v0 ) at ( 2,0 ) [ label = below:@xmath358 ; ( v0 ) at ( 0,2 ) [ label = below:@xmath350 ; ( v0 ) at ( 0,0 ) [ label = below:@xmath351 ; ( v1 ) at ( 2,2 ) [ label = below:@xmath359 ; ( v1 ) at ( 4,2 ) [ label = below:@xmath352 ; ( v1 ) at ( 4,0 ) [ label = below:@xmath353 ; ( v2 ) at ( 3,4.7 ) [ label = above:@xmath354 ; ( v2 ) at ( 6,2 ) [ label = below:@xmath355 ; ( v2 ) at ( 6,0 ) [ label = below:@xmath356 ; ( v2 ) ( v0 ) ; ( v2 ) ( v1 ) ; ( v2 ) ( v1 ) ; ( v2 ) ( v2 ) ; ( v0 ) ( v0 ) ( v1 ) ( v1 )
( v1 ) ; ( v0 ) ( v1 ) ; ( v2 )
( v2 ) ; ( 5,4.3 ) node @xmath360 ; = [ draw , circle , inner sep=1.55pt , fill = gray!30 ] ; = [ draw , circle , inner sep=1.55pt ] ( v0 ) at ( 0,2 ) [ label = left:@xmath209 ; ( v0 ) at ( 2,2 ) [ label = below:@xmath350 ; ( v0 ) at ( 2,0 ) [ label = below:@xmath351 ; ( v1 ) at ( 4,2 ) [ label = below:@xmath206 ; ( v1 ) at ( 6,2 ) [ label = below:@xmath352 ; ( v1 ) at ( 6,0 ) [ label = below:@xmath353 ; ( v2 ) at ( 8,2 ) [ label = below:@xmath354 ; ( v2 ) at ( 10,2 ) [ label = below:@xmath355 ; ( v2 ) at ( 10,0 ) [ label = below:@xmath356 ; ( v3 ) at ( 6,5.6 ) [ label = above:@xmath361 ; ( v3 ) at ( 12,2 ) [ label = below:@xmath362 ; ( v3 ) at ( 12,0 ) [ label = below:@xmath363 ; ( v3 ) ( v0 ) ; ( v3 ) ( v0 ) ; ( v3 ) ( v1 ) ; ( v3 ) ( v1 ) ; ( v3 ) ( v2 ) ; ( v3 ) ( v2 ) ; ( v3 ) ( v3 ) ( v3 ) ; ( v0 ) ( v0 ) ( v1 ) ( v1 ) ( v2 ) ( v2 ) ( v2 ) ; ( v1 ) ( v1 ) ; ( v0 ) to[bend angle=20 , bend left ] ( v1 ) ; ( v0 ) to[bend angle=25 , bend left ] ( v2 ) ; ( v0 ) to[bend angle=22.5 , bend left ] ( v2 ) ; ( v1 ) to[bend angle=20 , bend left ] ( v2 ) ; ( 10,5 ) node @xmath364 ; = [ draw , circle , inner sep=1.55pt , fill = gray!30 ] ; = [ draw , circle , inner sep=1.55pt ] ( v0 ) at ( 2,0 ) [ label = below:@xmath358 ; ( v0 ) at ( 0,2 ) [ label = below:@xmath350 ; ( v0 ) at ( 0,0 ) [ label = below:@xmath351 ; ( v1 ) at ( 2,2 ) [ label = below:@xmath359 ; ( v1 ) at ( 4,2 ) [ label = below:@xmath352 ; ( v1 ) at ( 4,0 ) [ label = below:@xmath353 ; ( v2 ) at ( 6,2 ) [ label = below:@xmath354 ; ( v2 ) at ( 8,2 ) [ label = below:@xmath355 ; ( v2 ) at ( 8,0 ) [ label = below:@xmath356 ; ( v3 ) at ( 5,5.6 ) [ label = above:@xmath361 ; ( v3 ) at ( 10,2 ) [ label = below:@xmath362 ; ( v3 ) at ( 10,0 ) [ label = below:@xmath363 ; ( v3 ) ( v0 ) ; ( v3 ) ( v1 ) ; ( v3 ) ( v1 ) ; ( v3 ) ( v2 ) ; ( v3 ) ( v2 ) ; ( v3 ) ( v3 )
( v3 ) ; ( v0 ) ( v0 ) ( v1 ) ( v1 ) ( v2 ) ( v2 ) ( v2 ) ; ( v0 ) ( v1 ) ; ( v1 ) ( v1 ) ; ( v0 ) to[bend angle=22.5 , bend left ] ( v2 ) ; ( v1 ) to[bend angle=20 , bend left ] ( v2 ) ; ( 8,5 ) node @xmath365 ; = [ draw , circle , inner sep=1.55pt , fill = gray!30 ] ; = [ draw , circle , inner sep=1.55pt ] ( v0 ) at ( 0,2 ) [ label = left:@xmath209 ; ( v0 ) at ( 2,2 ) [ label = below:@xmath366 ; ( v0 ) at ( 2,0 ) [ label = below:@xmath367 ; ( v1 ) at ( 4,2 ) [ label = below:@xmath206 ; ( v1 ) at ( 6,2 ) [ label = below:@xmath368 ; ( v1 ) at ( 6,0 ) [ label = below:@xmath369 ; ( v2 ) at ( 8,2 ) [ label = below:@xmath354 ; ( v2 ) at ( 10,2 ) [ label = below:@xmath370 ; ( v2 ) at ( 10,0 ) [ label = below:@xmath371 ; ( v3 ) at ( 12,2 ) [ label = below:@xmath361 ; ( v3 ) at ( 14,2 ) [ label = below:@xmath372 ; ( v3 ) at ( 14,0 ) [ label = below:@xmath373 ; ( v4 ) at ( 8,7.5 ) [ label = above:@xmath374 ; ( v4 ) at ( 16,2 ) [ label = below:@xmath375 ; ( v4 ) at ( 16,0 ) [ label = below:@xmath376 ; ( v4 ) ( v0 ) ; ( v4 ) ( v0 ) ; ( v4 ) ( v1 ) ; ( v4 ) ( v1 ) ; ( v4 ) ( v2 ) ; ( v4 ) ( v2 ) ; ( v4 ) ( v3 ) ; ( v4 ) ( v3 ) ; ( v4 ) ( v4 )
( v4 ) ; ( v0 ) ( v0 )
( v1 ) ( v1 ) ( v2 ) ( v2 ) ( v3 )
( v3 ) ; ( v1 )
( v1 ) ; ( v2 )
( v2 ) ; ( v0 ) to[bend angle=20 , bend left ] ( v1 ) ; ( v0 ) to[bend angle=25 , bend left ] ( v2 ) ; ( v0 ) to[bend left ] ( v3 ) ; ( v0 ) to[bend angle=22.5 , bend left ] ( v2 ) ; ( v0 ) to[bend angle=27.5 , bend left ] ( v3 ) ; ( v1 ) to[bend angle=20 , bend left ] ( v2 ) ; ( v1 ) to[bend angle=25 , bend left ] ( v3 ) ; ( v1 ) to[bend angle=22.5 , bend left ] ( v3 ) ; ( v2 ) to[bend angle=20 , bend left ] ( v3 ) ; ( 13,7 ) node @xmath377 ; = [ draw , circle , inner sep=1.55pt , fill = gray!30 ] ; = [ draw , circle , inner sep=1.55pt ] ( v0 ) at ( 2,0 ) [ label = below:@xmath378 ; ( v0 ) at ( 0,2 ) [ label = below:@xmath366 ; ( v0 ) at ( 0,0 ) [ label = below:@xmath367 ; ( v1 ) at ( 2,2 ) [ label = below:@xmath379 ; ( v1 ) at ( 4,2 ) [ label = below:@xmath368 ; ( v1 ) at ( 4,0 ) [ label = below:@xmath369 ; ( v2 ) at ( 6,2 ) [ label = below:@xmath354 ; ( v2 ) at ( 8,2 ) [ label = below:@xmath370 ; ( v2 ) at ( 8,0 ) [ label = below:@xmath371 ; ( v3 ) at ( 10,2 ) [ label = below:@xmath361 ; ( v3 ) at ( 12,2 ) [ label = below:@xmath372 ; ( v3 ) at ( 12,0 ) [ label = below:@xmath373 ; ( v4 ) at ( 7,6.5 ) [ label = above:@xmath374 ; ( v4 ) at ( 14,2 ) [ label = below:@xmath375 ; ( v4 ) at ( 14,0 ) [ label = below:@xmath376 ; ( v4 ) ( v0 ) ; ( v4 ) ( v1 ) ; ( v4 ) ( v1 ) ; ( v4 ) ( v2 ) ; ( v4 ) ( v2 ) ; ( v4 ) ( v3 ) ; ( v4 ) ( v3 ) ; ( v4 ) ( v4 )
( v4 ) ; ( v0 ) ( v0 ) ( v1 ) ( v1 ) ( v2 ) ( v2 ) ( v3 ) ( v3 )
( v3 ) ; ( v1 )
( v1 ) ; ( v2 )
( v2 ) ; ( v1 ) ( v0 ) ; ( v0 ) to[bend angle=22.5 , bend left ] ( v2 ) ; ( v0 ) to[bend angle=27.5 , bend left ] ( v3 ) ; ( v1 ) to[bend angle=20 , bend left ] ( v2 ) ; ( v1 ) to[bend angle=25 , bend left ] ( v3 ) ; ( v1 ) to[bend angle=22.5 , bend left ] ( v3 ) ; ( v2 ) to[bend angle=20 , bend left ] ( v3 ) ; ( 11,6 ) node @xmath380 ; now , let @xmath336 be the largest integer such that @xmath2 contains an induced @xmath381 or an induced @xmath382 .
then let @xmath210 be the vertex @xmath383 of such an @xmath3 .
again , write @xmath261 for @xmath212 .
we first show : [ obs2 ] let @xmath275 be a vertex in @xmath384 such that there exists a vertex @xmath385 and a vertex @xmath386 .
then @xmath275 is non - adjacent to a vertex @xmath387 for some @xmath388 .
_ proof of observation [ obs2 ] : _ note that @xmath389 is the only vertex in @xmath3 adjacent to all @xmath390 .
hence the statement of the observation is trivially true in case @xmath391 .
so , let @xmath392 , and suppose the contrary that @xmath275 is adjacent to all @xmath390 .
then @xmath393 for , if @xmath275 was non - adjacent to a vertex @xmath394 for some @xmath395 , then @xmath396 would induce a @xmath202 , and @xmath397 for , if @xmath275 was adjacent to a vertex @xmath398 for some @xmath399 , then @xmath400 ( if @xmath401 ) or @xmath402 ( if @xmath403 ) would induce a gem . then , @xmath404 since @xmath405 .
moreover , @xmath406 for , if @xmath309 was adjacent to a vertex @xmath394 for some @xmath407 , then @xmath309 would be non - adjacent to @xmath408 ( otherwise @xmath409 $ ] would be a @xmath202 ) and @xmath410 $ ] would be a gem , and @xmath411 for , if @xmath309 was adjacent to @xmath412 for some @xmath395 , then @xmath309 would be non - adjacent to @xmath413 ( otherwise @xmath414 $ ] would be a @xmath202 ) and @xmath415 , @xmath416 $ ] would be a gem .
it follows that @xmath309 is non - adjacent to @xmath417 , otherwise @xmath418 $ ] would be a gem .
furthermore @xmath419 for , if @xmath309 was adjacent to @xmath398 for some @xmath399 , then @xmath420 and @xmath275 would induce a @xmath202 .
similarly , we can show that @xmath421 is non - adjacent to all @xmath422 , and that , in case @xmath382 , @xmath421 is also non - adjacent to @xmath423 .
it follows that @xmath424 we now distinguish two cases .
suppose @xmath425 . in this case ,
@xmath3 and @xmath426 induce an @xmath357 ( if @xmath275 is adjacent to @xmath423 ) or an @xmath360 ( otherwise ) , contradicting the choice of @xmath425 .
suppose @xmath344 . in this case , if @xmath381 , then @xmath275 must be adjacent to @xmath423 , otherwise @xmath427 and @xmath347 would induce a gem .
but then @xmath3 and @xmath428 would induce an @xmath334 , contradicting the choice of @xmath429 .
if @xmath382 , then @xmath275 must be non - adjacent to @xmath423 , otherwise @xmath430 and @xmath275 would induce a gem .
but then @xmath3 and @xmath428 would induce an @xmath335 , a contradiction again .
thus , @xmath275 can not be adjacent to all @xmath390 , and observation [ obs2 ] is proved .
we next show that any induced @xmath16 of type ( c ) is good .
this will be implied by the definition of the orientation and the following fact .
[ typec ] let @xmath281 be an induced @xmath16 of type ( c ) and @xmath211 such that @xmath263 , and @xmath264 .
then @xmath117 is even .
_ proof of observation [ typec ] : _ suppose , for a contradiction , that @xmath117 is odd .
let @xmath431 be a chordless path of odd length @xmath117 , connecting @xmath432 and @xmath433 ( thus , @xmath434 for all @xmath435 ) .
set @xmath436 .
consider first the case @xmath437 .
suppose @xmath438 for some @xmath439 .
then , since @xmath440 and @xmath441 for all @xmath442 , @xmath404 . moreover , there are no edges between @xmath443 and @xmath444 ( otherwise @xmath445 $ ] would contain a @xmath446 , or a @xmath71 ) .
now , if @xmath401 , then @xmath447 induce a @xmath448 .
if @xmath403 , then there are no edges between @xmath449 and @xmath450 ( otherwise @xmath451 $ ] would contain a gem , a @xmath202 or a @xmath131 ) , hence @xmath452 induce a @xmath194 . in any case we have a contradiction .
suppose @xmath453 for some @xmath399 . then , @xmath454 ( as @xmath440 and @xmath455 for all @xmath456 ) , and there are no edges between @xmath449 and @xmath457 for any @xmath458 ( otherwise @xmath459 $ ] would contain a @xmath446 or a @xmath71 ) .
thus , if @xmath344 , then @xmath460 induce a @xmath448 for any pair @xmath461 . in case @xmath425 ,
@xmath462 induce a @xmath448 ( @xmath309 and @xmath421 are non - adjacent to @xmath423 because @xmath2 is chordal ) .
again , in any case , we have a contradiction .
so , we may assume that @xmath463 .
let @xmath464 be the set of all vertices in @xmath465 such that @xmath275 is non - adjacent to all vertices in @xmath464 . by observation [ obs2 ] , @xmath466 .
set @xmath467 .
note that @xmath468 ( if @xmath469 for some @xmath470 , then @xmath471 $ ] is a @xmath202 ) , and there are no edges between @xmath449 and @xmath472 ( otherwise @xmath473 $ ] would contain a @xmath446 or a @xmath71 ) .
thus , if @xmath474 , then @xmath475 induce a @xmath448 for any @xmath458 with @xmath476 , a contradiction .
so , let @xmath477 for some @xmath388 .
suppose @xmath344 , and let @xmath478 .
then , since @xmath275 is adjacent to @xmath479 for all @xmath480 , @xmath309 is non - adjacent to @xmath481 and @xmath482 ( otherwise @xmath483 $ ] would contain a @xmath202 or a gem ) , implying @xmath484 .
furthermore , @xmath449 are non - adjacent to @xmath485 ( otherwise @xmath486 $ ] would contain a @xmath202 or @xmath131 ) . thus , @xmath487 induce a @xmath194 , a contradiction .
so , let @xmath425 , and assume without loss of generality that @xmath488 . now ,
if @xmath275 is non - adjacent to @xmath423 , then @xmath449 are non - adjacent to @xmath423 ( otherwise @xmath489 $ ] would contain a @xmath202 or @xmath131 ) , and @xmath490 induce a @xmath448 .
if @xmath275 is adjacent to @xmath423 , then @xmath309 is non - adjacent to @xmath491 ( otherwise @xmath492 $ ] would contain a @xmath202 or a gem ) , hence @xmath493 , and @xmath309 and @xmath421 are non - adjacent to @xmath494 ( otherwise @xmath495 $ ] would contain a @xmath202 or @xmath131 ) .
thus , @xmath496 together with @xmath497 induce a @xmath194 . in any case
, we have a contradiction .
thus , @xmath117 must be even , and the proof of observation [ typec ] is complete .
it follows from the definition of the orientation and observation [ typec ] that all induced @xmath16s of type ( c ) are good .
we finally show that the orientation for edges @xmath231 with @xmath280 for some @xmath214 is well defined ( that is , all other induced @xmath16s of types ( d ) and ( e ) are good ) .
let @xmath231 be an end - edge of an induced @xmath16 , say @xmath281 . if @xmath121 is the only @xmath16 containing @xmath231 , then the orientation of @xmath231 , implied by the one of @xmath12 , is well defined
otherwise , we have two cases below .
let us assume that @xmath12 is oriented by @xmath105 .
case 1 : @xmath121 is of type ( e ) . in this case ,
@xmath282 and @xmath283 .
+ by observation [ obs ] , there is a vertex @xmath284 adjacent to @xmath285 and @xmath13 .
thus , @xmath286 is an induced @xmath16 of type ( c ) which has been shown to be good . in particular , @xmath287 is oriented @xmath288 .
now consider another induced @xmath16 containing @xmath231 , say @xmath219 .
assume @xmath289 is of type ( d ) .
then @xmath290 is of type ( a ) , hence @xmath291 is oriented @xmath292 .
so @xmath121 and @xmath219 imply the same orientation of @xmath231 .
assume @xmath293 is of type ( d ) .
note that there are no edges between @xmath498 and @xmath83 ( as @xmath2 is gem - free chordal ) .
hence @xmath499 , otherwise @xmath500 , a vertex @xmath501 and @xmath502 would together induce a @xmath194 .
moreover , observation [ typec ] ( for the @xmath503 ) implies that @xmath228 .
thus , @xmath504 . by observation [ obs2 ] ,
@xmath4 is non - adjacent to a vertex @xmath505 for some @xmath388 .
then @xmath506 and @xmath507 $ ] is a @xmath194 or contains an induced gem or an induced cycle of length at least @xmath0 , a contradiction .
so , @xmath219 is not of type ( d ) , and we may assume that @xmath219 is of type ( e ) . if @xmath219 is @xmath298 , then @xmath299 induce a @xmath202 ( as no @xmath16 has all vertices in @xmath261 ) which is impossible as @xmath2 is chordal .
if @xmath219 is @xmath300 , then @xmath121 and @xmath219 imply the same orientation of @xmath231 .
case 1 is settled .
case 2 : @xmath121 is of type ( d ) . in this case ,
@xmath301 and @xmath264 .
+ by observation [ obs ] , there is a vertex @xmath284 adjacent to @xmath5 and @xmath4 .
let @xmath219 be another induced @xmath16 containing @xmath231 . by case 1
, we may assume that @xmath219 is of type ( d ) .
let @xmath219 be @xmath298 . then , as @xmath2 is chordal , @xmath508 and there are no edges between @xmath305 and @xmath296 .
moreover , @xmath302 , otherwise there would exist vertices @xmath303 such that @xmath304 induce a @xmath193 .
we will see that the cases @xmath306 are also impossible .
. then @xmath504 . by observation [ obs2 ] ,
@xmath5 is non - adjacent to some @xmath509 and @xmath4 is non - adjacent to some @xmath510 in @xmath3 . if @xmath511 , then @xmath512 $ ] is a @xmath193 ( note that , as @xmath2 is chordal , @xmath513 and there are no edges between @xmath514 and @xmath515 ) .
so , we may assume that @xmath516 and @xmath5 is adjacent to @xmath510 , @xmath4 is adjacent to @xmath509 .
but then @xmath517 induce a gem , a contradiction .
. then @xmath518 , and by observation [ obs2 ] , @xmath269 is non - adjacent to a vertex @xmath519 in @xmath3 . as @xmath2 is chordal , @xmath519 is non - adjacent to @xmath285 . but
then @xmath520 $ ] is a @xmath193 , a contradiction .
case 2 is settled , and the proof of lemma [ lem : final ] is completed .
a tree is an opposition graph if and only if it is @xmath185-free , @xmath336 .
if we define generalized coalition graphs by omitting the condition acyclic in the definition [ def : coali ] , we unfortunately lose the perfectness of the graphs we obtain , as any cycle admits a generalized coalition orientation .
nevertheless , we do define a similar auxiliary graph , and it turns out that this is useful when discussing certain _ hole - free _ coalition graphs . given a graph @xmath75 , the graph @xmath521 has * @xmath77 as vertex set , * two vertices @xmath78 and @xmath79 are adjacent in @xmath521 if @xmath80 , or else @xmath522 or @xmath523 is an induced @xmath16 in @xmath2 . as an example , if @xmath2 is the graph @xmath524 depicted in figure [ fig : n ] , then @xmath521 is the complement of @xmath71 ( the graph on the left side in figure [ fig : genoppo ] ) .
= [ draw , circle , inner sep=1.5pt ] ( 1 ) at ( 0,1 ) ; ( 2 ) at ( 1,1 ) ; ( 3 ) at ( 2,1 ) ; ( 4 ) at ( 3,1 ) ; ( 5 ) at ( 1.5,2 ) ; ( 6 ) at ( 1.5,3 ) ; \(1 ) ( 2 ) ( 3 ) ( 4 ) ; ( 2 ) ( 5 ) ( 3 ) ; ( 5 ) ( 6 ) ; ( 1.5,.3 ) node @xmath524 ; now it can be shown ( cf . theorem [ thm : genoppo ] ) that @xmath2 is a generalized coalition graph if and only if @xmath521 is bipartite . moreover , we have the following theorem , similar to theorem [ thm : gemhouse ] about @xmath30-free opposition graphs . [ thm : gemhouseholecoali ]
let @xmath2 be @xmath31-free .
then @xmath2 is a coalition graph if and only if @xmath521 is bipartite . for the non - trivial direction ,
let @xmath521 be bipartite and consider a bipartition @xmath525 of @xmath521 .
following the proof of theorem [ thm : gemhouse ] , let @xmath107 be the ( partial ) orientation of @xmath2 such that an edge @xmath83 of @xmath2 that is an end - edge of an induced @xmath16 is oriented @xmath105 in @xmath107 if @xmath106 , otherwise @xmath110 . recall that any induced @xmath111 in @xmath2 is good
: it has @xmath105 and @xmath113 in @xmath107 or else @xmath63 and @xmath112 in @xmath107 . as in the proof of theorem [ thm : gemhouse ] , we will show that @xmath107 is acyclic .
assume , by contradiction , that @xmath114 is a directed cycle in @xmath107 .
note that , since @xmath115 , by definition of @xmath107 , @xmath116 for all @xmath117 .
choose such a cycle @xmath118 with minimum length @xmath119 .
then , @xmath526 otherwise there would be a shorter directed cycle in @xmath107 .
let @xmath121 be an induced @xmath16 containing @xmath122 as an end - edge .
let , without loss of generality , @xmath121 be @xmath123 for some vertices @xmath124 and @xmath125 . since @xmath126 , @xmath527 , and hence @xmath528 .
note that , as @xmath529 and @xmath530 are adjacent , @xmath531 .
moreover , @xmath532 ( otherwise @xmath533 $ ] would be a @xmath131 , a house or a gem . ) if @xmath139 , then @xmath138 ( if @xmath124 and @xmath534 are non - adjacent ) or @xmath535 ( if @xmath124 and @xmath534 are adjacent ) is a bad @xmath16 , a contradiction .
if @xmath148 , then @xmath124 and @xmath536 are non - adjacent ( otherwise @xmath537 would be a bad @xmath16 ) .
then , @xmath538
. moreover , @xmath539 and @xmath536 are non - adjacent ( otherwise @xmath540 would be a chord of @xmath118 that is an end - edge of the induced @xmath541 , contradicting ( [ eq : coalition ] ) ) , and @xmath124 and @xmath534 are non - adjacent ( otherwise @xmath542 would be a house or a gem ) .
but then @xmath543 or @xmath544 is a bad @xmath16 , a contradiction .
so , let us assume that @xmath545 .
then , as @xmath2 is hole - free , @xmath118 has a chord connecting @xmath348 and @xmath546 ( a _ @xmath19-chord _ ) or @xmath348 and @xmath547 ( a _ @xmath0-chord _ ) for some @xmath117 ( modulo @xmath548 ) .
we distinguish two cases .
case 1 : @xmath118 has a @xmath19-chord .
we may assume that @xmath549 is a @xmath19-chord of @xmath118 .
then @xmath529 and @xmath124 are adjacent ( otherwise the chord @xmath549 of @xmath118 would be an end - edge of the induced @xmath142 , contradicting ( [ eq : coalition ] ) ) .
but now @xmath550 and @xmath125 induce a bad @xmath16 . case 1 is setteld .
case 2 : @xmath118 has no @xmath19-chord . in this case , we may assume that @xmath551 is a @xmath0-chord of @xmath118 ( and notice that @xmath348 and @xmath546 are non - adjacent for all @xmath117 ) .
then @xmath529 and @xmath124 are non - adjacent ( otherwise @xmath550 and @xmath125 would induce a bad @xmath16 ) .
then , @xmath552
. moreover , @xmath124 and @xmath534 are non - adjacent ( otherwise @xmath553 $ ] would be a house ) .
but then the chord @xmath551 of @xmath118 would be an end - edge of the induced @xmath554 , contradicting ( [ eq : coalition ] ) .
case 2 is settled .
the proof of theorem [ thm : gemhouseholecoali ] is complete . given a @xmath31-free graph @xmath2 with @xmath29 edges , it can be recognized in @xmath555 time if @xmath2 is a coalition graph , and if so , a coalition orientation of @xmath2 can be obtained in the same time .
for the smaller class of distance - hereditary graphs , we can do better .
note that the distance - hereditary graph @xmath524 depicted in figure [ fig : n ] is a minimal non - coalition graph .
[ thm : dhcoali ] the following statements are equivalent for a distance - hereditary graph @xmath2 : * @xmath2 is a coalition graph ; * @xmath521 is bipartite ; * @xmath2 is @xmath524-free @xmath556 ; * @xmath2 is a comparability graph .
\(i ) @xmath109 ( ii ) follows from theorem [ thm : gemhouseholecoali ] .
the implications ( ii ) @xmath109 ( iii ) and ( iv ) @xmath109 ( i ) are obvious ( recall that comparability graphs are coalition graphs ) .
the implication ( iii ) @xmath109 ( iv ) has been proved in ( * ? ? ? * theorem 3.1 ) .
we remark that ( iii ) @xmath109 ( iv ) can also be proved by induction on the number of vertices of @xmath2 as follows .
let @xmath2 be an @xmath524-free distance - hereditary graph .
then @xmath2 has a pair of twins or a pendant vertex ( @xcite ) .
if @xmath296 are twins , then by induction , @xmath557 admits a transitive orientation . adopting the directions of the @xmath14-edges for the @xmath13-edge and orienting the edge @xmath12 ( if any ) arbitrarily
, we obtain a transitive orientation for @xmath2 .
so , let @xmath2 have a pendant vertex @xmath5 and let @xmath4 be the neighbor of @xmath5 . by induction
again , @xmath42 is a comparability graph .
as @xmath2 has no gem , no house , and no @xmath524 , @xmath4 is a regular vertex of @xmath42 in the sense of @xcite .
it follows from ( * ? ? ?
* theorem 1 ) that @xmath42 admits a transitive orientation in which @xmath4 is a source .
then orienting @xmath231 from @xmath4 to @xmath5 we obtain a transitive orientation for @xmath2 .
hence ( iv ) . theorem
[ thm : dhcoali ] implies that distance - hereditary coalition graphs can be recognized in time needed to recognize comparability graphs which requires time proportional to matrix multiplication ( see @xcite for more information ) .
however , ( * ? ? ?
* theorem 4.14 ) provides linear time recognition algorithm for distance - hereditary comparability graphs .
thus , we have given a distance - hereditary graph @xmath2 with @xmath28 vertices and @xmath29 edges , it can be recognized in @xmath558 time if @xmath2 is a coalition graph , and if so , a coalition orientation of @xmath2 can be obtained in the same time .
a distance - hereditary graph is a comparability graph if and only if it is @xmath524-free .
recognizing and characterizing coalition graphs and opposition graphs remain long - standing open problems in structural and algorithmic graph theory . in this paper
we have characterized generalized opposition graphs , and our characterization leads to a polynomial time recognition algorithm .
we also give algorithmic good characterizations , as well as forbidden subgraph characterizations for some opposition graph classes and coalition graph classes , including distance - hereditary opposition and coalition graphs . 1 .
determine further ( perfect ) graph classes @xmath559 such that opposition graphs and generalized opposition graphs in @xmath559 coincide , i.e. , for all @xmath560 , @xmath2 is an opposition graph if and only if @xmath76 is bipartite . for such classes ,
opposition graphs can be recognized in polynomial time .
this is done for @xmath561 in @xcite and for @xmath562 in this paper .
2 . similarly , we know that if @xmath2 is a coalition graph , then @xmath521 is bipartite . for which graph classes
does the converse hold true ? in this paper , it is shown that this is the case for distance - hereditary graphs .
3 . characterize and recognize chordal opposition / coalition graphs . in this paper
we are able to solve the case of ptolemaic graphs .
characterize and recognize comparability graphs that are opposition graphs , and co - comparability graphs that are opposition / coalition graphs . in @xcite ,
the case of bipartite graphs and co - bipartite graphs are settled .
vaek chvtal , generalized opposition graphs , in : _ open problems on perfect graphs _ , 1 .
perfection of special classes of berge graphs .
http://www.cs.rutgers.edu/~chvatal/perfect/problems.html , august 24 , 2000 . | a graph is an opposition graph , respectively , a coalition graph , if it admits an acyclic orientation which puts the two end - edges of every chordless @xmath0-vertex path in opposition , respectively , in the same direction .
opposition and coalition graphs have been introduced and investigated in connection to perfect graphs . recognizing and characterizing opposition and coalition graphs are long - standing open problems .
this paper gives characterizations for opposition graphs and coalition graphs on some restricted graph classes .
implicit in our arguments are polynomial time recognition algorithms for these graphs .
we also give a good characterization for the so - called generalized opposition graphs .
perfectly orderable graph , one - in - one - out graph , opposition graph , coalition graph , perfect graph 05c75 , 05c17 , 05c62 |
the identification of nuclear effects in neutrino scattering is essential for modern neutrino oscillation experiments @xcite . in particular the sensitivity of the neutrino energy reconstruction to multi - nucleon events
has been stressed in recent data analyses @xcite . in the minerva neutrino experiment
an enhanced population of multi - proton states has been observed between the quasielastic and @xmath3 peaks .
on the other hand , observation of events with a pair of energetic protons at the interaction vertex accompanying the muon in @xmath4ar@xmath0 reaction has been reported in the argoneut experiment @xcite . from these events
several back - to - back nucleon configurations have been identified and associated with nuclear mechanisms involving short - range correlated ( src ) neutron - proton ( np ) pairs in the nucleus @xcite .
however in @xcite these `` hammer events '' have been modeled by a simple pion production and reabsorption model without nucleon - nucleon correlations , suggesting that the distribution of pp pairs in the final state is less sensitive to details of the initial pair configuration . in the opinion of the authors of @xcite ,
the events can not teach us anything significant about src .
the nuwro event generator supports that the excess of back - to - back events in argoneut has a kinematic origin and is not directly related to src @xcite .
src with back - to - back configurations have been also identified in two - nucleon knock - out electron scattering experiments on @xmath1c for high momentum transfer and missing momentum @xcite . in this case one
expects an excess of np pairs over pp pairs @xcite .
the experiment reported a number of np pairs 18 times larger than their pp counterparts .
the analysis of these experiments is compatible with theoretical single - nucleon and nucleon pair momentum distributions in variational monte carlo calculations , where the importance of the tensor forces in the ground - state correlations of nuclei has been emphasized @xcite . while the kinematics of the experiments have been selected to minimize the contribution from other mechanisms that can induce two - particle emission , such as meson - exchange currents ( mec ) and isobar excitations @xcite , the contribution of mec can not be ruled out _ a priori _ @xcite . in this work
we investigate the relative effect of mec on the separate pp and np channels in the inclusive 2p-2h neutrino cross section .
it has been emphasized that the separate charge distributions of 2p-2h events are useful .
one of the reasons is for their use in monte carlo event generators @xcite .
for instance in nuwro configurations the mec 2p-2h excitations are assumed to occur 95% of the time for events where the interaction occurs in initial np pairs @xcite ( or final pp pairs for charged current neutrino scattering ) .
this value was estimated based on the assumption , claimed also in @xcite , that neutrinos interact mostly with correlated np pairs . from a naive calculation
this value agrees with a factor 18/19 , corresponding to the extracted value of np/(np+pp ) in the @xmath1c@xmath5 experiment of @xcite .
however this neutrino generator uses a 2p-2h model that does not give separate pp and np contributions , and therefore this choice is not fully consistent from the theoretical point of view @xcite . on the other hand
it is expected that the ratio between np and pp interactions should be kinematics dependent and not only a global factor .
thus a theoretical quantification of the np / pp ratio and its dependence on the typical kinematics would be desired for each implementation of 2p-2h cross sections .
results for the separate pp and np contributions due to short - range correlations have been presented in @xcite , for the @xmath6 and @xmath7 response functions , and for @xmath8 mev / c , but not for the differential cross section .
the contribution of initial np pairs to the t response found in @xcite is about twice that of the initial nn pairs .
we have recently developed a fully relativistic model of meson - exchange currents in the 2p-2h channel for electron and neutrino scattering @xcite .
this model is an extension of the relativistic mec model of @xcite to the weak sector .
it has been recently validated by comparing to the @xmath1c@xmath9 inclusive cross section data for a wide kinematic range within the superscaling approach ( susa ) @xcite .
this model describes jointly the quasielastic and inelastic regions using two scaling functions fitted to reproduce the data , while the 2p-2h mec contribution properly fills the dip region in between , resulting in excellent global agreement with the data .
the model has been recently extended to the description of neutrino scattering reactions for a variety of experiments providing an excellent agreement with data @xcite . with this benchmark model
we are able to study the separate np and pp channels in the response functions and cross section for the three @xmath9 , @xmath10 and @xmath11 reactions . while this analysis was performed in @xcite for electron scattering , in this work we consider neutrino reactions .
our model includes the contributions of pion - in - flight , seagull , pion - pole and @xmath3(1232 ) excitation diagrams of the mec .
the two - body matrix elements between relativistic spinors were presented in our recent work @xcite , where they have been deduced from the weak pion production amplitudes of @xcite .
the formalism of 2p-2h cross section including mec in the relativistic fermi gas was given in @xcite .
we write the charged current ( cc ) ) cross section as @xmath12 , \label{sigma}\end{aligned}\ ] ] where @xmath13 is a kinematic factor including the weak couplings defined in @xcite .
note that there is a linear combination of five response functions , labeled as @xmath14 and @xmath15 .
the @xmath15 response function contributes differently for neutrinos ( plus sign ) than for antineutrinos ( minus sign ) .
the @xmath16 factors are kinematic functions that were defined in @xcite . the response functions @xmath17
depend on the energy and momentum transfer .
they are computed here in a relativistic fermi gas ( rfg ) model , with fermi momentum @xmath18 , where they can be expanded as the sum of one - particle one - hole ( 1p-1h ) , two - particle two - hole ( 2p-2h ) , plus additional channels .
here we are interested in the 2p-2h channel , where two nucleons with momenta @xmath19 and @xmath20 are ejected out of the fermi sea , @xmath21 , leaving two hole states in the daughter nucleus , with momenta @xmath22 and @xmath23 ( with @xmath24 ) .
the 2p-2h response functions are computed as @xmath25 where the momentum of the second nucleon is fixed by momentum conservation inside the integral sign , @xmath26 , @xmath27 is the volume of the system , @xmath28 is the nucleon mass , while @xmath29 and @xmath30 are the energies of the holes and particles , respectively . using energy conservation , the calculation of the inclusive 2p-2h responses of eq .
( [ hadronic ] ) for given energy and momentum transfer @xmath2 , is reduced to a seven - dimensional integral that is computed numerically following the methods developed in @xcite .
the main ingredient of the calculation is the set of five response functions @xmath31 , for the elementary 2p-2h transition .
these elementary response functions are written in terms of the two - body mec antisymmetrized matrix elements , summed over spin .
we separate the contributions of the different charge channels to the response functions .
these can be @xmath32 for neutrinos , and @xmath33 for antineutrinos .
in @xcite we derived general formulae for the separate np and pp response functions .
the total cc mec for neutrino scattering can be written as @xmath34 where @xmath35 and we have defined the isospin operators @xmath36 that stands for the @xmath37-component of the two - body isovector operator @xmath38 .
\label{isospin_seagull}\ ] ] the isospin - independent two - body currents @xmath39 , @xmath40 , and @xmath41 , follow from the amplitudes of weak pion production model of @xcite and are written in @xcite .
in other models of neutrino scattering @xcite , only the direct diagrams ( a , b ) of fig . [ selfenergy ] are included , while the direct - exchange contribution corresponding to the diagrams ( c , d ) are disregarded . in our model , on the contrary , both contributions are considered .
the elementary 2p-2h transverse response function is given for pp emission by @xmath42 where @xmath43 is the effective two - body current for pp emission with neutrinos given by @xmath44 the first term in the transverse response is the `` direct '' contribution , and the second one is the `` exchange '' contribution , actually being the interference between the direct and exchange matrix elements .
diagrams ( a , b ) represent the direct contribution .
diagrams ( c , d ) are the exchange contributions . ]
the elementary transverse response for np emission has a similar expression @xmath45 but with the effective current @xmath46 for np emission @xmath47 which is in general different from the pp one because of the distinct isospin matrix elements .
the direct contribution corresponds to neglecting the second terms in eqs .
( [ wpp],[wnp ] ) .
in the following we present results for the semi - inclusive @xmath1c@xmath48 and @xmath1c@xmath49 reactions , corresponding to the 2p-2h channel of the @xmath1c@xmath0 reaction in the two separate charge channels in the final hadronic state corresponding to two - nucleon knockout .
our mec model and its parameters were obtained from the pion production amplitudes of @xcite .
this is an extension of the electromagnetic mec model of @xcite to the weak sector .
the five mec - induced 2p-2h responses for cc neutrino interactions at fixed momentum transfer are shown in fig .
the coulomb @xmath7 and transverse @xmath6 responses are also present in electron scattering . in general the @xmath50 dependence of the 2p-2h responses
shows a broad peak coming from the @xmath3 excitation .
the strength of the mec peak weakens with @xmath51 due to the decrease of the electroweak form factor with @xmath52 , especially the @xmath3 form factors .
note also that the most important contribution to the neutrino cross section comes form the two transverse responses @xmath6 and @xmath53 . in the figure
we only show the separate pp and np 2p-2h response functions , the total responses being the sum of the two . for all the cases in fig .
[ fig1 ] we observe that the pp response functions are much larger than the np ones by a factor 6 or less depending on the kinematics . the pp / np ratio in the present neutrino calculation can be compared to the np / pp ratio in the @xmath9 reaction studied in @xcite , because they correspond to the same pairs in the initial state . for the transverse response
that ratio for neutrino scattering is roughly a factor of two smaller than for the electron case . for @xmath8 mev
/ c our results for @xmath6 can be compared to those of the src model of @xcite .
our mec response at the maximum is one order of magnitude larger than that of the src one . in our calculation
, the pp pair emission transverse response induced by mec is about a factor of 6 larger than the np one .
in contrast , the mentioned src model shows at most a factor of 3 between the two contributions .
the order of magnitude of the @xmath7 , on the other hand , is small in both mec and src 2p-2h responses , but still the mec results are about twice those of @xcite .
the pp pairs in the final state continue to dominate the @xmath7 mec response , while in the src case , both pairs contribute similarly . and @xmath15 2p-2h response functions of @xmath1c , for @xmath54 mev / c , compared to the direct contributions obtained by neglecting the direct - exchange interferences . ]
although the behaviors of mec and src 2p-2h responses are completely different , one should emphasize that the cross section is dominated by the transverse responses .
one could conclude from this comparison that the mec are the largest contribution to the 2n knockout strength , and that the influence of src is around 10% . in a previous work @xcite we showed that the interference diagrams ( c , d ) of fig .
1 can amount to @xmath55 of the total 2p-2h responses .
but for the separate charge channels the interference influence can be truly different , as we show in fig .
while the interference for pp emission produces a reduction of 20% , for np emission the reduction factor is about 1/2 .
thus the ratio pp / np critically depends on the treatment on the interference contributions .
the same conclusion was found for the electron scattering responses in @xcite .
the effect from the interference contribution is of the same size for the @xmath15 response , as shown in the lower panel of fig .
[ fig3 ] . ,
@xmath15 and @xmath56 2p-2h response functions of @xmath1c , for @xmath57 , 200 , ... , 2000 mev / c . ]
2p-2h response function of @xmath1c , for @xmath57 , 200 , ... , 2000 mev / c , including only the axial mec . ] in fig .
[ fig4 ] we compare the separate np and pp contributions to the response functions @xmath6 , @xmath53 , @xmath7 , in the full range of momentum transfer from @xmath57 to 2000 mev / c .
this corresponds to the typical kinematic range of the neutrino experiments operating in the few - gev region .
the @xmath58 response is more than twice the @xmath15 both for np and for pp channels .
this seems to indicate that the axial mec contribution is larger than the vector one .
indeed , this can be truly observed in fig .
[ fig5 ] , where only the axial mec current is included in the calculation .
note that the @xmath56 response is much smaller than the other two , except for the @xmath59 point , where the @xmath60 and @xmath61 contributions are approximately cancelled
. they would be exactly cancelled if the total current was conserved . as a matter of fact , near the photon point
the seagull current dominates the mec for high @xmath51 when the @xmath3 resonance is far away .
the axial seagull contribution is mainly longitudinal .
that is why the @xmath56 response in fig .
[ fig4 ] is large at the photon point for high @xmath51 and the @xmath58 response in fig .
[ fig5 ] is so small .
c , for fixed muon scattering angle and for three neutrino energies , as a function of the muon kinetic energy .
the separate 1p-1h and 2p-2h cross sections are displayed for comparison . ]
c , for fixed muon scattering angle and for three neutrino energies , as a function of the muon kinetic energy .
the separate np and pp channels are shown . ] to appreciate the size of the mec 2p-2h contribution , in fig .
[ fig6 ] we plot the double differential neutrino cross section per neutron , @xmath62 , of @xmath1c , as a function of the muon kinetic energy , for @xmath63 and for three values of the incident neutrino energy .
we show the separate contributions of 1p-1h and 2p-2h channels in the rfg .
the relative contribution of 2p-2h increases with the neutrino energy , and the mec and quasielastic peaks get closer . here
the neutrino energy is fixed , while in the experiments the neutrino energy is not fixed and the flux produces an average of all the contributions around the mean energy . for typical peak energies of 1 gev , the results of fig .
[ fig6 ] indicate that one can expect a contribution of roughly 20% to the cross section from 2p-2h .
c , @xmath64 , in units of @xmath65 , as a function of @xmath66 for fixed neutrino energy @xmath67 gev .
the separate pp and np channels are shown in the middle and bottom panels , respectively .
, title="fig : " ] + c , @xmath64 , in units of @xmath65 , as a function of @xmath66 for fixed neutrino energy @xmath67 gev .
the separate pp and np channels are shown in the middle and bottom panels , respectively .
, title="fig : " ] + c , @xmath64 , in units of @xmath65 , as a function of @xmath66 for fixed neutrino energy @xmath67 gev .
the separate pp and np channels are shown in the middle and bottom panels , respectively .
, title="fig : " ] the separate pp and np channels in the differential neutrino cross section are shown in fig .
[ fig7 ] for the same kinematics .
the pp channel clearly dominates the 2p-2h cross section .
the pp / np ratio is around 5 - 6 near the maximum , but its precise value depends on the kinematics . note that the np distribution is shifted towards higher muon energies compared with the pp case .
this effect can be further observed in fig .
[ fig8 ] , where we show the @xmath68 dependence of the 2p-2h double differential cross section , for @xmath67 gev .
indeed the second and third panels show the separate pp and np distributions .
the np is much smaller than the pp one , and it is clearly shifted towards higher @xmath69 and smaller angles .
it can be seen that for this neutrino energy , the absolute maximum of the cross section is located around @xmath70 0.85 and @xmath71 mev , and corresponds approximately to the maximum shown in the middle panel of fig .
the 2p-2h strength is concentrated in the top - right corner of fig .
[ fig8 ] corresponding to small angles and large muon kinetic energies , meaning low energy transfer , around @xmath72 mev .
this corresponds to the excitation energy of the @xmath73 , which gives the main contribution to the mec .
our calculation predicts that , when the lepton scattering angle increases , two particle emission implies a decrease of the kinetic energy of the muon or larger @xmath50 .
in this work we have studied the separate charge channels @xmath48 and @xmath49 , from @xmath1c , integrated over the two emitted nucleons , that contribute to the 2p-2h cross section in quasielastic - like cc neutrino scattering .
we have computed the response functions and double differential cross sections for several kinematics .
the pp channel dominates over the np contribution in the whole domain .
the pp / np ratio is about 5 - 6 for a wide range of neutrino energies .
future plans are to fold the cross section with the neutrino fluxes for the various neutrino oscillation experiments . having the separate isospin contributions will allow us to apply this formalism to asymmetric nuclei @xmath74 .
this will be of interest for neutrino experiments based , for instance , on @xmath4ar , @xmath75fe or @xmath76pb .
this work was supported by spanish direccion general de investigacion cientifica y tecnica and feder funds ( grants no .
fis2014 - 59386-p and no .
fis2014 - 53448-c2 - 1 ) , by the agencia de innovacion y desarrollo de andalucia ( grants no .
fqm225 , fqm160 ) , by infn under project manybody , and part ( twd ) by u.s .
department of energy under cooperative agreement de - fc02 - 94er40818 .
irs acknowledges support from a juan de la cierva fellowship from spanish mineco .
gdm acknowledges support from a junta de andalucia fellowship ( fqm7632 , proyectos de excelencia 2011 ) .
i. r. simo , j. e. amaro , m. b. barbaro , a. de pace , j. a. caballero and t. w. donnelly , arxiv:1604.08423 [ nucl - th ] .
a. de pace , m. nardi , w.m .
alberico , t.w .
donnelly and a. molinari , nucl .
a726 * ( 2003 ) 303 .
g. d. megias , j. e. amaro , m. b. barbaro , j. a. caballero and t. w. donnelly , phys . rev .
d 94 , 013012 ( 2016 ) .
g. d. megias , j. e. amaro , m. b. barbaro , j. a. caballero , t. w. donnelly and i. r. simo , to be published . | we use a recently developed model of relativistic meson - exchange currents to compute the neutron - proton and proton - proton yields in @xmath0 scattering from @xmath1c in the 2p-2h channel .
we compute the response functions and cross sections with the relativistic fermi gas model for different kinematics from intermediate to high momentum transfers .
we find a large contribution of neutron - proton configurations in the initial state , as compared to proton - proton pairs . in the case of charge - changing neutrino scattering the 2p-2h cross section of proton - proton emission ( _ i.e. ,
_ np in the initial state ) is much larger than for neutron - proton emission ( _ i.e. , _ two neutrons in the initial state ) by a @xmath2-dependent factor . the different emission probabilities of distinct species of nucleon pairs are produced in our model only by meson - exchange currents , mainly by the @xmath3 isobar current .
we also analyze other effects including exchange contributions and the effect of the axial and vector currents . |
lattice qcd calculations have predicted a transition from quark - gluon matter ( qgp ) to hadronic matter at high temperature or density @xcite .
this transition was believed to occur around ten microseconds after big bang .
practically , one may obtain this piece of information by studying the relativistic heavy ion collisions at brookhaven national laboratory now .
some possible probes have been proposed for searching for it @xcite .
elliptic flow @xmath1 is one of these probes , from which one may understand some early information in relativistic heavy ion collision through measuring anisotropy of momentum space of particles in final state . some experimental results from bnl - rhic have been published recently @xcite . as we know ,
hydrodynamical model , which is one method to study relativistic heavy ion collisions , can reproduce @xmath1 data very well below @xmath2 2 gev / c of transverse momentum ( @xmath3 ) @xcite .
however , the hadronical rescattering effect on elliptic flow after the hadronization following by the hydrodynamics , at least in our knowledge , has not been investigated in details so far .
based on this motivation , we shall focus the influence of rescattering in hadronic state on elliptic flow @xmath1 .
it should be pointed that hadronic phase space before rescattering is produced directly by an anisotropic momentum space of produced hadrons in order to produce @xmath1 in our calculation , which will be described in detail in following . as a result
, we find that @xmath1 of final particles is suppressed due to the rescattering among hadrons to a considerable extent .
it may indicate that the effect of rescattering among hadrons on @xmath1 is not ignored in the research on @xmath1 .
hydrodynamics of relativistic heavy ion collision @xcite is defined by ( local ) energy - momentum and net charge conservation , @xmath4 where @xmath5 denotes the energy - momentum tensor and @xmath6 the net four - current of the @xmath7th conserved charge . here
will only take the net baryon number .
we implicitly assume that all other charges which are conserved on strong - interaction time scale , e.g. , strangeness , charm , and electric charge , vanish locally .
the corresponding four - currents are therefore identically zero , and the conservation equations are trivial
. for ideal fluids , the energy - momentum tensor and net baryon current assume the simple form : @xmath8 where @xmath9 , p , and @xmath10 are energy density , pressure , and net baryon density in the local rest frame of the fluid , which is defined by @xmath11 = ( @xmath10,@xmath12 ) , respectively . here
= diag(+,-,-,- ) is the metric tensor , and @xmath14 = @xmath15(1,@xmath16 ) the four - velocity of the fluid ( @xmath16 is the three - velocity and @xmath15 = @xmath17 , the lorentz factor ) . for simplicity ,
the hydrodynamics that we are using assumes cylindrically symmetric transverse expansion with a longitudinal scaling flow profile , @xmath18 though we are investigating the non - central collision .
we recognize that our initial phase space of hadrons in this way is not perfect since we put an initial anisotropy of phase space by hand just after cooper - frye hadronization as illustrated in the following .
nevertheless , we think that it is still useful and interesting to investigate the hadronic rescattering effect on spectra and flow based on this treatment . in this point ,
our simulation is somehow like that we start from an assumed initial phase space of hadrons and then study the hadronic rescattering effect .
this phase space is similar to a sudden hadronization phase space which the partonic anisotropy turns to hadron anisotropy . for hydrodynamics , at @xmath19 = 0 , eqs ( 1 ) reduces to @xmath20 = -(\frac{v_{t}}{r_{t}}+\frac{1}{t})(e+p),\ ] ] @xmath21 = -(\frac{v_{t}}{r_{t}}+\frac{1}{t})m,\ ] ] @xmath22 = -(\frac{v_{t}}{r_{t}}+\frac{1}{t})r,\ ] ] where we defined e @xmath23 @xmath24 , m @xmath23 @xmath25 , and r @xmath23 @xmath26 . in the above expressions
, the index @xmath27 refers to the transverse component of the corresponding quantity .
the equations ( 3)-(5 ) describe the evolution in the @xmath19 = 0 plane . because of the assumption of longitudinal scaling
, the solution at any other @xmath28 0 can be simply obtained by a lorentz boost .
initial conditions for au + au at 200 gev / c in 20 - 40@xmath0 centrality is supposed that initial energy density = 10.3 @xmath29 and initial time = 0.6 @xmath30 , and a hydrodynamical evolution of cylindrically symmetric transverse expansion with a longitudinal scaling flow profile is adopted .
when temperature of liquid reaches to a critical energy density ( = 0.42 @xmath29 ) , a hadronization of qgp which is described by employing the mit bag model eos @xcite will take place . and an ideal hadron gas is established and the corresponding eos is applied .
hadronization in hydrodynamics usually takes the method of cooper and frye @xcite . in this work , however , we make an assumption that the coordinate and momentum spaces of hadronized particles are not isotropic but have shapes as plotted in figure 1 in order to simulate au + au at 200 gev / c in 20 - 40@xmath0 centrality and get @xmath1 in our calculation . here
the mean anisotropy in momentum space , @xmath31px / py@xmath32 , takes 1.18 in our simulation for all produced hadrons . for the details of the simulation to result in the above anisotropic distributions
, we wrote @xmath33 @xmath34 @xmath35 @xmath36 @xmath37 @xmath38 where @xmath39 , ( @xmath40 ) and ( @xmath41 ) are the 4-vectors of produced hadrons in coordinate and momentum spaces respectively .
we assume that hadrons are randomly distributed in the overlap coordinate space of two spherical nucleus in terms of the left column of equations ( 7 ) by hadronization . @xmath42 and y are random numbers which are respectively randomly distributed in [ 0,14 ] fm / c and [ -5,5 ] . as for momentum and position of hadrons , @xmath43 and
@xmath44 denote a random variable distributed in [ 0,4 ] gev and [ 0,7 ] fm , then @xmath45 and @xmath46 can be obtained by the anisotropic factor in momentum and coordinate space , namely @xmath47 , which makes our momentum and coordinate space distribution in transverse plane elliptic and anti - elliptic .
if we make @xmath47 equal to 1 , our hadronization method will go back to the method of cooper and frye . @xmath48 and @xmath49 are the azimuths distributed in [ 0,2@xmath50 in coordinate and momentum spaces .
but here we assume these produced hadrons are from a radial source in which @xmath48 and @xmath49 have a correlation like the equation ( 8) , which means that once a @xmath49 has been decided , the corresponding @xmath48 will be selected in a gaussian distribution whose mean is @xmath49 and whose width is always assumed to be 0.2@xmath51 in our calculation .
the set of equations results in the coordinate and momentum space as figure 1 shows .
here we choose the anisotropic factor @xmath47 in momentum space , equals to 1.18 for 200 gev / c au + au in 20 - 40@xmath0 centrality .
after the hadrons are produced , they enter next rescattering process among them .
we here use the rescattering model from luciae model @xcite .
two particles will collide if their minimum distance @xmath52 satisfies @xmath53 where @xmath54 is the total cross section in @xmath55 and the minimum distance is calculated in the c.m.s .
frame of the two colliding particles .
if these two particles are moving towards each other at the time when both of them are formed , the minimum distance is defined as the distance perpendicular to the momentum of both particles .
if the two particles are moving back - to - back , the minimum distance is defined as the distance at the moment when both of them are formed .
assuming that the hadrons move along straight - line classical trajectories between two consecutive collisions it is possible to calculate the collision time when two hadrons reach their minimum distance and order all the possible collision pairs according to the collision time sequence .
if the total and the elastic cross section satisfies @xmath56 where @xmath57 is a random number then the particles will be elastically scattered or else the collision will be considered as an inelastic collision . the distribution of the momentum transfer , @xmath58 , is taken as @xmath59 where @xmath60 , for an elastic scattering , depends on the masses of two scattering particles .
the azimuthal angle will be isotropically distributed . during the rescattering process
, the following inelastic reactions are included in our calculation : + [ cols="^,^ " , ] where the hyperons @xmath61 = @xmath62 or @xmath63 .
the relative probabilities for the different channels , e.g. in ( @xmath51 n)-scattering , is used to determine the outcome of the inelastic encounter .
as the reactions introduced above do not make up the full inelastic cross section , the remainder is again treated as elastic encounters .
figure 2 shows the transverse momentum distribution of @xmath64 , @xmath65 , @xmath66 before and after rescattering in au + au at 200gev / c in 20 - 40@xmath0 centrality in our hydrodynamical calculation followed by the rescattering model . , @xmath65 , @xmath66 in au + au at 200gev / c in 20 - 40@xmath0 centrality in our hydrodynamical + rescattering model .
full lines and dash lines represent @xmath3 spectra after and before hadronic rescattering , respectively ; solid symbols are experimental data which come from @xcite . ] at the same time , we obtain the @xmath1 of hadrons before and after rescattering in terms of formula ( 12 ) , which indicates the anisotropy of hadrons momentum space . @xmath67
figure 3 shows dependence of @xmath1 of @xmath64+@xmath65 , @xmath68+@xmath69 , @xmath66 , @xmath62+@xmath70 on @xmath3 before and after rescattering in hydrodynamics and experimental data in au+au at 200gev / c in 20 - 40@xmath0 centrality @xcite .
dependences of @xmath1 of @xmath64+@xmath65 , @xmath68+@xmath69 , @xmath66 and @xmath62 + @xmath70 before and after rescattering in hydrodynamical+rescattering model and experimental data which come from @xcite and @xcite in au+au at 200 gev / c in 20 - 40@xmath0 centrality . ] in order to study anisotropy of hadrons coordinate space , the parameter named as @xmath71 is defined by formula ( 13 ) .
@xmath72 dependences of spectra , @xmath1 and @xmath71 of @xmath73+@xmath74 ( square ) , @xmath75+@xmath65 ( circle ) , p+@xmath66 ( up - triangle ) and @xmath62+@xmath70 ( down - triangle ) before ( open symbols ) and after ( full symbols ) rescattering in hydrodynamical + rescattering model in au+au at 200 gev / c in 20 - 40@xmath0 centrality ; ( b),(d ) and(f ) : the @xmath3 dependences of @xmath76 , @xmath77 and @xmath78 .
see more details in text . ]
the plots ( a ) ( c ) and ( e ) in figure 4 give @xmath3 spectra , @xmath1 and @xmath71 of four types of particles ( ( @xmath73+@xmath74 ) , ( @xmath75+@xmath65 ) , ( p+@xmath66 ) and ( @xmath62+@xmath70 ) ) vs @xmath3 before and after rescattering in hydrodynamics+rescattering model .
the open and filled points , respectively , stand for the status before and after hadrons rescattering .
we can see that @xmath3 yield and @xmath71 of hadrons is enhanced by rescattering , and also note @xmath71 changes its sign through rescattering . on the other hand ,
elliptic flow @xmath1 of hadrons is suppressed by the rescattering . in order to quantify the extent of enhancement ( or suppression ) at different @xmath3 bins , we define three factors named as @xmath76 , @xmath77 and @xmath78 by formula ( 14),(15 ) and ( 16 ) .
@xmath79 @xmath80 @xmath81 the plots ( b ) ( d ) and ( f ) in figure 4 show respective dependence of @xmath76 @xmath77 and @xmath78on @xmath3 . in ( b ) , we find that the yield of hadrons is increased by hadrons rescattering , especially for higher @xmath3 yield .
it indicates the rescattering produced many secondary hadrons , whose @xmath3 seems harder than those hadrons in last generation . in ( d ) , @xmath77 increases with @xmath3 , which indicates that the suppression effect on @xmath1 from hadrons rescattering becomes weaker and weaker with the increasing of @xmath3 . in ( e ) and ( f )
, we see the rescattering turns the shape of coordinate space , and the effect becomes stronger with the increasing of @xmath3 .
it indicates the rescattering makes hadrons momentum space less anisotropy and more heavier particles , more suppression the @xmath1 . while the changes of the tropism of coordinate space occurs in the same time . in our calculation , at least 20 @xmath2 40% @xmath1 will disappear through final hadrons rescattering .
so the effect from final hadrons rescattering on @xmath1 maybe should be considered in order to research early state information in relativistic heavy ion collisions .
dependences of spectra , @xmath1 and @xmath71 of charged particle before ( open symbols ) and after ( full symbols ) rescattering in hydrodynamical + rescattering model in au + au at 200 gev / c in 20 - 40@xmath0 centrality .
the circles and squares are for the hadronization with both initial @xmath1 and @xmath71 , and the circles correspond to 0.2@xmath51 of the width of @xmath49-@xmath82 correlation and squares correspond to 0.5@xmath51 .
the diamonds are from the hadronization whose particles are only with initial @xmath71and the triangles are from the hadronization whose particles are only with initial @xmath1 ( see the insert of ( b ) ) ; ( b),(d ) and ( f ) : the @xmath3 dependences of @xmath76 , @xmath77 and @xmath78 in these four hadronaztion situations . ] in order to investigate the initial geometrical dependence of our results farther , we next attempt to take cooper - frye hadronization method but with three different anisotropic conditions to see the change of effect from final hadrons rescattering . since the status of momentum and coordinate space before the hadrons rescattering can not be measured by us , we assume anther three conditions here .
\(i ) we decrease the correlation between momentum space and coordinate space , i.e. increase @xmath83= 0.2@xmath51 to @xmath83 = 0.5@xmath51 .
as we see the circles and squares in figure 5 , the stronger correlation seems to produce more hadrons and suppress @xmath1 more strongly , but its change for the anisotropy of coordinate space @xmath71 is slight .
\(ii ) we only remain anisotropy of coordinate space when hadronization takes place .
what will happen after rescattering among hadrons ?
the diamons in figure 5 show our results .
obviously , only hadronic rescatterring can not induce @xmath1 if the interaction system has not anisotropy of momentum space before the hadronic rescattering , though its high @xmath76 indicates that more secondary hadrons are produced in rescattering . on the other hand ,
the anisotropy of coordinate space also disappear after hadrons rescattering .
it is also reasonable because the rescattering plays a role that it decreases @xmath1 and tends to make the reaction system isotropic from above , which also is consistent with the isotropic feature of rescattering angle in luciae rescattering model .
so the anisotropy of momentum space , i.e. @xmath1 , may be produced in partonic state before hadronization if the hadronic rescattering has no direction - sense .
in other words , flow must be formed before the hadronization .
\(iii ) we only remain initial anisotropy of momentum space but lack initial anisotropy of coordinate space when hadronization takes place as shown with triangles in figure 5 , the rescattering still decrease hadrons @xmath1 and creat a big anisotropy of coordinate space @xmath71 . not only in above three conditions but also in our default condition , we all find that hadrons rescattering produces many secondary hadrons and suppresses the elliptic flow @xmath1 of hadrons to a certain extent .
when the two anisotropy of coordinate and momentum space tend to be homologic with time , the rescattering ends and the reaction system freezes out
. however , we note that the effect of rescattering on @xmath1 in high @xmath3 domain is not perfect from figure 3 , i.e. , @xmath1 does not overlap with experimental data which saturates above @xmath842 gev / c .
we think that one cause may be from our anisotropic factor in the momentum space .
even though the experimental data can be fitted well in lower @xmath3 , it is a bit artificial since the nucleon distribution of transverse plane ( x , y ) and anisotropic hydrodynamical evolution are not taken into account in this hydrodynamics @xcite .
the other cause could be the partonic effect on flow , such as partons cascade @xcite and jet quenching @xcite , is also very important for the character of particles with high @xmath3 , which we also do not take into account .
of course , the main goal of this work is to investigate the hadronic rescattering effect on the final state hadronic elliptic flow .
in conclusion , we apply hydrodynamics model followed by the rescattering model to investigate the effect of hadronic rescattering on elliptic flow @xmath1 for 200 gev / c au + au at 20 - 40@xmath0 centrality .
hadronization is treated by cooper - frye method and an initial anisotropic phase space of hadrons is assumed .
we find that the yield of hadrons is increased after the rescattering . while , the rescattering among the hadrons plays a suppression role for @xmath1 , which makes an asymmetric system in momentum space less anisotropic . the suppression becomes weaker with the increasing of transverse momentum .
we miss at least 20 @xmath2 40% @xmath1 after hadronic rescattering in our hydrodynamics + rescattering model .
this may , however , change if hadronization occurs when partons make their last scattering as treated in texas multi - phase transport model when the spatial anisotropic is small @xcite instead of when the energy reaches certain critical energy density presented here . on the other hand ,
the hadronic rescattering makes the coordinate space of hadrons tend to the similar directions to the momentum space . in our model
, we also find that the rescattering does not induce the elliptic flow if only initial anisotropy of coordinate space exists in the hadronic phase .
considering that the hadronic system created in rhic collision is very dense , the following hadronic rescattering may happen potentially and play an important role .
of course , only information of the momentum space after the freeze - out can be measured in experiment .
but anyway , the effect of elliptic flow @xmath1 from final hadronic rescattering deserve attention if the hadronization takes place when the energy reaches certain critical energy density presented here in hydrodynamics .
we would like to thank dr .
x. n. wang and h. z. huang for discussions in earlier stage of this work .
this work was supported partially by the national natural science foundation of china under grant no 10328259 and 10135030 , the chinese academy of sciences grant for the distinguished young scholars of the national natural science foundation of china under grant no 19725521 and the major state basic research development program of china under contract no g200077404 .
s. s. adler et al . , phys .
lett . * 91 * , 18231 ( 2003 ) .
k. h. ackermann et al . , phys .
lett . * 86 * , 402 ( 2001 ) . c. alder et al .
lett . * 87 * , 182301 ( 2001 ) . c. alder et al .
lett . * 89 * , 132301 ( 2002 ) . c. alder et al .
rev . lett . * 90 * , 032301 ( 2003 ) .
k. adcox et al . , phys .
lett . * 89 * , 212301 ( 2002 ) .
j. adams et al .
, arxiv : nucl - ex/0306007(2003 ) .
p. f. kolb et al . , phys .
b * 459 * , 667 ( 1999 ) .
p. f. kolb and u. heinz , arxiv : nucl - th/0305084(2003 ) .
p. huovinen , p. f. kolb and u. heinz et al . ,
phys . lett .
b * 503 * , 58 ( 2001 ) .
m. csanad , t. csorgo et al .
, arxiv : nucl - th/0310040(2003 ) .
t. hirano , phys . rev
. lett . * 86 * , 2754 ( 2001 ) .
d. h. rischke and s. bernard and j. a. maruhn , nucl .
a * 595 * , 346 ( 1995 ) .
d. h. rischke and y. pursun and j. a. maruhn , nucl .
a * 595 * , 383 ( 1995 ) .
a. dumitru and d. h. rischke , phys .
c * 59 * , 354 ( 1999 ) .
a. dumitru and s. a. bass et al .
b*460 * , 411 ( 1999 ) .
a. chodos , r. l. jaffe et al .
d * 9 * , 3471 ( 1974 ) .
f. cooper and g. frye , phys .
d * 11 * , 192 ( 1975 ) .
a. tai and b. h. sa , comput .
commun * 116 * , 353 ( 1999 )
. b. h. sa and x. r. wang et al .
c * 60 * , 047901 ( 1999 ) .
s. s. adler et al .
, arxiv : nucl - ex/0307022(2003 ) .
z. w. lin and c. m. ko , phys . rev .
c * 65 * , 034904 ( 2002 ) . m. gyulassy , i. vitev and x. n. wang , phys .
86 * , 2537 ( 2001 ) . | the effect of hadronic rescattering on the elliptic flow has been investigated by the cooper - frye hadronization model from hydrodynamic evolution following by the afterburner hadronic rescattering model for 200 gev / c au + au at 20 - 40@xmath0 centrality .
it is found that the hadronic rescattering can suppress elliptic flow @xmath1 in momentum space especially in lower transverse momentum region .
in addition , the hadronic rescattering effects on transverse momentum spectra and anisotropy coordinate space of hadrons are studied . |
one of the signatures suggested for the formation of the quark - gluon plasma ( qgp ) in relativistic heavy ion collisions is the suppressed production of @xmath2 @xcite .
this was based on the idea that once the qgp is formed in heavy - ion collisions , the @xmath2 produced from initial hard scattering would be dissociated in the qgp because of the color screening between its charm and anticharm quark , reducing thus the number of produced @xmath2 .
however , recent results from lattice quantum chromodynamics ( lqcd ) calculations have suggested that @xmath2 can exist above the critical temperature for the phase transition between the qgp and the hadron gas ( hg ) @xcite .
this thus not only reduces the dissociation of @xmath2 but also can lead to the production of @xmath2 from the recombination of charm and anticharm quarks in qgp @xcite .
the latter effect becomes more important when the collision energy is high as more charm and anticharm quarks are produced from initial hard scattering . experimentally , suppressed production of @xmath2 in heavy - ion collisions compared with that is expected from initial hard scattering has been observed at both super proton synchrotron ( sps ) and relativistic heavy ion collider ( rhic ) @xcite .
many theoretical explanations have been proposed to understand these experimental observations @xcite . among them
are the two - component model @xcite and the statistical model @xcite . in the two - component model , some of the initially produced @xmath2
were found to survive during the collision if the @xmath2 dissociation temperature is sufficiently high @xcite . in the statistical model , it is assumed , however , that all initially produced @xmath2 are dissociated in qgp and the detected @xmath2 are entirely due to regeneration from charm and anticharm quarks in the qgp @xcite .
both models have been able to successfully reproduce the measured nuclear modification factor @xmath0 of @xmath2 as a function of the number of participants @xcite as well as of the transverse momentum of @xmath2 @xcite . in the two - component model ,
the number of regenerated @xmath2 is smaller than that in the statistical model , because not all charm quarks produced in a collision are thermalized in qgp , the finally produced @xmath2 include , however , those produced from initial hard scattering and having survived thermal decay in the produced hot dense matter .
although the transverse momentum of thermally produced @xmath2 in the statistical model is smaller than that of @xmath2 produced from initial hard scattering , it is enhanced by the transverse flow of @xmath2 .
therefore , it is difficult to discriminate the two models based on the number of finally produced @xmath2 and their transverse momentum distribution .
previously , two of us have studied in the two - component model the @xmath0 of @xmath2 as a function of the number of participants in heavy - ion collisions @xcite . in that study ,
the initial local temperature of produced qgp was determined from the local entropy density , which was taken to be proportional to the linear combination of the number of participants and the number of binary collisions in that region .
the initial fireball thus separates into two regions with temperatures higher and lower than the dissociation temperature of @xmath2 .
while the survival rate of initially produced @xmath2 vanishes in the high temperature region , its value in the low temperature region is finite and depends on the temperature .
the different values of the @xmath2 survival rate in the two regions were found to give rise to a kink in the dependence of the @xmath0 of @xmath2 on the number of participants as seen in the experimental data . in the above mentioned work , both the screening mass between charm and anticharm quark and the relaxation factor of charm quarks in qgp were calculated in the leading order in pqcd , while the dissociation cross section of charmonia in qgp were calculated up to the next - to - leading order in pqcd . according to lqcd
, the strong coupling constant remains large up to around 2 @xmath3 , so below this temperature the qgp is nonperturbative and higher - order corrections in pqcd may not be negligible .
the higher - order corrections are expected to enhance in qgp both the thermal decay rate of charmonia , which would decrease the number of initially produced @xmath2 , and the relaxation of charm quarks , which would increase the number of regenerated @xmath2 .
the cancelation of these two effects is probably one of the reasons that both the two - component model and the statistical model can reproduce the measured @xmath0 of @xmath2 as a function of the number of participants in heavy - ion collisions . in the present work ,
we improve our previous two - component model in two ways .
firstly , instead of assuming that the transverse acceleration of produced fireball is constant and then vanishes at a certain time as in the previous work , the acceleration in the present work is obtained from the pressure of the fireball and continues to be present until the fireball freezes out .
secondly , we take into account the azimuthally anisotropic expansion in noncentral collisions by using a simple parametrization with a transverse - momentum - dependent flow velocity to reproduce the measured @xmath1 of light hadrons .
this schematic fireball expansion model is then applied to study the @xmath0 of @xmath2 as a function of both the number of participants and the transverse momentum of @xmath2 as well as to study the @xmath1 of @xmath2 . for the chemical and thermal freeze - out temperatures , we follow those used in the statistical model . in this model ,
the chemical freeze - out occurs at temperature around 160 mev , after which chemical equilibrium is reached and no additional effective chemical reactions are present and the kinetic freeze - out takes place later when the temperature drops to 120 mev . since the kinetic freeze - out depends on the elastic scattering cross sections of hadrons and unlike light hadrons , the elastic scattering cross sections of @xmath2 in hadronic matter are small , we assume that the kinetic freeze - out temperature for @xmath2 is the same as the chemical freeze - out temperature . finally , we take into account the higher - order corrections by multiplying the pqcd cross sections with arbitrary constants to study their effects on the @xmath0 and @xmath1 of @xmath2 .
as shown later , although the fraction of @xmath2 produced from regeneration increases with increasing larger higher - order corrections , the resulting @xmath0 of @xmath2 remains essentially unchanged . on the other hand ,
the @xmath1 of @xmath2 at low transverse momentum becomes more negative with larger higher - order corrections .
therefore , studying the @xmath1 of @xmath2 can help to discriminate between the two @xmath2 production mechanisms of from initial hard scattering and regeneration in the qgp .
this paper is organized as follows . in sec . [ two ] , we briefly describe the two - component model used in ref . @xcite .
we then introduce in sec .
[ fireball ] the schematic model for the expanding fireball produced in heavy - ion collision to obtain the time evolution of the thermal properties as well as the spatial and kinematic information of the fireball for both central and non - central collisions . in sec .
[ raa ] , the @xmath0 of @xmath2 as a function of the number of participants and the transverse momentum of @xmath2 are calculated with the schematic model .
both the cronin effect for the initial transverse momentum broadening and the leakage effect on initially produced @xmath2 are considered .
results on the elliptic flow of @xmath2 are given in sec .
[ elliptic ] . in sec .
[ high ] , higher - order corrections are introduced by multiplying the pqcd results with constant factors , and their effects on the @xmath0 and @xmath1 of @xmath2 are investigated . finally , the summary and conclusion are given in sec .
[ summary ] .
also , details on the the azimuthal angle dependence of the flow velocity are given in the appendix .
in the two - component model for @xmath2 production in relativistic heavy ion collisions , some of initially produced @xmath2 from hard scattering can survive through the collisions and contribute to the final observed @xmath2 in experiments , besides those regenerated from the charm and anticharm quarks in the produced qgp . in this section ,
we briefly describe the method used in ref @xcite to evaluate these two contributions .
the number of initially produced @xmath2 in a heavy ion collision at impact parameter @xmath4 can be determined by using the glauber model , that is @xmath5 where @xmath6 is the cross section for @xmath2 production in a @xmath7 collision , which has a value of 0.774 @xmath8 per rapidity near midrapidity at collision energy @xmath9 gev @xcite , and a is the mass number of colliding nuclei .
the overlap function @xmath10 in the above equation is defined by @xmath11 in terms of the thickness function @xmath12 where @xmath13 is the nucleon distribution in the colliding nucleus .
the survival rate of initially produced @xmath2 in relativistic heavy ion collisions is then given by @xmath14 where @xmath15 is the time the @xmath2 is produced from initial hard scattering and @xmath16 is the time the @xmath2 freezes out from the hot dense matter .
the @xmath2 survival rate depends on its thermal decay width @xmath17 in the produced hot dense matter , @xmath18 in the above , the summation is over the particle species @xmath19 , which includes the quark and gluon in qgp as well as the baryon and meson in hg , with its density denoted by @xmath20 ; @xmath21 is the relative velocity between the @xmath2 and the particle ; and @xmath22 is the dissociation cross section of the @xmath2 by the particle .
since some directly produced @xmath2 are from the decay of directly produced excited charmonium states @xmath23 and @xmath24 , the survival rate of all initially produced @xmath2 from thermal decay is the linear combination of the survival rates for @xmath2 , @xmath23 , and @xmath24 , that is @xmath25 where we have assumed that 25% and 8% of @xmath2 are , respectively , from the feed - down of @xmath23 and @xmath24 as usually adopted in the literature @xcite .
we note that in ref @xcite the dissociate cross sections of charmonia in qgp were calculated in pqcd up to the next - to - leading order and those in hg were calculated via the factorization formula with the bethe - salpeter amplitude for the bound charmonium states @xcite . for regenerated @xmath2 , their number
is similarly determined as in the statistical model , that is @xmath26 in the above , @xmath27 , @xmath28 , and @xmath29 are , respectively , the equilibrium densities of @xmath2 , @xmath23 , and @xmath24 at chemical freeze - out when the system has volume @xmath30 . the charm fugacity parameter @xmath31 takes into account the fact that the total charm quark number is not in chemical equilibrium .
it is determined instead by @xmath32 where @xmath33 and @xmath34 are modified bessel functions with @xmath35 and @xmath36 being , respectively , the density of open and hidden charms , and @xmath37 is the total charm quark pairs produced from initial hard scattering given by @xmath38 where @xmath39 is the cross section for @xmath40 pair production in a @xmath7 collision , which is 63.7 @xmath41b per rapidity near midrapidity at @xmath9 gev in perturbative qcd @xcite .
the factor @xmath42 in eq .
( [ reg ] ) takes into account the nonequilibrium effect of charm quarks in qgp and is evaluated according to @xmath43 where @xmath15 and @xmath44 are , respectively , the time charm quark pairs are produced from initial hard scattering and the time qgp phase ends , and @xmath45 is the thermal scattering width of charm quarks in qgp and is calculated in ref @xcite using the leading - order pqcd .
in our previous work @xcite , it was assumed for simplicity that the fireball formed in heavy - ion collisions expands transversely with constant acceleration of 0.1 @xmath46/fm until certain time . besides , the expansion is assumed to be azimuthally isotropic even in noncentral collisions , neglecting thus the initial elliptic shape of the collision geometry . since the transverse acceleration of the fireball is expected to be a smooth and continuous function of time and also to be azimuthally anisotropic in non - central collisions , we improve the treatment of the acceleration of the fireball by using the thermodynamical consideration and also introduce a simple parametrization with a transverse - momentum - dependent flow velocity for the anisotropic expansion . to treat dynamically the transverse acceleration of the fireball formed in relativistic heavy ion collisions , we consider the pressure inside the fireball . in terms of the energy density @xmath47 , entropy density @xmath48 , and temperature @xmath49 of the fireball as well as the chemical potential @xmath50 and number density @xmath51 of particle species @xmath19 in the fireball at a certain time , the pressure @xmath52 is given by @xmath53 the outward force due to the pressure is then @xmath54 where @xmath55 is the area of the cylindrical side of the fireball .
letting @xmath56 be the inertia mass against the transverse expansion of the fireball , the transverse acceleration is then @xmath57 where @xmath58 is the pressure of the fireball at freeze - out , if we take into consideration of the fact that the acceleration should vanish at freeze - out , as particles undergo only free streaming afterwards .
it is assumed in the above that the fireball is globally in thermal equilibrium .
we have thus ignored the fact that in more realistic hydrodynamic approach only local thermal equilibrium is achieved and each fluid element of the fireball moves collectively . assuming the transverse acceleration @xmath59 acts in the local frame , the transverse velocity of the boundary of the fireball in the laboratory frame at `` i+1''th time step can then be obtained from its value at `` i''th time step by adding the effect of acceleration relativistically , i.e. , @xmath60 where @xmath61 . for the thermal quantities in eq .
( [ pressure ] ) , we use the quasiparticle picture for the qgp , which assumes that strongly interacting massless partons ( quarks and gluons ) can be substituted by noninteracting massive partons to reproduce the energy density and pressure extracted from lqcd . in this approach , the pressure , energy density , and entropy density are given , respectively , by @xcite @xmath62 in the above , @xmath63 , @xmath64 , and @xmath50 are , respectively , the thermal mass , degeneracy factor and chemical potential of parton species @xmath19 with @xmath50 obtained from the flavor conservation @xcite . the parton distribution function is denoted by @xmath65\pm 1}\ ] ] with the plus and minus signs in the denominator for quarks and gluons , respectively , and @xmath66 . for the bag pressure @xmath67
, it is obtained from the relation @xmath68 such that @xmath69 where @xmath70 mev is the critical temperature for the qgp to hg phase transition and @xmath71 is the bag pressure at @xmath3 with its value taken to be 0.1 times the energy density of qgp at @xmath3 to make the pressure continuous at @xmath3 . from the above thermal quantities
, the parton thermal masses can be determined by fitting the qgp equation of state from lqcd as given in details in ref .
@xcite . for thermal quantities in the hg phase , they are calculated in the resonance gas model by including all mesons of masses less than 1.5 gev and all baryons of masses less than 2.0 gev , and for simplicity they are assumed to be noninteracting . figure [ ep ]
shows the resulting energy density and pressure , divided by temperature to the fourth power , as functions of temperature .
we note that the pressure at the thermal freeze - out temperature 120 mev is @xmath72 .
the inertia mass @xmath56 in eq .
( [ acc ] ) is a free parameter in our model and is determined by fitting the average transverse momenta of light hadrons at thermal freeze out to the measured data .
assuming particles freeze out at certain proper time , their number is then @xmath73 where @xmath74 and @xmath75 are the hypersurface vector of the fireball and the current density vector of particles , respectively ; @xmath76 and @xmath77 are , respectively , the 4-velocity of the fluid element and the number density of particles . in obtaining the last step of the above equation , we have used the boltzmann distribution for evaluating the number density and the fact that the 4-velocity of a fluid element is equal to the average velocity of particles in the fluid element .
ignoring the transverse boost and keeping only the longitudinal boost , the hypersurface vector becomes @xmath78 where @xmath79 for the 4-velocity @xmath80 of the fluid element , both boosts are included , with the transverse boost followed by the longitudinal boost , leading to @xmath81 where @xmath82 is the transverse rapidity defined as @xmath83 with @xmath84 being the transverse velocity .
the transverse rapidity of the fluid element is assumed to be proportional to the radial distance from the center of the fireball , based on the result from hydrodynamics @xcite .
for the 4-momentum @xmath85 of a particle , it is @xmath86 where @xmath87 and @xmath88 are , respectively , the rapidity and the transverse energy . in general ,
three different azimuthal angles are required : one for the hypersurface vector @xmath89 ; one for the 4-velocity @xmath76 of the fluid element ; and the other for the momentum vector @xmath85 . in central collisions ,
the azimuthal angle for @xmath89 is , however , the same as that for @xmath76 .
this is also the case in noncentral collisions if only the radial flow is considered and this is assumed in present study .
substituting eqs .
( [ hypersurface ] ) into eq .
( [ number ] ) leads to @xcite @xmath90 k_1\bigg[\frac{m_t \cosh \rho}{t}\bigg ] , % \label{number}\end{aligned}\ ] ] where @xmath91 is the transverse area , and @xmath92 and @xmath93 are modified bessel functions . from eq .
( [ number ] ) , the averaged transverse momentum of particles is given by @xmath94 we find that with the inertia mass @xmath95 gev , the transverse velocity of fireball at freeze - out is 0.71 , and the average transverse momenta of pions , kaons and protons are 462 , 668 and 953 mev/@xmath96 , respectively , which are comparable to measured values of 451@xmath9733 , 670@xmath9778 and 949@xmath9785 mev/@xmath96 for @xmath98 , @xmath99 and proton , respectively , in au+au collisions as @xmath9 gev and of 0 - 5% centrality @xcite .
we note that the measured average transverse momenta of @xmath100 , @xmath101 and antiproton have similar values as those for their antiparticles . in fig .
[ comparison](a ) , we show the transverse velocity of the fireball obtained with the acceleration using the above thermodynamical approach as a function of time and compare it with that using the constant acceleration as in our previous study .
the transverse velocity is seen to increase rapidly at first and then slowly in the thermodynamical approach , whereas it increases linearly with time and stays constant after a certain time in our previous study with constant acceleration .
for the temperature of the fireball , shown in fig .
[ comparison](b ) , it evolves in time similarly in the two cases in spite of their different evolutions in the transverse velocity . in the thermodynamic approach
, the qgp phase terminates at @xmath102 fm/@xmath96 , the mixed phase at @xmath103 fm/@xmath96 , and the hg phase reaches the freeze - out temperature at @xmath104 fm/@xmath96 . in the case of constant acceleration ,
the corresponding times are @xmath105 fm/@xmath96 , @xmath106 fm/@xmath96 , and @xmath107 fm/@xmath96 , respectively , and are thus similar .
therefore , the modified acceleration introduced in the present study is not expected to change our previous results on the @xmath0 of @xmath2 as a function of the number of participants @xcite . in noncentral collisions ,
the length of the overlapping region of two colliding nuclei is shorter in the direction of impact parameter than in its perpendicular direction on the transverse plane , which are designated , respectively , as the @xmath108-direction and the @xmath87-direction .
the difference in the overlapping lengths along these two directions leads to difference in the pressure gradient in these two directions , and as a result fluid velocities along these two directions are different as well , with the fluid velocity in the @xmath108-direction higher than the one in the @xmath87-direction . to include such an anisotropic expansion of the fireball , we add a term proportional to the eccentricity of the fireball to the acceleration [ eq .
( [ acc ] ) ] as follows @xmath109 where the eccentricity @xmath110 is defined as @xmath111 and @xmath112 is a fitting parameter to take into account the different accelerations in and perpendicular to the reaction plane in non - central collisions . in the above , @xmath113 and @xmath114 are , respectively , the half length of the fireball in the @xmath108-direction and in the @xmath87-direction on the transverse plane . if the shape of the overlapping region is a perfect ellipse , they then correspond to the semiminor and semimajor of the ellipse , respectively . for simplicity
, we transform the initial shape of the fireball into a ellipse by keeping the same transverse area and the same @xmath113 to @xmath114 ratio .
the velocities in the @xmath108-direction and in the @xmath87-direction are then , respectively , @xmath115 where @xmath116 and @xmath117 .
assuming that the fireball keeps the elliptic shape on transverse plane during expansion , the fluid velocity , which is assumed to be along the radial direction , as a function of the azimuthal angle @xmath118 is then @xmath119,\end{aligned}\ ] ] where @xmath120^{-1/2}\nonumber\ ] ] is the radial distance from the center of the fireball to its surface along the azimuthal angle @xmath118 .
details in the derivation of the above result are given in the appendix .
for the transverse acceleration @xmath59 in eq .
( [ acc2 ] ) , it is obtained in the same way as in central collisions .
the constant @xmath112 in eq .
( [ acc2 ] ) is determined by requiring that the final @xmath1 of light hadrons , given by the expectation value of @xmath121 as @xmath122 reproduce the measured data . and @xmath123 ( b ) , @xmath113 and @xmath114 ( c ) for au+au collisions at @xmath9 gev and impact parameter
@xmath1249 fm.,width=302 ] in fig
. [ evolution ] , we show the time evolution of the temperature of the fireball and of @xmath125 , @xmath123 , @xmath113 and @xmath114 for au+au collisions at @xmath9 gev and impact parameter @xmath124 9 fm . these results are obtained with the anisotropic constant @xmath126 in eq .
( [ acc2 ] ) and the inertia mass @xmath127 gev in eq .
( [ acc ] ) .
it is seen that the qgp phase ends at @xmath128 , the mixed phase lasts until @xmath129 , and the freeze - out temperature is reached at @xmath130 .
the flow velocities @xmath125 and @xmath123 are 0.51 @xmath96 and 0.27 @xmath96 at the phase transition to hg and reach 0.76 @xmath96 and 0.57 @xmath96 at freeze - out , respectively .
we note that the length of @xmath113 becomes slightly longer than that of @xmath114 at freeze - out .
of pions ( solid line ) , kaons ( dashed line ) , and protons ( dotted line ) for au+au collisions at @xmath9 gev and impact parameter @xmath131 fm before the high @xmath132 correction ( a ) and after the correction ( b ) . experimental data from collisions at 20 - 60% centrality @xcite are shown by rectangles , circles , and triangles for pions , kaons , and protons , respectively.,width=302 ] the elliptic flow @xmath1 of pions ( solid line ) , kaons ( dashed line ) and protons ( dotted line ) at @xmath131 fm , obtained from eq .
( [ v2 ] ) and using the results of fig . [ evolution ] , are shown in fig .
[ v2-light](a ) . except in the low @xmath132 region
, they are seen to deviate from the experimental data from collisions at 20 - 60% centrality @xcite , shown by rectangles , circles , and triangles for pions , kaons , and protons , respectively .
the reason for this deviation is because high-@xmath132 particles are not likely to be completely thermalized as assumed in our fireball model . to take into the non - equilibrium effect
, we multiply @xmath133 $ ] to @xmath134 , where @xmath135 is a fitting parameter and @xmath77 is the number of constituent quarks in a hadron , as described in the appendix .
figure [ v2-light](b ) shows the resulting @xmath1 of light hadrons after this high @xmath132 correction , and they are seen to reproduce very well the experimental data at all @xmath132 . since the high @xmath132 correction factor modifies the boltzmann distribution in eq .
( [ number ] ) , it can be considered as an effective way to include the viscous effect on the fireball expansion . in the quark - counting picture for hadron scattering ,
the ratio of the baryon to the meson scattering cross sections is 3/2 .
this then leads to a smaller viscous effect on baryons than mesons , which is consistent with the quark number dependence in the above high @xmath132 correction factor .
( solid line ) , @xmath99 ( dashed line ) , and proton ( dotted line ) from the fireball model with corresponding experimental data shown by the rectangle , circle and triangle.,width=302 ] the average transverse momenta of pions , kaons , and protons in the above case of @xmath131 fm are @xmath136 , @xmath137 and @xmath138 mev/@xmath96 , respectively , which are comparable to measured values of 436@xmath9735 , 655@xmath9754 , and 901@xmath9788 mev/@xmath96 in collisions at 30 - 40% centrality as well as 426@xmath9735 , 636@xmath9754 and 868@xmath9788 mev/@xmath96 in collisions at 40 - 50% centrality for @xmath98 , @xmath99 and proton , respectively @xcite . for other values of the impact parameter , it is found that the inertia mass parameterized as @xmath139 with @xmath140 being the number of participants in a collision , works well as shown in fig .
[ resistance ] , where the participant number dependence of the average transverse momenta of @xmath98 , @xmath99 and protons obtained with the parameterized inertia mass are compared with the experimental data .
the nuclear modification factor @xmath0 of @xmath2 is a measure of its production in heavy ion collisions relative to that in @xmath7 collisions multiplied by the number of binary collisions .
its value is less than one if there is a suppression and greater than one if there is an enhancement . besides effects due to the produced hot dense matter
, the @xmath0 of @xmath2 is also affected by two other important effects , namely the cronin effect and the leakage effect .
charmonium is produced mainly through the gluon - gluon fusion in high - energy nuclear collisions .
if a gluon in one nucleus is scattered by the nucleon in the other colliding nucleus before it produces the charmonium , the gluon obtains an additional transverse momentum , and the charmonium produced from the gluons then also acquires this additional transverse momentum . as a result
, the distribution of the transverse momentum of @xmath2 becomes different from that in p+p collisions , and this is called the cronin effect . taking into account of this effect ,
the average transverse momentum of @xmath2 in a+a collisions is modified as following : @xmath141 where @xmath142 and @xmath143 are average squared transverse momenta of charmonium in an a+a collision and in a p+p collision , respectively ; @xmath144 is the average of added squared transverse momentum by one gluon - nucleon collision , @xmath145 is the mean free path of a gluon in uniform nuclear matter , @xmath146 is the average path length of the gluon in uniform nuclear matter before it produces a charmonium .
the prefactor @xmath147 is fitted to the experimental data of p+a or d+a collisions and @xmath146 is calculated with the glauber model according to @xmath148 where @xmath149\bigg).\nonumber \label{k}\end{aligned}\ ] ] in the above , @xmath150 and @xmath151 are the impact parameter and transverse vector in configuration space , respectively ; @xmath152 and @xmath153 are longitudinal positions where @xmath2 is produced from the center of nucleus a and from the center of nucleus b , respectively ; @xmath154 and @xmath155 are the distribution functions of nucleons in nucleus a and in nucleus b , respectively , and they are normalized to their mass numbers ; @xmath156 is the saturation density of nuclear matter .
the function @xmath157 is the sum of the path length in nucleus a of a gluon in nucleus b and the path length in nucleus b of a gluon in nucleus a , which correspond to the first term and the second term on the right hand side of @xmath157 , respectively .
the probability that a charmonium is produced at the position @xmath152 from the center of nucleus a and @xmath153 from the center of nucleus b is given by @xmath158 .
the exponential factor in @xmath158 is due to the effect of nuclear absorption or nuclear destruction as produced @xmath2 can be absorbed or destroyed by nucleons with the absorption cross section @xmath159 before it gets out of the colliding nuclei .
notice that the integration range for the cronin effect is from @xmath160 to @xmath161 and for nuclear absorption from @xmath161 to @xmath162 in nucleus a(b ) , as the cronin effect is present before the production of the primordial charmonium and the nuclear absorption happens after the creation of the @xmath2 . in the present work
, @xmath163 is calculated at each point on the transverse plane as following : @xmath164 the cronin effect thus depends on the transverse position at which a @xmath2 is produced . in our previous work
, it was assumed that charmonia produced inside the fireball could not get out and charmonia produced outside could not get into the fireball .
in realistic situation , it is possible for charmonia of high @xmath132 to escape from the fireball , leading to the so - called leakage effect . to take into account this effect ,
we consider the thermal decay width of a charmonium inside the fireball , i.e. , its transverse position ( @xmath108 and @xmath87 ) satisfies @xmath175 ^ 2+[y / r_y(t)]^2<1 $ ] , @xmath176 where @xmath19 denotes the particle species that dissociates the charmonium with @xmath64 and @xmath51 being its degeneracy factor and number density , respectively ; @xmath21 is the relative velocity between the charmonium and the particle ; @xmath22 is the dissociation cross section of charmonium by the particle . the charmonium thermal decay width vanishes once it moves out of the fireball when its transverse position satisfies @xmath175 ^ 2+[y / r_y(t)]^2>1 $ ] .
furthermore , we set the charmonium decay width to infinite if it is produced in a region with temperatures higher than its dissociation temperature . of @xmath2 as a function of the number of participants including both the cronin and the leakage effect .
the dotted line is the @xmath0 for constant acceleration and the solid line is that for acceleration determined thermodynamically , which are , however , almost indistinguishable.,width=302 ] results obtained with the two - component model described in sec .
[ two ] for the @xmath0 of @xmath2 as a function of the number of participants including both the cronin and the leakage effect are shown in fig .
[ raa - np1 ] .
the dotted line is the @xmath0 with constant acceleration of the fireball and the solid line is that with the fireball acceleration determined thermodynamically . in the latter case , the cronin effect and the leakage effect
are included . as shown in fig .
[ comparison](b ) , the lifetime of qgp is slightly longer in the former case than in the latter case , and this leads to a somewhat smaller @xmath0 in the former case . since the cronin effect only changes the transverse momentum of @xmath2 , not the number of @xmath2 , it does not affect the @xmath0 of @xmath2 . the leakage effect , on the other hand , may increase @xmath0 or decrease @xmath0 , depending on whether @xmath2 initially produced outside the fireball can move into the fireball and on whether @xmath2 initially produced inside the fireball can escape to the outside . the net leakage effect on @xmath0
is , however , found to be small , and the difference between the results for constant acceleration and those for acceleration determined thermodynamically is thus hardly seen in fig .
[ raa - np1 ] . of @xmath2 as a function of transverse momentum in collisions of 0 - 20% ( a ) , 20 - 40% ( b ) , 40 - 60% ( c ) and 60 - 92% ( d ) centralities .
dashed and dotted lines are , respectively , the @xmath0 from initially produced @xmath2 and regenerated @xmath2 , and the solid line is the sum of the two.,width=302 ] figure [ raa - pt ] shows the @xmath0 of @xmath2 as a function of transverse momentum obtained with the fireball acceleration calculated thermodynamically .
it is seen that the @xmath0 of initially produced @xmath2 , given by the dashed line , increases with transverse momentum because of the cronin effect . on the contrary , the @xmath0 of regenerated @xmath2 decreases with transverse momentum at high @xmath132 as shown by the dotted line .
we note that the @xmath0 of @xmath2 is slightly enhanced and suppressed at high and low @xmath132 , respectively . in au+au collisions at @xmath177 gev from the two - component model ( solid line ) and experimental data ( filled squares).,width=302 ] in figure [ pt2 ] ,
we show the participant number dependence of the average squared transverse momentum of @xmath2 in au+au collisions from the two - component model .
our results reproduce the experimental data for peripheral and mid - central collisions . for central collisions
, the two - component model overestimates the measured average squared transverse momentum of @xmath2 .
one possible reason for this is that we may have overestimated the diffusion of charm quarks in the fireball .
the number of charm quarks produced in heavy - ion collision is proportional to the number of binary collisions of nucleons , which is more concentrated in the central region of the fireball .
the number of light quarks is , on the other hand , proportional to the number of participants .
since the relative concentration of the number of binary collisions to the number of participants is larger in central collisions than in semi - central or peripheral collisions , charm quarks are initially more concentrated in the central region than light quarks .
if the diffusion of charm quarks is slow , they would remain near the central region even at hadronization and their transverse velocity would then be smaller than those of light quarks .
the elliptic flow @xmath1 of @xmath2 in the two - component model can be different from that of light hadrons as a result of following two effects .
firstly , @xmath2 s are produced from two different mechanisms : the initial hard scattering and the regeneration , while light hadrons are produced only through regeneration .
secondly , the freeze - out temperature for @xmath2 could be different from that of light hadrons .
since initially produced @xmath2 s have transverse momenta that are azimuthally isotropic , they are expected to have vanishing @xmath1 . however , the anisotropic expansion of the fireball could give rise to nonzero @xmath1 for @xmath2 .
in noncentral collisions , the size of the fireball is shorter in the @xmath108-direction than in the @xmath87-direction . as a result
, the survival rate of @xmath2 moving in the @xmath108-direction will be higher than that moving in the @xmath87-direction . in this case
, the @xmath1 of @xmath2 is positive , because the number of @xmath2 moving in the @xmath108-direction is more than that moving in the @xmath87-direction . on the contrary ,
the @xmath1 of regenerated @xmath2 is negative at low @xmath132 for the following reason .
assuming particles are locally thermalized , they then have a thermal distribution in the local frame .
since the thermal distribution is peaked at @xmath178 , where @xmath49 and @xmath179 are the temperature and mass of the particle , respectively , more massive particles thus peak at higher momenta .
if the thermalized matter is boosted with a certain velocity , the peak is shifted to higher momentum in the laboratory frame . in noncentral collisions ,
the difference in the length of the overlapping region in the @xmath108-direction and the @xmath87-direction leads to a difference in the pressure gradient and as a result a difference in fluid velocity along these two directions .
because the length of the overlapping region is shorter in the @xmath108-direction , the fluid velocity is higher and the peak of the momentum distribution is more shifted in that direction . as the peak shifts to higher momentum , the number density at momentum lower than the peak momentum becomes smaller than the number density in the unshifted or less shifted momentum distribution , while the number density at momentum higher than the peak becomes larger . as a result
, the number density of @xmath2 moving in the @xmath108-direction is smaller than that of @xmath2 moving in the @xmath87-direction at small @xmath132 , and this becomes opposite at high @xmath132 , leading thus to a negative @xmath1 . of @xmath2 for collisions at @xmath131 fm in the two - component model .
the dashed , dotted and solid lines are , respectively , the @xmath1 of initially produced , regenerated and total @xmath2.,width=302 ] figure [ v2-jpsi1 ] shows the results from the two - component model for the @xmath1 of initially produced , regenerated and total @xmath2 in au+au collisions at @xmath177 gev and @xmath131 fm .
these results are obtained with the high-@xmath132 correction to the flow velocity as in the case for light hadrons .
it is seen that the @xmath1 of initially produced @xmath2 is essentially zero , whereas the @xmath1 of regenerated @xmath2 has a minimum value of @xmath180 at @xmath181 gev .
the @xmath1 of total @xmath2 is , however , slightly negative , because only 12% of @xmath2 are from regeneration for collisions at @xmath131 fm in the two - component model .
we note that the @xmath1 of @xmath2 has been recently measured by the phenix collaboration for au+au collisions at 20 - 60% centralities @xcite and it has a value of [email protected] between @xmath182 and @xmath183 gev , which is much much more negative than our value .
in our two - component model , the screening mass between charm quarks and the thermal decay widths of charmonia as well as the relaxation factor of charm quarks in qgp are calculated in pqcd .
specifically , the dissociation cross section of charmonium used for determining its thermal decay width is calculated up to the next - to - leading order , and the elastic cross section of charm quarks for calculating the relaxation factor , and the screening mass are calculated in the leading order . since it is known that pqcd is not reliable up to around @xmath184 , corrections due to higher orders are expected to be important for the temperature regime considered in the present study .
the higher - order correction for the dissociation cross sections of charmonia is expected to suppress the survival rate of charmonia in hot dense nuclear matter while the correction for the relaxation factor of charm quarks would enhance the regeneration of charmonia .
therefore , there could be a balance between the dissociation and the regeneration of charmonia in calculations including higher - order corrections . to include higher - order corrections , we simply introduce in the present study some multiplication factors to the dissociation cross sections of charmonia and to the elastic cross section of charm quarks as following : @xmath185 where @xmath55 and @xmath186 are arbitrary constants . of @xmath2 as a function of the number of participants with different values of the multiplicative parameteres ( a , b)=(1,1 ) , ( 2,2 ) , ( 5,4 ) and ( 10,6.5 ) for including the effect of higher - order corrections in pqcd .
corresponding results are shown by solid , dashed , dotted , and dot - dashed lines , respectively.,width=302 ] in fig .
[ raa2 ] , the @xmath0 of @xmath2 is shown as a function of the number of participants .
solid , dashed , dotted , and dash - dotted lines correspond , respectively , to results obtained in the two - component model with ( a , b)=(1,1 ) , ( 2,2 ) , ( 5,4 ) , and ( 10,6.5 ) for the multiplicative factors used to include high - order effects in pqcd .
these results show that higher - order corrections do not change much the value of the @xmath0 of @xmath2 even with a large correction factor of ten .
this is so because of the balance between the dissociation and the regeneration of @xmath2 as mentioned previously , with the suppression of @xmath2 caused by enhanced dissociation compensated by the enhancement of regenerated @xmath2 .
however , as the correction factor becomes large , the kink in the @xmath0 curve around @xmath187 becomes less noticeable . to understand this
, we note that the kink in the curve is caused by the discontinuity in the survival rate of @xmath2 in the region of the fireball where the temperature is higher than the dissociation temperature and in the region where the temperature is lower than the dissociation temperature .
as higher - order corrections become large , the thermal decay width of @xmath2 increases and the survival rate of @xmath2 in the low temperature region becomes comparable to the survival rate in the high temperature region . as a result
, the kink disappears when higher - order corrections are large . among total @xmath2 in heavy - ion collisions .
solid , dashed , dotted , and dot - dashed lines correspond , respectively , to multiplicative factors ( a , b)=(1,1 ) , ( 2,2 ) , ( 5,4 ) and ( 10,6.5 ) for higher - order corrections.,width=302 ] figure [ portion ] shows the fraction of regenerated @xmath2 among total @xmath2 in heavy - ion collisions in the two - component model . without higher - order corrections ( solid line ) , the regenerated @xmath2 is about 31% of total @xmath2 produced in central collisions .
including higher - order corrections , the percentage increases to 52 , 81 and 96% , respectively , for multiplicative factors ( a , b)=(2,2 ) , ( 5,4 ) and ( 10,6.5 ) , shown by dashed , dotted , and dot - dashed lines , respectively .
we note that the faction of regenerated @xmath2 is smaller in semi - central and peripheral collisions than in central collisions . of @xmath2 as a function of transverse momentum in collisions of 0 - 20% ( a ) , 20 - 40% ( b ) , 40 - 60% ( c ) and 60 - 92% ( d ) centralities .
solid , dashed , dotted , and dot - dashed lines correspond , respectively , to multiplicative factors ( a , b)=(1,1 ) , ( 2,2 ) , ( 5,4 ) and ( 10,6.5 ) for the higher - order corrections.,width=302 ] figure [ raa - pt2 ] shows the @xmath0 of @xmath2 as a function of the transverse momentum in collisions of 0 - 20% ( a ) , 20 - 40% ( b ) , 40 - 60% ( c ) and 60 - 92% ( d ) centralities .
the lines in each figure correspond , respectively , to multiplicative factors ( a , b)=(1,1 ) , ( 2,2 ) , ( 5,4 ) and ( 10,6.5 ) for the higher - order corrections .
as we can see , @xmath0 does not change much even with large higher - order corrections as a result of the large transverse flow of regenerated @xmath2 . as shown in fig .
[ raa - pt ] , the @xmath132 dependence of the @xmath0 of initially produced @xmath2 is similar to that of regenerated @xmath2 except at high @xmath132 where the @xmath0 of initially produced @xmath2 increases while that of regenerated @xmath2 decreases . as a result ,
the @xmath0 at high @xmath132 decreases when higher - order corrections are included , and this seems to contradict with the @xmath132 dependence of measured @xmath0 , which is an increasing function at high @xmath132 . although higher - order corrections are small at high @xmath132 , they are relatively large at low @xmath132 , and this seems to reproduce the observed decreasing @xmath0 with increasing @xmath132 , when @xmath132 is low , in peripheral collisions as shown in fig .
[ raa - pt2](d ) .
this is mainly due to the fact that in peripheral collisions , the time interval between the thermalization of fireball and the freeze - out of @xmath2 is short , so the transverse flow is not sufficiently developed to give a large transverse flow velocity to the regenerated @xmath2 . in au+au collision .
solid , dashed , dotted , and dot - dashed lines correspond , respectively , to multiplicative factors ( a , b)=(1,1 ) , ( 2,2 ) , ( 5,4 ) and ( 10,6.5 ) for the higher - order corrections.,width=302 ] in fig .
[ pt2b ] , we show the average squared transverse momentum of @xmath2 before and after including higher - order corrections .
it is seen that higher - order corrections have very little effect on the results .
of @xmath2 in collisions at @xmath131 fm with higher - order corrections .
solid , dashed , dotted , and dot - dashed lines correspond , respectively , to multiplicative factors ( a , b)=(1,1 ) , ( 2,2 ) , ( 5,4 ) and ( 10,6.5 ) for the higher - order corrections .
the double - dot - dashed line is the @xmath1 of @xmath2 if they are totally from regeneration.,width=302 ] figure [ v2-jpsi2 ] shows the @xmath1 of @xmath2 in collisions at @xmath131 fm without and with higher - order corrections .
different from the @xmath0 of @xmath2 , the change in @xmath1 due to higher - order corrections is significant . for increasing large correction factors listed in the above ,
the fraction of regenerated @xmath2 among total @xmath2 is 12% , 23% , 42% and 63% respectively .
this can be easily understood from the fact that the @xmath1 of initially produced @xmath2 is very small . in terms of the @xmath1 of
initially produced @xmath2 and that of regenerated @xmath2 , the @xmath1 of total @xmath2 is given by @xmath188 where @xmath189 , @xmath190 and @xmath191 are , respectively , the total number of @xmath2 , the number of initially produced @xmath2 and that of regenerated @xmath2 . equation ( [ combination ] ) shows that the @xmath1 of total @xmath2 in heavy - ion collisions can be approximated by the @xmath1 of regenerated @xmath2 multiplied by its fraction among total @xmath2 .
the double - dot - dashed line in fig .
[ v2-jpsi2 ] is the @xmath1 of @xmath2 if all @xmath2 are produced from regeneration .
it is worthwhile to point out that even in this extreme case , the @xmath1 of @xmath2 still only reaches the upper error bar of experimental data , which has a value of @xmath192 between @xmath182 and @xmath183 gev @xcite .
if the experimental data on @xmath1 of @xmath2 is correct , it then means that the fraction of regenerated @xmath2 is not small even in semi - central collisions .
we have investigated in the two - component model the @xmath0 of @xmath2 as a function of transverse momentum and as a function of the number of participants as well as the @xmath1 of @xmath2 in relativistic heavy ion collisions .
firstly , a schematic model was constructed to describe the expansion of the fireball formed in these collisions .
the transverse acceleration of the fireball was assumed to be proportional to the outgoing force given by the product of the pressure of the hot dense matter inside the fireball and the surface area of the fireball , and inversely proportional to an inertia mass , which was taken to be a free parameter to fit the experimental data on the measured transverse momentum spectra of light hadrons .
the anisotropic expansion in non - central collisions was modeled by simply adding or subtracting an additional term proportional to the eccentricity of the fireball to the acceleration in @xmath108-direction or @xmath87-direction .
the coefficient of the additional term was determined such that the @xmath1 of light hadrons was reproduced after including the high-@xmath132 correction to the flow velocity due to the non - equilibrium effect .
next , the @xmath0 and @xmath1 of @xmath2 were calculated with this schematic model .
both the cronin effect due to the initial transverse momentum broadening of gluons in nuclei and the leakage effect due to the escape of initially produced @xmath2 from produced hot dense matter were included , and it was found that the @xmath0 was suppressed at low @xmath132 and enhanced at high @xmath132 by both effects . however , the leakage effect was much smaller than the cronin effect . for the kinetic freeze - out temperature of @xmath2 ,
it was taken to be the same as the chemical freeze - out temperature , because the elastic cross section of @xmath2 is much smaller than that of light hadrons , which normally has a lower kinetic freeze - out temperature than the chemical freeze - out temperature .
finally , higher - order corrections in pqcd were included simply by multiplying separate constant factors to the dissociation cross sections of charmonia and to the elastic cross section of charm quarks .
the former is related to the thermal decay widths of charmonia and the latter to the relaxation factor of charm quarks .
higher - order corrections were , however , not applied to the screening mass .
we found that higher - order corrections enhanced both the dissociation of charmonia and the relaxation of charm quarks .
the former decreases the number of initially produced @xmath2 while the latter increases the number of regenerated @xmath2 . as a result
, the total number of @xmath2 was found to change very little after including higher - order corrections . however , the fraction of initially produced @xmath2 and that of regenerated ones were affected significantly by higher - order corrections with larger higher - order corrections leading to a larger fraction of regenerated @xmath2 among total @xmath2 .
this is the reason why the @xmath0 of @xmath2 as a function of the number of participants was insensitive to the higher - order corrections , and both the two - component model and the statistical model can describe successfully the dependence of @xmath0 on the number of participants . applying same higher - order corrections to the transverse momentum dependence of the @xmath0 of @xmath2 , we found that it was also insensitive to the fraction of regenerated @xmath2 among the total @xmath2 except at high @xmath132 , as the small transverse momentum of regenerated @xmath2 is compensated by the large transverse flow velocity . on the other hand ,
the @xmath1 of @xmath2 is sensitive to the fraction of regenerated @xmath2 because the @xmath1 of initially produced @xmath2 is almost zero .
if the elastic cross section of @xmath2 in qgp is not small , the initially produced @xmath2 can then acquire nonnegligible @xmath1 through scattering with thermal partons .
however , initially produced @xmath2 is expected to have a smaller binding energy and larger radius because of the high temperature .
the loosely bound @xmath2 is expected to be dissociated rather than scattered elastically when colliding with thermal partons . for regenerated @xmath2
, they were assume to be produced from thermalized charm quarks in qgp .
it is not clear yet how close charm quarks are from thermal equilibrium in heavy - ion collisions .
recent experimental data on non - photonic single electrons , which are produced through the decay of heavy mesons , suggest that charm quarks are thermalized significantly @xcite .
if charm quarks are thermalized , resulting thermally produced @xmath2 would share their thermal properties and thus acquire appreciable @xmath1 .
a possible future analysis would be to include the effect of abrupt decrease in mass of the charmonia near @xmath3 @xcite , as decreased mass will enhance the yield of regenerated @xmath2 and hence change its total contribution to @xmath1 .
presently , the @xmath1 of @xmath2 has large error bars in the experimental data .
if refined and precise data on the @xmath1 of @xmath2 become available , they are expected to play an important role in discriminating between the two production mechanisms for @xmath2 , i.e. , initial production and regeneration .
this work was supported in part by the u.s . national science foundation under grant no .
phy-0758115 , the welch foundation under grant no .
a-1358 , the korean ministry of education through the bk21 program , and the korea research foundation under grant no .
krf-2006-c00011 .
let the semimajor and the semiminor of an elliptic fireball produced in heavy - ion collisions be @xmath114 and @xmath113 , respectively , and the fluid velocities along these two axes be @xmath123 and @xmath125 , respectively , at a certain time during the evolution of the fireball .
a short time @xmath193 later , the semimajor becomes @xmath194 and the semiminor @xmath195 .
suppose that a certain point on the ellipse , @xmath196 , moves to a slightly expanded ellipse , @xmath197 , in radial direction keeping the same polar angle @xmath118 .
then the velocity of the point @xmath198 satisfies where @xmath200^{-1/2}\nonumber\end{aligned}\ ] ] is the distance from the center of the ellipse to the point on the ellipse along the polar angle @xmath118 . expanding eq .
( [ ellipse ] ) with respect to the infinitesimal time interval @xmath193 gives @xmath201=1,\nonumber\end{aligned}\ ] ] defining @xmath203 and @xmath134 , eq .
( [ flow - v ] ) can be rewritten as @xmath204.\end{aligned}\ ] ] because high-@xmath132 particles are unlikely to follow the collective motion , both @xmath205 and @xmath206 can be functions of @xmath132 . for simplicity , we modify only @xmath206 by multiplying it with the factor @xmath133 $ ] , where @xmath207 is a fitting parameter @xcite . | using a two - component model for charmonium production , which includes contributions from both the initial hard nucleon - nucleon scattering and from the regeneration in the quark - gluon plasma , we study the nuclear modification factor @xmath0 and elliptic flow @xmath1 of @xmath2 in relativistic heavy ion collisions . for the expansion dynamics of produced hot dense matter , we introduce a schematic fireball model with its transverse acceleration determined from the pressure gradient inside the fireball and azimuthally anisotropic expansion parameterized to reproduce measured @xmath1 of light hadrons .
we assume that light hadrons freeze out at the temperature of 120 mev while charmonia at 160 mev , similar to the kinetic and chemical freeze - out temperatures in the statistical model , respectively .
for the properties of charmonia in the quark - gluon plasma , we use the screening mass between their charm and anticharm quarks and their dissociation cross sections given by the perturbative qcd ( pqcd ) in the leading order and up to the next - to - leading order , respectively . for the relaxation time of charm and anticharm quarks in the quark - gluon plasma
, we also use the one calculated in the leading order of pqcd . modeling the effect of higher - order corrections in pqcd by introducing multiplicative factors to the dissociation cross section of charmonia and the elastic scattering cross sections of charm and anticharm quarks
, we find that this effect is small for the @xmath0 of @xmath2 as they suppress the number of initially produced @xmath2 but enhance the number of regenerated ones .
the higher - order corrections increase , however , the @xmath1 of @xmath2 .
our results suggest that the @xmath1 of @xmath2 can play an important role in discriminating between @xmath2 production from the initial hard collisions and from the regeneration in the quark - gluon plasma . |
categorification of representations of affine quantum groups has proved to be an important tool for understanding many classic objects arising from modular group representation theory , among which hecke algebras and rational cherednik algebras of cyclotomic type , and finite classical groups .
more precisely , the study of crystals and canonical bases of the level @xmath2 fock space representations @xmath3 of @xmath4 gives answers to several classical problems in combinatorial terms .
in particular , we know that the @xmath4-crystal graph of @xmath3 can be categorified in the following ways : * by the parabolic branching rule for modular cyclotomic hecke algebras @xcite , when restricting to the connected component containing the empty @xmath2-partition , * by bezrukavnikov and etingof s parabolic branching rule for cyclotomic rational cherednik algebras @xcite , * by the weak harish - chandra modular branching rule on unipotent representations for finite unitary groups @xcite , for @xmath5 and @xmath6 varying . in each case , the branching rule depends on some parameters that are explicitly determined by the parameters @xmath7 and @xmath6 of the fock space .
recently , there has been some important developments when shan and vasserot @xcite categorified the action of the heisenberg algebra on a certain direct sum of fock spaces , in order to prove a conjecture by etingof @xcite .
losev gave in @xcite a formulation of shan and vasserot s results in terms of crystals , as well as an explicit formula for computing it in a particular case . independently and using different methods , the author defined a notion of heisenberg crystal for higher level fock spaces @xcite , that turns out to coincide with losev s crystal .
an explicit formula was also given in another particular case , using level - rank duality .
like the @xmath4-crystal , the heisenberg crystal gives some information at the categorical level .
in particular , it yields a characterisation of * the finite dimensional irreducible modules in the cyclotomic cherednik category @xmath8 by @xcite and @xcite , * the usual cuspidal irreducible unipotent modular representations of finite unitary groups @xcite .
this paper solves two remaining problems about the heisenberg crystal .
firstly , even though it originally arises from the study of cyclotomic rational cherednik algebras ( it is determined by combinatorial versions of certain adjoint functors defined on the bounded derived category @xmath8 ) , the heisenberg crystal has an intrinsic existence as shown in @xcite .
therefore , it is natural to ask for an algebraic construction of the heisenberg crystal which would be independent of any categorical interpretation .
this is achieved via the boson - fermion correspondence and the use of the schur functions , acting on uglov s canonical basis of @xmath3 .
this gives a new realisation of the heisenberg crystal , analogous to kashiwara s crystals for quantum group .
secondly , we give an explicit formula for computing the heisenberg crystal in full generality .
this generalises and completes the particular results of @xcite and @xcite .
this is done in the spirit of @xcite where formulas for the @xmath4-crystal were explicited .
the present paper has the following structure . in section 2 , we start by introducing in detail several combinatorial objects indexing the basis of the wedge space ( namely charged multipartitions , abaci and wedges ) and the different ways to identify them
. then , we quickly recall some essential facts about the @xmath4-structure of the fock spaces @xmath3 .
section 3 focuses on the heisenberg action on the wedge space @xmath9 , seen as a certain direct sum of level @xmath2 fock spaces @xmath3 .
in particular , we recall the definition of the heisenberg crystal given in @xcite .
then , we give in section 4 a solution to the first problem mentioned above .
more precisely , we recall the quantum boson - fermion correspondence and fundamental facts about symmetric functions . inspired by shan and vasserot @xcite and leclerc and thibon @xcite , we study the action of the schur functions on the wedge space and use a result of iijima @xcite to construct the heisenberg crystal as a version of this action at @xmath1 ( theorem [ thmschur ] ) , resembling kashiwara s philosophy of crystal and canonical bases .
most importantly , by doing so , we bypass entirely shan and vasserot s original categorical construction .
section 5 is devoted to the explicit computation of the heisenberg crystal .
we introduce level @xmath2 vertical @xmath7-strips , as well as the notion of good addable and removable vertical strips by defining an appropriate order .
the action of the heisenberg crystal operators is then given in terms of adding / removing good level @xmath2 vertical @xmath7-strips ( theorem [ thmheis ] ) , which is reminiscent of the explicit formula for the kashiwara crystal operators first given in @xcite ( see also @xcite ) .
we also give some examples of explicit computations of the heisenberg crystal .
finally , we recall some useful facts about level - rank duality in appendix , enabling the definition of the heisenberg crystal .
fix once and for all elements @xmath10 and @xmath11 .
an _ @xmath2-partition _ ( or simply _ multipartition _ ) is an @xmath2-tuple of partitions @xmath12 .
these will be represented using young diagrams .
denote @xmath13 the set of @xmath2-partitions and @xmath14 the set of partitions .
partitions will sometimes be denoted multiplicatively for convenience , e.g. @xmath15 .
an _ @xmath2-charge _ ( or simply _
( multi)charge _ ) is an @xmath2-tuple of integers @xmath16 .
we write @xmath17 .
we call _ charged @xmath2-partition _ the data consisting of an @xmath2-partition @xmath18 and an @xmath2-charge @xmath6 , and denote it by @xmath19 . for a box @xmath20 in the young diagram of @xmath18 ( where @xmath21 is the row index of the box , @xmath22 the column index and @xmath23 the coordinate ) , let @xmath24 , the _ content _ of @xmath25
. we will represent @xmath19 by filling the boxes of the young diagram of @xmath18 with their contents .
[ exachargedmp ] take @xmath5 , @xmath26 and @xmath27 .
we have @xmath28 and @xmath29 in the following , we will only consider multicharges @xmath6 verifying @xmath30 . for a partition @xmath31 ,
let @xmath32 denote its conjugate , that is the partition obtained by swapping rows and columns in the young diagram of @xmath31 .
we extend this to charged multipartitions by setting @xmath33 where @xmath34 and @xmath35 .
a charged @xmath2-partition @xmath19 can also be represented by the @xmath36-graded @xmath2-abacus @xmath37 where @xmath38 denotes the @xmath39-th part of @xmath40 . in the rest of this paper
, we will sometimes identify @xmath19 with @xmath41 .
[ exaabacus ] with @xmath19 as in example [ exachargedmp ] , we get the following corresponding abacus @xmath42 which we picture as follows , by putting a ( black ) bead at position @xmath43 where @xmath44 and @xmath23 is the row index ( numbered from bottom to top ) : [ ab2 ] note that the conjugate a multipartition can be easily described on the abacus : it suffices to rotate it by 180 degrees around the point of coordinates @xmath45 and swap the roles of the white and black beads . using the abaci realisation of charged multipartitions
, we define below a bijection @xmath46 which can be seen as a twisted version of taking the @xmath2-quotient and the @xmath2-core of a partition , see ( * ? ? ?
* chapter 1 ) .
however , unlike the usual @xmath2-quotient and @xmath2-core , @xmath47 and @xmath48 will depend not only on @xmath2 , but also on @xmath7 .
writing down the euclidean division first by @xmath49 and then by @xmath7 , one can decompose any @xmath50 as @xmath51 with @xmath52 , @xmath53 and @xmath54 .
we can then associate to each pair @xmath55 the pair @xmath56 where @xmath57 in particular , @xmath47 sends the bead in position @xmath58 into the rectangle @xmath59 , on the row @xmath60 and column @xmath61 ( numbered from right to left within each rectangle ) .
the map @xmath47 is bijective and we can see @xmath48 as the following procedure : 1 .
divide the @xmath2-abacus into rectangles of size @xmath62 , where the @xmath63-th rectangle ( @xmath64 ) contains the positions @xmath65 for all @xmath66 and @xmath67 .
2 . relabel each @xmath65 by the second coordinate of @xmath68 , see figure [ ab9 ] for an example .
replace the newly indexed beads on a @xmath69-abacus according to this new labeling .
-abacus according to @xmath48 , for @xmath70 and @xmath71 . ]
explicitly , @xmath47 and @xmath48 verify the following formulas : @xmath72 and @xmath73 we go back to example [ exaabacus ] .
the action of @xmath47 and @xmath48 are represented below .
[ ab4 ] let @xmath74 denote the set of sequences of integers @xmath75 such that @xmath76 for @xmath39 sufficiently large , and set @xmath77 [ defiwp ] an _ elementary wedge _
( respectively _ ordered wedge _ ) is a formal element @xmath78 where @xmath79 ( respectively @xmath80 ) .
let @xmath19 be a charged @xmath2-partition with @xmath81 , and let @xmath82 be the partition such that @xmath83 .
set @xmath84 where @xmath85 for all @xmath44 . in other terms
, @xmath86 is the sequence of integers indexing the beads of @xmath87 .
we clearly have @xmath88 , so we can consider the elementary wedge @xmath89 .
in fact , we will also identify @xmath19 with @xmath89 . to sum up
, we will allow the following identifications : @xmath90 we will denote @xmath91 and @xmath92 let @xmath0 be an indeterminate .
fix an @xmath2-charge @xmath6 .
the _ fock space _ associated to @xmath6 is the @xmath93-vector space @xmath3 with basis @xmath94 .
[ thmjmmo ] the space @xmath3 is an integrable @xmath4-module of level @xmath2 .
the action of the generators of @xmath4 depends on @xmath6 and @xmath7 , and is given explicitly in terms of addable / removable boxes , see e.g. ( * ? ? ?
* section 6.2 ) . in turn
, this induces a @xmath4-crystal structure ( also called kashiwara crystal ) @xcite , usually encoded in a graph @xmath95 , whose vertices are the elements of @xmath94 , and with colored arrows representing the action of the kashiwara operators . an explicit ( recursive ) formula for computing this graph was first given in @xcite : two vertices are joined by an arrow if and only if one is obtained from the other by adding / removing a _ good _ box , see ( * ? ? ? * section 6.2 ) for details .
\(a ) at ( 5,6 ) @xmath96 ; ( b1 ) at ( 2.5,4 ) @xmath97 ; ( b2 ) at ( 7.5,4 ) @xmath98 ; ( c1 ) at ( 1,2 ) @xmath99 ; ( c2 ) at ( 4,2 ) @xmath100 ; ( c3 ) at ( 6,2 ) @xmath101 ; ( c4 ) at ( 9,2 ) @xmath102 ; ( c5 ) at ( 12,2 ) @xmath103 ; ( d1 ) at ( 0,0 ) @xmath104 ; ( d2 ) at ( 2,0 ) @xmath105 ; ( d3 ) at ( 4,0 ) @xmath106 ; ( d4 ) at ( 6,0 ) @xmath107 ; ( d5 ) at ( 8,0 ) @xmath108 ; ( d6 ) at ( 10,0 ) @xmath109 ; ( d7 ) at ( 12,0 ) @xmath110 ; ( d8 ) at ( 14,0 ) @xmath111 ; ( d9 ) at ( 16,0 ) @xmath112 ; \(a ) node [ font= , left ] ( b1 ) ; ( a ) node [ font= , above ] ( b2 ) ; ( b1 ) node [ font=,left ] ( c1 ) ; ( b1 ) node [ font= , above ] ( c2 ) ; ( b2 ) node [ font= , above ] ( c3 ) ; ( b2 ) node [ font= , above ] ( c4 ) ; ( c5 ) node [ font= , above ] ( d7 ) ; ( c1 ) node [ font= , above ] ( d1 ) ; ( c1 ) node [ font= , above ] ( d2 ) ; ( c2 ) node [ font= , above ] ( d3 ) ; ( c3 ) node [ font= , above ] ( d4 ) ; ( c4 ) node [ font= , above ] ( d5 ) ; ( c4 ) node [ font= , above ] ( d6 ) ; . [ uecrystal ] it has several connected components , each of which parametrised by its only source vertex , or _ highest weight vertex_. the decomposition of this graph in connected components reflects the decomposition of @xmath3 into irreducible representations ( which exists because @xmath3 is integrable according to theorem [ thmjmmo ] )
. denote @xmath9 the @xmath93-vector space spanned by the elementary wedges , and subject to the straightening relations defined in ( * ? ? ?
* proposition 3.16 ) .
this is called the _ space of semi - infinite @xmath0-wedges _ , or simply the _ wedge space _ , and the elements of @xmath9 will be called _
wedges_. using the straightening relations , a @xmath0-wedge can be expressed as a @xmath93-linear combination of ordered wedges .
in fact , one can show ( see ( * ? ? ?
* proposition 4.1 ) ) that the set of ordered wedges @xmath113 forms a basis of @xmath9 . 1 .
the terminology `` wedge space '' is justified by the original construction of this vector space in the level one case by kashiwara , miwa and stern @xcite . in this context
, one can first construct a quantum version of the usual @xmath39-fold wedge product ( or exterior power ) @xmath114 , where @xmath115 is the natural @xmath4-representation ( an affinisation of @xmath116 ) .
the space @xmath9 is then defined as the projective limit ( taking @xmath117 ) of @xmath118 . in the higher level case ,
the analogous construction was achieved by uglov @xcite .
using the bijection between ordered wedges and partitions charged by @xmath119 , one can see @xmath9 as a level @xmath69 fock space , whence the notation .
in fact , it is sometimes called the _ fermionic _ fock space .
[ thmwedgeuemod ] the wedge space @xmath9 is an integrable @xmath4-module of level @xmath2 .
the identification @xmath120 of section [ wedge ] yields the @xmath93-vector space decomposition @xmath121 by theorem [ jmmofockspace ] , this decomposition still holds as integrable @xmath4-module .
it turns out that the wedge space has some additional structure , namely that of a @xmath122-module , where @xmath122 is the ( quantum ) heisenberg algebra .
this has been first observed by uglov @xcite , generalising some results of lascoux , leclerc and thibon @xcite .
let us recall the definition of this algebra .
let us start by recalling the definition of the quantum heisenberg algebra .
[ defheis ] the _ ( quantum ) heisenberg algebra _ is the unital @xmath93-algebra @xmath122 with generators @xmath123 and defining relations @xmath124 = { \delta}_{m ,- m ' } m \frac{1-q^{-2me}}{1-q^{-2m}}\times\frac{1-q^{2m\ell}}{1-q^{2m}}\ ] ] for @xmath125 .
the elements @xmath126 are called _ bosons_. note that this is a @xmath0-deformation of the usual heisenberg algebra with relations @xmath127={\delta}_{m ,- m'}m$ ] .
[ thmboson ] the formula @xmath128 endows @xmath9 with the structure of @xmath122-module . for a proof
, we refer to ( * ? ? ?
* proposition 4.4 and 4.5 ) .
this is quite technical and is done in two distinct steps , the second of which requires the notion of _ asymptotic _ wedge .
this will be crucial in section [ cbschur ] .
[ corboson ] the action of the bosons preserves the level @xmath2 fock spaces @xmath3 for @xmath81 .
one can use the explicit action of @xmath126 on @xmath9 to show that it preserves the @xmath2-charges @xmath6 , or rely on uglov s argument ( * ? ? ?
* section 4.3 ) . as a consequence of this corollary
, @xmath129 also holds as @xmath122-module decomposition . for two partitions @xmath31 and @xmath130 , denote @xmath131 the reordering of @xmath132 and @xmath133 the partition @xmath134 .
extend these notations to @xmath2-partitions by doing these operations coordinatewise .
[ remadd ] this is not the usual notation for adding partitions or multiplying by an integer .
one recovers the usual one by conjugating we have @xmath135 we also define the notion of asymptotic charges , which will mostly be useful in section [ cbschur ] .
[ defasymptotic ] a wedge @xmath136 is called _ @xmath137-asymptotic _ if there exists @xmath138 such that the corresponding charged multipartition @xmath19 verifies @xmath139 for all @xmath140 .
a notion of crystal for the heisenberg algebra , or @xmath122-crystal , has been indepedently introduced by losev @xcite and the author @xcite .
explicit formulas for computing this crystal have been given in some particular cases : asymptotic in @xcite and _ doubly highest weight _ in @xcite .
recall the definition of the @xmath122-crystal according to @xcite .
this requires the crystal level - rank duality exposed in appendix .
[ defdhw ] let @xmath141 , which we identify with its level - rank dual charged @xmath7-partition .
we call @xmath19 a _ doubly highest weight vertex _ if it is a highest weight vertex simultaneously in the @xmath4-crystal and in the @xmath142-crystal
. some important properties of doubly highest weight vertices are exposed in ( * ? ? ?
* section 5.2 ) .
in particular , an element @xmath19 is a doubly highest weight vertex if and only if it has a totally @xmath7-periodic @xmath2-abacus and a totally @xmath2-periodic @xmath7-abacus , according to a result by jacon and lecouvey @xcite , see the definition therein .
moreover , every bead of a given period encodes a same part size in @xmath18 , and one can define the partition @xmath143 as @xmath144 where @xmath145 is the part encoded by the @xmath39-th period .
[ defbsigma ] let @xmath146 .
the heisenberg crystal operator @xmath147 ( respectively @xmath148 ) is the uniquely determined map @xmath149 ( respectively @xmath150 ) such that 1 .
if @xmath19 is a doubly highest weight vertex , then @xmath151 is obtained from @xmath19 by shifting the @xmath39-th period of @xmath2-abaci of @xmath19 by @xmath152 steps to the right ( respectively to the left when possible , and @xmath153 otherwise ) , 2 .
it commutes with the kashiwara crystal operators of @xmath4 and of @xmath142 .
[ defhhw ] we say that @xmath154 is a highest weight vertex for @xmath122 if @xmath153 for all @xmath146 . note that each @xmath155 is well defined since ( 2 ) allows to define @xmath156 even when @xmath19 is not a doubly highest weight vertex .
moreover , ( * ? ? ?
* theorem 7.6 ) claims that in the asymptotic case , the heisenberg crystal operators coincide with the combinatorial maps introduced by losev @xcite .
note also that the definition of @xmath157 for doubly highest weight vertices induces ( by ( 2 ) of definition [ defbsigma ] ) a surjective map @xmath158 such that @xmath159 is a highest weight vertex for @xmath122 for all @xmath160 .
in order to have a description of the @xmath122-crystal similar to the @xmath4-crystal , see figure [ uecrystal ] , we wish to define the @xmath122-crystal as a graph whose arrows have minimal length . therefore , we define the following maps @xmath161 where @xmath162 with @xmath163 being the addable box of @xmath157 verifying @xmath164 and where @xmath165 with @xmath163 being the removable box of @xmath157 verifying @xmath166 , [ defhcrystal ] the @xmath122-crystal of the wedge space @xmath9 is the graph @xmath167 with 1 .
vertices : all charged @xmath2-partitions @xmath19 with @xmath160 and @xmath81 .
arrows : @xmath168 if @xmath169 . by definition ,
each @xmath170 is a composition of maps @xmath155 .
conversely , each @xmath147 is a composition of maps @xmath171 , see ( * ? ? ?
* formula ( 6.17 ) ) .
we will also call the maps @xmath171 heisenberg crystal operators .
[ prophcrystal ] 1 . the @xmath122-module decomposition @xmath172 induces a decomposition of @xmath167 .
each connected component of @xmath167 is isomorphic to the young graph .
a vertex @xmath19 is source in @xmath167 if and only if @xmath19 is a highest weight vertex for @xmath122 .
the depth of an element @xmath19 in @xmath167 is @xmath173 . we know from definition [ defbsigma ] that the action of the crystal heisenberg operators preserves the multicharge , proving ( 1 ) . in particular
, there is a notion of @xmath122-crystal for @xmath3 , which we denote @xmath174 .
moreover , the @xmath122-crystal is characterised by being the preimage of a @xmath175-crystal on the set of partitions under a certain bijection , depending on @xmath157 , see ( * ? ? ?
* remark 6.16 ) .
the @xmath175-crystal is exactly the young graph on partitions , see figure [ younggraph ] , where the arrows are colored by the contents of the added boxes , which proves ( 2 ) .
point ( 3 ) is clear by definition , and ( 4 ) follows from the definition of @xmath157 .
\(a ) at ( 4,8 ) @xmath176 ; \(b ) at ( 4,6 ) @xmath177 ; ( c1 ) at ( 2,4 ) @xmath178 ; ( c2 ) at ( 6,4 ) @xmath179 ; ( d1 ) at ( 1,2 ) @xmath180 ; ( d2 ) at ( 4,2 ) @xmath181 ; ( d3 ) at ( 7,2 ) @xmath182 ; ( e1 ) at ( 0,0 ) @xmath183 ; ( e2 ) at ( 2,0 ) @xmath184 ; ( e3 ) at ( 4,0 ) @xmath185 ; ( e4 ) at ( 6,0 ) @xmath186 ; ( e5 ) at ( 8,0 ) @xmath187 ; \(a ) node [ font= , left ] ( b ) ; ( b ) node [ font=,left ] ( c1 ) ; ( b ) node [ font= , above ] ( c2 ) ; ( c1 ) node [ font= , above ] ( d1 ) ; ( c1 ) node [ font= , above ] ( d2 ) ; ( c2 ) node [ font= , above ] ( d3 ) ; ( c2 ) node [ font= , above ] ( d2 ) ; ( d1 ) node [ font= , above ] ( e1 ) ; ( d1 ) node [ font= , above ] ( e2 ) ; ( d2 ) node [ font= , above ] ( e2 ) ; ( d2 ) node [ font= , above ] ( e3 ) ; ( d2 ) node [ font= , above ] ( e4 ) ; ( d3 ) node [ font= , above ] ( e4 ) ; ( d3 ) node [ font= , above ] ( e5 ) ;
denote @xmath188 the algebra of symmetric functions , that is , the projective limit of the @xmath93-algebras of symmetric polynomials in finitely many indeterminates ( * ? ? ?
* chapter 1 ) : @xmath189^{{\mathfrak{s}}_\infty}.\ ] ] the space @xmath188 has several natural linear bases , among which : * the monomial functions @xmath190 where @xmath191 where the sum runs over all permutations @xmath192 of @xmath193 , * the complete functions @xmath194 where @xmath195 and @xmath196 for @xmath197 , * the power sums @xmath198 where @xmath199 and @xmath200 for @xmath197 , * the schur functions @xmath201 where @xmath202 where @xmath203 are the kostka numbers , see ( * ? ? ?
* chapter 7 ) .
the expansion of @xmath204 in the basis of the power sums is given by @xmath205 where @xmath206 with @xmath207 , for all @xmath208 .
moreover , by duality , the kostka numbers also appear as the coefficients of the complete functions in the basis of the schur functions : @xmath209 by a result of miwa , jimbo and date @xcite , there is a vector space isomorphism @xmath210 called the _ boson - fermion _ correspondence .
in fact , when refering to the symmetric realisation @xmath188 , one sometimes uses the term _ bosonic _ fock space , as opposed to the fermionic , antisymmetric definition of @xmath9 .
the following result is ( * ? ? ?
* section 4.2 and proposition 4.6 ) .
[ propheis ] there is a @xmath122-module isomorphism @xmath211 , where , at @xmath212 , the action of @xmath126 on @xmath9 is mapped to the multiplication by @xmath213 on @xmath188 .
in general , the action of @xmath126 is mapped to a @xmath0-deformation of the multiplication by @xmath213 , and @xmath214 to a @xmath0-deformation of the derivation with respect to @xmath213 , see ( * ? ? ? * section 5 ) and ( * ? ? ? * section 5.1 ) . in @xcite , leclerc and
thibon studied the action of the heisenberg algebra on level @xmath69 fock spaces in order to give an analogue of lusztig s version @xcite of the steinberg tensor product theorem .
their idea was to use a different basis of @xmath188 to compute the heisenberg action in a simpler way , namely that of schur functions .
this result has been generalised to the level @xmath2 case by iijima in a particular case @xcite .
independently , the schur functions have been used by shan and vasserot to categorify the heisenberg action in the context of cherednik algebras .
more precisely , they constructed a functor on the category @xmath8 for cyclotomic rational cherednik algebras corresponding to the multiplication by a schur function on the bosonic fock space @xmath188 ( * ? ? ?
* proposition 5.13 ) .
the aim of this section is to use iijima s result to recover in a direct , simple way the results of @xcite and @xcite and , by doing so , bypassing shan and vasserot s categorical constructions . in the early nineties , kashiwara and lusztig
have independently introduced the notion of canonical bases for irreducible highest weight representations of quantum groups , see e.g. @xcite .
they are characterised by their invariance under a certain involution .
uglov @xcite has proved an analogous result for the fock spaces @xmath3 , even though it is no longer irreducible .
we recall the definition of the involution on @xmath9 and uglov s theorem . for any @xmath197 and @xmath215 , let @xmath216 that is , the number of repetitions in @xmath217 . [ defibar ]
the bar involution is the @xmath93-vector space automorphism @xmath218 with @xmath219 where @xmath220 and @xmath221 for all @xmath222 according to the notations of section [ abaci ] .
the bar involution behaves nicely on the wedge space , in particular it preserves the level @xmath2 fock spaces @xmath3 for @xmath81 , and commutes with the bosons @xmath126 , this is ( * ? ? ?
* section 4.4 ) .
we can now state uglov s result , that is derived from the fact that the matrix of the bar involution is unitriangular .
[ thmcanonicalbases ] let @xmath223 such that @xmath81 . there
exist unique bases and of @xmath3 such that , for @xmath224 , 1 .
@xmath225 2 .
@xmath226 where @xmath227|{\boldsymbol{{\lambda}}},{\mathbf{s}}\rangle$ ] .
this result is compatible with kashiwara s crystal theory .
more precisely , each integrable irreducible highest weight @xmath4-representation is contained in @xmath9 for some @xmath119 , by taking the span of the vectors @xmath228 for @xmath81 , and it is proved in ( * ? ? ?
* section 4.4 ) that the bases @xmath229 restricted to @xmath230 coincide with kashiwara s lower and upper canonical bases .
therefore , we also call @xmath229 the ( lower or upper if @xmath231 or @xmath232 respectively ) canonical basis of @xmath3
. define the following operators on @xmath9 @xmath233 where @xmath234 are the inverse kostka numbers , that is , the entries of the inverse of the matrix of kostka numbers . by proposition [ propheis ] , at @xmath212 , the action of @xmath235 ( respectively @xmath236 ) corresponds to the multiplication by a complete function ( respectively schur function ) on @xmath188 through the boson - fermion correspondence .
[ thmschur ] 1 .
the operators @xmath236 induce maps @xmath237 on @xmath113 , preserving @xmath94 for @xmath81 .
2 . let @xmath238 be @xmath137-asymptotic .
we have @xmath239 with @xmath240 where @xmath241 if @xmath140 and @xmath242 , provided @xmath151 is still @xmath137-asymptotic . in particular , @xmath237 coincides with the heisenberg crystal operator @xmath147 in this case .
when @xmath243 , this says that @xmath237 shifts the @xmath7 rightmost beads in the @xmath137-th runner of @xmath41 , and one recovers losev s result , see for instance ( * ? ?
* example 7.3 ) . in a general manner
, we can identify a crystal map @xmath244 with an operator on the wedge space @xmath9 mapping an element of @xmath229 to an element of @xmath229 , for @xmath224 .
indeed , an element @xmath154 can be identified with @xmath245 . another way to see
this is to look at the action on @xmath245 and put @xmath1 or @xmath246 respectively in the resulting vector .
this provides the identification because of condition ( 2 ) of theorem [ thmcanonicalbases ] . as a matter of fact
, the action of @xmath236 on a canonical basis element turns out to have the desired form . indeed , by ( * ? ? ? * theorem 4.12 ) , we know that @xmath247 verifies @xmath248 where @xmath249 ( with @xmath250 as in the statement of theorem [ thmschur ] ) provided : * @xmath251 is @xmath7-regular for all @xmath252 * @xmath19 and @xmath151 are @xmath137-asymptotic . here ,
iijima s original statement has been twisted by conjugation , because the level - rank duality used in his result is the reverse of that of the present paper .
thus , the notion of @xmath7-restricted multipartition is replaced by @xmath7-regular . accordingly
, we use the upper canonical basis instead of the lower one .
this statement can be extended to an arbitrary @xmath137-asymptotic element @xmath253 , that is , without the restriction that each @xmath40 is @xmath7-regular . to do this , for all @xmath160 , let @xmath254 where @xmath255 is @xmath7-regular for all @xmath256 and @xmath257 if @xmath140 and @xmath258 .
then , for all @xmath146 , @xmath259 where @xmath260 , which we can compute by the preceding formula .
note that @xmath47 is the common addition of @xmath193 and @xmath192 , see remark [ remadd ] . for a general @xmath137-asymptotic vector @xmath19
, we write again @xmath261 . as explained in the beginning of the proof , this induces a crystal map @xmath262 , by additionnaly requiring that @xmath237 commutes with the kashiwara crystal operators .
in fact , when @xmath19 is @xmath137-asymptotic , the formula for @xmath237 is precisely the formula for the heisenberg crystal operator given in @xcite ( provided again that one twists by conjugating ) , which coincides with @xmath147 by ( * ? ? ? * theorem 7.6 ) .
this completes the proof .
as explained in the proof , with this approach , @xmath263 is identified with @xmath264 .
the map @xmath236 being an actual operator ( on the vector space @xmath9 ) , this justifies the terminology `` operator '' for the maps @xmath265 , thereby completing the analogy with the kashiwara crystal operators .
in this section , we give the combinatorial formula for computing the heisenberg crystal in full generality .
this completes the results of @xcite , where the asymptotic case is treated , and of @xcite where the case of doubly highest weight vertices ( in the level - rank duality ) is treated , see appendix for details . in the spirit of @xcite and @xcite
, we will express the action of the heisenberg crystal operators in terms of adding / removing certain vertical strips .
[ defvertstrip ] let @xmath19 be a charged @xmath2-partition . 1 . a _
( level @xmath2 ) vertical @xmath7-strip _ is a sequence of @xmath7 boxes @xmath266 such that no horizontal domino appears , i.e. there is no @xmath267 such that @xmath268 and @xmath269 .
moreover , a vertical @xmath7-strip is called _ admissible _ if : 1 . the contents of @xmath270
are consecutive , say @xmath271 for all @xmath267 .
2 . for all @xmath272
, we have @xmath273 .
a vertical @xmath7-strip @xmath274 is called _ addable _ ( respectively _ removable _ ) for @xmath19 if @xmath275 and @xmath276 is still an @xmath2-partition ( respectively @xmath277 and @xmath278 is still an @xmath2-partition ) . in the following
, we may drop the adjective `` admissible '' , since these are the only vertical strips we will consider .
[ remverticalstrip ] this is a generalisation to multipartitions of the usual notion of vertical strip , see for instance ( * ? ? ?
* chapter i ) .
[ exaverticalstrip ] let @xmath279 , @xmath280 and @xmath281 the only removable vertical @xmath7-strips of @xmath19 is [ cols="^,^ , < " , ] for a charged @xmath2-partition @xmath19 , let @xmath283 and @xmath284 be the set of addable and removable admissible vertical @xmath7-strips of @xmath19 respectively , and let @xmath285 be either @xmath283 or @xmath284 .
we will define a total order on @xmath285 . to do that
, we use a total order on the set of addable ( respectively removable ) boxes of @xmath19 .
let @xmath20 and @xmath286 be two boxes of @xmath19 , either both addable or removable . write @xmath287 if @xmath288 or @xmath289 and @xmath290 .
note that this is the total order used to define the good boxes in a charged @xmath2-partition , which characterises the kashiwara crystals , see ( * ? ? ?
* chapter 6 ) . by extension ,
let @xmath291 denote the lexicographic order induced by @xmath291 on @xmath7-tuples of boxes in a given charged @xmath2-partition .
this restricts to a total order on @xmath285 .
[ defgoodverticalstrip ] let @xmath19 be a charged @xmath2-partition .
* the first good addable ( respectively removable ) vertical @xmath7-strip of @xmath19 is the maximal element @xmath292 of @xmath285 with respect to @xmath291 , where @xmath293 ( respectively @xmath284 ) .
* let @xmath294 .
the @xmath39-th good addable ( respectively removable ) vertical @xmath7-strip of @xmath19 is the maximal element @xmath295 of @xmath296 with respect to @xmath291 , where @xmath293 ( respectively @xmath284 ) . in other terms , the greatest ( with respect to @xmath291 ) vertical strip is good , and another admissible vertical strip is good except if one of its boxes already belongs to a previous good vertical strip .
[ remgoodverticalstrip ] clearly , given a charged @xmath2-partition , there is a finite number of good vertical @xmath7-strips , say @xmath297 .
moreover , the @xmath39-th good vertical strip is well - defined and uniquely determined for each @xmath298 .
we go back to example [ exaverticalstrip ]
. then we have @xmath299 for all @xmath300 .
moreover , @xmath301 is obviously the only good removable vertical @xmath7-strip , and there are only 3 good addable vertical strips , namely @xmath302 , @xmath303 and @xmath304 .
[ deftc ] let @xmath305 .
define @xmath306 and @xmath307 by @xmath308 where @xmath309 ( respectively @xmath310 ) is obtained from @xmath18 by adding ( respectively removing , when possible ) recursively @xmath152 times the @xmath39-th good addable vertical @xmath7-strip , for all @xmath311
. if it is impossible to do the removing operation , we set @xmath312 .
[ lemheis1 ] the map @xmath313 is well defined .
we need to prove that there is always at least a good addable ( admissible ) vertical @xmath7-strip .
indeed , let @xmath19 be a charged @xmath2-partition and consider the sequence of boxes @xmath314 where @xmath315 is the height of the first component of @xmath18 .
then @xmath316 is clearly a vertical strip , admissible , and addable .
consider the set @xmath317 of addable admissible vertical strips .
@xmath317 is not empty since @xmath318 and is finite because @xmath18 has a finite number of boxes .
therefore , @xmath317 has a maximal element with respect to @xmath291 , which is obviously good .
[ corheis1 ] for all @xmath146 , the map @xmath319 is well defined . if @xmath320 , then @xmath321 and so is well defined .
so we assume @xmath322 , say @xmath323 .
first of all , lemma [ lemheis1 ] and the definition of @xmath324 implies that @xmath325 is well defined for all @xmath326 ( so in particular for @xmath327 ) .
it remains to prove that for all @xmath160 and for all @xmath2-charge @xmath6 , @xmath328 has a second good addable vertical @xmath7-strip .
the result follows then by induction because of the recursive definition of @xmath319 .
denote @xmath329 and @xmath330
. then @xmath331 by definition .
let @xmath332 denote the good first addable vertical strip of @xmath151 .
then the first good addable vertical strip of @xmath333 is @xmath334 with @xmath335 for some @xmath336 . in particular ,
since @xmath338 , the partition @xmath333 has at least @xmath7 boxes .
therefore , the sequence @xmath339 , where @xmath340 is the height of the first component of @xmath310 , is an addable admissible vertical strip verifying @xmath341 if @xmath342 denotes the content of the first element of @xmath274 . in particular , @xmath343 and @xmath344 .
consider the set @xmath345 of addable admissible vertical strips @xmath274 of @xmath333 such that @xmath346 and @xmath347 .
@xmath345 is not empty since @xmath348 and is finite because @xmath310 has a finite number of boxes .
therefore , @xmath345 has a maximal element with respect to @xmath291 , which is good .
that means that @xmath333 has a second good addable vertical strip .
take again the values of example [ exaverticalstrip ] .
then , for @xmath349 , we have @xmath350 each box of @xmath193 corresponds to a vertical @xmath7-strip , the matching being given by the colors
. we will prove the following theorem .
[ thmheis ] for all @xmath146 , we have @xmath351 the strategy for proving this result consists in starting from the doubly highest weight case , in which we know an explicit formula for @xmath155 .
then , we show that the @xmath352 commute with the kashiwara crystal operators for @xmath142 , then @xmath4 , and use the commutation of the kashiwara crystals and the @xmath122-crystal ( see definition [ defbsigma ] ) to conclude .
[ propheis1 ] let @xmath19 be a doubly highest weight vertex .
then , for all @xmath146 , we have @xmath353 we need to translate the explicit formula for @xmath155 , given in terms of abaci , in terms of @xmath2-partitions . by the correspondence given in section [ mpwedge ] ,
an @xmath7-period in the @xmath2-abacus @xmath354 * corresponds to a removable admissible vertical @xmath7-strip if @xmath355 where @xmath356 is the last bead of the period .
* yields an addable admissible vertical @xmath7-strip if @xmath357 where @xmath358 is the first bead of the period .
therefore , shifting the @xmath39-th shiftable period of @xmath359 by @xmath152 steps to the right ( respectively to the left ) amounts to adding ( respectively removing ) the @xmath39-th addable ( respectively removable ) vertical strip of @xmath19 .
in other terms @xmath155 is the same as @xmath352 when restricted to doubly highest weight vertices .
[ propheis2 ] let @xmath19 be a highest weight vertex for @xmath4 .
then , for all @xmath146 , we have @xmath353 write @xmath360 where @xmath361 is the highest weight vertex for @xmath142 associated to @xmath19 , and where @xmath362 is a sequence of kashiwara crystal operators of @xmath142 . because of theorem [ thmlevelrank ] , the two kashiwara crystals commute , and thus @xmath361 is a doubly highest weight vertex .
we prove the result by induction on @xmath197 .
if @xmath363 , then @xmath19 is a doubly highest weight vertex and this is proposition [ propheis1 ] .
suppose that the result holds for a fixed @xmath364 .
write @xmath365 , so that @xmath366 .
because the crystal level - rank duality is realised in terms of abaci , we need to investigate one last time the action of @xmath367 in the abacus . we know that @xmath367 acts on an @xmath7-partition by adding its good addable @xmath368-box ( i.e. of content @xmath368 modulo @xmath2 ) .
this corresponds to shifting a particular ( white ) bead one step up in the @xmath7-abacus of @xmath333 , see example [ exalevelrank ] for an illustration .
now , if @xmath369 , then this corresponds to moving down a ( black ) bead in the @xmath2-abacus .
since the resulting abacus @xmath41 is again totally @xmath7-periodic ( the two kashiwara crystals commute ) , this preserves the @xmath7-period containing this bead . if @xmath370 , then moving this white bead up corresponds to moving a black bead in position @xmath371 in the @xmath2-abacus ( which is the first element of its @xmath7-period ) down to position @xmath372 .
again , this preserves the @xmath7-period . in both cases , the reduced @xmath368-word in the @xmath7-abacus ( see ( * ? ? ?
* section 4.2 ) for details ) is unchanged and @xmath373 therefore , we have @xmath374 finally , since @xmath148 and @xmath375 have the property to be inverse to @xmath147 and @xmath319 respectively , we also have @xmath376 .
we are now ready to prove theorem [ thmheis ] .
it remains to investigate the action of the kashiwara crystal operators of @xmath4 .
_ proof of theorem [ thmheis ] . _ as in the proof of proposition [ propheis2 ] , it suffices to show it for @xmath147 and @xmath319 .
write @xmath377 where @xmath378 is the highest weight vertex for @xmath4 associated to @xmath19 , and where @xmath379 is a sequence of kashiwara crystal operators of @xmath4 .
we prove the result by induction on @xmath197 .
if @xmath363 , then @xmath19 is a highest weight vertex for @xmath4 and this is proposition [ propheis2 ] .
suppose that the result holds for a fixed @xmath364 .
write @xmath380 , so that @xmath381 .
consider the reduced @xmath382-word for @xmath333 .
again , it is preserved by the action of @xmath319 by property ( 1 ) of definition [ defvertstrip ] , and we have @xmath383 the commutation of @xmath147 with @xmath384 together with the induction hypothesis completes the proof .
@xmath385 [ remheis ] theorem [ thmheis ] also yields an explicit description of the operators @xmath171 .
it acts on any charged @xmath2-partition by adding its @xmath39-th good vertical @xmath7-strip , where @xmath386 is the content of the @xmath39-th addable box of @xmath157 ( with respect to the order @xmath291 on boxes ) .
we end this section by mentioning an alternative realisation of the @xmath122-crystal .
in fact , some of the combinatorial procedures require some conventional choices , such as the maps @xmath47 and @xmath387 yielding the level - rank duality , or the order on the boxes or on the vertical strips of multipartition .
like in the case of kashiwara crystals , see e.g. ( * ? ? ?
* remark 3.17 ) or ( * ? ? ?
* remark 4.9 ) choosing a different convention yields to a slightly different version of the heisenberg crystal .
we have already seen in the proof of theorem [ thmschur ] that conventions can be adjusted to fit losev s @xcite or iijima s @xcite results about the action of a heisenberg crystal operator or a schur function respectively .
this is done by using the conjugation of multipartitions .
more precisely , one can decide to identify a charged @xmath2-partition @xmath19 with @xmath388 instead of @xmath389 with the notations of section [ abaci ] .
then , the action of the heisenberg crystal operators is expressed in terms of addable / removable horizontal @xmath7-strips in the @xmath2-partition .
this is equivalent to changing the order on the vertical strips , applying a heisenberg crystal operator , and then conjugate .
one could also decide to exchange the role of @xmath47 and @xmath387 in the level - rank duality , see appendix .
then , applying a heisenberg crystal operator on an @xmath2-abacus would de decribed in terms of the corresponding @xmath7-abacus .
more precisely , @xmath171 would then consists in shifting a @xmath2-period in the @xmath7-abacus in the particular case where it is totally @xmath2-periodic ( see also the proof of proposition [ propheis1 ] ) .
this permits to give an interpretation of tingley s tightening procedure on descending @xmath2-abaci ( * ? ? ?
* definition 3.8 ) , thereby answering question 1 of ( * ? ? ? * section 6 ) .
[ proptighten ] let @xmath359 be a descending @xmath2-abacus .
for all @xmath311 , denote @xmath390 the @xmath39-th tightening operator associated to @xmath359 .
we have @xmath391 where @xmath386 is determined by @xmath39 .
recall that for this statement , we have considered the realisation of the heisenberg crystal with the alternative version of level - rank - duality , swapping the roles of @xmath47 and @xmath387 , so that the operators @xmath392 act on the @xmath7-abacus ( rather than the @xmath2-abacus ) by removing a vertical @xmath2-strip ( theorem [ thmheis ] ) .
remember that @xmath393 where @xmath394 depends on @xmath386 .
now , if an @xmath2-abacus is descending ( * ? ? ?
* definition 3.6 ) , its corresponding @xmath7-abacus is totally @xmath2-periodic . in particular ,
the corresponding @xmath7-partition is a highest weight vertex in the @xmath142-crystal , and we can use proposition [ propheis2 ]
. in particular , @xmath395 and @xmath396 act by shifting @xmath2-periods in the @xmath7-abacus , and @xmath392 shifts one @xmath2-period , say the @xmath39-th one , one step to the left in the @xmath7-abacus .
this precisely what @xmath390 does .
finally , we mention that in another particular case , the heisenberg crystal operator @xmath395 coincide with the canonical @xmath4-isomorphism @xmath397 ( up to cyclage ) of ( * ? ? ?
* theorem 5.26 ) used to construct an affine robinson - schensted correspondence .
[ propreduc ] let @xmath154 be a doubly highest weight vertex .
write @xmath143 .
we have @xmath398 where @xmath399 is the number of parts of @xmath157 and @xmath400 is the cyclage isomorphism , see ( * ? ? ?
* proposition 4.4 ) .
we use the notations of @xcite .
first of all , because of ( * ? ? ?
* proposition 5.7 ) , doubly highest weight vertices are cylindric in the sense of ( * ? ? ?
* definition 2.3 ) , and @xmath401 is therefore well defined , and simply verifies @xmath402 where @xmath403 is the number of pseudoperiods in @xmath19 and @xmath404 is the reduction isomorphism .
in fact , by definition of @xmath157 , we have @xmath405 , the number of ( non - zero ) parts of @xmath157 , and it suffices to apply the cyclage @xmath399 times to @xmath19 to match the formulas for @xmath395 and @xmath406 . by remark [ remheis ] ,
the heisenberg crystal can be computed recursively from its highest weight vertices , each of which yields a unique connected component , isomorphic to the young graph by proposition [ prophcrystal ] .
the empty multipartition is obviously a highest weight vertex for @xmath122 , and so is every multipartition with less than @xmath7 boxes .
for instance , if @xmath5 , @xmath407 and @xmath71 , we can compute the connected components of the heisenberg crystal of @xmath3 with highest weight vertex @xmath408 and @xmath409 . up to rank @xmath410 , we get the following subgraph of the @xmath122-crystal .
[ ex1 ] for @xmath411 and @xmath412 , the @xmath413-partition @xmath414 is clearly a highest weight vertex for @xmath122 by theorem [ thmheis ] .
the corresponding connected component , up to rank @xmath415 is the following graph .
[ ] we also wish to give an example of the asymptotic case .
take @xmath279 , @xmath416 , @xmath417 and @xmath71 , so that @xmath19 is a highest weight vertex for @xmath122 and is @xmath413-asymptotic .
we see in the following corresponding @xmath122-crystal that the elements @xmath282 , for @xmath418 , only act on the third component of @xmath19 , but that @xmath419 acts already on the second component .
this illustrates theorem [ thmschur ] ( 2 ) .
we recall , using a slightly different presentation , the results of ( * ? ? ?
* section 4 ) concerning the crystal version of level - rank duality .
there is a double affine quantum group action on the wedge space . in section [ jmmofockspace ] ,
we have explained how @xmath4 acts on @xmath9 .
it turns out that @xmath142 , where @xmath420 , acts on @xmath9 in a similar way .
more precisely we will : 1 . index ordered wedges by charged @xmath7-partitions , using an alternative bijection @xmath387 , 2 .
make @xmath142 act on @xmath9 via this new indexation by swapping the roles of @xmath7 and @xmath2 and replacing @xmath0 by @xmath421 . 1 .
divide the @xmath2-abacus into rectangles of size @xmath62 , where the @xmath63-th rectangle ( @xmath64 ) contains the positions @xmath65 for all @xmath66 and @xmath67 .
2 . relabel each @xmath65 by the second coordinate of @xmath425 , see figure [ ab8 ] for an example .
replace the newly indexed beads on a @xmath69-abacus according to this new labeling .
we see that @xmath424 , so also @xmath387 , actually only depends on @xmath7 , not on @xmath2 .
in fact , explicit formulas for @xmath387 and @xmath426 are given by @xmath427 and @xmath428 and one notices that @xmath387 corresponds to taking the usual @xmath7-quotient and the @xmath7-core of a partition , where the @xmath7-core is encoded in the @xmath7-charge .
more precisely , the renumbering of the beads according to @xmath387 is the well - know `` folding '' procedure used to compute the @xmath7-quotient , see @xcite .
the map @xmath430 already defines a level - rank duality , this was the one studied by uglov @xcite .
however , the twisted version defined above ( i.e. where the conjugation is added ) is the one that is relevant when it comes to crystals , see theorem [ thmlevelrank ] below . there is a convenient way to picture the crystal level - rank duality as follows . starting from an @xmath2-abacus , stack copies on top of each other by translating by @xmath7 to the right . then , extract a vertical slice of the resulting picture , and read off the corresponding @xmath7-partition by looking at the white beads ( instead of black beads ) in each column , starting from the leftmost one .
we have the following induced maps @xmath434{\tau^{-1 } } & { \mathcal{f}}_s & \xrightleftharpoons[\quad ' \quad ] { ' } & { \mathcal{f}}_{-s } & \xrightleftharpoons[{\dot{\tau}}^{-1}]{\hspace{4mm}{\dot{\tau}}\hspace{4 mm } } & { \displaystyle}\bigoplus_{|{\dot{{\mathbf{s}}}}|=s } { \mathcal{f}}_{{\dot{{\mathbf{s } } } } } \\ like @xmath47 , the bijection @xmath387 induces a @xmath142-module isomorphism , and @xmath9 has a @xmath142-crystal , given by the same rule as the @xmath4-crystal by swapping the roles of @xmath7 and @xmath2 and replacing @xmath0 by @xmath421 .
the first point is essentially due to uglov ( * ? ? ?
* proposition 4.6 ) , where he uses the non - twisted version of level - rank duality , see ( * ? ? ?
* theorem 3.9 ) for the justification in the twisted case .
the second point is ( * ? ? ?
* theorem 4.8 ) . | we explain how the action of the heisenberg algebra on the space of @xmath0-deformed wedges yields the heisenberg crystal structure on charged multipartitions , by using the boson - fermion correspondence and looking at the action of the schur functions at @xmath1 .
in addition , we give the explicit formula for computing this crystal in full generality . |
the energy of a map from one manifold to another is a measure of the total stretching of the map .
energy minimizing harmonic maps have found numerous applications in geometry , analysis , algebra and topology .
the two survey papers of eells and lemaire @xcite give an introduction to the extensive literature of this subject .
one of the fundamental existence theorems states that a nontrivial map of a riemannian manifold @xmath0 to a negatively curved manifold @xmath1 is homotopic to a unique harmonic map .
when @xmath0 is a surface , each conformal class of metrics on @xmath0 gives rise to the same map , and smooth families of domains give rise to smooth families of harmonic maps @xcite . in this paper
we introduce a new type of energy that leads to what we call a _ simplicial harmonic map_. the key idea is to give a new , more combinatorial , definition of area , and then find a corresponding definition for energy .
the payoff is that the existence and regularity of simplicial harmonic maps are simple to prove , but they retain enough of the features of smooth harmonic maps to be useful in applications . in this paper
we focus on maps from surfaces into non - positively curved manifolds and spaces .
we briefly discuss extensions to maps between manifolds of any dimension .
one motivation for studying energy from a more combinatorial point of view is to obtain numerical methods for computing minimal and harmonic surfaces .
this was carried out in the work of pinkall and polthier @xcite .
see @xcite for a survey of related work .
the discrete harmonic maps of pinkall and polthier are maps of a triangulation of a domain surface into @xmath2 that are linear on each @xmath3simplex .
their formula gives the natural energy for a piecewise - linear approximation of a smooth map , and has found many uses in computational geometry @xcite , where harmonic maps give a preferred choice of surface map for purposes such as texturing and meshing .
while the pinkall - polthier harmonic maps retain many of the useful features of smooth harmonic maps , particularly when the @xmath3simplices in the domain are acute angled , they do not in general satisfy the convex hull or mean value properties .
( see section [ meanvalueandconvexhullproperties ] for a discussion of these properties and their consequences . )
an example of polthier and rossman @xcite shows that a discrete harmonic map may fail to satisfy the convex hull property , with an interactive demonstration available at @xcite .
the simplicial harmonic maps we describe in this paper are defined for very general target spaces and do satisfy a convex hull property .
a consequence is that if @xmath4 is a simplicial harmonic immersion of a surface @xmath0 into a riemannian manifold @xmath1 , then the induced curvature on @xmath0 is more negative than the sectional curvature of @xmath1 .
when @xmath1 itself is negatively curved , the gauss - bonnet theorem then implies that the area of @xmath4 is bounded above by a constant times the euler characteristic of @xmath0 .
when mapped into euclidean space , simplicial harmonic maps also satisfy a mean value property and the maximum principle . in this paper
, we use the completeness and non - positive curvature of a riemannian manifold @xmath1 to conclude that each homotopy class of arcs in @xmath1 contains a unique geodesic representative ( rel boundary ) .
there are other conditions that imply this property , and our construction then immediately applies .
for example , teichmuller space with either the teichmuller metric or the weil - peterson metric has a unique geodesic connecting any two points @xcite , although the weil - peterson metric is not complete .
similarly , any two points in the interior of a hemisphere are connected by a unique geodesic arc .
more generally , our methods work equally well if @xmath1 is a metrized graph , a @xmath5 space , or a path metric space with the property that any two points are joined by a unique geodesic in each homotopy class .
this paper is organized as follows .
initially we consider maps of triangulated compact surfaces into manifolds . in section [ simplicialarea ]
we define the simplicial area of a map , and in section [ simplicialenergy ] we define simplicial energy . in section [ polygonaldecompositions ]
we briefly discuss how to define simplicial area and energy of maps when the source surface is divided into polygons which need not be triangles . in section [ simplicialharmonic ]
we define simplicial harmonic maps . in section [ meanvalueandconvexhullproperties ]
we establish the mean value and convex hull properties of simplicial harmonic maps .
in section [ families ] we consider the problem of deforming families of surfaces in @xmath6manifolds to surfaces of small area . given any surface in a @xmath6manifold , it is possible to find a homotopy that deforms the surface to a collection of minimal surfaces joined by thin tubes . in a hyperbolic @xmath6manifold , there is a universal upper bound for the area of such a surface in terms of its euler characteristic @xmath7 ; a surface with @xmath8 can be homotoped to have area at most @xmath9 .
but there is no canonical homotopy to such a small area surface . for many applications one would like to start with a family of surfaces and
continuously homotope the entire family of surfaces so that each surface in the family has small area .
area deformation techniques based on mean curvature and related flows are not well suited to such a process , due to the formation of singularities and the non - uniqueness of least area surfaces in a homotopy class . in @xcite
it was shown that smooth energy deformation techniques based on analytic results of eells and sampson @xcite can be applied successfully in this context .
we show that a simplified theory based on simplicial energy suffices to give a deformation of a family of surfaces to a family of small area surfaces
. then we extend this discussion to explain how the smooth energy deformation techniques in @xcite can be replaced by simpler simplicial techniques . in section [ genus ]
we introduce the idea of the genus of an @xmath10dimensional manifold , extending the notion of heegaard genus from dimension three to all dimensions .
we find connections between this @xmath10dimensional genus and the areas of surfaces in a sweepout of an @xmath10manifold .
the simplicial energy of a map depends on the choice of a simplicial metric on its domain . in section
[ limits ] we show the existence of maps which are global minimizers of simplicial energy over the space of all simplicial metrics . in section [ diffeo ]
we consider the problem of finding canonical parametrizations , or canonical triangulations of a surface .
we solve this by showing that if a simplicial harmonic map from a closed surface of negative euler characteristic to a non - positively curved riemannian surface is homotopic to a homeomorphism , then it is itself a homeomorphism .
then we give applications of this result to study the space of straight triangulations of a compact surface @xmath0 with a riemannian metric of non - positive curvature .
( a triangulation of @xmath0 is called straight if each edge is a geodesic arc . )
we show that the space of straight triangulations of @xmath0 of a fixed combinatorial type is connected .
we note that in the case when @xmath0 is the disk with a flat metric , bloch , connelly and henderson @xcite proved much more , showing that this space is homeomorphic to some euclidean space and so contractible . we thank igor rivin for telling us about this work , and for suggesting that our ideas could be useful for studying the space of straight triangulations of more general surfaces .
finally , in section [ higherdimensions ] we briefly discuss generalizations to maps of higher dimensional manifolds .
in this section we define the simplicial area of a compact @xmath3dimensional surface @xmath0 . in place of a riemannian metric on @xmath0 , we specify a triangulation @xmath11 and a map @xmath12 that assigns to each edge @xmath13 , @xmath14 , of @xmath11 a length @xmath15 , with the lengths @xmath16 satisfying the triangle inequality for each triangle of @xmath11 .
we call such an assignment a _ simplicial metric _ on @xmath0 , and denote it by @xmath17 , or just @xmath12 when the context is clear
. note that we allow the possibility that for some triangles one of the triangle inequalities is an equality .
we do not require the triangulation @xmath11 to be combinatorial .
thus we allow triangulations where two , or even all three , vertices of a single triangle coincide in @xmath0 , or where a pair of triangles intersect in two or three vertices .
but we assume that @xmath11 has the structure of a @xmath18complex @xcite .
this is not a serious restriction .
this means that @xmath0 is constructed from a disjoint union of triangles , each of whose vertices is ordered , by identifying faces of the triangles in an order preserving way .
thus the given orderings can be extended to an ordering of all the vertices of @xmath11 . for simplicity we will always discuss triangles in @xmath19complexes as if they were embedded in @xmath0 .
if a triangle is not embedded in @xmath0 , one needs to consider the source triangle in the initial disjoint union of triangles .
each simplex in a @xmath19complex has a minimal ( in the given order ) vertex , which we call its preferred vertex . when considering limits of simplicial metrics , it is convenient to also allow edges in the triangulation that are assigned zero length .
we call such an extension a _ simplicial quasi - metric _ on @xmath0 .
we wish to assign a meaning to the area of @xmath20 .
a natural choice for the area of a triangle @xmath21 with edge lengths @xmath22 is the area of the euclidean triangle with the same edge lengths .
this is given by the classical heron s formula , see @xcite for example , @xmath23 we will however use an alternate formula , one that is much simpler , but still useful . we define the _ simplicial area _
@xmath24 of the triangle @xmath21 by the formula@xmath25 we note that the euclidean area @xmath26 of a triangle is at most @xmath27 , for any pair @xmath28 , @xmath29 of edge lengths .
so @xmath30 , @xmath31 and @xmath32 .
equality occurs when an appropriate angle is a right angle .
adding these inequalities yields the inequality@xmath33 one defect of our definition of the simplicial area of a triangle , is that if we subdivide a triangle into smaller triangles , the simplicial area is not , in general , additive .
there is however a special case where additivity holds .
this is when we add a vertex at the midpoint of each edge of a euclidean triangle @xmath34 and subdivide @xmath21 into four similar triangles .
we call this _
conformal subdivision_. another defect of our definition is that the simplicial area of a triangle can only be zero when @xmath35 .
for some purposes , this is not a problem . in the cases where this is a serious problem , we use quadrilaterals rather than triangles .
see section [ polygonaldecompositions ] .
[ defnofsimplicialarea]the _ simplicial area _ of @xmath20 is @xmath36 where @xmath37 are the lengths of the edges of the @xmath3simplex @xmath21 of @xmath11 . unlike heron s formula , this formula makes sense even when the triangle inequality does not hold .
however the triangle inequality will prove useful , so we have built it into the definition of a simplicial metric .
we will see that simplicial area retains enough similarities to the standard notion of area to be useful in many applications .
we now use this idea to define a simplicial analogue of the area of a map .
let @xmath38 be a piecewise - smooth map of a metrized triangulated compact surface @xmath20 to a riemannian manifold @xmath1 and let @xmath39 denote the length of @xmath40 .
of course , this can only be defined when @xmath40 is rectifiable .
the simplicial area of a map @xmath41 is @xmath42 where the sum is taken over all @xmath3simplices of @xmath11 .
the simplicial area of @xmath4 does not depend on the simplicial metric @xmath12 , but it does depend on the triangulation of @xmath0 .
the energy of a smooth map @xmath43 of a riemannian surface @xmath44 to a riemannian manifold @xmath45 is @xmath46 in local coordinates the energy can be obtained by integrating the @xmath3form @xmath47 if we take @xmath48 to be an orthonormal frame in a neighborhood on @xmath44 , then this expression simplifies to @xmath49 we now define a corresponding notion of simplicial energy .
suppose that @xmath4 is a map from a metrized or quasi - metrized triangulated compact surface @xmath0 to a riemannian manifold @xmath1 .
if an edge @xmath13 has @xmath50 then we denote the ratio @xmath51 by @xmath52 , and call it the _ stretch factor _ of the edge @xmath13 under @xmath4 .
if @xmath13 has length @xmath53 then we define the stretch factor by @xmath54 if @xmath55 , and @xmath56 if @xmath57 .
by analogy with the definition of smooth energy , we define the formula for the simplicial energy of a map on a triangle by multiplying the squares of stretch factors by the local contribution to the area .
[ defnofsimplicialenergy]the _ simplicial energy _ of a piecewise - smooth map @xmath4 of a metrized triangulated compact surface @xmath20 to a riemannian manifold @xmath1 is @xmath58 where @xmath37 are the edge lengths of the @xmath3simplex @xmath21 of @xmath11 . in the case where some edge @xmath13 has length @xmath53 in a quasi - metric and this edge
is not mapped by @xmath4 to a point , so that @xmath59 and @xmath56 , we define the simplicial energy of @xmath4 to be infinite . for all other cases the above formula gives a well defined finite energy .
the simplicial energy of @xmath4 is invariant under scale change of the domain simplicial metric . if the simplicial metric @xmath12 is replaced by @xmath60 , resulting in multiplying each @xmath16 by a constant @xmath61 , then the energy is unchanged .
it is sometimes convenient to rewrite the simplicial energy as a sum over the edges of @xmath11 . for a finite energy map @xmath41
, this gives @xmath62 where the sum is over the edges @xmath63 of @xmath11 , and @xmath64 , @xmath65 , @xmath66 and @xmath67 are the lengths of the four edges of @xmath11 which are adjacent to @xmath13 in the two triangles of which @xmath13 is a face .
see figure [ triangle.fig ] . ,
width=192 ] the coefficient @xmath68 of @xmath69 is called the _ weight _ of @xmath13 .
note that for any simplicial metric on @xmath70 , each weight @xmath71 is positive
. other approaches to combinatorial harmonic maps have studied a variety of ways of assigning such weights @xcite .
for example the combinatorial energy of pinkall and polthier @xcite assigns to the edge @xmath13 the weight @xmath72 , where @xmath73 and @xmath74 are the two euclidean angles opposite the edge @xmath13 . see figure [ triangle.fig ] .
the cotangent formula goes back at least to duffin @xcite .
note that some of these weights will be negative if any of the triangles of @xmath11 has an obtuse angle .
our choice of weights was made because of its simplicity and its close connection to the previously defined simplicial area . for our purposes
, it also turns out to be important that all our weights are positive .
in fact , many of the results in this paper hold for any energy functional of the form @xmath75 , so long as all the weights are positive , though the connection between simplicial energy and area does not . for our next result , it will be convenient to introduce the following notation .
let @xmath0 be a surface with triangulation @xmath11 , and let @xmath12 be a simplicial quasi - metric on @xmath70 .
then we denote by @xmath76 the subcomplex of @xmath11 which consists of all zero length edges .
[ e > a ] let @xmath38 be a piecewise - smooth map of a connected triangulated compact surface @xmath70 to a riemannian manifold @xmath1 .
let @xmath12 be a simplicial quasi - metric on @xmath70 .
then @xmath77 further equality holds if and only the stretch factors @xmath52 are constant on the edges of each component of @xmath78 . in particular , if @xmath12 is a simplicial metric , so that no @xmath16 is zero , then equality holds if and only if all stretch factors @xmath52 take the same value on all edges of @xmath11 .
this result is the combinatorial analogue of the fact that if @xmath79and @xmath1 have riemannian metrics and @xmath38 is a smooth map , then the energy of @xmath4 is always at least as large as its area , with equality if and only if @xmath4 is conformal . the condition that the stretch factors @xmath52 take the same value on all edges of @xmath11 is a much weaker condition than conformality , but this may be an advantage for some purposes .
suppose that edges @xmath13 and @xmath80 lie on the boundary of a triangle in @xmath11 .
if @xmath16 and @xmath81 are positive , we apply the inequality @xmath82 , with equality precisely when @xmath83 .
we obtain @xmath84 with equality precisely when @xmath85 .
if some @xmath16 is zero , we have two cases .
if @xmath53 , and @xmath57 , then @xmath52 is infinite , so that @xmath86 . if @xmath87 is finite , @xmath53 , and @xmath55 , then @xmath54 and the same inequality holds as both sides are zero .
thus this inequality holds in all cases .
summing over all triangles of @xmath0 shows that @xmath88 .
if @xmath89 , and @xmath76 is empty , the connectedness of the @xmath90skeleton of @xmath11 immediately implies that all stretch factors @xmath52 take the same value on all edges of @xmath11 .
if @xmath76 is non - empty , the connectedness of the @xmath90skeleton of the dual cellulation of @xmath11 implies that all the stretch factors @xmath52 are constant on the edges of each component of @xmath78 .
we now give an example that shows some of the difficulties in achieving useful simplicial versions of area and energy .
the example illustrates that it is not possible to capture all aspects of how a map stretches a surface using only information about the stretching of edges in a fixed triangulation .
[ example.stretch ] measuring stretching along the boundary of a triangle fails to capture dirichlet energy .
let @xmath91 be a simplicial map from an isosceles triangle with edges of length @xmath92 to an equilateral triangle with edge lengths all equal to one .
the euclidean area of @xmath93 is close to zero , while that of @xmath94 is @xmath95 .
the affine map @xmath4 that takes @xmath93 to @xmath94 stretches each of these edges by a factor of @xmath90 or less
. if we use a combinatorial definition of area that makes the area of @xmath93 close to zero , and a combinatorial definition of the energy of @xmath4 which depends only on how much @xmath4 stretches the edges of @xmath93 , then the energy of @xmath4 would also be close to zero .
thus the energy of @xmath4 would be much less than its area , which is the area of the image triangle @xmath94 , unlike the situation of lemma [ e > a ] .
the problem is that measuring the stretching of @xmath4 only on the boundary of @xmath93 ignores the large stretching by @xmath4 of segments in @xmath93 orthogonal to the long edge .
we conclude that reasonable formulas defining simplicial area and energy can not simultaneously satisfy all of the following : 1 .
simplicial energy is evaluated using stretching of edges and simplicial area .
simplicial energy is greater or equal to simplicial area .
the simplicial area of a triangle contained in a line is zero .
as we pointed out in section [ simplicialarea ] our definition of simplicial area only has the third property when every edge of a triangle has zero length . in section [ families ] , we will see how using quadrilaterals instead of triangles removes some of these difficulties .
in this section , we generalise the preceding two sections to the situation where a compact surface @xmath0 is divided into polygons rather than just triangles .
we again use @xmath11 to denote such a division .
we again have a map @xmath12 that assigns to each edge @xmath13 , @xmath96 , of @xmath11 a length @xmath15 . for each polygon of @xmath11 , we require that each edge has length no more than the sum of all the other edges of that polygon .
this suffices to ensure the existence of a convex embedding of the polygon in the euclidean plane .
if the polygon has more than three edges , we can not expect this embedding to be unique , but this is not important .
note that we allow the possibility that for some polygons one of these inequalities is an equality .
we call such an assignment a _ simplicial metric _ on @xmath0 , and denote it by @xmath20 , or just @xmath12 when the context is clear . as in the previous sections
, we do not require that each polygon of @xmath11 be embedded in @xmath0 .
we will also extend this to the case where some edges of @xmath11 are assigned zero length .
we call such an extension a _ simplicial quasi - metric _ on @xmath0 . throughout this section
, @xmath0 will be a compact surface with a polygonal decomposition @xmath11 and a simplicial metric or quasi - metric @xmath12 .
let @xmath97be a @xmath10gon , whose edges are @xmath98 with lengths @xmath99 reading round @xmath100 . thus @xmath16 and @xmath101 are lengths of adjacent edges , where we define @xmath102 to equal @xmath103 .
we define the _ simplicial area _ @xmath24 of this polygon by the formula@xmath104 as in the case of triangles , the simplicial area of polygons is not , in general , additive .
however for quadrilaterals there is a natural analogue of the conformal subdivision of a triangle which we described near the start of section [ simplicialarea ] .
but note that it may be that the process we describe does not correspond to subdividing a quadrilateral embedded in the plane .
given a quadrilateral @xmath100 , there is a natural way to divide it into four sub quadrilaterals , by cutting along two lines which join interior points of opposite edges of @xmath100 .
we now assign lengths to the new edges as shown in figure [ quads ] . as for triangles ,
the area of a polygon can be non - zero even if several edges have length zero .
however the following fact will be very useful .
if @xmath100 is a quadrilateral , and two opposite edges of @xmath100 are assigned zero length , then the simplicial area of @xmath100 is zero . for the applications in sections [ families ] and [ genus ] , this is all we will need .
the _ simplicial area _ of @xmath20 is @xmath105 let @xmath38 be a piecewise - smooth map to a riemannian manifold @xmath1 and let @xmath39 denote the length of @xmath40 . of course , this can only be defined when @xmath40 is rectifiable . for a @xmath10gon @xmath100 of @xmath11 , the _ simplicial area _ of a map @xmath106
is @xmath107 , which we denote by @xmath108 .
the _ simplicial area _ of a map @xmath41 is @xmath109 next we define the simplicial energy of @xmath4 . as for triangles ,
we use the stretch factor @xmath110 , when @xmath50 .
if @xmath53 we define the stretch factor by @xmath54 if @xmath55 , and @xmath111 if @xmath57 . for a @xmath10gon @xmath100 of @xmath11 ,
the _ simplicial energy _ of a map @xmath106 is @xmath112 , where @xmath113 equals @xmath114 .
the _ simplicial energy _ of a piecewise - smooth map @xmath4 of @xmath20 to a riemannian manifold @xmath1 is @xmath115 as for triangles , the above formula gives a well defined finite energy , unless some @xmath56 , when @xmath116 .
the simplicial energy of @xmath4 is invariant under scale change of the domain simplicial metric .
if the simplicial metric @xmath12 is replaced by @xmath60 , resulting in multiplying each @xmath16 by a constant @xmath61 , then the energy is unchanged .
it is sometimes convenient to rewrite the simplicial energy as a sum over the edges of @xmath11 . for a finite energy map @xmath41
, this gives @xmath117 where the sum is over the edges @xmath63 of @xmath11 , and @xmath64 , @xmath65 , @xmath66 and @xmath67 are the lengths of the four edges of @xmath11 which are adjacent to @xmath13 in the two polygons of which @xmath13 is a face .
see figure [ triangle.fig ] for the case when the polygons are triangles .
if @xmath12 is a simplicial quasi - metric on @xmath70 , then we denote by @xmath118 the subcomplex of @xmath11 which consists of all zero length edges .
the following result is the natural generalisation of lemma [ e > a ] , and is proved in the same way .
[ e > aforpolygons]let @xmath41 be a piecewise - smooth map of @xmath0 to a riemannian manifold @xmath1 .
then @xmath77 further equality holds if and only the stretch factors @xmath52 are constant on the edges of each component of @xmath78 .
in particular , if @xmath12 is a simplicial metric , so that no @xmath16 is zero , then equality holds if and only if all stretch factors @xmath52 take the same value on all edges of @xmath11 .
we call a map of a surface @xmath0 into a manifold @xmath1 _ trivial _ if the image of @xmath119 is contained in a cyclic subgroup of @xmath120 . when @xmath1 is non - positively curved , or more generally when @xmath121 , this is equivalent to the map factoring through a map of the circle into @xmath1 . in this section
we show that any homotopy class of maps from a metrized triangulated closed surface @xmath20 into a closed non - positively curved riemannian manifold @xmath1 contains a simplicial - energy minimizing map .
this map is unique if the map is nontrivial and the curvature of @xmath1 is strictly negative .
note that nontrivial maps never exist if @xmath0 is the sphere or projective plane .
further if the metric on @xmath1 is strictly negatively curved , then nontrivial maps do not exist if @xmath0 is the torus or klein bottle .
since @xmath1 is geodesically complete , any map @xmath122 can be homotoped , fixing the vertices of @xmath0 , to a new map that sends each edge of @xmath11 to a geodesic arc in @xmath1 .
if @xmath1 is non - positively curved , such an arc is unique . after picking an identification of each edge @xmath13 of @xmath11 with the interval
@xmath123 $ ] of the real line , this yields a canonical map from the @xmath90skeleton of @xmath11 to @xmath1 . the resulting map @xmath124 has energy no greater than the original map @xmath125 , and it suffices to consider such maps in searching for an energy minimizer .
the choice of an extension of @xmath124 to the @xmath3simplices of @xmath0 does not affect its energy , so we are free to make this choice in any convenient way .
to uniquely describe an extension , we make some further choices .
since the edge lengths of @xmath11 satisfy the triangle inequality , we can identify each triangle of @xmath11 with the euclidean triangle having the same edge lengths . as discussed at the start of section [ simplicialarea
] , each triangle of @xmath11 has a preferred vertex , and we fix the natural cone structure on the triangle with this vertex as cone point .
there is a unique extension of @xmath124 to a map of @xmath0 that sends each cone line in a triangle to a constant speed geodesic arc in @xmath1 .
if @xmath11 is a polygonal decomposition of @xmath0 , we proceed in the same way .
the only difference is that we need to choose a preferred vertex for each polygon , choose a convex embedding of each polygon in the euclidean plane , and then use the natural cone structure on the polygon with this vertex as cone point .
we make the following definition .
[ defnofsimplicialmap]a map @xmath41 from a triangulated surface @xmath0 with a simplicial metric @xmath12 to a riemannian manifold @xmath1 is _ simplicial _ if each edge of @xmath11 and each cone line in each @xmath3simplex of @xmath11 is mapped as a geodesic arc .
we can use the same definition and terminology if @xmath11 is a polygonal decomposition of @xmath0 .
the preceding discussion shows that if @xmath1 is non - positively curved , then any map @xmath122 can be homotoped , leaving the vertices fixed , to a unique simplicial map @xmath124 and that the simplicial energy of @xmath124 is no greater than that of @xmath125 .
thus to establish the existence of simplicial - energy minimizing maps into a non - positively curved manifold , we need only consider simplicial maps . given a simplicial map @xmath4 we define a _ variation _ of @xmath4 to be a @xmath90parameter family of maps @xmath126 with @xmath127 .
note that @xmath128 need not be simplicial when @xmath129 . a map @xmath41 of a closed surface with a simplicial metric or simplicial quasi - metric to a riemannian manifold @xmath1 is _ simplicial harmonic _ if it is simplicial and is a critical point of simplicial energy under all variations @xmath128 of @xmath4 .
if @xmath0 is compact with non - empty boundary , then @xmath4 is _ simplicial harmonic _ if it is simplicial and is a critical point of simplicial energy under all variations @xmath128 of @xmath4 which leave @xmath130 fixed .
the map @xmath4 is _ simplicial conformal _ if the stretch factors @xmath52 are constant on the edges of each component of @xmath78 , where @xmath76 denotes the subcomplex of @xmath11 which consists of all zero length edges .
lemmas [ e > a ] and [ e > aforpolygons ] tell us that if @xmath4 is a simplicial map , then @xmath89 if and only if @xmath4 is simplicial conformal . the above definition of @xmath4 being simplicial harmonic is equivalent to the one obtained by restricting attention to variations in which each map @xmath128 is simplicial . for most of this paper
, we will consider only the case when the surface @xmath0 is closed , but in the next section it will be important to consider the case when @xmath0 has non - empty boundary .
[ uniqueharmonicmapsexist ] let @xmath20 be a closed triangulated surface with a simplicial metric .
let @xmath41 be a continuous map of @xmath0 to a non - positively curved , closed riemannian manifold @xmath1
. then 1 .
@xmath4 is homotopic to a simplicial map @xmath125 that minimizes simplicial energy among all maps homotopic to @xmath4 , and hence is simplicial harmonic .
if @xmath4 is nontrivial and @xmath1 is negatively curved , then @xmath125 is the unique simplicial harmonic map in the homotopy class of @xmath4 .
furthermore , if @xmath12 is a simplicial quasi - metric and @xmath4 is homotopic to a map with finite energy , then ( 1 ) and ( 2 ) continue to hold .
\(1 ) any homotopy of a map @xmath38 that decreases some image length @xmath39 without increasing any @xmath131 reduces the simplicial energy ( see equation [ edgesum ] ) .
it follows that a map that minimizes simplicial energy sends each edge of @xmath11 to an arc in @xmath1 that minimizes length in its relative homotopy class .
a simplicial map is not determined by the images of the vertices of @xmath11 , since distinct relative homotopy classes of arcs connecting two points will yield distinct geodesic arcs connecting those points .
however a simplicial map @xmath4 is determined uniquely by the images of the vertices and the homotopy class ( rel boundary ) of @xmath4 on each edge of @xmath11 .
we select a vertex @xmath132 as a basepoint and also fix a fundamental region @xmath133 in the universal cover @xmath134 of @xmath1 .
since @xmath1 is compact , so is @xmath133 .
let @xmath135 be the set of simplicial maps homotopic to @xmath4 , let @xmath136 be a constant and let @xmath137 be the subset of @xmath135 consisting of maps with simplicial energy less than or equal to @xmath138 .
we will show that @xmath137 is compact .
let @xmath4 be a map in @xmath137 , let @xmath100 denote @xmath139 and let @xmath140 be a point of @xmath133 projecting to @xmath100 .
let @xmath141 denote the universal cover of @xmath0 , equipped with the natural triangulation lifted from @xmath11 , and let @xmath142 denote a point of @xmath141 above @xmath143 .
there is a unique map @xmath144 that covers @xmath4 and sends @xmath142 to @xmath140 .
as @xmath134 is simply connected , the simplicial map @xmath145 is determined completely by the images of the vertices of @xmath141 . each edge @xmath13 in @xmath11 is part of a simple arc @xmath146 contained in the @xmath90skeleton of @xmath11 and connecting @xmath13 to the base point @xmath143 .
the restriction of @xmath4 to @xmath146 lifts to a path @xmath147 in @xmath134 which starts at @xmath140 .
let @xmath148 be the length of the shortest edge in the given simplicial metric on @xmath11 , let @xmath149 be the length of the longest edge and let @xmath39 be the length of the restriction of @xmath4 to @xmath13 . as @xmath4 lies in @xmath137 , we have @xmath150 so that @xmath151 let @xmath152 denote the quantity @xmath153 , so that @xmath154 , for all @xmath155
. since @xmath146 is simple it contains at most @xmath156 edges , where @xmath156 is the total number of edges in @xmath11 .
hence the restriction of @xmath4 to @xmath146 has length at most @xmath157 .
note that this bound applies for any element @xmath4 of @xmath137 .
as the point @xmath158 lies in the compact fundamental domain @xmath133 , there is a ball @xmath159 in @xmath134 such that each @xmath147 lies in @xmath159 , for any @xmath160 .
it follows that the simplicial map @xmath145 is completely determined by the finite number of points in @xmath159 that are vertices of the @xmath147 s .
the set of all such is compact , therefore a minimum energy map exists , which establishes part ( 1 ) .
note that the above argument does not really use the non - positive curvature assumption on @xmath1 . without that assumption ,
one can still find an energy minimizing map homotopic to @xmath4 , but may not be able to arrange that this map is simplicial as defined in definition [ defnofsimplicialmap ] .
\(2 ) now suppose that @xmath4 is nontrivial , and that @xmath125 and @xmath124 are simplicial harmonic maps in the homotopy class @xmath135 of @xmath4 .
further suppose that @xmath1 is negatively curved .
as @xmath125 and @xmath124 are homotopic , there is a straight homotopy @xmath128 between them , in which @xmath128 is a simplicial map for each @xmath161 and the images of vertices of @xmath11 move along geodesic arcs in @xmath1 at constant speed .
we will make use of the following key fact .
let @xmath162 and @xmath163 denote two geodesic arcs in the negatively curved manifold @xmath1 and let @xmath164 be the shortest geodesic arc from @xmath162 to @xmath163 .
then the length @xmath165 of @xmath164 is a convex function of @xmath166 .
further it is strictly convex except possibly when the arcs @xmath162 , @xmath163 and @xmath164 are all contained in a single geodesic @xcite . since each @xmath167
is convex we see that the simplicial energy function @xmath168 given in equation [ edgesum ] is also a convex function of @xmath166 .
now suppose that @xmath125 and @xmath124 do not coincide .
then there is a vertex @xmath143 of @xmath11 such that the homotopy @xmath128 from @xmath125 to @xmath124 moves @xmath143 along a nontrivial geodesic arc @xmath169 .
we let @xmath170 denote the unique complete geodesic of @xmath1 which contains @xmath169 .
as @xmath4 is nontrivial , there is an edge @xmath13 of @xmath11 whose image under @xmath125 or @xmath124 is not contained in @xmath170 .
first we suppose that @xmath143 is a vertex of @xmath13 . in this case
, the function @xmath167 is a strictly convex function of @xmath166 , and it follows immediately that @xmath171 is also strictly convex , and so can have only one critical point . since both @xmath125 and @xmath124 are critical points , they must coincide , proving part ( 2 ) of the proposition in this case . however it may be the case that every edge of @xmath11 incident to @xmath143 has image which remains in @xmath170 under each map @xmath128 . here
we need a slightly refined version of the above key fact .
let @xmath162 and @xmath163 denote two geodesic arcs in the negatively curved manifold @xmath1 which are contained in a single geodesic @xmath170 , and suppose , in addition , that @xmath172 and @xmath173 each has constant speed .
let @xmath164 be the shortest geodesic arc from @xmath162 to @xmath163 .
then the length @xmath165 of @xmath164 is a linear function of @xmath166 .
hence @xmath174 is a strictly convex function of @xmath166 , unless @xmath165 is constant .
now we return to the situation where @xmath143 is a vertex of @xmath11 such that the homotopy @xmath128 moves @xmath143 along a nontrivial geodesic arc contained in @xmath170 , there is an edge @xmath13 of @xmath11 whose image under @xmath125 or @xmath124 is not contained in @xmath170 , and @xmath143 is not a vertex of @xmath13 .
choose a simple path in the @xmath90-skeleton of @xmath11 that connects @xmath143 to @xmath13 , let @xmath80 denote the first edge of this path and let @xmath175 denote the other vertex of @xmath80 .
the preceding discussion shows that @xmath176 is a strictly convex function of @xmath166 unless @xmath177 stays in @xmath170 for all @xmath166 , and @xmath178 is constant . in the first case
, it follows as before that @xmath171 is a strictly convex function of @xmath166 . in the second case
, it follows that @xmath175 , like @xmath143 , is a vertex of @xmath11 such that the homotopy @xmath128 from @xmath125 to @xmath124 moves @xmath175 along a geodesic arc contained in @xmath170 .
further this geodesic arc is nontrivial because @xmath178 is constant .
as @xmath175 is joined to @xmath13 by a shorter edge path than @xmath143 , a simple inductive argument completes the proof that @xmath171 is a strictly convex function of @xmath166 in all cases . as above this
implies that @xmath125 and @xmath124 coincide , which completes the proof of part ( 2 ) of the proposition . finally , we consider the case when @xmath12 is a quasi - metric
. it may no longer be true that a harmonic map homotopic to @xmath4 exists .
for example , @xmath12 might assign zero length to each edge of @xmath11 , which would mean that all maps homotopic to @xmath4 have infinite energy unless @xmath4 is homotopic to a point .
however if @xmath4 is homotopic to a map with finite energy , then we can take @xmath148 to be the smallest positive length of the edges in @xmath11 .
all zero length edges must be mapped to a point if the total energy is finite , and so they make no contribution to the energy . the previous argument
then applies and establishes that parts ( 1 ) and ( 2 ) of the theorem continue to hold in this case .
in this section , we show that simplicial harmonic maps share some of the classical properties of smooth harmonic maps .
let @xmath0 be a riemannian manifold .
a map @xmath179 has the _ mean value property _
if for any point @xmath180 , and any ball @xmath159 of radius @xmath156 centered at @xmath28 , the value of @xmath4 at @xmath28 is the average of the values of @xmath4 over @xmath181 . if @xmath4 is harmonic it has the mean value property .
harmonic maps to @xmath2 also have the following _ convex hull property _ : for @xmath182 any compact connected submanifold of @xmath0 , the image @xmath183 is contained in the convex hull of @xmath184 .
note that if @xmath0 is closed , then any harmonic map from @xmath0 to @xmath2 must be constant , so the discussion in this section is only of interest when @xmath79has non - empty boundary .
finally a real valued function @xmath4 on @xmath0 is said to satisfy the _
maximum principle _ if , for any compact submanifold @xmath182 of @xmath0 , the restriction of @xmath4 to @xmath182 attains its upper and lower bounds only on @xmath185 .
non - constant harmonic functions also satisfy the maximum principle . as mentioned in the introduction
, the discrete harmonic maps of pinkall and polthier @xcite do not satisfy the convex hull or mean value properties .
see @xcite for an interactive demonstration of an example found by polthier and rossman @xcite which shows that discrete harmonic maps do not satisfy the convex hull property .
next we show that simplicial harmonic maps do satisfy the convex hull and mean value properties .
let @xmath143 be a vertex of a surface triangulation @xmath11 , and let @xmath182 denote the closed star of @xmath143 , so that @xmath182 is the union of all the simplices of @xmath11 which contain @xmath143 .
we let @xmath186 denote the other vertices of @xmath182 .
[ convexhullproperty](euclidean mean value and convex hull properties ) let @xmath4 be a simplicial harmonic map of a connected metrized triangulated compact surface @xmath20 into @xmath2
. then @xmath139 lies in the convex hull of @xmath187 .
further @xmath139 is the centre of mass of a collection of positive weights placed at @xmath187 . equation ( 1 ) shows that the simplicial energy @xmath188 has the form @xmath189 , for some positive constants @xmath71 . as @xmath4 is harmonic ,
it is a critical point of the functional @xmath190 .
we will apply this fact to variations of @xmath4 through simplicial maps which equal @xmath4 on all vertices apart from @xmath143 . for @xmath191 ,
let @xmath80 denote the edge joining @xmath143 to @xmath192 , and consider the terms @xmath193 of @xmath171 .
let @xmath194 be the coordinates of @xmath139 , and let @xmath195 be the coordinates of @xmath196 , @xmath197 .
thus @xmath198 . as @xmath4 is a critical point of @xmath190 , we know that , for @xmath199 , the partial derivative of this summation with respect to @xmath200 must be zero .
hence , for @xmath199 , we have @xmath201 , so that @xmath202 . as
each @xmath203 is positive , @xmath200 is a linear combination of the @xmath204 s with coefficients which lie between @xmath205 and @xmath90 .
it follows that @xmath139 lies in the convex hull of @xmath187 , and that @xmath139 is the centre of mass of weights @xmath206 placed at @xmath187 . the special case of this result when @xmath207 gives the following form of the maximum principle .
( maximum principle ) let @xmath4 be a non - constant simplicial harmonic map of a connected metrized triangulated compact surface @xmath20 into @xmath208 .
then for any subsurface @xmath209 of @xmath0 which is also a subcomplex , the restriction of @xmath4 to @xmath209 attains its upper and lower bounds only on @xmath210 .
the convex hull of a subset of a general riemannian manifold @xmath1 may be rather complicated . for simplicity
we consider only the case of non - positive curvature and will restrict attention to convex balls in @xmath1 . in this
setting we obtain the following result . [ convexhull ] let @xmath4 be a simplicial harmonic map of a connected metrized triangulated compact surface @xmath20 into a convex ball @xmath159 with a non - positively curved riemannian metric .
let @xmath143 be an interior vertex of @xmath11 and let @xmath182 denote the closed star of @xmath143 .
let @xmath186 denote the other vertices of @xmath182 .
then @xmath139 lies in the convex hull of @xmath187 .
if @xmath139 lies outside the convex hull of @xmath187 , then there is a shortest geodesic arc @xmath172 from @xmath139 to the convex hull .
take the tangent plane @xmath100 at @xmath139 perpendicular to @xmath172 .
the geodesics joining @xmath139 to @xmath196 have tangent vectors at @xmath139 that all lie on the same side of @xmath100 . moving @xmath139 away from @xmath100 towards this side will reduce each @xmath131 and so reduce the energy of @xmath4 , contradicting the assumption that @xmath4 is simplicial harmonic .
lemma [ convexhull ] implies a universal area bound for homotopy classes of maps of a fixed surface into a negatively curved manifold .
we first recall the well - known corresponding bound for smooth maps .
[ arealessthan4pi(g-1)]let @xmath1 be a hyperbolic manifold , let @xmath0 be the closed orientable surface of genus @xmath211 , and let @xmath38 be any map .
then @xmath4 can be homotoped to a map @xmath125 with @xmath212 .
if @xmath1 has sectional curvature bounded above by @xmath213 then we have @xmath214 .
suppose that @xmath1 has sectional curvature bounded above by @xmath213 .
if @xmath4 sends a simple loop on @xmath4 to a trivial loop in @xmath1 , we compress @xmath4 .
otherwise @xmath4 is homotopic to a least area immersion in its homotopy class , using the existence theorems for minimal surfaces due to schoen and yau @xcite , and sacks and uhlenbeck @xcite .
note that as @xmath121 , the least area maps obtained in @xcite and @xcite must be homotopic to @xmath4 .
the normal curvature @xmath215 of a minimal immersion is non - positive .
its sectional curvature @xmath216 . by the gauss - bonnet theorem , with @xmath217 , @xmath218 so @xmath219 .
the homotopy class of the original uncompressed surface map can be recovered by adding tubes of zero area .
we reprove this result using simplicial harmonic maps . in doing
so we no longer need to use the existence results for minimal surfaces .
since our results apply equally well to non - orientable surfaces , we use the euler number rather than the genus of a surface .
[ arealessthan2pichi]let @xmath1 be a hyperbolic manifold , let @xmath0 be a closed surface with euler characteristic @xmath8 , and let @xmath38 be any map .
then @xmath4 can be homotoped to a map @xmath125 with @xmath220 .
if @xmath1 has sectional curvature bounded above by @xmath213 then we have @xmath221 .
suppose that @xmath1 has sectional curvature bounded above by @xmath213 .
choose a triangulation and a simplicial metric on @xmath0 , and let @xmath125 be a simplicial harmonic map in the homotopy class of @xmath4 .
if @xmath4 is trivial , then @xmath125 has image contained in some geodesic of @xmath1 and so has zero riemannian area . in this case
the required inequality is obvious .
otherwise @xmath4 is nontrivial , so that @xmath125 is unique and at least one @xmath3simplex of @xmath11 does not have image contained in a geodesic .
the simplicial map @xmath125 is a ruled smooth immersion on each such @xmath3simplex .
thus the riemannian metric induced on @xmath0 has curvature @xmath222 on the interior of such @xmath3simplices , and edges are sent to geodesics . at the vertices , the convex hull property in lemma [ convexhull ] shows that the angle sum around each vertex is greater than @xmath223 , and hence that the vertices contribute negative curvature to the total curvature when using the gauss - bonnet formula for simplicial complexes .
the argument now follows as in the smooth case .
the preceding lemma can be proved in essentially the same way if @xmath0 is divided into polygons . for
then each polygon has a cone structure , and so has a division into subcones which are triangles .
again each triangle either has image contained in a geodesic , or is mapped by a ruled smooth immersion into @xmath1 .
given an immersion of a surface into a riemannian manifold , the mean curvature flow gives an area decreasing flow that pushes the surface in the direction that decreases its area as rapidly as possible .
however singularities develop as the surface evolves under mean curvature , making it difficult to extend the flow beyond the time of singularity formation .
additional difficulties arise in extending mean curvature flow to families of surfaces , and these have proved an obstruction to topological applications .
harmonic maps offer an alternative approach to constructing area decreasing flows , a fact exploited in @xcite . in this section we show how to deform a multi - parameter family of simplicial maps to a family of simplicial harmonic maps . under this deformation the simplicial energy of each map in
the family decreases monotonically . the simplicial area may not be monotonically decreasing , but the simplicial area of each surface is at all times bounded above by its initial value .
although some of the results in this section work for triangulations , we will use only decompositions into quadrilaterals .
this is because of the applications we have in mind in this and the next section . for these applications ,
it is important that the riemannian area of a simplicial map is bounded by its simplicial energy .
note that no such bound can exist for general maps , as the simplicial energy of a map is determined solely by the restriction of that map to the @xmath90skeleton of the given triangulation .
[ e > standarda ] let @xmath0 be a compact surface with a decomposition @xmath11 into quadrilaterals and a simplicial quasi - metric @xmath12 .
let @xmath38 be a simplicial map to a non - positively curved riemannian manifold @xmath1 .
then the riemannian area , @xmath224 , satisfies the inequality@xmath225 let @xmath100 be a quadrilateral of @xmath11 , with edges @xmath226 cyclically ordered .
recall that there is a preferred vertex @xmath143 of @xmath100 , that we use a cone structure on @xmath100 with @xmath143 as cone point , and that a simplicial map from @xmath0 to @xmath1 sends each cone line in this cone structure to a geodesic in @xmath1 .
the diagonal of @xmath100 that has one end at @xmath143 divides @xmath100 into two triangles , and the restriction of @xmath4 to each triangle is a ruled @xmath3simplex in @xmath1 .
suppose that @xmath143 is the vertex @xmath227 , so that one triangle contains @xmath228 and @xmath229 , and the other triangle contains @xmath230 and @xmath231 .
as @xmath1 is non - positively curved , the induced curvature on each triangle is also non - positive and so the induced riemannian area is no greater than the euclidean area of a triangle with the same edge lengths .
the euclidean areas of the two triangles are at most @xmath232 and @xmath233 .
we conclude that the riemannian area of the restriction of @xmath4 to @xmath100 is at most @xmath234 , which is clearly at most one half of the simplicial area of the restriction of @xmath4 to @xmath100 .
now lemma [ e > aforpolygons ] implies that this is at most one half of the simplicial energy of the restriction of @xmath4 to @xmath100 . summing over all the quadrilaterals of @xmath11
immediately yields the inequality @xmath235 , as required .
let @xmath0 be a closed surface with a decomposition @xmath11 into quadrilaterals and a simplicial quasi - metric @xmath12 .
let @xmath41 be a continuous map of @xmath0 to a negatively curved , closed riemannian manifold @xmath1 .
if @xmath4 is nontrivial , then proposition [ uniqueharmonicmapsexist ] states that @xmath4 is homotopic to a unique simplicial harmonic map @xmath125 .
when @xmath4 itself is a simplicial map , there is a canonical way to deform it to the simplicial energy minimizing map @xmath125 .
the simplicial energy function is determined locally by the images under @xmath4 of the vertices of @xmath11 , and thus is a real valued function on a finite dimensional manifold .
the gradient of this function determines a direction of deformation for the vertices of @xmath236 .
this gives rise to an energy decreasing flow @xmath128 whose derivative at each @xmath166 equals the negative of the gradient of the simplicial energy function .
this flow moves the vertex images of @xmath237 so as to decrease simplicial energy as quickly as possible .
we define the _ simplicial energy gradient flow _ to be this flow of the map @xmath4 .
since there is a unique critical point for simplicial energy in any homotopy class , the gradient flow converges to @xmath125 , the minimizing simplicial harmonic map homotopic to @xmath4 .
we now show that the map @xmath125 obtained by minimizing energy among simplicial maps homotopic to an initial map @xmath4 depends continuously on the map @xmath4 and the simplicial metric @xmath12 .
this is a straightforward consequence of the uniqueness of simplicial harmonic maps within a homotopy class .
it was shown in @xcite that smooth harmonic maps depend continuously on the domain metric .
we use the term _ continuous family of maps @xmath238 of a surface @xmath0 to a manifold @xmath1 parametrized by a space _ @xmath239 to refer to a continuous function @xmath240 , with @xmath241 , for each @xmath242 and @xmath180 .
[ gradientflow]let @xmath0 be a closed surface with a decomposition @xmath11 into quadrilaterals , and let @xmath1 be a negatively curved closed riemannian manifold .
let @xmath243 be a continuous family of nontrivial simplicial maps parametrized by a subspace @xmath239 of @xmath244 , where @xmath245 is a continuous family of simplicial quasi - metrics .
then the gradient flow applied to each @xmath246 yields a continuous family of simplicial maps @xmath247 parametrized by @xmath248 , such that , for each @xmath249 , the map @xmath250 equals @xmath246 , and the family of maps @xmath251 converges to a simplicial harmonic map @xmath252 .
the simplicial harmonic maps @xmath252 depend continuously on the initial map @xmath246 and on the initial simplicial metric @xmath245 . for fixed @xmath245 ,
the simplicial harmonic map homotopic to @xmath243 is unique , so does nt vary when @xmath246 is changed by a homotopy .
the path of maps @xmath247 defined by the gradient flow , for each fixed @xmath249 , varies continuously with the map , since each point is following the trajectory of a solution to a first order differential equation whose initial conditions vary continuously .
it remains to show that the simplicial harmonic maps @xmath252 depend continuously on the choice of simplicial metric @xmath245 .
let @xmath253 be a sequence of points in @xmath239 converging to @xmath249 in @xmath239 , let @xmath254 denote the metric associated to @xmath255 , and let @xmath256 denote the metric associated to @xmath249 .
thus the sequence of simplicial quasi - metrics @xmath254 converges to @xmath256 .
arguing as in the proof of the first part of proposition [ uniqueharmonicmapsexist ] shows that the corresponding simplicial harmonic maps @xmath257 converge to a limiting map @xmath258 whose simplicial energy is equal to the limiting value of the simplicial energies of @xmath257 . with respect to the simplicial metric @xmath256 ,
the map @xmath258 has the same simplicial energy as @xmath252 , which is the unique harmonic map in its homotopy class with simplicial metric @xmath256
. therefore @xmath258 must be equal to @xmath252 .
this establishes that the resulting simplicial harmonic maps vary continuously with the simplicial metric , as claimed .
now we can apply the above result to show that we can homotope a family of maps from @xmath0 to @xmath1 to arrange that there is a uniform bound on the riemannian area of the maps in the new family , where the bound depends only on the euler characteristic of @xmath0 , and the curvature of @xmath1 .
[ monotone ] let @xmath0 be a closed surface with euler characteristic @xmath8 , and with a subdivision @xmath11 into quadrilaterals .
let @xmath238 , @xmath259 , be a continuous family of simplicial maps of @xmath0 to a riemannian manifold @xmath1 with sectional curvatures bounded above by @xmath213 , where @xmath260 is a path connected subset of @xmath2 .
further suppose that some , and hence every , @xmath246 is nontrivial .
then , for each @xmath259 , there is a simplicial quasi - metric @xmath261 on @xmath0 which varies continuously with @xmath249 , and a continuous family of maps @xmath262 parametrized by @xmath263 , satisfying 1 .
@xmath264 , 2 .
@xmath265 is a simplicial harmonic map , 3 .
the simplicial energy of @xmath266 is non - increasing with @xmath166 , 4 .
@xmath267 , for every @xmath249 . for each simplicial map @xmath246 in the given family
, we pick a simplicial quasi - metric @xmath261 on @xmath11 by setting the edge length @xmath268 of @xmath13 equal to the length @xmath269 of @xmath270 . for this quasi - metric ,
all the stretch factors @xmath52 are equal to @xmath90 on all edges with @xmath271 .
if @xmath272 , then @xmath246 must map @xmath13 to a point .
thus @xmath246 has finite simplicial energy .
now we apply proposition [ gradientflow ] .
this yields a continuous family of simplicial maps @xmath273 parametrized by @xmath248 , such that @xmath274 .
further , for each @xmath249 , the function @xmath275 is a non - increasing function of @xmath166 , and the family of maps @xmath251 converges to a simplicial harmonic map @xmath252 .
now we use a homeomorphism between @xmath276 and @xmath277 to obtain a new family of simplicial maps @xmath262 parametrized by @xmath263 , such that @xmath278 , and @xmath279 is simplicial harmonic .
this family is continuous by proposition [ gradientflow ] .
in particular , parts 1)-3 ) of the theorem hold .
since @xmath265 is simplicial harmonic , the proof of lemma [ arealessthan2pichi ] shows that @xmath280 for every @xmath249 , so that part 4 ) of the theorem holds .
the above theorem and its proof are very similar to what occurs in the smooth setting .
the crucial difference is that in the smooth setting , if @xmath38 is a map to a riemannian manifold @xmath1 , one can only define the riemannian metric on @xmath0 induced by @xmath4 when @xmath4 is an immersion .
in particular the energy of @xmath4 must be non - zero .
but in the above theorem maps with zero energy cause no problems , as we simply induce a quasi - metric on @xmath0 .
this greatly simplifies applications , as we will see later in this section and in the next section .
if we start with a family @xmath246 of smooth or piecewise - smooth maps from @xmath0 to @xmath1 , we need to approximate it by a homotopic family of simplicial maps before we can apply the above result .
this is easy to do by just choosing a subdivision @xmath11 of @xmath0 into quadrilaterals , and then replacing each @xmath246 by the unique simplicial map which agrees with @xmath246 on the vertices of @xmath11 .
theorem [ monotone ] generalizes from @xmath90parameter to @xmath281parameter families certain results of thurston @xcite , bonahon @xcite , canary @xcite , minsky @xcite and wolf @xcite , that utilized pleated surfaces and simplicial hyperbolic surfaces to control the area and diameter of a @xmath90parameter family of surfaces .
see also the recent use of bounded area @xmath90parameter families of surfaces by agol @xcite and calegari - gabai @xcite in proving the tameness conjecture for ends of negatively curved @xmath6manifolds .
area bounds for @xmath3parameter families of surfaces obtained using smooth harmonic maps were used to find counterexamples to the stabilization conjecture for heegaard splittings in @xcite .
see the discussion at the end of this section .
pitts and rubinstein @xcite gave a construction of unstable minimal surfaces that utilizes a minimax argument applied to a family of surfaces called a sweepout .
for later applications we give below a modified definition of a sweepout .
given a closed orientable surface @xmath0 , a closed orientable @xmath6manifold @xmath1 , and a map @xmath282 with the property that @xmath283 , there is a natural homomorphism @xmath284 we now explain the construction of @xmath285 .
since @xmath286 is a cyclic group generated by the fundamental class @xmath169 of @xmath287 , it suffices to define the action of @xmath285 on @xmath169 .
let @xmath172 denote a fundamental @xmath6chain for @xmath288 , so that @xmath289 , where @xmath290 denotes a fundamental @xmath3cycle for @xmath291 , @xmath292
. thus each @xmath293 is a singular @xmath3cycle in @xmath1 with zero volume .
a theorem of federer implies that , for @xmath292 , there is a singular @xmath6chain @xmath294 in @xmath1 such that @xmath295 , and @xmath294 also has zero volume @xcite .
thus the @xmath6chain @xmath296 is a cycle .
we want to define the homology class represented by this cycle to be @xmath297 .
we need to know that this is independent of the choices of the @xmath294 s .
suppose that @xmath298 and @xmath299 are zero volume singular @xmath6chains in @xmath1 such that @xmath300 .
then @xmath301 is a zero volume @xmath6cycle in @xmath1 , and a theorem of federer implies this cycle must be null homologous .
it follows that our definition of @xmath302 is independent of the choice of @xmath6chain @xmath298 .
similarly it is also independent of the choice of @xmath6chain @xmath303 .
it follows that @xmath302 is well defined .
let @xmath1 be a closed orientable @xmath6manifold , and let @xmath0 be a closed orientable surface .
a _ sweepout _ of @xmath1 by @xmath0 is a map @xmath304 with @xmath305 , and inducing an isomorphism @xmath284 one can equally well define a sweepout using homology with @xmath306 coefficients .
there is a natural way to associate a sweepout to a heegaard splitting of a closed orientable @xmath6manifold @xmath1 . in a handlebody @xmath307
, we can choose a homotopy which shrinks the boundary surface @xmath0 onto a @xmath90dimensional spine @xmath308 of the handlebody , so that the homotopy is through embeddings except at the end of the homotopy .
doing this for each of the two handlebodies of the heegaard splitting of @xmath1 yields a sweepout in which the surface @xmath309 is the heegaard surface and , as @xmath310 , the homotopy crushes @xmath0 down to the chosen spine of one handlebody , and , as @xmath311 , the homotopy crushes @xmath0 down to the chosen spine of the other handlebody .
pitts and rubinstein @xcite considered a closed @xmath6manifold with a strongly irreducible heegaard splitting .
they then constructed a minimal surface that has maximal area in some sweepout derived from the splitting . in a hyperbolic manifold @xmath1
, a minimal surface of genus @xmath312 has area less than @xmath313 , and this implies the same area bound for each surface in the sweepout .
the existence of a sweepout by bounded area surfaces has implications for the geometry of @xmath1 .
it gives upper bounds on the injectivity radius , as noted in rubinstein @xcite and in bachman - cooper - white @xcite .
we will show that one can use harmonic maps to obtain the same area bound given by pitts and rubinstein , but without proving the existence of an unstable minimal surface and without needing to assume that the heegaard splitting is strongly irreducible .
the minimal surface obtained by pitts and rubinstein is embedded , whereas harmonic maps need not yield embedded surfaces , but embeddedness is not needed to establish results such as a bound on injectivity radius .
the natural way to apply the theory of harmonic maps to a sweepout @xmath304 is to apply theorem [ monotone ] or its smooth analogue in order to obtain a sweepout by bounded area surfaces . in the smooth case
the difficulty is that at the ends of the sweepout , i.e. at @xmath314 and @xmath315 , the maps from @xmath0 to @xmath1 have @xmath90dimensional image and so there is no way to define the induced riemannian metric on @xmath0 and one can not directly use the gradient flow . in @xcite , the authors showed how to work round this difficulty . in the simplicial setting , this is not a problem , but a new difficulty arises .
it is not automatic that the maps from @xmath0 to @xmath1 at the ends of the sweepout have zero energy .
if they do not , then the gradient flow applied to such maps may yield maps of non - zero riemannian area , so that we no longer have a sweepout .
we resolve this difficulty in the following way .
first we consider maps from surfaces to graphs in general .
let @xmath316 denote projection onto the second factor .
let @xmath0 be a closed surface , and let @xmath308 be a graph .
a map @xmath317 is _ of standard type _ if there is a compact subsurface @xmath318 of @xmath0 , such that the following conditions hold : 1 .
each component of @xmath318 is an annulus .
each component of the closure of @xmath319 is sent by @xmath4 to a vertex of @xmath308 .
each component annulus @xmath26 of @xmath318 is sent by @xmath4 to an edge of @xmath308 , and @xmath320 is the composite of @xmath321 with the attaching map of the edge .
any map from @xmath0 to a graph @xmath308 is homotopic to a map of standard type . for future reference , we set out this result and its proof .
[ standardtypeforonemap]let @xmath0 be a closed surface , and let @xmath308 be a graph . then any map from @xmath0 to @xmath308 is homotopic to a map of standard type .
we denote the edges of @xmath308 by @xmath322 , @xmath323 .
the first step is to choose a point @xmath324 in the interior of each edge @xmath322 of @xmath308 , and to homotope the map @xmath317 to be transverse to each @xmath324 .
let @xmath325 denote @xmath326 , so that @xmath325 is a finite collection of disjoint simple closed curves in @xmath0 .
further there is a closed interval @xmath327 in the interior of @xmath322 , and a homeomorphism of @xmath328 with @xmath329 , such that the restriction of @xmath4 to each @xmath330 component of @xmath328 is the composite of @xmath321 , projection of @xmath330 onto the second factor , with a homeomorphism onto @xmath327 .
now consider a map @xmath331 which , for each edge @xmath322 of @xmath308 , maps @xmath327 to @xmath322 by a homeomorphism on their interiors and maps each component of @xmath332 to the vertex of @xmath308 in that component .
the composite @xmath333 is of standard type . as @xmath334 is clearly homotopic to the identity
, it follows that there is a homotopy of @xmath4 to a map of standard type , as required .
next we consider constructing subdivisions into quadrilaterals .
pick a subdivision of the circle @xmath335 into intervals . taking the product with @xmath336 yields a subdivision of the annulus @xmath330 into quadrilaterals , which we call a standard subdivision .
now given a map @xmath317 of standard type , we choose a subdivision @xmath11 of @xmath0 into quadrilaterals by starting with a standard subdivision of each component of @xmath318 , and then extending in any way to the rest of @xmath0 .
we will say that such a subdivision is _ compatible _ with @xmath4 .
next let @xmath12 be the simplicial quasi - metric on @xmath70 induced by @xmath4 , and suppose that @xmath308 has a path metric .
thus we can consider the simplicial energy @xmath188 of @xmath337 .
let @xmath100 be a quadrilateral of @xmath11 .
if @xmath100 is not contained in @xmath318 , it is mapped by @xmath4 to a point and so contributes zero to @xmath188 . if @xmath100 is contained in @xmath318 , two of its edges are mapped to vertices of @xmath308 and the other two edges are mapped onto edges of @xmath308 .
further the two edges mapped to vertices are non - adjacent edges of @xmath100 .
it follows that in this case also @xmath100 contributes zero to @xmath188 .
we conclude that @xmath338 . before returning to our discussion of sweepouts ,
we need one further refinement of lemma [ standardtypeforonemap ] . we consider a closed surface @xmath0 , a finite collection of graphs @xmath339 , and maps @xmath340 .
lemma [ standardtypeforonemap ] tells us that , for each @xmath155 , we can homotope @xmath341 to be of standard type and can then find a compatible subdivision @xmath342 of @xmath0 into quadrilaterals .
the following lemma shows how to find a single subdivision of @xmath0 into quadrilaterals which is compatible with each @xmath341 .
[ standardtypeforseveralmaps]let @xmath0 be a closed surface , and , for @xmath343 , let @xmath344 be a graph , and let @xmath340 be a continuous map .
then there is @xmath345 homotopic to @xmath341 , and a single subdivision of @xmath0 into quadrilaterals which is compatible with each @xmath345 . by applying lemma [ standardtypeforonemap ]
, we can assume that each @xmath341 is of standard type .
choose one point in the interior of each edge of @xmath344 , and let @xmath346 denote the union of all these points .
thus @xmath347 is a finite collection of disjoint simple closed curves on @xmath0 .
we can slightly homotope each @xmath341 to a new map @xmath345 , also of standard type , to arrange that the @xmath348 s intersect each other transversely and that there are no triple points .
we let @xmath349 . for each @xmath155
, we choose a regular neighborhood @xmath350 of @xmath351 so that the following conditions hold : 1 .
the intersection of three distinct @xmath350 s is empty .
2 . for distinct @xmath155 and @xmath352
, the boundaries of @xmath350 and @xmath353 meet transversely .
3 . for distinct @xmath155 and @xmath352 ,
each component of @xmath354 is a disc which contains exactly one point of @xmath355 , exactly one sub - arc of @xmath351 and exactly one sub - arc of @xmath356 .
the picture must be as shown in figure [ strips ] . and
@xmath353 , width=192 ] now we can construct a single subdivision @xmath11 of @xmath0 into quadrilaterals which is simultaneously compatible with each @xmath341 . for distinct @xmath155 and @xmath352 ,
consider a component of @xmath354 .
as in figure [ strips ] , there is a natural way to regard this as a quadrilateral , with two edges in @xmath357 and two in @xmath358 .
we start by choosing all these quadrilaterals .
next we consider a component of @xmath350 .
this is an annulus , and some pieces of this annulus have already been declared to be quadrilaterals .
it is now trivial to subdivide the rest so as to obtain a standard subdivision of the annulus .
having made these choices for each @xmath155 , we will have a decomposition of the union of all the @xmath350 s .
we now complete this to a subdivision of @xmath0 in any way .
it is easy to see that , after a further homotopy , @xmath11 is compatible with each @xmath345 , as required .
next we apply theorem [ monotone ] to sweepouts .
[ boundedarea ] if @xmath1 is a closed orientable hyperbolic @xmath6manifold with a genus @xmath312 heegaard splitting , then @xmath1 has a sweepout by genus @xmath312 surfaces with riemannian area bounded above by @xmath313 .
let @xmath0 denote the heegaard surface of @xmath1 , and let @xmath359 and @xmath360 denote the two handlebodies into which @xmath0 cuts @xmath1 . for @xmath292 , pick a @xmath90dimensional spine @xmath344 for @xmath361 . by subdividing the edges of these graphs
, we can arrange that each edge is a geodesic segment in @xmath1 .
then there is a sweepout @xmath362 of @xmath1 by @xmath0 such that @xmath363 collapses @xmath0 to @xmath364 , and @xmath365 collapses @xmath0 to @xmath366 . for each @xmath367 or @xmath90 , there is a map @xmath340 , which is of standard type , and is homotopic to @xmath345 . we homotope the given sweepout @xmath312 so that @xmath368 and @xmath369 .
the above discussion yields a subdivision @xmath11 of @xmath0 into quadrilaterals which is compatible with both @xmath363 and @xmath365 , perhaps after further small homotopies of @xmath363 and @xmath365 . as @xmath363 and @xmath365
are now of standard type , and the edges of @xmath364 and @xmath366 are geodesic , it follows that @xmath363 and @xmath365 are simplicial .
now we simultaneously homotope each @xmath370 to the simplicial map which agrees with @xmath370 on the vertices .
this will not alter @xmath363 and @xmath365 , so these maps still have zero simplicial energy with respect to the induced simplicial quasi - metric .
finally we can apply the proof of theorem [ monotone ] to this sweepout by simplicial maps to obtain a @xmath90parameter family of simplicial harmonic maps .
the crucial fact is that as the energies of @xmath363 and @xmath365 are each zero , the gradient flow leaves them unchanged .
thus we have homotoped the original sweepout , through sweepouts , to a sweepout by simplicial harmonic maps , each of which has riemannian area less than @xmath313 , by lemma [ e > standarda ] .
the scharlemann - thompson genus of @xmath1 can be used in place of the heegaard genus @xmath312 in corollary [ boundedarea ] , and may be smaller @xcite .
we end this section by discussing how to replace the use of smooth harmonic maps by simplicial harmonic maps in the work of hass , thompson and thurston @xcite . in that paper , the authors considered a particular closed hyperbolic @xmath6manifold @xmath1 with a heegaard splitting @xmath371 and heegaard surface @xmath0 .
they showed that there could not be an ambient isotopy of @xmath1 which interchanged @xmath359 and @xmath360 , and a key role in their argument was played by applying the gradient flow to a @xmath3parameter family of smooth maps from @xmath0 to @xmath1 .
we will describe how to replace this part of their argument by the work in this paper . as in the proof of lemma [ boundedarea ] , for @xmath292
, we pick a @xmath90dimensional spine @xmath344 for @xmath361 , so that each edge of @xmath344 is a geodesic segment in @xmath1 . as in that lemma , we can find a sweepout @xmath372 by simplicial harmonic maps such that , for @xmath292 , the map @xmath373 sends @xmath0 to @xmath344 and is of standard type . in particular @xmath374 and @xmath375 each has zero simplicial energy .
next suppose that @xmath376 is an ambient isotopy of @xmath1 which interchanges @xmath359 and @xmath360 .
thus @xmath377 is a homeomorphism , @xmath378 is the identity map of @xmath1 , and @xmath379 and @xmath380 .
let @xmath381 denote projection .
then the composite map @xmath382 is a @xmath3parameter family of maps @xmath383 from @xmath0 to @xmath1 , with @xmath384 , for all @xmath249 . as @xmath374 collapses @xmath0 to the spine @xmath364 of @xmath359 , the map @xmath385 collapses @xmath0 to the graph @xmath386 embedded in @xmath1 .
further each map @xmath385 is of standard type .
similarly each map @xmath387 collapses @xmath0 to the graph @xmath388 embedded in @xmath1 , and is of standard type .
note that as @xmath389 interchanges @xmath359 and @xmath360 , we have @xmath390 is a spine of @xmath360 and @xmath391 is a spine of @xmath359 .
the next step is to replace each map @xmath383 by the unique simplicial map which agrees with @xmath383 on the vertices of @xmath11 .
as each @xmath392 is already simplicial , the maps @xmath384 do not change .
the maps @xmath385 and @xmath387 may change , but they will continue to have zero simplicial energy .
this is because any edge of @xmath11 which was mapped to a point before the change will continue to be mapped to a point after the change .
thus the gradient flow will leave all the maps @xmath385 and @xmath387 unchanged .
hence when we apply theorem [ monotone ] to this @xmath3parameter family of maps , we will obtain a family @xmath251 of harmonic maps , each with riemannian area less than @xmath393 , such that all the maps @xmath394 and @xmath395 have zero simplicial energy and hence zero riemannian area . in @xcite ,
the authors considered hyperbolic @xmath6manifolds of fixed heegaard genus which have arbitrarily high volume .
this implies that a sweepout @xmath372 of @xmath1 by @xmath0 in which each surface has riemannian area less than @xmath396 , must be `` long '' in a certain sense .
they show that this implies that in any @xmath3parameter family that flips the orientation of the heegaard splitting , there must be a surface of large area .
this in turn was used to show that it is not possible to connect such a heegaard sweepout to the same sweepout with opposite orientation without stabilizing until the genus is doubled .
in this section we define the genus of an @xmath10dimensional closed orientable manifold , extending the idea of the heegaard genus of a @xmath6manifold to closed manifolds of arbitrary dimension .
to do so we extend the definition of a sweepout of a @xmath6manifold by a @xmath90parameter family of surfaces to a sweepout of a closed @xmath10dimensional manifold @xmath1 by an @xmath397parameter family of surfaces .
we replace the unit interval @xmath336 in a heegaard sweepout with an arbitrary compact @xmath397dimensional manifold which parametrizes the surface maps . for this construction
we let @xmath398 be the orientable total space of a bundle over a compact @xmath397manifold @xmath159 with fiber a closed orientable surface @xmath0 , with projection map @xmath399 , and with @xmath181 possibly empty .
we will consider an appropriately defined degree one map @xmath400 such that , for all @xmath401,@xmath402 as before , we need to define a homomorphism @xmath403 to make sense of the notion of degree for such maps . again
note that we could also use homology with @xmath306 coefficients .
since @xmath404 is a cyclic group generated by the relative fundamental class @xmath169 , a homomorphism @xmath285 to @xmath405 is determined by the action of @xmath285 on @xmath169 .
so long as @xmath406 satisfies some mild smoothness condition , the @xmath407volume of @xmath408 will equal zero because @xmath409 for @xmath401 .
so @xmath408 gives a zero volume @xmath407chain in @xmath1 which in turn is the boundary of a zero volume @xmath10chain @xmath410 in @xmath1 .
thus @xmath411 is a @xmath10cycle in @xmath1 whose homology class @xmath412 is independent of the choice of @xmath410 .
we set @xmath413 .
this leads to the following definition .
a _ sweepout _ of a closed orientable @xmath10manifold @xmath1 by a closed orientable surface @xmath0 is a map @xmath400 from a bundle @xmath398 with orientable total space , fiber @xmath0 , and base space a compact @xmath397manifold @xmath159 , whose boundary may be empty , such that 1 .
@xmath409 , for all @xmath401 , and the @xmath407volume of @xmath408 also equals zero , and 2 .
@xmath406 induces an isomorphism@xmath414 if @xmath1 is a closed orientable @xmath10manifold which admits a sweepout by genus @xmath312 surfaces , for some @xmath312 , then the _ genus _ of @xmath1 is the smallest such number @xmath312 . for @xmath6manifolds
this number is clearly less than or equal to the heegaard genus , since the above definition allows non - embedded and non - disjoint surfaces to sweep over the @xmath6manifold .
in addition , it allows the consideration of sweepouts by bundles over a circle , as well as over an interval . we do not know whether or not equality always holds for a given @xmath159 . here are some examples of @xmath10manifolds which admit a sweepout by surfaces .
our first examples are simply the product of a @xmath6manifold with another manifold .
let @xmath415 be a closed @xmath416dimensional manifold , let @xmath417 be a closed @xmath6manifold , and let @xmath418 .
choose a sweepout @xmath419 of @xmath417 by a closed surface @xmath0 , let @xmath420 , and let @xmath421 .
then the map @xmath422 is a sweepout of @xmath1 by @xmath0 .
our next examples will play an important role in our discussion below .
let @xmath159 be a compact @xmath397dimensional manifold , and let @xmath0 be a closed orientable surface of genus @xmath312 .
let @xmath398 be a bundle over @xmath159 with fiber @xmath0 whose restriction to @xmath181 is trivial .
choose a homeomorphism @xmath4 from @xmath0 to the boundary of a handlebody @xmath307 , and let @xmath312 denote the homeomorphism @xmath423 .
we form a closed @xmath10manifold @xmath1 by attaching @xmath398 to @xmath424 using the homeomorphism of their boundaries @xmath425 . as in the previous section
, there is a homotopy which shrinks the boundary surface @xmath0 of the handlebody @xmath307 onto a @xmath90dimensional spine @xmath308 of @xmath307 , such that the homotopy is through embeddings except at the end of the homotopy .
taking the product of this homotopy with the identity map of @xmath159 yields a map @xmath426 which , for each @xmath427 in @xmath181 , shrinks @xmath428 onto @xmath429 .
together with the identity map of @xmath398 , this yields a map from @xmath430 to @xmath1 which is clearly a sweepout of @xmath1 by @xmath0 .
of course @xmath431 is homeomorphic to @xmath398 , so we can and will regard this sweepout as a map from @xmath398 to @xmath1 .
this sweepout has the special property that for all interior points of @xmath159 , the associated map from @xmath0 to @xmath1 is an embedding , and that for all boundary points of @xmath159 , the associated map from @xmath0 to @xmath1 has image a copy of the graph @xmath308 .
if @xmath181 is not connected , one can make this construction slightly more general by separately choosing a homeomorphism from @xmath0 to the boundary of the handlebody @xmath307 for each component of @xmath181 .
when @xmath159 is the @xmath397-ball , the bundle must be trivial over @xmath181 .
@xmath10-dimensional heegaard splittings _ therefore fall into this class of examples .
now suppose that we have a sweepout @xmath400 of a closed orientable @xmath10manifold @xmath1 by a closed orientable surface @xmath0 .
in addition suppose that @xmath1 has a metric of negative curvature .
we would like to apply the ideas of the previous section to show that @xmath1 admits a sweepout by small area maps from @xmath0 to @xmath1 .
as in that section , the key to doing this is being able to find a sweepout such that , for each @xmath427 in @xmath181 , the associated map from @xmath0 to @xmath1 has zero energy .
the result we obtain is the following .
let @xmath1 be a closed riemannian @xmath10manifold with sectional curvatures bounded above by @xmath213 , let @xmath0 be a closed orientable surface of genus @xmath432 , and let @xmath11 be a subdivision of @xmath0 into quadrilaterals .
suppose that @xmath400 is a sweepout of @xmath1 by @xmath0 such that , for each @xmath427 in @xmath181 , the associated map @xmath433 is simplicial and has zero simplicial energy with respect to the induced quasi - metric on @xmath70 .
then @xmath406 can be homotoped , fixing @xmath434 , to a new sweepout so that , for each @xmath427 in @xmath159 , the associated map from @xmath0 to @xmath1 has riemannian area at most @xmath435 .
we will use the same ideas as in the proof of lemma [ boundedarea ] in the case of dimension @xmath6 .
if @xmath398 is a trivial bundle over @xmath159 , we can simply apply theorem [ monotone ] to homotope the sweepout to a family of simplicial harmonic maps from @xmath0 to @xmath1 .
this homotopy will be fixed for fibers over @xmath181 , so still gives a degree one map from @xmath398 to @xmath1 , and is therefore still a sweepout .
theorem [ monotone ] also tells us that in this new sweepout the area of each surface is less than @xmath435 , as required .
if @xmath398 is a nontrivial bundle over @xmath159 , it is still locally trivial , and that suffices to apply theorem [ monotone ] . given a sweepout , it is not obvious that one can find a subdivision into quadrilaterals that satisfies the hypotheses of the above theorem .
but it is easy to do this in the cases discussed just before the theorem . as in the second set of examples , we let @xmath159 be a compact @xmath397dimensional manifold ,
let @xmath0 be a closed orientable surface of genus @xmath312 , and let @xmath398 be a bundle over @xmath159 with fiber @xmath0 whose restriction to @xmath181 is trivial .
finally let @xmath1 be a closed @xmath10manifold formed by attaching @xmath398 to @xmath436 using a product homeomorphism of their boundaries .
then there is a sweepout @xmath400 such that , for all interior points of @xmath159 , the associated map from @xmath0 to @xmath1 is an embedding , and , for all boundary points of @xmath159 , the associated map from @xmath0 to @xmath1 has image a graph which depends only on the component of @xmath181 .
further , for a component @xmath437 of @xmath181 , the maps from @xmath0 to the graph @xmath308 associated to the points of @xmath437 are all equal .
as @xmath181 has only finitely many components , lemma [ standardtypeforseveralmaps ] tells us that after a suitable homotopy of @xmath406 , there is a single subdivision @xmath11 of @xmath0 into quadrilaterals which is compatible with all of these maps .
now it is immediate that , for each @xmath427 in @xmath181 , the associated map from @xmath0 to @xmath1 has zero simplicial energy with respect to the induced quasi - metric on @xmath70 .
if we assume that @xmath1 has negative curvature , then we can homotope this sweepout to the sweepout by simplicial maps which agrees with the original sweepout on the vertices of @xmath11 .
thus , for each @xmath427 in @xmath181 , the associated simplicial map from @xmath0 to @xmath1 has zero simplicial energy with respect to the induced quasi - metric on @xmath70 , verifying all the hypotheses of the above theorem .
we can generalize the previous discussion as follows .
let @xmath1 be as in the previous paragraph , but do not assume that @xmath1 has negative curvature .
suppose that there a map @xmath281 of degree @xmath90 from @xmath1 to some negatively curved @xmath10manifold @xmath438 .
then the composite map @xmath439 is a sweepout such that , for each @xmath427 in @xmath181 , the associated map from @xmath20 to @xmath438 has zero simplicial energy .
further , we can homotope this sweepout to the sweepout by simplicial maps which agrees with the original sweepout on the vertices of @xmath11 , again verifying all the hypotheses of the above theorem .
in this section we examine global minimizers of simplicial energy over the space of all simplicial metrics for a fixed triangulation .
it was known to courant and rado that one way to find a least area map in a homotopy class is to find a map that minimizes energy not just for a given metric on the domain but among all possible metrics @xcite .
we now present this argument in the simplicial setting , where it becomes considerably easier .
let @xmath38 be a map of a triangulated closed surface @xmath0 to a closed manifold @xmath1 with a riemannian metric .
let @xmath440 denote the simplicial energy of @xmath4 with respect to a simplicial quasi - metric @xmath12 on @xmath0 . 1 .
if @xmath441 has least possible value among all maps homotopic to @xmath4 , then there is a simplicial quasi - metric @xmath12 on @xmath0 such that @xmath440 has the least possible value among all maps homotopic to @xmath4 and over all simplicial quasi - metrics on @xmath0 .
further @xmath442 .
2 . if @xmath12 is a simplicial quasi - metric on @xmath0 , and @xmath4 is a map such that @xmath440 has least possible value among all maps homotopic to @xmath4 and over all simplicial quasi - metrics on @xmath0 , then @xmath441 also has the least possible value , and @xmath442 .
\(1 ) let @xmath12 denote the simplicial quasi - metric on @xmath70 induced by @xmath4 , i.e. for each edge @xmath13 of @xmath11 , we set @xmath443 . for this quasi - metric ,
all the stretch factors @xmath52 are equal to @xmath90 on all edges with @xmath50 and therefore @xmath442 , by lemma [ e > a ] .
the hypothesis that @xmath441 has least possible value implies that , for any map @xmath444 homotopic to @xmath4 , we have @xmath445 .
thus for any simplicial quasi - metric @xmath281 on @xmath0 , we have the inequalities @xmath446 it follows that @xmath440 has the least possible value over all maps homotopic to @xmath4 and over all simplicial quasi - metrics on @xmath0 , and that @xmath442 , as claimed .
\(2 ) suppose that @xmath441 does not have the least possible value .
thus there is a map @xmath444 homotopic to @xmath4 with @xmath447 .
we define a simplicial quasi - metric @xmath281 on @xmath0 by setting @xmath448 equal to @xmath449 . for this quasi - metric and the map @xmath444 , the stretch factors @xmath52 are equal to @xmath90 on all edges with @xmath50 .
thus , by lemma [ e > a ] , we have @xmath450 which contradicts the hypothesis that @xmath440 has least possible value .
we conclude that @xmath441 must have the least possible value .
now part ( 1 ) implies that there is a simplicial quasi - metric @xmath281 on @xmath0 such that @xmath451 has the least possible value among all maps homotopic to @xmath4 and over all simplicial quasi - metrics on @xmath0 .
further @xmath452 .
as @xmath440 also has least possible value , it follows that @xmath440 must equal @xmath451 , so that @xmath442 , as required . the proof of the above lemma is much simpler than in the smooth setting .
this is because any map of @xmath0 into @xmath1 for which one can measure the lengths of the edges of @xmath11 induces a simplicial quasi - metric on @xmath0 , whereas in the smooth setting , one needs @xmath4 to be an immersion in order to induce a riemannian metric on @xmath0 .
in section [ simplicialharmonic ] we saw that a finite energy map from a surface with a fixed simplicial metric or quasi - metric to a manifold of non - positive curvature can be homotoped to a simplicial harmonic map that minimizes simplicial energy .
we now examine what happens when the simplicial metric on the domain surface is allowed to change .
we will show that the class of maps homotopic to a given map @xmath38 contains a map with minimal simplicial energy among all maps homotopic to @xmath4 and all possible simplicial quasi - metrics @xmath20 . in the smooth setting
this gives a least area map in the homotopy class of @xmath4 .
the existence of a least area map in that setting was shown by sacks and uhlenbeck @xcite and by schoen and yau @xcite , with the hypotheses that the surface is orientable with genus at least @xmath90 , and its fundamental group injects into the fundamental group of @xmath1 .
it would seem more natural to consider only simplicial metrics rather than allow quasi - metrics , but the following discussion explains why this is unreasonable . in fact this is why we introduced the idea of a simplicial quasi - metric .
consider a map @xmath453 from a triangulated surface @xmath70 to a riemannian manifold and let @xmath454 where the infimum is taken over all maps @xmath4 homotopic to @xmath124 and over all simplicial metrics on @xmath11 .
we would like to establish the existence of a simplicial harmonic map homotopic to @xmath124 with simplicial energy equal to @xmath455 .
there is a sequence @xmath456 of simplicial metrics on @xmath0 , and of homotopic simplicial maps @xmath457 such that @xmath458 , as @xmath459 .
we call such a sequence a _ simplicial - energy minimizing _ sequence .
the main difficulty we face is that even if the sequence @xmath456 converges to a function @xmath460 on the edges of @xmath11 , it may be that @xmath461 , for some edge @xmath13 of @xmath11 .
it is clear that @xmath462 , for all @xmath155 , and that @xmath460 satisfies the triangle inequality for each triangle of @xmath11 .
thus @xmath460 is a simplicial quasi - metric on @xmath0 .
there is a further difficulty with this approach which is that it is possible that @xmath461 , for every edge @xmath13 of @xmath11 .
for example , suppose that @xmath12 is a simplicial metric on @xmath0 , and @xmath38 is a simplicial map such that @xmath463 .
for each @xmath464 , we define the simplicial metric @xmath456 by @xmath465 , for each edge @xmath13 of @xmath11 , and the simplicial map @xmath257 to equal @xmath4 .
the scale invariance of simplicial energy implies that @xmath466 , for each @xmath464 , but the sequence @xmath456 converges to the simplicial quasi - metric @xmath460 such that @xmath461 , for every edge @xmath13 of @xmath11 .
the analogous phenomenon occurs in the smooth setting , and as in that setting we avoid this problem by changing our initial choice of minimizing sequence .
the following lemma shows that allowing quasi - metrics does not change the minimal energy for a homotopy class .
let @xmath453 be a map of a triangulated closed surface @xmath70 to a closed riemannian manifold @xmath1 .
let @xmath455 denote the infimum of simplicial energies over the family @xmath467 of all maps @xmath4 homotopic to @xmath124 and over all simplicial metrics on @xmath11 , and let @xmath468 denote the infimum of simplicial energies over the family @xmath469 of all finite energy maps @xmath4 homotopic to @xmath124 and over all simplicial quasi - metrics on @xmath11
. then @xmath470 . as @xmath471
, it is immediate that @xmath472
. now let @xmath460 be a simplicial quasi - metric on @xmath0 , and let @xmath473 be a finite energy map homotopic to @xmath124 .
consider the family of edge length functions @xmath474 , defined for @xmath475 by @xmath476 . as @xmath474 is positive and satisfies all triangle inequalities in @xmath11 , it follows that @xmath474 is a simplicial metric for each @xmath475 .
let @xmath477 denote the energy of the map @xmath38 computed using the simplicial quasi - metric @xmath474 on @xmath11 .
for @xmath475 we have @xmath478 if either of @xmath479 or @xmath480 is zero then so is @xmath39 or @xmath131 , so that @xmath481 it follows immediately that @xmath482 , so that @xmath470 as required
. next we show the existence of a simplicial quasi - metric and a map which minimizes energy in a homotopy class .
note that , unlike the smooth case , this theorem places no requirements on the given map @xmath124 or on its homotopy class . in the smooth case , an incompressibility condition is required .
[ absolutelyleastenergymapexists]let @xmath453 be a map of a triangulated closed surface @xmath70 to a closed non - positively curved manifold @xmath1 .
then there is a simplicial quasi - metric @xmath460 on @xmath0 , and a simplicial map @xmath125 homotopic to @xmath124 , such that @xmath125 minimizes simplicial energy among all maps homotopic to @xmath124 and all choices of simplicial quasi - metric on @xmath70 .
further @xmath125 minimizes simplicial area among all maps homotopic to @xmath124 .
let @xmath455 denote the infimum of simplicial energies over the family of all finite energy maps @xmath4 homotopic to @xmath124 and over all simplicial quasi - metrics on @xmath11 .
consider a sequence @xmath456 of simplicial quasi - metrics on @xmath70 and finite - energy simplicial maps @xmath457 , each homotopic to @xmath4 , such that @xmath483 .
for each @xmath10 , let @xmath484 denote the length of @xmath485 and define a new simplicial quasi - metric @xmath486 on @xmath70 by @xmath487 .
for this quasi - metric , the map @xmath257 is a simplicial isometry , with @xmath488 and @xmath489 .
now @xmath490 is independent of the simplicial quasi - metric on @xmath0 and @xmath491 by lemma [ e > a ] .
hence @xmath492 , and replacing each simplicial quasi - metric @xmath456 by @xmath486 retains the property that @xmath483 , so that @xmath257 is still an energy minimizing sequence . in particular , the sequences @xmath493 and @xmath490 are bounded .
we now claim that , after passing to a subsequence , the sequence @xmath484 is bounded for each @xmath155 . first note that by passing to a subsequence we can arrange that each sequence
@xmath484 is either convergent or unbounded . now consider an edge @xmath494 and suppose that @xmath495 is unbounded . by passing to a further subsequence
we can assume that @xmath495 is increasing and unbounded .
as @xmath494 meets one triangle of @xmath11 on each of its two sides , there are four additional edges of @xmath11 which belong to a triangle containing @xmath494 .
we denote these edges by @xmath496 ( they may not all be distinct in @xmath0 ) .
the terms in @xmath490 which involve @xmath497 are @xmath498 , and this sequence of sums is bounded , as we noted above . as @xmath497 is increasing and unbounded , the sum @xmath499 must approach @xmath205 as @xmath459 .
since each @xmath500 , we deduce that @xmath501 as @xmath502 , for @xmath503 .
now the triangle inequality implies that @xmath497 must also converge to @xmath205 as @xmath459 , a contradiction .
this establishes that @xmath484 is bounded for each @xmath155 , as claimed .
it follows that there is a subsequence of @xmath257 for which each sequence @xmath484 is convergent . as @xmath1 is compact , we can find a further subsequence such that for each vertex @xmath143 of @xmath11 , the sequence @xmath504 converges . as @xmath487 , the sequence @xmath505 also converges and the limit is a simplicial quasi - metric @xmath506 on @xmath0 .
so the sequence of finite - energy simplicial maps @xmath257 converges to a finite - energy simplicial map @xmath125 with @xmath507 .
this completes the proof of the theorem .
we say that a map of a surface into @xmath1 is _ simply essential _ if it sends nontrivial simple loops on the surface to nontrivial elements of @xmath120 . for two - sided embedded surfaces in a @xmath6manifold
this is equivalent to being @xmath508injective , but for singular surfaces in a @xmath6manifold this equivalence is unknown , and is called the simple loop conjecture .
the result below shows that if we assume that the map @xmath453 in theorem [ absolutelyleastenergymapexists ] is simply essential , then we can refine the result to assert the existence of a simplicial metric ( not just quasi - metric ) on @xmath0 and an energy minimizing map .
[ simplyessential ] let @xmath122 be a simply essential simplicial map from a triangulated closed surface @xmath70 to a closed non - positively curved manifold @xmath1 , and suppose that @xmath12 is a simplicial quasi - metric on @xmath70 such that @xmath507 . then there is a new triangulation @xmath509 of @xmath0 , obtained by collapsing edges and triangles of @xmath11 , a new simplicial map @xmath510 homotopic to @xmath125 , and a simplicial metric on @xmath509 for which @xmath511 . as @xmath125
is simplicial , it maps any edge of @xmath11 to a point or to a geodesic arc , and it maps any triangle of @xmath11 to a point or to a geodesic arc or to a triangle . if an edge @xmath512 of @xmath11 is zero length , i.e. @xmath513 , the fact that @xmath125 has finite energy implies that it must map @xmath512 to a point .
if a triangle @xmath514 of @xmath11 has @xmath515 consisting entirely of zero length edges , the fact that @xmath125 is simplicial implies that it must map @xmath514 to a point . if @xmath515 has two zero length edges , the fact that @xmath12 satisfies the triangle inequality implies that all three edges of @xmath515 are zero length .
finally if @xmath515 has one zero length edge , the fact that @xmath12 satisfies the triangle inequality implies that the other two edges of @xmath515 have the same @xmath12length , and the fact that @xmath125 is simplicial implies that these edges have the same image under @xmath125 .
now we let @xmath19 denote the subcomplex of @xmath11 consisting of all zero length edges of @xmath11 together with all triangles of @xmath11 whose boundary consists of zero length edges .
thus @xmath125 maps each component of @xmath19 to a point .
as @xmath125 is simply essential , if @xmath318 denotes a regular neighborhood of a component of @xmath19 , all except one of the components of the closure of @xmath516 must be a disc , and the union of @xmath318 with these discs must itself be a disc .
suppose there is a disc component @xmath159 of the closure of @xmath516 , and let @xmath517denote the component of @xmath518 which contains @xmath159 . then @xmath519 is contained in @xmath19 , so is mapped to a point by @xmath125 . the facts that @xmath125 maps @xmath519 to a point , is simplicial , and has least energy implies that it must map all of @xmath520 to a point .
thus @xmath520 is contained in @xmath19 , which is a contradiction .
this shows that no component of the closure of @xmath516 can be a disc , so that @xmath318 itself must be a disc .
we form the quotient @xmath521 of @xmath0 by collapsing every simplex of @xmath19 to a point .
note that @xmath125 must factor through a map @xmath522 .
the preceding discussion implies that @xmath521 is homeomorphic to @xmath0 .
the image of @xmath11 in @xmath521 need not be a triangulation . for
if @xmath514 is a triangle of @xmath11 such that @xmath515 has one zero length edge , then the image of @xmath514 in @xmath521 is a @xmath3gon .
but the image of @xmath514 in @xmath1 is a geodesic arc , and so we can take a further quotient @xmath523 of @xmath521 , by collapsing each such @xmath3gon to an arc , and @xmath125 must factor through a map @xmath524
. again @xmath523 must be homeomorphic to @xmath0 , and now the image of @xmath11 in @xmath523 is naturally a triangulation @xmath509 of @xmath523 .
( note that even if @xmath11 is a simplicial triangulation of @xmath0 , the triangulation @xmath509 of @xmath523 may not be simplicial . )
distinct edges of @xmath11 can not map to the same edge in @xmath521 .
thus if distinct edges @xmath228 and @xmath229 of @xmath11 map to the same edge in @xmath523 , it must be because certain @xmath3gons in @xmath521 were collapsed , so that @xmath525 .
it follows that there is a well defined function @xmath526 on the edges of @xmath509 , given by @xmath527 , where @xmath528 is any edge of @xmath529 and @xmath512 is any edge of @xmath11 which maps to @xmath528 .
clearly @xmath526 is never zero and so is a simplicial metric on @xmath509 .
further , as @xmath125 factors through the simplicial map @xmath530 , it follows that @xmath511 .
finally , as @xmath523 is obtained from @xmath0 by collapsing disjoint contractible subsets , @xmath510 must be homotopic to @xmath125 .
a consequence of corollary [ simplyessential ] is that any potential counterexample for the simple loop conjecture in a hyperbolic @xmath6manifold can be chosen to be geometrically controlled ; it can be homotoped to be simplicial with area less than @xmath9 .
since the simple loop conjecture has been proven for seifert fiber spaces @xcite and for graph manifolds @xcite , the hyperbolic case is central to this problem .
note that in @xcite , cooper and manning gave an example of a representation of a surface group into @xmath531 which has nontrivial kernel , but no element of the kernel is represented by an essential simple closed curve .
however this is not a counterexample to the simple loop conjecture as their representation is not known to be discrete , nor to have image in the fundamental group of a hyperbolic @xmath6manifold .
in this section we consider a metrized triangulated closed surface @xmath532 , and a closed surface @xmath533 with euler characteristic @xmath8 , and with a riemannian metric of non - positive curvature .
we will consider maps from @xmath534 to @xmath533 which are homotopic to a given homeomorphism .
we show that in the homotopy class of such a map there is a simplicial harmonic map which is a homeomorphism .
if @xmath533 is negatively curved , this simplicial harmonic map is then unique in its homotopy class .
thus the simplicial harmonic map never contains unnecessary singularities or folds .
hence if @xmath11 is a fixed triangulation of @xmath534 , this map gives a canonical triangulation of @xmath533 that varies continuously as we vary the metric @xmath406 .
note that if @xmath534 and @xmath533 were orientable , with @xmath534 of genus @xmath6 , and @xmath533 of genus @xmath3 , and if @xmath4 was a map of degree @xmath90 from @xmath534 to @xmath533 , then @xmath4 would be homotopic to a simplicial harmonic map @xmath312 , by proposition [ uniqueharmonicmapsexist ] , but @xmath312 could not be a homeomorphism nor could it be a covering map .
thus in general simplicial harmonic maps between surfaces must have some type of singularity .
see sampson @xcite for an analysis of the possible singularities in the smooth case . in order to prove the results in this section , we need to slightly restrict the type of triangulations we allow .
a triangulation @xmath11 of a compact surface @xmath0 is _ good _ if no edge of @xmath11 forms a null homotopic loop , and no union of two edges of @xmath11 forms a null homotopic loop .
clearly if a triangulation of @xmath0 is not good and @xmath0 has a metric of non - positive curvature , it is impossible to have a homeomorphism of @xmath0 which sends every edge to a geodesic . for the rest of this section , we fix a metrized triangulated surface @xmath532 , a closed surface @xmath535 with euler characteristic @xmath8 , and with a non - positively curved riemannian metric @xmath406 , and a homeomorphism @xmath536 .
let @xmath537 denote the infimum of the simplicial energies of all homeomorphisms isotopic to @xmath4 , and let @xmath538 denote the @xmath90skeleton of @xmath11 .
note that we can isotope @xmath4 to be piecewise geodesic on @xmath538 , so that the simplicial energy is now defined .
in particular , the isotopy class of @xmath4 always contains homeomorphisms whose simplicial energy is defined .
we will consider an energy minimizing sequence of homeomorphisms @xmath539 from @xmath534 to @xmath533 each of which is isotopic to @xmath4 , and will show that there is a subsequence which yields a simplicial homeomorphism @xmath540 whose energy equals @xmath537 .
the special case where @xmath11 is a @xmath90vertex triangulation and @xmath406 is negatively curved turns out to be much simpler and will be treated separately after the following lemma .
note that this first lemma does not require that @xmath11 be good , nor that the metric on @xmath533 be non - positively curved .
[ edgesaremappedtogeodesics]suppose that @xmath534 and @xmath533 are closed surfaces with negative euler number .
let @xmath541 be a homeomorphism from @xmath534 , with simplicial metric @xmath12 , to @xmath533 with riemannian metric @xmath406 .
let @xmath539 be a sequence of homeomorphisms from @xmath534 to @xmath533 , each of which is isotopic to @xmath4 and whose simplicial energies approach @xmath537 .
then there is a subsequence of the @xmath257 s whose restrictions to @xmath538 converges to a map @xmath542 .
further any map @xmath125 obtained in this way maps each edge of @xmath11 to a point or to a geodesic arc .
as @xmath257 has finite simplicial energy , it must map each edge of @xmath11 to a rectifiable arc .
as the sequence of simplicial energies of the @xmath257 s is convergent , it is bounded .
as @xmath533 is compact , ascoli s theorem shows that the sequence of @xmath257 s has a subsequence whose restriction to @xmath538 converges to a continuous function @xmath125 .
if @xmath512 is an edge of @xmath11 , then the sequence of lengths of the arcs @xmath543 is also bounded , so that @xmath544 must have finite length .
thus @xmath125 maps each edge of @xmath11 to a rectifiable arc .
further for any extension of @xmath125 to all of @xmath534 , we have @xmath545 .
we claim that if @xmath546 is any point of @xmath547 which is not the image of a vertex of @xmath11 , then there must be a disc @xmath182 centered on @xmath546 such that @xmath547 meets @xmath182 in geodesic arcs .
assuming this claim , the fact that there can be only finitely many images of vertices of @xmath11 immediately implies that @xmath125 must map each edge of @xmath11 to a point or to a geodesic arc , as required . to prove our claim , suppose that @xmath546 is a point of @xmath547 which is not the image of a vertex of @xmath11 .
then suitably small neighborhoods of @xmath546 will also not contain the image of a vertex of @xmath11 . as @xmath125 maps
each edge of @xmath11 to a rectifiable arc , it follows that there is a small convex disc @xmath182 centered on @xmath546 such that @xmath548 consists of a finite number of points @xmath549 . as @xmath182 does not contain the image of a vertex , the intersection of the image of @xmath547 with @xmath182 consists of isolated points in @xmath185 together with arcs joining the images of the remaining @xmath71 s in pairs .
we replace these arcs by geodesic arcs joining the images of the @xmath71 s in pairs in the same way , to obtain a new map @xmath510 of @xmath538 into @xmath533 .
as @xmath125 is the limit of the homeomorphisms @xmath539 , we could make the analogous changes to @xmath257 , for all large enough values of @xmath10 .
the geodesic arcs obtained when we do this for @xmath257 will all be disjoint .
we conclude that there is a new sequence @xmath550 of homeomorphisms such that the limit of @xmath551 is @xmath510 . as no edge of @xmath11
is lengthened by this construction , we have @xmath552 , for each @xmath553 if there is any edge of @xmath11 whose image under @xmath125 meets @xmath182 in a non - geodesic arc , our construction of @xmath510 will strictly shorten the image of this edge , so that the limit of the sequence @xmath554 is strictly less than @xmath537 , which is a contradiction .
we deduce that @xmath547 must meet @xmath182 in geodesic arcs , which completes the proof of the claim . before proceeding further
, we consider the special case when @xmath11 is a @xmath90vertex triangulation of @xmath534 , and @xmath533 has a riemannian metric of negative curvature .
[ 1-vertextriangulation]let @xmath534 and @xmath533 be closed surfaces with negative euler number .
let @xmath11 be a good @xmath90vertex triangulation of @xmath534 , let @xmath406 be a negatively curved riemannian metric on @xmath533 , and let @xmath555 be a homeomorphism . then the following hold : 1 .
@xmath4 is isotopic to a simplicial homeomorphism .
the unique least energy simplicial map homotopic to @xmath4 is a simplicial homeomorphism and is isotopic to @xmath4 .
any homeomorphism from @xmath534 to @xmath533 which is homotopic to @xmath4 must be isotopic to @xmath4 . by taking @xmath534 equal to @xmath533 in part 3 ) , we see that we have a new proof of the classical fact that homotopic homeomorphisms of @xmath0 are isotopic , so long as @xmath79has negative euler number .
\1 ) lemma [ edgesaremappedtogeodesics ] tells us that there is a sequence @xmath539 of homeomorphisms from @xmath534 to @xmath533 each of which is isotopic to @xmath4 , and whose simplicial energies approach @xmath537 , such that the restrictions of the @xmath257 s to @xmath538 converges to a map @xmath125 which sends each edge of @xmath11 to a point or to a geodesic arc .
let @xmath143 denote the vertex of @xmath11 .
as @xmath11 is a good @xmath90vertex triangulation , @xmath125 can not send an edge of @xmath11 to a point , and it must send distinct edges of @xmath11 to distinct geodesics in @xmath533 . in particular , the restriction of @xmath125 to each edge of @xmath11 is an immersion .
if @xmath125 fails to embed @xmath538 in @xmath533 , there must be an edge @xmath512 of @xmath11 with an interior point @xmath175 such that @xmath556 equals @xmath557 .
see figure [ 1vertex ] . if @xmath558 is a suitably small neighborhood of @xmath557 in @xmath533 , the image of @xmath538 in @xmath558 consists of the image of a neighborhood @xmath133 of @xmath175 in @xmath512 , together with the image of a neighborhood @xmath559 of @xmath143 in @xmath538 . as the @xmath257 s are all homeomorphisms
, @xmath560 lies on one side of @xmath561 , as shown in figure [ 1vertex ] .
now consider a homotopy of @xmath125 which moves @xmath557 a small distance towards this side of @xmath561 , and let @xmath510 denote the map so obtained .
we further choose @xmath510 to be equal to @xmath125 outside @xmath559 , and so that @xmath562 consists of a union of geodesic arcs joining @xmath563 to @xmath564 .
this homotopy will shorten the image of every edge of @xmath11 , so that @xmath510 has strictly less simplicial energy than @xmath125 .
further @xmath510 can be extended to a homeomorphism which is isotopic to @xmath4 .
as the energy of @xmath510 is less than @xmath537 , this is a contradiction .
we conclude that @xmath125 must embed @xmath538 in @xmath533 .
as the @xmath257 s are all homeomorphisms , and @xmath125 is the limit of the sequence @xmath565 , there is a unique extension of @xmath125 to a simplicial homeomorphism @xmath312 from @xmath532 to @xmath535 .
the alexander trick applied to each triangle then implies that @xmath4 is isotopic to this simplicial homeomorphism , as required .
\2 ) by construction , @xmath312 is a homeomorphism which is isotopic to @xmath4 , and @xmath312 minimizes @xmath190 among all homeomorphisms isotopic to @xmath4 .
in particular , @xmath312 is a critical point for @xmath190 under all variations through homeomorphisms .
as any small deformation of @xmath312 must still be a homeomorphism , this implies that @xmath312 is a critical point for @xmath190 under all variations .
as @xmath406 is negatively curved , it follows that @xmath312 is the unique least energy simplicial map homotopic to @xmath4 , as required .
\3 ) let @xmath124 denote a homeomorphism from @xmath534 to @xmath533 which is homotopic to @xmath4 .
by parts 1 ) and 2 ) , @xmath124 is isotopic to an energy minimizing simplicial homeomorphism which must be @xmath312 . in particular @xmath124
must be isotopic to @xmath4 , as required .
now we return to the case of a general triangulation .
the result we obtain is similar to previous results obtained in the smooth and orientable setting .
sampson @xcite and schoen and yau @xcite showed that there is a least energy harmonic diffeomorphism in each homotopy class of diffeomorphisms from a riemannian surface to a surface of non - positive curvature .
later it was shown by jost and schoen @xcite that there is a least energy harmonic diffeomorphism in each homotopy class of diffeomorphisms between two riemannian surfaces of genus at least @xmath3 , even if the target surface has some positive curvature . in @xcite ,
coron and helein showed that , in all cases , any harmonic diffeomorphism between two riemannian surfaces of genus at least @xmath3 , must be energy minimizing in its homotopy class , and is the unique energy minimizing map in that homotopy class .
[ harmonicsurfacehomeoexists ] suppose that @xmath534 and @xmath533 are closed surfaces with negative euler number .
let @xmath541 be a homeomorphism from @xmath534 , with simplicial metric @xmath12 , to @xmath533 with a riemannian metric of non - positive curvature .
further assume that @xmath11 is a good triangulation .
then @xmath4 is isotopic to a simplicial harmonic homeomorphism @xmath566 that minimizes @xmath190 among all homeomorphisms isotopic to @xmath4 .
further , if @xmath535 is negatively curved , then @xmath312 is the unique simplicial harmonic map in the homotopy class of @xmath4 , and so minimizes energy in this homotopy class . recall that @xmath537 denotes the infimum of the simplicial energies of all homeomorphisms isotopic to @xmath4 , and that @xmath538 denotes the @xmath90skeleton of @xmath11 .
lemma [ edgesaremappedtogeodesics ] tells us that there is a sequence @xmath539 of homeomorphisms from @xmath534 to @xmath533 each of which is isotopic to @xmath4 , and whose simplicial energies approach @xmath537 , such that the restrictions of the @xmath257 s to @xmath538 converge to a map @xmath125 which sends each edge of @xmath11 to a point or to a geodesic arc . in lemma [ f0isahomeomorphism ]
we will show that @xmath125 must be an embedding .
as the @xmath257 s are all homeomorphisms , and @xmath125 is the limit of the sequence @xmath565 , there is an extension of @xmath125 to a simplicial homeomorphism @xmath312 from @xmath532 to @xmath535 . by construction
, @xmath312 is a homeomorphism which is isotopic to @xmath4 , and @xmath312 minimizes @xmath190 among all homeomorphisms isotopic to @xmath4 .
in particular , @xmath312 is a critical point for @xmath190 under all variations through homeomorphisms . as any small deformation of @xmath312
must still be a homeomorphism , this implies that @xmath312 is a critical point for @xmath190 under all variations .
thus @xmath312 is simplicial harmonic .
this completes the proof of the first part of the theorem .
if @xmath535 has negative curvature , proposition [ uniqueharmonicmapsexist ] tells us that the homotopy class of @xmath4 contains only one simplicial harmonic map , and that this map minimizes @xmath190 in the homotopy class .
it follows that @xmath312 is this map , which completes the proof of the theorem subject to giving the proof of lemma [ f0isahomeomorphism ] . in order to show that @xmath125 must be an embedding
, we need to analyze how this might fail .
first we consider what happens if an edge of @xmath11 is collapsed to a point by @xmath125 .
note that , as @xmath11 is good , this can not happen for an edge which is a loop . before dealing with the general case , we consider the following special case .
[ singleedgecollapsed]suppose that @xmath534 and @xmath533 are closed surfaces with negative euler number .
let @xmath541 be a homeomorphism from @xmath534 , with simplicial metric @xmath12 , to @xmath533 with riemannian metric @xmath406 , of non - positive curvature .
let @xmath567 be a map which is the limit of the restriction to @xmath538 of a sequence of homeomorphisms from @xmath534 to @xmath533 , each of which is isotopic to @xmath4 , and whose simplicial energies approach @xmath537 .
suppose that there is an edge @xmath512 of @xmath11 which is mapped to a point by @xmath125 , such that none of the edges incident to @xmath512 is mapped to a point .
then there is a geodesic @xmath170 through @xmath568 such that every edge of @xmath11 incident to @xmath512 has image contained in @xmath170 .
hence @xmath125 maps every edge in the star of @xmath512 into @xmath170 .
let @xmath143 and @xmath175 denote the vertices of @xmath512 .
let @xmath437 denote the unit circle in the tangent space to @xmath533 at @xmath568 .
each edge @xmath528 of @xmath11 which is incident to @xmath512 determines a point of @xmath437 which corresponds to the tangent vector to the image of @xmath528 . as
@xmath125 is a limit of homeomorphisms , we can cut @xmath437 at two points into two closed intervals @xmath569and @xmath159 such that one contains the points determined by edges of @xmath11 incident to @xmath143 , and the other contains the points determined by edges of @xmath11 incident to @xmath175 .
suppose that these endpoints of @xmath26 and @xmath159 are not diametrically opposite points of @xmath437 .
then we can cut @xmath437 at a pair of diametrically opposite points into two closed intervals @xmath570and @xmath520 one of which contains one of @xmath26 or @xmath159 in its interior . without loss of generality , we can suppose that @xmath571 contains @xmath26 in its interior .
now consider a homotopy of @xmath125 which moves @xmath557 a small distance towards the midpoint of @xmath571 , and does not move any other vertices of @xmath11 , and let @xmath510 denote the map so obtained . if @xmath528 is a geodesic segment in @xmath0 , and we move one end @xmath100 of @xmath528 along a geodesic making an angle @xmath572 with @xmath528 , then the derivative of the length of @xmath528 with respect to the distance moved by @xmath100 is @xmath573 .
in particular , if @xmath528 is an edge of @xmath11 other than @xmath512 , which is incident to @xmath143 , it must be shortened by such a homotopy , as @xmath572 is not equal to @xmath574 .
the homotopy does increase the length of @xmath512 , but the contribution of the length of @xmath512 to the energy of @xmath510 remains zero to first order .
it follows that @xmath510 has strictly less simplicial energy than @xmath125 , if we move @xmath557 a suitably small distance .
further @xmath510 is a limit of homeomorphisms @xmath550 . but
this contradicts the energy minimizing property of @xmath125 .
hence the endpoints of @xmath26 and @xmath159 must be diametrically opposite in @xmath437 .
if the interior of @xmath26 or the interior of @xmath159 contains any points corresponding to edges incident to @xmath512 , we can again obtain a contradiction by essentially the same argument .
note that , as we move the vertex @xmath143 , any edge @xmath528 of @xmath11 incident to @xmath143 whose tangent vector is an endpoint of @xmath26 will not change in length to first order , as the angle @xmath572 is equal to @xmath574 for such @xmath528 .
we conclude that there are only two tangent vectors determined by edges of @xmath11 incident to @xmath512 and that these two points of @xmath437 are diametrically opposite .
if @xmath170 denotes the geodesic through @xmath568 with these tangent vectors , then it follows that every edge incident to @xmath512 has image contained in @xmath170 .
if @xmath528 is an edge in the star of @xmath512 which is not incident to @xmath512 , then there are two edges of the star of @xmath512 which are incident to @xmath512 , and which together with @xmath575 bound a triangle of @xmath11 .
hence each vertex of @xmath528 is mapped into @xmath170 by @xmath125 , and @xmath576 is homotopic into @xmath170 fixing the vertices of @xmath528 .
now the non - positive curvature of @xmath533 implies that @xmath125 must map @xmath528 into @xmath170 , which completes the proof of lemma [ singleedgecollapsed ] . in the above result ,
the boundary @xmath515 of any triangle @xmath514 in the star of @xmath512 is mapped by @xmath125 into the geodesic @xmath170 , and @xmath515 is not mapped to a point . intuitively , the picture is that as @xmath459 , the sequence @xmath257 is collapsing the entire triangle @xmath514 into the geodesic @xmath170 . unfortunately , by itself the condition that @xmath125 maps the boundary of @xmath514 into @xmath170 is not enough to ensure that there is an extension of @xmath125 to a map of @xmath514 into @xmath170 .
for example if @xmath170 were a simple closed curve , @xmath125 could send @xmath515 to an essential loop in @xmath170 , and if @xmath170 were singular , @xmath125 need not even send @xmath515 to a loop in @xmath170 . in order to clarify the situation , we use the fact that if @xmath577 is a geodesic in the universal cover of @xmath533 which lies above @xmath170 , then @xmath577 is an embedded line .
this is because of the non - positive curvature of @xmath533 .
this leads us to the following definition .
let @xmath170 be a geodesic in @xmath533 .
then a triangle @xmath514 in @xmath11 is @xmath170__degenerate under _ _
@xmath125 if 1 . @xmath125 maps @xmath515 into @xmath170 , and 2 . the restriction of @xmath125 to @xmath515 lifts to a map of @xmath515 into @xmath577
. note that if @xmath514 is @xmath170degenerate under @xmath125 , then @xmath578 extends to a map of @xmath514 into @xmath170 .
the proof of the preceding lemma shows that if @xmath514 is a triangle in the star of @xmath512 , then @xmath514 is @xmath170degenerate under @xmath125 , but @xmath515 is not mapped to a point by @xmath125 . next we consider the general case when edges adjacent to @xmath512 may also be collapsed to the point @xmath568 .
we consider the subcomplex of @xmath11 consisting of all edges which are mapped to this point by @xmath125 together with each @xmath3simplex @xmath514 such that @xmath579 . as the image of @xmath125
can not be a point , this subcomplex is not equal to @xmath580
. let @xmath581 denote the component of this subcomplex which contains @xmath512 .
thus @xmath581 is not a single point .
lemma [ singleedgecollapsed ] is a special case of the following result , and much of the proof is similar .
we proved lemma [ singleedgecollapsed ] separately to clarify the argument .
note that @xmath125 is only defined on the @xmath90skeleton of @xmath581 , but in this setting we will abuse notation and denote @xmath568 by @xmath582 .
[ manyedgescollapsed]using the above notation , there is a geodesic @xmath170 through @xmath582 such that every triangle in the star of @xmath581 is @xmath170degenerate under @xmath125 .
as @xmath125 is homotopic to the restriction of a homeomorphism , the inclusion map of @xmath581 into @xmath534 must induce the trivial map of fundamental groups .
thus if @xmath318 denotes a regular neighborhood of @xmath581 , all except one of the components of the closure of @xmath583 must be a disc , and the union of @xmath318 with these discs must itself be a disc .
suppose there is a disc component @xmath159 of the closure of @xmath583 , and let @xmath517denote the component of @xmath584 which contains @xmath159 .
then @xmath519 is contained in @xmath585 , so is mapped to @xmath582 by @xmath125 . hence for @xmath10 large , @xmath519 is mapped arbitrarily close to @xmath582 by @xmath257 . as @xmath159 is a
disc and @xmath257 is a homeomorphism , it follows that the same holds for any edge of @xmath538 which is contained in @xmath520 , so @xmath125 must map each such edge to @xmath582 .
it follows that @xmath520 itself must be contained in @xmath581 , which is a contradiction .
we conclude that no component of the closure of @xmath583 can be a disc , so that @xmath318 itself is a disc . in particular , @xmath581 must be simply connected .
now let @xmath586 denote the subcomplex in which @xmath581 intersects the closure of @xmath584 . as @xmath581 is simply connected , a vertex of @xmath587separates @xmath581 if and only if it locally separates @xmath581 .
as @xmath581 is finite , there must be at least two vertices of @xmath586 which are `` extreme '' and so do not separate @xmath581 . as any such vertex does not locally separate @xmath581
, it must have connected link in @xmath581 .
let @xmath143 and @xmath175 denote two such vertices of @xmath586 .
now we will argue very much as in the proof of lemma [ singleedgecollapsed ] .
see figure [ kcollapse ] .
let @xmath437 denote the unit circle in the tangent space to @xmath533 at @xmath582 .
note that each edge of @xmath588 which is incident to @xmath581 can not be mapped to a point by @xmath125 .
thus each such edge @xmath512 determines a point of @xmath437 which corresponds to the tangent vector to the image of @xmath512 . as
@xmath125 is a limit of homeomorphisms , and @xmath143 and @xmath175 have connected link in @xmath581 , we can cut @xmath437 at two points into two closed intervals @xmath569and @xmath159 such that one contains the points determined by edges of @xmath588 incident to @xmath143 , and the other contains the points determined by edges of @xmath588 incident to @xmath175 .
suppose that these endpoints of @xmath26 and @xmath159 are not diametrically opposite points of @xmath437 .
then we can cut @xmath437 at a pair of diametrically opposite points into two intervals @xmath570and @xmath520 one of which contains one of @xmath26 or @xmath159 in its interior . without loss of generality , we can suppose that @xmath571 contains @xmath26 in its interior .
now a homotopy of @xmath125 which moves @xmath557 a suitably small distance towards the midpoint of @xmath571 , and does not move any other vertices of @xmath11 , will yield a map @xmath510 with strictly less simplicial energy than @xmath125 . for each edge of @xmath588 which is incident to @xmath143
is shortened , and although edges of @xmath581 which are incident to @xmath143 increase in length , the contribution of the length of each such edge to the energy of @xmath510 remains zero to first order .
also all other edges are unchanged .
further @xmath510 is a limit of homeomorphisms @xmath550 .
as before , this contradicts the energy minimizing property of @xmath125 , so that the endpoints of @xmath26 and @xmath159 must be diametrically opposite in @xmath437 .
but if the interior of @xmath26 or the interior of @xmath159 contains any points corresponding to edges of @xmath588 incident to @xmath143 or @xmath175 respectively , we can obtain a contradiction by essentially the same argument .
note that any edge of @xmath588 incident to @xmath143 whose tangent vector is an endpoint of @xmath26 will not change in length to first order .
we conclude that there are only two tangent vectors determined by edges of @xmath588 incident to @xmath143 or @xmath175 and that these two points of @xmath437 are diametrically opposite .
if @xmath170 denotes the geodesic through @xmath582 with these tangent vectors , then it follows that every edge of @xmath588 incident to @xmath143 or to @xmath175 has image contained in @xmath170 .
as @xmath125 is a limit of homeomorphisms , it follows that every edge of @xmath588 incident to @xmath581 has image contained in @xmath170 . as in the proof of lemma [ singleedgecollapsed ]
, it follows that @xmath125 maps every edge in the star of @xmath581 into @xmath170 .
it also follows that every triangle in the star of @xmath581 is @xmath170degenerate under @xmath125 , which completes the proof of lemma [ manyedgescollapsed ] .
note that , in the above result , if @xmath514 is a triangle in the star of @xmath581 which does not lie in @xmath581 , then @xmath514 is @xmath170degenerate under @xmath125 , but @xmath515 is not mapped to a point by @xmath125 .
now we simply suppose that we have a geodesic @xmath170 in @xmath533 and a triangle @xmath514 of @xmath11 such that @xmath514 is @xmath170degenerate under @xmath125 , but @xmath515 is not mapped to a point by @xmath125 .
we consider the subcomplex of @xmath11 which consists of the union of all triangles which are @xmath170degenerate under @xmath125 , and let @xmath571 be the component of this subcomplex which contains @xmath514 .
thus the restriction of @xmath125 to the @xmath90skeleton @xmath589 of @xmath571 lifts to a map of @xmath589 into @xmath577 .
let @xmath590 denote the intersection of @xmath571 with the closure of @xmath591 .
[ noedgeincidenttodlcollapses]using the above notation , if @xmath143 is a point of @xmath590 , and @xmath512 is an edge of @xmath11 which is incident to @xmath143 , then @xmath125 does not collapse @xmath512 to a point .
if @xmath125 collapses @xmath512 to a point , we consider the subcomplex of @xmath11 consisting of all edges mapped to this point together with all @xmath3simplices whose boundary is mapped to this point , and let @xmath581 denote the component of this subcomplex which contains @xmath512 .
lemma [ manyedgescollapsed ] tells us that @xmath125 maps every edge in the star of @xmath581 into some geodesic @xmath169 . as @xmath515
is not collapsed to a point , it follows that @xmath581 does not contain some edge @xmath209 of @xmath515 .
pick a path in @xmath571 which joins @xmath581 to @xmath209 .
some edge @xmath592 of this path must lie in the star of @xmath581 but not in @xmath581 .
this edge @xmath592 is not collapsed to a point by @xmath125 .
as @xmath592 lies in @xmath571 and in the star of @xmath581 , the image under @xmath125 of @xmath592 must lie in @xmath170 and in @xmath169 .
it follows that @xmath593 .
as the star of @xmath512 is contained in the star of @xmath581 , it follows that every triangle of the star of @xmath512 is @xmath170degenerate under @xmath125 , which contradicts the fact that @xmath143 lies in @xmath590 .
this contradiction completes the proof of the lemma . of course @xmath571 itself need
not be a surface , but it will fail to be a surface precisely at those vertices whose link is not connected .
we temporarily remove all such vertices and let @xmath594 denote the closure of the component of the resulting object which contains @xmath514 .
thus @xmath594 also need not be a surface but is obtained from a connected surface @xmath558 by ( possibly ) identifying certain vertices in its boundary . now we pick a boundary component @xmath437 of @xmath558 . note that any edge of @xmath437 also lies in @xmath590 , and so is not collapsed to a point by @xmath125 . in what follows
we will assume that @xmath437 embeds in @xmath534 .
if this is not the case , we can push @xmath437 slightly into the interior of @xmath558 to obtain an embedded circle @xmath595 in @xmath534 , and then apply the following arguments using @xmath595 in place of @xmath437 .
let @xmath596 and @xmath597 denote the universal covers of @xmath534 and @xmath533 respectively , where @xmath596 has the triangulation @xmath598 induced from @xmath11 and @xmath597 has the riemannian metric induced from @xmath406 .
let @xmath599 denote the @xmath90skeleton of @xmath598 , let @xmath600 be a map which covers @xmath125 , and let @xmath577 be a geodesic in @xmath597 above @xmath170 .
recall that , as @xmath533 is non - positively curved , @xmath577 is an embedded line .
let @xmath601 denote a component of the pre - image in @xmath596 of @xmath437 , chosen so that @xmath602 , and let @xmath603 denote the component of the pre - image in @xmath596 of @xmath558 which has @xmath601 as a boundary component . note that @xmath601 is a simple closed curve or a line embedded in @xmath604 .
recall that the restriction of @xmath125 to the @xmath90skeleton @xmath589 of @xmath571 lifts to a map of @xmath589 into @xmath577 .
in particular , it follows that the restriction of @xmath605 to @xmath601 maps @xmath601 into @xmath577 . [ chas0or2criticalpoints ] using the above notation , one of the following holds : 1 .
the map @xmath606 is a homeomorphism , or 2 .
@xmath601 is a circle , and the map @xmath606 has precisely two critical points . to prove this ,
choose a homeomorphism @xmath257 , for large @xmath10 , lift to the universal cover and consider the induced map from @xmath601 to @xmath597 .
this map embeds @xmath601 very close to @xmath577 , and edges of @xmath601 are almost geodesic . at each vertex @xmath143 of @xmath601 , either @xmath601 is almost straight , or it turns through an angle of approximately @xmath607 .
the vertices of @xmath601 where the turn angle is approximately @xmath607 correspond to the critical points of @xmath605 .
at such a vertex , @xmath603 must lie on the side of @xmath601 where the internal angle is almost zero . for any triangle of @xmath598 which is incident to @xmath143 and on that side of @xmath608
must be mapped to @xmath577 by @xmath605 and hence must lie in @xmath603 . if @xmath601 is a simple closed curve , we know it must have total curvature close to @xmath205 or to @xmath609 .
now each sharp bend contributes essentially @xmath610 , and every other vertex and every edge contributes essentially zero .
further the contributions of the sharp bends all have the same sign , as the side of @xmath601 where the internal angle is almost zero always lies in @xmath603 .
it follows immediately that @xmath605 has exactly zero or two critical points . in the first case
, @xmath605 would map @xmath601 to @xmath577 by an immersion which is impossible .
thus in this case , part 2 ) of the lemma must hold . if @xmath601 is a line ,
it must be properly embedded , so for any subinterval @xmath336 of @xmath577 , there is a subinterval @xmath611 of @xmath601 which starts at one end of @xmath336 and ends at the other .
for such an interval @xmath611 , the total curvature must be close to zero .
again the contributions of the sharp bends all have the same sign , so it follows immediately that @xmath611 has no critical points .
hence @xmath605 has no critical points , so that @xmath605 maps @xmath601 to @xmath577 by a homeomorphism , showing that part 1 ) of the lemma holds .
this completes the proof of the lemma . in the above lemma
, we showed that in all cases , either @xmath605 maps @xmath601 to @xmath577 by a homeomorphism , or @xmath601 is a circle and @xmath605 maps @xmath601 to @xmath577 with exactly two critical points . recall that @xmath437 is a simple closed curve .
in the first case , it follows that @xmath170 is a simple closed curve in @xmath533 , and that @xmath612 is a covering map .
as @xmath437 is embedded in @xmath534 , and @xmath534 is orientable , @xmath612 must be a homeomorphism .
now we are ready to prove that @xmath125 must be an embedding .
[ f0isahomeomorphism ] suppose that @xmath534 and @xmath533 are closed surfaces with negative euler number .
let @xmath541 be a homeomorphism from @xmath534 , with simplicial metric @xmath12 , to @xmath533 with a riemannian metric of non - positive curvature .
further assume that @xmath11 is a good triangulation .
let @xmath567 be a map which is the limit of the restriction to @xmath538 of a sequence of homeomorphisms from @xmath534 to @xmath533 each of which is isotopic to @xmath4 , and whose simplicial energies approach @xmath537
. then @xmath125 is an embedding .
recall from lemma [ edgesaremappedtogeodesics ] that @xmath125 maps each edge of @xmath11 to a point or to a geodesic arc .
suppose that @xmath125 is not an embedding . as @xmath125 is a limit of homeomorphisms , it must fail to embed the boundary of some triangle of @xmath11 .
thus one of the following cases occurs . 1 .
some vertex is mapped to the image of the interior of an immersed edge of @xmath11 , 2 .
some edge of @xmath11 is collapsed to a point by @xmath125 , or 3 .
some triangle of @xmath11 is mapped to a geodesic segment , not a point . in case 1 ) , there is a vertex @xmath143 and an immersed edge @xmath512 of @xmath11 with an interior point @xmath175 such that @xmath556 equals @xmath557 .
the proof of part 1 ) of theorem [ 1-vertextriangulation ] provides a contradiction if no edge of @xmath11 incident to @xmath143 is mapped into @xmath568 .
see figure [ 1vertex ] .
it follows that if case 1 ) occurs , we must also be in case 3 ) . in case 2 ) , if an edge @xmath512 of @xmath11 is collapsed to a point by @xmath125 , we define the subcomplex @xmath581 as we did just before lemma [ manyedgescollapsed ] .
that lemma implies that there is a geodesic @xmath170 such that if @xmath514 is a triangle in the star of @xmath581 which does not lie in @xmath581 , then @xmath514 is @xmath170degenerate under @xmath125 , but @xmath515 is not mapped to a point by @xmath125 .
thus in all cases , there is a geodesic @xmath170 in @xmath533 and a triangle @xmath514 of @xmath11 such that @xmath515 is not collapsed to a point by @xmath125 , and @xmath514 is @xmath170degenerate under @xmath125 . as in the preceding two lemmas
, we define the subcomplexes @xmath571 and @xmath594 of @xmath534 , and the surface @xmath558 , and we pick a boundary component @xmath437 of @xmath558 , which we will assume is embedded in @xmath534 .
lemma [ chas0or2criticalpoints ] implies that either @xmath125 maps @xmath437 to @xmath170 by a homeomorphism , or that @xmath601 is a circle and @xmath605 maps @xmath601 to @xmath577 with exactly two critical points . in either case , we claim there is a vertex @xmath143 of @xmath437 which is not a critical point of the map @xmath612 .
this claim is trivial in the first case . in the case when @xmath605 has two critical points
, @xmath601 is a circle so that @xmath437 must be null homotopic in @xmath534 . now
our hypothesis that @xmath11 is a good triangulation implies that @xmath437 must contain at least three vertices , so that at least one of these vertices is not a critical point , as claimed .
we let @xmath613 denote the star of @xmath143 in @xmath534 . as @xmath437 is a boundary component of @xmath558 , the two edges of @xmath437 incident to @xmath143 together divide @xmath613 into two discs , one of which
is contained in @xmath558 , and hence is contained in @xmath594 .
we denote this disc by @xmath182 , and let @xmath614 denote the other disc .
thus @xmath613 equals @xmath615 , and @xmath616 equals two edges of @xmath437 .
we claim that no edge of @xmath617 which is incident to @xmath143 can lie in @xmath571 .
let @xmath512 be an edge of @xmath617 which is incident to @xmath143 .
thus @xmath512 splits @xmath614 into two subdiscs @xmath618 and @xmath619 .
now suppose that @xmath512 lies in @xmath571 , so that @xmath620 .
recall from lemma [ noedgeincidenttodlcollapses ] , that @xmath125 does not collapse @xmath512 to a point . as @xmath125 is a limit of homeomorphisms
, it follows that @xmath125 maps all the edges in one of @xmath618 or @xmath619 into @xmath170 .
if @xmath621 , then @xmath622 must be contained in @xmath571 . hence @xmath623 is contained in @xmath571 . but
this implies that the interior of an edge of @xmath437 lies in the interior of @xmath571 which contradicts the definition of @xmath437 .
it follows that no edge of @xmath617 which is incident to @xmath143 can lie in @xmath571 , as claimed .
recall again that lemma [ noedgeincidenttodlcollapses ] shows that every edge incident to @xmath143 is not mapped to a point by @xmath125 . as @xmath125 is a limit of homeomorphisms , all the edges of @xmath617 which are incident to @xmath143 must be mapped by @xmath125 to the same side of @xmath170 .
note that this statement makes crucial use of the fact that @xmath143 is not a critical point of the map from @xmath437 to @xmath170 . consider a homotopy of @xmath125 which moves @xmath557 a small distance towards this side of @xmath170 , and does not move any other vertices of @xmath11 , and let @xmath510 denote the map so obtained .
each edge of @xmath617 which is incident to @xmath143 is not mapped into @xmath170 , and so is shortened by such a homotopy .
each edge of @xmath571 which is incident to @xmath143 is mapped into @xmath170 , so this homotopy does not change its length to first order .
it follows that @xmath510 has strictly less simplicial energy than @xmath125 , if we move @xmath143 a suitably small distance .
further @xmath510 is a limit of homeomorphisms . but this contradicts the energy minimizing property of @xmath125 .
this contradiction completes the proof of lemma [ f0isahomeomorphism ] .
we will end this section by discussing an application of the preceding arguments .
but first we note that the arguments of this section work perfectly well if @xmath534 and @xmath533 have non - empty boundary , and in this case , one need not insist that the metrics be strictly negatively curved to obtain uniqueness .
the result we obtain is the following .
[ harmonicsurfacehomeosexistrelboundary]suppose that @xmath534 and @xmath533 are compact surfaces with non - empty boundary .
let @xmath624 be a homeomorphism from @xmath534 , with simplicial metric @xmath12 , to @xmath533 with a non - positively curved riemannian metric @xmath406 such that @xmath625 is locally convex .
further assume that @xmath11 is a good triangulation .
then @xmath4 is isotopic rel @xmath626 to a simplicial harmonic homeomorphism @xmath566 that minimizes @xmath190 among all homeomorphisms isotopic to @xmath4 rel @xmath626 .
further , @xmath312 is the unique simplicial harmonic map in the homotopy class of @xmath4 , and so minimizes energy in this homotopy class .
now let @xmath0 be a compact surface with a riemannian metric @xmath406 , and let @xmath11 denote a good triangulation of @xmath0 .
we will say that a triangulation of @xmath0 is _ straight _ if each edge is a geodesic arc .
we will be interested in the number and homotopy type of the components of the space of straight triangulations of @xmath0 of a fixed combinatorial type .
note that theorems [ harmonicsurfacehomeoexists ] and [ harmonicsurfacehomeosexistrelboundary ] tell us that for any good triangulation @xmath11 of @xmath0 , the space of straight triangulations of @xmath0 modelled on @xmath11 is non - empty .
we will show that , with certain curvature restrictions , there is a natural bijection between the components of this space and the group of isotopy classes of homeomorphisms of @xmath0 .
when @xmath0 is the @xmath3disk with a flat metric , bloch , connelly and henderson @xcite proved this , and they proved , in addition , that the components of this space are homeomorphic to some euclidean space . in related work ,
awartani and henderson @xcite considered the problem of showing that when @xmath0 is the @xmath3sphere , the analogous space has the homotopy type of @xmath627 .
and bloch @xcite showed that the space of embeddings with geodesic edges and convex image of a triangulated @xmath3disk into the plane has the homotopy type of @xmath628 .
if we have a simplicial metric @xmath12 on @xmath11 , and if @xmath312 is any simplicial homeomorphism from @xmath20 to @xmath629 , then @xmath630 is a straight triangulation of @xmath0 .
conversely a straight triangulation of @xmath629 by @xmath11 determines a unique simplicial homeomorphism from @xmath20 to @xmath629 .
thus , after fixing @xmath11 and its simplicial metric , we can identify the space of straight triangulations of @xmath0 by @xmath11 with the space of simplicial homeomorphisms of @xmath0 .
note that any triangulation @xmath11 of @xmath0 admits a simplicial metric @xmath12 by specifying that @xmath12 takes the value @xmath90 on every edge of @xmath11 .
having fixed a good triangulation @xmath11 of @xmath0 and its simplicial metric @xmath12 , let @xmath631 denote the space of simplicial homeomorphisms from @xmath20 to @xmath629 .
this is a subset of the space @xmath632 of all homeomorphisms from @xmath0 to @xmath0 .
the components of @xmath632 naturally correspond to the isotopy classes of homeomorphisms of @xmath0 .
[ straighttriangulationsarestraightisotopic ] let @xmath0 be a compact surface with a riemannian metric @xmath406 of non - positive curvature such that @xmath130 is locally convex .
if @xmath0 is closed we suppose that @xmath406 has negative curvature .
let @xmath11 denote a good triangulation of @xmath0 , with simplicial metric @xmath12 .
using the above notation , each component of @xmath632 contains exactly one component of @xmath631 .
theorems [ harmonicsurfacehomeoexists ] and [ harmonicsurfacehomeosexistrelboundary ] imply that any homeomorphism @xmath633 is isotopic rel @xmath130 to a unique simplicial harmonic homeomorphism @xmath634 that minimizes @xmath190 among all homeomorphisms isotopic to @xmath4 .
in particular , each component of @xmath632 contains a point of @xmath631 .
note that the fact that @xmath631 is non - empty is already nontrivial to prove .
next we concentrate on homeomorphisms isotopic to @xmath4 .
let @xmath307 denote a component of the space of simplicial homeomorphisms from @xmath20 to @xmath629 which are isotopic to @xmath4 .
we will show that @xmath307 contains the energy minimizing simplicial harmonic homeomorphism @xmath312 .
as this holds for all such components , it will follow that the space of simplicial homeomorphisms from @xmath20 to @xmath629 which are isotopic to @xmath4 must be connected , which will prove the first part of the theorem .
now @xmath307 is contained in the space , @xmath635 , of all simplicial maps from @xmath20 to @xmath629 which are homotopic to @xmath4 .
the gradient flow of the energy functional on @xmath635 yields a deformation retraction of @xmath635 to the unique energy minimizer @xmath312 , which we know to be a homeomorphism .
it would be extremely convenient if this flow induced a deformation retraction of @xmath307 to this same energy minimizer .
one would need to show that if one starts with a simplicial homeomorphism from @xmath20 to @xmath629 , then the result of following the gradient flow remains a homeomorphism for all time .
unfortunately this is not the case .
see the end of this section for a discussion and example .
this means that we are unable to show that the components of @xmath631 are contractible , although we believe this to be the case .
instead we consider the closure @xmath636 of @xmath307 in @xmath635 .
as @xmath307 is open in @xmath635 , it follows that @xmath637 is closed in @xmath635 .
this implies that the restriction of the energy functional to @xmath637 attains a minimum . to see this ,
recall that the proof of proposition [ uniqueharmonicmapsexist ] showed that given any constant @xmath581 , the subspace of @xmath635 consisting of maps with energy @xmath638 is compact .
thus the same holds for the intersection of this subspace with @xmath639 .
now let @xmath363 be a point of @xmath637 of least energy .
thus @xmath363 is not a homeomorphism , and is a limit of simplicial homeomorphisms @xmath257 in @xmath307 .
we make the following claim .
there is @xmath10 such that we can isotope @xmath257 to a simplicial homeomorphism @xmath550 whose energy is strictly less than that of @xmath363 .
of course , such @xmath550 also lies in @xmath307 .
assuming this claim , consider the gradient flow starting at @xmath550 .
any map in this flow has energy less than that of @xmath550 , and so less than that of @xmath363 .
hence no map in this flow can lie in @xmath637 .
hence the flow stays in @xmath307 and so determines an isotopy of @xmath550 to the unique energy minimizer @xmath312 .
in particular it follows that @xmath307 contains @xmath312 , which completes the proof of the theorem .
it remains to prove the above claim .
we would like to prove this in much the same way as we proved lemma [ f0isahomeomorphism ] .
specifically we want to use lemmas [ manyedgescollapsed ] , [ noedgeincidenttodlcollapses ] , [ chas0or2criticalpoints ] and [ f0isahomeomorphism ] .
as stated , these lemmas do not apply to our situation , as @xmath363 is not an energy minimizer for all homeomorphisms .
but we can apply the arguments in these lemmas as we now describe .
recall that all these lemmas consider a map @xmath567 which is the limit of the restriction to @xmath538 of a sequence of homeomorphisms from @xmath534 to @xmath533 , each of which is isotopic to @xmath4 , and whose simplicial energies approach @xmath537 .
lemma [ edgesaremappedtogeodesics ] shows that any such map sends each edge of @xmath11 to a point or a geodesic segment . in the present situation
, @xmath363 is a limit of simplicial homeomorphisms , so it is immediate that it sends each edge of @xmath11 to a point or a geodesic segment , but it need not have the minimal energy @xmath537 .
lemma [ manyedgescollapsed ] used the least energy assumption on @xmath125 to show that if @xmath125 collapsed some edge @xmath512 to a point , then it must map the star of some complex @xmath581 which contains @xmath512 into some geodesic @xmath170 .
now suppose that @xmath363 collapses some edge @xmath512 to a point .
the arguments in the proof of lemma [ manyedgescollapsed ] show that either we get the same conclusion for @xmath363 , or we can isotope some @xmath257 to a simplicial homeomorphism @xmath550 whose energy is strictly less than that of @xmath363 . an alternative way of putting this is to say that either the claim holds or the result of lemma [ manyedgescollapsed ] holds for @xmath363 .
lemmas [ noedgeincidenttodlcollapses ] and [ chas0or2criticalpoints ] depend on lemma [ manyedgescollapsed ] but do not otherwise use the energy minimizing assumption on @xmath125 .
thus again we have that either the claim holds or their results also hold for @xmath363 .
finally the arguments in the proof of lemma [ f0isahomeomorphism ] use the conclusions of lemmas [ noedgeincidenttodlcollapses ] and [ chas0or2criticalpoints ] to show that if @xmath125 is not a homeomorphism we can isotope @xmath257 to a simplicial homeomorphism @xmath550 whose energy is strictly less than that of @xmath125 . as @xmath363 is not a homeomorphism ,
when we apply these arguments to @xmath363 , it follows that the claim must hold .
this completes the proof of the claim , and hence of theorem [ straighttriangulationsarestraightisotopic ] .
we end this section by discussing the fact that the gradient flow need not preserve the property of being a homeomorphism .
recall from section [ simplicialenergy ] the formula for the simplicial energy of a map .
if we consider varying a single vertex @xmath100 , only those edges incident to @xmath100 can contribute to any change in the energy .
to each such edge @xmath512 we associate a vector @xmath640 in the tangent space at @xmath100 so that @xmath640 is tangent to @xmath512 , has magnitude equal to the length of @xmath512 , and points away from @xmath100 .
suppose that every edge of @xmath11 has length @xmath90 in our simplicial metric .
then a simple calculation shows that under the gradient flow , the flow vector of @xmath100 is simply the sum of the vectors @xmath640 .
thus @xmath100 is `` pulled '' much more by long edges than by short ones .
if a configuration like the one in figure [ nonembedded ] is part of a triangulation of a surface , it will quickly cease to be embedded under the gradient flow .
moves rightward faster than vertex @xmath427.,width=192 ]
in the previous sections we considered maps of a triangulated closed surface @xmath0 into a non - positively curved closed manifold @xmath1 whose dimension was unrestricted . in this section
we observe that all the preceding work can be generalized to maps of a triangulated closed @xmath281manifold @xmath0 into a non - positively curved manifold @xmath1 , for any @xmath641 .
we do this by restricting attention to the @xmath3skeleton @xmath642 of the triangulation @xmath11 of @xmath0 .
this seems reasonable as the homotopy class of a map of @xmath0 into @xmath1 is determined by its restriction to @xmath642 , as @xmath1 is aspherical . in the smooth setting
, much work has been done in this setting .
see @xcite@xcite . for simplicity
we discuss only the case when @xmath643 . as in the @xmath3dimensional case
, we specify a triangulation @xmath11 of the closed @xmath6manifold @xmath0 , and a map @xmath12 that assigns to each edge @xmath13 , @xmath96 , of @xmath11 a length @xmath15 , with the lengths @xmath16 realizable by a euclidean tetrahedron for each tetrahedron of @xmath11 .
we call such an assignment a _ simplicial metric _ on @xmath0 , and denote it by @xmath20 , or just @xmath12 when the context is clear
. note that we allow the possibility that for some triangles one of the triangle inequalities is an equality . as before we order the vertices of @xmath11 using a @xmath19-complex structure .
given a map @xmath38 , let @xmath39 denote the length of the restriction of @xmath4 to the edge @xmath13 , and as before define the @xmath155th stretch factor to be @xmath110 .
the _ simplicial @xmath3-volume of @xmath4 _ is defined to be @xmath645 , and the _
simplicial @xmath3-energy of the map @xmath4 _ to be @xmath646 , where the sum is taken over all corners of all the @xmath3simplices of @xmath11 .
next we need to define when a map from @xmath0 to @xmath1 is simplicial . as before we insist that each edge of the triangulation @xmath11 is mapped to @xmath1 as a geodesic arc . in order to determine @xmath4 canonically on the remainder of @xmath0 , we use the ordering of the vertices of @xmath11 . for each triangle of @xmath11
this determines a cone structure , with cone point the minimal vertex of the triangle , and as before this determines a canonical extension of @xmath4 to the @xmath3skeleton of @xmath0 .
we give each tetrahedron @xmath21 of @xmath11 the cone structure from its minimal vertex @xmath143 .
note that all the faces of @xmath21 which meet @xmath143 also have cone structure with @xmath143 as the cone point .
thus the canonical extension of @xmath4 to the tetrahedra of @xmath11 yields a well defined simplicial map to @xmath1 .
as in lemma [ e > a ] , for any simplicial metric on @xmath0 , we have the inequality @xmath647 with equality if and only if all the stretch factors are equal .
we say that @xmath4 is _ @xmath3-simplicial harmonic _ if it is a critical point of the @xmath3-energy functional .
as in proposition [ uniqueharmonicmapsexist ] , if @xmath1 is closed , the energy functional has a minimum among all maps homotopic to @xmath4 , and we can choose any minimizing map to be simplicial . further
if @xmath1 is negatively curved and @xmath4 is nontrivial , i.e. @xmath4 can not be homotoped to have image contained in a closed geodesic , then this minimum energy simplicial map @xmath312 is unique . as in section [ families ] , the fact that any map @xmath38 can be homotoped to a unique simplicial map without moving the vertices allows one to construct families of harmonic maps with uniform area bounds . in the preceding paragraphs , we prefixed @xmath3 to our definitions , because there seem to be some natural alternative definitions .
for example , we could define the _
simplicial @xmath6-volume _ of @xmath70 to be the sum @xmath648 , where the sum is taken over all corners of all the @xmath6simplices of @xmath11
. then we would define the _
simplicial @xmath6-volume of @xmath4 _ to be @xmath649 , and the _ simplicial @xmath6-energy of @xmath4 _ to be @xmath650 . in general ,
if the dimension of @xmath0 equals @xmath651 , then one could similarly define the simplicial @xmath281volume of @xmath0 , and the simplicial @xmath281energy of @xmath4 , for any @xmath281 such that @xmath652 .
however one can not expect there to be any connection between simplicial @xmath281volume and simplicial @xmath281energy , except when @xmath653 .
m. desbrun , m. meyer , p. schroder , a. barr , _ implicit fairing of irregular meshes using diffusion and curvature flow _ , proceedings of the 26th annual conference on computer graphics and interactive techniques , ( 1999 ) , 317324 . | we introduce a combinatorial energy for maps of triangulated surfaces with simplicial metrics and analyze the existence and uniqueness properties of the corresponding harmonic maps .
we show that some important applications of smooth harmonic maps can be obtained in this setting . |
chromospheric wave propagation has been an extensively studied , but poorly understood , subject in solar physics for some time .
one puzzle has been the presence of propagating waves with periods on the order of 5 minutes or more .
such waves were observed by , e.g. , @xcite and @xcite ; this was considered surprising since the acoustic cutoff period in the chromosphere , above which waves should not be able to propagate , is on the order of 200 s. such long - period propagation was later found to be widespread , generally occurring wherever the local magnetic field is strong .
@xcite show an example of observations of neighboring internetwork and network regions that have a marked difference in their fourier spectra .
it was realized @xcite that , since magnetoacoustic waves in a strongly magnetized medium are restricted to propagating along field lines , the effective gravity ( i.e. , the component of gravity along the magnetic field ) would be reduced in magnetic regions . since the cutoff frequency ( in an isothermal atmosphere ) is given by @xmath0 where @xmath1 is the sound speed and @xmath2 is the ratio of specific heats , the cutoff frequency would also be lower in regions of strong inclined field , potentially allowing 5-minute ( @xmath33 mhz ) waves to propagate .
this hypothesis has since been tested in a number of increasingly advanced numerical simulations , e.g. @xcite , @xcite , and @xcite .
all have found that it is an effective mechanism for transmitting 5-minute power through the chromosphere ; some models have suggested that the leakage of 5-minute waves into the chromosphere can also explain the presence of 5-minute oscillatory signal in the corona @xcite .
one criticism of this explanation has been that not all observations of 5-minute propagation are in regions of obviously inclined field .
an alternative explanation has therefore been suggested , originally by @xcite and later developed and tested by @xcite and @xcite , in which changes in the radiative relaxation time associated with small scale magnetic structures are responsible for increasing the cutoff period .
this mechanism has also been demonstrated to enable propagation of 5-minute oscillations , even in vertical magnetic structures . however , the underlying basis of this theory and the related simulations is a highly simplified energy equation in which radiative losses are approximated by newton s law of cooling ( @xcite @xcite and references therein ) .
a related subject that has been studied extensively is the generation and propagation of chromospheric jets such as spicules ( type i and ii ) , macrospicules , surges , fibrils , and mottles .
a variety of models have been proposed over the years ( see @xcite @xcite for a review ) , and although it remains unclear whether the many types of jets are different manifestations of the same underlying physical phenomenon or not , there are at least strong indications that the jets known as dynamic fibrils are driven by shock waves traveling through the chromosphere @xcite . many of these jets have lifetimes around 5 minutes , and so the waves driving them need to have been channeled into the upper chromosphere via one of the processes outlined above . in this paper , we present the results of two - dimensional simulations of wave propagation from the convection zone to the transition region and corona .
the simulations include a sophisticated treatment of radiative losses and heat conduction , and study the effects of different magnetic field geometries and strengths on the propagation of waves through the chromosphere .
we also look at the jets that these waves produce once they reach the transition region , and perform a statistical comparison of the jets produced in a model with vertical field and in one with inclined field . in 2
we describe the different simulations and the code .
3 contains an analysis of wave propagation and periodicities in the various models , and the results are discussed in 4 .
5 looks at jet formation and properties , and a summary and conclusions follow in 6 .
the simulations have been run using the _ bifrost _ code ; a detailed code description has recently been published @xcite .
it is a 3d mhd code that is designed to model the solar atmosphere to a high degree of realism by including as many of the relevant physical processes as practical and possible .
the code is staggered such that scalars are defined on the grid interior while fluxes are defined on the edges between computational cells .
sixth order polynomials are used to calculate derivatives .
high order interpolation is also used when a given variable needs to be evaluated half a grid cell away . in order to maintain stability , artificial viscosity and resistivity terms with strengths that can be set by the user
are included following a hyperdiffusive scheme , such that regions of large gradients are also those with the largest viscosity and magnetic diffusivity .
this allows the diffusivity to be low throughout most of the simulation box , only becoming significant where the conditions require it .
lccl case a & @xmath4 & @xmath5 & extremely weak + case b & @xmath4 & @xmath5 & moderate , vertical + case c & @xmath6 & @xmath7 & strong expanding tube + case d & @xmath4 & @xmath5 & moderate , inclined + case e & @xmath4 & @xmath5 & strong , inclined the code includes thermal conduction along magnetic field lines , computed implicitly using a multigrid algorithm .
a realistic equation of state is pre - computed based on lte , and stored in tabular form to compute temperatures , pressures , and radiation quantities , given the mass density and internal energy .
the code solves the equations of radiative transfer for the photosphere and lower chromosphere , including scattering , using the multigroup opacity methods developed by @xcite and @xcite in short characteristic form @xcite . in the upper chromosphere , transition region , and corona ,
we utilize an advanced radiative loss function consisting of several parts . in the corona and transition region ,
the optically thin approximation is used , taking into account radiation from hydrogen , helium , carbon , oxygen , neon , and iron using atomic collisional excitation rates from the hao - diaper atom data package @xcite . in the upper chromosphere ,
most important lines are not in local equilibrium , and excitation rates are precomputed and tabulated from 1d chromospheric models with radiative transfer calculated in detail ( e.g. , * ? ? ?
* ; * ? ? ?
more details on these radiative loss functions as well as on the code itself can be found in @xcite and @xcite . in sum ,
the code thus includes most physics that are important in the solar atmosphere ; the main omission in the present work is time - dependent ionization , which mainly has an effect in the upper chromosphere and transition region , above a height of @xmath31 mm .
see @xcite for a discussion of these effects .
the horizontal boundary conditions are periodic , while both the vertical boundaries are open .
the upper one is based on characteristic extrapolation of the magnetohydrodynamic variables ; the lower one allows outflowing material to leave the box , while the entropy of the inflowing material is set .
there is no piston or other imposed external driver producing wave motions ; all oscillations are produced self - consistently by the turbulent motion in the convection zone .
, while the gray lines are magnetic field lines .
case a is high-@xmath8 throughout and contains only a very weak magnetic field . the dotted vertical black line in case c marks the location of the horizontal boundary ;
as this model has a smaller horizontal extent than the others , the rightmost 5.5 mm have been repeated on the left . ] , while the gray lines are magnetic field lines . ]
we will be analyzing five different simulations , using model atmospheres that are about 16 mm high .
they extend from the upper layers of the convection zone at the lower boundary ( @xmath9 mm ) , through the photosphere ( @xmath10 mm corresponding to the average height where @xmath11 ) , chromosphere , transition region and corona , with the upper boundary at @xmath12 mm .
the simulation boxes are two - dimensional and contain @xmath4 grid cells , using a uniform horizontal spacing of 32.5 km . in the vertical direction
, we use a non - uniform spacing which is 28 km between each cell from the lower boundary to @xmath13 mm , increasing gradually until @xmath14 mm ; from there to the upper boundary it stays at a constant value of 150 km .
the properties of our five different simulations are summarized in table [ simtab ] , and figure [ initfig ] shows their initial temperature and magnetic field structures .
case c has slightly smaller dimensions and higher resolution than the others ; the other models differ mainly in the strengths and orientations of their magnetic fields .
all the simulations have been run for 4500 s , or 75 minutes , of solar time ( after an initial relaxation period ) .
the turbulent motion in the convection zone is able to move magnetic flux tubes around , which means that the field configurations in the high-@xmath8 sections ( where @xmath8 is the ratio between the gas and magnetic pressures ) of all five simulations are rather dynamic .
in contrast , the field in the upper chromosphere and corona in all models except case a is rather stable and changes little with time .
the motion in the convection zone also generates waves that propagate upwards ( see figure [ propagation ] ) and create jets of various lengths and lifetimes once they reach the transition region .
many such jets can be seen in figure [ initfig ] .
they tend to follow the magnetic field , taking on relatively thin , elongated shapes , but in the non - magnetic case a , they look more like extended spherical waves moving the transition region up and down . since we are primarily interested in wave propagation in the chromosphere , not the corona , we omit from our analysis the regions above @xmath15 mm , the height of the peaks of the tallest jets . in figure [ tprc ]
we show the temperature , pressure , and density structures of case d in greater detail as functions of height , as well as the cutoff frequency ( given by equation [ cutoffeq ] ) .
the other cases look similar .
the temperature plot clearly shows the significant movement the transition region undergoes as jets are continually formed and push it upwards .
the three prominent horizontal lines in the temperature structure correspond to the ionization temperatures of hydrogen and helium ( the latter twice ) , and so are an effect of the tabulated equation of state @xcite .
the plot of the cutoff frequency shows the region , on average between about @xmath10 mm and @xmath16 mm but with large variation in time and space , where the cutoff is above @xmath17 mhz , corresponding to a period of 5 minutes .
this is the region where such long - period waves are evanescent , and which they will have to `` tunnel '' through if they are to survive into the upper layers of the atmosphere . in figure [ cutoffval ]
we have plotted the average cutoff frequency in our case d and the cutoff calculated from the semi - empirical val3c model @xcite .
the val model has a somewhat flatter cutoff in the high chromosphere than our average model , but agrees that the cutoff frequency remains high until one reaches the transition region at about @xmath18 mm .
equation [ cutoffeq ] , which we used in calculating the cutoff , is strictly speaking only valid in an isothermal atmosphere .
there are many possible definitions of the cutoff frequency in a stratified atmosphere @xcite , and the values in figures [ tprc ] and [ cutoffval ] should not be considered exact .
the main point is to illustrate that the solar atmosphere contains a height range where the cutoff frequency is higher than the frequencies associated with photospheric oscillations such as @xmath19-modes .
we should point out that such long - period waves have very long wavelengths , so `` local '' conditions are not a very well defined term when analyzing them
. a wave with a 5-minute period will have a wavelength of 2100 - 3000 km , given typical sound speeds in the lower atmosphere of 7 - 10 km s@xmath20 .
the entire height of the photosphere and chromosphere therefore amounts to less than one wavelength . from this
, we may expect that even in the absence of any channeling , through field inclination or other processes , a portion of the 5-minute power could make it through to the upper layers .
another expectation is that such channeling does not necessarily have to be active throughout the entire 2 mm high region , though the effect will likely be greater the more prevalent conditions conducive to channeling are ( in both space and time ) .
this paper aims to bring greater clarity to issues such as the effectiveness of channeling mechanisms and the effects of propagating long - period waves on higher atmospheric layers . to that end ,
our analysis focuses on two separate but related phenomena : first , the propagation of waves of different periodicities through the atmosphere and the influence upon them by the magnetic field , and second , the jets produced by these waves once they reach the transition region . to study the former , we perform an extensive fourier and wavelet analysis of the five simulations . for the latter , we individually measure properties such as the lifetimes , lengths and maximum velocities of the jets , and make a statistical comparison between case b ( vertical field ) and case d ( inclined field )
mm in case a , with superimposed contours showing the regions where @xmath21 at @xmath10 mm is greater than @xmath22 g. the white dashed line shows the location of the periodic boundary ; the leftmost @xmath23 mm are repeated on the right . ]
we start by analyzing case a , a reference model that contains only a very weak magnetic field .
it is everywhere weaker than 1 g , and the model s behavior is expected to be essentially non - magnetic . for consistency with the other models ,
however , the magnetic field is included in the analysis below .
the initial temperature state is shown in the top panel of figure [ initfig ] .
this model is meant to be representative of conditions in the internetwork .
the turbulent motion of the convection zone generates waves and flows at many different periods and in many locations .
figure [ weakbuz ] shows a plot of the vertical velocity ( grayscale ) as a function of horizontal position ( @xmath24 ) and time , at a constant height ( @xmath25 ) of 1 mm .
the overplotted black contours , for consistency with the other cases , mark the locations where the magnetic field is strongest ; specifically , where the vertical component of the magnetic field at @xmath10 mm ( 1 mm below the height at which the velocity is plotted ) is @xmath22 g or greater .
obviously , these `` flux concentrations '' are still very weak .
since one of the flux concentrations lies right on the ( periodic ) boundary , the plot shows the leftmost 3.5 mm of the box again on the right ; the location of the boundary is marked with a vertical white dashed line . in this case , the wavefronts appear to be rather uniformly distributed , with no strong clustering into specific regions either in time or space .
they are typically horizontally coherent on length scales of 3 - 5 mm .
figure [ weakbuz2 ] shows the vertical velocity at the horizontal position @xmath26 mm , allowing us to see the wave shapes in more detail .
the waves are notably asymmetric , showing the characteristic `` n''-shape of compressive waves that have steepened into shocks as a result of the density stratification .
the amplitude varies quite a bit with time due to the randomness of the photospheric driver .
the strongest shock trains appear around @xmath27 s , @xmath28 s , and towards the end of the simulation . at these times , the shocks reach peak - to - peak amplitudes of up to 30 km s@xmath20 , generally with stronger downflows than upflows .
in fact , the time - averaged velocity at this location is a downflow of about 3 km s@xmath20 , although this does not imply a net downward mass flux , since the shock waves propagating upward have greater densities .
the time between velocity peaks is typically around 200 s. we have created power spectra of the simulations by performing fourier transforms of the velocity at each grid cell .
an example of the resulting spectra is shown in figure [ cuzpow ] .
the lower panel shows the power spectrum itself as a function of @xmath24 and frequency @xmath29 at our analysis height of @xmath30 mm , the same height where we plotted the velocity field in figures [ weakbuz ] and [ weakbuz2 ] .
this height is also marked by the light gray horizontal line in the upper panels .
the bell - shaped black curves on either side delineate the two dominant frequency bands used in the subsequent analysis , the `` 5-minute '' band centered at @xmath31 mhz and the `` 3-minute '' band centered at @xmath32 mhz .
both are gaussian bands including contributions at frequencies up to 1 mhz above and below the central frequency .
the central frequencies correspond to actual periods of 333 s and 200 s , respectively .
the upper panel shows the ratio between the power in the 3-minute band and that in the 5-minute band , showing 3-minute dominance as pink and white , 5-minute dominance as green and black , and approximately equal power as blue .
superimposed on this plot are the height where the power spectrum in the lower panel is plotted ( light gray horizontal line ) , and magnetic field lines calculated from the time - averaged field ( dark gray ) .
the small panel to the right shows the height profile of the total velocity fourier power across all frequencies , multiplied by the @xmath24-averaged density and sound speed to create a measure of the energy flux density rather than the straight velocity power .
the middle panel of the figure shows the total time - averaged magnetic field at @xmath30 mm ( black ) and at @xmath10 mm ( gray ) . in this model , the 3-minute band is dominant throughout the chromosphere ( @xmath33 mm ) . only in a small area near @xmath34 mm
does the 5-minute band achieve rough parity , but a look at the spectrum ( lower panel ) shows that the dominant frequencies at that point are at or above 6 mhz , and begin to fall outside the 3-minute band .
in fact , there is significant power at frequencies of @xmath35 mhz across most of the simulation box .
this picture , with 3-minute domination throughout , is typical of the classical results from observations of weakly magnetic regions in the internetwork ( e.g. , @xcite @xcite and references therein ) .
it is a natural consequence of the influence of the acoustic cutoff frequency and the radiative damping of higher - frequency waves . under chromospheric conditions ,
the acoustic cutoff frequency is typically around @xmath36 - 5 mhz , corresponding to periods of 200 - 220 s ( see figure [ cutoffval ] ) , and waves at lower frequencies will have difficulty propagating upwards .
it is , however , possible to see small amounts of 5-minute power across much of the box , particularly around @xmath37 mm , @xmath38 mm , and @xmath39 mm .
this shows that although the long - period waves are damped , their long wavelength still ensures that part of their power reaches @xmath30 mm .
mm in case b , with superimposed contours showing the regions where @xmath21 at @xmath10 mm is greater than 300 g. the white dashed line shows the location of the periodic boundary ; the leftmost @xmath23 mm are repeated on the right .
( an mpeg animation of this figure is available in the online journal . ) ] having studied an essentially non - magnetic model , we now turn our attention to models with magnetic fields of varying strengths and orientations .
case b is a model with several flux concentrations in the photosphere .
these flux concentrations are relatively narrow ( a few hundred km ) and have strengths of @xmath40 g. such concentrations are a result of the convective motions below the photosphere , which tend to concentrate the magnetic field in relatively narrow regions , often connected with an average downflow , such as intergranular lanes . as we move upwards , the field expands , creating inclined field lines on the edges of the flux concentrations in the chromosphere , while once we reach coronal heights , the field fills all space and the orientation is close to vertical .
these effects can be seen in the second panel from the top of figure [ initfig ] , which shows the initial state of this case .
this model is intended to represent solar conditions in and around network or plage regions . , but showing the field - aligned ( rather than vertical ) velocity at @xmath30 mm .
this is the velocity that is used in the subsequent fourier and wavelet analysis . ]
the velocity field of this case is shown in figures [ vertbuz ] and [ vertbud ] . as previously , the grayscale image shows the velocity , the black contours outline the field concentrations at @xmath10 mm , now using a threshold of 300 g , while the white dashed line shows the location of the periodic boundary .
figure [ vertbuz ] shows the vertical velocity at @xmath41 mm . at photospheric heights ,
the velocity field is dominated by `` global '' oscillations with large horizontal wavelengths and periods around 5 minutes . at @xmath41 mm , there are still some remnants of these , for example at @xmath42 mm after 2500 s , but the shocks generated in and around the flux concentrations are stronger and have begun to fan out .
this figure is available as an animation in the online journal , showing how the velocity field develops from @xmath10 mm to @xmath18 mm .
the animation shows quite clearly how the waves propagate upwards and fan outwards .
figure [ vertbud ] shows the velocity _ along the magnetic field _ at @xmath30 mm , and it is this velocity component we will be performing our analysis on in this and all subsequent cases . since waves are forced to propagate along the magnetic field in regions where the field is dominant , this should give us a cleaner velocity signal than the vertical velocity ; the reason the field - aligned velocity was not used in case a is that it is not a very meaningful quantity in weak - field ( high-@xmath8 ) regions . unlike in case
a , the velocity field at @xmath30 mm is now far from uniformly distributed . as we see , all strong flux concentrations are associated with a region of high velocity amplitude higher up , and nearly all regions of high amplitude are connected with flux concentrations .
one main exception is the region of relatively high - frequency signal that starts out at @xmath37 mm , eventually moving to @xmath43 mm .
this area is situated between two flux concentrations and is a region of interference between waves coming from either side one can see the interference pattern developing in the animated version of figure [ vertbuz ] , which shows the evolution of the wave pattern with height . some other regions between flux concentrations , for example around @xmath44 mm and for the last 1000
s around @xmath37 mm , show quite weak signals .
the many curved wavefronts are a consequence of the outward propagation of waves from the flux concentrations ; particularly notable examples are found at @xmath45 mm , @xmath43 mm , and near the boundary at @xmath46 mm .
it is also interesting to note how much the flux concentrations are moved around by the convective motion one flux tube starts out at @xmath45 mm and ends up at @xmath43 mm , and several flux concentrations merge during the simulation .
this horizontal movement is on the order of 1 km s@xmath20 , which is within the ranges reported from observations of the movement of bright points in intergranular lanes @xcite . as in case a
, we have performed a fourier analysis to investigate the periodicity of the velocity signal .
the results are shown in figure [ auzpow ] , which , like figure [ cuzpow ] for case a , shows the power spectrum in the bottom panel , the time - averaged magnetic field in the middle panel , and the ratio of power in the 3-minute and 5-minute bands in the top panel .
as before , the horizontal line in the top panels shows the analysis height ( @xmath30 mm ) , and the dark gray lines are time - averaged field lines ; the new thick white line indicates the time - averaged height where @xmath47 , the dividing line between the regions dominated by the magnetic field ( above the line ) and by gas pressure forces ( below it ) .
although the central flux tube appears very dominant when looking at the time - averaged field strength , it is in fact not that much stronger than the others ; however , it moves much less horizontally , making it stronger when averaging over the whole simulation .
the top panel shows that the 3-minute band remains dominant in the chromosphere .
this is as expected from classical theory , since the acoustic cutoff period is shortest there , and 5-minute disturbances are expected to be evanescent .
it is notable that the peaks in the 3-minute power seem to correspond very well with the peaks in the magnetic field strength .
if the effects of the reduced radiative relaxation time in strongly magnetized regions were important , these are locations where 5-minute propagation would be expected . there are , however , several windows in the chromosphere where there is also significant power in the 5-minute band , notably around @xmath43 mm , @xmath48 mm , and @xmath49 mm .
the latter two are regions of inclined field on either side of the central strong flux concentration .
this fits well with the results of many previous simulations @xcite that have found that otherwise evanescent long - period disturbances can still propagate upwards along strong , inclined field , since the reduced effective gravity increases the acoustic cutoff period .
we should point out that the previous simulations have generally assumed strong , inclined flux tubes all the way down to the photosphere , while we find that long - period propagation occurs even when only a part of the propagation is along strong , inclined field at the edges of a mostly vertical flux concentration .
on the other hand , @xmath43 mm is a region where the average field inclination is not particularly high , at least not compared to the regions surrounding it . however , as we have seen from figure [ vertbud ] , the field is quite dynamic , and the time - averaged field inclination shown in the top panel of figure [ auzpow ] does not tell the whole story .
in fact , a look at figure [ vertbud ] shows that @xmath43 mm connects to two different photospheric flux concentrations over the course of the simulation , and it seems to be particularly during the time when the region is on the outer edge of the central flux tube ( from 1000 to 2500 s ) that it is dominated by longer - period waves .
so we need to take the variations with time into account , not just the averages over the whole simulation . the fourier analysis deals with averages over the whole analyzed time period , and in order to get sufficient spectral resolution ,
one typically needs time series on the order of one hour .
the solar atmosphere , however , is dynamic on timescales much shorter than that , and when conditions at one location change significantly during the analysis time , the fourier analysis averages this out and can be misleading .
a wavelet analysis , which takes variations in time into account , is a more precise tool ; we have performed such an analysis at some selected locations . in figure
[ vertpwrudang185 ] , the top panel shows the power in the 3 and 5-minute bands taken from the fourier analysis , with a vertical line marking the horizontal location , @xmath43 mm , where we are plotting the data in the three lower panels .
the upper middle panel shows the time series of the field - aligned velocity , while the lower middle panel shows the variation of the ( absolute ) field inclination means the field is vertical . ] , both at @xmath30 mm . since the appearance of the wave pattern does not just depend on conditions at this height , but is a function of the conditions throughout the lower atmosphere , we have also traced the field lines downwards at each timestep and plotted , in the same panel , the inclination 250 km ( dashed line ) and 500 km ( dotted line ) lower along the field . the bottom panel shows the wavelet power spectrum of the velocity signal at @xmath30 mm , calculated using the morlet wavelet ( see * ? ? ?
* ) , with a black line showing the cone of influence ; values below this line are subject to edge effects .
darker color corresponds to higher power . in these plots
we see how the period of the velocity signal increases significantly between about 1000 and 2600 s , which is the time when this region is located towards the outer edge of the central flux concentration and the inclination is much greater than earlier and later . at this time we get a large peak in the power spectrum at periods around 5 minutes , while at other times , most of the power is in 3-minute waves .
this again strongly supports the results in the earlier literature , showing that field inclination is important for long - period wave propagation .
so , do inclination increases always result in increased 5-minute power ?
yes and no .
the peaks in the 5-minute power tend to be associated with periods of increased inclination , but the area around @xmath43 mm is unique in having such a strong , relatively pure signal coherent over five wave periods ; hence the large peak in the power spectrum there . but even in other locations where the 5-minute power does not increase markedly , the wave dynamics are still quite different when the inclination is large .
an example can be seen in figure [ vertpwrudang360 ] , showing the velocity signal and inclination at @xmath50 mm . here , the inclination is large for the first 2000 s , and the velocity signal at that time is weaker and more irregular than later in the simulation , when the field is more vertical .
one explanation for this can be deduced from a comparison with figure [ vertbud ] : the times when the inclination is large are , in general , the times when the region is located at the interface between two flux concentrations . since the flux concentrations are the main sources of wave power in the system , such interface regions are also regions of interference between the waves coming from the two flux concentrations .
in many cases , this interference is destructive and reduces the amplitude of the disturbances . the periodicity is also perturbed since the signal is now a superposition of the waves coming from the two sources .
examples of regions with such destructive interference are at @xmath37 mm after 3000 s ( a case of triple interference , as there is a weaker , @xmath51 g flux concentration acting as a wave source below this location ) , and the plotted @xmath52 mm for the first 2000 s. in contrast , the regions of most dominant 5-minute power , at @xmath43 mm and @xmath49 mm , are located towards the sides of the region dominated by the strongest central flux tube , with the interference regions located on the outside of these .
it is also worth noting that the regions of strongest 3-minute dominance in the chromosphere are at @xmath53 mm and @xmath54 mm , both of which are directly above flux concentrations that show relatively little horizontal movement ( cf .
figure [ vertbud ] ) .
mm ( top panel ) and vertical velocity at @xmath30 mm ( bottom panel ) in case c , with superimposed contours showing the regions where @xmath21 at @xmath10 mm is greater than 150 g. the white dashed line shows the location of the periodic boundary ; the rightmost @xmath23 mm are repeated on the left.,title="fig : " ] + mm ( top panel ) and vertical velocity at @xmath30 mm ( bottom panel ) in case c , with superimposed contours showing the regions where @xmath21 at @xmath10 mm is greater than 150 g. the white dashed line shows the location of the periodic boundary ; the rightmost @xmath23 mm are repeated on the left.,title="fig : " ] while case b had several magnetic flux concentrations of comparable strength , case c contains only one dominant flux tube . unhindered by the influence of neighboring flux tubes , it can expand freely and fill the entire simulation box at coronal heights .
this expansion creates a large region of inclined field at the edges of the flux concentration , which was where we preferentially saw 5-minute propagation in case b. the initial state of this case is shown in the middle panel of figure [ initfig ] .
although the main flux concentration is strong , with a @xmath47 height reaching down to @xmath55 mm ( far below the photosphere ) at any given time , the magnetic field in the rest of this model is rather weak . as a result ,
the average height where @xmath47 is higher in this case than in case b ( see figure [ poreafourier ] ) ; between @xmath56 mm and @xmath48 mm , it is above our analysis height of 1 mm , and the field - aligned velocity is not a meaningful quantity there . because of this , we will be performing the analysis of this case at @xmath57 mm .
figure [ poreabuz ] ( top ) shows the field - aligned velocity at that height ( grayscale ) , with superimposed black contours showing where @xmath21 at @xmath10 mm is greater than 150 g. this value was chosen to point out the secondary flux concentration that merges with the primary at @xmath58 s ; the primary has a field strength on the order of 2000 g. since the primary flux concentration is located towards the left of the simulation box , we have plotted the rightmost 3.5 mm again on the left , with the white dashed line marking the location of the boundary . although there are some powerful shocks connected with the secondary flux concentration in the first 1200 s , the velocity field is dominated by the waves propagating outwards from the primary .
new oscillations are constantly being generated in the center of the flux concentration .
( figure [ propagation ] is an example of such wave propagation , taken from this case ; see also @xcite @xcite for a study of wave generation in a similar model . )
these propagate upwards and outwards , but there is large variability in how fast they move , and how far to the sides they reach . in general , the waves travel faster and longer in the direction that the flux concentration itself is moving at the time ; this is not unreasonable , as the waves from the lower layers will tend to align with the axis of the flux concentration , and this is itself inclined in the direction the flux concentration is moving .
we do occasionally see some wavefront merging as a result of these varying propagation patterns , and this could potentially affect the calculated periodicities ; however , most of the wavefronts `` absorbed '' this way are rather weak .
two things are obvious from the picture : although the flux concentration is only a few hundred km wide at photospheric heights , it dominates a very large area ( up to 4 mm to each side ) in the upper chromosphere , and the period of oscillations at its center is significantly shorter than that of the oscillations to the side . since the analysis in the other cases is performed at @xmath30 mm , we also show the velocity field at that height in this case ( figure [ poreabuz ] , bottom panel ) . as @xmath30
mm is partially below the average @xmath47 height , we here plot the vertical velocity rather than the velocity along the magnetic field . in this case , we see wavefronts that are coherent over several mm horizontal distance , especially in the first 2500 s. these sometimes become very strong shocks with amplitudes of more than 30 km s@xmath20 , and some of the strongest are connected with the secondary flux concentration .
the typical period of these oscillations is 200 - 250 s , though there is some variation .
meanwhile , in part superimposed on the larger `` global '' pattern , we see the wavefronts clearly propagating outwards from the primary flux concentration .
the area farthest from the flux concentration , around @xmath48 mm , shows very weak signal in the later part of the simulation . at this time , the magnetic field in the region is close to horizontal , and temporarily lower temperature makes the medium magnetically dominated .
this creates a `` lid '' which inhibits vertical wave propagation . as in the other cases ,
we have performed a fourier analysis , the results of which are shown in figure [ poreafourier ] . a pattern is clear : we see 3-minute propagation mainly above the center of the flux concentration , and to some extent in the very weak - field area around @xmath59 mm , and 5-minute propagation in the inclined field regions on the sides of the flux concentration .
although the ratio plot shows this even more clearly at greater heights , it is also plainly visible in the spectrum itself ; the central flux tube region between @xmath26 mm and @xmath60 mm has its whole spectrum shifted towards higher frequencies than in the regions surrounding it .
this is in very good agreement with what we found in case b. mm in case d , with superimposed contours showing the regions where @xmath21 at @xmath10 mm is greater than 300 g. the white dashed line shows the location of the periodic boundary ; the leftmost @xmath23 mm are repeated on the right . ]
we now turn our attention to case d , which has similar flux concentrations as case b at the photospheric level , but also includes an additional imposed uniform field of @xmath61 inclination ( slanting towards lower @xmath24 ) .
this uniform field is dominant in the corona and upper chromosphere , but is not generally strong enough to dominate dynamics at the photospheric or lower chromospheric levels .
the initial state of this model is shown in the second panel from the bottom of figure [ initfig ] .
this case is meant to represent a plage region with nearby field of opposite polarity .
figure [ 45degbud ] , like figure [ vertbud ] for case b , shows the field - aligned velocity at @xmath30 mm in case d as a function of @xmath24 and time , with overplotted black flux concentrations ( @xmath62 g ) at @xmath10 mm and the leftmost @xmath23 mm plotted again on the right due to the flux concentration on the boundary . as in case b
, we see that the regions with a powerful velocity signal are correlated with the regions of strong flux concentrations below .
in fact , the association seems even stronger in case d , with large areas between flux concentrations showing quite weak velocity signals ( e.g. between @xmath45 and @xmath37 mm in the later half of the simulation time ) .
other regions between flux concentrations , e.g. at @xmath48 mm and @xmath39 mm , become sites of less regular flows , especially early in the simulation .
the field inclination in these regions at those times is generally very large , even near - horizontal in the region around @xmath39 mm . the sudden outburst at @xmath63 mm between 3200 s and 3500 s is due to a reconnection event at @xmath64 mm .
in general , due to the left - leaning field , each flux concentration dominates the velocity in a large area to its left , but only a short distance to its right .
a fourier analysis of this case is shown in figure [ budpow ] .
as in case b , the 3-minute band is dominant in large parts of the chromosphere ( notably @xmath65 mm and @xmath66 mm ) , but there are some windows where the 5-minute band dominates .
the most prominent of these are around @xmath67 mm , @xmath48 mm , and @xmath39 mm .
the latter two also exhibit large amounts of low - frequency power , due to the non - periodic flows present there .
compared to case b ( figure [ auzpow ] ) , there is less of a connection between the strongest average photospheric fields and the peaks in the 3-minute power , even allowing for the slight leftward drift that would be expected due to the inclination of the additional homogeneous field .
this is likely primarily because the flux concentrations in this model display more horizontal movement than in case b ( cf .
figure [ budpow ] ) for example , the strong flux tube that starts out at @xmath38 mm moves first left to @xmath68 mm , and later right to @xmath69 mm .
meanwhile , there is a separate flux concentration at @xmath70 mm that eventually merges with the other one after 3000 s. the area connected to these is in general the area of greatest velocity power , in particular in the 3-minute band .
it is also the region of lowest average field inclination .
the flux concentration at the boundary , @xmath46 mm , moves comparatively little and this region does correspond to a peak in the 3-minute power .
overall , though , comparison with the average fields and inclinations is of less value in case d due to the horizontal movement . in order to figure out what is going on , it is necessary to take time variations into account .
the wavelet analysis is a tool better suited for this analysis , and we will again look at the time variation of the velocity and inclination at some specific horizontal positions .
figure [ bampang024 ] is a four - panel figure showing ( top ) the 3-minute ( solid ) and 5-minute ( dashed ) power , with a vertical line showing the current horizontal location ( @xmath71 mm ) ; ( upper middle ) the field - aligned velocity as a function of time at that location ; ( lower middle ) the field inclination as a function of time at that location ; and ( bottom ) the wavelet power spectrum .
all panels are plotted at the height @xmath30 mm ; as previously , the field inclination is also shown at two lower heights along the field .
the region around @xmath71 mm has the highest 5-minute power in the simulation , and it appears to be mainly due to the behavior in the first 1700 s. at that time , there is a quite regular train of long - period shock waves coming up . between 1700
s and 3000 s , the velocity signal is much more irregular , correlated with large and rapid variations in the inclination at lower heights , although the dominant periodicity remains 5 minutes .
a comparison with figure [ 45degbud ] shows that the change in behavior happens when the region becomes an interference region between the flux concentration at the boundary and the one starting out at @xmath26 mm after roughly 1700 s ; before that , it is dominated by waves coming from the latter flux concentration .
after about 3300 s , the flux concentration at the boundary has moved to a position just below our observation point ; the inclination becomes smaller and we get a wave train with shorter periodicity .
the general behavior is similar to what we saw in case b ; much of the difference seems to be that with an imposed , left - leaning @xmath61 field , each flux concentration tends to dominate the wave field in a large region to its left , but only in a small region to its right .
figure [ bampang338 ] shows the situation at @xmath72 mm , one of the largest peaks of 3-minute power . here , the inclination remains mostly below @xmath73 at @xmath30 mm and @xmath74 at lower heights , and we get a fairly regular short - period velocity signal , though with varying amplitude . in the last 1000 s , the inclination increases , and this does lead to a more irregular signal , with longer time between the peaks , though the effect seems noticeable only during the relatively short time when the inclination is above @xmath74 . finally , figure [ bampang456 ] shows the situation at @xmath75 mm , at the center of a wide region where the 3-minute power is low and the 5-minute power is dominant .
this is a region where the field inclination is generally very large during most of the simulation , it is more than @xmath76
. a few initial shocks , which actually dominate the power spectrum , are followed by a long period with near horizontal field and a steadily increasing downflow , culminating at 1800 s. then , after 2200 s , a rather irregular , mostly long - period signal sets in .
a comparison with figure [ budpow ] shows that this signal comes from the flux tube on the boundary to the right .
a new flux concentration appears ( at @xmath10 mm ) after around 2000 s , but it connects only to the left and does not dominate the velocity field directly above it .
the most notable effect at this location is the near complete absence of 3-minute signal due to the large field inclination . in sum
, this case shows a behavior that is rather similar to case b , which has a more vertical field .
we see 3-minute signal above the strongest flux concentrations , where the vertical component of the field is usually strong enough to keep the inclination low , and 5-minute signal towards the edges of the regions dominated by each flux concentration . among the differences between case b and case
d , we find that case d has less power at high frequencies , connected with interference regions in case b , and that the 3-minute signal is very weak in regions of heavily inclined field , which were absent in case b. the following case looks into what happens when the strength of the constant inclined field is increased , allowing it to dominate at lower heights .
mm in case e , with superimposed contours outlining the regions where @xmath21 at @xmath10 mm is greater than 1000 g. the white dashed line shows the location of the periodic boundary ; the leftmost @xmath23 mm are repeated on the right . ] in case e , the imposed field is still at a 45@xmath77 angle with respect to the vertical , but the field is significantly stronger than in case d. in the mid - chromosphere , around @xmath30 mm , it is about 8 times stronger ; lower down it is more variable . the field - aligned velocity at @xmath30 mm is plotted in figure [ caseebud ] , with overplotted contours outlining the regions where @xmath21 at @xmath10 mm is greater than 1000 g ( as compared to the 300 g used in the similar plot for case d , figure [ 45degbud ] ) . the initial state of the model is shown in the bottom panel of figure [ initfig ] .
this model represents conditions in very strong plage with nearby field of opposite polarity . as in the other cases , the regions of highest velocity amplitude seem to be located directly above the strongest flux concentrations
this may seem counterintuitive at first with such a strong field acting as a wave guide , we might expect the velocity signal to appear to the side of the locations of the flux concentrations .
but we should keep in mind that the flux concentrations themselves are fairly wide .
many have widths on the order of 1 mm , or the same as the height we are observing at , so even though the sideways propagation starts around @xmath78 mm ( see figure [ caseefourier ] ) , the waves coming from a flux concentration will still tend to be located above it .
as we move higher in the atmosphere , the velocity signal does move sideways along the magnetic field , but this effect is not yet very pronounced at @xmath30 mm
. between the flux concentrations , there are channels where almost no wave signal can be seen .
these neatly split the computational box into zones of influence for each flux concentration .
the channels are actually downflow regions , with slowly varying velocities of @xmath79 km s@xmath20 .
a single larger downflow event can be seen between @xmath67 and @xmath80 mm after 4000 s. figure [ caseefourier ] shows the fourier analysis of this case . at this field strength and inclination ,
the 5-minute band is dominant across most of the chromosphere ( top panel ) . only in two locations , around @xmath81 mm and @xmath38 mm ,
does the 3-minute band dominate .
the 3-minute power is highest above high-@xmath8 regions , i.e. where the magnetic field is weak .
this is the opposite of what was found in case b ( figure [ auzpow ] ) , where the strongest flux concentrations were the sites of highest 3-minute power . in that case , however , the strongest fields were also the most vertical , whereas here , the weaker fields are more vertical than the stronger fields .
the results therefore support the idea that the field inclination is central to wave propagation .
the full fourier spectrum at @xmath30 mm ( bottom panel ) shows the same situation : significant power in the 5-minute band across most locations , though the 3-minute peak at @xmath81 mm is the strongest in the simulation .
the 5-minute power is more evenly distributed , with one broad peak at @xmath82 mm and some smaller , also quite broad peaks .
we also see the very low - frequency power connected with the downflow regions .
although the strong - field case e has a more steady velocity field than the cases with weaker fields , there is still some drift and variation with time .
we once again perform a wavelet analysis in order to look more closely at the wave propagation .
figure [ ewvl073 ] shows the results at @xmath83 mm , the largest peak of 3-minute power .
the velocity amplitude ( upper middle panel ) is large in the initial stages , when the inclination is relatively low ( less than or around 30@xmath77 ) .
there is power at a rather wide range of frequencies at that time , though by far the strongest signal is found in the 3-minute band between 4 and 6 mhz .
after 2000 s , most of the signal dies out , as this vertical location becomes part of a rather quiescent downflow region with velocities of 4 - 5 km s@xmath20 .
after 4000 s , a single larger downflow event occurs , which shows up as power at very low frequencies ( below the line marking the cone of influence ) .
the largest peak of the 5-minute power is centered at @xmath84 mm , shown in figure [ ewvl334 ] .
this is directly above a flux concentration ( figure [ caseebud ] ) and there is some periodic velocity signal throughout the simulation , though the amplitude varies with time .
again , it is strongest early on , though the peak is mostly within the region subject to edge effects .
although the inclination is below 40@xmath77 in the beginning of the simulation , little coherent signal is visible in the 3-minute band .
the 5-minute band shows power throughout , with a second peak appearing towards the end ( again partially within the cone of influence ) .
the field inclination slowly increases with time and ends up at 53@xmath77 .
the main result from our simulations is that 5-minute waves are able to propagate through the chromosphere in regions where the magnetic field is inclined and sufficiently strong . in regions where the field is vertical or weak ,
the velocity field is dominated by waves with periods of around 3 minutes . as mentioned in the introduction , a model which uses the radiative relaxation time , rather than the inclination of the magnetic field , as the mechanism for increasing the cutoff period , has been proposed , and
has been demonstrated to work in numerical simulations @xcite . however , it relies on a simple newtonian cooling model for approximating radiative losses .
our simulations use an advanced and realistic method for computing the radiative losses @xcite that includes all the important mechanisms involved in radiative losses in the photosphere and chromosphere , and as such it is considerably more realistic than a simple newtonian cooling model .
while the radiative relaxation time model expects 5-minute propagation above all strong small - scale magnetic structures , regardless of the field inclination , our simulations show 3-minute propagation above the central region of flux concentrations , and 5-minute propagation in inclined - field regions to the sides .
this is exactly the result predicted by the field inclination model , but is in conflict with the predictions of the radiative relaxation model .
our conclusion , based on the simulations presented here , is therefore that the radiative relaxation model is not effective when a more realistic energy equation is considered .
while large - scale regions of homogeneous inclined field , like our cases d and e , represent idealizations of the conditions at the edge of plages , our case c represents realistic conditions for an isolated magnetic element or pore . in this model ,
the magnetic field is largely vertical in the photosphere and above the transition region , while the field expansion leads to significant inclination where it is needed the most , in the chromosphere .
this can then account for the 5-minute propagation , but we expect 3-minute propagation in the center of the flux concentration . is this supported by observations ?
the answer is in fact not very clear . @xcite
show an example of a network region which has very little power at frequencies above 4 mhz over most of its area .
there is , however , a pronounced spike of higher - frequency power located in the center of the network ( @xmath85 mm in their figure 4 ) .
the authors attribute this to noise from seeing , which it may very well be ; it is visible at frequencies up to 20 mhz , and a study of the coherence spectra also lends credence to this explanation .
however , the signal at frequencies of 5 - 6 mhz in this spike is much stronger than that at higher frequencies , and indeed is of the same magnitude as the ( real ) signal at those frequencies in the neighboring quiet regions .
it is notable that such a spike , with a spatial extent of 1 - 2 mm , is in fact exactly what we find in our case c. the central spike observed by @xcite could thus be real . in order to make more direct comparisons to their results , which were based on doppler shifts in observations of the ca ii h line at @xmath86 and the fe i line at @xmath87 ,
we have computed synthetic spectra in these lines from our simulations using the non - lte radiative transfer code multi @xcite .
the ca line is treated in full non - lte , while the fe line , which is formed in the wing of the ca line , is included in the same computation using additional localized opacity and source function terms based on lte .
the calculations have been performed column by column , i.e. , neglecting any radiative interaction in the @xmath24-direction .
the computed spectra have been smoothed over @xmath88 , which appears to be close to the effective resolution of the observations in @xcite , and we have then calculated doppler shifts from these smoothed spectra . it should be noted that ca ii h has a very complex line profile , usually with several emission peaks within the deeper absorption line . in a dynamic atmosphere , particularly in the presence of strong shocks
, these emission peaks can become very large and are often asymmetric ( so called bright grains ) , and the line can undergo central reversal as well . defining and identifying
the line center of such a line is a non - trivial task .
we have used a method that finds the center of the region where the intensity is below a certain threshold above the minimum intensity .
this method generally gives acceptable results , but can not be expected to correspond directly to the velocity at any one given height in the simulations .
the power spectra of the calculated doppler velocities are shown in the two right - hand panels of figure [ litesrad ] ; the two left - hand panels show the power spectra of the vertical velocity taken directly from our simulation data , at heights corresponding to the approximate formation heights of the lines .
since the observations of @xcite covered a larger area than our simulation boxes , we show a combination of the spectra from two of our simulations in the figure . in the center of all panels , between @xmath46 mm and @xmath89 mm ,
we show the spectrum of our case b , representing network conditions . on both sides , from @xmath90 mm to @xmath46 mm and from @xmath89 mm to @xmath91 mm , we show the spectrum of our case a , representing conditions in the weakly magnetized internetwork .
this figure should be compared with figure 4 in @xcite .
while our synthesized ca power spectrum ( upper right ) is not a perfect match to the observations of @xcite , there are many similarities .
of particular note is the difference in the dominant frequencies between the non - magnetic regions on both sides and the network region in the center .
the network ( case b ) is dominated by lower frequencies , particularly in the area between @xmath92 mm and @xmath93 mm , where the dominant frequencies are 3-@xmath23 mhz .
the internetwork ( case a ) has a more scattered spectrum , but the dominant frequencies are mostly between 5 and 7 mhz .
furthermore , the network has significant power at very low frequencies , below 3 mhz .
these effects are also found by @xcite .
this power at low frequencies is not found in the simulation velocity at @xmath30 mm ( upper left ) , and although the velocity spectrum and the ca spectrum are similar in many ways , there are also several differences .
these differences are partly due to the smoothing applied to the ca data , partly due to the difficulty of defining a meaningful doppler shift of the highly complex ca line profile , and partly due to the fact that the line is formed over a range of heights rather than at one given height , and this height range can also vary with both horizontal position and time . using the ca doppler velocity as a proxy for atmospheric velocity on the real sun can therefore be misleading .
the velocity spectrum of the fe line ( lower right ) shows a very good correspondence with the velocity at @xmath94 mm in the models ( lower left ) .
this line has a much simpler profile and is formed in a region with few strong shocks affecting local conditions .
the dominant frequency is 3 mhz in both the network and the internetwork , though there is also some power at 5 mhz in most locations .
@xcite also find that the power in this line is in the 3 - 5 mhz band , although most of it is between 3 and 4 mhz . like us
, they find no significant difference between the spectra in the internetwork and in the network in this line .
more recent observations related to the question of wave periodicity have been performed by @xcite , @xcite , and @xcite ( see also * ? ? ?
@xcite used the tenerife infrared polarimeter of the german vacuum tower telescope at the observatorio del teide , with a seeing - limited spatial resolution of @xmath95 .
they then found propagating 5-minute waves in the chromosphere above a facular region .
the photospheric magnetic field as determined from stokes inversions was within 20@xmath77 of the vertical .
@xcite , using the solar optical telescope on _ hinode _ with a resolution of @xmath96 , found 3-minute signal in the center of a plage region , but more 5-minute propagation towards the sides in the direction of the expanding field .
@xcite used a combination of data from _ hinode _ and the ibis instrument at the dunn solar telescope , with an estimated average resolution of @xmath97 .
they found propagating 5-minute waves along the inclined field on the edges of a pore , and some power in 3-minute oscillations at the center .
they also found both 5-minute and 3-minute propagation , though with more power in the 5-minute band , in a nearby region with small magnetic elements where they estimate that the chromospheric magnetic field is close to vertical , based on a force - free field extrapolation .
the results of our simulations are in agreement with @xcite , and with the propagation patterns observed by @xcite around their pore . in the more vertical magnetic structures observed by @xcite and @xcite
, we would expect more power in the 3-minute band than in the 5-minute band based on the results of our simulations , if indeed the field is mainly vertical and the flux tubes do not move around very much .
there are , however , several possible mechanisms that could explain the 5-minute dominance and resolve this apparent difference .
for one , all flux tubes naturally expand with height , and this expansion creates a region of inclined field ( as illustrated , on a large scale , by our case c ) .
thus , even if the field is close to vertical in the photosphere , there will be regions between the photosphere and the chromosphere where the field at the edges of the flux tubes is inclined , and the long - period waves can propagate there .
@xcite do study coherence spectra to look for signs of a possible horizontal shift in the signal as a result of propagation along inclined fields , and find good coherence between the photospheric and chromospheric signal , but our results show that the field does not need to be inclined throughout the photosphere and chromosphere in order to enable 5-minute wave propagation .
a few hundred km along the edges of an expanding flux tube may be enough , and any horizontal shift could then be less than one resolution element ( @xmath95 in their data ) . in the observations of @xcite ,
the photospheric field is not uniformly vertical , and a force - free extrapolation is not a very good approximation in the chromosphere .
furthermore , although there is more power in the 5-minute signal , they also find significant signal at periods around 3 minutes in the region with smaller magnetic elements .
we believe that our model , where field inclination is the dominant mechanism for allowing long - period wave propagation , is compatible with these findings . flux tube movement and limited resolution may also be partially responsible for the relative dominance of 5-minute power in the observations of @xcite . in case b
, we found that the strongest 3-minute power appeared above flux tubes that undergo little horizontal motion , while 5-minute power was found in inclined field regions at the edges of flux tubes .
if the flux tubes move around , both the 5-minute and 3-minute power will be spread out and one would not see a clear distinction between the ( average ) flux tube center and the sides in a fourier analysis .
the flux tube center also covers a rather small area at any given time , and this makes the related 3-minute waves difficult to observe in low - resolution data . in the higher - resolution data of @xcite , regions of 3-minute propagation are found , and this could possibly be because the flux tubes at these locations move around less
. we would encourage observers to look for differences in the periodicity of oscillations at the center and edges of flux tubes in future high - resolution datasets .
a third possibility is that heating may play a role .
the temperature structure of the chromosphere is in general not well known . in particular , the real sun may have more magnetic heating of the upper chromosphere than our 2d models
. higher temperature would reduce the cutoff frequency and allow 5-minute waves to propagate more easily .
yet another suggestion , as mentioned by @xcite , is that the field may be twisted .
the waves could then travel along field lines that are everywhere inclined with respect to the local vertical , but without significant horizontal displacement .
such field twist is a 3d effect and can not be tested in our 2d simulations , but should be considered in later work . a different point , that we have already mentioned in our analysis , is also worth making : although a fourier analysis can be a powerful tool , it has some important limitations . in order to achieve sufficient spectral resolution ,
one usually needs time series on the order of one hour . the solar atmosphere , however , is dynamic on timescales of minutes .
atmospheric conditions can and do change , and the fourier transform is not well suited for picking up such changes . at least in areas where the general signal is weak , non - recurring events can end up dominating the power spectrum ( see figure [ bampang456 ] for an example ; there are several other examples in the dataset ) .
in such cases , the fourier analysis does not say all that much about the general conditions at that location .
this can be particularly dangerous because the eye is naturally drawn to peaks in the power spectrum .
also , there can be times when local conditions are notably different from the time average , and these can be correlated with changes in the signal ( e.g. , figure [ vertpwrudang185 ] ) . a wavelet analysis , which takes time variations into account , or at least a careful examination of the time series , is essential for identifying the actual processes going on in the atmosphere .
so far we have been looking at the propagation of waves through the chromosphere .
but one of the reasons why we are interested in this propagation is the effect these waves have once they reach the transition region .
jets of chromospheric material go by many names , such as spicules , mottles , fibrils , straws , macrospicules , or surges .
there is no widespread agreement on whether all or some of these represent different aspects of the same underlying phenomenon , or on the driving mechanism or mechanisms behind them .
however , there is mounting evidence that at least dynamic fibrils and some mottles appear to be driven by waves coming from below @xcite .
such wave - driven jets also appear frequently in our simulations here , and we will look into their properties and make a statistical comparison between jets in case b ( mostly vertical field ) and case d ( field inclined 45@xmath77 ) . a large number of jets are formed in the simulations , and we have used semi - automatic routines to identify and measure them . we identify the jets as local maxima ( in both space and time ) of the height where the temperature is 40 000 k , corresponding to the lower transition region .
we have included snapshots of the tallest jets in the two simulations in figures [ vertspic9 ] ( case b ) and [ 45degspic9 ] ( case d ) .
we see that they can be quite tall , reaching heights of up to 6 mm , and that their axes follow the magnetic field , as we would expect . in order to analyze the properties of the jets , we want to look at their time evolution . a rising and falling jet will generally have a parabolic shape in a plot of distance vs. time ( e.g. , * ? ? ?
* ) , but for the most accurate results , the measurements should be taken close to the central axis of the jet .
the jet peaks found previously are used as starting points . in case d , where the magnetic field is inclined ,
we track the magnetic field line that passes through the jet peak , and measure the variables along this ( moving ) line through the jet s lifetime . in case
b , the magnetic field and the jet axes are generally close to vertical , and we use a simplified procedure taking vertical cuts through the locations of the jet peaks .
experiments using the more involved field - tracking procedure in case b show that the inaccuracies resulting from the simplified treatment are small . using a method similar to the one of @xcite and @xcite , working on observations , @xcite , working on 1d simulations , and @xcite , working on 3d simulations , we produce distance - time diagrams of the temperature and fit the position of the transition region with parabolas .
these fits are then used to calculate properties such as the maximum velocity , deceleration , maximum length , and duration of each jet .
jets that do not have a parabolic shape or that represent only minor disturbances are discarded from the analysis .
note that , as jets sometimes occur in quick succession in the same locations , occasionally they can overlap .
this can lead to situations where the descending phase of one jet is cut short by the next jet appearing .
this will mainly affect the calculated lifetime of the jet , though it can to a lesser extent also impact the maximum measured length and velocity .
the calculated deceleration depends only on the overall shape of the parabola and is largely unaffected .
overall , this does not happen frequently enough to significantly change the statistics , though it can lead to some outliers , particularly with respect to the durations . in total
, we have identified 192 parabolic jets in case b , and 129 in case d. their properties and various correlations are shown in figure [ spiccorr ] , which is in the same format as figure 1 in @xcite for ease of comparison .
they should also be compared with figures 12 and 13 in @xcite .
overall , the results of the different simulations and the observations are in good agreement . as we see , there are clear linear correlations between several of the jet properties , in particular between maximum velocity and maximum length ( bottom right ) , deceleration and maximum velocity ( bottom left ) , and maximum length and duration ( center right ) .
the same correlations are found in observations by @xcite , and in simulations by @xcite , who only plot deceleration vs. maximum velocity , and @xcite , who find a weaker correlation between length and duration .
in addition , we find a correlation between the maximum velocity and the duration ( top right ) in our simulations , whereas @xcite and @xcite find little evidence of a correlation here .
conversely , they both find a weak anticorrelation between deceleration and duration ( top left ) . in our simulations
, there is a quite weak anticorrelation in the distribution for case d ( blue diamonds ) , whereas case b ( red asterisks ) shows little evidence of a correlation .
the two distributions do however cluster in different regions , with the longer - duration jets in case d having lower decelerations than the shorter - lived jets in case b. this regional difference between the distributions was also found in the observations . neither we , @xcite , nor @xcite find any strong correlation between the deceleration and the maximum length of the jets ( center left ) .
as noted above , the properties of the jets formed in case b and case d can be somewhat different for example , at a given maximum velocity , the jets in case d ( inclined field ) tend to be longer ( bottom right panel ) and have a longer duration ( top right panel ) . in order to further investigate these differences ,
we have made histograms showing the distributions of the jet properties in the two cases .
the results are plotted in figure [ spichist ] , which shows the jets in case b as solid red columns , case d as hatched blue columns , and the sum as black outlines .
the distributions are notably different with respect to deceleration ( upper left panel ) and duration ( lower right ) , with the inclined jets lasting longer and having lower decelerations .
the distributions of maximum lengths ( upper right ) are broadly similar , but the inclined jets have somewhat lower maximum velocities ( lower left ) .
@xcite found similar regional differences in their observations , with jets in dense plage regions ( referred to as region 2 in their paper ) having higher decelerations and shorter lifetimes than jets in adjacent regions with weaker and more inclined field ( referred to as region 1 ) .
they suggested that the differences were due to the field inclination , which is supported by our results .
the results are also consistent with what we have found out about wave propagation in the different models : long - period waves are more prevalent in inclined field regions , and these waves then produce longer - lived jets with lower decelerations .
in addition to differences between the cases , there are also some regional differences within each case .
we see in figure [ vertspic9 ] that a number of the tallest spicules in case b are located in regions where field lines from different flux concentrations meet ( ref .
figure [ vertbud ] ) .
this is particularly noticeable in the cases of jets @xmath98 and @xmath99 , where the central axis is right along the interface region .
jets @xmath100 ,
@xmath101 , @xmath102 , and @xmath103 are also located close to the interface , while jets @xmath104 and @xmath105 are above the center of their connected flux concentrations . in figure
[ vertjetzvsx ] , we have plotted the maximum heights reached by the jets as a function of their horizontal location .
notably , the jets that form above the central , wide flux concentration ( @xmath106 mm ) reach lower heights than the ones forming in other locations , while the tallest jets form in interface regions between flux concentrations at @xmath107 mm and @xmath108 mm .
there are at least two possible explanations why this should be so .
one is that you can get constructive interference between the waves coming from different flux concentrations , leading to stronger shocks and longer jets .
for example , jet @xmath98 in figure [ vertspic9 ] can be traced to such an event happening at @xmath56 mm , @xmath109 s ( cf .
figure [ vertbud ] , which shows the velocity field ) .
another possible explanation is that when several flux concentrations ( of the same polarity ) are located near each other , the field lines connected with each can not spread out over such a large area , and the waves coming from below will be more concentrated and propagate more vertically .
conversely , above the large central flux concentration , the wave energy is spread over a much larger area as much of it follows the widening field .
this then leads to wider but less powerful jets .
we have presented several simulations investigating the periodicity of waves propagating through the chromosphere .
we find that waves with periods of around 5 minutes can propagate in regions where the magnetic field is strong and inclined , including at the edges of flux tubes . in regions where the magnetic field is weak or vertical ,
we find primarily 3-minute waves ; this also applies above vertical flux tubes and in the center of strong expanding flux tubes .
these results indicate that field inclination is critical to the propagation of long - period waves . where the flux tubes undergo significant horizontal motion ,
both the 5-minute and the 3-minute power is spread out and the distinction is not as clearly visible .
since we have included an advanced treatment of radiative losses in our simulations and find 3-minute propagation above vertical field regions , we conclude that variation in the radiative relaxation time is not an effective mechanism for increasing the cutoff period .
our simulations are in agreement with the results of recent high - resolution _ hinode _ observations .
we have also studied the jets produced by these waves once they reach the transition region .
we find systematic differences between the jets produced in a model with mostly vertical field , and in a model with mostly inclined field .
the results are in agreement with observations of dynamic fibrils .
it is also important to point out , for the purpose of future analyses of wave propagation and periodicity , whether from simulations or observations , that the fourier analysis can be misleading and hide important information about the state of the medium .
this is because the solar atmosphere is dynamic and changes on timescales much shorter than one hour , which is often the minimum timescale needed to achieve sufficient spectral resolution in a fourier analysis .
a wavelet analysis takes variations in time into account and can be an invaluable tool for figuring out what processes are important for the dynamics . | we present the results of numerical simulations of wave propagation and jet formation in solar atmosphere models with different magnetic field configurations .
the presence in the chromosphere of waves with periods longer than the acoustic cutoff period has been ascribed to either strong inclined magnetic fields , or changes in the radiative relaxation time .
our simulations include a sophisticated treatment of radiative losses , as well as fields with different strengths and inclinations . using fourier and wavelet analysis techniques
, we investigate the periodicity of the waves that travel through the chromosphere .
we find that the velocity signal is dominated by waves with periods around 5 minutes in regions of strong , inclined field , including at the edges of strong flux tubes where the field expands , whereas 3-minute waves dominate in regions of weak or vertically oriented fields .
our results show that the field inclination is very important for long - period wave propagation , whereas variations in the radiative relaxation time have little effect .
furthermore , we find that atmospheric conditions can vary significantly on timescales of a few minutes , meaning that a fourier analysis of wave propagation can be misleading .
wavelet techniques take variations with time into account and are more suitable analysis tools .
finally , we investigate the properties of jets formed by the propagating waves once they reach the transition region , and find systematic differences between the jets in inclined field regions and those in vertical field regions , in agreement with observations of dynamic fibrils . |
solar 5-min . oscillations are excited by turbulent convection ( downdrafts ) in subsurface layers of the sun .
these oscillations consist of acoustic and surface gravity waves with frequencies in the range of 2@xmath08 mhz and in a wide range of wave numbers .
the observed oscillations can be used for reconstruction of internal structure of the sun by methods of helioseismology .
there are several methods of investigation of the interaction of traveling acoustic waves with small perturbations to the background state .
one of them is the time - distance approach @xcite .
the key concept of this method is measuring and inversion of wave travel times .
propagation of acoustic waves in this approach is calculated using various approximations to the wave equations , such as the ray theory or first born approximation @xcite .
these approximations have been tested using simple models for point sources ( e.g. @xcite ) , but not for realistic solar conditions , e.g. realistic stratification and random excitation sources .
such tests , which require direct numerical simulations , are extremely important for validating the inferences from time - distance helioseismology and other local helioseismology methods .
there are two main directions in numerical simulations of solar oscillations and waves .
the first one is to use realistic non - linear simulations of solar convection . in such simulations ,
the waves are naturally excited by convective motions .
these simulations reproduce quite well the solar oscillation spectrum @xcite , and have been used for testing time - distance helioseismology @xcite .
the second approach is based on linearized euler equations describing wave propagation for a given background state .
the background state can be taken from non - linear numerical simulations or by perturbing the standard solar model . in this paper
, we describe a numerical method and initial simulation results developed for the second approach .
we developed a 3d code which utilizes a realistic physics and accurately simulates reflection of acoustic waves from the top boundary .
a realistic equation of state which takes into account partial ionizations and various corrections is used .
the equation of state was calculated by interpolation of the opal tables . in section 2.1
we give a detailed description of the underlying physics .
the main attention is paid to developing a consistent procedure for obtaining a convectively stable background model and establishing realistic top boundary conditions based on the perfectly match layer ( pml ) method . in section 2.2
we describe a semi - discrete numerical scheme of high order which preserves dispersion relations of the continuous problem .
the main attention is paid to developing stable high - order numerical boundary conditions consistent with the finite - difference scheme , in which the dispersion relation is preserved for the inner mesh points of the computational domain . in section 3 ,
we compare the numerical and analytical solutions of various 1d test problems with an isothermal background model to validate the code , investigate accuracy of the numerical scheme , and the non - reflecting boundary conditions in a gravitationally stratified medium . in section 4 , we present results of numerical three dimensional simulations of acoustic wave field for a standard solar background model .
we used different types of single or multiple acoustic sources : z - component of force or pressure , point or distributed with different time dependence .
the main goals are to study properties of solar waves for various excitation sources and generate artificial wave fields for testing accuracy of the born and ray approximations and local helioseismic diagnostics of the solar interior , currently used for soho / mdi and gong data .
the results of this testing will be presenting in future papers .
the numerical simulations are carried out on parallel supercomputers at nasa ames research center .
the propagation of adiabatic acoustic waves below the solar photosphere is described by the following system of linearized euler equations : @xmath1 where @xmath2 are the perturbations of @xmath3 velocity components , @xmath4 and @xmath5 are the density and pressure perturbations correspondingly .
quantities with subscript 0 such as the pressure @xmath6 , the density @xmath7 , and the gravitational acceleration @xmath8 correspond to the background reference model and depend only on radius .
christensen - dalsgaard s standard solar model s @xcite with smoothly joined chromosphere provided by @xcite was chosen as such a model . to close the system we use an adiabatic relation between eulerian variations of pressure @xmath5 and density @xmath4 @xmath9 where @xmath10 is the square of sound speed
, @xmath11 is the adiabatic exponent , @xmath12 is the brunt - visl frequency , @xmath13 is the vertical displacement .
so , from the background model we need only profiles of the sound speed and brunt - visl frequency .
the background state is convectively unstable , especially just below the photosphere , where the temperature gradient is super - adiabatic and convective motions are very intense and turbulent .
this leads to the instability of the solution of linear system ( [ lineuler ] ) .
convection instability is developed on a time scale of 30 - 40 minutes of solar time .
simulation of the acoustic spectrum of solar oscillations needs to perform calculations on a time interval of 5 - 8 hours of solar time , and this instability will badly distort the result . to make the background model stable against convection we slightly modified profiles of pressure and density in a thin ( 500 km in depth ) super - adiabatic layer just below the photosphere .
the condition for stability against convection requires that the square of brunt - visl frequency has to be positive @xmath14 where @xmath15 is the distance from the center of the sun .
the profile of @xmath16 near the solar surface is shown in the bottom left pane of figure [ bkgmodel_fig ] by the solid line .
if we replace negative values by zero ( or small positive ) value we guarantee a stability of the modified model against convection .
now we can recalculate the pressure and density from the modified profile of @xmath17 .
combining equation ( [ bruntvaisala ] ) with the condition of hydrostatic equilibrium , we get the following boundary value problem for @xmath18 and @xmath19 : @xmath20 \displaystyle \frac{{\mbox{d}}p}{{\mbox{d}}z } = -\rho g,\\[7pt ] 0\leq z\leq l,\quad \rho(0)=\rho_0(0),\quad\rho(l)=\rho_0(l),\quad p(l)=p_0(l ) , \end{array}\ ] ] where @xmath21 is the depth of the computational domain , @xmath22 is the vertical coordinate which is counted off from the bottom of the domain .
we introduced a free parameter @xmath23 that has to be determined , to make the modified model closer to the original one .
we established boundary conditions for the system ( [ modifiedmodel ] ) setting the density at the top and bottom boundaries equal to solar values .
parameter @xmath23 does not change the condition of convective stability if it remains positive .
introducing of an additional parameter permits us to establish yet another boundary condition and fix the pressure at the top boundary .
so the recipe to build a convectively stable background model close to the standard one is the following .
we smoothly join the density profiles of the standard solar model and the chromosphere , obtain the pressure profile from the condition of hydrostatic equilibrium , calculate square of the modified brunt - visl frequency profile @xmath24 , replacing negative values by a zero or small positive value , and substitute it into the right hand side of ( [ modifiedmodel ] ) .
parameter @xmath23 , profiles of density and pressure of the modified convectively stable model are obtained as a solution of the eigenvalue problem ( [ modifiedmodel ] ) .
the profile of @xmath25 , needed for the sound speed profile , is found from the realistic opal equation of state @xcite .
vertical profiles of pressure @xmath6 , density @xmath7 , sound speed @xmath26 , square of brunt - visl frequency @xmath16 , adiabatic exponent @xmath25 , and acoustic cut - off frequency @xmath27 for both models ( standard with joined chromosphere and modified convectively stable ) are shown in figure [ bkgmodel_fig ] .
the depth of the domain from the photosphere equals 30 mm , the height of the chromosphere equals 2 mm , @xmath28 .
the dashed curves represent profiles in the convectively stable model , the solid lines show profiles of the standard solar model .
the thin vertical line marks the position of the fitting point of the chromosphere and the standard solar model .
the background model , obtained with the procedure described above , is convectively stable and self - consistent .
the system ( [ lineuler ] ) is written in a conservative form @xmath29 where @xmath30 is the vector of independent variables , @xmath31 is the source term containing the gravity term and acoustic sources which depend on time explicitly .
the source term does not contain spatial derivatives .
explicit expressions for components of vectors @xmath32 , @xmath33 , and @xmath34 can be easily found from the system ( [ lineuler ] ) .
we used a semi - discrete numerical scheme . in semi - discrete
approach the space and time discretization processes are separated .
first the spatial discretization is performed , leaving the problem continuous in time .
the spatial derivatives have been approximated by the finite difference method reducing the system of partial differential equations to the system of ordinary differential equations @xmath35
\displaystyle l_{ikj}({\mbox{\boldmath$q$}},t)&=&\displaystyle -\frac{1}{\delta x}\sum_{l=-3}^3a_l{\mbox{\boldmath$f$}}_{i , k , j+l } -\frac{1}{\delta y}\sum_{l=-3}^3a_l{\mbox{\boldmath$g$}}_{i , k+l , j}\\[7pt ] & -&\displaystyle\frac{1}{\delta z}\sum_{l=-3}^3a_l{\mbox{\boldmath$h$}}_{i+l , k , j}+ { \mbox{\boldmath$s$}}_{ikj}(t ) \end{array}\ ] ] which can be solved by any stable time advancing method .
we used four stage , 3rd order strong - stability - preserving runge - kutta method @xcite with courant number @xmath36 : @xmath37 \displaystyle{\mbox{\boldmath$q$}}^{(2)}={\mbox{\boldmath$q$}}^{(1)}+\frac{1}{2}\delta t\ , l({\mbox{\boldmath$q$}}^{(1)},t^n+\frac{1}{2}\delta t),\\[7pt ] \displaystyle{\mbox{\boldmath$q$}}^{(3)}=\frac{2}{3}{\mbox{\boldmath$q$}}^{(n ) } + \frac{1}{3}{\mbox{\boldmath$q$}}^{(2)}+\frac{1}{6}\delta t\ , l({\mbox{\boldmath$q$}}^{(2)},t^n+\delta t),\\[7pt ] \displaystyle{\mbox{\boldmath$q$}}^{(n+1)}={\mbox{\boldmath$q$}}^{(3)}+\frac{1}{2}\delta t\ , l({\mbox{\boldmath$q$}}^{(3)},t^n+\frac{1}{2}\delta t ) .
\end{array}\ ] ] the source term @xmath31 which depends on time explicitly does not require a special treatment .
high - order dispersion - relation - preserving ( drp ) scheme developed by @xcite was used for spatial discretization .
coefficients @xmath38 of the finite difference scheme @xmath39 are chosen from the requirement that the difference between fourier transform of the numerical scheme and fourier transform of the spatial derivative has to be minimal .
taking the fourier transform of both sides of ( [ ddx_approx ] ) one can get the effective wave number @xmath40 of the fourier transform of the numerical scheme ( [ ddx_approx ] ) @xmath41 an assumption that the integral error @xmath42 of fourier transform of the finite difference scheme ( [ ddx_approx ] ) is minimal for waves with wavelength @xmath43 leads to the following equation @xmath44 or in the explicit form @xmath45 the rest five equations @xmath46 are obtained from a requirement that the numerical scheme ( [ ddx_approx ] ) approximates a spatial derivative with the 4th order . now calculation of the coefficients of 7-dot symmetrical stencil is straightforward .
the explicit expressions for the coefficients are the following : @xmath47 the plots of numerical wave number @xmath48 versus @xmath49 for different numerical schemes are shown in figure [ wavenum_fig ] .
dotted , dash - dotted , dashed , and solid curves represent classic 2nd , 4th , 6th , and drp 4th order schemes correspondingly .
one can see that the 4th - order drp scheme describes short waves more accurately than the classic 6th - order scheme .
waves with wavelength less than @xmath50 are not resolved by the numerical scheme ( [ ddx_approx ] ) .
they lead to point - to - point oscillations of the solution that can cause a numerical instability .
such waves have to be filtered out .
we used the following digital filter of the 6th - order to eliminate unresolved short wave component from the solution : @xmath51 where @xmath52 is the original grid function , @xmath53 is the filtered grid function , @xmath54 is the damping function , @xmath55 is the constant between 0 and 1 , determining the filter strength .
the frequency response function @xmath56 of the filter relates the fourier images of the original @xmath57 and filtered @xmath58 grid functions as follows @xmath59 . in this paper , the coefficients @xmath60 of the digital filter
have been chosen in such a way that @xmath61 coefficients @xmath60 of the digital filter are symmetric : @xmath62 using the technique proposed by @xcite we have found a stable 3rd - order boundary closure of the explicit dispersion - relation - preserving inner scheme ( [ ddx_approx ] ) with coefficients given by ( [ spatialcofs ] ) . obtained boundary closure has summation - by - parts properties .
this approach is based on the implicit pad approximation of spatial derivatives near the boundaries @xmath63 where matrices @xmath64 and @xmath65 satisfy the following conditions : 1 .
@xmath64 is symmetric non - singular matrix ( @xmath66 ) , 2 .
@xmath64 is positive - definite matrix ( @xmath67 for @xmath68 ) , 3 .
@xmath65 is almost skew - symmetric matrix , except corner top left and bottom right elements ( @xmath69 ) 4 .
@xmath70 . taking into account these properties
, one can write explicitly the top left corners of matrices @xmath64 and @xmath65 : @xmath71 where coefficients @xmath72 of the inner scheme are defined in ( [ spatialcofs ] ) .
expanding the left and right - hand sides of ( [ padederiv ] ) in taylor series at the top boundary and equating terms of the same order of @xmath73 , one can obtain a system of linear equations for coefficients @xmath74 and @xmath75 .
not all of these equations are independent , so the solution depends on two free parameters @xmath76 and @xmath77 : @xmath78 to satisfy a condition of positive definiteness it is sufficient to choose matrix elements @xmath76 and @xmath77 in such a way , that the signs of coefficients of a characteristic polynomial alternate .
however , this property does not guarantee a boundedness of the solution for all times , which is called asymptotic stability . to make a solution
be bounded for all times , all eigenvalues of the spatial discretization operator @xmath79 from eq.([pde2ode ] ) , incorporated with the boundary conditions , must have non - positive real parts .
details of this procedure can be found in @xcite . due to complexity of the original 3d problem
, we have tested a stability of the scheme on 1d advection problem .
distribution of eigenvalues of the drp spatial discretization operator in the complex plane for different choices of the pairs of coefficients ( @xmath77 , @xmath76 ) is shown in fig .
[ eigs_fig ] .
symbols plus correspond to the scheme @xmath80 , which does not exhibit asymptotic stability . circles and crosses represent choices ( 1/80 , 125/128 ) and ( -1/10 , 65/64 ) of coefficients @xmath81 correspondingly .
both schemes are asymptotically stable .
acoustic sources have been added to the right - hand side of equations ( [ lineuler ] ) .
we used sources of two types .
if we add a scalar function @xmath82 to the right - hand side of z - component of momentum equation , this term can be combined with the gravity term and interpreted as a source of z - component of force . if we add a gradient of a scalar function @xmath83 to the right - hand side of all momentum equations , these terms
can be combined with components of the pressure gradient and interpreted as a pressure source .
acoustic sources are spatially localized and have finite lifetime . spatial dependence is given by gaussian spherically symmetric function with semi - width of 2 - 3 grid nodes .
we experimented with two different time dependencies of acoustic sources : one period of sin function @xmath84,\ ; t_0\leq t\leq t_0 + 2\pi/\omega$ ] and ricker s wavelet @xmath85,\ ; t_0\leq~t\leq~t_0 + 4\pi/\omega$ ] .
such time dependencies were chosen because such sources are not monochromatic and have spectral power localized around central frequency @xmath86 , but spectral power is not too spread out . single or multiple acoustic sources can be added to the right hand side of equations .
multiple sources are randomly distributed at some depth ( in our case it was 350 km ) and initiated independently at arbitrary moments of time .
amplitudes and frequencies are randomly distributed on intervals [ 0 , 1 ] and [ 2 mhz , 8 mhz ] correspondingly .
thickness of both ( top and bottom ) pmls equals 5 grid nodes .
calculation of the acoustic spectrum requires long term simulations , so besides the asymptotic stability we have to prevent a spurious reflection of acoustic waves from the boundaries back to the computational domain . in this paper
we follow @xcite , who proposed a procedure to construct the pml for the euler equations .
it can be proofed that for a homogeneous medium and uniform mean flow without gravity pml absorbs waves without reflection for any angle of incidence and frequency .
we set non - reflecting boundary conditions based on the pml at the top and bottom boundaries of the domain .
the lateral boundary conditions are periodic . inside the pml independent variables
@xmath87 are split into sub - components @xmath88 such that @xmath89 .
thus , in the pml 3d system ( [ lineuler ] ) is split into 1d+1d+1d system of coupled locally one dimensional equations @xmath90 \displaystyle { \frac{\partial { \mbox{\boldmath$q$}}_2}{\partial t } } + { { \frac{\partial { \mbox{\boldmath$g$}}({\mbox{\boldmath$q$}})}{\partial y } } } = 0,\\[7pt ] \displaystyle { \frac{\partial { \mbox{\boldmath$q$}}_3}{\partial t } } + { { \frac{\partial { \mbox{\boldmath$h$}}({\mbox{\boldmath$q$}})}{\partial z } } } = { \mbox{\boldmath$s$}}({\mbox{\boldmath$q$}},t ) - \sigma_z{\mbox{\boldmath$q$}}_3 , \end{array}\ ] ] where @xmath91 is the damping factor , @xmath92 is the vertical coordinate inside the pml counted off from the interface of the pml with the inner region , d is the depth of the pml .
values of @xmath93 at the top and bottom boundaries are 0.3 and 1.0 correspondingly . in the paper of f.q .
hu a quadratic dependence of @xmath94 on the coordinate @xmath92 is used .
we were forced to add a small constant term to stabilize the pml in the presence of gravity .
it is important to note , that vectors @xmath32 , @xmath33 , and @xmath34 depend only on unsplit variable @xmath87 .
although @xmath95 , @xmath96 , and @xmath97 are not defined outside the pml , the variable @xmath87 , which is used for calculation of the spatial derivatives , is defined everywhere in the computational domain .
hence , inside the pml near the interface with the inner region we can use the same centered stencil as for the inner points . near the top and bottom boundaries the implicit pade approximation ( [ padederiv ] ) is used which guarantees numerical stability of the scheme .
we smoothly joined the chromospheric model provided by @xcite with the top of the standard solar model by christensen dalsgaard and established the top non - reflecting boundary condition based on the pml in the chromosphere above the temperature minimum .
this simulates a realistic situation when not all waves are reflected by the photosphere .
waves with frequencies higher than the acoustic cut - off frequency pass through the photosphere and will be absorbed by the pml layer .
for validation of the code we chose 1d initial boundary value problem ( ibvp ) for linearized euler equations with constant gravity @xmath98 : @xmath99 \displaystyle { { \frac{\partial } { \partial t}}}(\rho_0u')+{{\frac{\partial p'}{\partial x}}}=g_0\rho',\\[9pt ] \displaystyle { { \frac{\partial } { \partial t}}}(\rho_0\xi)=\rho_0u',\\[9pt ] \displaystyle p'=a_0 ^ 2\rho'+(\gamma-1)g_0(\rho_0\xi ) , \end{array } \begin{array}{l } 0\leq x\leq 1,\quad t\geq 0,\\[9pt ] \rho'(0,t)=\rho'(1,t)=0,\\[9pt ] \rho'(x,0)=f(x),\quad \rho_0(x)u'(x,0)=0,\\[9pt ] \displaystyle \rho_0(x)\xi(x,0)=-\int\limits_0^xf(\eta)\;{\mbox{d}}\eta , \end{array}\ ] ] where @xmath100 is the displacement .
we need to know @xmath100 to calculate the eulerian perturbation of pressure @xmath5 .
actually , the combination @xmath101 is used as a variable , so we group them together in equations .
initial conditions for @xmath100 must be consistent with the initial conditions for @xmath4 .
they are related by the continuity equation .
waves are adiabatic , the background model @xmath102 is hydrostatic and isothermal @xmath103 .
the last equation in the system ( [ ibvp_1d ] ) is the adiabatic relation ( [ adiabat ] ) written for the isothermal background model .
variable @xmath104 represents here the depth from the surface .
the system ( [ ibvp_1d ] ) is written in the same conservative form as the original system ( [ lineuler ] ) .
this problem was chosen for testing the code because it shows all characteristic behavior of the realistic solution and yet not too complicated and can be solved analytically . formally , the variable @xmath101 can be eliminated using the continuity equation in 1d , and system ( [ ibvp_1d ] ) can be reduced to the system of two equations . however , in 3d @xmath101 can not be eliminated . to have the test example as close to the real case as possible
, we left this variable and solved numerically the full system ( [ ibvp_1d ] ) .
analytical solution of these equations can be obtained by the method of separation of variables @xmath105 \displaystyle \xi(x , t)=e^{-x/2h}\sum_{n=1}^\infty b_n(\sin\pi
nx-2\pi nh\cos\pi nx ) \cos \lambda_na_0t,\\[9pt ] \displaystyle a_n=2\int_0 ^ 1 f(\eta)e^{-\eta/2h}\sin\pi n\eta { \mbox{d}}\eta,\quad b_n = -\frac{2ha_n}{1 + 4\pi^2n^2h^2},\\[9pt ] \displaystyle \lambda_n=\sqrt{\frac{1}{4h^2}+\pi^2n^2},\quad h^{-1}=\frac{\gamma g_0}{a_0 ^ 2}. \end{array}\ ] ] the following distribution of density perturbation @xmath106 was chosen as the initial condition for @xmath4 : @xmath107 ^ 2\quad \mbox{if}\;\ ; 0.4\leq x\leq 0.6 \\ 0 \end{array } \right.\ ] ] solution of the initial boundary value problem ( [ ibvp_1d ] ) for different moments of time with the initial density distribution given by ( [ ibvp_rhoini ] ) and parameters @xmath108 , @xmath109 , @xmath110 , @xmath111 , @xmath112 ( number of grid nodes ) is shown in figure [ ibvp_fig ] .
the left column represents the density perturbation , the right one shows the vertical displacement .
the solid curve is the exact solution ( [ ibvp_rhoxi ] ) .
the dashed curve represents the low - order ( classic 2/1 ) numerical solution which uses the 2nd - order classic central difference approximation of spatial derivatives for inner points with the one sided 1st - order scheme at the boundaries .
the high - order numerical solution is indistinguishable from the exact one .
it uses the drp 4/3 scheme ( dispersion - relation - preserving spatial discretization of the 4th order for inner points with the stable boundary closure of the 3rd order consistent with the inner scheme ) . the bottom panels give the profiles of density and displacement after reflection from the bottom boundary .
one can see , that the second order solution approximates the exact one well enough before the wave hits the boundary .
after this the accuracy of the solution switches from the second to the first order .
solution becomes too dispersive which causes nonphysical oscillations .
the high - order solution based on drp 4/3 scheme reproduces the exact solution well even after @xmath113 iterations and 20@xmath030 reflections from boundaries .
this test shows that the high - order drp numerical scheme does not introduce a noticeable damping or dispersion even on big intervals of integration .
these simulations also test an accuracy and stability of the numerical boundary conditions .
to test the efficiency of the pml for non - uniform isothermal background model we compared the numerical solution of ( [ ibvp_1d ] ) with the pml established at the top boundary with the exact solution of the same problem for infinite interval @xmath114 : @xmath115 bottom boundary condition for the numerical solution remains reflecting , because the bottom pml is inconsistent with the initial conditions for @xmath116 . at the bottom boundary in the initial moment of time @xmath117 ( see the top right panel of figure [ ibvp_fig ] ) .
if we established pml at the bottom , it would damp @xmath100 to zero value , generating non - physical perturbations near the bottom boundary which corrupt the solution . the analytical solution ( [ inf_1d ] )
does not contain reflected waves , because all initial perturbations propagate to infinity .
this solution can be used as a reference solution for determining the damping properties of the top pml .
results are shown in figure [ pml_fig ] at the moments of time @xmath118 0 , 0.2 , 0.4 , and 0.64 .
the value @xmath119 ( density perturbation with removed exponential factor ) is plotted .
the solid line represents the exact solution ( [ inf_1d ] ) , the dash - dotted line represents the numerical solution with pml at the top boundary , and the the dashed line represents the exact solution ( [ ibvp_rhoxi ] ) for the reflecting top boundary . the solid vertical line marks position of the interface between the top pml and inner region .
the dashed vertical line shows position of the initial perturbation .
the top pml reduces the amplitude of reflected wave by factor 20@xmath040 .
our original 3d system contains acoustic sources explicitly depending on time in the right hand side of the momentum equations .
we have tested the code in presence of sources on the same problem ( [ ibvp_1d ] ) with the pressure source term @xmath120 \displaystyle { { \frac{\partial } { \partial t}}}(\rho_0u')+{{\frac{\partial p'}{\partial x}}}=g_0\rho'-{{\frac{\partial \phi(x , t)}{\partial x}}},\\[9pt ] \displaystyle { { \frac{\partial } { \partial t}}}(\rho_0\xi)=\rho_0u',\\[9pt ] \displaystyle p'=a_0 ^ 2\rho'+(\gamma-1)g_0(\rho_0\xi ) , \end{array } \begin{array}{l } 0\leq x\leq 1,\quad t\geq 0,\\[9pt ] \rho'(0,t)=\rho'(1,t)=0,\\[9pt ] \rho'(x,0)=0,\quad \rho_0(x)u'(x,0)=0,\\[9pt ] \displaystyle \rho_0(x)\xi(x,0)=0 . \end{array}\ ] ] for test purposes we chose gaussian shaped harmonic source function as follows @xmath121 the system ( [ src_1d ] ) can be solved analytically by quadratures @xmath122
\displaystyle g(x,\eta,\tau ) = 2\;e^{(x-\eta)/2h } \sum_{n=1}^\infty \frac{\sin\lambda_na_0\tau}{\lambda_na_0 } \sin\pi nx \sin\pi n\eta,\\[15pt ] \displaystyle \lambda_n=\sqrt{\frac{1}{4h^2}+\pi^2 n^2}. \end{array}\ ] ] the results of numerical simulations with the source function given by ( [ src_func ] ) and parameters @xmath123 , @xmath124 , @xmath108 , @xmath109 , @xmath110 , @xmath125 , @xmath126 , @xmath127 are shown in figure [ src_1d_fig ] .
the non - reflecting boundary conditions are established on the top and bottom boundaries for numerical solution .
the solid curve represents the exact solution of ( [ src_1d_fig ] ) with zero boundary conditions for @xmath4 established at @xmath128 and @xmath129 .
the dashed line represents drp 4/3 numerical solution of ( [ src_1d_fig ] ) with non - reflecting top and bottom boundaries .
the vertical dashed line marks the position of the source .
the vertical solid line shows the position of the interface between the inner region of the computational domain and the non - reflecting pml .
the numerical solution reproduces the exact one well in the inner region , and is effectively damped by the absorbing layer , preventing unwanted reflection from the bottom boundary .
numerical simulations of propagation of waves in 3d from a single source inside the sun are shown in figure [ ss_3d_fig ] .
the brunt - visl frequency of the standard solar model with smoothly joined chromosphere was modified near the surface to make the model stable against convection .
such modified model was chosen as the background model .
non - reflecting boundary conditions ( pmls ) were established at the top and bottom boundaries .
the top layer was established at the height of 500 km above the photosphere in the region of the temperature minimum .
this layer absorbs all waves with frequencies higher than the acoustic cut - off frequency which pass to the chromosphere and do not affect reflection of waves with lower frequencies , because these waves are reflected from layers below the photosphere .
lateral boundary conditions are periodic .
the computational domain of size 120@xmath130120@xmath13050 mm@xmath131 was covered by the uniform grid of size 720@xmath130720@xmath130300 nodes with spatial intervals @xmath132 km
. the time step @xmath133 sec was chosen from stability condition .
the gaussian spherically symmetric pulse source of z - component of force @xmath134\sin(\omega_0 t),\quad 0\leq t\leq 2\pi/\omega_0.\ ] ] with @xmath135 mm was placed at the depth of 3.4 mm below the photosphere .
for such a choice of @xmath94 semi - width of the source equals approximately 4 grid nodes .
the time dependent part is just one period of sin - function with @xmath136 mhz .
figure [ ss_3d_fig ] shows snapshots of the density perturbation from such a source at @xmath137 min ( left column ) and @xmath138 min ( right column ) .
the top row represents the vertical slices of the computational domain , the bottom row shows the horizontal slices at a height of 350 km above the photosphere .
the thin horizontal line at @xmath139 represents the photosphere .
the left column shows the disturbance from the direct wave , generated by the source .
the right column shows the wave reflected from the photosphere .
the reflected wave front is broader and has less amplitude , because our source has finite lifetime ( one period of sine ) and generates high frequency waves which pass through the photosphere .
such waves are absorbed by the top non - reflecting layer and do not make a contribution to the amplitude of the reflected wave .
the developed numerical method and code have been used for simulation of the acoustic wave field generated by multiple acoustic sources inside the sun .
we found that the height of the pml affects absorbing properties of the top boundary and the shape of the acoustic spectrum ( k-@xmath140 diagram ) .
reflection from the top boundary is a wave process .
the waves are reflected not from the fixed level but from some vertical region .
region with the acoustic cut - off frequency greater than the wave frequency @xmath141 acts as a potential barrier for such waves . even if the wave frequency is less than the acoustic cut - off frequency waves penetrate to this region with exponentially decaying amplitude .
if the thickness of the barrier is finite waves can leak through it .
this process is similar to the tunneling effect in quantum mechanics .
this happens in the real sun as well .
we studied behavior of the solution for different heights of the top boundary .
all depths and heights are calculated from the level of the photosphere ( @xmath142 in the christensen - dalsgaard s standard solar model s ) .
the background model varies fast in the region above the temperature minimum . to be able to simulate propagation of acoustic waves in the chromosphere
we were forced to reduce the vertical spatial step to @xmath143 km to preserve the numerical stability . to keep the horizontal size of the domain as in previous simulations without significant increasing number of grid nodes the horizontal spatial steps were chosen 3 times bigger @xmath144 km . to satisfy the courant stability condition for the explicit scheme , the time step was reduced to @xmath145 sec .
the computational domain of size 122.2 mm @xmath130 122.2 mm @xmath130 32 mm is covered by the uniform grid of size 816@xmath130816@xmath130640 .
sources of z - component of force with random frequencies were randomly distributed at the depth of 350 km .
sources are initiated at random moments of time ( one source per time step ) and depend on time as ricker s wavelet with central frequency from range 2@xmath08 mhz . in the sun
we have some damping due to the turbulent viscosity .
we simulated this additional viscosity by adding a damping term @xmath146 to the right hand side of z - momentum equation in the region above the photosphere and smoothly fading to zero below .
the time dependence of rms ( root mean square ) wave amplitude averaged along the horizontal plane at the height of 300 km above the photosphere for different heights @xmath147 of the top boundary and different values of the damping coefficient @xmath148 is shown in figure [ meanampl_fig ] .
the rms amplitude for the high pml established at the height of @xmath149 km without additional damping @xmath148 = 0 is shown by the solid curve i. the rms amplitudes for the same height of the top pml and @xmath148 = 0.3 , 0.6 , and 1.0 are plotted by solid lines ii , iii , and v correspondingly .
in the last case the rms amplitude reaches an equilibrium state .
the curve iv corresponds to the low pml , established at the height of 500 km above the photosphere without additional damping in the inner region @xmath148 = 0 .
rms amplitude reaches an equilibrium state in this case as well , because the acoustic modes leak through the acoustic potential barrier and their exponential tails reach the top absorbing boundary , which adds an additional damping and stabilizes the amplitude .
the top boundary with the height of 1750 km is set high enough and does not affect modes with frequencies less than the acoustic cut - off frequency . energy is continuously pumped to the system by acoustic sources . the bottom boundary is deep enough that some modes resolved by numerical scheme have turning points above the bottom boundary .
such modes are trapped in the domain , the total energy increases , and the rms amplitude does not reach an equilibrium state .
this distorts the acoustic power spectrum and changes the amplitude ratio of trapped modes and modes that can be absorbed at the top and/or bottom boundaries .
the left panes in figure [ kw_fig ] show the acoustic power spectra obtained from observations ( top ) , simulations without damping with low ( 500 km ) top boundary ( middle ) , and high ( 1750 km ) top boundary ( bottom ) .
the right panes show the vertical cuts of corresponding k-@xmath140 diagrams at @xmath150 .
the bottom left pane ( high top boundary ) shows presence of g- modes in simulations .
they appear because our background model is convectively stable in the thin layer below the photosphere . in the real sun
, this layer is convectively unstable , and thus the g- modes do not propagate .
energy leakage through the acoustic potential barrier in the case of low top boundary ( middle row ) damps all modes uniformly and does not change the shape of the acoustic spectrum .
so the height of the top boundary can be used for controlling the damping rate without distortion of the acoustic spectrum .
developed linear 3d code for propagation of acoustic waves inside the sun uses the realistic equation of state and realistic non - reflecting boundary conditions which permits to simulate accurately reflection of waves from the top boundary . waves with frequencies less then the acoustic cut - off frequency are reflected from the photosphere , and waves with higher frequencies pass to the chromosphere .
the top non - reflecting boundary absorbs such waves .
establishing the top boundary high enough in the chromosphere leads to the trapping of some modes in the domain and increasing their amplitudes with time , which distorts the shape of the acoustic spectrum .
energy leakage through the acoustic potential barrier in case of the low ( 500 km ) top boundary leads to an additional uniform damping of all modes which stabilizes the amplitude and does not distort the spectrum .
the height of the non - reflecting top boundary can be used as a parameter for controlling of the damping rate in the system .
the acoustic spectrum obtained from simulated wave field shows existence of p- , and f - modes .
the simulated acoustic spectrum is good agreement with observations .
this code has been used by @xcite to model the effects of non - uniform spatial distribution of acoustic sources in sunspot regions .
their results showed that this effect can explain at least a half of the observed amplitude reduction in sunspots .
the code can be used for studying details of interaction of waves with inhomogeneities of solar structure and for producing artificial data for testing an accuracy of helioseismic inversion as well , as for studying of propagation of acoustic waves in the chromosphere and reflecting properties of the photosphere .
future simulations will include subsurface flows and magnetic field .
this research is supported by the living with the star nasa grant nng05gm85 g .
the calculations were performed on columbia supercomputer at nasa ames research center ( nasa advanced supercomputing division ) .
kosovichev a. g. , duvall t. l. jr .
1997 , in score96 : solar convection and oscillations and their relationship , eds .
f. p. pijpers , j. christensen - dalsgaard , & c. s. rosenthal ( kluwer academic publishers , dordrecht ) , 241 . | results of numerical 3d simulations of propagation of acoustic waves inside the sun are presented . a linear 3d code which utilizes realistic opal equation of state
was developed by authors .
modified convectively stable standard solar model with smoothly joined chromosphere was used as a background model .
high order dispersion relation preserving numerical scheme was used to calculate spatial derivatives . the top non - reflecting boundary condition established in
the chromosphere absorbs waves with frequencies greater than the acoustic cut - off frequency which pass to the chromosphere , simulating a realistic situation .
the acoustic power spectra obtained from the wave field generated by sources randomly distributed below the photosphere are in good agreement with observations .
the influence of the height of the top boundary on results of simulation was studied .
it was shown that the energy leakage through the acoustic potential barrier damps all modes uniformly and does not change the shape of the acoustic spectrum .
so the height of the top boundary can be used for controlling a damping rate without distortion of the acoustic spectrum .
the developed simulations provide an important tool for testing local helioseismology . |
the nature and consequences for confinement and chiral - symmetry breaking of long - range gluon fields in the vacuum have been the subject of extended investigation . while many studies rely on one or more of the specific configurations of monopoles , vortices and instantons , there are arguments @xcite that in fact the complete hierarchy of singular gauge fields must play a role , especially if these objects condense in the vacuum .
it is natural then to seek a model in which the bulk properties of this hierarchy can be effectively represented .
we partially do this here by introducing a model for such fields in which singularities in vector potentials are concentrated on three - dimensional hypersurfaces @xmath0 in euclidean space , in the vicinity of which gauge fields can be divided into a sum of a singular pure gauge @xmath1 and a regular fluctuation part @xmath2 , and where a certain @xmath3 colour vector @xmath4 can be associated with @xmath1 .
thus we have idealised the situation by ignoring lower dimensional singular objects which would appear as `` dislocations '' at the abelian domain wall singularities .
the requirement of finiteness of the action ( density ) for singular background plus fluctuation parts for configurations dominating the functional integral @xcite can be formalised in our case as specific boundary conditions on @xmath5 for the fluctuation fields charged with respect to @xmath6 . the interiors of these regions thus constitute `` domains '' @xmath7 .
gauge field modes neutral with respect to @xmath8 are not restricted and generate interactions between domains . within the present framework
then , we decompose a general gauge field @xmath9 and demanding finiteness of the classical action density at the singularity we come to the conditions @xmath10 for gluons , and @xmath11 for quarks .
the adjoint matrix @xmath12 appears in the condition for gluons , and a bag - like boundary condition arises for quarks , with a unit vector @xmath13 normal to @xmath0 . in a given domain @xmath7 the effect of fluctuations in neighbouring regions
can be seen as an external gauge field @xmath14 neutral with respect to @xmath4 .
this enables an approximation in which different domains are assumed to be decoupled from each other but , to compensate , a certain mean field is introduced in domain interiors . to make the model analytically tractable we consider spherical domains with fixed radius @xmath15 and approximate the mean field in @xmath7 by a configuration with the field strength @xmath16 , which is ( anti-)self - dual @xmath17 so that @xmath18 , with @xmath19 a constant the same for all domains .
the constant colour matrix @xmath20 belongs to the cartan subalgebra with angles @xmath21 corresponding to the weyl subgroup .
there is no source for this field on the boundary and therefore it should be treated as strictly homogeneous in all further calculations .
the homogeneity itself appears here just as an approximation .
thus the partition function we will deal with can be written @xmath22 \delta_{\rm fp}[\breve{\cal b}^{(i)},q^{(i ) } ] \nonumber \\ { } & { } & \times \exp\left\ { - s_{v_i}^{\rm qcd } \left[q^{(i)}+{\cal b}^{(i)},\psi^{(i)},\bar\psi^{(i)}\right ] \right\ } \nonumber\end{aligned}\ ] ] where the gluons and quarks are integrated over the spaces @xmath23 and @xmath24 respectively , and the thermodynamic limit assumes @xmath25 but with the density @xmath26 taken fixed and finite .
the measure of integration over parameters characterising domains is defined as @xmath27 here @xmath28 and @xmath29 are spherical angles of the chromomagnetic field , @xmath30 is the angle between the chromomagnetic and chromoelectric fields , @xmath31 is the angle in the colour matrix @xmath32 , @xmath33 is the chiral angle and @xmath34 is the centre of the domain @xmath35 .
this partition function describes a statistical system of density @xmath36 composed of noninteracting extended domain - like structures , each of which is characterised by a set of internal parameters and whose internal dynamics are represented by the fluctuation fields .
in this model the connected @xmath37point correlator of field strength tensors , @xmath38 can be calculated explicitly using the measure , eq . ( [ measure ] ) .
translation - invariant functions @xmath39 emerge and can be seen as the volume of the region of overlap of @xmath40 hyperspheres of radius @xmath15 and centres ( @xmath41 ) , normalised to the volume of a single hypersphere @xmath42 .
the functions @xmath43 are continuous and vanish if @xmath44 .
correlations in the background field have finite range @xmath45 .
the fourier transform of @xmath43 is then an entire analytical function and thus correlations do not have a particle interpretation .
the statistical ensemble of background fields is not gaussian since all connected correlators are independent of each other and can not be reduced to the two - point correlations .
the simplest application of the above correlators gives a gluon condensate density which to this approximation is @xmath46 another vacuum parameter which plays a significant role in the resolution of the @xmath47 problem is the topological susceptibility @xcite . to define this we consider first the topological charge density @xmath48 which to lowest order is @xmath49\cos\omega_j , \nonumber\end{aligned}\ ] ] where @xmath50 depends on the duality of the @xmath51-th domain .
for a given field configuration the topological charge is additive @xmath52 for @xmath53 where @xmath54 is a ` unit ' topological charge , namely the absolute value of the topological charge of a single domain , and @xmath55 @xmath56 is the number of domains with ( anti-)self - dual field , @xmath57 .
the probability of finding the topological charge @xmath58 in a given configuration is defined by the ratio of the number of configurations @xmath59 with a charge @xmath58 and the total number of configurations @xmath60 , @xmath61 the distribution is symmetric about @xmath62 , where it has a maximum for @xmath63 even .
for @xmath63 odd the maximum is at @xmath64 . averaged topological charge is zero . the topological susceptibility @xmath65 is determined by the two - point correlator of topological charge density , which in the lowest approximation gives @xmath66
to zeroth order in fluctuations the wilson loop can be written @xmath67 where stokes theorem has been used .
note that path ordering in our case is not necessary since the matrices @xmath68 are assumed to be in the cartan subalgebra .
computationally it is convenient to consider a circular contour in the @xmath69 plane of radius @xmath70 with centre at the origin .
details of the calculation are given in @xcite .
put briefly , the colour trace is evaluated exactly followed by the integration over orientations of the vacuum field . integrating over the domain centres and taking the thermodynamic limit ( @xmath71 , @xmath72 ) , one obtains for a large wilson loop @xmath73 @xmath74^n = e^{-u(l ) } \nonumber\end{aligned}\ ] ] with the exponent @xmath75 displaying an area law with string tension for @xmath76 given by @xmath77 with the function @xmath78 having a purely geometrical origin .
this function is positive for @xmath79 and has a maximum for @xmath80 .
we choose this maximum to estimate the model parameters by fitting the string constant to the lattice result , @xmath81 and get for the gluonic parameters of the vacuum @xmath82 and @xmath83 , while the density of the system is @xmath84 .
there is no separation of characteristic scales @xmath85 , hence an approximation based on large or small domains is not justifiable and has been not used here .
if @xmath19 goes to zero then the string constant vanishes .
this underscores the role of the gluon condensate in the confinement of static charges . on the other hand , if the number of domains is fixed and the thermodynamic limit is defined as @xmath86 , namely if the domains are macroscopically large , then @xmath87 , which indicates the absence of a linear potential between infinitely massive charges in a purely homogeneous field .
an area law obviously does not occur for adjoint charges because of the presence of zero eigenvalues of the adjoint matrix @xmath88 .
confinement of the fluctuation fields can be seen in the analytical properties of their propagators . for the above boundary conditions the fluctuation quark and gluon propagators can be analytically calculated by reduction to the charged scalar field problem .
this in turn is essentially just that of a four - dimensional harmonic oscillator with the orbital momentum coupled to the external field , and the general solution has been found exactly by decomposition over hyperspherical harmonics @xcite . due to the above boundary conditions the @xmath89space propagators of charged fields are defined in regions of finite support where they have integrable singularities so that their fourier transforms are entire functions in the complex momentum plane
this is consistent with a confinement of dynamical fields@xcite .
entire propagators have the physically appealing property of a regge spectrum of relativistic bound states@xcite .
turning to quarks specifically , we address the eigenvalue problem for the massless dirac operator in a domain , @xmath90 subject to the boundary condition eq.([bc ] ) .
we introduce projectors in the colour direction @xmath91 and spin projection with respect to the magnetic field @xmath92 with @xmath93 .
it is also useful to introduce the mixed projector @xmath94 .
substituting @xmath95 shows that @xmath96 satisfies @xmath97 whose exact solution is given in @xcite .
it turns out that it is impossible to fulfill the boundary condition eq.([bc ] ) with the term @xmath98 in eq.([q - pr1 ] ) .
thus @xmath99 must have definite chirality corresponding to the duality of the vacuum gluon field .
the boundary condition enforces a discretisation of the eigenvalues @xmath100 .
the states are then labelled by the set of quantum numbers : @xmath101 with @xmath102 a radial quantum number labelling the solutions of the boundary condition for given @xmath103 angular quantum numbers @xmath104 , @xmath105 corresponding to the colour projection with respect to @xmath106 and @xmath107 representing the spin projection with respect to the magnetic field .
zero modes are absent under chosen boundary conditions .
the eigenspinors have the structure @xmath108 where spinors @xmath109 and @xmath110 as well as the constraints defining the corresponding eigenvalues are known in analytical form and will be discussed in detail elsewhere . for illustration we give the eigenvalue equations for the purely radial @xmath111 modes . with @xmath112 , @xmath113
we have the equation for @xmath114 @xmath115 and that for @xmath116 @xmath117 \nonumber \\ + e^{\mp i\alpha}\lambda m\left(\lambda^2 + 2,2,z\right)=0 \nonumber \end{aligned}\ ] ] and @xmath118 and @xmath119 . here
@xmath120 is the confluent hypergeometric function .
the sign minus ( plus ) in front of @xmath121 corresponds to the ( anti-)self - dual domain . with these solutions
we investigate the chirality of the eigenmodes and also how their probability density correlates with the underlying domain .
following @xcite we introduce the local chirality parameter @xmath122 defined via @xmath123 which will give extremal values @xmath124 at positions @xmath125 where @xmath126 is purely right(left ) handed . for lattice overlap fermions , even in uncooled gauge fields , histograms of @xmath127 measured at probability density maxima for low - lying overlap - dirac eigenmodes
show peaks close to the extremal values @xcite .
this indicates that low - lying modes are strongly chiral and are thus a useful filter for the duality of `` objects '' underlying the gauge field fluctuations . within the present model we observe
that at the centres of domains all of the eigenmodes are exactly chiral and probability densities are maximal .
the `` width '' of the peaks for the lowest modes at half - maximum varies for different values @xmath129 and is of the order of @xmath130 , consistent with the lattice observations of @xcite . as is seen from fig.([fig : x ] ) the chirality of the lowest mode ( @xmath131 ) monotonically decreases with distance from the centre .
the chirality parameter for the excited modes alternates between extremal values , the number of alternations is correlated with the radial number @xmath40 , and the half - width decreases with growing @xmath40 .
the chirality parameter is zero at the boundary for all modes .
qualitatively this picture does not depend on angle @xmath129 .
the degree of chirality of the modes can be conveniently characterized via a histogram if we average x(x ) over a small neighborhood of the domain centre .
then , by assuming that all values of @xmath129 are equally probable we compute the probability to find a given value of smeared @xmath127 among a set of modes .
the result given in fig.([fig : x - hist ] ) was obtained for the lowest modes with all possible spin - color orientaions .
we observe the double peaking close to the extrema with @xmath132 . including higher modes
will broaden the peaks and build up a central plateau .
this feature as well as the above mentioned values for the half - decay and the density of domains is in qualitative and quantitative agreement with recent lattice results @xcite .
however , at this stage , it would be premature to go beyond purely numerical comparison of our results and those of the lattice .
finally , we quote the estimate of the quark condensate density averaged over ( anti-)self - dual configurations and @xmath129 obtained in @xcite directly from the quark propagator . for @xmath19 and @xmath15 as in eq .
( [ par - val ] ) the density at the domain center is equal to @xmath134 it should be stressed that this nonzero result is due to the non - integral topological charge @xmath135 associated with domains .
this and our results for the chirality parameter @xmath127 suggest that the underlying mechanism of chiral symmetry breaking here is the strong chirality of the low - lying non - zero modes , which in turn is a consequence of the ( anti-)self - duality of the background gluonic configurations and their non - integral topological charge .
p. van baal , hep - ph/0008206 , at the frontier of particle physics , vol . 2 683 - 760 , ed . by m. shifman ,
world scientific , 2001 .
f. lenz , o. jahn , phys .
d58 ( 1998 ) 085006 ; f. lenz , s. worlen , hep - ph/0010099 , at the frontier of particle physics , vol . 2 761 - 821 , ed . by m.
shifman , world scientific , 2001 .
crewther , phys.lett .
b70 ( 1977 ) 349 ; g. veneziano , nucl . phys .
b159 ( 1979 ) 213 ; e. witten , nucl .
b156 ( 1979 ) 269 .
a.c.kalloniatis , s.n .
nedelko , phys .
d64 ( 2001 ) 114025 .
h. leutwyler , phys .
b96 ( 1980 ) 154 .
efimov , s.n .
nedelko , phys .
d51 ( 1995 ) 174 ; g.v .
efimov , g. ganbold , hep - ph/0103101 .
kalloniatis , s.n .
nedelko , _ in preparation_. i. horvath et al .
, hep - lat/0102003 .
edwards , u.m .
heller , hep - lat/0105004 .
dong et al . , hep - lat/0110037 .
s. duerr , a. wipf , nucl .
b443 ( 1995 ) 201 . | singular gauge fields in the partition function for qcd can lead to a domain - like picture for the qcd vacuum by virtue of constraints on quantum fluctuations at the singularities . with a simple model of hyperspherical domains with interiors of constant field strength we show that the main features of gluon condensation and an area law for static quarks can be realised .
the dirac operator in such a background is exactly soluble .
chirality properties of the solutions show agreement with recent lattice results . |
high - energy x - rays can penetrate extremely large column densities of gas and dust ; hence , hard x - ray surveys are our best current method for obtaining the least biased samples of active galactic nuclei ( agns ) .
ultradeep surveys with the _ chandra x - ray observatory _ detect the faintest x - ray sources in both the soft ( @xmath15 kev ) and hard ( @xmath6 kev ) x - ray energy bands and have revealed large numbers of agns that were missed in optical surveys . in the present paper
we describe the optical and near - infrared properties of the 503 x - ray point - sources detected in the @xmath0 ms _ chandra _ exposure of the region around the hubble deep field - north called the _
chandra _ deep field - north ( cdf - n ) .
this exposure is the deepest x - ray image to date .
barger et al . ( 2002 ) presented a similar catalog of the 370 x - ray point - sources detected in the @xmath5 ms cdf - n exposure ( brandt et al .
they had very complete redshift identifications ( 78% ) for the @xmath4 galaxy sources in a @xmath16 radius around the approximate x - ray image center and found spectroscopic evidence for large - scale structure in the field , which could account for a part of the field - to - field variation seen in the x - ray number counts ( e.g. , cowie et al .
2002 ; yang et al .
such large - scale structure is also seen in the spectroscopic redshift distribution ( hasinger 2002 ; gilli et al .
2003 ) of the x - ray sources in the @xmath17 ms exposure of the _ chandra _ deep field - south ( giacconi et al .
2002 ) .
cowie et al . ( 2003 ) constructed rest - frame @xmath6 kev luminosity functions versus redshift for the agns in a number of _ chandra _ ( including the 1 ms cdf - n ) , _ rosat _ , and _ asca _ deep fields .
these authors found that at @xmath18 , most of the @xmath6 kev energy density arises in sources with luminosities in the @xmath19 ergs s@xmath3 range .
they also showed that the number density of sources in this luminosity range rises with decreasing redshift , while the number density of higher luminosity sources exhibit the well - known peak at @xmath20 .
cowie et al . ( 2003 ) argued that the dominant supermassive black hole formation occurred at recent times in sources with low accretion , rather than at earlier times in more x - ray luminous sources with high accretion .
the 1 ms _ chandra _ surveys have also detected apparently normal galaxies with x - ray - to - optical flux ratios lower than those of agns , e.g. , @xmath21 ( barger et al .
2002 ; rosati et al .
the majority of x - ray sources with @xmath22 have been shown to be consistent with infrared and radio - emitting starburst galaxies ( alexander et al .
2002a ; bauer et al . 2002 ) , while those with even lower x - ray - to - optical flux ratios are generally consistent with quiescent galaxies with low x - ray luminosities ( @xmath23 ergs s@xmath3 ; hornschemeier et al .
2003 ) . these apparently normal galaxies are distinct from the x - ray luminous , `` optically normal '' galaxies that have been discovered in _
chandra _ and _ xmm - newton _ surveys ( e.g. , barger et al .
2001b ; comastri et al .
the latter sources show no high ionization signatures of agn activity in their optical spectra , and their hard to soft x - ray flux ratios suggest they are highly absorbed systems whose column densities could effectively extinguish the ultraviolet , optical , and near - infrared continua from the agns .
alternatively , moran , filippenko , & chornock ( 2002 ) propose that , since the entire galaxy often falls within the spectrograph slit for these distant systems , the host galaxy light from stars and hii regions could be overwhelming the emission - line signatures of the agn activity .
the 1 ms exposure of the cdf - n has since been extended to a second megasecond ( alexander et al .
2003 ) , enabling the detection of increasing numbers of faint agns and apparently normal galaxies . in this paper
we present multicolor imaging and optical spectroscopy of the x - ray sources detected in the @xmath24 ms cdf - n exposure .
we use these data to characterize the redshift distribution and properties of the faint x - ray sources .
we take @xmath25 km s@xmath3 mpc@xmath3 and use an @xmath26 , @xmath27 cosmology .
alexander et al . ( 2003 ; hereafter , a03 ) presented the 1.945 ms cdf - n x - ray images , along with details of the observations , data reduction , and technical analysis .
a03 merged their point - source lists in seven x - ray bands , @xmath28 kev ( full band ) , @xmath29 kev ( soft band ; sb ) , @xmath6 kev ( hard band ; hb ) , @xmath30 kev ( sb1 ) , @xmath31 kev ( sb2 ) , @xmath32 kev ( hb1 ) , and @xmath33 kev ( hb2 ) , into a catalog ( tables 3a and 3b in a03 ) of 503 significantly detected point sources over an area of about 460 arcmin@xmath34 . near the aim point , the data reach limiting fluxes of @xmath35 ( @xmath15 kev ) and @xmath36 ergs @xmath37 s@xmath3 ( @xmath6 kev ) . in table
a1 of the appendix we present optical magnitudes and spectroscopic measurements , where available , for the full x - ray point source catalog of a03 .
the optical imaging data consist of johnson @xmath38 , johnson @xmath39 , cousins @xmath40 , cousins @xmath41 , and sloan @xmath42 observations obtained with the subaru prime - focus camera suprime - cam ( miyazaki et al .
2002 ) on the subaru 8.2 m telescope during february - april of 2001 and 2002 .
the suprime - cam observations and reductions are described in capak et al .
( 2003a ) , where catalogs of the entire sample of galaxies and stars in the field can be found .
@xmath43 band observations were obtained with the quick infrared camera quirc ( hodapp et al .
1996 ) on the university of hawaii 2.2 m telescope .
the notched @xmath43 filter has a central wavelength of 1.8 @xmath44 m and covers the longer wavelength region of the @xmath45 band and the shorter wavelength region of the @xmath46 band ( roughly the @xmath47 filter ) . because of its broad bandpass and low sky background ,
this filter is extremely fast and is roughly twice as sensitive as the @xmath45 , @xmath47 , or @xmath48 filters .
barger et al . ( 1999 ) found the empirical relation @xmath49 to convert between the @xmath43 and @xmath46 bands ; for galaxies at most redshifts , @xmath50 .
@xmath51 band observations were obtained with the mosaic prime - focus camera ( jacoby et al .
1998 ; muller et al .
1998 ; wolfe et al . 1998 ) on the kitt peak national observatory 4 m telescope .
the @xmath43 and @xmath51 band data are described in capak et al .
( 2003a ) .
the x - ray catalog of a03 has extremely good positional accuracy , which simplifies the problem of matching the x - ray sources to their optical counterparts . for individual x - ray sources ,
the x - ray positional uncertainties are expected primarily to reflect the centroiding accuracies and the overall distortions in the x - ray data , and are dependent on the off - axis angles and the detected counts . for each source
, a03 gives an @xmath52% confidence x - ray positional uncertainty in column [ 4 ] of their table 3a .
the largest such positional uncertainty is just under @xmath53 for a small number of sources at the outside edge of the _ chandra _ field , while about 90% of the sources have positional uncertainties less than @xmath54 .
the absolute astrometry of both the x - ray catalog of a03 and the optical catalog of capak et al .
( 2003a ) have been matched to the vla 20 cm catalog of richards ( 2000 ) .
we tested the accuracy of the individual optical source positions relative to the vla 20 cm positions for the 20 cm sources lying in the cdf - n region and found a @xmath55 deviation of @xmath56 .
eighty percent of the optical sources with 20 cm counterparts lie within @xmath57 of the radio positions .
since the radio and optical morphologies may differ , these numbers should represent an upper bound on the accuracy of the optical positions , which are generally more accurate than the x - ray positions .
given the x - ray positional uncertainties , we identified an x - ray source with an optical counterpart if an optical source brighter than @xmath58 ( @xmath59 ) or @xmath60 ( @xmath59 ) is within @xmath61 of the x - ray position .
if more than one such optical counterpart is within the search radius , then we identified the x - ray source with the nearest optical neighbor . using this criterion , there are 429 x - ray sources with optical counterparts in the a03 sample , 278 of which have magnitudes @xmath4 . from capak
et al . ( 2003a ) , the surface density of galaxies and stars in the cdf - n region to @xmath62 is approximately 16.6 arcmin@xmath63 .
thus , with a search radius of @xmath61 and a sample of 503 sources , we may expect a random field contamination of about 30 sources to @xmath62 , or roughly 10% .
( a more sophisticated analysis allowing for clustering would yield a slightly higher value , e.g. , mcmahon et al . 2002 .
) this contamination rises at fainter magnitudes . for @xmath64 , where the field surface density is 50.1 arcmin@xmath63
, we may expect that about 34 of the 140 optical identifications , or roughly 25% , are chance projections , based on the 225 sources that are not identified with @xmath65 sources .
the high contamination rate at faint optical magnitudes emphasizes the need for the highest possible spatial accuracy in the x - ray sample .
if we restrict to the 403 sources where the x - ray positional uncertainties are less than @xmath66 , and if we use a smaller search radius of @xmath66 to match to the optical sources , then we find that 307 of the x - ray sources have @xmath67 counterparts , and less than 10% of these will be random field sources .
seven sources lie within the envelopes of bright galaxies but are separated by more than @xmath61 and less than @xmath68 from the nuclear positions . since these may be sources lying within the galaxies , we have placed the optical positions of these off - axis x - ray sources ( sources 121 , 197 , 270 , 316 , 404 , 410 , and 479 ) at the bright galaxy nuclei . however , the number of examples is sufficiently small that some of these may well just be chance projections of sources lying behind the bright galaxies .
four of these seven sources lie within @xmath69 of the x - ray positions , the faintest having @xmath70 , while two lie within @xmath71 ( @xmath72 ) and one within @xmath68 ( @xmath73 ) . with 503 sources in our x - ray sample ,
we expect three sources to lie within @xmath71 of an @xmath74 galaxy and three sources to lie within @xmath69 of an @xmath75 galaxy , based on the surface density of @xmath40 band selected sources in the cdf - n region ( capak et al .
2003a ) .
since many of the x - ray sources have positions which are much more accurate than our chosen @xmath61 search radius , we can examine how accurate the a03 estimates of the x - ray positional uncertainties are , and how reliable our counterpart identifications are , by plotting the distribution of @xmath4 counterparts versus radial x - ray - optical separation in arcseconds for three different ranges of x - ray positional uncertainty . in figures
[ fig1]a , 1b , and 1c we plot the @xmath4 sources with x - ray positional uncertainties @xmath76 , @xmath77 , and @xmath78 , respectively .
dashed lines show how many random @xmath65 sources are expected as a function of radial x - ray - optical separation , assuming a uniform spatial distribution .
if the a03 x - ray positional uncertainties are accurate , then we should see tighter distributions when we look at smaller x - ray positional uncertainty ranges , and this is indeed what we see . for the two highest accuracy bins , random field contamination is negligible and less than 20% of the counterparts lie beyond the 80 % x - ray positional uncertainty limits in both cases . for the sources with optical counterparts , we measured the magnitudes centered on the optical positions .
( the optical positions are set to the peak of the optical emission . ) for the 67 x - ray sources with optically fainter counterparts , we centered on the x - ray positions . in all cases we measured the optical magnitudes in @xmath79 diameter apertures and corrected these to total magnitudes using an average offset in color . for the @xmath40 band
we also give isophotal magnitudes .
further details may be found in capak et al .
( 2003a ) .
table a1 of the appendix gives @xmath38 and @xmath39 magnitudes in the johnson system , @xmath40 and @xmath41 magnitudes in the kron - cousins system , @xmath43 magnitudes in the wainscoat - cowie system ( wainscoat & cowie 1992 ) , and @xmath51 and @xmath42 magnitudes in the ab magnitude system .
an ab magnitude is defined by @xmath80 , where @xmath81 is the flux of the source in units of ergs @xmath37 s@xmath3 hz@xmath3 .
the @xmath59 magnitude limits of the images in the above systems are 27.1 ( @xmath51 ) , 27.0 ( @xmath38 ) , 26.8 ( @xmath39 ) , 26.4 ( @xmath40 ) , 25.1 ( @xmath41 ) , 25.4 ( @xmath42 ) , and 20.5 ( @xmath43 ) .
we show @xmath42 band thumbnail images of the cdf - n sources in figure a1 of the appendix . in a few cases
there is complex structure in the neighborhood of the x - ray position that complicates the photometry and counterpart identification .
we have noted in table a1 the sources for which this might be a problem .
in figure [ fig2]a we plot @xmath40 magnitude versus @xmath15 kev flux for the cdf - n soft x - ray sources . at bright x - ray fluxes
, we also plot the lehmann et al .
( 2001 ) _ rosat _ ultra deep survey data ( _ solid diamonds _ ; groups and clusters and one optically unidentified source have been excluded ) .
spectroscopically identified stars are denoted by asterisks .
we note in passing that flares of at least factors of several are observed in some of the stars ( e. d. feigelson et al .
, in preparation ) , as can be seen from their high @xmath82 ratios . in figure [ fig2]b we plot @xmath40 magnitude versus @xmath6 kev flux for the cdf - n hard x - ray sources . at bright x - ray fluxes ,
we include the akiyama et al .
( 2000 ) _ asca _ large sky survey data ( _ filled diamonds _ ; the two clusters and one source without an optical identification have been excluded , and the star is off - scale ) , after converting their @xmath83 kev fluxes to @xmath6 kev assuming @xmath84 ( the value akiyama et al .
assumed in their paper for the intrinsic photon index ) .
the shaded regions in figure [ fig2 ] indicate typical x - ray - to - optical flux ratio ranges for different source classes .
x - ray - to - optical flux ratios were found to be a good way to discriminate between source classes at bright x - ray fluxes ( @xmath85 ergs @xmath37 s@xmath3 ) from the _ einstein observatory _ extended medium - sensitivity survey data ( e.g. , maccacaro et al . 1988 ) ; their discriminating power was later confirmed at greater depths using ultradeep _ rosat _
( schmidt et al . 1998 ) and , finally , _
chandra _ data ( hornschemeier et al . 2001 ;
barger et al . 2002 ) .
the flux in the @xmath40 band , @xmath86 , is related to the @xmath40 magnitude by @xmath87 .
agns typically lie in the regions defined by the loci @xmath88 ( _ lightest shading _ ) .
we see from figure [ fig2 ] that this trend continues to hold for a large number of sources down to very faint optical magnitudes and x - ray fluxes .
median optical magnitudes for the cdf - n x - ray sources ( _ large open squares _ ) and for the lehmann et al . (
2001 ) ( fig .
[ fig2]a ) or akiyama et al . ( 2000 ) ( fig .
[ fig2]b ) sources ( _ large open diamonds _ ) are also shown .
the horizontal bars show the widths of the flux bins , while the vertical bars show the 68% confidence ranges in the medians computed using the number of sources in each bin ( gehrels 1986 ) . in figure [ fig2]a , at bright x - ray fluxes ,
@xmath89 matches the median optical magnitudes because the total light output from the sources is dominated by unobscured agns . at @xmath90 ergs @xmath37 s@xmath3 ,
quite a few sources whose optical counterparts are brighter than expected for agns begin to populate the sample .
the median optical magnitudes drop as host galaxy light from this growing population of normal galaxies begins to dominate the total light output and the ratio of the x - ray to optical light decreases .
by @xmath91 ( _ darkest shading _ ) , most of the sources are fairly normal galaxies whose x - ray emission is dominated by processes associated with star formation and accreting binary systems ( hornschemeier et al .
2003 ) . in figure [ fig2]b , at the brightest x - ray fluxes
, @xmath92 also matches the median optical magnitudes .
however , at fainter x - ray fluxes , the medians deviate above @xmath92 as the sources become obscured in the optical due to dust and gas . at the faintest x - ray fluxes ,
the medians flatten as the optical light from the host galaxy begins to dominate the total light output from each source .
the spectroscopic observations described in barger et al . ( 2002 ) were made with the low - resolution imaging spectrograph ( lris ; oke et al .
1995 ) on the keck 10 m telescopes and with the hydra spectrograph ( barden et al .
1994 ) on the wiyn 3.5 m telescope . in this paper
we present new spectroscopic observations that were obtained with the deep extragalactic imaging multi - object spectrograph ( deimos ; faber et al .
2002 ) on keck ii the nights of ut 2003 january 2930 , march 27 , and april 2527 .
the observations were made with the 600 lines per mm grating , giving a resolution of @xmath93 and a wavelength coverage of @xmath94 .
the spectra were centered at an average wavelength of @xmath95 , although the exact wavelength range for each spectrum depends on the slit position .
each @xmath96 hr exposure was broken into three subsets , with the objects stepped along the slit by @xmath97 in each direction .
the spectra were reduced in the same way as previous lris spectra ( cowie et al .
1996 ) . in the vast majority of cases , only spectra that could be confidently identified based on multiple emission and/or absorption lines were included .
however , we have assigned redshifts to thirteen sources based primarily on a single line and the continuum shape
. these sources may be less reliable than the other sources in the catalog .
we have noted these sources in table a1 .
we have only cross - identified x - ray sources with spectroscopic counterparts if the radial offsets are @xmath98 .
the only exceptions are the seven off - axis x - ray sources discussed in [ secimaging ] .
we have placed the optical positions of these off - axis sources ( sources 121 , 197 , 270 , 316 , 404 , 410 , and 479 ) at the bright galaxy nuclei .
[ we note that the redshift for object 316 was given incorrectly in barger et al .
( 2002 ; their object 218 ) to be @xmath99 ; the correct redshift is @xmath100 . ] with this selection , 284 of the 503 x - ray sources ( 56% ) have spectroscopic identifications , including 14 stars .
thirty - six of the sources have broad emission lines .
three of the redshifts ( sources 173 , 219 , and 378 ) are taken from the compilation of cohen et al .
( 2000 ) . in figure
[ fig3]a we show the @xmath40 magnitude distribution for all 503 sources ( _ open _ ) .
we indicate which of the sources have spectroscopic identifications with the light shaded region . of the 299 sources with @xmath4 , 261 ( 87% )
have been spectroscopically identified , and the bulk of those that have not been identified lie in the magnitude range @xmath101 . in figure
[ fig3]b we show the @xmath40 magnitude distribution for the 219 cdf - n x - ray sources that either were not observed spectroscopically ( _ open _ ) or were observed but could not be identified ( _ medium shading _ ) .
broadband galaxy colors have been used in recent years to obtain photometric redshift estimates for galaxies ( see bolzonella , miralles , & pell 2000 for a review ) .
x - ray sources may have complex spectral energy distributions that arise from both the host galaxy and the agn . consequently , estimating photometric redshifts for x - ray sources using standard galaxy templates may be difficult . however , we are most interested in obtaining photometric redshifts for obscured agns whose spectroscopic signatures are not easily visible , and for those sources , photometric redshift estimates based on standard galaxy templates may be expected to work well .
capak et al . ( 2003b ) estimated photometric redshifts for sources in the cdf - n region using the template fitting method and photometry in the @xmath51 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 , and @xmath43 bands .
this method should avoid the biases at high redshifts that are introduced by the training set method ( connolly et al .
1995 ) or the principal component analysis method ( cabanac , de lapparent , & hickson 2002 ) .
capak et al .
used their large spectroscopic database to compare results from two publicly available codes , hyperz ( bolzonella et al .
2000 ) and bpz ( bayesian photometric redshift estimation ; bentez 2000 ) .
the major difference between the codes is the inclusion of a weighting function in bpz that reduces the number of degenerate fits at different redshifts .
additional differences are that hyperz includes a range of templates from the updated bruzual & charlot evolutionary code ( gissel98 ; bruzual & charlot 1993 ) , as well as the coleman , wu , & weedman ( 1980 ; hereafter , cww ) templates , while bpz only uses the templates of cww and kinney et al . (
1996 ) ; however , bpz does interpolate between the templates to produce intermediate templates . also ,
an intrinsic reddening term is included in hyperz but not in bpz , and the two codes treat non - detections differently . for most of the galaxies in the @xmath102 range , capak et al .
( 2003b ) found that the two codes produced similar results ; however , hyperz confused a larger number of sources at @xmath103 with @xmath104 sources , which is not surprising since the weighting function in bpz was introduced to solve this problem .
capak et al .
( 2003b ) were able to improve the results of both codes by tuning the photometry .
they fitted templates at the known redshift for each source and then calculated the mean offset between their photometry and the template photometry .
iterating on this process produced small photometric offsets ( 0.049 in @xmath51 , 0.022 in @xmath39 , 0.069 in @xmath40 , @xmath105 in @xmath42 , and @xmath106 in @xmath43 ) that significantly improved the results ( see capak et al .
2003b for details ) ; the @xmath38 and @xmath41 band magnitudes were fixed since the authors were confident in the photometry in those bands .
the offsets are likely due to differences between real galaxies and the templates and between the real instrument filter profiles and the assumed profiles .
the bpz code outputs the probability @xmath107 of @xmath108 , where @xmath109 is the galaxy redshift and @xmath110 is the `` best '' redshift estimate .
when the value of @xmath107 is low , the redshift probability is spread over a large range in redshift , and the prediction is likely not reliable , so we applied a @xmath111 threshold to remove galaxies with catastrophic redshift errors ( see capak et al .
2003b for more details ) .
we also applied a @xmath112 threshold to eliminate saturated sources .
this is slightly fainter than the saturation limit in @xmath40 to also remove sources that may only be saturated in other bands .
such sources would not have good photometric redshift estimates , but since they are already spectroscopically identified , photometric redshift estimates are not needed . in figure [ fig4 ]
we plot the ratio of the photometric redshift estimate from bpz to the spectroscopic redshift versus the spectroscopic redshift for the 193 remaining x - ray sources ( 172 non - broad - line and 21 broad - line sources ) with both measurements .
we plot the redshift ratio because what we are interested in is how accurate our distances are .
this means that photometric redshift errors at low redshifts are accentuated .
some of the discrepant objects at low redshifts have complex optical structures near the x - ray source positions that contaminate the photometry .
the broad - line agns with both spectroscopic and photometric redshifts are denoted by open squares .
the photometric redshift method fails for many of them , but broad - line agns are straightforward to identify spectroscopically , even in the redshift range @xmath113 , and we now have observed the bulk of the sources in the cdf - n x - ray sample ( see figure [ fig3]b ) , so we do not need to worry that broad - line agns will contaminate our photometric redshift sample .
the photometric redshifts for the non - broad - line sources are clearly quite robust , with 94% having photometric redshifts within 25% ( _ horizontal dashed lines _ ) of their spectroscopic redshifts .
only one very low redshift source has a badly discrepant photometric redshift ( spectroscopic redshift of @xmath114 and photometric redshift of 2.58 ) and hence is not within the vertical range of the plot .
we have not attempted to refine the photometric redshifts with spectral information ( e.g. , there may be cases where we had insufficient spectral information to reliably identify the source spectroscopically but that information could have been used to try to improve the photometric redshifts ) , since for most purposes the photometric redshifts are accurate enough .
the bpz photometric redshifts presented in this paper are an improvement over the photometric redshift estimates made with hyperz for the 1 ms cdf - n x - ray sources ( barger et al .
in figure [ fig5 ] we show redshift versus hard x - ray flux for the spectroscopically ( _ solid squares _ ) and photometrically ( _ open triangles _ ) identified cdf - n hard x - ray sources .
to avoid incompleteness at the faint end when determining the fractional light contribution from each redshift interval to the measured hard x - ray light , we consider two restricted uniform flux - limited hard x - ray subsamples , which we refer to as our `` bright '' and `` deep '' subsamples . for the bright subsample we consider a @xmath16 radius region around the approximate center of the x - ray image and select sources detected with fluxes above @xmath115 ergs @xmath37 s@xmath3 . [ we note , however , that above @xmath116 ergs @xmath37 s@xmath3 the number density of sources in the cdf - n becomes too low for an accurate measurement . ] for the deep subsample we consider a @xmath117 radius high image quality and high exposure time region and select sources detected with fluxes between @xmath118 and @xmath115 ergs @xmath37 s@xmath3 . the numbers at the right of figure [ fig5 ] are the percentages of the measured hard x - ray light from the combined deep and bright subsamples that come from only spectroscopically identified sources in each redshift interval .
the spectroscopically identified sources already comprise 75% of the measured hard x - ray light , of which at least 54% arises at redshifts below @xmath119 and at least 68% below @xmath120 .
these percentages increase to 58% and 76% , respectively , if we include the sources with photometric redshifts , and they rise even higher if the bright x - ray sources found in deep _ asca _ surveys ( akiyama et al .
2000 ) are considered ( see figure 14a of barger et al .
thus , an impressive amount of the @xmath6 kev extragalactic background light arises at recent times , as was first shown by barger et al .
( 2001a ) . in figure [ fig6 ]
we show the redshift - magnitude relation for the cdf - n x - ray sources ; sources without any redshift information are plotted below @xmath121 .
broad - line agns ( _ large solid squares _ ) are systematically the most optically luminous of the x - ray sources because of their agn contribution to the light .
the sources with photometric redshifts show a smooth continuation of the spectroscopic trend towards fainter @xmath40 magnitudes with increasing redshift . in figure [ fig7 ]
we show the @xmath6 kev flux distribution for the cdf - n x - ray sources . labeled arrows at the top of the figure show the flux of a source with a rest - frame @xmath6 kev luminosity of @xmath122 ergs s@xmath3 at a redshift of @xmath119 , 2 , or 3 , computed using the @xmath6 kev flux and a @xmath46-correction for an intrinsic @xmath123 power - law spectrum , where @xmath124 is the photon index ( see barger et al .
2002 for the reason behind this choice of @xmath124 ) .
the arrow at @xmath125 also roughly corresponds to the @xmath6 kev flux limit of @xmath126 ergs @xmath37 s@xmath3 where the fitted x - ray number counts give @xmath127% of the @xmath6 kev light ( cowie et al .
2002 ) . above the flux limit where roughly 80% of the light arises
, there are 150 sources . of these ,
the 113 with spectroscopic or photometric identifications account for most of the light ; only 15% of the light is not accounted for by the identified sources .
sources with rest - frame @xmath6 kev luminosities greater than @xmath122 ergs s@xmath3 ( often called quasars ) contribute 39% of the light , while fainter sources contribute 45% .
of the 25 sources in the quasar luminosity range , 15 are broad - line agns , and these 15 contribute 26% of the light .
six of the sources in the quasar luminosity range ( sources 89 , 240 , 259 , 390 , 420 , and 495 ) are type ii agns , based on the presence of [ neiii ] , [ nev ] , or civ emission lines in the spectra , though in many cases these lines are weak features superimposed on a much brighter galaxy spectrum . in one of these sources ( source 420 )
the civ line shows a p - cygni profile .
these six type ii quasars contribute only 11% of the light and hence are not the dominant contributors to the x - ray background at these energies .
one additional source in the quasar luminosity range ( source 398 ) is a ly@xmath128 emitter with absorption at civ , while the remaining three ( sources 108 , 165 , and 463 ) have only photometric redshifts . in figure
[ fig8]a we show the spectroscopic ( _ open _ ) and photometric ( _ hatched _ ) redshift distributions for all the cdf - n x - ray sources with redshift information using a low resolution ( @xmath129 ) binning . the gap in the spectroscopic redshift distribution between @xmath130 and 2 reflects the difficulty of identifying galaxies with redshifts in this range , where [ oii ] 3727 has moved out of the optical window and ly@xmath128 1216 has not yet entered in .
the two broad redshift spikes seen in the 1 ms exposure ( barger et al .
2002 ) have grown substantially in the 2 ms exposure .
fifty - four spectroscopically identified sources lie between @xmath131 and @xmath132 , with a median redshift of @xmath133 , and 79 lie between @xmath134 and @xmath135 , with a median redshift of @xmath136 .
thus , these two redshift intervals contain nearly half of the identified galaxies .
it is hard to quantify the significance of these peaks without knowing the true redshift distribution .
if the galaxies between @xmath121 and @xmath137 were uniformly distributed , then we would expect 44 galaxies in the lower redshift bin and 55 in the higher .
the @xmath138 peak could simply be the maximum in the redshift distribution , but it appears likely that the @xmath139 peak is an unique feature of the cdf - n region . in figures [ fig8]b and [ fig8]c we show the separate redshift distributions for the soft and hard samples , respectively , including only sources within @xmath140 of the approximate x - ray image center with @xmath15 kev fluxes above @xmath141 ergs @xmath37 s@xmath3 ( figure [ fig8]b ) and @xmath6 kev fluxes above @xmath126 ergs @xmath37 s@xmath3 ( figure [ fig8]c ) .
these limiting fluxes were chosen to match roughly the value where the fitted x - ray number counts give @xmath127% of the light ( cowie et al .
the solid arrows show the median redshifts of all the sources with spectroscopic and photometric redshifts in each sample , and the dashed arrows show the median redshifts after all the sources without any redshift information have been arbitrarily placed at high redshifts .
the dashed arrows are only shifted to slightly higher redshifts than the solid arrows because most of the sources to these flux levels have already been identified . in figure [ fig9 ]
we show the redshift distribution for all the spectroscopically identified x - ray sources between @xmath121 and @xmath120 using a higher resolution ( @xmath142 ) binning .
the strongest single feature in the distribution is at @xmath143 , where 16 galaxies lie within a thousand km s@xmath3 of the redshift .
( these are split between two bins in fig .
[ fig9 ] but do not contain all of the galaxies in each of the bins . ) six of these sources lie in a tight spatial clump ( source 116 is the brightest galaxy in this clump ) . because of the apparent velocity sheets and complex structures in the redshift distributions of the x - ray sources in individual _
fields , we will need to observe a large number of fields in order to get the average true redshift distribution .
in figure [ fig10]a we show @xmath144 color versus @xmath41 magnitude for the x - ray sources with both x - ray positional uncertainties and optical counterpart separations @xmath145 .
( the tight positional selection is used in order to keep the field contamination low ; see [ secimaging ] . )
median @xmath144 colors for the non - broad - line sources in each magnitude bin are shown as large open diamonds .
the 68% confidence ranges in the medians were computed using the number of sources in each bin ( gehrels 1986 ) .
the median @xmath144 color for all the broad - line sources is shown as a dashed line .
the figure shows a general trend of the non - broad line sources towards bluer colors at fainter magnitudes , with the colors of the @xmath146 sources approaching the median blue color of the broad - line agns .
schreier et al . ( 2001 ) , and subsequently koekemoer et al . ( 2002 ) , also noted this color trend with magnitude from _ hst _ data on the cdf - s .
they interpreted it as meaning that there are two distinct populations of sources , an optically bright red population that consists of normal galaxies at @xmath7 , and an optically faint blue population that consists of type 2 agns of low to moderate luminosity located at @xmath147 .
our interpretation is instead that the optically faint blue population consists of normal galaxies at @xmath1 .
the galaxies simply appear blue in the optical because the 4000 break has moved beyond the @xmath41 band .
we illustrate this in figure [ fig10]b with a @xmath144 versus @xmath148 plot . here
the solid ( open ) squares denote the @xmath149 ( @xmath150 ) sources from figure [ fig10]a .
color - color tracks for an evolved elliptical galaxy ( _ upper track _ ) , sb galaxy ( _ middle track _ ) , and irregular galaxy ( _ bottom track _ ) from cww are plotted as solid lines for @xmath7 and as dashed lines for @xmath1 .
the figure shows that @xmath119 is a natural dividing line between the optically bright x - ray sources , which are redder in @xmath144 , and the optically faint x - ray sources , which are bluer in @xmath144 . finally , in figure
[ fig10]c we plot @xmath144 versus redshift for the sources with spectroscopic ( _ solid squares _ ) or photometric ( _ open triangles _ ) redshifts to show that we have both photometric and some spectroscopic redshift confirmation for our interpretation . the sources with optically faint magnitudes @xmath150 are denoted by the second , larger symbols .
in fact , x - ray sources may be effective markers of _ red _ galaxies at high redshifts , as suggested ( and spectroscopically confirmed for one source at @xmath151 from near - infrared spectroscopy ) by cowie et al .
in figure [ fig11 ] we plot @xmath9 color versus redshift for the cdf - n x - ray sources with ( a ) non - broad - line spectroscopic ( _ solid squares _ ) or photometric ( _ open triangles _ ) redshifts , and ( b ) broad - lines ( _ large solid squares _ ) .
the overlays are cww galaxy redshift tracks for an evolved elliptical galaxy ( _ solid curve _ ) , sb galaxy ( _ dashed curve _ ) , and irregular galaxy ( _ dotted curve _ ) .
sources with photometric redshifts smoothly extend the spectroscopically identified source population to higher redshifts and redder colors .
the upper envelope of the photometrically identified sources tracks the elliptical galaxy curve , and at the higher redshifts , many of the galaxies fall into the extremely red object ( ero ) color range ( see also alexander et al .
2002b and mainieri et al .
2002 ) . here
we conservatively define eros as having @xmath10 , which is equivalent to @xmath152 ( @xmath50 ; barger et al .
1999 ) . just as for field eros
, these x - ray selected eros could be explained by either old stellar populations at @xmath1 or by dust - enshrouded galaxies .
most of the eros in the cdf - n sample ( _ circles _ in figure [ fig2 ] ) have high x - ray - to - optical flux ratios because these sources are faint in the optical , and perhaps this signature could be used to pick out the higher redshift x - ray sources . although we do not have spectroscopic redshifts for any of the sources redder than an sb galaxy , and consequently our photometric redshift estimates are unconfirmed , spectroscopic identifications of high - redshift ( @xmath153 ) red galaxies have recently been made , and the sources have been interpreted as the evolved descendants of galaxies that started forming stars at redshifts @xmath154 ( franx et al .
2003 ; van dokkum et al .
. it would be even more interesting to see how the _ rest - frame _ colors of the x - ray counterparts compare , and whether there are any intrinsic differences in the source properties between the brighter low - redshift sources and the fainter high - redshift ones . with our wide wavelength coverage , we are able to determine the rest - frame colors for the cdf - n x - ray sources with redshift identifications . in figure
[ fig12]a we show rest - frame ab @xmath155 color versus redshift for the x - ray sources with either spectroscopic or photometric redshifts .
beyond @xmath11 the colors are extrapolated and may be more uncertain .
the horizontal lines show the colors of an elliptical galaxy ( _ solid _ ) and an sb galaxy ( _ dashed _ ) .
the figure shows that many of the x - ray sources have the colors of evolved red galaxies and that there does not appear to be much evolution in the intrinsic colors of the galaxies with redshift .
this is a rather surprising result , since passive evolution of the host galaxies would be expected to make the x - ray sources at high redshifts considerably bluer . in figure
[ fig12]b we show the ab @xmath155 color distribution for the @xmath156 non - broad - line sources ( _ open _ ) .
the vertical lines show the rest - frame colors of an elliptical galaxy ( _ solid _ ) and an sb galaxy ( _ dashed _ ) .
here we can more clearly see that a very large fraction of the x - ray population has rest - frame colors within a narrow red color range .
in contrast , the @xmath156 broad - line sources ( _ hatched _ ) are very blue . this latter result is expected , since the presence of a central agn should make the rest - frame colors bluer than those of evolved galaxies .
the agn contribution to the light is particularly strong in the ultraviolet and hence should increase the 2800 flux relative to the 8140 flux .
however , this effect does not seem to be occurring in the non - broad - line sources .
in fact , the number of sources with intermediate colors is rather small .
the situation therefore seems to be that either the agns dominate the colors , as in the case of the broad - line sources , or the agns do not have much effect on the colors , leaving the sources to have the colors of evolved red galaxies .
we can examine the evolution of the rest - frame colors of the x - ray sources with either spectroscopic or photometric redshifts more quantitatively by dividing the sample into two redshift bins , @xmath157 and @xmath158 .
in figure [ fig13 ] we show the rest - frame ab @xmath155 color distributions .
there are 212 sources in the @xmath157 ( _ solid _ ) interval and 95 in the @xmath158 ( _ dashed _ ) interval .
a kolmogorov - smirnov test shows that the maximum deviation of 0.14 is consistent with the two distributions being the same with a 95% confidence criterion .
the median color of the @xmath157 sources is 3.0 , and the median color of the @xmath158 sources is 3.3 .
one of the most striking features of figure [ fig10]b is that there is a clear color - color separation as a function of apparent magnitude . we might ask why apparent magnitude seems to be a reasonably good redshift diagnostic for the x - ray sources .
the most likely answer is that most of the host galaxies of the x - ray sources have similar absolute magnitudes , and the x - ray sources lie in a rather uniform host galaxy population .
we examine this possibility in figure [ fig14 ] , where we plot absolute ab 8140 magnitude versus redshift for the spectroscopically identified non - broad - line sources ( _ solid squares _ ) , photometrically identified sources ( _ open triangles _ ) , and broad - line agns ( _ large solid squares _ ) .
the bottom envelope is just a selection effect against very faint optical magnitudes .
the dashed line shows the @xmath159 galaxy from loveday et al .
( 1992 ) , which roughly corresponds to an sb galaxy with @xmath160 . from figure [ fig14 ]
we see that many of the non - broad - line sources are very luminous ( @xmath161 ) , even at high redshifts .
thus , x - ray surveys may provide an effective way to find galaxies more luminous than @xmath14 with colors similar to present - day early - type galaxies to high redshifts , though at present this interpretation is based on photometric redshift estimates that currently have relatively little confirmation at these redshifts .
1 . we presented a catalog of optical and near - infrared magnitudes and optical spectroscopy for the x - ray sources identified in the @xmath24 ms _
chandra _ exposure of the hubble deep field - north region .
we now have redshifts for 284 of the 503 x - ray point sources .
the redshift identifications are very complete ( 87% ) for the @xmath4 sources . 2 .
we also presented photometric redshifts , obtained using seven broadband colors and a bayesian photometric redshift estimation code , for 78 spectroscopically unidentified sources .
we expect these photometric redshift estimates to be reliable since we found that the photometric redshifts of all but 11 of the 172 spectroscopically identified non - broad - line sources with photometric redshift estimates are within 25% of the spectroscopic redshifts .
since broad - line agns are straightforward to identify spectroscopically , and we have now observed nearly all the x - ray sources , it is unlikely that broad - line agns are contaminating our photometric redshifts . 3 .
we showed that x - ray - to - optical flux ratios continue to be a good discriminator between x - ray source classes down to very faint optical magnitudes and x - ray fluxes , with agns typically falling within the region defined by the loci @xmath88 and quiescent galaxies lying at @xmath162 .
we constructed redshift slices of the hard x - ray sources versus hard x - ray flux and found that the spectroscopically identified sources already comprise 75% of the measured light , of which 54% arises from @xmath7 sources and 68% from @xmath8 sources .
our spectroscopic evidence for two broad redshift spikes in the x - ray source population discovered by barger et al .
( 2002 ) has continued to grow , illustrating that the cdf - n is dominated by a rather small number of independent velocity elements . because of the apparent velocity sheets and complex structures in the redshift distributions of the x - ray sources in individual _
chandra _ fields , we will need to observe a large number of fields in order to get the average true redshift distribution .
we found that x - ray sources may be effective markers of red galaxies at high redshifts . in an @xmath9 versus redshift plot
, the sources with photometric redshifts smoothly extend the spectroscopically identified source population to higher redshifts and redder colors .
the upper envelope of the photometrically identified sources follows that of an elliptical galaxy track , and fourteen of these sources have @xmath9 colors that would classify them as eros with redshifts between @xmath11 and @xmath12 . 6 . with our wide wavelength coverage , we were also able to determine rest - frame colors for the x - ray sources with redshift identifications .
we found that many of the non - broad - line x - ray sources have intrinsic colors of evolved bulge - dominated galaxies and that there is very little evolution in these colors with redshift , as confirmed with a kolmogorov - smirnov test between two redshift bins , @xmath157 and @xmath158 .
if anything , the galaxies in the higher redshift interval are redder than those in the lower redshift interval .
this is surprising both because passive evolution of the host galaxies would be expected to make the x - ray sources at higher redshifts bluer and because any agn contributions to the light would make the sources bluer .
thus , it seems that either the agn light dominates the colors , as is the case for the broad - line sources , or the agn presence does not have much effect on the colors .
we showed that apparent magnitude seems to be a reasonably good redshift diagnostic for the non - broad - line x - ray sources , with there being a clear separation at a redshift of @xmath163 between optically bright and optically faint sources in a @xmath144 versus @xmath148 color - color plot .
we suggest that this results from the fact that the x - ray sources lie in a rather uniform host galaxy population with similar absolute magnitudes .
we found that many of the non - broad - line sources are very luminous ( @xmath161 ) , even at high redshifts .
we therefore infer that x - ray surveys may provide an effective way to find @xmath14 galaxies with colors similar to present - day early - type galaxies to high redshifts , though at present this interpretation is based on photometric redshift estimates that currently have relatively little confirmation at these redshift values .
we thank the referee , david helfand , for helpful comments that improved the manuscript .
we are grateful to herv aussel for providing supplemental optical spectra .
we gratefully acknowledge support from the university of wisconsin research committee with funds granted by the wisconsin alumni research foundation ( a. j. b. ) , the alfred p. sloan foundation ( a. j. b. ) , nsf grants ast-0084847 ( a. j.
b ) , ast-0084816 ( l. l. c. ) , and ast 99 - 83783 ( d. m. a. , w. n. b. ) , and nasa grants df1 - 2001x ( l. l. c. ) , go2 - 3187b ( l. l. c. ) , nas 8 - 01128 ( g. p. g. ) , and g02 - 3187a
( d. m. a. , f. e. b. , w. n. b. ) . in table a1 we present optical magnitudes and spectroscopic measurements , where available , for the full x - ray point source catalog of a03 .
the 503 x - ray sources are identified by a03 number ( col .
[ 1 ] ) and by right ascension and declination coordinates in decimal degrees ( cols . [ 2 ] and [ 3 ] ) .
we also give right ascension and declination coordinates where the aperture magnitudes were measured ( cols .
[ 4 ] and [ 5 ] ) . in columns
@xmath164 we give the soft and hard x - ray fluxes from a03 in units of @xmath165 ergs @xmath37 s@xmath3 . in columns
@xmath166 we give isophotal @xmath40 magnitudes , @xmath43 , @xmath42 , @xmath41 , @xmath40 , @xmath39 , @xmath38 , and @xmath51 corrected - aperture magnitudes , spectroscopic redshifts , and photometric redshifts .
objects where a spectrum was obtained but no identification could be made are marked `` obs '' for observed in the spectroscopic redshift column ( col .
the final column contains a `` b '' if the spectrum has broad emission lines , an `` s '' if the spectroscopic identification is at all suspect , and an `` m '' if the source is complex or if a bright nearby neighbor may be contaminating the photometry .
source 267 is more likely associated with an extremely red object ( dickinson et al .
2000 ) to the northeast of the @xmath167 galaxy identified here , but the adopted automatic procedure gives the @xmath167 galaxy as the counterpart .
we note that v. l. sarajedini ( priv .
comm . ) has identified the @xmath167 galaxy as variable , which may suggest that some part of the x - ray emission may also be coming from this object .
@xmath168 source 91 : the new x - ray position has moved the source south of the southern component of this merging pair , whose @xmath169 redshift was given as the redshift for source 71 in barger et al .
the x - ray source now lies on the tidal tail .
it appears likely that the @xmath169 redshift is still correct , but the positional offset means that it no longer meets the formal criteria for inclusion in the table .
@xmath168 sources 119 and 120 : source 92 of barger et al .
has now split into these two sources .
source 119 has the redshift of the barger et al .
source 92 ( @xmath170 ) , while source 120 has a redshift of @xmath171 .
@xmath168 source 259 : in barger et al .
the redshift for this source ( their source 184 ) was given as @xmath172 .
newer spectra show this redshift was based on an incorrect identification of an emission line as ly@xmath128 , when it is really civ .
the present identification is based on [ oii ] 3727 , mgii 2800 , ciii ] 1909 , and civ 1550 and appears unambiguous .
@xmath168 source 443 : this source has two possible counterparts , a brighter one at @xmath173 and a fainter one at @xmath174 .
we have adopted the @xmath174 redshift of the closer source instead of the @xmath173 redshift given in barger et al .
( their source 331 ) .
the higher resolution images were smoothed to match the seeing in the lower resolution images in order to give the most precise color measurements .
the fluxes for each source were measured using @xmath79 diameter apertures on the smoothed images .
the suprime - cam photometry was scaled to match the hdf - n proper photometry of fernndez - soto , lanzetta , & yahil ( 1999 ) , both for absolute calibration and to approximate total magnitudes ( the fernndez - soto et al .
measurements are isophotal ) .
the @xmath43 magnitudes were corrected to approximate total magnitudes using an offset calculated as the median difference between @xmath79 and @xmath175 aperture magnitudes measured for a complete @xmath176 sample in the field .
we give the @xmath38 and @xmath39 magnitudes in the johnson system , the @xmath40 and @xmath41 magnitudes in the kron - cousins system , the @xmath43 magnitudes in the wainscoat - cowie system ( wainscoat & cowie 1992 ) , and the @xmath51 and @xmath42 magnitudes in the ab magnitude system .
offsets of @xmath177 , 0.023 , 0.228 , 0.453 , and 1.595 for the @xmath38 , @xmath39 , @xmath40 , @xmath41 , and @xmath43 bands , respectively , can be added to the magnitudes in table a1 to get ab magnitudes .
saturation occurs in the images at magnitudes of about 16 ( @xmath51 ) , 20.6 ( @xmath38 ) , 19.8 ( @xmath39 ) , 20 ( @xmath40 ) , 19.7 ( @xmath41 ) , 19.4 ( @xmath42 ) , and 14.4 ( @xmath43 ) . magnitudes brighter than these values in table a1 are likely to be underestimated .
sources that had a negative flux in the aperture are listed with a negative magnitude ; in these cases , the absolute value of the magnitude was computed from the absolute value of the flux .
if no magnitude is given , then this indicates that the magnitude was severely contaminated , usually by a nearby bright star .
we show @xmath42 band thumbnail images of the cdf - n sources in figure a1 , ordered by a03 catalog number .
we use the @xmath42 band because many of the sources are red and hence more easily seen in this band .
each thumbnail is @xmath178 on a side .
the expected x - ray position of each source is shown with a @xmath71 diameter circle ( representing the @xmath61 search radius ) , and the position where the @xmath79 diameter aperture magnitude was measured is shown with a @xmath79 diameter circle .
the number of the source from table a1 is printed in the top left corner .
if there is a spectroscopic identification , then the redshift ( or the word `` star '' for a stellar spectrum ) is printed in the bottom left corner .
if a nearby faint source and a more distant bright source with a spectroscopic redshift are both found to lie within the @xmath61 search radius , then we have identified the x - ray source with the faint source and zeroed out the spectroscopic redshift .
if there is a spectroscopic identification for a source that lies outside our chosen cross - identification radius of @xmath61 but still within a @xmath68 radius about the x - ray position , then the redshift of that source is printed in fainter text to indicate that the redshift has been given for interest but has not been used .
the only exceptions are the seven off - axis x - ray sources discussed in [ secimaging ] . if there is no spectroscopic redshift for a source , but there is a photometric redshift estimate , then the photometric redshift estimate is printed in parentheses . | we present an optical and near - infrared catalog for the x - ray sources in the @xmath0 ms _ chandra _ observation of the hubble deep field - north region .
we have high - quality multicolor imaging data for all 503 x - ray point sources in the x - ray - selected catalog and reliable spectroscopic redshifts for 284 .
we spectroscopically identify six high - redshift ( @xmath1 ) type ii quasars ( @xmath2 ergs s@xmath3 ) in our sample .
our spectroscopic completeness for the @xmath4 sources is 87% .
the spectroscopic redshift distribution shows two broad redshift spikes that have clearly grown over those originally seen in the @xmath5 ms exposure .
the spectroscopically identified extragalactic sources already comprise 75% of the measured @xmath6 kev light .
redshift slices versus @xmath6 kev flux show that an impressive 54% of the measured @xmath6 kev light arises from sources at @xmath7 and 68% from sources at @xmath8 .
thus , major accretion onto supermassive black holes has occurred since the universe was half its present age .
we use seven broadband colors and a bayesian photometric redshift estimation code to obtain photometric redshifts for the x - ray sources .
we find that the photometric redshifts are within 25% of the spectroscopic redshifts for 94% of the non - broad - line sources with both photometric and spectroscopic measurements .
the photometrically identified sources show a smooth continuation of the spectroscopically identified sources to redder @xmath9 color with increasing redshift , consistent with the galaxy tracks of evolved bulge - dominated galaxies .
fourteen have colors @xmath10 that would classify them as extremely red objects ( eros ) .
the photometric redshifts of these eros are all between @xmath11 and @xmath12 .
we use our wide wavelength coverage to determine rest - frame colors for the x - ray sources with spectroscopic or photometric redshifts .
we find that many of the x - ray sources have the rest - frame colors of evolved red galaxies and that there is very little evolution in these colors with redshift .
we also determine absolute magnitudes and find that many of the non - broad - line sources are more luminous than @xmath13 , even at high redshifts .
we therefore infer that deep x - ray observations may provide an effective way of locating @xmath14 galaxies with colors similar to present - day early - type galaxies to high redshifts . |
inflation has become the standard paradigm for explaining the homogeneity and the isotropy of our observed universe @xcite . during this epoch the energy density of the universe was dominated by some scalar field ( the inflaton ) , with negligible kinetic energy density , such that the corresponding vacuum energy density was the responsible for the exponential growth of the scale factor of the universe . during this phase a small and smooth region of the order of size of the hubble radius grew so large that it easily encompassed the comoving volume of the entire presently observed universe .
this is the reason for which the observed universe is so homogeneous and isotropic .
furthermore , it is now clear that structure in the universe comes primarily from an almost scale - invariant super - horizon curvature perturbation .
this perturbation originates presumably from the vacuum fluctuation , during the almost - exponential inflation , of some field with mass much less than the hubble parameter @xmath2 .
indeed , any scalar field whose mass is lighter than @xmath2 suffered , in a ( quasi ) de sitter epoch , with a scale independent quantum fluctuations spectrum@xcite . during inflation ,
particle production can only occur for particles that are light compared to the hubble scale without classical conformal invariance ; gravitons and massless minimally coupled scalars and light fermions are unique in that respect .
particle creation during inflation was studied many years ago in the framework of standard inflation@xcite .
the process appeared to be straightforward : in models like new inflation and chaotic inflation@xcite .
these models incorporate a second order phase transition to end inflation , the inflaton field would wind up oscillating around the minimum of its potential near the end of inflation .
these oscillations would produce a sea of relativistic particles , if one added ( by hand ) interaction terms between the inflaton and these lighter species .
the direct production from vacuum fluctuations during inflation of x - particle was considered in@xcite .
particle creation has been also considered in warm inflationary@xcite and fresh inflationary@xcite scenarios .
more recently , particle creation during inflation was considered in a model of brane inflation where two stacks of mobile branes are moving ultra relativistically in a warped throat@xcite .
after inflation , the universe could have suffered a phase transition from a de sitter vacuum dominated state to a decelerated radiation dominated stage . at this moment , a great amount of the potential density energy is transferred to radiation energy density due to the decay of the inflaton field produced by its interaction with other boson fields@xcite .
some years ago kaiser@xcite demonstrated that particles produced from the parametric resonance effect when @xmath3 is fewer than in the @xmath4 case , but can still be exponentially greater than when the resonance is neglected altogether . on the other hand , in a previous work@xcite we studied the scalar metric fluctuations of a 5d spacetime background metric @xmath5 , which is riemann - flat and hence ricci - flat : @xmath6 . from the mathematical point of view
, the campbell - magaard theorem@xcite serves as a ladder to go between manifolds whose dimensionality differs by one .
this theorem , which is valid in any number of dimensions , implies that every solution of the 4d einstein equations with arbitrary energy momentum tensor can be embedded , at least locally , in a solution of the 5d einstein field equations in vacuum .
because of this , the stress - energy may be a 4d manifestation of the embedding geometry .
physically , the background metric there employed describes a 5d extension of an usual de sitter spacetime .
inflationary cosmology can be recovered from a 5d vacuum@xcite . other version of 5d gravity , which is mathematically similar , is the membrane theory . in this theory gravity propagates freely on the 5d bulk and the interactions of particles are confined to a 4d hypersurface called `` brane''@xcite
. both versions of 5d general relativity are in agreement with observations . in this work we study both ,
particle and graviton production after a phase transition occurred after inflation . to make it
, we shall consider the solutions obtained at the end of inflation from a 5d vacuum on an effective 4d hypersurface described by a de sitter metric .
after it , we shall use the bogoliubov formalism to calculate the modes solution for the fields after a phase transition to a radiation dominated epoch .
the topic here studied is very important , because the case of inflaton decay into further inflatons may be of interest for dark matter searches .
following their production , the inflatons would decouple from the rest of matter@xcite .
if the inflatons were given a tiny mass , then these bosons could serve as a natural candidate for the missing dark matter@xcite .
we consider a 5d manifold ( @xmath7 ) with a coordinate system @xmath8 .
we are interested in a 5d theory of gravity on which we define a 5d vacuum , such that the first variation of the action is @xmath9 , where @xmath10 the first term is the variation of the gravitational einstein action @xmath11 and the second one is the variation of the matter action @xmath12 . here
, @xmath13 is the gravitational constant , @xmath14 is the determinant of the covariant tensor metric @xmath15 run from @xmath16 to @xmath17 and greek letters rum from @xmath16 to @xmath18 . ] and @xmath19 is the 5d ricci scalar on the metric .
the energy - momentum tensor of matter , @xmath20 is defined from the variation of the matter action @xmath12 under a change of the metric , and will be considered as null to describe the 5d apparent vacuum .
the einstein equations on the 5d manifold @xmath21 ( we use @xmath22 units ) @xmath23 where @xmath24 and @xmath25 are respectively the einstein and energy momentum tensors on @xmath21 . furthermore , we can define a scalar function @xmath26 , which represents the foliation of the higher - dimensional manifold .
we shall consider that the extra coordinate is space - like .
each hypersurface , @xmath27 , is considered as a 4d lorentzian spacetime .
we denote @xmath28 as the normal vector to the hypersurface @xmath27 and the covariant derivative on @xmath27 with a semicolon : @xmath29 . ]
@xmath30 such that the expression @xmath31 normalizes @xmath32 .
now we can define a coordinate system @xmath33 on @xmath27 .
the basis vectors are @xmath34 these objects can be used to project 5d tensors ( for instance , the metric tensor ) into 4d ones ( which lives on the @xmath27 hypersurface ) @xmath35 the extrinsic curvature @xmath36 of the 4d hypersurface @xmath27 is a symmetric 2-range tensor given by the derivative of the induced metric in the normal direction to @xmath27 @xmath37 if now we consider an alternative coordinate system on @xmath38 , such that @xmath39 where the vector tangent to lines with constant @xmath40 can be decomposed into the sum of a part tangent to @xmath27 , and a part normal to @xmath27 @xmath41 such that @xmath42 .
the 4d vector @xmath43 is the shift vector , which describes how the @xmath40 coordinate system changes as we move from a given @xmath27 hypersurface to another .
finally , the 5d line element @xmath44 can be written as @xmath45 which , for a constant foliation @xmath46 , reduces to @xmath47 in this work we are interested in dealing with a 5d riemann - flat metric , so that the ricci tensor is null @xmath48 on each @xmath27 hypersurface , the gauss - codazzi equations are\beta}$ ] , and @xmath49}=0 $ ] . ]
@xmath50\beta } , \\ & & \hat{r}_{mabc } n^m e^a_{\alpha } e^b_{\beta } e^c_{\gamma } = 2 k_{\alpha [ \beta ; \gamma]}.\label{ub}\end{aligned}\ ] ] if we use the expression for the 5d ricci tensor @xmath51 we obtain the contractions of ( [ ua ] ) @xmath52 using ( [ uc ] ) into ( [ uu ] ) , and making use of ( [ ub ] ) , we obtain the expressions @xmath53 where @xmath54 , @xmath55 and @xmath56 . notice that the equation ( [ eea ] ) means that the second rank ( symmetric ) tensor @xmath57 is conserved on the 4d hypersurface @xmath27 .
it is important to notice that the ricci tensor on @xmath27 , related to the 5d riemann flat metric @xmath58 , is given by @xmath59 finally , the einstein tensor @xmath60 on a given @xmath27 , is @xmath61 notice that for a 4d metric induced from a 5d riemann - flat metric , the einstein tensor only depends on the extrinsic curvature @xmath62 in this work , we are interested in describing the primordial universe on a 4d hypersurface @xmath27 . in order to make it , we shall consider a massless single scalar field @xmath63 , which , we shall consider to describe the physical vacuum on the 5d manifold @xmath21 .
the kinetic lagrangian corresponding to this field is @xmath64 where @xmath65 is the determinant of the covariant metric tensor and @xmath66 is a dimensional constant .
hence , to describe the energy momentum tensor related to @xmath67 , on @xmath21 , we set @xmath68.\ ] ] it is easy to demonstrate that the @xmath69 projected on the hypersurface @xmath27 is @xmath70 finally , using ( [ em ] ) and ( [ ein ] ) , we obtain the einstein equations on @xmath27 obtained from a riemann - flat 5d vacuum @xmath71\right\}\right|_l,\ ] ] that describes the equations of motion for the fields on hypersurfaces with constant @xmath72 ( i.e. , for constant foliations ) .
notice that the equations of motion obtained from a ricci - flat 5d metric must include the einstein tensor ( [ ein1 ] ) , rather than ( [ ein ] ) , in the left side of the einstein equations ( [ eqq ] ) .
we consider the riemann - flat background metric@xcite @xmath73 where @xmath74 . here , @xmath75 are the 3d cartesian space - like dimensionless coordinates , @xmath76 is a dimensionless time - like coordinate and @xmath72 is the space - like non - compact extra coordinate , which has length units .
the non - perturbative metric fluctuations of the background metric ( [ s1 ] ) , are introduced in our analysis by the line element introduced in@xcite @xmath77 where the metric function @xmath78 describes the gauge - invariant metric fluctuations with respect to the background metric . in order to describe a 5d physical vacuum
, we shall consider a massless and free test scalar field @xmath79 defined on ( [ s2 ] ) .
the dynamics of @xmath67 can be derived from the action @xmath80 now we consider a semiclassical approximation for the 5d scalar field @xmath67 in the form @xmath81 , with @xmath82 denoting the background part of @xmath67 and @xmath83 denoting the quantum fluctuations of @xmath67 , which will be considered as very small .
this is consistent with a linear approximation for the scalar metric fluctuations .
thus , a first - order approximation in the gauge scalar fluctuations of the form @xmath84 , will be sufficient in order to have a good description of these fluctuations during inflation , in the present formalism . after making this approximation , one find that the dynamics on 5d of @xmath85 and @xmath86 , are @xmath87=0,\label{x3}\\ \frac{\partial^{2}\delta\varphi}{\partial n^2}&+ & 3\frac{\partial\delta\varphi}{\partial n}-e^{-2n}\nabla_{r}^{2}\delta\varphi-\left[l^{2}\frac{\partial^{2}\delta\varphi}{\partial l^2}+4l\frac{\partial\delta\varphi}{\partial l}\right ] \nonumber \\ & - & 2\psi\left[l^{2}\frac{\partial^{2}\varphi_{b}}{\partial l^2 } + 4l\frac{\partial\varphi_{b}}{\partial l}\right]=0 . \label{x22}\end{aligned}\ ] ] the expression ( [ x3 ] ) gives the dynamics on the background of @xmath88 , whereas eq .
( [ x22 ] ) describes the dynamics for the quantum fluctuations @xmath89 in terms of the scalar metric fluctuations @xmath90 and the background field @xmath82 . on the other hand ,
the expression for the energy momentum tensor components on the background is the background tensor metric and @xmath91 the solution for @xmath63 on @xmath92 . ]
@xmath93\right|_{g_{ab}=\bar{g}_{ab},\varphi = \varphi_b},\ ] ] where the lagrangian density corresponding to the inflaton field is @xmath94 one can find the _ linearized equation _ which describes the dynamics of the salar metric fluctuations @xmath90 on the linearized fluctuating metric @xmath95@xcite : @xmath96 \nonumber \\ & = & -\frac{8\pi
g}{3}\left[\frac{\partial\varphi_b}{\partial n}\frac{\partial\delta\varphi}{\partial n } + \psi l^{2}\left(\frac{\partial\varphi _ b}{\partial l}\right)^{2 } + l^{2}\frac{\partial\varphi_{b}}{\partial l}\frac{\partial\delta\varphi}{\partial l}\right ] , \label{x15}\end{aligned}\ ] ] where @xmath90 complies with the linearized non - diagonal einstein equations @xmath97 these equations provide us with the dynamics for both , the scalar field and the 5d gauge - invariant metric fluctuations@xcite .
we consider the background metric ( [ s1 ] ) .
if we take a static foliation @xmath98 , with the transformations @xmath99 we obtain the effective 4d de sitter background metric @xmath100 here , @xmath2 is a constant that represents the hubble parameter . for the relativistic point of view , this means that now we shall move with penta - velocities @xmath101 the effective 4d action on @xmath102 derived from the 5d action ( [ s3 ] ) , reads @xmath103 where @xmath104 is the massive scalar field induced on @xmath102 .
furthermore , @xmath105 is the determinant of the 4d induced metric , which in the case of background metric ( [ w1 ] ) yields @xmath106 .
moreover , in the case of the perturbed metric ( [ s2 ] ) gives @xmath107 $ ] , and @xmath108 is a dimensionless constant .
the induced 4d effective potential @xmath109 in the action ( [ w8 ] ) , has the form @xmath110 given the quantum nature of the fields @xmath111 and @xmath112 it seems convenient to use the quantization procedure . to do it ,
we impose the commutation relations @xmath113
= i\ , \delta^{(3)}\left(\vec r-\vec r'\right),\quad \left[\psi(t,\vec r ) , \pi^0_{(\psi)}(t,\vec r')\right ] = i\ , \delta^{(3)}\left(\vec r-\vec r'\right),\ ] ] where @xmath114 and @xmath115 are the canonical momentums for @xmath111 and @xmath112 , respectively
. we can derive the commutation relations on the effective 4d background metric .
if the inflaton and metric fluctuations are small , these commutators can be approximated to @xmath116 \simeq i \sqrt{\left|\frac{^{(4 ) } g_0}{^{(4 ) } \bar{g}}\right| } \delta^{(3)}\left(\vec r-\vec r'\right ) , \label{w11 } \\ & & \left[\psi(t,\vec r ) , \dot\psi(t,\vec r')\right ] \simeq i \frac{4}{9 } \pi g \sqrt{\left|\frac{^{(4 ) } g_0}{^{(4 ) } \bar{g}}\right| } \delta^{(3)}\left(\vec r-\vec r'\right ) .
\label{w22}\end{aligned}\ ] ] in this section we shall examine the evolution of the background inflaton field , jointly with the metric and inflaton fluctuations .
the dynamics of the background inflaton dynamics on the effective 4d background ( de sitter ) metric is described by the equation ( [ x3 ] ) with @xmath117 and @xmath118 . in what follows
we shall define @xmath119 . after making separation of variables in ( [ x3 ] ) ,
we obtain @xmath120 @xmath121 being a dimensionless separation constant .
the background friedmann equation @xmath122 , on the effective 4d hypersurface is @xmath123 the complete solution for eqs .
( [ af1 ] ) and ( [ af2 ] ) is with @xmath124 , such that the inflaton field is a constant of time : @xmath125@xcite .
however , at the end of inflation the second term in the equation ( [ af1 ] ) can be neglected and that equation can be approximated to @xmath126 in what follows we shall consider the dynamics of the inflaton field governed by the approximated equation ( [ af3 ] ) with the effective 4d friedman equation ( [ af2 ] ) .
we shall consider that at the end of inflation the mass of the inflaton field is of the order of the hubble parameter .
in particular , if we consider @xmath127 , the dynamics of the metric fluctuations described by ( [ x15 ] ) is simplified and we obtain the following solution for ( [ af3 ] ) @xmath128 where @xmath129 and @xmath130 is some constant of integration to be determined . using the result ( [ af4 ] ) , and combining equations ( [ x15 ] ) to ( [ x10 ] ) over the 4d hypersurfaces @xmath131 , we obtain an uncoupled equation for @xmath132 @xmath133 where @xmath134 , @xmath135 being a separation constant . following a canonical quantization process , the field @xmath112 can be expanded in fourier modes @xmath136,\ ] ] where the annihilation and creation operators @xmath137 and @xmath138 , satisfy the usual commutation algebra @xmath139=\delta ^{(3)}(\vec{k}-\vec{k'}),\quad [ a_{k},a_{k'}]=[a_{k}^{\dagger},a_{k'}^{\dagger}]=0.\ ] ] using the commutation relation ( [ m5 ] ) and the fourier expansion ( [ m4 ] ) , we obtain the normalization condition for the modes @xmath140 @xmath141 where @xmath142 is the scale factor in the 4d metric and @xmath143 , being @xmath144 the cosmic time at the end of inflation . inserting ( [ m4 ] ) in ( [ af5 ] )
, we obtain that the modes @xmath140 satisfy @xmath145 lets us consider the phase @xmath130 so that at the end of inflation @xmath146 . in this case
, the equation ( [ m7 ] ) leads to @xmath147 by means of the equation ( [ m6 ] ) and choosing the bunch - davies vacuum condition , the normalized solution of ( [ m8 ] ) , is @xmath148,\ ] ] where @xmath149 $ ] is the second kind hankel function , @xmath150 and @xmath151 .
now , employing again the equations ( [ x15 ] ) to ( [ x10 ] ) over the 4d hypersurfaces @xmath152 and incorporating the solution ( [ af4 ] ) , we obtain that the 4d quantum fluctuations for the inflaton field are determined by the equation@xcite @xmath153 \right|_{l = h_0^{-1 } } + 8 h_0 ^ 2 \psi \phi_b=0,\ ] ] performing a fourier expansion of @xmath154 in the form @xmath155,\ ] ] where the annihilation and creation operators @xmath156 and @xmath157 satisfy the commutation algebra given by ( [ m5 ] ) . inserting ( [ m10 ] ) and ( [ m4 ] ) in ( [ m9 ] )
, it leads to @xmath158 \right|_{l = h_0^{-1 } } + 8 h_0 ^ 2 \xi_{k } \phi_b=0.\ ] ] the previous equation is inhomogeneous , non separable and presents a source term .
the general solution will be the solution of the homogeneous equation plus a particular solution @xmath159 + \frac{e^{3 h_0 ( t_0-t)}}{2 } \sqrt{\frac{\pi}{h_0 } } \ : { \cal h}^{(2)}_{\mu}\left[x(t)\right],\ ] ] where @xmath160 and @xmath161 are polinomial functions of @xmath162 , and @xmath163 $ ] is the second kind hankel function with @xmath164 and @xmath165 . here ,
we have used both , the bunch - davies vacuum and the normalization conditions @xmath166 for the fourier modes of the homogeneous solution . towards the end of inflation
one can neglect the lower terms of @xmath160 and @xmath161 , to take only those with higher order in @xmath167 .
the term with the hankel function can be neglected too , so that we obtain @xmath168 finally , the approximated solution for the modes of the inflaton field fluctuations at the end of inflation can be rewritten as @xmath169 where @xmath160 is given by ( [ af7 ] ) .
to consider the particle creation lets first rewrite the solution ( [ af8 ] ) in terms of the conformal time in a spatially flat 4d friedmann - robertson - walker cosmology @xmath170 .
furthermore , we obtain that the modes of the inflaton fluctuations are @xmath171 and their derivatives with respect to @xmath172 , which we denote with a prime , are @xmath173 where @xmath174 in order to calculate the number of particles produced for each mode ( @xmath175 ) , we match the solutions obtained at the end of inflation , with those of the long - wavelength modes during a phase transition from inflation to the radiation - dominated era .
to calculate this last solutions , we make use of the bogolyubov transformation ( @xcite ) . for the long - wavelength modes we can approximate the phase transition to be instantaneous , occurring at some time @xmath176 .
so , we have that @xmath177 where @xmath178 . for convenience ,
we set @xmath179 , so we obtain that @xmath180 , @xmath181 and @xmath182 . after the transition
, the hubble parameter is no longer constant , @xmath183 , and we get : @xmath184 in what follows we shall compute the number of particles and gravitons in the radiation dominated era , which takes place after a phase transition from the inflationary epoch . in order to calculated the number of created particles after a phase transition , we calculate the modes of the inflaton field fluctuations taking into account the bogolyubov transformations @xmath185,\ ] ] so that their temporal derivatives are @xmath186.\ ] ] to calculate the coefficients @xmath187 and @xmath188 , we use the continuity of @xmath189 and @xmath190 at @xmath191 , so that @xmath192 @xmath193 where @xmath194 finally , by using @xmath195 we obtain the number of created particles with a wavenumber @xmath196 @xmath197 to estimate numerically the number of particles created after the phase transition on cosmological scales ( small wavenumbers ) , we shall consider that the number of e - folds , @xmath198 , at the end of inflation is at least : @xmath199 .
the value of @xmath196 depends upon the details of the transition between inflation and the radiation dominated epoch . for long - wavelength modes ( on cosmological super hubble scales ) , the transition can be considered as abrupt . in general
one expects that @xmath200@xcite .
for instance , if we take @xmath201 and @xmath202 .
] , with @xmath203 ) , for the inflaton field fluctuations on cosmological scales . ]
, we obtain @xmath204 which , in cases where the wavenumber is small , is a number sufficiently important to agree with the desired values for entropy @xmath205 . in the figure ( [ figura1 ] )
was plotted the number of particles created for @xmath206 .
notice that the number of created particles @xmath207 increases dramatically for cosmological scales .
we can also estimate the number of gravitons per mode generated by the metric fluctuations at the end of inflation .
again we first must rewrite the solution of the metric fluctuations modes in conformal time @xmath172 .
the modes for the metric fluctuations ( [ m9 ] ) can be expressed @xmath208 so that the conformal time derivative is @xmath209 where the function @xmath210 and its conformal time derivative are given by @xmath211 @xmath212 let us consider again the phase transition ( [ transition1 ] ) , with @xmath179 , so that we have @xmath180 , @xmath181 and @xmath182 . after the phase transition
, the metric fluctuations modes can be written as @xmath213,\ ] ] so that @xmath214.\ ] ] to obtain the coefficients @xmath215 and @xmath216 we use the continuity of @xmath217 and @xmath218 at moment of the phase transition , @xmath219 , and we obtain @xmath220\right . \nonumber \\ & -&\left .
\sqrt{\frac{k \sqrt{3}}{2 } } \right\}.\end{aligned}\ ] ] using again @xmath221 , we can obtain the number of gravitons generated after the phase transition @xmath222 notice that the third term inside the first bracket is the dominant one .
now , if we consider the values of the parameters at the end of inflation , we can calculate the approximate number of gravitons created after the phase transition . in order to estimate a numerical value , we set @xmath223 and @xmath224 , so that we obtain @xmath225 which is a number very small with respect to the number of particles created ( [ pc ] ) after the phase transition . in the figure ( [ figura2 ] )
was plotted the number of gravitons created @xmath226 , which also increases dramatically on larger cosmological scales .
we have studied particle and graviton production after an abrupt phase transition from inflation to the radiation dominated era .
we have considered the bogoliubov formalism to calculate the modes solution for the fields after a phase transition to a radiation dominated epoch .
our approach is different from others developed in @xcite , because here resonance effects are not important for the inflaton and metric fluctuations .
furthermore , the solutions for the modes of the inflaton fluctuations and the scalar fluctuations of the metric were obtained using a recently developed formalism which is related to the induced matter theory of gravity . during inflation ,
such approach makes possible the simultaneous description of the coupled dynamics of the inflaton field fluctuations with the gauge - invariant fluctuations of the metric .
after inflation this solutions can be matched with those of the radiation dominated epoch to obtain the bogoliubov coefficients .
the results here obtained are very interesting , because the number of created particles after inflation results to be greater than @xmath0 at cosmological scales , and increases dramatically as @xmath227 .
furthermore , @xmath207 is very sensitive to the number of e - folds during inflation .
this is due to the fact that @xmath207 increases as @xmath228 . in what respect to the number of gravitons created after inflation , the results are qualitatively similar to those of created particles
however , the number of gravitons created are , at least , of the order of @xmath229 , which is much smaller ( @xmath1 times ) than @xmath207 .
all these results were obtained using the values @xmath230 , with @xmath231 , and wavenumbers @xmath232 .
the authors acknowledge unmdp and conicet argentina for financial support . 99 a. h. guth , phys . rev .
d 23 ( 1981 ) 347 .
d. h. lyth and a. riotto , phys . rept . * 314 * : 1 ( 1999 ) .
linde , physics and inflationary cosmology , harwood , chur , switzerland , 1990 .
a. r. liddle and d. h. lyth , cosmological inflation and large - scale structure , cambridge university press , 2000
. m. bellini , _ et al_. phys .
* d54 * : 7172 ( 1996 ) .
e. w. kolb , s. matarrese , a. notari and a. riotto , mod .
lett . * a20*:2705 ( 2005 ) .
esteban calzetta .
phys . rev .
* d44*:3043(1991 ) .
a. a. grib and y. v. pavlov .
. cosmol . * 11 * : 119 ( 2005 ) .
a. d. linde .
. lett . * b129 * : 177(1983 ) .
vadim kuzmin and igor tkachev , * jetp * lett . *
68 * : 271 ( 1998 ) .
gert aarts and anders tranberg , phys . lett . * b650 * : 65 ( 2007 ) .
mauricio bellini , nuovo cim . * b117 * : 653 ( 2002 ) .
hassan firouzjahi and salomeh khoeini - moghaddam , * jcap 1102 * : 012 ( 2011 )
. a. d. dolgov and a. d. linde .
phys . lett . * b116 * : 329 ( 1982 ) .
l. f. abbott , e. farhi and m. b. wise .
. lett . * b117 * : 29 ( 1982 ) .
d. v. nanopoulos , k. a. olive , m. srednicki .
lett . * b127 * : 30 ( 1983 ) .
y. shtanov , j. h. traschen , r. h. brandenberger .
rev . * d51 * : 5438 ( 1995 ) .
l. kofman , a. linde , a. a. starobinsky .
* 76 * : 1011 ( 1996 ) .
d. i. kaiser , phys . rev . * d53 * , 1776 ( 1996 ) .
m. anabitarte , m. bellini , j. e. madriz aguilar , eur .
j. * c65*:295 ( 2010 ) .
j. e. campbell , _ a course of differential geometry _
( charendon , oxford , 1926 ) ; + l. magaard , _ zur einbettung riemannscher raume in einstein - raume und konformeuclidische raume_. ( phd thesis , kiel , 1963 ) . s. rippl , c. romero , r. tavakol , class . quant . grav .
* 12 * : 2411 ( 1995 ) .
f. dahia , c. romero , j. math .
phys.*43 * : 5804 ( 2002 ) .
f. dahia , c. romero , class .
* 22 * : 5005 ( 2005 ) .
m. bellini .
. lett . * b609 * : 187 ( 2005 ) .
j. e. madriz aguilar and m. bellini
. phys . lett . * b619 * : 208 ( 2005 ) .
m. anabitarte and m. bellini .
phys.*47 * : 042502 ( 2006 ) .
l. randall and r. sundrum , mod .
a13*:2807 ( 1998 ) .
l. randall and r. sundrum , phys .
lett . * 83*:4690 ( 1999 ) .
l. f. p. da silva , j. e. madriz aguilar , mod .
. lett . * a23*:1213 ( 2008 ) .
l. kofman , a. linde and a. a. starobinsky , phys .
rev . lett . * 73 * : 3195 ( 1994 ) .
d. s. ledesma and m. bellini , phys
. lett . * b581*:1 ( 2004 ) .
n. d. birrell and p. c. w. davies , _ quantum fields in curved space_. cambridge university press , cambridge , england ( 1982 ) .
t. damour and a. vilenkin .
phys . rev . *
d53 * : 2981 ( 1996 ) . | we use the bogoliubov formalism to study both , particles and gravitons creation at the reheating epoch , after a phase transition from inflation to a radiation dominated universe .
the modes of the inflaton field fluctuations and the scalar fluctuations of the metric at the end of inflation are obtained by using a recently introduced formalism related to the induced matter theory of gravity .
the interesting result is that the number of created particles is bigger than @xmath0 on cosmological scales .
furthermore , the number of gravitons are nearly @xmath1 times smaller than the number of created particles . in both cases
, these numbers rapidly increase on cosmological scales . |
dwarf galaxies are by far the most numerous type of galaxies , not only in the local group but also in the nearby clusters which have been studied in detail ( eg . sandage & binggeli 1984 for virgo ; caldwell & bothun 1987 for fornax ) .
it should therefore be normal to assume that nearby groups of galaxies contain similarly a large population of dwarfs .
however it is often believed that these nearby groups are devoided of gas - rich dwarfs at least , based on earlier hi surveys of lo & sargent ( 1979 ) and haynes & roberts ( 1979 ) .
yet these surveys were done with only very sparse samplings in the case of the sculptor group , and to sensitivities of a few times 10@xmath3 m@xmath4 for the m81 , cvni , and ngc 1023 groups .
finding nearby dwarf galaxies is crucial in defining the faint - end of the luminosity function of galaxies , which has far - reaching consequences in the study of faint galaxy populations and galaxy formation in general .
but redshift surveys of apparent - magnitude - limited samples are , for various reasons , not very efficient at finding low - luminosity systems .
the group and the a group of galaxies are the most nearby groups in the southern hemisphere .
being roughly 3 magnitudes closer than the virgo and fornax clusters they bring us a unique opportunity to study the very faint end of the luminosity function .
the goal of this project was therefore to investigate the dwarf population in these two groups , especially the gas - rich dwarf irregulars ( dirrs ) .
it is much easier to get redshifts for dirrs ( in hi ) than to determine redshifts for des candidates , and also our main interest was to carry out dark matter studies on dirrs . the and a groups offer very different environments to their dwarf population .
the group is mainly composed of late - type spiral galaxies .
its five major members , ngc 55 , ngc 247 , ngc 253 , ngc 300 , and ngc 7793 , are almost all normal gas - rich systems , and their properties are listed in table [ t : sculm ] .
our adopted distance , based on numerous distance indicators ( listed in puche & carignan 1988 ) will be 2.5 mpc .
two more galaxies , ngc 45 and ngc 24 , also late - type spirals , are found in the same region but several mpcs further away , possibly forming an extension of the group . amongst the sculptor members only ngc 253 shows starburst activities while the other galaxies are quiescent .
their hi distributions have been mapped in details at the vla by carignan & puche ( 1990a , 1990b ) , puche ( 1990 , 1991a , 1991b ) .
miller ( 1996 ) has imaged in h@xmath5 the known dwarf members of the group and detected only two with current star formation .
the a group on the other hand is a loose chain of galaxies , a heterogeneous assembly of early to late - type galaxies , having the largest dispersion of morphological types amongst all of the 55 nearest groups ( de vaucouleurs 1979 ) .
almost all of the major members are abnormal .
not only is ngc 5128 , centaurus a itself , one of the most peculiar radio galaxies , but the other main members also show signs of activities .
for example ngc 5236 ( m83 ) has a starburst nucleus and possesses an anomalously high supernovae rate ( telesco & harper 1980 ) , and an asymmetric hi velocity field ( lewis 1969 ) .
ngc 5253 also shows starburst activities and a high supernovae rate ( eg . van den bergh 1980 ) , and a very disturbed h@xmath6 velocity field ( taylor 1992 ) .
ngc 5102 is an unusual s0 galaxy in a post - starburst phase ( pritchet 1979 ) .
and ngc 4945 has a seyfert nucleus emitting variable hard x - rays ( moorwood & olivia 1994 ) .
several authors have suggested that these peculiarities are induced by accretion of gas - rich dwarf systems ( graham 1979 ; van gorkom 1990 ) .
therefore the study of the gas - rich dwarfs in this group is of particular interest .
the main members are listed in table [ t : cenam ] , and the adopted distance for this group is 3.5 mpc based on the newly derived distances for these galaxies ( see the table ) .
it is therefore the second nearest group outside the local group , more or less ex - aequo with the m 81 group ( from cepheids m 81 is placed at 3.6 mpc , freedman 1994 ) .
this paper , the first of a series , presents the redshift surveys for finding dirrs in sculptor and centaurus a. further papers will present full b , r , i photometry of these nearby dwarfs ( ct 1997a ) , the neutral hydrogen kinematics of a sample of these objects ( ct 1997b ) and finally the dark matter studies for that sample of dirrs ( ct 1997c ) .
the first step was to find dwarf galaxy candidates in the region of the sculptor and centaurus a groups from visual inspection of src j survey films .
these groups , because of their proximity , subtend very large angles on the sky .
the areas on which to select candidates were determined from the sky coverage of the already - known galaxy members in each group ( main members as well as dwarf members known at the time ) .
the fields thus retained for visual scanning were those with centers between : @xmath7 @xmath8 for the and a groups respectively , where @xmath5 and @xmath9 are the right ascension and declination at epoch 1950 .
this corresponds to wide areas of approximately 50 src survey films in each case .
each film was inspected visually with a small magnifier , scanning once north to south , then west to east , to locate all objects which could be irregular galaxies at low redshift . a dwarf irregular should be easily identifiable at such a small distance of a few mpc , showing much structural details , with even the brightest stars resolved .
we selected particularly all low - surface - brightness galaxies which showed some degree of stellar resolution , or some distinct regions or condensations , including all objects with small bulges or some form of high - surface - brightness central regions , and those composed of bright knots embedded in low - surface - brightness envelopes .
the objects excluded from the candidate list were giant ellipticals and obvious distant giant spirals ( judged as such from the relative sizes of the bulges and the tightness of the spiral arms ) .
with these criteria we should have included in our candidate lists all gas - rich dwarfs like dwarf spirals , dwarf irregulars , and blue compact dwarfs ( bcd ) as well , although some bcds that appear as a single symmetric burnt - out clump on the src films with no associated low - surface - brightness envelope would have escaped us , being indistinguishable from background ellipticals . a total of 123 dwarf galaxy candidates for the sculptor group and 145 for the centaurus a group were thus identified , with no redshifts in existing galaxy catalogues at the time , including the rc2 ( de vaucouleurs 1976 ) , the _ nearby galaxies catalog _ ( tully 1988 ) , the _ eso catalog _ ( lauberts 1982 ) , the _ southern galaxy catalog _ ( corwin 1985 ) , the _ hi catalog of galaxies _ ( huchtmeier and richter 1989 ) , and the nasa / ipac extragalactic database ( ned ) .
these candidates are spread on the sky over approximatly 940 and 910 square degrees . amongst these , 58 objects are identified for the first time and are named with the sigla sc or cen ( see tables [ t : sback ] and [ t : cback ] ) . dwarf irregular galaxies being normally gas - rich , the easiest way to determine their redshift is with 21 cm hi line observations
. some of these galaxies like ddo 154 for example ( carignan & freeman 1988 ) have actually 5 times more mass in hi than in stars . and at such low redshift it is not very time - consuming to get down to an interestingly low detection limit ( in terms of hi solar mass ) using a single - dish telescope .
also since the groups are so widespread on the sky there is little chance of confusion even with a large beam .
the hi observations were carried out in august 1990 for the candidates and february 1991 for the a ones , at the 64 m parkes radiotelescope .
the telescope had a half - power beamwidth of 14.8 at 21 cm and a sensitivity of 0.63 k / jy . the front end was a cryogenic fet receiver yielding a system temperature of @xmath1040 k. the back end was the 1024 channel digital autocorrelator ( ables 1975 ) , configured to give two 512 channels spectra ( 1 spectrum for each of two orthogonal polarizations ) , with a bandwidth of 10 mhz .
this resulted in a channel separation of 4.1 and a velocity resolution after hanning smoothing of 8.2 , covering a velocity range of roughly @xmath11200 to 1800 .
this setup ensured that we were surveying a wide enough velocity interval for these two groups whose main members have heliocentric velocities between 116 and 669 , as well as preserving a good enough resolution to distinguish dwarf galaxies profiles .
these do not necessarily have large rotation velocities and therefore their profiles can often look like asymmetric gaussians , of much smaller widths than for spirals ( see karachentseva 1990 who classified 539 dwarf hi profiles ) .
the flux density calibration was carried out by observing hydra a ( pks 0915 - 18 ) , with an adopted flux density of 43.5 jy at 21 cm , and the velocity scale was checked on the source uks 1908 - 621 which has a radial heliocentric velocity @xmath12946 and a profile width of @xmath13 92 .
each observation consisted of a 10 minutes integration on source , alternated with a 10 minutes integration on a sky reference position lying 10 minutes away to the east , so that the same position ( in hour angle ) was observed during this reference scan to get the same background .
the reductions were performed with _ slap _ , an interactive spectral line reduction package developed at jodrell bank , and _ posp _ , developed at nrao .
the sky reference spectra were used to subtract the sky and remove large baseline variations , and the two polarization spectra were then co - added .
this baseline was interactively determined by masking out any emission and fitting a low order chebyshev polynomial .
the zeroth and first moments of the emission were calculated to obtain the total flux and heliocentric velocity of the galaxy .
line widths were determined at the 20% and 50% peak flux levels .
the resulting spectra had rms values typically around 30 mjy , and so our 3@xmath14 detection limits in and a are @xmath15 and @xmath1 ( for our adopted distances of 2.5 mpc and 3.5 mpc respectively ) . tables [ t : sback ] and [ t : cback ] list for all the hi detections the heliocentric velocity of the galaxy , the hi integrated flux , and the linewidth at the 20% level .
the velocity quoted is an average of the intensity - weighted velocity , and the midpoint velocities at the 50% and 20% levels .
errors on these velocities are typically 2 .
figures [ f : sallsp ] and [ f : callsp ] show the reduced parkes spectra of all the detected objects over the whole velocity range .
however one needs to be cautious when detecting hi emission in the velocity range 100 to 600 in the direction of the : many high velocity clouds ( hvc ) associated with our galaxy are known to have velocities sometimes as high as 400 ( wakker 1990 ) .
so it could happen that such a hvc is lying in the 15 parkes beam towards a candidate object ( which could actually be a background object ) , confusing us into believing that the hi detected belongs to that galaxy .
in fact there are some hvcs with high positive velocities known to lie in the ( _ l , b _ ) area of both groups ( wayte 1990 ; wakker 1990 ) .
these are not the only possible source of confusion : intra - group hi clouds have been detected in the group by haynes and roberts ( 1979 ) who made extensive mapping in that region with the green bank 140 ft and found several seemingly ` free - floating ' hi clouds . from their observations
it is not clear actually if these clouds are genuine intergalactic hi clouds , mere hvcs , or some component of the magellanic stream ( mathewson 1974 ) , but for our purposes all these possibilities are to be considered hi pollution .
therefore in order to discriminate genuine nearby dwarf hi emission from hi hvcs or clouds , further 5-point mappings were carried out in may 1991 and september 1991 .
this consists of integrating 15 away ( one beam width ) from the source to the north , south , east and west ( or in a cross pattern : north - west , north - east , etc ) . in the case of a dwarf galaxy
the hi emission will be mostly concentrated in the middle beam , and velocity variations between different pointings will indicate rotation ; for a hi cloud the extended emission is less ordered in velocity and on the sky .
it is also very useful for the dwarfs that lie near a hi - rich large galaxy like m83 or ngc 4945 , where one can detect hi flux in the sidelobes of the beam and so one component in the spectrum will be due to the hi envelope of the large galaxy ; in this case a 5-point mapping helps confirming that there is an independent object as well in the spectrum .
however this 5-points method works only if there is actually some emission 15 away and so is not always a useful discriminant between dwarfs and usurper hi clouds .
the clearest way to discriminate between hi pollution and genuine dwarfs is to try to get a redshift optically in h@xmath5 for these objects .
the presence of h@xmath5 at the same velocity as hi indicates the presence of star formation , therefore ruling out a neutral cloud .
the h@xmath5 spectroscopy observations were carried out in april 1991 , august 1991 and november 1991 , with the double beam spectrograph ( dbs ) on the mssso 2.3 m telescope .
not only did we observe the objects detected in hi at the groups velocities ( ie : the objects that needed confirming because of hvcs etc ) , but also all the objects on our catalog lists which were not detected at all in hi , in the hope to identify new dwarf galaxies weak in hi but detectable in h@xmath5 .
the long slit was about 6 and was set to a width of 3 , positioned when possible along the major axis of the object .
the detector used was a photon counting array with a spatial resolution of 0.67 per pixel .
the 1200 g / mm gratings were used in the blue and red arms , covering spectral ranges of 4950 to 5250 in the blue , and 6450 to 6750 in the red , at 0.4 per pixel ( 18 at h@xmath5 ) .
this blue arm setup was to target the 4959 and 5007lines , but these lines being always fainter than the h@xmath6 , in most cases the velocities were determined from the h@xmath5 alone .
exposures were 2000s per object , or were stopped before that when sufficient signal - to - noise was achieved for a redshift determination .
spectral reduction was carried out with the figaro package from k. shortridge ( aao ) .
each galaxy observation was bracketed with arc lamp exposures for wavelength calibration .
sky spectra were extracted from the frames at regions well outside galactic emission .
gaussian profiles were then fit to the emission lines , and a heliocentric correction was applied to the resulting velocities .
these are listed in tables [ t : sback ] to [ t : cfin ] , and have errors of typically 10 .
of the 123 and 145 original candidates in sculptor and centaurus a , all observed , we obtained new redshifts for 108 objects . for 61 other objects
, redshifts were found in the literature during the course of the project , for example when the rc3 became available , and for simplicity are not repeated in our tables [ t : sback ] and [ t : cback ] ; they can be found in the rc3 , maia ( 1993 ) , dressler ( 1991 ) , da costa ( 1991 ) , fouqu ( 1990 ) , rhee ( 1992 ) , and parker ( 1990 ) .
finally 74 objects remain redshiftless .
a total of 16 dwarfs are found to be members of the sculptor group and 20 of the centaurus a group , and are listed in tables [ t : sfin ] and [ t : cfin ] . amongst these ,
25 were already known in the literature ( we confirmed their redshift ) , 5 more were already identified and catalogued but without known redshifts , and finally 6 of these objects are identified for the first time .
none of these 6 new objects had been detected by iras .
these 36 members of sculptor and centaurus a include all our objects with confirmed redshifts in h@xmath5 , as well as those detected only in hi but with successful 5-points mappings . only 21 of these dwarfs had their hi emission confirmed with an h@xmath5 redshift . and
3 more objects , not detected in hi , were found in the h@xmath5 survey : am 0106 - 382 in , as well as ngc 5206 and eso 272-g025 in a. further deeper integrations in hi at parkes still did not detect any neutral gas content , bringing their hi upper limit to @xmath17 for and @xmath18 for a. several hvc s were indeed found in the line - of - sight of some of our candidates as suspected : in all , 15 hi detections were rejected from our dwarf lists , either after the 5-point mapping , or after obtaining an h@xmath5 detection at higher redshift .
these hvc hi profiles are considerably more messy that the confirmed galaxies profiles .
we are therefore confident that the 12 dwarfs which remain unconfirmed in h@xmath5 but were kept in the sample judging from their 5-point mapping are bona fide dwarfs .
moreover , subsequent surface photometry , to be presented in ct ( 1997a ) , shows that these objects have the same blue mean colours as our confirmed dwarfs sample . amongst other known objects in the literature to be found in the sculptor region
is ugca 438 ( alias uks 2323 - 326 ) , which we decided not to include as a sculptor member : longmore ( 1978 ) who discovered it estimated a distance of 1.3 mpc using the brightest stars , although van den bergh ( 1994 ) claims it does not belong to the local group , based on its location on the ( heliocentric velocity ) versus ( distance from the solar apex ) diagram .
so it might be lying in fact between the two groups .
two other controversial cases are eso 294-g010 and eso 410-g005 which were included in miller s ( 1994 ) work on sculptor dwarfs , but which we have not considered members here , because eso 294-g010 shows a 2.5@xmath14 detection at 4450 and eso 410-g005 was not detected neither in hi nor in h@xmath5 ( we have prefered to include only the objects proven to be in the groups rather than those that have not been proven to be in the background ) .
it should be stressed again that only positions of candidates selected visually on the films were observed , ie : no parkes hi integrations at random positions on the sky were performed in the regions of the two groups .
it is therefore more than probable that many more faint dwarfs are to be found in these regions .
a survey with the parkes 21 cm multibeam receiver for example could provide a more complete sample down to an interesting hi mass limit with full coverage over the area of the groups .
the spatial and velocity distribution of the 36 confirmed dwarfs of sculptor and centaurus a are shown in figures [ f : dis ] and [ f : vlg ] , and will be discussed at greater length in section 3.3 .
all the confirmed dwarfs are shown in figure [ f : atlas ] in a montage of enlarged src j prints .
their optical sizes range from about 40 up to 11 ( 0.5 kpc to 11 kpc ) .
this atlas reveals a large diversity of morphologies , although a few subclasses of dwarfs are noticeable .
eso 383-g087 is certainly a dwarf spiral ( type sm ) , being in fact one of the rare dwarf galaxies with clear spiral structure .
other dwarf spirals are eso 274-g001 , which even has a small nucleus , while ddo 6 , ddo 161 , and ddo 226 show hints of a bar , but no nuclei .
a few objects can be classified as bcds ( like eso 347-g017 , ngc 5264 or ngc 5408 ) , harbouring compact regions of very high surface brightness , although only eso 347-g017 satisfies the strict thuan and martin ( 1981 ) criterion that the localized star - forming region should be @xmath19 1 kpc .
eso 272-g025 and am0106 - 382 are possibly extreme bcds : they have the most compact star - forming regions , and are amongst the only three objects which were not detected in hi , perhaps because the gas has been succesfully expelled as is often believed to happen in low - mass bcds during their burst of star formation .
most of our objects are gas - rich dwarf irregulars , with sparse regions either concentrated in the center of the low - surface - brightness envelope ( eso 324-g024 ) , or dispersed through the whole envelope ( eso 381-g020 ) . amongst these lsb objects ,
some like ddo6 or eso 293-g035 have an almost ` cometary ' look , with a clump of regions at one extreme of a otherwise lsb disk .
these particular ` cometary dwarfs ' are actually very similar to many faint galaxies now revealed in the hst deep field .
bcds with similar morphologies were also found by loose & thuan ( 1985 ) , who suggested that it is due to self - propagating star formation which stopped at the edge of the galaxy .
these cometary dwarfs contrast with dirr objects like eso 444-g084 or sgc142448 - 4604.8 that are of extremely low surface - brightness overall .
some objects like sc2 or cen6 are just too ` dwarfish ' to discern any features but appear significantly distorted . finally ,
although this survey was focusing on gas - rich dwarfs , some early - types were caught in the sample : ngc 5206 , not detected in hi , is a rather symmetrical - looking de , whose light profile is well - fitted by a de vaucouleurs @xmath20 law ( prugniel 1993 ) . following binggeli and cameron s ( 1991 ) classification for early - type dwarfs , ngc 59 qualifies to be a ds0 type b ( for its high flattening ) and ngc 5237 a ds0 type d ( for its boxiness ) .
surface photometry results will be presented in ct ( 1997a ) , which will also present the luminosity function obtained , and discuss the selection and completeness of the survey .
the distribution on the sky of the member dwarfs is shown for each group in figure [ f : dis ] . in both groups
the dwarfs are more widely spread than the brighter main galaxies . despite the difference of environment in the two groups the spatial coverage of the dwarfs is found to be rather similar for both . in each case
an appendage of dwarfs is spearing away from the main members , while a few other objects have gathered near massive members , like ngc 247 or m83 , and even though these dwarfs are not all bound to these bright galaxies they probably respond more to these galaxies potential then to the overall - group potential .
interestingly just as the main a members are aligned in a long chain , the dwarf members show also such a elongated structure albeit heading more in a south - east direction . in the group
as well the dwarfs extend out to the south - east .
this is not due to some selection bias related to obscuration effects , because for the galaxy ngc 253 is very near the south galactic pole ( at @xmath21 ) , while for a the galactic plane lies to the south ( ngc 4945 is at @xmath22 ) .
in fact the distribution of candidate objects from our original lists is very uniform in the regions surveyed ( for it is even slightly skewed towards the north ) .
figure [ f : vlg ] shows the velocity distribution of the dwarfs in each group , where their declination is plotted against their velocity @xmath23 ( @xmath23 is the velocity relative to the local group , which is more useful than the heliocentric velocity because of the groups being so spread out on the sky : we used the yahil 1977 formula for the correction ) .
here again one can see how the dwarfs have a wider velocity coverage than their bright companions .
however it is quite clear from this figure where each group cuts off in velocity - space , : the dwarfs do not cover the whole observed velocity range uniformly .
both groups are well ` contained ' in velocity space , not overlapping with any string of stray dwarfs . only at @xmath24
appears another agglomeration of galaxies , which tully s catalogue ( 1988 ) refers to as the ` centaurus spur ' .
this is similar to what is observed in more extensive surveys , in which dwarfs are loosely following the bright galaxies distribution , and are found on the edges of the voids but do not seem to be filling these voids ( see for example thuan 1991 ) .
the fact that the dwarfs distribution is wider spatially than that of the more massive members is reminiscent of the situation in the local group , where most of the dirr galaxies sit in regions of low density , out at the fringes of the group .
a similar behaviour has been observed as well for the late - type dwarfs of nearby clusters , like virgo for which bothun ( 1985 ) conclude that the normal late - type spirals as well as the dirrs form an extended cluster population that has not yet fallen into the virgo core ( see also binggeli 1987 ) .
also observed in the virgo cluster is a general hi deficiency in the spirals , which also have smaller characteristic hi sizes than field spirals ( cayatte 1994 ) . here for and a there are no significant changes in the hi properties of dwarfs at the periphery of the group compared to those near the center of mass of the group or orbiting a bright member , and in such small and loose groups one does not expect to encounter the gas
removal processes that operate in denser groups or clusters .
the sample of radial velocities of the dwarfs and their projected distances from the center of mass of each group can be used to estimate their velocity dispersions and also their crossing times . the line - of - sight velocity dispersions obtained for sculptor and centaurus a are 202 and 150 respectively , which are typical for groups of this size . for the crossing times estimates ,
the centers of mass were adopted to be at the mean projected position and velocity of the most massive galaxies in each group : ngc 5128 and ngc 5236 for a , and ngc 253 and ngc 247 for sculptor .
ngc 45 and ngc 24 , being at least at twice the distance of the other sculptor members , were excluded from this estimate .
although ngc 247 is less luminous than ngc 7793 , its maximum rotational velocity is larger .
in fact from the mass - models of puche & carignan ( 1991 ) and carignan & puche ( 1990 ) , ngc 253 and ngc 247 together provide about 90% of the mass of the group .
the indicative crossing time ( see rood & dickel 1978 ) is then calculated by : @xmath25 where @xmath26 is the average projected radial distance from the center of mass , and @xmath27 is the mean absolute radial velocity from the center of mass .
table [ t : tcross ] show the results : the crossing times are quite long , @xmath28 years and @xmath29 years respectively , a considerable fraction of a hubble time , which shows that these two groups are probably not virialised .
the fact that the groups are unvirialised is hardly surprising : for a the spatial distribution of the objects in a long chain is easily noticeable , and for the velocity distribution is clearly non - gaussian .
figure [ f : velhis ] illustrates this point , where cumulative velocity histograms are shown for each group .
the histogram shows a plateau near @xmath30 ; ie rather few dwarfs have velocities close to the center of mass velocity . in figure
[ f : dvdr ] we plot @xmath31 against @xmath32 for each object ; here again it is verified that there is qualitatively no obvious signature of virialisation ; the points cover the whole ( @xmath31 @xmath32 ) area which indicates that these two loose groups are still collapsing . the situation in and a is however far from being unusual .
turner & gott ( 1976 ) have calculated crossing times for groups identified by sandage , tamman and de vaucouleurs , and concluded that most groups are just now entering the virialised regime .
giuricin ( 1988 ) finds that 75% of the groups in geller & huchra s ( 1983 ) catalogue are still in the phase of collapse and not yet virialised .
even cluster virialisation has now been put in doubt , as x - ray maps start revealing distinct subclustering , also seen optically as more redshifts come available like in the coma cluster ( colless & dunn 1996 ) .
our own local group is certainly still collapsing ( gunn 1974 ) .
in fact , looking in redshift space , the group seems to form one long appendage to the local group and on to the m81 group on the other side , as if these 3 groups were born out of the condensation of the same cloud , in which these 3 ` clumps ' have not yet quite collapsed ( it is referred to as the ` coma - sculptor cloud ' in tully s catalog 1988 ) .
this makes the determination of masses and mass - to - light ratios for these groups rather difficult .
if we navely apply the virial theorem to and a , we obtain mass - to - light ratios of @xmath33 and @xmath34 respectively ( using @xmath35 from rc3 b values and our adopted distances of 2.5 and 3.5 mpc ) .
but if we require the groups to be bound only , their masses are then a factor of two smaller than those obtained here from the virial theorem . that would bring the mass - to - light ratio of the whole a group into a regime similar to mass - to - light ratios of single galaxies .
in the case of however we have to conclude that the mass - to - light ratio of the group exceeds that of individual galaxies and suggests the presence of a large amount of dark matter in the group .
there are more sophisticated techniques to study the dynamics of loose groups .
peebles ( 1989 , 1990 , 1994 ) in a series of papers explored a new method of tracing nearby galaxies orbits back in time , using a numerical application of the action variational principle ( see also dunn & laflamme 1993 , 1995 ) .
his predicted velocities agree well with the observed velocities for the neighboring groups of galaxies within 3 mpc when they are assigned m / l of as much as @xmath36 150 @xmath37 .
our new dwarfs will be very valuable in testing this mass model .
a total of 123 and 145 dwarf galaxy candidates in the two nearest groups of galaxies sculptor ( 2.5 mpc ) and centaurus a ( 3.5 mpc ) were selected visually on src j films , over areas of 940 and 910 square degrees respectively . from an hi parkes survey , with detection limits of @xmath0 and @xmath1 ,
33 dwarf galaxies were detected in the redshift ranges of the groups . a follow - up h@xmath38 spectroscopic survey has confirmed 21 of these redshifts , and has found 3 more objects not detected in hi .
a total of 16 dwarfs are therefore found to be members of the sculptor group , and 20 of the centaurus a group .
most of these objects were already known in the literature , but 5 members have newly determined redshifts and 6 other members are identified for the first time
. these objects are mostly dirrs but with a variety of morphologies .
their global properties will be presented in subsequent papers .
the dwarf members have a broader distribution spatially and in velocity than the brighter members of each group . as most nearby groups , sculptor and centaurus a are not yet virialised .
it is a pleasure to thank bruno binggeli , evan skillman , and jacqueline van gorkom for many interesting comments .
this survey was started at universit de montral with the help of serge demers , pierre chastenay , and especially nathalie martimbeau who produced all the finding charts .
sc thanks tom broadhurst for his constant criticism .
thanks to j.english , a.koekemoer , h.liang , c.lidman , r.vaile and especially e.troup for their help on the parkes observing run .
the parkes telescope is part of the australia telescope national facility , which is operated in association with the division of radiophysics by csiro .
this project would have been a nightmare without the help of the nasa / ipac extragalactic database ( ned ) , which is operated by the jet propulsion laboratory , caltech , under contract with the national aeronautics and space administration .
financial support for this work was provided by an australian national university postgraduate scholarship , and by fonds fcar qubec .
lccccccc ngc 7793 & 23 55 15 -32 52 06 & sa(s)d & 230 & 9.61 & 3.38 & -18.03 & 1 ngc 55 & 00 12 24 -39 28 00 & sb(s)m & 125 & 8.39 & 1.6 & -17.72 & 1 ngc 247 & 00 44 39 -21 02 00 & sab(s)d & 159 & 9.60 & 2.53 & -17.41 & 1 ngc 253 & 00 45 07 -25 33 42 & sab(s)c & 251 & 7.99 & 2.58 & -19.07 & 1 ngc 300 & 00 52 31 -37 57 24 & sa(s)d & 142 & 8.70 & 2.1 & -17.91 & 2 ngc 24 & 00 07 24 -25 14 36 & sa(s)c & 554 & 12.13 & 11.08 & -18.09 & 1 ngc 45 & 00 11 32 -23 27 35 & sa(s)dm & 468 & 11.32 & 4.35 & -16.93 & 1 lccccccc ngc 4945 & 13 02 31 -49 12 12 & sb(s)cd sp & 560 & 8.45 & 3.18 & -19.06 & 1 ngc 5068 & 13 16 13 -20 46 36 & sab(rs)cd & 672 & 10.37 & 5.11 & -18.17 & 1 ngc 5102 & 13 19 07 -36 22 12 & sa0- & 467 & 10.13 & 3.12 & -17.34 & 2 ngc 5128 & 13 22 32 -42 45 33 & s0 pec & 562 & 7.32 & 3.50 & -20.4 & 3 a1332 - 45 & 13 31 39 -45 17 06 & sb(s)m & 826 & 11.27 & 5.25 & -17.33 & 1 ngc 5236 & 13 34 12 -29 36 48 & sab(s)c & 516 & 8.05 & 3.70 & -19.79 & 1 ngc 5253 & 13 37 05 -31 23 30 & i m pec & 404 & 10.67 & 4.09 & -17.39 & 4 llrclc eso 406-g040 & 22 57 33 -37 28.1 & 1248 & 4.2 & 67 & eso 406-g042 & 22 59 25 -37 21.2 & 1377 & 7.1 & 143 & mcg-05 - 54 - 024 & 23 11 07 -29 51.5 & & & & eso 407-g011 & 23 12 13 -34 02.7 & & & & 1869 sc1 & 23 16 54 -31 37 & & & & sc2 & 23 17 56 -32 10.8 & 68 & 10.4 & 49 & eso 347-g008 & 23 18 09 -42 00 & 1622 & 8.6 & 76 & 1606 eso 347-g017 & 23 24 16 -37 37.3 & 702 & 10.5 & 104 & 659 sc3 & 23 25 51 -27 05.8 & & & & sc4 & 23 26 40 -33 35.3 & & & & eso 470-g016 & 23 28
18 -27 56.8 & & & & mcg-05 - 55 - 020 & 23 28 21 -27 48 & & & & sc5 & 23 28 59 -27 28.5 & & & & sc6 & 23 30 13 -28 47.2 & & & & eso 291-g031 & 23 31 39 -46 16 & 1494 & 6.5 & 96 & mcg-05 - 55 - 026 & 23 32 59 -26 57.5 & & & & sc7 & 23 33 48 -23 18 & & & & eso
292-g002 & 23 34 51 -43 53.8 & & & & eso 537-g001 & 23 40 23 -26 36 & & & & 3780 ugca 442 & 23 41 10 -32 13.8 & 274 & 54.3 & 114 & 283 sc8 & 23 43 57 -32 50.5 & & & & eso 348-g009 & 23 46 47 -38 03 & 656 & 8.4 & 98 & 628 sc9 & 23 48 19 -23 16.1 & & & & sc10 & 23 48 25 -29 48.5 & & & & sc11 & 23 51 01 -26 37.5 & & & & 3006 sc12 & 23 51 42 -22 57.5 & & & & sc13 & 23 53 45 -42 01.7 & & & & sc14 & 23 54 23 -31 11.2 & & & & sc15 & 23 56 03 -24 49 & & & & sc16 & 23 56 07 -31 44.6 & & & & sc17 & 23 58 15 -26 45.4 & & & & sc18 & 23 58 22 -41 25.6 & 151 & 4.6 & 46 & eso 409-g008 & 00 00 18 -32 08.4 & & & & 7427 sc19 & 00 00 57 -26 51.3 & & & & eso 538-g021 & 00 03 13 -22 21.3 & & & & 3148 eso 472-g015 & 00 04 14 -25 13.3 & & & & eso 293-g035 & 00 04 19 -42 07 & 110 & 6.9 & 52 & eso 293-g040 & 00 05 00 -37 44.4 & & & & sdig & 00
05 41 -34 51.3 & 229 & 2.7 & 31 & ugca 003 & 00 07 45 -18 32.5 & 1549 & 9.6 & 50 & 1551 ngc 59 & 00 12 53 -21 43.2 & 367 & 3.7 & 87 & 357 eso 410-g005 & 00 13 00 -32 27.5 & & & & eso 194-g002 & 00 16 02 -47 56 & & & & 1494 sc20 & 00 16 07 -24 17 & & & & ugca 005 & 00 16 17 -19 17 & & & & 3187 sc21 & 00 19 33 -19 07.1 & & & & sc22 & 00 21 20 -24 58.7 & & & & eso 410-g011 & 00 22 27 -27 34.2 & & & & eso 294-g010 & 00 24 06 -42 07.8 & & & & 4450 eso 473-g020 & 00 24 55 -25 26.5 & & & & eso 410-g012 & 00 25 49 -28 15.3 & 1545 & 5.9 & 131 & 1557 eso 473-g024 & 00 28 50 -23 02.6 & 553 & 7.7 & 74 & eso 294-g020 & 00 29 45 -40 32.3 & & & & 1431 sc23 & 00 30 53 -22 08.8 & & & & eso 410-g017 & 00 31 10 -28 04.5 & 1473 & 7.0 & 126 & 1535 sc24 & 00 34 12 -32 50.8 & 79 & 11.8 & 92 & sc25 & 00 34 33 -27 09 & & & & am0035 - 434 & 00 35 21 -43 46.5 & & & & sc26 & 00 37 44 -18 05.7 & & & & eso 540-g012 & 00 37 49 -20 58 & & & & 3933 sc27 & 00 38 32 -26 32.5 & & & & 2694 ddo 226 & 00 40 36 -22 31.4 & 372 & 7.5 & 59 & sc28 & 00 41 12 -17 47.8 & & & & sc29 & 00 42 43 -25 32.4 & & & & eso 411-g013 & 00 44 42 -31 51 & & & & eso 411-g018 & 00 45 35 -32 15 & 1728 & 5.2 & 161 & 1712 sc30 & 00 45 56 -25 44.5 & & & & eso 411-g019 & 00 46 41 -29 23.6 & & & & 1817 eso 540-g030 & 00 46 53 -18 20.8 & & & & 0047 - 21 &
00 47 18 -21 56.8 & & & & ddo 006 & 00 47 20 -21 17.5 & 304 & 4.5 & 58 & eso 540-g032 & 00 47 56 -20 10.8 & & & & eso 411-g027 & 00 50 26 -27 35.8 & & & & 1829 sc31 & 00 51 11 -25 44.9 & & & & eso 540-g033 & 00 51 22 -19 49 & & & & 6398 eso 351-g029 & 00 57 46 -35 48.6 & & & & eso 541-ig012 & 00 59 51 -19 56 & & & & 16954 am0106 - 382 & 01 06 06 -38 28.4 & & & & 645 eso 195-g032 & 01 07 19 -48 03.5 & & & & 7685 eso 243-g050 & 01 08 34 -42 38.5 & 1477 & 2.2 & 72 & 1490 sc32 & 01 10 27 -38 31.1 & & & & 6607 sc33 & 01 15 20 -21 56.2 & & & & sc34 & 01 19 44 -27 39.7 & & & & sc35 & 01 20 11 -26 16 & & & & eso 476-g010 & 01 24 23 -25 34.4 & 1607 & 4.2 & 134 & sc36 & 01 24 31 -15 48.5 & & & & sc37 & 01 26 32 -41 31.8 & & & & eso 542-g022 & 01 28 50 -17 57.5 & & & & 5442 sc38 & 01 30 33 -39 46.5 & & & & sc39 & 01 31 27 -29 57 & & & & eso 476-g023 & 01 31 29 -25 28 & & & & sc40 & 01 31 31 -19 19.7 & & & & eso 476-g023 & 01 31 31 -25 27.9 & & & & eso 353-g017 & 01 31 50 -34 41.9 & & & & 3820
ngc 625 & 01 32 56 -41 41.4 & 406 & 32.5 & 106 & 415 sc41 & 01 35 20 -34 48.7 & & & & sc42 & 01 37 11 -47 33.1 & 162 & 8 & 64 & mcg-03 - 05 - 014 & 01 39 09 -16 23.8 & 1632 & 5.9 & 143 & sc43 & 01 41 51 -27 55.5 & & & & eso 245-g005 & 01 42 58 -43 50.5 & 399 & 87.3 & 94 & 389 sc44 & 01 48 00 -17 02.7 & & & & phoenix & 01 49 02 -44 41.3 & 56 & 2.4 & 29 & eso 477-g012 & 01 51 28 -23 21.1 & & & & 1463 eso 477-g017 & 01 54 36 -25 46 & & & & llrclc eso 321-g018 & 12 13 10 -37 50 & & & & 3163 eso 322-g015 & 12 24 57 -37 50 & & & & eso 574-g001 & 12 25 55 -21 57.5 & & & & 6763 eso 218-g012 & 12 37 10 -51 57.5 & & & & 1809 eso 268-g033 & 12 39 45 -47 17 & & & & 5437 eso 172-g006 & 12 40 21
-52 30.5 & & & & 1910 eso 381-g020 & 12 43 16 -33 34 & 596 & 31.9 & 103 & 599 eso 443-g001 & 12 51 09 -27 36 & & & & 3187 eso 443-g006 & 12 51 55 -31 36 & & & & 3563 cen1 & 12 53 50 -42 02 & & & & eso 381-g031 & 12 55 30 -32 56 & & & & 2467 eso 323-g054 & 12 56 35 -40 42 & & & & 2944 eso 269-g024 & 12 56 40 -42 53 & & & & 3584 eso 219-g017 & 12 56 55 -47 59 & & & & 3532 eso 323-g056 & 12 57 20 -38 08 & & & & 4550 cen2 & 12 58 05 -47 26.5 & & & & 2115
cen3 & 12 58 12 -47 27.3 & & & & eso 507-g065 & 12 58 21 -25 49 & & & & 2607 cen4 & 12 58 50 -25 55 & & & & sgc1259.6 - 1659 & 12 59 33 -16 58 & 732 & 5.2 & 63 & 720 ddo 161 & 13 00 37 -17 09 & 747 & 110.1 & 136 & 750 cen5 & 13 02 02 -49 08 & 122 & 5.5 & 53 & cen6 & 13 02 12 -39 48 & 619 & 4.4 & 43 & 602 eso 219-g027 & 13 03 39 -49 35 & 1281 & 11.7 & 112 eso 219-g028 & 13 03 41 -49 25 & 1438 & 9.9 & 143 & eso 508-g004 & 13 04 10 -22 34 & & & & 2896 eso 269-g053 & 13 05 57 -42 39 & & & & 2002 eso 443-g075 & 13 06 25 -28 22 & 1425 & 25 & 172 eso 219-ig032 & 13 06 30 -47 42.7 & & & & 3489 uks 1307 - 429 & 13 07 10 -42 56 & & & & 2126 eso 443-g079 & 13 07 38 -27 42 & & & & 2138 cen7 & 13 08 24 -38 38 & & & & eso 443-g080 & 13 08 21 -27 44 & & & & 2137 eso 443-g083 & 13 10 10 -32 25 & & & & 2389 eso 323-g097 & 13 11 10 -39 00 & & & & 5010 eso 508-g030 & 13 12 12 -22 52 & 1502 & 19.3 & 151 & eso 443-g085 & 13 12 25 -32 00 & & & & 2384 eso 444-g002 & 13 14 00 -27 37 & 1636 & 10.2 & 122 & eso 382-g030 & 13 14 10 -37 24 & & & & eso 444-g006 & 13 15 43 -31 21 & & & & 3653 eso 382-g040 & 13 16 35 -33 00.5 & 1682 & 14.5 & 144 & eso 269-g087 & 13 17 20 -47 25.5 & & & & 3000 eso 382-g045 & 13 17 25 -35 46 & 1460 & 33.6 & 155 & am 1317 - 425 & 13 17 42 -42 50 & & & & 3323 cen8 & 13 20 04 -33 17 & & & & am1321 - 304 & 13 21 49 -30 42 & & & & eso 382-g061 & 13 22 41 -37 07 & 1437 & 7.3 & 105 & cen9 & 13 22 48 -29 49 & & & & 4413 eso 576-g059 & 13 23 52 -21 57.5 & 1431 & 10.2 & 118 & eso 220-g011 & 13 24 00 -48 28 & & & & 2925 eso 324-g023 & 13 24 36 -37 55 & 1439 & 71.9 & 218 & eso 324-g024 & 13 24 40 -41 13 & 526 & 52.1 & 113 & 524 cen10 & 13 26 06 -30 13.5 & & & & eso 444-g059 & 13 27 42 -31 57 & & & & eso 383-g017 & 13 30 25 -34 12 & & & & 3471 ngc 5206 & 13 30 41 -47 53.7 & & & & 571 eso 509-g059 & 13 31 05 -24 30 & & & & 4664 fourc - figu & 13 31 45 -45 16 & 831 & 201.6 & 157 & cen11 & 13 31 48 -28 58.5 & & & & 1347 ugca 365 & 13 33 42 -28 59 & 582 & 4.6 & 61 & eso 444-g084 & 13 34 30 -27 47 & 591 & 19.6 & 82 & 578 ngc 5237 & 13 34 40 -42 35.5 & 369 & 7.6 & 89 & 371 eso 324-g044 & 13 35 15 -39 35 & & & & 2532 ic 4316 & 13 37 28 -28 38 & 589 & 7.8 & 120 & 589 ngc 5264 & 13 38 47 -29 39.7 & 487 & 13.7 & 84 & 476 eso 220-g032 & 13 38 58 -48 04 & 1364 & 7.7 & 114 & eso 383-g062 & 13 39 05 -35 26 & & & & eso 383-g064 & 13 39 15 -36 06 & & & & eso 509-g096 & 13 40 07 -24 52 & & & & 5467 eso 445-g023 & 13 41 52 -29 49 & & & & 4620 eso 325-g011 & 13 42 00 -41 36 & 550 & 25.4 & 77 & 554 eso 383-g087 & 13 46 20 -35 50 & 333 & 27.4 & 68 & 342 eso 383-g091 & 13 47 36 -37 03 & 1077 & 9.4 & 155 & eso 384-g002 & 13 48 28 -33 34 & 1391 & 65.5 & 146 & eso 445-g061 & 13 49 15 -31 35 & & & & am1357 - 504 & 13 57 30 -50 48.5 & & & & -48 cen12 & 14 00 15 -50 31.5 & & & & ngc 5408 & 14 00 17 -41 09 & 506 & 65.5 & 123 & 497 eso 446-g020 & 14 06 38 -30 02 & & & & 2629 eso 579-g003 & 14 10 48 -17 45 & 1445 & 20.5 & 170 & eso 271-g025 & 14 11 50 -43 43 & 1786 & 25.2 & 116 & eso 222-g001 & 14 15 30 -47 30 & 1284 & 22.1 & 170 & eso 272-g004 & 14 16 00 -45 05 & 1653 & 20.5 & 171 & eso 579-g019 & 14 16 43 -21 31 & & & & eso 446-g053 & 14 18 22 -29 02 & 1391 & 23.2 & 169 & eso 579-g021 & 14 19 15 -18 44 & & & & eso 385-g014 & 14 19 20 -35 48 & & & & 3659 eso 222-g004 & 14 20 24 -49 26 & 1555 & 26.9 & 115 & eso 385-g019 & 14 20 45 -35 18 & & & & 3493 sgc 142448 - 4604.8 & 14 24 48 -46 04.8 & 397 & 19.5 & 68 & eso 511-g050 & 14 28 40 -25 10 & & & & 2555 eso 222-g010 & 14 31 42 -49 12.5 & 632 & 9.1 & 83 & 603 cen13 & 14 35 15 -46 40 & & & & ic 4472 & 14 36 45 -44 06 & & & & 2869 eso 272-g025 & 14 40 10 -44 29 & & & & 624 eso 222-g015 & 14 41 05 -49 11 & & & & 2254 eso 386-g022 & 14 43 52 -35 18 & & & & 6120 ic 4501 & 14 44 34 -22 12 & & & & 3319 cen14 & 14 45 45 -28 07 & & & & sgc145501 - 4730 & 14 55 05 -47 28 & 1056 & 155 & 205 & eso 223-g009 & 14 57 35 -48 04 & 593 & 96.2 & 103 & 593 ngc 5810 & 14 59 54 -17 40 & & & & 3297 eso 274-g001 & 15 10 45 -46 36.5 & 528 & 117 & 177 & 507 llr sc2 & 23 17 56 -32 10 49 & 15.3eso 347-g017 & 23 24 17 -37 37 18 & 15.4ugca 442 & 23 41 10 -32 13 58 & 82.6eso 348-g009 & 23 46 47 -38 02 55 & 12.4sc18 & 23 58 22 -41 25 40 & 6.8eso 293-g035 & 00 04 19 -42 07 01 & 10.2sdig & 00 05 41 -34 51 24 & 4.0ngc 59 & 00 12 53 -21 43 12 & 5.4eso 473-g024 & 00 28 50 -23 02 37 & 11.3sc24 & 00 34 12 -32 50 52 & 17.4ddo 226 & 00 40 36 -22 31 20 & 11.0ddo 006 & 00 47 20 -21 17 27 & 6.6am0106 - 382 & 01 06 06 -38 28 21 & @xmath192.8ngc 625 & 01 32 56 -41 41 24 & 47.8sc42 & 01 37 11 -47 33 10 & 11.8eso 245-g005 & 01 42 58 -43 50 33 & 128.4 llr eso 381-g020 & 12 43 18 -33 33 54 & 92.0sgc1259.6 - 1659 & 12 59 35 -16 58 13 & 15.0ddo 161 & 13 00 38 -17 09 14 & 317.5cen5 & 13 02 02 -49 08 00 & 15.9cen6 & 13 02 12 -39 48 00 & 12.7eso 324-g024 & 13 24 40 -41 13 18 & 150.2ngc 5206 & 13 30 41 -47 53 42 & @xmath195.5ugca 365 & 13 33 42 -28 58 54 & 13.3eso 444-g084 & 13 34 32 -27 47 30 & 56.5ngc 5237 & 13 34 40 -42 35 36 & 21.9ic 4316 & 13 37 29 -28 38 30 & 22.5ngc 5264 & 13 38 47 -29 39 42 & 39.5eso 325-g011 & 13 42 01 -41 36 30 & 73.2eso 383-g087 & 13 46 23 -35 48 48 & 79.0ngc 5408 & 14 00 18 -41 08 12 & 188.9sgc142448 - 4604.8 & 14 24 48 -46 04 48 & 56.2eso 222-g010 & 14 31 41 -49 12 12 & 26.2eso 272-g025 & 14 40 09 -44 29 36 & @xmath195.5eso 223-g009 & 14 57 35 -48 04 00 & 277.4eso 274-g001 & 15 10 45 -46 36 30 & 337.4 | we report the discovery of a large population of dwarf irregular galaxies in the two nearest groups of galaxies outside the local group , the sculptor and the centaurus a groups ( 2.5 and 3.5 mpc ) .
total areas of approximately 940 and 910 square degrees in these groups were scanned visually on the src j films to find dwarf candidates .
redshifts were obtained by an hi survey carried out at parkes , with detection limits of @xmath0 and @xmath1 , and 33 dwarf galaxies were successfully detected in the groups . a follow - up optical survey ( spectroscopy )
detected a few more , and confirmed most of the hi redshifts .
a total of 16 and 20 dwarf galaxies are found in sculptor and centaurus a , of which 6 are newly identified objects , and 5 more have newly determined redshifts . in both groups
the dwarf members show a wider spatial and velocity distribution than the brighter members . from their radial velocities and projected distances we estimate the crossing times of the groups , which confirm that they are not yet virialised .
2 3 2 2 = cmr12 = cmr9 = 11 versim#1#22.45pt = 12 # 10.16ex @xmath2 ] |
in quantum field theory , ultraviolet ( uv ) divergences always emerge from momentum integration in the loops . in order to identify the physical content of the theory under consideration ,
one has to subtract these divergences with a definite regularization procedure . among them ,
the dimensional regularization is probably the most popular one because it is easy to use and respects all the symmetry properties of the initial lagrangian @xcite .
the latter aspect is of fundamental importance and could justify by itself the choice of dimensional regularization .
nevertheless , the introduction of an explicit cut - off is sometimes advantageous , for example as soon as one considers renormalization group equations in the wilsonian approach @xcite or effective theories @xcite . in that context
, the problem concerning the non - respect of symmetries arises .
a three - momentum cut - off has been proved to be useful , e.g. , in the case of the nambu - jona - lasinio model ( see , e.g. , ref .
@xcite ) but it does not respect lorentz invariance .
this can be avoided by using a four - momentum cut - off , but the latter suffers from the fact that it violates gauge invariance , too .
this can be seen for instance from the fact that the one - loop photon polarization tensor in qed is no longer transverse .
the problem has been solved by schwinger within his proper - time approach @xcite . however , his formalism is very different from the usual feynman diagram technique . in the literature there
exists a modified version of the proper - time regularization which can directly be applied to feynman diagrams @xcite , but as shown later in this paper , the results are ambiguous since they depend on the way the propagators are written . on purely phenomenological grounds , the problem is sometimes `` solved '' by simply imposing gauge invariance at hand , i.e. , if we keep the example of the photon polarization tensor in qed , by projecting it onto the transverse part .
this method is of course not very satisfactory .
finally , another method which respects gauge invariance is that by pauli and villars @xcite .
the masses of the fictitious particles which are introduced by this method can be interpreted in the sense of a cut - off @xcite . however , as soon as the theory allows for particles having different masses propagating in the same loop , this method becomes impractible . in conclusion , a regularization scheme with an explicit cut - off dependence and preserving symmetries would be very desirable . in this paper
, we will not give the ultimate solution of this long - standing problem .
however , within the context of renormalization group equations , it is often sufficient to look at the divergent parts of the integrals .
a prescription to calculate these divergent parts within a cutoff scheme preserving symmetries is thus on its own an interesting problem .
we emphasize in particular the importance of a consistent treatment of quadratic divergences , which in the case of qed always cancel at the end of the calculation , but which survive in other theories and can give rise to a very strong running of the model parameters within the renormalization group @xcite .
we will show through several examples how we can conciliate the explicit dependence on a cut - off and the requirements of symmetry . in sect .
2 , we will start by giving an example where the proper - time method fails . in sect .
3 , we give a prescription which allows to conciliate all constraints .
then , in the remaining sections , we will give some illustrations through several examples ( qed in sect .
4 , a phenomenological chiral model in sect . 5 ) .
as already mentioned in the introduction , there are essentially three regularization methods implying a cut - off : the naive ( three or four dimensional ) momentum cut - off , the method by pauli - villars and the proper - time one . for the reasons detailed in the introduction we are concentrating on a modified version of the latter , which can immediately be applied to the calculation of feynman diagrams . however , as we are going to show , ambiguities may arise with this method . to be more explicit ,
let us start by giving two expressions which can appear in a one loop calculation : @xmath0 although both expressions are divergent , we expect a useful regularization procedure to guarantee the equality @xmath1 .
once the wick rotation is done , the proper - time regularization method as stated in the literature @xcite consists in rewriting the integrals and introducing a cut - off @xmath2 _ via _ : @xmath3 where @xmath4 is the usual euler function .
in this paper we are only interested in the divergent part of the integrals , i.e. the parts which stay finite in the limit @xmath5 . applying the proper - time method to ( [ int_a ] ) and ( [ int_b ] ) , one obtains for the divergent contributions : + @xmath6 i.e.
, the proper - time regularization procedure breaks the formal equality @xmath1 .
the logarithmic term is the same but the coefficient for the quadratic divergence depends on the way the integrals are written . of course
, one might argue that rewriting ( [ int_a ] ) in two parts is not correct , since it is divergent , and decide to forbid such a manipulation for the computation of divergences .
such a rule would be arbitrary but acceptable if one could be sure that the results are in agreement with all symmetry requirements .
actually , we will see in the following sections that this is not the case .
more generally , this example exhibits the fact that some ambiguities can arise , so that a special care has to be taken in the computation of divergences
. therefore , we have to proceed differently in order to obtain a consistent dependence on @xmath2 .
such a method is developed in the next section .
when dealing with a quantum field theory containing scalar and vector fields , we are confronted with integrals of the typical form @xmath7 here it is supposed that feynman parameters have already been introduced , such that the `` mass '' @xmath8 might depend on several parameters @xmath9 .
the integrals over the feynman parameters are omitted since they are irrelevant for our discussion .
tensor integrals with @xmath10 in the numerator are transformed using @xmath11 and analogous rules in the case of more than four indices . in the case of fermion fields ,
the same integrals appear after the traces over gamma matrices have been evaluated . for our purposes , @xmath12 and @xmath13 can take the values @xmath14 and @xmath15 , such that the integral ( [ general_int ] ) is at most quadratically divergent .
despite we want to calculate our results in @xmath16 , we leave @xmath17 unevaluated for reasons which will be explained below .
the key point is not to try to regularize all these integrals separately , but to deduce them from a single one .
we introduce @xmath18 @xmath19 and @xmath20 being some arbitrary parameters ( which will be taken equal to @xmath21 at the end of the calculation ) .
it is easy to see that all the integrals of the form ( [ general_int ] ) can be rewritten as linear combinations of @xmath22 and its derivatives .
for instance , the expressions @xmath23 and @xmath24 defined by eqs .
( [ int_a ] ) and ( [ int_b ] ) can be written as @xmath25 and @xmath26 it is straightforward to show the formal relation @xmath27 and the remaining integral @xmath28 can be evaluated in @xmath16 with the help of the proper - time method @xcite @xmath29 ( the dots represent terms which stay finite for @xmath30 ) , and hence @xmath31 for reasons which will become clear later , we have not yet set @xmath16 in the prefactor @xmath32 .
one could hope that , if one derived all needed integrals from this single one by taking derivatives with respect to @xmath19 and @xmath20 , everything would be consistent .
however , there still remains a subtlety as we will see now . from eq .
( [ aderiv ] ) we obtain @xmath33 and from eq .
( [ bderiv ] ) @xmath34 if we set @xmath16 , we see that the formal equality @xmath1 is violated .
however , we can correct this by keeping @xmath16 only in the term @xmath35 , but setting @xmath36 in the coefficient of the term @xmath37 .
if we do so , the expressions for @xmath38 and @xmath39 coincide .
this prescription can be generalized to all other integrals under consideration .
although this prescription seems to be rather arbitrary , it allows us to obtain a consistent set of regularized elementary integrals such that the final result is independent of the way the non - regularized integrals are written .
these integrals are listed in the appendix . to generalize the preceding result to loops containing derivative couplings ,
let us now consider a more complicated example : @xmath40 with the help of eqs .
( [ tensorintegral ] ) and ( [ ialphabeta1 ] ) , this integral can be written as @xmath41 note that we leave @xmath17 undetermined until the end of the calculation .
( [ ialphabeta2 ] ) , we obtain @xmath42 and then , following the prescription given above , we finally set @xmath36 in the term @xmath37 and @xmath16 in the term @xmath43 : @xmath44 this method seems very unusual concerning the way to treat the dimensionality of the divergences .
however , veltman @xcite already noticed that quadratic divergences are associated with @xmath36 in the context of dimensional regularization , whereas the logarithmic part has to be treated in dimension @xmath16 .
more recently , harada _
_ used a similar trick @xcite in order to derive the renormalisation group equations for different models of hadronic effective theories .
they show @xcite that one can not obtain correct results if one simply works in dimension @xmath16 during the calculation .
in fact , the first integrals listed in our appendix have been obtained in ref .
@xcite by calculating them in dimensional regularization , expanding the results around @xmath36 and @xmath16 , and identifying the resulting poles according to the rules @xmath45 we mention that with the help of eqs .
( [ tensorintegral ] ) and ( [ tensorintegral2 ] ) we can reproduce in this way all the other integrals in our appendix , too . within this dimensional regularization method , the remaining finite parts of the integrals can in principle be calculated , but they are independent of @xmath2 and therefore irrelevant for the renormalization group equations .
although we are not giving a formal proof , the examples in the next sections show that our method satisfies symmetry requirements automatically .
let us now come to the first example , the qed lagrangian .
in order to test the procedure previously developed , it is important to consider a simple gauge theory .
this is why we start with the regularization of the qed lagrangian in this part . with standard notations ,
the lagrangian reads : @xmath46 when one writes explicitly the gauge invariance of the generating functional , one can obtain some general relations and constraints for @xmath47-points green functions , called the ward identities .
one of these identities concerns the transversality of the photon correlation function represented in fig .
1a . [ cols="^,^,^ " , ] using standard feynman rules , it reads : @xmath48\ , .
\label{qed_se}\ ] ] after introducing the feynman parameters , we can rewrite ( [ qed_se ] ) as follows : @xmath49 g_{\mu\nu } -2 x ( 1-x ) p_\mu p_\nu}{(k^2-\delta^2)^2}\,,\ ] ] where @xmath50 . using the integrals listed in the appendix
, we finally obtain @xmath51 ( again , the dots represent terms which stay finite for @xmath30 ) , so that we recover immediately the transversality of the photon , as required by the @xmath52 gauge symmetry . a direct computation with a naive cut - off or the proper - time method does not give this result .
it should also be noticed that during the calculation quadratic divergences appear , but in the final result they cancel , as it should be .
another quite simple and remarkable consequence of gauge invariance is the relation between the vertex function @xmath53 and the electron self - energy @xmath54 , depicted in figs .
1b and 1c , respectively .
the corresponding ward identity reads : @xmath55 where @xmath56 \gamma^\mu \frac{i}{p \hbox{\hskip -0.2 true cm}\slash - k \hbox{\hskip -0.2 true cm}\slash -m } \gamma^\nu\ ] ] and @xmath57 \gamma^\mu \frac{i}{p \hbox{\hskip -0.2 true cm}\slash - k \hbox{\hskip -0.2 true cm}\slash -m } \gamma_\sigma\frac{i}{p \hbox{\hskip -0.2 true cm}\slash - k \hbox{\hskip -0.2 true cm}\slash -m } \gamma^\nu\,.\ ] ] after some manipulations on gamma matrices and using the formulas listed in the appendix , one obtains @xmath58 and @xmath59 so that the ward identity connecting three - point and two - point green functions is satisfied .
there are of course other ward identities to be checked , even at one loop level , but the examples given above are considered as significant tests for the consistency of our procedure .
we will therefore go on to our second example , which is a phenomenological chiral model .
as a last illustration , we consider in this section a model which reproduces very well the phenomenology of the hadronic vector and axial - vector correlators in the low - energy region ( see @xcite ) .
this model contains @xmath60 , @xmath61 , @xmath62 , and @xmath63 mesons as elementary fields . using the notations defined in @xcite , we write the lagrangian in the following form @xmath64 where @xmath65 are the scalar and pseudoscalar fields , and @xmath66 are the vector and axial vector fields . @xmath67 and @xmath68 are @xmath69 matrices related to the global @xmath70 chiral symmetry ,
fulfilling the commutation relations @xmath71 = i \varepsilon_{ijk}\ , t_k\ , , \qquad [ t_i\,,t_j^5 ] = i \varepsilon_{ijk}\ , t_k^5\ , , \qquad [ t_i^5\,,t_j^5 ] = i \varepsilon_{ijk}\ , t_k\ , .
\label{tcommutators}\ ] ] experimental information about @xmath62 and @xmath63 mesons is obtained from electromagnetic processes and @xmath72 decay data . as a consequence ,
it is necessary to introduce the photon and the @xmath73 fields and thus replace the ordinary derivatives by covariant ones , following @xmath74 in order to determine the photon self - energy , one has to evaluate the different contributions depicted in ref .
@xcite figures 10 and 11 .
thanks to the method described in the preceding section , it is possible to show after some tedious calculations that the total @xmath75 self - energy is simply proportional to @xmath76 , as it should be .
furthermore , we have explicitly checked that this result can not be obtained by a direct use of the proper - time method .
another important test of our method is the goldstone theorem . if there is no explicit symmetry breaking term ( @xmath77 in our case ) , but the global symmetry is spontaneously broken in the vacuum , then there has to be a massless particle which corresponds to the pion in our case
. moreover , we have to take into account the mixing between @xmath60 and @xmath63 .
this calculation is highly non trivial ; one has to evaluate all the diagrams depicted in ref .
@xcite ( see figure 6 ) .
writing the self - energy for the pion in this model , one finally obtains in the chiral limit ( @xmath78 ) @xmath79 i.e. , the pion self - energy is simply proportional to @xmath80 and therefore does not destroy the goldstone character of the pion .
it is also interesting to note that , as in the case of the photon , all quadratic divergences have canceled in the final result .
the above expression for @xmath81 does not include the resummation of the @xmath82 mixing term . when including this mixing we also recover the goldstone theorem ( @xmath83 ) , but the expression for the total pion self - energy is far more complicated , such that we refrain from giving the full expression here .
again , this result is non trivial and we should note that , up to now , it has only been obtained in a dimensional regularization approach ( see @xcite ) . in all these examples
, we have seen that the way of computing divergences described in the third section respects symmetry requirements .
we have seen in this paper how to handle logarithmic and quadratic divergences in a cut - off regularization scheme in a consistent way , i.e. , in a way satisfying constraints from symmetry requirements .
for example , we have shown that our approach was preserving the transversality of the photon polarization tensor in qed as well as the transversality of the photon polarization tensor and the goldstone theorem in the case of a phenomenological chiral model including @xmath60 , @xmath61 , @xmath62 , and @xmath63 mesons . in all these examples ,
the quadratic divergences cancel at the end of the calculation due to the symmetry properties of the model .
recently our method has also been applied to qed with one extra dimension @xcite , where the quadratic divergences survive and give rise to a power - law in the running of the effective four - dimensional gauge coupling .
we have to stress , however , that the present method is only designed to compute the divergent parts of the integrals .
our aim is now to use this regularization scheme in order to derive renormalization group equations for the phenomenological chiral model described in the last section .
in fact , it is interesting to incorporate the scalar degree of freedom to extend the results of harada _
@xcite concerning the chiral symmetry restoration .
we gratefully acknowledge stimulating discussions with m. ericson , g. chanfray and r. stora .
we present in this appendix a detailed and complete list of regularized integrals necessary for the calculation of all integrals described in the main text and in @xcite .
as already mentioned , only the divergent parts are given .
% \int \frac{d^4k}{i(2\pi)^4}\frac{1}{(k^2- m^2)^2 } = & \frac{1}{(4\pi)^2}\ln\lambda^2+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^2}{(k^2- m^2)^2 } = & -\frac{1}{(4\pi)^2}\left [ \lambda^2 - 2 m^2\ln\lambda^2 \right]+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^\mu k^\nu}{(k^2- m^2)^2 } = & -\frac{g^{\mu\nu}}{2(4\pi)^2}\left [ \lambda^2 - m^2\ln\lambda^2 \right ] + \dots \label{itensor } \\
% \int \frac{d^4k}{i(2\pi)^4}\frac{k^2}{(k^2- m^2)^3 } = & \frac{1}{(4\pi)^2}\ln\lambda^2+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^\mu k^\nu}{(k^2- m^2)^3 } = & \frac{g^{\mu\nu}}{4(4\pi)^2}\ln\lambda^2+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^4}{(k^2- m^2)^3 } = & -\frac{1}{(4\pi)^2}\left [ \lambda^2 - 3 m^2\ln\lambda^2 \right]+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^2k^\mu k^\nu}{(k^2-m^2)^3 } = & -\frac{1}{4(4\pi)^2}\left [ 2 \lambda^2 -3 m^2\ln\lambda^2 \right]+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^\mu k^\nu k^\rho k^\eta}{(k^2- m^2)^3 } = & -\frac{1}{8(4\pi)^2}\left[g^{\mu\nu}g^{\rho\eta}+ g^{\mu\rho}g^{\nu\eta}+g^{\mu\eta}g^{\nu\rho } \right ] \left [ \lambda^2 - m^2\ln\lambda^2 \right ] + \dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^4}{(k^2- m^2)^4 } = & \frac{1}{(4\pi)^2}\ln\lambda^2+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^2 k^\mu k^\nu}{(k^2- m^2)^4 } = & \frac{g^{\mu\nu}}{4(4\pi)^2}\ln\lambda^2+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^\mu k^\nu k^\rho k^\eta}{(k^2- m^2)^4 } = & \frac{1}{24(4\pi)^2}\left[g^{\mu\nu}g^{\rho\eta}+ g^{\mu\rho}g^{\nu\eta } + g^{\mu\eta}g^{\nu\rho } \right ] \ln\lambda^2+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^6}{(k^2- m^2)^4 } = & -\frac{1}{(4\pi)^2}\left [ \lambda^2 - 4 m^2\ln\lambda^2 \right]+\dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^4 k^\mu k^\nu}{(k^2- m^2)^4 } = & -\frac{g^{\mu\nu}}{2(4\pi)^2}\left [ \lambda^2 - 2 m^2\ln\lambda^2 \right ] + \dots \\ % \int \frac{d^4k}{i(2\pi)^4}\frac{k^2 k^\mu k^\nu k^\rho k^\eta}{(k^2- m^2)^4 } = & -\frac{1}{24(4\pi)^2}\left[g^{\mu\nu}g^{\rho\eta}+ g^{\mu\rho}g^{\nu\eta } + g^{\mu\eta}g^{\nu\rho}\right ] \left [ 3 \lambda^2 - 4 m^2\ln\lambda^2 \right]+\dots \end{aligned}\ ] ] g. t hooft and m.j.g .
veltman , nucl . phys .
b * 44 * ( 1972 ) 189 .
s. weinberg , _ the quantum theory of fields _ , vol . 1 : _ foundations _ ( cambridge university press , 1995 ) .
s. weinberg , _ the quantum theory of fields _ , vol . 2 : _ modern applications _ ( cambridge university press , 1996 ) .
m. harada and k. yamawaki , phys . rep .
* 381 * , 1 ( 2003 ) .
h. georgi , ann .
nucl . part .
* 43 * ( 1993 ) 209 .
s. p. klevansky , rev .
phys . * 64 * ( 1992 ) 649 . j. schwinger , phys . rev .
* 74 * ( 1948 ) 1439 .
j. zinn - justin , _ quantum field theory and critical phenomena _ ( oxford science publications , 1993 ) .
w. pauli and f. villars , rev . mod .
* 21 * ( 1949 ) 434 . c. itzykson and j.b .
zuber , _ quantum field theory _
( mc graw - hill , 1980 ) .
dienes , e. dudas , and t. gherghetta , nucl .
b * 537 * ( 1999 ) 47 .
t. varin , j. welzel , a. deandrea , and d. davesne , arxiv : hep - ph/0610130 ( 2006 ) .
veltman , acta phys .
b * 12 * ( 1981 ) 437 .
m. harada and k. yamawaki , phys .
d * 64 * ( 2001 ) 014023 .
m. harada and k. yamawaki , phys . rev .
* 87 * ( 2001 ) 152001 .
y. hidaka , o. morimatsu and m. ohtani , phys .
d * 73 * ( 2006 ) 036004 .
m. urban , m. buballa and j. wambach , nucl .
a * 697 * ( 2002 ) 338 . | we present a prescription to calculate the quadratic and logarithmic divergent parts of several integrals employing a cutoff in a coherent way , i.e. in total agreement with symmetry requirements .
as examples we consider one - loop ward identities for qed and a phenomenological chiral model . |
the experimental measurement setup is shown in fig . s1 .
two - tone spectroscopy or dispersive readout is a standard measurement technique which uses the dispersive shift of a mode to precisely determine the frequency of the qubit @xcite . in our massively multimode strong coupling experiment ,
such spectroscopy shows that the mode frequency is affected not only by the qubit state , but also by populating the modes nearby to the qubit .
this is yet another manifestation of mode - mode interactions mediated via the qubit .
the magnitude of transmission , a proxy for the dispersive shift , is indicated by the colorbar in fig .
s2 . as the qubit frequency approaches a mode , the strength of interaction between that mode and the measurement mode steadily increases .
this interaction is so strong that the measurement mode shifts substantially even when the qubit is over 200 mhz away .
thus the frequency of the measurement mode needed to be recalibrated for each flux voltage in fig . s2 .
furthermore , to track the qubit in this frequency range with sufficient contrast , the designated measurement mode needed to be changed such that it was farther and farther away from the qubit . for each change of measurement mode , there is a horizontal black line in fig .
we find that ultranarrow fluorescence is very robust to the flux bias voltage of the transmon qubit as shown in fig .
s3 , where we display the linewidth as a function of the flux bias voltage applied to the qubit .
this observed robustness will be discussed in the following theoretical analysis .
also displayed is the integrated fluorescence power against the same bias .
the power and linewidth are seen to be inversely related .
the drive frequency and power here are both fixed to that which results in the minimum linewidth in fig .
4 of the main text .
in order to better interpret the intriguing dynamical phenomena observed in our experiment , here we provide a simplified model to illustrate key aspects of the physics . even though the model is far from a complete representation of the experimental system , we argue that qualitatively similar physics emerges in the massively multimode system .
@xmath46 + \frac{\gamma}{2 } \left ( 2\sigma^{-}\rho\sigma^{+ } - \sigma^{+}\sigma^{-}\rho - \rho\sigma^{+}\sigma{- } \right ) \:+\ : \sum_{m } \frac{\kappa_m}{2 } \left ( 2a_m\rho a^{\dagger}_m - a^{\dagger}_ma_m\rho - \rho a^{\dagger}_ma_m \right ) \ , \\ h & = h_0 + i\eta \left ( a^{\dagger}_{r } - a_{r } \right ) \ , \\ h_0 & = \sum_{m } \ : \delta_m \ : a_{m}^{\dagger}a_{m } + \delta_a \ : \sigma^{z } + g_m \left ( a^{\dagger}_{m}\sigma^{- } + \text{h.c . } \right ) \ , \end{aligned}\ ] ] with @xmath32 the photon loss rate for mode @xmath3 , and @xmath47 the qubit spontaneous decay rate .
the above hamiltonian is written in the rotating frame of the drive , and the rotating wave approximation ( rwa ) has been used .
@xmath48 is the detuning between the @xmath3th mode and the drive .
likewise , @xmath49 is the detuning between the qubit and the drive .
rwa is a good approximation if we focus on high harmonics ( @xmath50 in the experiment ) of this multimode system .
@xmath51 is the amplitude of the external drive , and the drive is assumed to be resonant with the @xmath52th mode .
steady state photon numbers of other modes in the uncoupled system ( @xmath53 ) , are suppressed relative to the resonant term , by multiples of @xmath54 , and we are justified in keeping only the resonant drive . applying an active drive and dissipation dependent unitary displacement to the state of the system , via @xmath55 , with @xmath56 transfers the drive in the master equation onto the qubit .
the hamiltonian appearing in the master equation is now @xmath57 , where @xmath58 is the rabi amplitude .
we could in fact directly implement such a drive in our experiment by pumping the flux bias line to qubit . for the sake of simplicity
we start from this form of the hamiltonian ( externally driving the qubit ) in what follows , and study a restricted two mode system , also taking the drive to be resonant with the qubit .
we disregard the mode resonant with the qubit , focusing on the more interesting sideband phenomena .
@xmath59 + \frac{\gamma}{2 } \left(2\sigma^{-}\rho\sigma^{+ } - \sigma^{+}\sigma^{-}\rho - \rho\sigma^{+}\sigma{-}\right ) \:+\ : \sum_{m=1,2 } \frac{\kappa_m}{2 } \left(2a_m\rho a^{\dagger}_m - a^{\dagger}_ma_m\rho - \rho a^{\dagger}_ma_m \right ) \ , \\ h & = \delta_1 \ : a_{1}^{\dagger}a_{1 } + \delta_2 \ : a_{2}^{\dagger}a_{2 } + \frac{\omega}{2 } \left(\sigma^{+}+\sigma^{-}\right ) + g_1\left(a^{\dagger}_{1}\sigma^{- } + \text{h.c.}\right ) + g_2 \left(a^{\dagger}_{2}\sigma^{- } + \text{h.c . } \right ) \ .\end{aligned}\ ] ] without cavity modes , the qubit dynamics result in the famous mollow triplet .
while the single mode case , for a special choice of parameters , can be solved analytically ( @xmath60 ) @xcite , this is in general not possible for a general multimode system .
we thus resort to numerical simulation based on the above equation in the following discussion .
to motivate the numerical exploration of the two mode master equation , we first analyze some properties of this hamiltonian to gain intuition . introducing a new basis for the qubit by defining @xmath61 , yields the hamiltonian to further simplify the analysis , as in the rwa , we truncate the taylor expansion to remove fast oscillating terms . these approximations are only valid in the limit of @xmath67 , and near certain resonance conditions .
we thus arrive at a finite order multimode multiphoton hamiltonian .
for example , let us assume that @xmath68 , and that the qubit is resonant with the drive frequency
. in this special case , the resonance conditions are simply @xmath69 , for @xmath70 .
the effective hamiltonians for the two different rabi rates ( @xmath60 and @xmath71 ) become @xmath72 \ , & m=1\\ \label{two - photon - ham } h_{eff}^{(\omega=2\delta ) } & = \delta
a_{1}^{\dagger}a_{1 } - \delta a_{2}^{\dagger}a_{2 } + \delta\tilde{\sigma^{z } } + \frac{g^2}{2\delta}[-\tilde{\sigma}^{+}a_1 ^ 2 + \tilde{\sigma}^{+}{a^{\dagger}_2}^2 + \text{h.c . } ] \ . &
m=2\end{aligned}\ ] ] note that two photon processes in the second hamiltonian above are supressed relative to the single photon ones in by @xmath73 .
terms like @xmath74 arise because the second mode carries negative frequency in the rotating frame .
the second hamiltonian is the basis for two photon lasing of driven atoms in a cavity @xcite . in our case where only a single qubit is present
, keeping high - order quantum correlations beyond semiclassics is essential in explaining relevant coherence properties of the fluorescence light .
the full master equation could be derived by applying these transformations on the dissipative part , giving rise to complicated multimode multiphoton dissipative terms .
we instead jump directly to the numerical results derived from the two mode master equation .
we use the monte carlo wavefunction approach to numerically simulate the dynamics of this driven dissipative multimode system . in our simulation
, the size of the hilbert space needed grows as @xmath75 , with @xmath76 the photon cutoff for mode @xmath3 .
the rapid growth of the dimension of the hilbert limits the number of modes that can be incorporated into the simulation .
expectation values of relevant operators are followed and steady state values for the driven system are presented below . to compute two time correlation functions such as @xmath77 using this method , we have to employ a two step procedure @xcite , equivalent to the quantum regression theorem , but at the expense of requiring large numbers of walks to produce reliable results .
to explain the lobe structure and multimode fluorescence observed in our experiment , we first analyze two special limits of the two mode model . resonant enhancement by multiphoton processes in a single mode @xmath78 ( via the interaction @xmath79 , while tuning the qubit drive @xmath18 , leads to peaks in the steady state photon number near subharmonics of the rabi rate @xmath80 . this is the physical origin of the lobe structure at far detuned modes in our massively multimode system , as @xmath81 is a good approximation for these modes .
s4 presents numerical results for the special two mode case where the two modes are situated symmetrically around the qubit ( and the drive which is resonant with it ) .
fluorescence at these two modes are equally and simultaneously enhanced .
moreover , the two peaks in photon number and qubit polarization @xmath5 deviate from the bare resonance conditions @xmath82 . the qubit inversion @xmath83 can go above @xmath84 .
the results for a single mode model are also shown for comparison .
while multiphoton resonances are apparent in both cases , the two mode system reaches much larger photon numbers in the steady state .
the difference also appears in the spectral linewidth ( see fig
while the single mode fluorescence shows no spectral narrowing , the spectral linewidth in the two mode model can be much smaller than the natural linewidth . beside the simple two photon interaction in a single mode ( @xmath85 ) analyzed above , the interaction vertex coupling two different modes ( @xmath86 ) will result in resonant enhancement in modes near harmonics of the rabi rate @xmath87 .
this is demonstrated in fig . [
fig : photon number_ii ] where another special two mode case ( @xmath88 ) is investigated .
similarly to the previous case , this effect leads to co - enhancement of fluorescence at two modes , as evidenced by the brightening of the second mode at @xmath89 . as this is a second - order effect induced by the first mode
, @xmath90 is much smaller than @xmath91 . for stronger coupling , we would expect the difference to be smaller . in our strong coupling multimode system
, all multimode multiphoton terms have to be taken into account , which makes the analysis very cumbersome .
the competition between these processes will generally make resonances broader and less prominent .
the numerical simulations qualitatively agree with our experimental results .
ultranarrow resonance fluorescence has been predicted for the single mode cavity @xcite , where coherence between dressed states of the driven system arises from coupling to common vacuum fluctuations .
the resultant spectrum linewidth of the mode could be much smaller than the natural linewidth of the bare system .
in fact , the linewidth of the sideband is proportional to the inverse of the steady state photon number @xmath92 .
we provide numerical evidence that this is also true in the multimode system .
it is worth noting that reference @xcite only considered a special value of the parameters , where @xmath93 and @xmath94 .
here we generalize their result to explain our observations . in fig .
s6 , for the symmetric two mode system , it is shown that spectral narrowing appears for a range of the rabi rate @xmath18 .
moreover , the linewidth is inversely proportional to the photon number ( see also fig .
s4(a ) ) such that the linewidth reaches its minimum as the photon number reaches its maximum at @xmath95 .
the left mode also shows the same amount of spectral narrowing ( data not shown ) . the spectral narrowing in this simplified model does nt reach that of our experiments ( a factor of @xmath96 in the simulation vs. @xmath97 in the experiment ) , yet
it displays qualitively similar features .
moreover , the relation @xmath98 suggests that the linewidth could be made much narrower than that of the current simplified model , when reaching larger steady state photon numbers . for the interacting massively multimode system
, the proliferation of the dressed states in the high excitation manifolds leads to enhanced generation of coherence among the dressed states from a collective interference , which relies on the near resonance of many dressed state level spacings .
concomitant with the collective suppression of dissipation for the steady state then is the rise in the steady state occupation .
simulations of such a massively multimode system demands more computational resources due to the growth of the hilbert space .
more extensive results will be presented elsewhere .
spectral narrowing is generally very insensitive to the qubit frequency . for finite drive - qubit detuning @xmath99 ,
the effective rabi rate is @xmath100 , with @xmath101 the amplitude of the drive .
based on the above analysis , spectral narrowing occurs when @xmath18 is within the range @xmath102 , where @xmath103 is again the mode frequency measured with respect to the drive ( i.e. its detuning ) , and @xmath104 is generally determined by the number of modes , the coupling strength and the dissipation . for relatively small detuning @xmath99 and large rabi rate @xmath101 , this gives a large effective range of qubit frequency @xmath105 over which narrowing occurs .
it is worth noting that even if the drive is not resonant with the qubit and the two modes are not symmetric around the drive , spectral narrowing still exists in our simulation . for a single mode system
, it was argued that spectral narrowing requires that the qubit decay rate @xmath106 be comparable to the coupling strength @xmath107 , i.e. the moderate coupling regime @xcite .
we therefore study the effect of qubit decay rate @xmath106 on steady state properties of the two mode system .
s7(a ) presents photon number @xmath91 versus rabi rate for a different @xmath106 value . from this result
, we conclude that spectral narrowing survives when going from moderate coupling to the strong coupling regime , which we attribute to the fact that the eigenstates of the multimode system in high excitation manifolds are more densely spaced than in the single mode system .
dynamically generated coherence between these closely spaced states then results in spectral narrowing of many modes simultaneously . | the study of light - matter interaction has seen a resurgence in recent years , stimulated by highly controllable , precise , and modular experiments in cavity quantum electrodynamics ( qed ) @xcite . the achievement of strong coupling @xcite , where the coupling between a single atom and fundamental cavity mode exceeds the decay rates , was a major milestone that opened the doors to a multitude of new investigations @xcite . here
we investigate multimode strong coupling ( mmsc ) @xcite , where the coupling is comparable to the free spectral range ( fsr ) of the cavity , i.e. the rate at which a qubit can absorb a photon from the cavity is comparable to the round trip transit rate of a photon in the cavity .
we realize , via the circuit qed architecture @xcite , the first experiment accessing the mmsc regime , and report remarkably widespread and structured resonance fluorescence , whose origin extends beyond cavity enhancement of sidebands @xcite .
our results capture complex multimode , multiphoton processes , and the emergence of ultranarrow linewidths . beyond the novel phenomena presented here , mmsc opens a major new direction in the exploration of light - matter interactions .
interest in going beyond strong coupling has focused on the ultrastrong coupling limit , where the breakdown of the rotating - wave approximation results in excitation non - conserving terms @xcite .
in contrast , the direction which we pursue is the simultaneous strong coupling of the qubit to numerous modes , leading to qubit mediated mode - mode interactions and many - body physics not present in the single mode problem . unlike the spin - boson problem , where the continuum bosonic modes are treated as a bath for the qubit , in mmsc the dynamics of individual modes and finite time correlations between modes are essential .
furthermore , the study of mmsc can not rely on the integrability present in the single mode and continuum problems @xcite , thus requiring new theoretical ideas for its study .
the closed system is described by the hamiltonian : @xmath0 where @xmath1 is the qubit frequency , @xmath2 is the pauli operator , and @xmath3 represents cavity mode number . here , @xmath4
( @xmath5 ) and @xmath6 ( @xmath7 ) are mode ( qubit ) raising and lowering operators .
the coupling strength of the qubit to the @xmath3th harmonic is @xmath8 , where @xmath9 is the coupling rate to the fundamental cavity mode @xcite . to achieve mmsc , the coupling between qubit and cavity must be comparable to the free spectral range , which is made possible through use of a long cavity .
previously , long cavities have been used for novel comb generation , with a kerr nonlinearity provided by the bulk medium @xcite . in cavity
qed , unique multimode platforms have been studied , where degenerate modes are coupled via ensembles of atoms @xcite .
in contrast , here the nonlinearity is provided by a single qubit . in our setup ,
the coupling to the fundamental mode @xcite , @xmath9 , scales with cavity length @xmath10 as @xmath11 .
holding constant the qubit frequency , we are interested in the @xmath12 harmonic of the cavity which is nearly resonant with the qubit .
the ratio of the coupling for this mode to the fsr then scales as @xmath13 .
as there are many modes simultaneously interacting with the qubit , many modes can be coupled to each other during the typical lifetime of an excitation in the system , with the effective qubit mediated interaction rate between a pair of modes being @xmath14 , leading to a delocalization of the excitation across many modes .
[ fig1 ] = 92 mhz , is realized by fabricating a 0.68 m coplanar waveguide resonator on a 25 mm x 25 mm sapphire substrate .
capacitive coupling at the input and output ports allows coupling of radiation into and out of the resonator , at rates @xmath15 @xmath16 0.5 - 3 mhz .
red rectangles show coupling capacitors .
blue rectangle encloses qubit .
* b * , symmetric interdigitated capacitors couple cavity to transmission lines .
* c * , a transmon qubit is capacitively coupled to the center pin near the output capacitor , an antinode for all modes .
squid geometry allows for tuning qubit frequency , @xmath1 , via the flux bias line ( shown ) or the external magnet @xcite .
the grid - like pattern in ground plane pins flux vortices .
scale bars denote ( a ) 10 mm , ( b ) 100 @xmath17 m , and ( c ) 100 @xmath17m.,title="fig : " ] [ fig2 ] within this rich domain , we first focus our attention on resonance fluorescence , or the response of a qubit to coherent drive . strongly driving a qubit in free space results in emission forming a three peaked structure known as the mollow triplet , composed of a center peak and two symmetric sidebands @xcite .
the sidebands are displaced linearly from the central peak by the rabi amplitude @xcite @xmath18 .
this phenomenon , also observed experimentally with superconducting qubits @xcite , is typically explained using the dressed state picture approach @xcite .
coupling of a qubit to a single mode cavity strongly modifies the mollow triplet @xcite . for weak coupling ,
the cavity passively filters the fluorescence , while at strong coupling , the sideband width is proportional to the coupling @xcite .
we design a very low frequency cavity such that the @xmath1 of a standard qubit falls near a high harmonic ( @xmath16 50th-75th ) of the cavity , unlike traditional circuit qed experiments .
additionally , this approach allows us to adapt to the typical measurement range ( 3 - 8ghz ) of existing circuit qed setups ( see methods ) .
the superconducting microwave cavity comprises a 0.68 m transmission line with identical capacitors on either end , resulting in a fundamental frequency , @xmath19mhz ( see fig . 1 and methods section for details ) .
we couple a flux - tunable transmon qubit @xcite near one end of the resonator , an anti - node for all the modes of the cavity , and achieve a qubit - mode coupling strength for the 75th cavity harmonic exceeding 30 mhz .
coherent drive , at frequency @xmath20 , is introduced only via the input port ( fig .
1b ) , while radiative decay of cavity photons occurs via both input and output ports . a unitary transformation maps the coherent drive from the cavity to the qubit , yielding a qubit driving term @xmath21 , with @xmath18 the rabi drive amplitude ( see supplement ) .
the mode linewidths , @xmath15 , span from 0.5 - 3 mhz to provide reasonable separation between modes .
the qubit spontaneous decay rate , @xmath22 , is of the order of 1.6 mhz , due to the multimode purcell effect @xcite .
measuring transmission while tuning the qubit frequency reveals that the qubit strongly couples to a vast number of modes ( fig .
2a ) @xcite .
in fact , we are able to follow the trail of avoided crossings for several ghz as the qubit energy tunes through the uniformly dense mode structure .
the smooth transitions between subsequent avoided crossings show that the qubit is perpetually near resonance with some mode and as such the dispersive approximation is never valid . in figs .
2b - c , the marked difference in the magnitude of the avoided crossings , and thus coupling strength , is apparent .
we apply a fit to find @xmath23 mhz , giving rise to @xmath24 32 mhz .
[ fig3 ] in fig .
3 we observe fluorescence across more than 50 modes , when the drive and qubit are resonant with a high harmonic of the cavity .
enhanced fluorescence is observed at mode @xmath3 when it is resonant with the rabi sideband .
as the displacement of mollow sidebands is proportional to drive amplitude @xmath18 , this occurs @xcite when detuning from the @xmath3th mode @xmath25 . for far detuned modes the drive power needed to reach the peak fluorescence at these modes @xmath26 , and hence @xmath27 ( fig .
3b ) . at fixed drive power
, emission is spread over many modes , with the simultaneous enhancement at many nearby modes , differentiating the multimode fluorescence from the single mode case . with varying drive power , far detuned modes exhibit a multi - lobed fluorescence structure ( e.g. fig .
3c ) .
the drive power for the peak of the secondary lobe approaches twice that of the first lobe for increasing detuning @xmath28 ( fig .
3d ) , originating from multiphoton decays .
two photon decay processes in single mode cavities , where @xmath29 , have previously been observed @xcite . for mmsc , the multiphoton decay processes can progress via combinations of many different modes .
the resonance condition for single and multiphoton processes is thus shifted away from @xmath30 , for @xmath31 ( see supplement ) . at smaller detuning ,
the drive power for the peak fluorescence deviates from a quadratic dependence on detuning .
multimode effects become increasingly dominant as we approach modes closer to the qubit , with the appearance of a different intricate multi - lobed structure .
when the resonant mode is very weakly driven , there is no incoherent scattering to the other modes .
however , at a higher drive power , fluorescence appears simultaneously at multiple nearby modes ( three of which are displayed in fig .
3e - g ) .
in addition to detecting bright fluorescence at these nearby modes ( three on either side of the center mode ) we also observe incoherent scattering at the center mode itself .
furthermore , there is a change in emission frequency at each of these modes , approaching the bare cavity mode from above with increasing drive power .
this shift is qualitatively the same for modes on either side of the drive , and can not be attributed to a simple ac stark shift . as drive power
is further increased , the simultaneous fluorescence at the nearby modes fades away . at an even higher power , the emission at these modes reappears , but at the bare cavity mode frequencies .
this power coincides with that at which the far detuned modes fluoresce , suggesting the emergence of qubit mediated massively multimode correlation .
a remarkable observation of our experiment is the appearance of ultranarrow linewidths in emission .
the fluorescence at far detuned modes , captured in fig . 3 , have linewidths that are roughly equal to that of the bare modes .
for modes nearest the qubit the fluorescence linewidth is smaller than @xmath32 , e.g. for the nearest mode above being respectively 650 khz vs. 1.1 mhz .
detuning the drive from the center mode unveils here a linewidth narrowing by over an order of magnitude to 65 khz ( fig .
4a ) .
the narrowest fluorescence occurs when the drive is detuned by @xmath33 mhz from the resonant mode .
this unexpected behavior is robust to qubit frequency ; the narrowed fluorescence produced with the @xmath33 mhz detuned drive remains even as the qubit frequency is tuned over @xmath34 mhz away from the cavity mode ( see supplement ) . ) for varying drive frequency ( @xmath35 ) , at fixed drive power , reveals linewidth narrowing .
minimum linewidth , 65khz , corresponds to @xmath36 mhz . *
b * , normalized spectra along dashed lines in ( a ) are plotted in blue , vertically displaced . for comparison , low power transmission is plotted in solid gray and the bare mode frequency is denoted with a vertical gray line .
* c * , emitted power ( over @xmath16 3 mhz bandwidth ) vs. @xmath35 shows simultaneous brightening of @xmath37 modes . *
d * , fluorescence linewidths vs. @xmath35 .
linewidths vary concurrently for @xmath37 and are inversely proportional to emitted power . for comparison ,
dashed lines show @xmath38.,width=331 ] we attribute this narrowing to the spontaneous generation of coherence arising from the coupling of many dressed states through a common vacuum .
such behavior has been predicted and observed in atoms with multiple closely spaced bare atomic levels coupled to vacuum or a single mode of a cavity @xcite , where the theoretical analysis is aided by the simpler dressed state structure . in our setup , unlike the single mode case , the number of states in a single excitation manifold is not a constant but rather grows as @xmath39 , where @xmath40 is the number of modes and @xmath41 is the excitation manifold .
the rapid growth of the number of states per manifold hinders a simplified analysis via the dressed state picture , but makes possible the near resonance of many level spacings ( bohr frequencies ) , leading to a collective enhancement of the coherences from their mutual couplings .
an external coherent drive allows us to dynamically access regions of the dressed states with large generated coherences , as evident in the drive dependence of the narrowing in fig .
4 . in the region where narrowing is observed , the linewidth and steady state photon number
are observed to be inversely proportional , as has been previously suggested for the case of a single mode @xcite . for our multimode cavity ,
the narrowing is in fact observed simultaneously for multiple modes at the same drive frequency , as depicted in figs .
4c - d for the two modes nearest the drive .
while the modes directly neighboring the center mode exhibit the sharpest linewidths of @xmath42 khz , the next nearest modes also display a narrowed fluorescence , @xmath43 khz .
the multi - lobed power dependent structure of fluorescence also appears at the optimal detuned drive frequency , however , the linewidth of the second lobe is of order @xmath32 , for the nearby modes . for more distant modes , while fluorescence follows a similar @xmath44 dependence , there is no narrowing apparent .
finally , the drive dependent narrowing is not unique to these described experimental parameters .
when we translate the drive and qubit to another set of modes , we observe equally narrow linewidth fluorescence which indicates that spontaneous generation of coherence is generic in this system . in this letter , we realized a new regime of circuit qed , multimode strong coupling .
we demonstrated the creation of many - body steady states with long lived quantum coherence , as evidenced by multimode resonance fluorescence with ultranarrow linewidths .
dynamically generated quantum correlations between modes in this driven dissipative system is expected to lead to qualitatively new physics @xcite .
while this new complex regime poses many theoretical challenges , precise circuit qed experiments can guide this exploration .
we hope that access to multimode strong coupling will lead to many new advancements in circuit qed and our understanding of light - matter interactions .
* + methods + * the mmsc device was made using a combination of standard optical and electron - beam lithography techniques on a 25 mm x 25 mm sapphire wafer .
the cavity and larger features were etched from a 200 nm thin film of sputtered niobium , while the qubit was made with evaporated aluminum using the dolan bridge technique .
the cavity has a center pin to gap ratio of 10:4.186 .
the transmon qubit @xcite is capacitively coupled near the end of the cavity , at an antinode .
the qubit consists of two large islands and a pair of parallel junctions in a squid - loop geometry .
while there is an on - chip flux bias line , a superconducting magnet was installed around the device to tune the qubit frequency without detrimental thermal heating .
the device was extensively wire - bonded to the circuit board to ensure proper connection of all ground planes and reduce the effect of spurious modes .
the device was cooled to 11 mk in a dilution refrigerator , and measured using standard low - temperature microwave circuit elements including an iq mixer , isolators , and a hemt amplifier . +
* acknowledgements + * we thank c. eichler , s. m. girvin , j. koch , d. i. schuster , w. e. shanks , s. j. srinivasan , and j. m. taylor for helpful conversations and technical support . * + funding + * this work was supported by the us army research office ( w911nf-11 - 1 - 0086 ) , the us national science foundation through the princeton center for complex materials ( dmr-0819860 ) and career awards ( grant nos .
dmr-0953475 & dmr-1151810 ) , the david and lucile packard foundation , and the dod through the ndseg program .
there are no competing financial interests . *
+ author contributions + * n. m. s. fabricated the device , led the experiments , analyzed the data , and wrote the manuscript .
y. l. and d. s. provided key contributions in the experiments and data analysis , helped in writing the manuscript , and performed the numerical simulations .
l. j. s. contributed to cavity design .
d. u. provided technical advice and built the superconducting magnet .
m. m. and h.e.t . provided theoretical support .
a. a. h. advised and supervised the project .
10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 _ _ .
oxford graduate texts ( , ) . , & . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _
_ _ * * , ( ) . & .
_ _ * * , ( ) . & _
_ * * , ( ) . & , .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) .
& _ _ * * , ( ) .
, , , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
, , , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) . , & _ _ .
wiley - interscience publication ( , ) .
_ et al . _ . _
_ * * , ( ) . .
_ _ * * , ( ) . .
_ _ * * ( ) .
_ et al . _
_ _ * * , ( ) .
& . _
_ * * , ( ) .
_ et al . _ .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . , & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) .
* supplementary material * |
quantum brownian motion is much more involved than its classical analog since in general tractable equations of motion do not exist @xcite .
progress can be made in two limiting ranges , namely , in the realm of weak dissipation and in the opposite one of strong friction . in the former case
a perturbative treatment has led to a variety of markovian weak - coupling master equations @xcite , among them the famous lindblad @xcite and the redfield @xcite equations .
successful applications include quantum optical systems , decoherence for solid - state based quantum bits and nonadiabatic dynamics in molecular systems , to name but a few .
strong friction has been explored only recently @xcite with a growing amount of research since then @xcite . there , the quantum smoluchowski equation , a sort of markovian master equation as well , allows to investigate condensed phase dynamics at lower temperatures e.g. in soft matter and mesoscopic systems @xcite . a formally exact description of open quantum systems valid for all temperatures and damping strengths is provided by the path integral approach initiated by the work of feynman and vernon @xcite and developed in detail in the 1980s @xcite .
the approach has been utilized in numerous applications especially in condensed phase systems , e.g. to reveal the non - exponential decay of low temperature correlation functions , and has further been exploited to consistently derive the markovian master equations mentioned above @xcite .
a real challenge , however , has been to evaluate the formally exact expression for the reduced density matrix in parameter regions where analytical progress and perturbative simplifications are prohibitive . with the increasing complexity of designed and controllable quantum systems on the nanoscale ,
e.g. for quantum information processing or molecular electronics , the issue of efficient numerical procedures in the _ real - time domain _ becomes a very crucial one .
indeed , important achievements have been made in the last decade with the development of advanced path integral monte - carlo methods ( pimc ) @xcite , the quasi - adiabatic propagator scheme ( quapi ) @xcite , stochastic schrdinger equations @xcite , and basis set - methods @xcite . in this context
a certain class of systems has been extensively studied , namely , systems with discrete configuration space , also coined tight - binding systems ( tbss ) . for these systems
quantum diffusion takes place on a lattice , where the sites are coupled by tunneling amplitudes , an important case being the restriction to nearest neighbor coupling .
the simplest example is the well - known spin - boson model @xcite with applications from condensed matter physics to electron transfer reactions .
the fundamental role of tbss follows from the diversity of realizations in physics and chemistry @xcite .
transport properties in general multistable systems at sufficiently low temperatures can be described within tbs models leading to relations to the kondo problem and the luttinger liquid model .
further , tbss serve as archetypical models to study quantum phase transitions in correlated systems , as e.g. for hubbard type of models .
remarkably , even a large class of continuous systems can be mapped exactly onto tbss by means of duality transformations @xcite .
based on the path integral representation exact non - markovian master equations for tbss @xcite have been derived , which provide the starting point for perturbative treatments such as the non - interacting blip approximation ( niba ) and reduce to markovian ones even for moderate dissipation and low temperatures @xcite , i.e. far from the limiting ranges addressed above .
the pimc approach is particularly suited to treat the dissipative real - time evolution of tbss numerically exactly . while former applications were restricted mainly to two and three state models @xcite , recently , we substantially improved the approach to apply to more complex systems @xcite , such as single charge transfer across long one - dimensional molecular chains including impurities and external driving fields . moreover , the simulation time range could be extended to cover basically all relevant time scales of the dissipative dynamics .
we have been thus in a position to directly access the range of validity of markovian master equations for tbss , which in turn are extremely helpful to reveal the relevant physical processes behind the numerical data .
the goal of this paper is now twofold : on the one hand we push the pimc procedure even further and present first results for the dissipative dynamics of spatially continuous and of correlated many - body systems ; on the other hand , we give arguments to what extent markovian master equations can be used in these more involved situations .
we will see that this latter question will directly lead us to consider generalized tight - binding lattices with non - equidistant spacing between sites or in higher dimensions .
the article is organized as follows .
we start in sec .
[ sec1 ] with a brief summary of the path integral representation for open quantum systems and continue in sec .
[ sec2 ] to collect the main ideas of the pimc scheme .
[ sec3 ] deals with known results for non - markovian and markovian master equations for tbss , the applicability of which for one - dimensional chains is illustrated .
then , in sec
. [ sec4 ] spatially continuous systems are discussed , before we come to the correlated many - body dynamics in sec .
[ sec5 ] . at the end
some conclusions are given .
the standard approach @xcite for the inclusion of dissipation into a quantum mechanical formulation starts from a system+reservoir model @xmath0 with a system part @xmath1 , an environmental part @xmath2 , and a system - bath interaction @xmath3 .
the reservoir ( heat bath ) is mimicked by a quasi - continuum of harmonic oscillators bilinearly coupled to the system : @xmath4 \ ; , \ ] ] where @xmath5 denotes a system operator corresponding to a one - dimensional degree of freedom .
dissipation appears when one considers the reduced dynamics by properly eliminating the bath degrees of freedom , i.e. , @xmath6 with an initial density matrix @xmath7 of the total system .
it turns out that for the reduced dynamics the environmental parameters enter only via the spectral density @xmath8 which effectively becomes a continuous function of @xmath9 for a condensed - phase environment .
note that in the above definition of the spectral density we have included a factor @xmath10 with @xmath11 being a proper length scale which is convenient for the treatment of tbss .
the gaussian statistics of the isolated environment is determined by the complex - valued bath autocorrelation function which for real time @xmath12 reads @xmath13 \over \sinh(\hbar\beta\omega/2 ) } \;,\end{aligned}\ ] ] where @xmath14 . in the sequel we focus on systems evolving in a discretized configuration space with respect to the pointer variable @xmath15 and
thus consider the population @xmath16 of a `` lattice site '' @xmath17 determined by the diagonal part of the reduced density matrix , i.e. , @xmath18 which is normalized @xmath19 .
accordingly , the initial density matrix of the total compound @xmath7 in ( [ reduced ] ) is taken to be @xmath20 with the partition function of the isolated reservoir @xmath21 .
the bath is equilibrated according to a localized initial state of the system on a lattice site @xmath22 , where @xmath23 so that e.g. for electron transfer in a polar environment , @xmath24 is the electronic dipol moment and @xmath25 the collective dipol moment of the bath .
generalizations to delocalized initial states for the system are straightforward @xcite .
the path integral representation provides a formally exact expression for the reduced dynamics and is thus the starting point for a numerically exact monte - carlo ( mc ) scheme . along the lines sketched above ,
the bath degrees of freedom are eliminated exactly to arrive at the reduced dynamics . as shown in ref .
@xcite , one thus obtains for eq .
( [ populations ] ) @xmath26 - \phi[\tilde{s } ] \right\ } \;.\ ] ] here the path integration runs over closed paths @xmath27 connecting @xmath28 with @xmath29 along the real - time contour @xmath30 , which combines the forward and backward paths @xmath31 and @xmath32 , respectively .
furthermore , @xmath33 $ ] denotes the action of the free system .
the influence of the traced - out bath is completely encoded in the _ feynman - vernon influence functional _ @xmath34 $ ] @xcite @xmath35 & = & \int_0^{t}\!dt'\ !
\int_0^{t'}\!dt '' [ q(t ' ) - q'(t ' ) ] \ [ { l}(t'-t'')q(t '' ) \nonumber\\ & & \qquad { } - { l}^\ast(t'-t'')q'(t '' ) ] \nonumber\\ & & { } + i{{\hat{\mu } } \over 2}\int_0^{t}\!dt ' [ q^2(t ' ) - q'^2(t ' ) ] \;,\end{aligned}\ ] ] where @xmath36 the influence functional introduces long - ranged nonlocal interactions among the system paths so that in general an explicit evaluation of the remaining path integral in eq .
( [ populations path - integral ] ) is possible only numerically . in this situation
the pimc method has been proven as a very promising approach to obtain numerically exact results even in regions of parameter space where other approximate methods fail .
a prerequisite for an efficient numerical algorithm is an appropriate discretization of time and configuration space .
the latter one is intrinsically given for multistable systems in the tight binding limit , where the relevant states are strongly localized in position , only very weakly coupled by tunneling , and energetically well separated from the rest of the spectrum . for spatially continuous systems supporting delocalized states the situation is less obvious , but in case of a discrete energy spectrum a mapping onto a generalized tight - binding lattice applies as well for lower temperatures @xcite . in any case , the configuration space variable can then be written as @xmath37 with a typical length scale @xmath11 and a dimensionless variable @xmath38 with @xmath39 according to a @xmath40-level system ( @xmath40ls ) .
hence , the system hamiltonian reads @xmath41 where @xmath42 describes the energetic distribution of the sites according to @xmath43 and @xmath44 the couplings between them @xmath45 . in particular , in case of @xmath46 and nearest neighbor coupling
only one recovers a @xmath47-spin - boson model . for the discretization in time ,
the time axis is sliced via @xmath48 uniformly spaced points with discretization steps @xmath49 .
the path integral in eq .
( [ populations path - integral ] ) then becomes @xmath50\ ] ] with @xmath51 = \left[\prod_{k=1}^{2 r } k(s_k , s_{k+1})\right]e^{-\phi[\{s_j\ } ] } \;.\ ] ] the sum runs over all realizations of the discretized spin path @xmath52 , and @xmath53 denotes the coordinate representation of the free @xmath40ls propagation over the time interval @xmath54 , i.e. , @xmath55 this propagator of the @xmath40ls hamiltonian can be obtained from the eigenstates @xmath56 as @xmath57 which can be easily computed numerically once the @xmath40ls s parameters are specified . to arrive at a discretized form ( in time ) of the influence functional ( [ influence - exponent ] ) , the sum and difference coordinates @xmath58 are introduced , which read @xmath59 ( @xmath60 ) for @xmath61 $ ] in their discretized form .
the sum paths are also considered as `` quasi - classical '' , while the difference paths capture quantum fluctuations @xcite . equation ( [ influence - exponent ] ) finally can be written as @xmath62 & = & i\sum_{j=2}^r \xi_j \hat{x}^{(s_i)}_j\\ & & { } + \sum_{j \ge k=2}^r \xi_j ( i x_{j - k } \eta_k + \lambda_{j - k } \xi_k ) \nonumber\end{aligned}\ ] ] where the kernels @xmath63 , @xmath64 , and @xmath65 follow from discretizing the twice integrated bath autocorrelation function @xmath66 defined by @xmath67 , @xmath68 with @xmath69 . in the sequel we consider a spectral density of the form @xmath70 which is equivalent to ohmic damping with a cut - off frequency @xmath71 . in this case @xmath66 can be calculated analytically and one obtains @xmath72\ ] ] with @xmath73 and the gamma function @xmath74 .
equations ( [ pop - discretized ] ) , ( [ density ] ) and ( [ influence ] ) constitute a discretized form of the populations ( [ populations ] ) and thus provide a starting point for pimc simulations .
as it is well - known this method is handicapped by the dynamical sign problem @xcite .
it originates from quantum interferences between different system paths @xmath75 , causing a small signal - to - noise ratio of the stochastic averaging procedure .
one approach to deal with this problem is based on the observation that the quasi - classical paths @xmath76 in eq .
( [ influence ] ) can be integrated out as a series of @xmath77 matrix multiplications @xcite .
this reduces the degrees of freedom from the @xmath78 variables @xmath79 to the @xmath77 quantum variables @xmath80 and therefore significantly improves the numerical stability of the corresponding mc simulations .
while this technique , which in fact is yet another example of a blocking approach @xcite , works greatly for dissipative two- and three - state systems @xcite , however , the increasing size of the corresponding matrices with the number of electronic sites turns these multiplications into a quite time consuming task .
since they have to be performed for every single update of the mc trajectory , the investigation of larger systems again requires exceptionally long cpu times .
nevertheless , this severe computational bottleneck can be profoundly alleviated on physical grounds @xcite . upon closer inspection
one finds that the possibility of rewriting the integration over the quasi - classical coordinates in terms of simple matrix multiplications is due to the fact that the real - valued part of the bath autocorrelation function ( [ lz ] ) governs only the _ quantum coordinates _ but not the quasi - classical ones .
this real - valued part , which eventually leads to a damping out of quantum coherences , introduces a non - local self - coupling among the quantum coordinates and is thus directly related to retardation effects , a main complication for treating dissipative quantum systems .
accordingly , the retardation effects influence the evolution of the quasi - classical coordinates contained in the imaginary part of the influence functional to a significantly weaker extend than that of the quantum coordinates . neglecting them
while generating the mc trajectories leads to an only minor impairment of the sampling statistics , but causes an almost complete decoupling between quantum and quasi - classical coordinates .
this in turn allows for an enormous speed - up of the matrix multiplications .
accordingly , the sampling process could so far be accelerated by a factor of approximately 100 with respect to the original approach @xcite ( for further details , we refer to refs .
@xcite ) , thus opening the door to treat the reduced dynamics of much larger and even many - body systems over sufficiently long times .
from the exact expression ( [ populations path - integral ] ) for the reduced density simplifications can be derived in certain limits in terms of time - local master equations . in the context discussed here , these are of crucial importance , mainly since ( i ) master equations allow to study ranges in parameter space where the monte carlo sampling is rather expensive , e.g. for very weak or very strong friction , and ( ii ) they provide a basis to access the physical processes underlying the numerical data also by means of analytical techniques . in the weak friction limit one imposes that the level broadening due to friction is much smaller than @xmath81 _ and _ the typical level separation .
it is thus natural to work in the basis spanned by the energy eigenstates of @xmath1 and to treat the coupling @xmath3 perturbatively .
with increasing coupling , however , the environment drives the system to its pointer basis in which the system - bath coupling operator @xmath15 is diagonal . as we have seen above in sec .
[ sec1 ] , in this basis the path integral formulation allows for a non - perturbative elimination of the reservoir degrees of freedom . in case that friction is very strong and thus level broadening much larger than level spacing and temperature , a master equation complementary to the weak friction range , the quantum smoluchowski equation , is again available @xcite .
what s about the intermediate regime ?
for tbss discussed here , substantial findings have been gained in the past decade @xcite .
namely , it was shown that an _
exact _ retarded master equation exists , i.e. , @xmath82 where the kernels obey @xmath83 .
basically they represent a power series in the couplings @xmath84 with corresponding spin - path integrals .
to make the latter tractable , one applies the niba @xcite or its generalization , the non - interacting cluster approximation ( nica ) @xcite , which has been shown to be accurate in a variety of parameter ranges .
for instance , in case of an ohmic spectral density with a high frequency cut - off , it captures quantum coherence as well as pure relaxation dynamics .
the corresponding kernels @xmath85 read in lowest order in the intersite couplings @xmath86\nonumber\\ & & \cos\left[(\epsilon_\mu-\epsilon_\nu ) t+(q_\mu - q_\nu)^2\ , q''(t)\right]\,\end{aligned}\ ] ] with @xmath87 and @xmath88 .
the practical use of ( [ exactme ] ) is limited though due to the retardation .
now , for sufficiently strong dissipation and fast enough bath modes the kernel falls off on a time scale much shorter than the time scale on which the relevant reduced dynamics occurs so that we may set in ( [ exactme ] ) @xmath89 and @xmath90 accordingly , ( [ exactme ] ) reduces to a simple markovian rate equation @xmath91 with a rate matrix @xmath92 consisting of the individual rates @xmath93 . for nearest neighbor coupling ,
the golden rule rates @xmath94 describe a sequential hopping process , while all higher order contributions to @xmath93 capture long - range hopping termed superexchange .
note that in contrast to weak coupling master equations as e.g.the redfield equations , ( [ exactme ] ) and ( [ differential rate equation ] ) together with the corresponding transition rates apply also to the range of moderate to strong bath coupling and very low temperatures . as an explicit example of ( [ differential rate equation ] ) we consider a one - dimensional tight - binding lattice with @xmath95 sites , spacing 1 , and constant nearest neighbor coupling corresponding to a spin-3-boson system ( for details see @xcite ) .
this model can be seen as a simple realization of electron transfer along a molecular chain subject to a dissipative environment @xcite .
initially , the charge is localized at the donor site @xmath96 and @xmath97 monitors the dynamics towards the acceptor at @xmath98 .
specifically , donor and acceptor are linked by a bridge with an impurity at its center .
donor / acceptor have vanishing onsite energy , the bridge sites are elevated by @xmath99 , and the impurity site has @xmath100 . as seen in fig .
[ fig1 ] the exact pimc data are very accurately described by the time - local master equation ( [ differential rate equation ] ) despite the fact that all parameters are chosen such that one is close to a coherent / incoherent transition ( for the given values of @xmath71 and @xmath101 coherences appear at inverse temperatures @xmath102 and larger ) and @xmath103 is far from the scaling limit .
in order to go beyond the spin - boson type of models ( one - dimensional tight binding lattice , equidistant spacing , constant nearest coupling ) , we start by presenting an approach to capture the dynamics of spatially continuous systems within the pimc procedure outlined above .
it turns out that this method applies to systems with a discrete energy spectrum .
the basic idea is simple : for sufficiently low temperatures only the lowest lying states of the isolated system can be assumed to take part in the dynamics ; thus , the full hilbert space @xmath1 can effectively be truncated to a subspace @xmath104 spanned by the @xmath105 lowest lying eigenstates .
one then has to find a proper basis in this subspace , obtained by a unitary transformation from the eigenstate basis , in which a stochastic sampling of the path integrals ( [ populations path - integral ] ) can be performed .
this sort of reduction is not new .
in fact , the spin - boson model can be seen as originating from a double well potential , where only the two lowest lying states are taken into account .
of course , the goal here is to go beyond : we want to capture the dissipative dynamics from very low up to moderate temperatures .
for very high temperatures semi - classical or classical methods apply anyway .
further , we are interested in the regime of moderate friction , where neither of the known approximate formulations work so that a numerical treatment has to start from the exact reduced dynamics ( [ populations path - integral ] ) .
the proper basis for the sampling in the restricted hilbert space is then the basis that diagonalizes the operator @xmath5 in @xmath104 .
this idea to treat spatially continuous systems has been first applied in the numerical quapi approach @xcite and has been recently analyzed in @xcite to derive a generalized master equation of the type given above ( [ exactme ] ) .
hence , we sketch here only the central results briefly and refer to these previous works for further details . in the @xmath105-dimensional subspace @xmath104 the pointer basis is defined as @xmath106 and obtained from the eigenstate basis @xmath107 by diagonalizing a matrix with entries @xmath108 , @xmath109 .
its eigenvalues provide the dimensionless @xmath110 , while the eigenvectors are given as @xmath111 .
the basis set @xmath112 is also known as the dvr - basis ( discrete value representation ) . by representing the hamiltonian in this dvr - basis
, one obtains `` onsite energies '' @xmath113 and `` intersite - couplings '' @xmath114 .
the original system is thus mapped onto a generalized @xmath105-dimensional tight - binding lattice with _ non - equidistant _ sites at @xmath115 and _ non - nearest neighbor _
couplings @xmath116 .
now , to implement this tbs into the pimc algorithm , one has to take into account the non - equidistant lattice by re - defining the @xmath117 and @xmath118 variables properly .
a rough estimate for the validity of the truncation procedure to the lowest @xmath105 eigenstates of the isolated system is given by the conditions that ( i ) the level broadening due to friction is of the same order as the level spacing or smaller and that ( ii ) the temperature is sufficiently low @xmath119 . while the second condition is obvious , the first one originates from the fact that for very strong friction the system tends to the classical limit again , see @xcite . by way of example
, we consider in the sequel a harmonic oscillator with mass @xmath120 and frequency @xmath121 , so that the typical length scale is @xmath122 .
while a detailed account of the implementation into the pimc algorithm and results of simulations for specific observables , particularly in comparison with analytical results , will be given elsewhere , here , our main interest is to explicitly show that apart from the conditions for the applicability of time local master equations known for spin - boson models , the treatment of continuous systems via the described mapping comes with an additional complication .
formally , this issue has also been addressed recently in @xcite . here
, we directly compare numerically exact pimc data with results from the simplified master equation . for this purpose the initial state is chosen as one of the dvr - states which suffices to study the time evolution of the populations on the sites @xmath110 .
results for @xmath123 and @xmath124 are depicted in figs .
[ fig2 ] , [ fig3 ] for a bath with @xmath125 , @xmath126 , and @xmath127 . for this parameter set
we have shown recently @xcite that time local master equations capture the exact dynamics of linear chains rather accurate even though one is close to an incoherent / coherent transition for the reduced dynamics .
the same is true here for @xmath123 , where we used ( [ differential rate equation ] ) together with the hopping rates ( [ grfull ] ) .
in contrast , for @xmath124 deviations appear in the short to intermediate time regime . for times of the order of @xmath128 and shorter ( @xmath129 ) ,
these are caused by adiabatic effects in the bath well - known from the dynamics in ordinary tight binding lattices . in the intermediate time range , however , where for certain populations pronounced maxima occur , deviations must be ascribed a specific property related to the mapping onto a tight binding lattice with non - equidistant spacing . as mentioned above
, this mapping requires a re - definition of the @xmath130 and @xmath131 variables , originally defined for a tight binding lattice with equidistant spacing of length @xmath11 .
qualitatively , the situation is the following : since the variance of the @xmath105th eigenstate is roughly @xmath132 , the box in position space covered by @xmath105 states has a typical width @xmath133 .
the mean spacing between adjacent lattice sites is thus @xmath134 .
hence , compared to a lattice with equidistant spacing of length @xmath135 , the re - definition of the @xmath130 and @xmath131 variables in the influence functional ( [ influence ] ) comes with an additional factor of the order @xmath136 , which basically renormalizes the damping kernel .
consequently , for fixed bath parameters and increasing @xmath105 , the effective coupling constant between bath and discrete system is not constant , but decreases as @xmath137 .
the system is thus driven into the range of weak coupling to the bath @xcite , where retardation effects due to quantum coherences become more substantial .
hence , the limit to a spatially continuous system @xmath138 is non - trivial and must be performed by keeping @xmath139 constant . specifically , by comparing the spectral density ( [ spectral ] ) with the one usually introduced for spatially continuous systems @xcite , one finds @xmath140 where @xmath141 denotes the macroscopic damping constant appearing in the classical langevin equation .
note that the above relation can not be seen as a strict equality since the spacing between adjacent dvr - sites varies slightly and is typically smaller deep inside the potential well .
the value @xmath126 chosen above corresponds for @xmath123 to @xmath142 and for @xmath124 to @xmath143 .
now , a rough estimate for the validity of time local master equations can be gained by assuming that the exponential in ( [ grrate ] ) must fall off on a time scale sufficiently shorter than that of the dynamics of @xmath144 . for moderate and low temperatures ,
this leads to @xmath145 and thus @xmath146 . in accordance with this relation , the pimc data presented in fig .
[ fig2 ] , [ fig3 ] for @xmath123 can still be captured quantitatively by ( [ differential rate equation ] ) , while for @xmath124 only a qualitative agreement can be seen . to perform
the continuum limit is thus not an easy task and certainly deserves further research .
the dissipative real - time evolution of interacting many body systems has been left basically untouched so far . in certain limits ,
e.g. for very strong repulsive interactions , results could be derived from a simple hopping model @xcite , but the intimate interplay between direct interaction , interaction mediated by the environment and dissipation has not been accessible . here , for the first time we present pimc results for the dynamics of two interacting particles along the lines described above . as in the previous section , in the sequel
the formulation is outlined only briefly and we focus on conceptual properties related to time - local master equations .
further , we consider the case of indistinguishable particles , called charges henceforth , so that the quantum nature of the particle statistics matters .
physically , the situation refers to interacting spinless fermions or interacting bosons .
the free system is taken as a tight - binding lattice with spacing 1 and nearest neighbor coupling , where the two charges can be placed on @xmath147 sites .
accordingly , the two - charge hamilton operator reads @xmath148 + { u}\right\ } \label{2hamilton}\ ] ] where @xmath149 and @xmath150 are straightforward generalizations of the corresponding operators introduced above for the single particle case , while @xmath151 describes a site dependent coulomb interaction specified below .
the coupling to the bath is determined by the total dipol - moment of the system and thus follows directly from ( [ hamilton ] ) : @xmath152 ^ 2\right\ } \label{total2 } \end{aligned}\ ] ] with @xmath153 . since the two charges are indistinguishable it is convenient for the pimc simulation to work not in the single particle product basis , but rather in the basis of the many body states .
for this purpose we introduce @xmath154 with @xmath155 .
the advantage of this representation is that the @xmath156 states @xmath157 are orthogonal in contrast to the original ones .
accordingly , @xmath158 takes the form @xmath159\ ] ] where now @xmath160 for an onsite coulomb energy this model is identical to a dissipative hubbard model with two charges .
note , however , that for the pimc simulations a non - local interaction can easily be taken into account . now , as seen above , the free system dynamics enters in the discretized path integral formulation only through its short time propagator which is conveniently represented in the energy eigenbasis of @xmath158 in ( [ dhamilton ] ) , see ( [ sysprop2 ] ) .
further , since in ( [ total2 ] ) the two charges interact with the bath only via the sum @xmath161 the corresponding change to the many body basis in the influence functional is straightforward .
it amounts to introduce discretized sum and difference paths according to @xmath162 such that the influence functional depends only on @xmath163 and @xmath164 . in principle
, we could now start to implement the above representation into a pimc algorithm .
however , before doing so we go one step further and exploit the following crucial property : _ the dissipative dynamics of two indistinguishable charges on a one - dimensional lattice with @xmath40 sites is equivalent to the dissipative dynamics of a single particle on a two - dimensional triangular lattice with @xmath156 sites_. to see this , one realizes that the coupling to the bath @xmath165 with @xmath166 as in ( [ initial_perp_p ] ) can be written as @xmath167 with the vector operators @xmath168 and @xmath169 , thus being identical to the system - bath coupling of a single particle on a surface .
eventually , by a proper re - labeling of the populations @xmath170 one formally obtains the desired mapping .
this is illustrated in fig .
[ fig4 ] where the actual coupling with the bath at a certain site is given by the projection of the corresponding `` spin - vector '' @xmath171 onto @xmath172 as just described and each site carries an onsite energy depending on @xmath173 . due to the nearest neighbor - coupling
, transitions can only occur in the vertical and the horizontal direction , respectively .
note that this mapping can be generalized to fermionic systems as well .
to summarize , the advantages of the many body basis and the subsequent mapping onto a single particle 2@xmath40-lattice are : ( i ) the pimc algorithm developed for single particles can be simply adapted to the case of two particles , ( ii ) the configuration space to be sampled is reduced from @xmath174 to @xmath156 , and ( iii ) the master equations introduced in the previous section can be directly applied .
let us now analyze ( iii ) in more detail based on ( [ differential rate equation ] ) with the corresponding transition rates ( [ grfull ] ) .
the simplest case is @xmath175 with only three available sites @xmath176 .
obviously , for this case the triangular lattice coincides with a linear chain with three sites so that if a time local master equation applies for this latter case , it also applies for the two - charge case . for larger @xmath40 , however , the situation changes fundamentally due to a different topology of the two - dimensional lattices which then contain bulk - sites with four adjacent sites ( apart from @xmath177 where there is no bulk site , but two edge - sites with three connections ) .
accordingly , the typical dwell time on a bulk - site is considerably shorter than on a site with two connections , roughly , by a factor of 2 .
an environment which is able to destroy the phase coherence of the wave function on each site of a linear chain before a jump to an adjacent site takes place , may be too slow to achieve the same on bulk - sites in a 2d - lattice .
hence , entangled many body states can survive on such a long time scale that a description based on a master equation local in time fails .
in fact , this can be seen in fig .
[ fig5 ] , where pimc data for @xmath178 are depicted together with the corresponding dynamics of the master equation ( [ differential rate equation ] ) . by successively removing more and more bulk sites from the lattice , the agreement of the pimc data with the prediction of the master equation , which is rather poor in case of the full lattice , increases , until upon creating a linear chain by removing all bulk states an almost perfect match is regained .
the situation becomes even worse for a larger number of charges @xmath179 since then the bulk sites of the @xmath179-dimensional cube attain @xmath180 decay channels to adjacent sites , thus reducing the average dwell time by a factor of about @xmath180 compared to the one - dimensional case @xmath181 .
the conclusion is the following : if for a bath characterized by @xmath71 and @xmath182 a markovian approximation applies for the dynamics on a 1d - lattice with nearest neighbor coupling @xmath183 , i.e. @xmath184 and @xmath185 , this approximation fails for the multi - charge dynamics , unless @xmath71 is taken to be very large and temperatures are sufficiently high , roughly , @xmath186 and @xmath187 .
hence , for most cases only numerical approaches like the pimc procedure presented here , seem to be applicable .
the numerically exact pimc approach has been pushed further to deal also with spatially continuous systems and correlated many - body dynamics . to better understand the numerical data in certain parameter ranges , to estimate the dissipative quantum dynamics before starting an involved pimc calculation , and even to develop simplified models capturing the relevant processes , markovian master equations are of great practical use . to explore their applicability in the above situations
, we had to consider generalized tight - binding lattices with either non - equidistant spacing or in higher dimensions . in both cases ,
additional restrictions must be imposed beyond the constraints known from equidistant one - dimensional tbss so that for broader ranges of bath parameters no simple description seems to be possible .
however , it is surprising that in other domains comprising stronger dissipation and lower temperatures , markovian master equations can still be found to work quite well , at least qualitatively .
thus , markovian master equations together with numerically exact pimc simulations , which often provide the sole mean to check for their validity in a certain scenario , provide a powerful means to study dynamical properties over the full time range . in particular , for the correlated many - body dynamics in presence of dissipation this may open the door to examine the intimate interplay between coulomb interaction , particle statistics , and phonon baths , e.g. for the charge transport in molecular and mesoscopic structures .
this work benefited from fruitful discussions with a. komnik .
we acknowledge financial support from the dfg through grant no .
an336 - 1 and the landesstiftung baden - wrttemberg ggmbh .
j.a . is a heisenberg fellow of the dfg .
99 u. weiss , _ quantum dissipative systems _ , series in modern condensed matter physics , vol .
2 ( world scientific , singapore , 1998 ) . c.w .
gardiner , _ quantum noise _
( springer , berlin , 1991 ) .
g. lindblad , commun .
phys . * 48 * , 119 ( 1976 ) .
a.g . redfield , adv .
. reson . * 1 * , 1 ( 1965 ) .
j. ankerhold , p. pechukas , and h. grabert , phys .
. lett . * 87 * , 086802 ( 2001 ) .
b. vacchini , phys .
e * 66 * , 027107 ( 2002 ) ; d. banerjee , b.c .
bag , s.k .
banik , d.s . ray , physica a * 318 * , 6 ( 2003 ) .
j. ankerhold , europhys . lett . * 61 * , 301 ( 2003 ) ; j. ucka , r. rudnicki , and p. hnggi , physica a * 351 * , 60 ( 2005 )
. j. ankerhold , europhys . lett . * 67 * , 280 ( 2004 ) ; l. machura , m. kostur , p. hnggi , p. talkner , and j. ucka , phys . rev .
e * 70 * , 031107 ( 2004 ) ; j. ankerhold , h. grabert , and p. pechukas , in : _ 100 years of brownian motion _ , chaos * 15 * , 026106 ( 2005 ) .
r.p . feynman and f.l .
vernon , ann .
( n.y . ) * 24 * , 118 ( 1963 ) .
a.o . caldeira and a.j .
leggett , physica * 121a * , 587 ( 1983 ) .
h. grabert , p. schramm , and g .- l .
ingold , phys . rep . * 168 * , 115 ( 1988 ) .
r. karrlein and h. grabert , phys .
e * 55 * , 153 ( 1997 ) .
mak and r. egger , adv .
phys . * 93*,39 ( 1996 ) .
d. e. makarov and n. makri , chem .
* 221 * , 482 ( 1994 ) .
strunz , l. disi , and n. gisin , phys .
lett . * 82 * , 1801 ( 1999 ) ; j.t .
stockburger and h. grabert , phys .
* 88 * , 170407 ( 2002 ) .
beck , a. jckle , g.a .
worth , and h .- d .
meyer , phys . rep . * 324 * , 1 ( 2000 ) .
leggett , s. chakravarty , a.t .
dorsey , m.p.a .
fisher , a. garg , and w. zwerger , rev .
phys . * 57 * , 1 ( 1987 ) .
a. schmid , phys .
lett . * 51 * , 1506 ( 1983 ) ; m.p.a .
fisher and w. zwerger , phys .
b * 32 * , 6190 ( 1985 ) .
r. egger , c.h .
mak , and u. weiss , phys .
e * 50 * , r655 ( 1994 ) .
r. egger and c. h. mak , j. phys .
chem . * 98 * , 9903 ( 1994 ) ; l. mhlbacher and r. egger , j. chem .
phys . * 118 * , 179 ( 2003 ) .
l. mhlbacher , j. ankerhold , and c. escher , j. chem . phys . * 121 * , 12696 ( 2004 ) .
l. mhlbacher , j. ankerhold , j. chem . phys . * 122 * , 184715 ( 2005 ) . m. thorwart , m. grifoni , and p. hnggi , phys . rev . lett .
* 85 * , 860 ( 2000 ) ; ann .
( new york ) * 293 * , 15 ( 2001 ) . , m. suzuki ( ed . ) ( world scientific , singapore , 1993 ) , and references therein .
r. egger and c. h. mak , phys .
b * 50 * , 15210 ( 1994 ) . c. h. mak , r. egger , and h. weber - gottschick , phys . rev .
lett . * 81 * , 4533 ( 1998 ) .
j. ankerhold and h. lehle , j. chem . phys . *
120 * , 1436 ( 2004 ) . j. jortner and m. bixon , eds . ,
phys . * 106 * , 107 ( 1999 ) ; a. nitzan , ann .
phys . chem . *
52 * , 681 ( 2001 ) .
e.g. petrov and p. hnggi , phys .
* 86 * , 2862 ( 2001 ) ; j. lehmann , g .-
ingold , and p. hnggi , chem .
phys . * 281 * , 199 ( 2002 ) . | the numerically exact path integral monte carlo approach for the real - time evolution of dissipative quantum systems ( pimc ) , particularly suited for systems with discrete configuration space ( tight - binding systems ) , is extended to treat spatially continuous and correlated many - body systems .
this way , one has to consider generalized tight - binding lattices with either non - equidistant spacing or in higher dimensions , which in turn allows to analyze to what extent markovian master equations can be applied beyond the usually studied spin - boson type of models . |
it is well known that the potential energy curves of diatomic molecules do not cross ( in fact , they even avoid each other ) when the states have the same symmetry .
this property of the electronic energies , commonly known as the non - crossing rule , has proved useful for the interpretation of many experiments in molecular spectroscopy and photochemistry@xcite . the theoretical explanation outlined by teller@xcite , and typically reproduced in most textbooks on quantum chemistry@xcite , was criticized by razi naqvi and byers brown@xcite . after arguing that such a proof is based on a _ non sequitur _ the authors proposed an alternative justification of the non - crossing rule .
their argument is closely related to the hellmann - feynman theorem@xcite in its more general off - diagonal form@xcite . as a result of a recent investigation on non - hermitian hamiltonians
we were led to revise the proofs on the non - crossing rule and the purpose of this paper is to put forward our analysis and discussion of the arguments given by razi naqvi and byers brown@xcite . in section
[ sec : h - f ] we derive similar equations by means of the off - diagonal hellmann - feynman theorem@xcite ( and references therein ) . in section [ sec : examples ] we illustrate the main theoretical conclusions by means of two simple examples .
finally , in section [ sec : conclusions ] we summarize the main results and draw conclusions .
the starting point is the time - independent schrdinger equation @xmath0 it follows from @xmath1 and @xmath2 , where * stands for complex conjugation , that @xmath3 from this expression we conclude that @xmath4 when @xmath5 .
this textbook result is well known but we write it here because it will be useful later on .
if @xmath6 depends on a parameter @xmath7 , then the eigenfunctions and eigenvalues will depend on this parameter too .
if we differentiate equation ( [ eq : schro ] ) with respect to @xmath7 and then apply the bra @xmath8 from the left we obtain the well known off - diagonal hellmann - feynman relation@xcite @xmath9 where the prime denotes differentiation with respect to @xmath7 .
suppose that @xmath10 and @xmath11 approach each other and cross at @xmath12 : @xmath13 . when @xmath14 @xmath15 by virtue of equation ( [ eq : pre_orthogonality ] ) and because of continuity we should also have @xmath16 it follows from this equation and ( [ eq : h - f ] ) that @xmath17 without this condition the approaching energy levels will not cross giving rise to an avoided crossing that looks like an energy - level repulsion .
since the two levels approach each other and then move apart the quantity @xmath18 should exhibit a minimum at some @xmath19 .
this particular value of the parameter is determined by the condition @xmath20 if the symmetries of @xmath21 and @xmath22 are different , then equation ( [ eq : hmn=0 ] ) holds for all @xmath7 and nothing prevents the approaching energy levels from crossing . throughout the discussion above we have tacitly assumed that the symmetry of @xmath6 is the same for all @xmath7 ( at least in the neighbourhood of @xmath23 under analysis ) . in other words , we have assumed that both @xmath6 and @xmath24 have the same symmetry .
suppose that the point group@xcite that describes the symmetry of @xmath6 is @xmath25 when @xmath26 and @xmath27 when @xmath28 and that the order @xmath29 of @xmath25 is smaller than the order @xmath30 of @xmath27 .
under such conditions the dimension of the subspaces of @xmath31 may be greater than those for @xmath32 @xmath33 and we therefore expect some level crossings at @xmath28 .
obviously , equation ( [ eq : hmn=0 ] ) applies to those states that become degenerate at this point .
razi naqvi@xcite took into account such symmetry changes in a discussion of the crossing of potential - energy surfaces of polyatomic molecules .
it is clear that no further discussion is necessary for proving equation ( [ eq : orthogonality ] ) that was required for deriving equation ( [ eq : hmn=0 ] ) from ( [ eq : h - f ] ) .
however , razi naqvi and byers brown@xcite criticized the continuity argument implied by equation ( [ eq : orthogonality ] ) . in order to discuss the additional steps in their proof
we first derive another equation . if we differentiate the eigenvalue equation for @xmath21 with respect to @xmath7 and apply @xmath34 from the left we arrive at an equation similar to ( [ eq : h - f ] ) : @xmath35 subtracting the complex conjugate of equation ( [ eq : h - f_2 ] ) from equation ( [ eq : h - f ] ) we obtain @xmath36 which is obviously the derivative of equation ( [ eq : pre_orthogonality ] ) with respect to @xmath7 . when @xmath37 equation ( [ eq : deriv_orthog ] ) reduces to @xmath38 \left\langle \psi _ { m}\right .
\left| \psi _ { n}\right\rangle ( \lambda _ { 0})=0 . \label{eq : rnbb1}\ ] ] razi naqvi and byers brown@xcite considered two electronic states @xmath39 and @xmath40 of a diatomic molecule such that the corresponding electronic energy levels @xmath41 and @xmath42 , where @xmath43 is the internuclear distance , cross at @xmath44 .
they derived an equation similar to ( [ eq : rnbb1 ] ) that reads : @xmath45 \left\langle \psi _ { 1}^{0}\right .
\left| \psi _ { 2}^{0}\right\rangle = 0 , \label{eq : rnbb2}\ ] ] where @xmath46 and @xmath47 are the electronic states at @xmath44 .
they invoked this equation to prove that @xmath48 if @xmath49 .
the reason of this _ circumlocution _ was their concern about the continuity argument expressed in the statement : `` it will be well to pause here momentarily and discuss the implications of equation ( [ eq : ortho_rnbb ] ) .
our demand that the two potential curves intersect at @xmath50 , forces us , to conclude that the overlap integral must vanish even when @xmath51 .
it is tempting to argue that , since @xmath52 for all @xmath43 in the vicinity of @xmath44 , it seems likely , on account of continuity , that it would also be true at @xmath50 .
however , this argument is not only unnecessary but misleading , for we know that degenerate eigenfunctions need not be orthogonal ; indeed we can choose them at will and make them to be non - orthogonal , if we so desire . ''
the reader may convince himself that the argument leading to equation ( [ eq : orthogonality ] ) clearly implies that we do not choose those functions `` at will '' because the states at @xmath44 are just the ones that result from the limit @xmath53 and , therefore , should remain orthogonal . in order to prove that equation ( [ eq : ortho_rnbb ] )
holds even when @xmath54 , @xmath55 , provided that @xmath56 , the authors differentiate equation ( [ eq : rnbb2 ] ) with respect to @xmath43 as many times as necessary@xcite .
however , it is obvious that equation ( [ eq : rnbb2 ] ) is valid only for @xmath44 because we have discarded a term from the general equation valid for all @xmath43 ( see ( [ eq : deriv_orthog ] ) ) . in order to carry out this proof
correctly we should differentiate an equation like ( [ eq : deriv_orthog ] ) as many times as necessary which is equivalent to differentiating an equation similar to ( [ eq : pre_orthogonality ] ) with respect to @xmath43 just one more time .
more precisely , if we define @xmath57 and @xmath58 then equation ( [ eq : pre_orthogonality ] ) becomes @xmath59 .
differentiating it @xmath60 times with respect to @xmath43 and substituting @xmath50 for @xmath43 we obtain @xmath61 from which it follows that @xmath62 when @xmath63 .
in addition to being simpler and clearer , this argument is free from the flaw in the additional steps of the proof attempted by razi naqvi and byers brown@xcite .
however , in our opinion this discussion is unnecessary because , as argued above , @xmath64 always applies when @xmath65 .
in section [ sec : h - f ] we mentioned the possibility that the symmetry of the system may change at @xmath28 . in order to illustrate this point here we choose an extremely simple model , the quantum - mechanical harmonic oscillator @xmath66 the eigenvalues and eigenfunctions of this dimensionless hamiltonian operator are given by @xmath67 respectively , where @xmath68 is an eigenfunction of the one - dimensional harmonic oscillator @xmath69 @xmath70 . when @xmath71 the symmetry of the system is described by the abelian point group @xmath72 that exhibits only one - dimensional irreducible representations@xcite .
therefore , its states are expected to be nondegenerate , except for accidental degeneracies that may occur when @xmath73 is rational . on the other hand , when @xmath74 we have an isotropic two - dimensional oscillator so that its symmetry is described by the full two - dimensional rotation group . since all its eigenstates @xmath75 , @xmath76 are degenerate we expect and infinite number of crossings at @xmath28 .
obviously , the off - diagonal matrix elements @xmath77 vanish when @xmath28 in agreement with the argument given is section [ sec : h - f ] .
the second example is even simpler but most interesting in some respects . in this case
we choose a two - level system given by the matrix representation @xmath78 the diagonal elements intersect at @xmath79 but the eigenvalues @xmath80 exhibit an avoided crossing as shown in figure [ fig : eh ] .
figure [ fig : h12 ] shows that the off - diagonal matrix element @xmath81 , where @xmath82 and @xmath83 are the two column eigenvectors of @xmath84 , non only does not vanish but even exhibits a maximum precisely at @xmath79 . it is well known that avoided crossings are associated to exceptional points in the complex plane@xcite .
present case is not an exception as the eigenvalues ( [ eq : e1,e2_hermit ] ) obviously cross in the complex @xmath85-plane at @xmath86 . by means of the change of variables
@xmath87 we obtain a parity - time - symmetric non - hermitian hamiltonian@xcite : @xmath88 with eigenvalues @xmath89 in this case the eigenvalues are real for all @xmath90 ( unbroken parity - time ( pt ) symmetry@xcite ) approach each other as @xmath91 , coalesce at the exceptional points @xmath92 and become a pair of complex conjugate numbers for @xmath93 ( broken pt symmetry ) .
this behaviour is shown in figure [ fig : enh ] . at the exceptional points the two eigenvectors
are linearly dependent@xcite .
it is most interesting to consider the more general case in which @xmath94 that leads to an hermitian hamiltonian when @xmath95 and a pt - symmetric one when @xmath96 .
figure [ fig:3d ] shows that @xmath97 is given by two intersecting surfaces that leave a hole where they do not touch .
the intersection of the whole composite surface with the plane @xmath98 yields the curves shown in figure [ fig : eh ] for the hermitian hamiltonian ( [ eq : h_hermit ] ) .
on the other hand , the intersection with the plane @xmath99 yields the curve in figure [ fig : enh ] for the eigenvalues of the pt - symmetric hamiltonian ( [ eq : k_non - hermit ] ) , where @xmath100 .
the arguments put forward by razi naqvi and byers brown@xcite are basically correct , except for the discussion of the orthogonality of the states at the crossing point @xmath44 . in the first place
, there is no problem with the orthogonality of the states at this point if one chooses them to be the result of the limit @xmath53 .
such states are not at all arbitrary and conserve their orthogonality even at the point of degeneracy . if one had any doubt about the orthogonality of the states at the crossing point one could in fact prove it as shown in equation ( [ eq : add_proof_ort ] ) that is an improvement on the argument given by those authors that leads to the correct answer but is based on an inadequate equation .
this report has been financially supported by pip no .
11420110100062 ( consejo
nacional de investigaciones cientificas y tecnicas , repblica argentina ) | the non - crossing rule for the energy levels of a parameter - dependent hamiltonian is revisited and a flaw in a commonly accepted proof is revealed .
some aspects of avoided crossings are illustrated by means of simple models .
one of them shows the close relationship between avoided crossings and exceptional points . |
the interaction of ultraintense laser pulses with solid targets leads to the formation of plasmas with unusual properties in terms of particle energy distributions and of spatial mass and charge density distributions .
such properties can be exploited in order to obtain sources of high energy electromagnetic radiation and charged particle beams with unprecedented intensities and time and space resolutions . for the intensities of present day ultrashort , superintense laser pulses ,
the energy that the ions in a target acquire due to direct interaction with the electromagnetic fields of the laser pulse is usually small , while the energy of the plasma electrons can be of the order of tens of @xmath0 .
these `` hot '' electrons expand until their `` pressure '' is balanced by the electrostatic field that sets up due to spatial charge separation @xcite .
this process leads to a steady state configuration ( ssc ) which is reached in a time of the order of some electron plasma periods .
thus , since the ion response time is much longer than that of the electrons , ssc can be achieved before the ions can depart significantly from their initial configuration .
afterwards ion acceleration takes place , as predicted theoretically @xcite , and confirmed experimentally @xcite .
clearly , this description does not apply to highly relativistic regimes like those described in @xcite .
+ the aim of this paper is to present a combined analytical and numerical investigation of three dimensional ssc characterized by a hot electron plasma and a cold ( immobile ) ion core .
such configurations are especially appealing because , contrary to one dimensional geometries , in three dimensional cases charged ssc are expected to set up .
in fact , while in the former case an infinite energy is required in order to bring a charge to infinity , in the latter the energy necessary for electron evaporation is finite . in particular , we show that the ssc charging up and the energy distribution of the electrons depend on the ratio between the radius of the ion core and the electron debye length and on the history of the electron expansion .
the understanding of the ssc charging up with immobile ions and of the electron energy distribution is relevant to many experimental conditions as these processes affect the way in which ions are accelerated on longer time scales when the constraint of a fixed ion core is removed . in the case of the coulomb explosion of a small cluster @xcite these processes affect the value of the maximum energy that the ions can acquire in the acceleration process .
furthermore , in applications related to proton imaging @xcite and to the propagation of ion beams in solid targets @xcite , the coulomb repulsion and the screening effect of the electrons can strongly affect the proton trajectories .
this is also the case for applications of proton laser acceleration to hadron therapy @xcite , where a very precise collimation of the proton beam and a high energy resolution are essential .
moreover , as discussed in @xcite , the topic of the charging of a spherical plasma in less extreme conditions can play a key role in the study of dusty plasmas @xcite , and aerosols @xcite . + the paper is organized as follows . in sec.[sl ] simple relationships are derived analytically on the basis of two schematic models that are introduced in order to grasp the main features of the plasma charging process in different collisionality regimes . in sec.[nr ] we present the results of a series of numerical simulations obtained with a one - dimensional particle in cell code ( pic ) in a spherical geometry and with fixed ions .
we then compare the numerically obtained charge values and electron energy distributions with those obtained from the analytical models .
finally , conclusions are presented in sec.[conclusions ] .
in this section we discuss two simplified models that allow us to identify the main physical features of the plasma charging process .
these models rely on assigning a simplified condition for the electrons to leave the ion core and on two different rules for the energy redistribution of the remaining electrons . + as a starting configuration we assume the following electron and ion density profiles @xmath1 with @xmath2 for @xmath3 and @xmath4 for @xmath5 . here
@xmath6 is the radius of the ion plasma core and @xmath7 the ion and electron initial particle number .
we denote by @xmath8 the time dependent number of electrons inside the ion core .
+ initially electrons have a maxwellian energy distribution with temperature @xmath9 .
moreover , in these models the electron density is taken to be uniform inside the ion core . as a further simplification , we assume that , on average , the radial crossing of the electron trajectories does not lead to a relative redistribution of the charge in front and behind each electron outside the core .
this simplification allows us to assume that , after leaving the ion core each electron moves as if in an effective time independent coulomb field .
hence , the condition for an electron to reach infinity is that it has a positive total energy when it reaches the ion core surface at @xmath10 . in this model
the particles which satisfy this condition are said to `` evaporate '' and are assumed to be lost when they reach @xmath10 . on the contrary , the electrons that have a negative energy at @xmath10 are assumed to remain inside the ion core .
furthermore , we assume that inside the ion core the electrons move as free particles .
the evaporation of the electrons with positive total energy at @xmath10 changes the number of electrons @xmath8 , the total energy of the system inside the ion core and causes an energy redistribution of the remaining electrons .
+ we discuss the `` collisional '' regime and the `` collisionless '' one . in the first one ,
the electrons which are not evaporated thermalize at a temperature @xmath11 , which turns out to be a decreasing function of time . in the second regime
no thermalization occurs , and the evaporation causes a progressive depletion in the high energy tail of the electron distribution function , which remains isotropic in velocity space . + in what follows , lengths will be measured in units of the initial debye length @xmath12 , with @xmath13 the absolute value of the electronic charge , time @xmath14 in units of @xmath15 , with @xmath16 the electron mass , energies in units of the initial electron temperature @xmath9 , velocities in units of the initial electron thermal speed @xmath17 , mass in units of the electron mass and particle numbers in units of @xmath18 . since inside the ion core the electron density
is taken to be uniform , with the adopted normalization the normalized electron density @xmath19 and the normalized total number of electrons @xmath8 are numerically equal .
+ if the electrons inside the ion core are re - thermalized by collisions , their velocity distribution function remains maxwellian and their time dependent kinetic energy is given by @xmath20 .
the electron evaporation rate is obtained by calculating the flux of electrons with positive total energy through the core surface . defining the time dependent quantity @xmath21 , with @xmath22 the electrostatic potential at the ion core surface
, the positive total energy condition corresponds to @xmath23 .
thus we obtain @xmath24 where @xmath25 is the electron crossing time inside the ion core .
analogously , the energy flux @xmath26 of the evaporating particle can be computed by noting that each evaporating electron carries away the residual energy @xmath27 .
then we obtain @xmath28 the total energy of the system can be written as @xmath29 where @xmath30 is the electrostatic energy of the charged configuration which increases as the electrons evaporate as @xmath31 thus , from the total energy balance we obtain for the time change of the kinetic energy of the system @xmath32 @xmath33 which , coupled to eq . ( [ thermal rate ] ) , gives the time evolution of the temperature @xmath11 . + if on the contrary we assume that plasma electrons inside the ion core are not significantly affected by collisions , their distribution function becomes non - maxwellian .
we assume that the electron distribution remains homogeneous in coordinate space and isotropic in velocity space .
thus , denoting by @xmath34 the time dependent number of electrons with kinetic energy ( normalized on the initial temperatute @xmath9 ) in the interval @xmath35 $ ] , and introducing the time dependent quantity @xmath36 , which differs from @xmath37 in the previous by the normalized temperature factor @xmath38 , we obtain @xmath39 this implies that the evaporation of the electron population with energy @xmath40 stops at a well defined time @xmath41 , where @xmath42 is such that @xmath43 .
therefore , @xmath44 with @xmath45 the @xmath40-dependent decay time and @xmath46 the electron kinetic energy distribution at the initial time @xmath47 .
we assume the electron velocities at time @xmath47 to be maxwellian distributed , hence the initial electron kinetic energy distribution @xmath48 is given by @xmath49 the electron number @xmath8 , and therefore @xmath50 , can thus be calculated performing numerically , at fixed @xmath14 , the integral @xmath51 note that in this collisionless model a rough estimate of the asymptotic electron number could be obtained by approximating the final electron distribution function with the initial one for @xmath52 , and with zero for @xmath53 .
we can then determine the cutoff energy @xmath54 self consistently by equating its value to the electrostatic energy of the configuration with charge @xmath55 , @xmath56 .\ ] ] numerical integration of eqs .
( [ thermal rate ] , [ kinetic spectrum ] ) shows that both in the collisionless and in the collisional regime the electron evaporation rate is significantly reduced from its initial value when the quantities @xmath57 become of order unity .
this means that , in the collisional regime , the evaporation nearly stops when the electron electrostatic energy , which is an increasing function of time , is of the order of the electron temperature which decreases with time .
whereas in the collisionless case the electron evaporation is significant only until the electron electrostatic energy is of the order of the initial average electron kinetic energy .
+ the initial evaporation rate , which is obviously the same in both , the collisional and in the collisionless case , is obtained from either eq .
( [ thermal rate ] ) or eq .
( [ energy decay ] ) which give @xmath58 a linear approximation of @xmath8 is feasible as long as where @xmath59 is the initial characteristic evaporation time and @xmath60 is defined in the two different collisionality regimes by either the condition @xmath61 or @xmath62 and which can be roughly evaluated as @xmath63 therefore , if @xmath64 , i.e. , for ion core radii satisfying @xmath65 , the charging process continues until almost complete depletion of the electron population .
+ finally , we note that the time dependent electron energy distribution predicted in the collisionless regime is highly non thermal .
the fact that the electron evaporation only occurs for those particles with kinetic energy @xmath40 satisfying @xmath66 causes a depletion of the high energy tail of the electron distribution function , as will be examined in detail in sec .
our pic simulations are initialized with the electron and ion density profiles @xmath67 given by eq .
( [ initial profile ] ) .
the initial electron distribution function is maxwellian with temperature @xmath9 . at time
@xmath47 the electrons are allowed to move . during their expansion ,
the electrons that reach the border of the simulation box with positive total energy are removed .
therefore , as the total number of electrons decreases with time , the plasma acquires a net positive charge and an electrostatic potential sets up .
finally a stationary state is reached where no more electrons reach the boundary .
we denote by @xmath68 the electron number , the electron density , the electric field profile , and the electron kinetic energy distribution once ssc has been reached . as expected , our simulations show that these quantities depend on the ion core radius @xmath6 .
+ the results regarding the profile of both the electric field @xmath69 and of the electron density @xmath70 for two different ion plasma radii @xmath6 are presented in fig.[field_and_density.eps ] . as one can see in fig.[field_and_density.eps]b , the electrons which are outside the ion sphere
are confined in a region whose width is of the order of a few debye lengths .
the numerical results regarding the electron kinetic energy distribution in the stationary state are presented in fig.[cut.eps ] . in the figure , the ratio @xmath71 ,
is shown versus the electron kinetic energy @xmath40 in semi - logarithmic scale , for several values of the ion plasma radius @xmath6 .
these results show that the electron kinetic energy distribution is highly non - thermal .
it exhibits a cut in its high energy tail , and the cutoff energy increases with the ion plasma radius @xmath6 .
+ regarding the comparison with the semi - infinite , planar case , our results show that both the electric field and the electron density are very similar to those presented in @xcite only as long as the ion core radius is greater than several tens of debye lengths . since , contrary to a one - dimensional configuration , in the case of a three - dimensional configuration the energy required for the electrons to evaporate is finite , the differences observed are mainly due to the cutoff in the electron high energy tail .
such cutoff is responsible for the electron density depletion observed outside the ion core ( see fig.[field_and_density.eps]b ) and , consequently , for the corresponding electric field profile . as shown in fig.[field_and_density.eps]a , in the limit @xmath72 the value of the dimensionless electric field at the surface of the ion core is almost independent of @xmath6 , thus the net dimensionless charge confined within the ion core scales approximately as @xmath73 .
+ the comparison between the charge value @xmath74 obtained numerically and that predicted analytically is shown in fig.[rfinale.eps ] , for different values of the ion core radius @xmath6 .
note that the value of @xmath75 obtained numerically includes all ssc electrons , i.e. , the electrons inside the ionic sphere and those in the surrounding halo .
it is seen that the agreement among the numerical results and the values obtained in the collisionless regime is very good in the whole range @xmath76 . with regard to the thermal model adopted for the collisional regime
we remark that for small radii it predicts a moderately larger value of @xmath77 , but the two different regimes lead to very close values @xmath77 in case of large radii .
+ our results indicate that collisions can affect the charging up process only for small ion core radii .
this result can be explained by noting that the potential due to the electron expansion scales as @xmath78 .
thus , as the potential barrier increases , the fraction of the electron population that , because of coulomb collisions , reaches a positive total energy and can thus leave the ion core decreases .
however , the collisional thermalization of the distribution function assumed in sec.[colregime ] can only be expected to to apply when the ion core radius is much larger than the electron mean free path , whereas in most plasma conditions the mean free path due to coulomb collisions is much greater than the debye length .
+ it is worth to note that a precise fit of the numerically - obtained plasma charge state is given by the following pad approximation @xmath79 with @xmath80 , @xmath81 and @xmath82 .
+ with regard to the high tail of the electron kinetic energy distribution we define the cutoff energy @xmath83 as the energy satisfying the relation @xmath84 .
the analytical and numerical results obtained for @xmath83 are shown in fig.[cut_comparison.eps ] for different values of the ion core radius @xmath6 .
taking into account these results , we remark that the collisionless model , although not capable of reproducing the fine details of the whole electron kinetic energy distribution observed in the simulations , nevertheless it predicts the cut in the distribution high energy tail with great accuracy .
in this paper we have investigated the charging up process of a spherically symmetric plasma configuration in vacuum in the limit of immobile ions .
two different simplified models have been presented . with the help of these models
we have established scaling laws relating the steady state total charge and electron energy distribution on the radius of the ion core , normalized in terms of the initial electron debye length .
these scalings have been validated by mean of spherical one - dimensional pic simulations .
the agreement we find is overall very good .
+ charged plasma configurations such as those investigated in this paper occur naturally in the interaction of ultraintense laser pulses with matter and are of interest , e.g. , for setting the initial conditions in the study of coulomb explosions and ion acceleration from small clusters irradiated by ultraintense laser pulses .
+ in particular , regarding the problem of cluster expansion the analytical and numerical results that we have presented show that a spherically symmetric configuration of cold ions and hot electrons , which is the typical starting configuration in cluster expansion experiments , does not evolve towards a neutral configuration , in contrast with the one - dimensional planar case .
this charging up effect can strongly modify the maximum energy that the ions can gain and the typical timescale on which their acceleration occurs .
j. e. allen and j. g. andrews , j. plasma physics , * 4 * , 187 ( 1970 ) .
j. e. crow , p. l. auer and j. e. allen , j. plasma physics , * 14 * , 65 ( 1975 ) .
p. d. prewett and j. e. allen , j. plasma physics , * 10 * , 451 ( 1973 ) .
t. nedelea and h. m. urbassek , phys .
e * 69 * , 0546408 ( 2004 ) .
m. passoni and m. lontano , phys .
e * 69 * , 026411 ( 2004 ) .
p. mora , phys .
lett . * 90 * , 185002 ( 2003 ) .
a. a. ivanov and k.s .
serebrennikov , jetp lett . * 78 * , 123 ( 2003 ) .
v. f. kovalev and v. yu .
bychenkov , phys .
. lett . * 90 * , 185004 ( 2003 ) .
v. f. kovalev , v. yu .
bychenkov and v. t. tikhounchuk , jetp lett . * 74 * , 10 ( 2001 ) .
d. s. dorozhkina and v. e. semenov , phys .
lett . * 81 * , 2691 ( 1998 ) .
d. s. dorozhkina and v. e. semenov , jetp lett .
* 67 * , 573 ( 1998 ) .
a. v. gurevitch , l. v. pariskaya and l.p .
pitaievskii , j. plasma physics , * 14 * , 65 ( 1975 ) .
m. widner , i. alexeff and w. d. jones , phys .
fluids * 14 * , 795 ( 1971 ) .
s. betti , f. ceccherini , f. cornolti and f. pegoraro , submitted for publication .
available at http://xxx.lanl.gov/abs/physics/0405030 .
m. hegelich _
et al . _ ,
lett . * 89 * , 085002 ( 2002 ) .
j. badziak _
et al . _ ,
lett . * 87 * , 215001 ( 2001 ) . a. j. mackinnon _ et al .
lett . * 86 * , 1769 ( 2001 ) .
a. j. mackinnon _ et al .
lett . * 88 * , 215006 ( 2002 ) .
a. maksimchuk and v. yu bychenkov , phys .
lett . * 84 * , 4108 ( 2000 ) . e. l. clark _
et al . _ ,
. lett . * 84 * , 670 ( 2000 ) .
e. l. clark _
et al . _ ,
85 * , 1654 ( 2000 ) .
k. krushelnick , phys .
plasmas * 7 * , 2055 ( 2000 ) .
s. p. hatchett _
et al . _ ,
plasmas * 7 * , 2076 ( 2000 ) .
snavely _ et al .
lett . * 85 * , 2945 ( 2000 ) .
habara _ et al .
e * 69 * , 036407 ( 2004 ) .
p. mckenna _
et al . _ ,
e * 70 * , 036405 ( 2004 ) .
* 92 * , 175003 ( 2004 ) .
v. kumarappan , m. krishnamurthy and d. mathur , phys .
a * 66 * , 033203 ( 2003 ) .
v. kumarappan , m. krishnamurthy and d. mathur , phys .
lett . * 87 * , 085005 ( 2001 ) . v.
p. krainov , m.b smirnov , phys .
rep . * 370 * , 237 ( 2002 ) .
h. m. milchberg , s. j. mcnaught and e. parra , phys .
e * 64 * , 056402 ( 2001 ) .
j. zweiback , t. ditmire and m. d. perry , phys .
a * 59 * , r3166 ( 1999 ) .
t. ditmire _
et al . _ , nature ( london ) * 398 * , 489 ( 1999 ) .
t. ditmire _
et al . _ , nature ( london ) * 386 * , 54 ( 1997 ) .
t. ditmire , r. a. smith , j. w. g. tisch and m. h. r. hutchinson , phys .
* 78 * , 3121 ( 1997 ) .
t. ditmire _
a * 53 * , 3379 ( 1996 ) .
y. l. shao _
et al . _ ,
lett . * 77 * 3343 ( 1996 ) .
e. m. snyder , s. a. buzza and a. w. castleman jr .
. lett . * 77 * 3347 ( 1996 ) .
m. borghesi _
et al . _ ,
lett . * 92 * , 055003 ( 2004 ) .
m. borghesi _
et al . _ ,
rev . lett . * 88 * , 135002 ( 2004 ) .
m. borghesi _
et al . _ ,
plasmas * 9 * , 2214 ( 2002 ) .
f. califano , f. pegoraro and s. v. bulanov , phys .
e * 68 * , 066406 ( 2003 ) .
s.v . bulanov _
phys . lett .
a * 299 * , 240 ( 2002 ) .
s.v . bulanov and v. s. khoroshkov , plasma .
phys . rep .
* 28 * , 453 ( 2002 ) .
et al . _ ,
* 89 * , 175003 ( 2002 ) .
r. orecchia _ et al .
_ , clinical reviews in oncology / hemathology * 51 * , 81 ( 2004 ) .
a. brahme , int .
j. radiat .
biol . phys . *
58 * , 603 ( 2004 ) .
w. k. weyrather and j. debus , clinical oncology * 15 * , s23 ( 2003 ) .
h. tsujii , eur . j. cancer * 37 * , s251 ( 2004 ) .
r. orecchia _ et al .
_ , eur . j. cancer * 34 * , 459 ( 1998 ) .
u. amaldi , analysis * 1 * , 1 ( 2003 ) .
m. kanapathipillai _
_ , phys . of plasmas ,
* 11 * , 3911 ( 2004 ) .
v. e. fortov _
_ , new . j. of phys . * 5 * , 102 ( 2003 ) .
j. kasparian _
_ , science * 301 * , 61 ( 2003 ) . | the stationary state of a spherically symmetric plasma configuration is investigated in the limit of immobile ions and weak collisions .
configurations with small radii are positively charged as a significant fraction of the electron population evaporates during the equilibration process , leaving behind an electron distribution function with an energy cutoff .
such charged plasma configurations are of interest for the study of coulomb explosions and ion acceleration from small clusters irradiated by ultraintense laser pulses and for the investigation of ion bunches propagation in a plasma . |
finding and understanding the infra - red structure of holographic rg flows of @xmath3 yang - mills theory and of abjm theory @xcite remains an immensely rich but rather challenging subject that still has the capacity to surprise . in this context ,
gauged @xmath4 supergravity in four and five - dimensions has proven to be a very powerful tool that continues to be extremely useful and yield interesting new results .
the catalog of physically interesting holographic solutions and flows that have been explicitly constructed in gauged supergravity is a very long one , whose early examples included the flows to highly non - trivial @xmath5 supersymmetric `` leigh - strassler '' fixed points @xcite and its abjm analog ( @xcite and @xcite ) , through examples of @xmath2 seiberg - witten flows @xcite , to maximally supersymmetric flows @xcite .
there are many more examples , perhaps the most recent of which is the supersymmetric flow that we studied in @xcite , where the large-@xmath6 theory on a stack of m2 branes flows to a new , `` nearly conformal '' supersymmetric theory in @xmath7 dimensions .
one of our purposes here is to discuss these new families of flows and related janus solutions in some detail .
another purpose of this paper is to highlight and explain some of the new techniques that were used in @xcite .
as has been noted in many places ( see , for example , @xcite ) , finding a holographic flow solution in lower dimensions , perhaps in some consistent truncation , does not often give direct insight into the underlying physics .
more specifically , a holographic flow in the lower - dimensional theory may be singular and fields may flow to infinite values .
it is only when these solutions are uplifted to m theory or iib supergravity that one can give a proper interpretation of the singular behavior in terms of a distribution of branes and fluxes . in this way
, singular low - dimensional solutions may actually encode very interesting physics in higher dimensions .
this , of course , raises the obvious question as to why one does not simply start in the higher - dimensional theory from the outset .
the answer is straightforward : the lower - dimensional theory encodes fields much more simply and computably ; complicated fluxes and metric deformations on internal manifolds become scalars described in terms of a potential .
the supersymmetries may also involve these internal fluxes and geometry in non - trivial ways .
the practical algorithms for solving supersymmetry variations directly in higher - dimensions therefore typically require the imposition of a high level of symmetry and supersymmetry .
the power of using the low - dimensional theory and its potential structure is that one can handle solutions that have much lower levels of symmetry and supersymmetry . as will become evident ,
finding the solutions that we construct here directly in m theory is a truly daunting task , even when one knows exactly where to look .
the price of working in lower - dimensional gauged supergravity is that it describes a very restricted family of deformations .
on the other hand they are some of the most interesting deformations since they are dual to marginal and relevant operators .
what is surprising is that even after fifteen years since the first holographic flows in gauged supergravity @xcite , there are still interesting new physical flows to be found ( like the one in @xcite ) and new janus solutions that can be constructed explicitly in gauged supergravity ( like those in @xcite ) .
since gauged supergravity continues to give us new very interesting , physical solutions , while their interpretation usually requires the `` uplift '' to m theory or iib supergravity , it becomes ever more important to understand and develop the precise relationships between the gauged supergravity in low dimensions and the higher - dimensional supergravities .
in particular , one wants to develop explicit uplift formulae that provide the exact m theory of iib solution in terms of the gauged supergravity fields .
there is also a vast literature on this subject and there has also been some remarkable progress on this in the last two or three years . for simplicity , we will only give a very brief review here and restrict our attention entirely to m theory and its relation to gauged @xmath4 supergravity in four dimensions .
gauged @xmath4 supergravity in four dimensions was first constructed in @xcite and there was a great deal of subsequent work that argued how this must be related to the @xmath8-compactification of m theory .
( for a review , see @xcite . )
there was extensive discussion as to whether gauged supergravity was a consistent truncation or merely a low - energy effective field theory .
the former is a much stronger statement in that it means that solving the equations of motion in gauged supergravity _ guarantees _ that one has also solved the equations of motion of the higher dimensional theory . over the years
it has become evident that the gauged theory is indeed a consistent truncation and formulae have emerged showing precisely how gauged supergravity encodes solutions to m theory .
one of the first general formula was given in @xcite where it was shown how to compute the exact deformed metric on the @xmath8 in terms of all the supergravity scalars .
this knowledge alone was immensely useful in finding uplifted solutions explicitly , see for example @xcite .
exact formulae for fluxes proved to be a much greater challenge .
indeed , the formulae for the components of the 4-form field strength obtained as part of the original proof of consistent truncation of m theory on @xmath8 in @xcite were prohibitively difficult to use and also suffered from an ambiguity that would lead to some components having a wrong symmetry @xcite .
it is only recently that a new set of considerably more workable uplift anstze for the internal 3-form potential have been proposed in @xcite and then extended to the other components of the flux @xcite .
however , explicit tests of those new formulae @xcite were confined to uplifts of simplest solutions of four - dimensional @xmath1 supergravity , namely the @xmath9 solutions for the stationary points of the potential . hence , it is important to perform non - trivial tests for solutions with varying scalar fields , such as holographic flows . in looking at the holographic
flows described in this paper and in @xcite , we first tried to find the uplift based entirely on knowing the internal metric through the formula of @xcite .
this turned out to be impossible and the flux uplift formula became an essential part of constructing the flow solution .
moreover , the janus solutions are intrinsically even more complicated and we certainly could not have constructed them without knowing how to uplift the fluxes .
we find that the new formula to uplift the fluxes @xcite do indeed generate the exact solution .
the only down - side is that they involve some heavy computations to arrive at a relatively simple result . while our focus in this paper will be mainly on the details of how to construct the uplifts
, one should not lose sight of the interesting physics of the solutions that we construct . as we described in @xcite ,
our flow solutions start from a uv fixed point of m2 branes that , under a relevant perturbation , go to solutions sourced largely , or even entirely by m5 branes in the ir .
unlike many flows to the ir , these flows , when uplifted , have only mild orbifold singularities .
moreover , there is a special class of flows that go to pure m5 branes and can be interpreted as describing an `` almost conformal '' fixed point in @xmath7 dimensions .
this was the focus of our earlier paper @xcite . in this paper
we will also look at janus solutions that delve into the backgrounds described by the flows and so may be interpreted as describing interfaces between phases described by the holographic ir flows . to date
, much of the discussion of janus solutions has been done directly in iib supergravity @xcite ( see , however , @xcite ) or m theory @xcite . as we remarked above
, such direct constructions usually require a high level of symmetry , or supersymmetry to make the computations feasible .
in particular , until @xcite , there were very few ( and no ) janus solutions in m theory known .
these new janus solutions were obtained in gauged supergravity and so we want to take one of the most non - trivial families of such flows and uplift to m theory so as to reveal the underlying geometric structure . in section [
sect : truncation ] we describe sector of gauged @xmath4 supergravity upon which we will focus and describe the bps flow and janus solutions from the four - dimensional perspective . in section [ sect : uplift ] we give the details of how this sector of gauged supergravity uplifts to m theory , while in section [ sect : eoms ] we show how this uplifted solution solves the equations of motion in m theory .
the supersymmetry structure of the solutions is studied from the eleven - dimensional perspective in section [ sec : susy11 ] . in section [ sec : largelambda ] we return to studying the ir limits of our holographic flows and how they are related to distributions of m5 and m2 branes .
we discuss general features of our solutions and how one might obtain more general families of solutions in section [ sect : general ] .
section [ sect : conclusions ] contains our concluding remarks .
our conventions and the tabulation of some of the more complicated formulae are given in the appendices .
in this section we summarize some explicit results for the truncation of four - dimensional , @xmath4 supergravity @xcite to the -invariant sector that we will need for the uplift to m theory in section [ sect : uplift ] .
our discussion here is based on @xcite and @xcite .
the lagrangian for the truncation can also be read - off from a more general @xmath10-invariant truncation in @xcite and @xcite .
the @xmath11 symmetry group of the truncation is defined by its action on the supersymmetries , @xmath12 , of the @xmath4 theory .
we choose @xmath10 and the first @xmath13 to act on the indices @xmath14 , while the second @xmath13 on the indices @xmath15 .
this corresponds to the branching @xmath16 the resulting truncation is particularly simple since , as observed in @xcite , the commutant of the symmetry group in @xmath17 consists of a single @xmath18 .
the invariant fields are : the graviton , @xmath19 , the gauge field , @xmath20 , for the two @xmath13 s , a scalar , @xmath21 , and a pseudoscalar , @xmath22 .
as we will describe below , this may be viewed as the bosonic sector of @xmath23 supergravity coupled to a vector multiplet .
the two non - compact generators of @xmath18 in the fundamental representation of @xmath17 can be chosen as follows : @xmath24 where @xmath25 are self - dual ( @xmath26 ) and antiself - dual ( @xmath27 ) @xmath28 tensors , respectively .
then the scalar ` 56-bein ' is @xmath29 where the scalar , @xmath30 , and the pseudoscalar , @xmath31 , parametrize the coset @xmath32 with the canonical complex coordinate , @xmath33 , given by @xmath34 given the explicit generators , it is easy to check that the exponential reduces to a polynomial , @xmath35 where @xmath36 note that the order of this polynomial coincides with the index of embedding of @xmath18 in @xmath17 . using the 56-bein , it is now straightforward to obtain the full bosonic action of the truncated theory @xcite . in particular
, we find that it is consistent to set the vector fields , @xmath20 , to zero .
then the resulting lagrangian for the gravity coupled to the scalar fields is : . ]
@xmath37 & \eql { 1\over 2}\,r -3\,\partial_\mu\lambda\partial^\mu\lambda
-{3\over 4}\sinh^2(2\lambda)\,\partial_\mu\zeta\partial^\mu\zeta+6g^2\cosh(2\lambda)\ , .
\end{split}\ ] ] the lagrangian has no explicit dependence on the phase , @xmath38 , and hence there is a conserved noether current @xmath39 with the corresponding @xmath40 symmetry being simply a rotation between the scalar and the pseudoscalar . it was shown in @xcite that by keeping the @xmath10-invariant fermions , the truncation yields a @xmath23 supergravity in four dimensions .
its @xmath41-symmetry is a combination of the two @xmath42 s and , from the supersymmetry variations , @xmath43 the real superpotential , @xmath44 , is given by an eigenvalue of the @xmath45-tensor , @xmath46 , see appendix [ appendix : b ] . substituting the real fields , @xmath47 and @xmath38 , in , we then find @xmath48 in terms of the superpotential , @xmath44 , the potential @xmath49 is given by @xmath50 - 3\,w^2\,.\ ] ] note that unlike the potential , @xmath51 , the superpotential , @xmath44 , is invariant only under a @xmath52 subgroup of @xmath53 .
in this paper we are interested in a special class of solutions corresponding to rg - flows and one - dimensional defects in the dual abjm theory .
thus we take the metric given by a domain wall ansatz @xmath54 and where the metric function , @xmath55 , and the scalar fields , @xmath56 and @xmath57 , are functions of the radial coordinate , @xmath58 , only .
furthermore , @xmath59 , is either a minkowski metric ( rg - flows ) or a metric on @xmath60 of radius @xmath61 ( janus solutions ) , @xmath62 since , at least formally , the equations for the rg - flows can be obtained by taking the radius @xmath63 , throughout much of the discussion we will write only the more general formulae for the janus solutions .
the equations of motion for the metric and the scalar fields that follow from the lagrangian are @xmath64 \zeta '' & \eql -3\,\,a'\zeta'-4\coth(2\lambda)\,\zeta'\,\lambda'\,,\\[6 pt ] a '' & \eql -{3\over 2}\,(a')^2-{3\over 2}\,(\lambda')^2-{3\over 8}\,\sinh^2(2\lambda)\,(\zeta')^2-{e^{-2a}\over 2\ell^2}\,,\\[6 pt ] \end{split}\ ] ] and @xmath65 where last are two equations are independent combinations of the einstein equations ) follows from ( [ d4eqs ] ) . ] . imposing an unbroken supersymmetry along the flow ,
one obtains a first order system of the bps equations .
we refer the reader to @xcite for further details and here only quote the final result : @xmath66 \zeta ' & \eql-{4\over 3}\,\bigg({a'\over w}\bigg ) \,{1 \over\sinh^2(2\lambda ) } \,{\partial w\over\partial\zeta } ~-~ { 2 \kappa \over 3}\ , \bigg({e^{-a}\over\ell}\bigg ) \,{1\over\sinh(2\lambda)}\,\frac{1}{w}\,{\partial w\over\partial\lambda } \,,\label{janusbps:2}\end{aligned}\ ] ] together with @xmath67 the constant @xmath68 is determined by the chirality of the unbroken supersymmetry , with @xmath69 for @xmath70 and @xmath71 for @xmath72 . in the following we set @xmath70 .
note that is the same as after one eliminates the derivatives of the scalar fields using and .
it is also straightforward to verify that the equations of motion follow from the bps equations .
finally , the bps equations for supersymmetric rg - flows @xmath73 and @xmath74 are obtained from , and by taking the @xmath63 limit .
there is no constraint on the chirality of the unbroken @xmath23 supersymmetry .
the background contours are of the superpotential @xmath75 .
a red dot denotes the `` central point '' of a flow at @xmath76 where @xmath77.,width=302 ] , and the scalar fields , @xmath56 and @xmath57 , for the different types of flows in figure [ spagh].,title="fig : " ] + , and the scalar fields , @xmath56 and @xmath57 , for the different types of flows in figure [ spagh].,title="fig : " ] + , and the scalar fields , @xmath56 and @xmath57 , for the different types of flows in figure [ spagh].,title="fig : " ] -plane .
the background contours are of the real superpotential @xmath75 .
the ridge trajectories have constant @xmath38 with @xmath78 ( green ) and @xmath79 ( red ) , respectively .
, width=302 ] ( red ) and @xmath78 ( green ) with @xmath80.,title="fig:",width=7 ] ( red ) and @xmath78 ( green ) with @xmath80.,title="fig:",width=6 ] the janus solutions to the bps equations have been studied in @xcite where it was shown by a numerical analysis that there are three classes of solutions shown in figures [ spagh ] and [ spprofiles ] : regular janus solutions ( shown in green ) interpolating between two @xmath9 regions corresponding to the same @xmath28 stationary point of the potential and singular solutions that diverge on either one side ( shown in red ) or both sides ( shown in blue ) of the flow .
the janus solutions are characterized by the presence of a special central point along a flow where the solution passes from one branch of to another .
this point is marked by a red dot and the corresponding values of the scalar fields are denoted by @xmath81 and @xmath82 , respectively .
the position of this point for a given flow determines the type of a solution , see figure 3 in @xcite .
in particular , for @xmath83 , all solutions are singular provided @xmath81 is large enough .
it is only when @xmath84 that all solutions are regular janus solutions irrespective of the value of @xmath81 .
in addition , there are solutions akin to rg - flows , which asymptote to @xmath9 on one side and become singular on the other , while remaining on a single branch of .
they can be thought of as a singular limit of janus solutions where the central point is moved off to infinity .
simplest examples of such flows are obtained by taking constant @xmath85 with @xmath86 . solving for @xmath87 , @xmath88 and @xmath89 , and then imposing consistency between them , one is left with @xmath90 { e^{-2a}\over\ell^2 } & \eql { g^2\over 2}\,{\sin^2(3\zeta_0)\,\sinh(2\lambda)\over \cos(3\zeta_0)+\coth(2\lambda)}\ , .
\end{split}\ ] ] choosing the top sign in , we can impose the @xmath9 boundary condition in the uv , that is @xmath91 as @xmath92 , to integrate the first equation for @xmath56 , and then solve the second equation for @xmath55 .
the resulting solutions are similar to the ones in figure [ ridgeflows ] , which we will discuss shortly .
the situation simplifies considerably in the rg - flow limit where the scalar equations do not involve the metric function , @xmath89 , and can be solved first .
choosing the top sign in , which corresponds to the uv region at @xmath93 , one then finds flows shown in figures [ rgflows ] and [ ridgeflows ] @xcite .
in fact , as we have discussed in @xcite , generic solutions for the rg - flows can be determined analytically using two constants of motion : the general one @xmath94 valid for any @xmath61 and corresponding to the conserved current , and the second constant , one would not expect to find such an additional constant of motion for the bps equations at finite @xmath61 . ]
@xmath95 ~=~ & \frac{(4 \ , \cos^2 \zeta -1)}{2
\,\sin^2 \zeta } \,\frac{(3+(\cosh 2\lambda -2 \cos\zeta \ , \sinh 2\lambda ) ^2 ) } { ( \cosh 2\lambda + \cos\zeta \ , \sinh 2\lambda ) ^2 } \ , , \end{split } \label{iom2}\ ] ] for the first order system . using @xmath96 and @xmath97 , one can then determine @xmath89 and @xmath38 as a function of @xmath47 , which is sufficient given the reparametrization invariance for the coordinate along the flow
. however , this method of integration fails for the special flows with constant @xmath85 . for an rg - flow , one must then have @xmath98 ( see , green and red ridge flows in figure [ rgflows ] ) .
the resulting equations can be obatained from by taking the limit , @xmath99 where @xmath100 is a constant .
one can then integrate those equations directly to obtain @xmath101 and @xmath102 where @xmath103 and @xmath100 are integration constants .
the solutions to those ridge flows are shown in figure [ ridgeflows ] . comparing with the trajectories in figure [ rgflows ]
, we expect that the solutions with @xmath104 are representative of the generic rg - flows , while those with @xmath105 are special .
this expectation is confirmed by the asymptotic expansions that we will now discuss .
the holographic rg flows , governed by ( [ rga ] ) and ( [ rgflows ] ) , have @xmath106 at some finite value of @xmath58 . generically ,
such solutions are dual to a massive flow toward some new infra - red limit .
as was observed in @xcite , the rg flow solutions considered here can encode rich and interesting infra - red physics once one examines them in m theory .
the janus flows , governed by ( [ janusbps:1])([apsq ] ) , can either form a loop starting and finishing at @xmath107 or start at @xmath107 and ultimately flow with @xmath106 .
there are also janus solutions that begin and end with @xmath106 .
the janus flows that involve large @xmath47 may be viewed as interfaces that explore the infra - red structure of the holographic flow solutions .
we will therefore examine the limiting behaviors of these flows as @xmath106 .
in section [ sec : largelambda ] we will uplift these results to m theory to see more precisely how they may be interpreted in terms of m branes . from the explicit solutions in figures [ spprofiles ] and [ ridgeflows ] , we see that the limit is characterized by @xmath108 where @xmath109 is a constant asymptotic angle for a given flow .
this observation is confirmed by a more careful expansion of the bps equations . expanding the superpotential as @xmath110 for a generic @xmath38 , we have @xmath111 while for the special ridge flows , @xmath112 the flow equations ( [ rgflows ] ) and ( [ rga ] ) implify @xmath113 { d \zeta \over d\lambda } & ~=~ { 4 \over \sinh^2(2\lambda ) } \,{\partial w\over\partial\zeta}\ , \bigg({\partial w\over\partial\lambda}\bigg)^{-1 } \quad \to \quad 0 \,.\label{dzetadlambda}\end{aligned}\ ] ] the latter confirms the constancy of @xmath38 at infinity while the former shows the rate of divergence of @xmath89 depends upon that angle .
this will translate into different physics once we uplift those flows to m theory .
we have obtained the lagrangian and the bps equations by a consistent truncation of the bosonic lagrangian and the supersymmetry variations of four - dimensional , @xmath4 gauged supergravity . since the latter theory is a consistent truncation of m theory on @xmath8 @xcite , any solution of the equations of motion for the lagrangian can be uplifted to a solution of the eleven - dimensional supergravity . in the next two sections we verify this using explicit uplift formula for the metric @xcite and the recently obtained uplift formulae for the flux @xcite
. similar calculation verifying the new uplift anstze for the flux has been carried out recently for a special class of solutions of the four - dimensional , @xmath4 theory given by some of the stationary points of the potential : @xmath28 , @xmath114 , @xmath115 , @xmath116 @xcite , and so(3)@xmath0so(3 ) @xcite , for which the four - dimensional space - time is @xmath9 and the scalar fields are constant .
our construction of the uplift is similar as in those references , which the reader should consult for any omitted background material .
the construction of an uplift inevitably leads to rather complicated formulae . both to organize the calculation and to write down the result in a sucinct form , it is convenient to express the internal components of the fields in terms of canonical @xmath117 tensors on @xmath8 that are associated with the @xmath17 generators of the scalar fields in the truncation . to this end
let us define : @xmath118 and @xmath119 where @xmath120 are the @xmath28 tensors defined in and @xmath121 are the @xmath28 killing vectors ( one - forms ) and two - forms on the round @xmath8 given in terms of an orthonormal basis of killing spinors , @xmath122 , @xmath123 the inverse radius of @xmath8 is denoted by @xmath124 and @xmath125 .
the circle indicates that @xmath126 is a siebenbein for the round metric on @xmath8 , @xmath127 where @xmath128 , and @xmath129 is the covariant derivative with respect to that metric . unless indicated otherwise , all indices on the @xmath8 tensors are raised and lowered with the _ round _ metric , for example @xmath130 .
the coordinates , @xmath131 , on @xmath8 are for the moment arbitrary .
however , one should note that @xmath132 defined in is a scalar harmonic on @xmath8 and may be thought of as providing a natural internal coordinate on the compactification manifold . by construction , the tensors and are invariant , _
i.e. _ have vanishing lie derivative , under the @xmath133 symmetry group of the truncation .
in particular , the killing vectors for the two @xmath13 s , @xmath134 @xmath135provide additional invariant one - forms on @xmath8 . in the following we will show that the metric and the flux for the uplift can be simply written in terms of the round metric , @xmath136 , the one - forms @xmath137 the three - form , @xmath138 and the scalar , @xmath132 .
the eleven - dimensional space - time for the uplifted solutions is a warped product , @xmath139 , with the metric @xmath140 where .
thus @xmath141 . ]
@xmath142 is the metric in four dimensions for a particular solution at hand .
the internal metric , @xmath143 , is determined by the celebrated formula for its densitized inverse @xcite : @xmath144\,,\ ] ] from which the warp factor , @xmath145 , can be calculated using @xmath146 while it is possible to express the densitized metric entirely using tensors and and their ( contracted ) products , the simplest expression is obtained by noting that the symmetric tensors resulting from such contractions can be rewritten using the round metric , @xmath136 , and bilinears in the one forms @xmath147 , @xmath148 and @xmath149 , as in the following examples : @xmath150 s_{mpq}s_n{}^{pq } & \eql { 1\over 4}(\omg_m-\varth_m)(\omg_n-\varth_n)\,,\\[6 pt ] s_{mpr}s_{nq}{}^{r}\xi^p\xi^q & \eql { 9\over 4 } \,\omg_m\omg_n+ { 3\over 4}(9 - 2\xi)\,\varth_m\varth_n+{3\xi\over 2 } \,\omg_{(m}\varth_{n)}\ , .
\end{split}\ ] ] after some algebra , we then find @xmath151 where all the @xmath152 can be expressed in terms of four - dimensional quantities and the scalar , @xmath132 : @xmath153 c_2 & \eql \frac{1 } { 6 ( 3- \xi ) } \,\sinh ( 2 \lambda ) \cos ( \zeta ) \,,\\[6 pt ] \end{split}\ ] ] @xmath154\,,\\[6 pt ] c_4 & \eql{\sinh(2\lambda)\over 16(\xi-3 ) } \big[(\xi -3 ) \sinh ^2(2 \lambda ) \cos ( 3 \zeta ) + \frac{1}{2 } ( \xi -3 ) \sinh ( 4 \lambda ) -6 \cos \zeta \big]\,,\\[6 pt ] c_5 & \eql { \sinh(2\lambda)\over 16(\xi-3 ) } \big [ ( \xi -3 ) \sinh ( 4 \lambda ) \cos ( 2 \zeta ) + ( \xi -3 ) \cosh ( 4 \lambda ) \cos
\zeta -(\xi + 9 ) \cos \zeta
\end{split}\ ] ] note that the indices on the right hand side in are raised using the round metric , @xmath136 , which is the convention followed throughout this section .
all that is needed now to invert the densitized metric are contraction identities between the one - forms , which can be derived using the explicit form of the @xmath28 tensors and properties of the killing vectors summarized in @xcite and the references therein .
we have @xmath155 it is then straightforward to check that @xmath156 where @xmath157 g_2 & \eql { \cals d \over 36(\xi-3)}\,\cos \zeta \big[3 \sinh ( 4 \lambda ) -(\xi + 3 ) \sinh ^2(2 \lambda ) \cos \zeta\big]\,,\\[6 pt ] g_3 & \eql { \cals d\over 16(\xi-3)}\,\big[3 \sinh ( 4 \lambda ) \cos ( \zeta ) -\sinh ^2(2
\lambda ) ( \xi + 3 \cos ( 2 \zeta ) ) \big]\,,\\[6 pt ] g_4 & \eql { \cals d\over 8(\xi-3)}\,\big[3 \sinh ( 4 \lambda ) \cos ( \zeta ) -\sinh ^2(2 \lambda ) ( \xi \cos ( 2 \zeta ) + 3)\big]\ , , \end{split}\ ] ] and @xmath158 \big[6\cosh ( 2 \lambda ) -(\xi + 3 ) \sinh ( 2 \lambda ) \cos \zeta \big]^{2}}\,.\ ] ] using the contractions , one can also calculate the derivatives of the warp factor given by with respect to @xmath47 and @xmath38 .
then a simple integration yields @xmath159 with the overall normalization set by @xmath160 for the round sphere metric when @xmath107 . dividing out @xmath145 in , yields the internal metric , @xmath161 , in terms of the @xmath117 tensors associated with the truncation .
in addition to the metric tensor , we will also need the corresponding orthonormal frames and those turn out to be rather cumbersome to write down using the invariant tensors .
also , the formulae like tend to obscure the underlying geometry of the solution and its symmetry . to address both of these points , we will now choose a suitable set of coordinates , @xmath131 , on the internal manifold .
as usual , see for example @xcite section 7.1 , this can be done systematically as follows : first , we embed @xmath8 into the ambient @xmath162 as the surface @xmath163 such that the killing vectors @xmath164 defined in are related by triality to the familiar ones , that is @xmath165 where @xmath166
similarly , we have @xmath167 the action of the symmetry group in the ambient space is given by the branching , transform in @xmath168 , while the ambient coordinates , @xmath169 , in @xmath170 of @xmath28 . ] @xmath171 one can choose a representation of @xmath172-matrices such that the @xmath10 generators act in the subspace , @xmath173 , while the two @xmath13 generators have @xmath174 diagonal blocks . then a convenient choice for the coordinates , @xmath175 , on @xmath8 , that makes the symmetry manifest is as follows : @xmath176 y^3+i\ , y^4 & \eql m_7^{-1}\,\cos\chi\sin\theta\,\cos{\frac{\alpha_1}{2}}\,e^{-{i\over 2}(\alpha_2+\alpha_3 ) } e^{-i(\phi+\psi)}\,,\\[6 pt ] y^5+i\ , y^6 & \eql m_7^{-1}\cos\chi\cos\theta \,e^{-i(\phi+\psi ) } \,,\\[6 pt ] y^7+i\ , y^8 & \eql m_7^{-1 } \sin\chi\,e^{-i\phi }
\,,\\ \end{split}\ ] ] where @xmath177 are the @xmath178 euler angles , while the angles @xmath179 and @xmath180 parametrize the u(1)@xmath0u(1 ) isometry . and @xmath181 , respectively . ] in this parametrization ,
the round metric on @xmath8 with unit radius is are implicitly pulled - back onto @xmath8 using . ]
@xmath182\\[6 pt ] & \hspace{42 pt } + \big[d\phi+\cos^2\chi\,\big(d\psi+{1\over 2}\sin^2\theta\,\sigma_3\big)\big]^2\ , , \end{split}\ ] ] where @xmath183 is the metric on @xmath184 and @xmath185 are the @xmath178-invariant forms .
the first line in is the metric on @xmath186 and the second line is the hopf fiber .
the @xmath187 symmetry acts transitivly on @xmath188 and the @xmath179-fiber .
both the metric on @xmath184 and the one form @xmath189 are invariant . using and
, we can now express the invariants introduced in section [ subsec : invariants ] in terms of ambient and local coordinates .
we find that the invariant function , @xmath132 , is simply @xmath190\eql -9 + 12\,m_7 ^ 2\big[(y^1)^2+\ldots+(y^6)^2\big]\\[6 pt ] & \eql 3(1 - 4 \sin^2\chi)\ , , \end{split}\ ] ] while the invariant one - forms are @xmath191 \omg_{(1)}+\varth_{(1 ) } & \eql 8\ , m_7(y^7dy^8-y^8dy^7)\eql -8\,m_7^{-1}\sin^2\chi\,d\phi\,,\\[6 pt ] \omg_{(1)}-3\varth_{(1 ) } & \eql 8\ , m_7(y^1dy^2-y^2dy^1+\ldots+y^5dy^6-y^6dy^5)\\ & \qquad \eql -8 m_7^{-1}\cos^2\chi\,(d\phi+d\psi+{1\over 2}\sin^2\theta\sigma_3)\ , .
\end{split}\ ] ] we also have @xmath192 & \eql -m_7^{-3}\,j_{\cc\pp^3}\wedge \vartheta_{s^7}\ , , \end{split}\ ] ] where @xmath193 are , respectively , the complex structure on @xmath186 and the corresponding sasaki - einstein one - form on @xmath8 . finally , we substitute the invariants and into the warp factor and the metric . to simplify expressions
we define : @xmath194 which are functions of the space - time coordinates , @xmath195 , and the internal coordinate , @xmath196 .
then the internal metric can be written as @xmath197 & \hspace{85 pt}+ { 1\over\sigma^2}\,\big(d\phi+\cos^2\chi\,\big(d\psi+{1\over 2}\sin^2\theta\,\sigma_3\big)\big)^2\,\big]\ , , \end{split}\ ] ] where to simplify the notation we set @xmath198 and @xmath199 .
the warp factor is @xmath200 for @xmath107 , we have @xmath201 and the metric reduces to the metric on the sphere with radius @xmath202 .
the deformation clearly preserves the symmetry as well as the metric along the hopf fiber , which is now rescaled by @xmath203 with respect to the six - dimensional base .
this suggests that there might be some deformed khler geometry still present in the background .
we will return to this point below in section [ sec : largelambda ] .
it is rather remarkable that it took more than 25 years to obtain workable formulae for the four - form flux , @xmath204 .
indeed , while the general proof of the consistent truncation of eleven - dimensional supergravity on @xmath8 @xcite yielded explicit formulae for @xmath205 , those formulae were rather difficult if not impossible to use for all but the simplest stationary point solutions @xcite .
it is only recently that new anstze for various components of the four - form flux were found in @xcite whose complexity is comparable to that of the metric ansatz .
starting with a domain wall solution in four - dimensions with a metric as in and scalar fields depending only on the transverse coordinate , the corresponding four - form flux in eleven - dimensional supergravity can be decomposed into a sum of two terms @xmath206 where @xmath207 is the `` space - time '' flux and @xmath208 is the `` transverse '' flux .
a label @xmath209 indicates a @xmath210 order form along @xmath211 and a @xmath212 order form along the internal manifold , @xmath213 .
since by the poincar or conformal symmetry along the three - dimensional slices in @xmath211 there can be no @xmath214-form in , the bianchi identity , @xmath215 , implies that both @xmath216 and @xmath217 must be closed .
hence @xmath218 , where @xmath219 can have at most one `` leg '' along @xmath220 and thus can be always gauge transformed into a 3-form with all three legs along the internal manifold @xmath213 , that is @xmath221 .
the components @xmath222 are given by the new uplift ansatz @xcite , which , in our conventions , reads @xmath223\,.\ ] ] it is convenient to define a two - form @xmath224 .
evaluating in terms of invariants , we then find @xmath225 where the vector index on the right hand side is raised using the round metric and the coefficients are given by @xmath226 a_{12 } & \eql \frac{1}{64}\ , m_7^{-1}\sinh ^2(2 \lambda ) \sin\zeta\,\big[2\cosh(2\lambda)\cos\zeta -\sinh(2\lambda ) \big]\,,\\[6 pt ] a_{21 } & \eql - \frac{1}{64}\ , m_7^{-1}\sinh ^2(2 \lambda ) \sin\zeta\,\big[2\cosh(2\lambda)\cos\zeta + \sinh(2\lambda ) \big ] \,,\\
a_{22 } & \eql -{1\over 64}\ , m_7^{-1}\sin(3\zeta ) \sinh^3(2\lambda ) \,,\\[6 pt ] a_3 & \eql -{1\over 2}\sin\zeta \sinh(2\lambda ) \ , .
\end{split}\ ] ] contracting with the densitized metric and then using the contraction identities together with @xmath227 \omg^m s_m & \eql { 1\over 12}\,m_7^{-1}(\xi-6)\,d\omg_{(1)}+{1\over 4}\,m_7^{-1}\,d\varth_{(1)}-{1\over 6}\,\xi_{(1)}\wedge\omg_{(1)}\,,\\[6 pt ] \varth^m s_m & \eql { 1\over 12}\,m_7^{-1}\xi\,d\omg_{(1)}-{1\over 4}\,m_7^{-1}\,d\varth_{(1)}-{1\over 6}\,\xi_{(1)}\wedge\omg_{(1)}\ , , \end{split}\ ] ] we find that the internal potential is simply given by @xmath228 where @xmath229 rewriting in local coordinates using and yields @xmath230 \wedge \vartheta_{s^7}\,.\ ] ] note that the @xmath219 has only components along the internal manifold , @xmath213 , so that its field strength , @xmath217 , can have at most one leg along the four - dimensional space - time .
we now turn to the second part of the flux , @xmath216 , which , as shown recently in @xcite , can be determined from the uplift for the transverse dual potential , @xmath231 .
the starting point is the maxwell equation in eleven dimensions , which by setting @xmath204 can be written locally as for definitions and properties of the various duals used in this section . ]
@xmath232 from which the dual potential , @xmath233 , is defined by @xmath234 the space time flux , @xmath216 , is determined by the transverse part of @xmath233 , that is @xmath235 where @xmath236 the six - form , @xmath237 , is given by the uplift ansatz @xmath238}{}^{ij}k^{pkl}(u_{mn}{}^{ij}+v_{mnij})(u^{mn}{}_{kl}+v^{mnkl})\,,\ ] ] where @xmath239 , while @xmath240 is the potential for the volume of the round @xmath8 , @xmath241 evaluating , we find @xmath242 where @xmath243 is the dual on @xmath8 with respect to the round metric . in terms of the local coordinates , @xmath244 then @xmath245\wedge \imath_{\partial_\chi } { \rm v\oo l}_{s^7}\\[10 pt ] & \qquad + 8\,m_7\,\frac { 4 \coth ( 2 \lambda ) \sec \zeta ( 1 - 2 \cos ( 2 \chi ) ) -4 \cos ( 2 \chi ) + 3 \cos ( 4 \chi ) + 5}{\left ( \coth ( 2 \lambda )
\sec \zeta -\cos ( 2 \chi ) \right)^2}\,{\rm
v\oo l}_{s^7}\ , .
\end{split}\ ] ] from , we have @xmath246}{(\cos \zeta + \coth ( 2 \lambda ) ) ( \coth ( 2 \lambda ) -\cos \zeta \cos ( 2 \chi ) ) ^2}\ , { \rm v\oo l}_{s^7}\,.\ ] ] substituting , and in , we get @xmath247 & \hspace{50 pt } -{m_7\over \dx\sigma^2}\big(\cos \zeta \sinh ( 2 \lambda ) ( 2 \cos ( 2 \chi ) -1)-3 \cosh ( 2 \lambda ) \big)\ , { \rm v\oo l}_{s^7}\big]\ , , \end{split}\ ] ] where the dual is with respect to the full metric . using identities and in appendix [ appendix :
a ] and @xmath248 we find that the space time flux is @xmath249 & \qquad\qquad + { m_7\delta^{-3}\over \dx\sigma^2}\big(\cos \zeta \sinh ( 2 \lambda ) ( 2 \cos ( 2 \chi ) -1)-3 \cosh ( 2 \lambda ) \big)\ , { \rm v\oo l}_{1,3}\ , .
\end{split}\ ] ] for the flow solutions where the scalar fields depend only on the radial coordinate , @xmath58 , we have @xmath250 , @xmath251 and evaluates to a very simple expression , @xmath252 where @xmath253 v(r,\chi ) & ~=~ { 3\over 4m_7 ^ 2}\,\sin 2\chi\,(4\cos\zeta\,\lambda'-\sinh ( 4\lambda)\sin\zeta\,\zeta')\ , , \end{split}\ ] ] and @xmath254 is the volume along the @xmath255 or @xmath60 slices .
it is straightforward to verify that @xmath216 given in satisfies the bianchi identity , @xmath256 , when the four - dimensional fields , @xmath55 , @xmath56 and @xmath57 , are on - shell , that is they satisfy the equations of motion in four dimensions .
the calculation above illustrates the point we have raised before , namely , that a rather long and complicated derivation using uplift formulae yields a relatively simple final result . in fact , after we have completed this calculation a paper @xcite appeared where a more direct ansatz for the freund - rubin flux , namely the term in @xmath216 proportional to the volume of the four - dimensional space - time , is proposed . in the present context , the key observation is that the second term in @xmath257 in is the scalar potential of the four - dimensional theory , while the first term is proportional to a derivative of the potential .
this can be generalized to a more efficient uplift formula , which is summarized in appendix [ appendix : fr ] .
we conclude this section with a brief summary of the eleven - dimensional fields constructed in sections [ subsec : metric][subsec : stflux ] .
while the formulae for the uplifted fields are valid for any field configuration in four dimensions , here we will specialize them to the four - dimensional flows we are interested in .
it turns out that the simplest form of the flux is obtained when we use suitable frames for the metric .
we will also need those frames later in the proof of supersymmetry of the rg flows and janus solutions in section [ sec : susy11 ] . given and ,
a natural choice for the frames , @xmath258 , @xmath259 , is to set @xmath260 e^5 & \eql m_7\,\dx ^{-\frac{1}{3 } } \ , \sigma^{\frac{1}{3 } } \ , d\chi\,,\\[6 pt ] e^6 & \eql m_7\ , \dx ^\frac{1}{6 } \ , \sigma^{-\frac{1}{6 } } \ , \cos\chi \ , d\theta \,,\\[6 pt ] e^{7,8 } & \eql \frac{m_7}{2}\ , \dx ^\frac{1}{6 } \ , \sigma^{-\frac{1}{6 } } \ , \cos\chi \,\sin \theta \ , \sigma_{1,2 } \ , , \\[6 pt ] e^9 & \eql \frac{m_7}{2}\ , \dx ^\frac{1}{6 } \ , \sigma^{-\frac{1}{6 } } \ , \cos\chi \,\sin \theta \,\cos \theta \ , \sigma_3 \ , , \\[6 pt ] e^{10 } & \eql m_7 \ , \dx ^\frac{2}{3 } \ , \sigma^{-\frac{2}{3 } } \ , \sin \chi \ , \cos\chi \,\big((d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 ) + \frac{\xi}{\dx } \ , d \phi \big ) \ , , \\[6 pt ] e^{11 } & \eql m_7 \ , \dx ^{-\frac{1}{3 } } \ , \sigma^{-\frac{2}{3 } } \ , \big(d \phi + \cos^2 \chi \,(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 ) \big ) \ , , \end{split}\ ] ] where @xmath261 , @xmath262 , are the frames for the @xmath255 or @xmath60 slices , @xmath263 \sigma(r,\chi ) & \eql \cosh(2\lambda)-\,\cos\zeta\sinh(2\lambda)\cos(2\chi)\ , . \end{split}\ ] ] then the transverse potential , @xmath219 , given in becomes surprisingly simple , @xmath264 where @xmath265 note that the coefficient function , @xmath266 , depends only on the four - dimensional space time radial coordinate . all dependence in on the internal geometry and coordinates enters only through the frames . finally , the space time flux is given in and .
this completes the constriction of the uplift .
in this section we verify explicitly that the metric and the four - form flux in the uplift satisfy the equations of motion of eleven - dimensional supergravity when the four - dimensional metric and the scalar fields satisfy the four - dimensional equations of motion . we start with some technical preliminaries that will help us simplify the algebra in the calculations that follow .
the main idea is to work directly with the functions that appear in the metric and the flux , in particular , with @xmath267 and @xmath266 given in and , respectively , rather than with the scalar fields , @xmath56 and @xmath57 . to this end
we use @xmath268 and @xmath269 to eliminate @xmath38 and @xmath47 in terms of @xmath270 and @xmath271 .
this converts complicated trigonometric expressions into rational functions of the new fields @xmath270 and @xmath271 that are typically easier to evaluate and simplify .
in particular , the four - dimensional equations of motion in the rationalized form are given by @xmath272 -2 \,g^2 \left(p^2+x^2 - 1\right)\\[10 pt ] p '' & \eql -3 \,p'a'+\frac{p}{x^2 } { \big[\left(p^2 + 1\right ) \left(x'\right)^2+x^2 \left(p'\right)^2 - 2 p x p ' x'\,\big]}-2\ ,
g^2\frac { p}{x } { \left(p^2+x^2 + 1\right)}\,,\\[10 pt ] a '' & \eql -{3\over 2}\,(a')^2-{3\over 8\dx^2}\,\big[\left(p^2 + 1\right ) \left(x'\right)^2+x^2 \left(p'\right)^2 - 2 p x p ' x'\big]\\[6 pt ] & \hspace{200pt } + { 3\,g^2\over 2\dx}\,\left(1+(p')^2+(\dx')^2\right)-{e^{-2a}\over 2\ell^2 } \ , , \end{split}\ ] ] and @xmath273 -{g^2\over \dx}\left ( p^2+\dx^2 + 1\right ) + { e^{-2 a}\over\ell^2}\eql 0\,.\ ] ] similarly , we find that the superpotential is given by @xmath274\ ] ] and the bps equations and for the scalars become @xmath275\,a'+\frac{e^{-a}}{\ell } \ , { p \left(3 p^2-x^2 + 3\right)\over w^2}\,,\\[10 pt ]
p ' & \eql - { p\over 4\ , \dx\,w^2}\ , \left[9 p^4 - 6 p^2 \left(x^2 - 3\right)+x^4 - 2 x^2 + 9\right]\ , a'\\[6 pt ] & \hspace{150 pt}+{e^{- a}\over\ell}\,{3 p^4 + 2 p^2 \left(x^2 + 3\right)-x^4 - 2 x^2 + 3\over 4\dx w^2}\ , .
\end{split}\ ] ] as a consistency check one can verify once more that the first order equations and with @xmath44 given in imply the second order equations and that is equivalent to . finally , the other metric and the flux functions are : @xmath276\,,\qquad \xi\eql -{1\over\dx}\,\left(p^2-\dx^2 + 1\right)\,,\ ] ] and @xmath277 v & \eql { 3 \over 4 m_7 ^ 2}\,\sin 2\chi\,\big[\,2(1+p^2)\,{\dx'\over\dx}-2p\,p'\big]\ , .
\end{split}\ ] ] this shows that indeed both the metric and the flux can almost entirely be written down using , up to overall factors , only rational functions of @xmath271 and @xmath270 , and their derivatives
! finally , we will be often able to eliminate trigonometric functions of @xmath196 using @xmath278 which follows from .
the first place where using the rationalized parametrization becomes clearly advantageous is the calculation of the components , @xmath279 , of the four - form flux , @xmath205 .
indeed , for the space - time part of the flux given in we simply have @xmath280 f_{1235 } & \eql { \tan\chi\over \dx^{7/6}\sigma^{4/3}\xi}\,(\sigma
-x ) \big[\left(p^2 + 1\right ) x'-p x p'\big]\ , .
\end{split}\ ] ] turning to the transverse flux , @xmath218 , we note that the part of the three - form potential along @xmath184 in has the complex structure , @xmath281 , as a factor .
thus the corresponding components of the field stength must satisfy @xmath282 modulo this identity , the non - vanishing components of the transverse part of the flux are : @xmath283 f_{569\,10 } & \eql -\frac{m_7 p ( \sigma + x)}{x^{2/3}\sigma ^{4/3 } } \ , , \qquad f_{6789 } \eql \frac{2 m_7 p } { x^{2/3}\sigma^{1/3}}\ , , \end{split}\ ] ] where @xmath284 & \hspace{150 pt}+\sigma \left(p^2 x'-2 p x p'+\left(x^2 + 1\right ) x'\right ) \big]\ , .
\end{split}\ ] ] later we will also need @xmath285 it appears that the flux produced through the uplift formulae is rather special , in particular , we find that the following components @xmath286 accidentally vanish , that is not due to the underlying symmetry of the construction .
the einstein equations of eleven - dimensional supergravity in our conventions . ]
are : @xmath287 we start by evaluating the components of the ricci tensor , @xmath288 , in the basis of frames .
the symmetries of the metric and the dependence of the scalar fields and the metric function in four dimensions on the radial coordinate only , imply that the non - vanisnhing components of the ricci tensor can be at most the following ones : @xmath289 this agrees with the explicit result .
indeed , we find that after imposing the four - dimensional equations of motion in the rational parametrization introduced above , the diagonal components of the ricci tensor can be written in the form @xmath290 where @xmath291 , @xmath292 , @xmath293 and @xmath294 are functions of @xmath270 , @xmath295 and @xmath196 ( or , equivalently , @xmath296 ) . in particular , we find that the cross - terms @xmath297 and @xmath298 are absent .
similarly , the off - diagonal components are of the form @xmath299 explicit formulae for all non - vanishing coefficient functions are given in appendix [ appendix : c ] .
evaluating the energy - momentum tensor on the right hand side in is straightforward .
we will forego the details and just look at one specific equation , the off - diagonal einstein equation with @xmath300 and @xmath301 .
on the one side , we have@xmath302 however , as one can see by inspection of the non - vanishing flux components , the other side must be zero .
this verifies the relation between the four - dimensional coupling constant , @xmath303 , and the inverse radius of the internal manifold , @xmath304 , @xcite @xmath305 given this relation , it is easy to check that all the remaining einstein equations are satisfied as expected .
the maxwell equations are @xmath306 for the flux , they yield seven independent equations : four first order and three second order .
the first order equations are along the components @xmath307 $ ] , @xmath308 $ ] , @xmath309 $ ] , and @xmath310 $ ] , and all have the same structure as this last one @xmath311 namely , they come with an overall factor of @xmath312 .
the second order equations come from the components @xmath313 $ ] , @xmath314 $ ] and @xmath315 $ ] in .
the first two equations are somewhat involved , but the last one is quite simple .
there we find @xmath316 \\[10 pt ] & \eql \frac{2 p \left(g^2 - 2 m_7 ^ 2\right ) } { x^{4/3}\sigma ^{2/3}}\ , , \end{split}\ ] ] where in going from the first to the second line we have used the four - dimensional equations of motion .
similarly , upon using , the other two equations reduce to the same expression modulo an overall factor of @xmath317
. thus the maxwell equations are satisfied if holds . to summarize ,
we have shown explicitly that the metric , @xmath318 , and the four - form flux , @xmath205 , constructed using the uplift formulae in section [ sect : uplift ] indeed satisfy the equations of motion of the eleven - dimensional supergravity when the scalar fields , @xmath56 and @xmath57 , and the metric function , @xmath55 , are on - shell in four dimensions .
it is important to note that to verify that we have used only the equations of motion in four dimensions or , equivalently , , but not the bps equations !
this means that also non - supersymmetric solutions of the same type will uplift to solutions of m theory .
we now turn to the janus and rg - flow solutions of the bps equations and , respectively , to demonstrate explicitly the @xmath71 and @xmath23 supersymmetry of the corresponding uplifts in m theory .
this has been discussed already in some detail in @xcite , where we have argued that the @xmath23 supersymmetry of the rg flows is achieved by brane polarization and is naturally defined through projectors that reflect the underlying almost - complex structure and a dielectric projector much like those encountered in @xcite .
the defect in the janus solutions leads to additional chiral projector that is also present in four dimensions .
the result for the rg - flows is then recovered by keeping both chiralities and taking the @xmath63 limit .
the bps equations in eleven dimensions are obtained by setting the supersymmetry variations of the gravitinos to zero , @xmath319 where the algebraic operators , @xmath320 , are given by @xmath321 the killing spinors of unbroken supersymmetries are invariant under the poincar transformations in the @xmath322-plane and are singlets of @xmath10 acting along @xmath188 .
hence @xmath323 does not depend on the cordinates @xmath324 , @xmath21 and @xmath325 , as well as the euler angles , @xmath326 .
this means that the corresponding equations are purely algebraic : label components with respect to the frames , while @xmath327 , or @xmath328 , with respect to the local coordinates and/or forms . ]
@xmath329 similarly , the dependence of @xmath323 on the @xmath330 angles , @xmath180 and @xmath179 , @xmath331 is determined by the charges , @xmath332 and @xmath333 , respectively .
let us now consider the first equation in , written in the form @xmath334 the matrix @xmath335 , expanded into the basis of @xmath172-matrices , is given by @xmath336\,\gamma^4 + { m_7 \dx ^{1/3}\over 6\sigma^{4/3}}\,{\partial\sigma\over\partial\chi}\,\gamma^5\\[6 pt ] & \qquad + { 1\over
3}\,\left(f_{1234}\gamma^4+f_{1235}\gamma^5\right)\,\gamma^{123}+{1\over 6}\,f_{45\,10\,11}\,\gamma^{45\,10\,11}+{1\over 6}\,f_{6789}\,\gamma^{6789}\\[6 pt ] & \qquad + { 1\over 6}\ , \left ( f_{469\,11}\gamma^{4\,11}+ f_{569\,10}\gamma^{5\,10}\right ) \left(\gamma^{69}+\gamma^{78}\right)\ , .
\end{split}\ ] ] together with the explicit formulae for the flux , this gives us a homogenous system of linear equation for the thirty two components of @xmath323 .
it is clear that after substituting the expressions for the flux components and expanding the derivatives of @xmath296 , see and , the operator @xmath335 , as well as the other operators , @xmath337 , become quite complicated .
hence , before we proceed with the analytic calculation , we first explore numerically the space of solutions to .
to do that , we first eliminate the derivatives @xmath338 , @xmath339 and @xmath87 using the bps equations and , and set @xmath340 .
next we assign random values to the fields @xmath271 , @xmath270 , @xmath89 , the angle @xmath196 , and the constants @xmath304 and @xmath61 upon which becomes a purely numerical system that can be solved for the components of the killing spinor , @xmath323 .
note that our numerical assignment amounts simply to choosing random initial conditions for the four - dimensional bps equations and thus is not constrained in any way .
those numerical solutions yield us some information about the subspace of allowed killing spinors , which confirms what one could also infer from an analysis in four dimensions and the symmetry .
more importantly , it allows us to short cut quite a bit of tedious analysis by fixing some of the signs in the projectors below that we would have to keep track of otherwise . for finite @xmath61 ,
the space of numerical solutions is generically two - dimensional in agreement with @xmath69 supersymmetry in four dimensions .
the unbroken supersymmetries , @xmath323 , must thus satisfy four conditions @xmath341 where @xmath342 are mutually commuting projectors . from the numerical analysis we also find that two of these projectors are constant . to conform with the conventions in @xcite , we will denote them by @xmath343 and @xmath344 . the first projector , @xmath345 arises from the fact that the killing spinor , @xmath323 , must be a singlet under the holonomy group , @xmath10 , of @xmath188 .
it depends on the choice of orientation of @xmath188 defined by the frames @xmath346 . the second projector
, @xmath347 is just an uplift of the corresponding chirality projector in four dimensions . in particular
, choosing @xmath72 in and changes the sign in .
finally , we find that on the subspace of the killing spinors satisfying , @xmath348 together with , this gives us two additional algebraic equations , which as we will see simplifies the calculations significantly .
we should also note that both projectors do not depend on the choice of the square root branch in used to eliminate @xmath87 . for the rg flows , taking the limit
@xmath63 eliminates the first term in .
the space of solutions includes then both @xmath349-chiralities ; the projector @xmath350 is thus absent and we have a four - dimensional space of solutions corresponding to @xmath23 supersymmetry .
we have shown in @xcite that the remaining two commuting projectors in this limit are @xmath351\,,\ ] ] and @xmath352\,,\ ] ] where the angles @xmath353 , @xmath354 and @xmath355 are some functions of @xmath58 and @xmath196 . in analogy with , the projector can be associated with an extension of the complex structure of @xmath188 to an almost complex structure with extra pair of complex frames . finally , is the dielectric deformation of the standard m2-brane projector at @xmath356 . for
the janus solutions , the projectors and must be deformed to account for the defect , which gives rise to additional terms in the supersymmetry variations , such as the first term in .
including such terms in and leads to the following ansatz for the projectors at finite @xmath61 : @xmath357\,,\ ] ] and @xmath358\,.\ ] ] those two operators form a pair of commuting projectors provided the vectors @xmath359 and @xmath360 are orthonormal .
such a pair of vectors can be parametrized by three angles , @xmath353 , @xmath354 and @xmath361 : @xmath362 a_3 & \eql \sin \alpha \sin \beta \cos \gamma + \cos \beta \sin \gamma \ , , \end{split}\ ] ] and @xmath363 together with @xmath355 those angles are some functions of @xmath58 and @xmath196 and will be determined by solving the supersymmetry variations . for @xmath364 ,
the projectors @xmath365 and @xmath366 reduce to @xmath367 and @xmath368 , respectively .
we can thus view the angle @xmath361 as the janus deformation parameter which goes to zero in the rg - flow limit .
there is still certain redundancy in our description of the projectors and . to see this ,
introduce a third vector , @xmath369 , so that @xmath370 are orthonormal and define @xmath371 .
observe that the product @xmath372 commutes with all the projectors , @xmath373 and so preserves the space of supersymmetries .
one is therefore free to rotate and using the action of @xmath372 and this induces a simultaneous rotation @xmath374 accompanied by a rotation of @xmath375 and @xmath369 by the angle @xmath376 .
in the following we will use this freedom to simplify our calculations .
we will now calculate all the projectors and the killing spinor , @xmath323 , by solving explicitly the bps equations . in principle , one should be able to determine all the projectors in , or equivalently solve for the angles @xmath377 and @xmath355 , directly from .
the problem is that this effectively amounts to obtaining the individual projectors @xmath342 from a particular linear combination of products of these projectors .
this , unsurprisingly , is not the best way to proceed .
instead , we will first solve algebraic equations that arise from judicious linear combinations of the variations in which the flux terms either cancel completely or are simple .
the first such equation arises from the `` magical combination '' of variations @xmath378 in which all flux terms cancel . after eliminating the derivatives with respect to the @xmath13 angles using and modulo terms annihilated by @xmath343 and @xmath344
, it reads @xmath379\,\epsilon\eql 0\,,\ ] ] where @xmath380 \fa_4 & \eql -\frac{m_7\,{\dx^{1/3 } } ( \cos ( 2 \chi ) -2 ) ) } { \sin(2\chi)\ , { \sigma ^{1/3}}}\,,\qquad \fa_5\eql \frac{m_7\,{\dx^{1/3 } } \left(\dx ( \cos ( 2 \chi ) -2)+3
\,\sigma \right)}{2 \cos^2\chi\ , { \sigma^{1/3 } } } \ , . \end{split}\ ] ] iterating one finds a single consistency condition @xmath381 which is satisfied by virtue of and .
this condition also means that , up to an invertible factor , is in fact the projector with @xmath382 where @xmath383 using and , we then read off @xmath384 & \hspace{1 in}\sin\alpha \eql -{a\sin(2\chi)\,a'\over\omega^{1/2}}\ , , \end{split}\ ] ] and @xmath385 where @xmath386 before proceeding we note that the rotation of the gamma matrices that define the projectors is equivalent to a rotation of the frames .
in particular , the rotation by @xmath355 is equivalent to starting with the frames : @xmath387 using ( [ omegares1 ] ) we find a rather simple result for one of these frames : @xmath388 note that the mixing of @xmath180 and @xmath179 does not involve functions of @xmath58 and furthermore ( [ mphpseqs ] ) implies that the supersymmetries only depend upon angles in precisely the combination @xmath389 .
we will return to this observation later .
continuing with the supersymmetry analysis , since the projectors and commute , we can not obtain any information from about the dielectric polarization angle , @xmath354 .
for that we turn to another magical combination , @xmath390 which has no derivatives of @xmath323 and no terms with components of the internal flux . after imposing the constant projections , it reads @xmath391\,\epsilon\eql 0\,,\ ] ] where @xmath392\,,\\[6 pt ] \fb_3 & \eql \frac{2 \dx a'+\dx'}{4\ , \sigma^{3/2 } \dx^{7/6}}\ , , \qquad \fb_4 \eql -\fb_5\eql -\frac{m_7\,{\dx^{1/3 } } \tan \chi } { \sigma^{1/3}}\,,\qquad \fb_6\eql{m_7 \over \sigma^{1/3}\dx^{2/3}}\ , .
\end{split}\ ] ] note that the presence of the @xmath393-term in , with an analogous term absent in , prevents from being a projector .
still , by iteration one finds a consistency condition @xmath394 which is indeed satisfied by virtue of and . using the projectors and in
, one is left with three independent products of @xmath172-matrices which yield the following equations : @xmath395 @xmath396 @xmath397 however , only one of those equations is independent , which can be seen by solving one of them for @xmath398 and then verifying that the other two are satisfied .
equivalently , one can solve the first two for @xmath399 and @xmath400 and then check that their squares indeed add up to one . substituting the result into the third equations yields a consistency condition @xmath401 this equation has a simple geometrical interpretation , namely that the ( non - unit ) vector @xmath402 satisfies @xmath403 which is the consistency condition between the operator in and the projector . similarly as for the equations of motion in section [ sect : eoms ] , the solution for @xmath399 and @xmath400 above can be simplified using rationalized bps equations .
after some algebra , we find the following result : @xmath404 where @xmath405\,{a'\over w^2}\,,\\[10 pt ] s_0 & \eql { p \,\omega^{1/2}\over \sqrt 2\,\,\sigma}{a'\over w^2}\ , , \end{split}\ ] ] and @xmath406\ , , \\[10 pt ] s_1 & \eql -\frac{\sqrt{\omega } \csc ( 2 \chi ) } { 4 \,\sigma \,\dx ^{3/2 } \ , w^2 } \big[\dx ^4 \sin
^2(\chi ) + \dx ^2 \left(p ^2 - 1\right ) ( \cos ( 2 \chi ) -2)+3 \left(p ^2 + 1\right){}^2 \cos ^2(\chi ) \big]\ , .
\end{split}\ ] ] this completes the calculation of all the angles in the projectors @xmath365 and @xmath366 . to determine explicitly the killing spinors for unbroken supersymmetries ,
let us introduce rotations @xmath407 and define @xmath408 which commute with the projectors @xmath343 and @xmath344 .
it is straightforward to check that the projectors and are then simply @xmath409 where @xmath410 thus any solution @xmath323 to can be written as @xmath411 where @xmath412 is in the kernel of the constant projectors , and .
from the supersymmetry variations along the radial direction , @xmath22 , we find @xmath413 which is the correct radial dependence for the killing spinor along @xmath60 .
this leaves two variations along @xmath58 and @xmath196 , which are solved as usual by setting @xmath414 where @xmath415 is the warp factor of the `` time '' frame , @xmath416 , and @xmath417 is a constant spinor along the internal manifold and with the standard dependence along @xmath60 , which satisfies the same constant projections as @xmath418 .
the supersymmetry analysis simplifies significantly for the holographic flow solution . for this one
simply imposes the projectors ( [ proj1 ] ) , ( [ proj2inf ] ) and ( [ proj0inf ] ) but does not impose a helicity projector like ( [ smproj ] ) .
we then find , taking the upper signs in ( [ rga ] ) and ( [ rgflows ] ) : @xmath419 @xmath420 and @xmath421 \,,\\[10 pt ] \sin\beta & \eql { p \,\omega^{1/2}\over \sqrt 2\,w\,\sigma}\ , . \end{split}\ ] ] the space - time components of the maxwell fields also simplify and we obtain a seemingly standard relation for holographic flows : @xmath422
having constructed the uplift in detail , we now examine the infra - red limits of the holographic rg flows described by ( [ rga ] ) and ( [ rgflows ] ) from the perspective of m theory . in an earlier paper @xcite we focussed upon the special flow with @xmath423 since this led to a very interesting new result . here
we will complete the asymptotic analysis for all flows .
first recall that @xmath38 limits to a constant value as @xmath424 and so the various warp factors behave as follows : @xmath425 for such a generic @xmath38 one has @xmath427 for some constant , @xmath41 .
thus the warp factor for the branes and the corresponding frames are finite and smooth for @xmath428 : @xmath429 thus the metric parallel to the branes is simply : @xmath430 from ( [ d11frames ] ) and ( [ 11dasymp1 ] ) one has @xmath431
e^{11 } ~\sim~ & \frac{2\ , m_7}{(1 + \cos \zeta)^{\frac{1}{3 } } \ , \widehat \sigma^{\frac{2}{3 } } } \ , e^{-2 \lambda } \ , \big(d \phi + \cos^2 \chi \,(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 ) \big ) \ , , \end{split}\ ] ] these are the only two frames to depend on @xmath47 in this limit .
the remaining frames limit to : @xmath432 it is instructive to rewrite @xmath433 in terms of the one - form appearing in @xmath434 : @xmath435 \ , , \label{newe11asymp}\ ] ] then one has @xmath436 & \hspace{70 pt } \times\ , \big(d \phi+ \frac{(1+\cos \zeta ) \cos^2 \chi } { \widehat \sigma }
\,\big(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 + \frac{2 \ , \cos \zeta } { ( 1+\cos \zeta ) } \ , d \phi\big ) \big)^2 \big ] \ , , \label{r2met } \end{split}\ ] ] where @xmath437 .
the remaining part of the metric is @xmath438 & \qquad\qquad ~+~ \frac{(1+\cos \zeta ) ^2 } { \widehat \sigma^2 } \
, \sin^2 \chi \,\cos^2 \chi \,\big(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 + \frac{2 \ , \cos \zeta } { ( 1+\cos \zeta ) } \ , d \phi\big ) \big ] \ , , \end{split}\ ] ] the full eleven - dimensional metric limits to the sum of ( [ 3met1 ] ) , ( [ r2met ] ) and ( [ 6met1 ] ) .
observe that ( [ 6met1 ] ) is conformally khler .
that is , the metric @xmath439 & ~+~ \frac{(1+\cos \zeta ) } { \widehat \sigma } \
, \sin^2 \chi \,\cos^2 \chi \,\big(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 + \frac{2 \ , \cos \zeta } { ( 1+\cos \zeta ) } \ , d \phi\big ) \ , , \label{6met2 } \end{split}\ ] ] has a khler form : @xmath440 ~=~ & d\ , \big[\coeff{1}{2}\ , \cos^2 \chi \ , \big(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 + \frac{2 \ , \cos \zeta } { ( 1+\cos \zeta ) } \ , d \phi\big ) \big ] \ , , \end{split}\ ] ] where @xmath441 is the khler form on @xmath442 .
here we are , of course , taking @xmath38 to be constant at its asymptotic value .
one can also easily verify that as @xmath443 this manifold is smooth , and is locally like the origin of @xmath444 .
the only singular parts of the metric occur at @xmath445 and at @xmath446 , where there are orbifold singularities in two different @xmath447 planes in ( [ r2met ] ) and ( [ 6met1 ] ) respectively . as we will discuss below , these loci represent the intersections of the various branes that are present in the infra - red limit .
the non - zero components of the maxwell field are given by : @xmath448 with @xmath449 \hat e^{11 } ~\equiv~ & \frac{2\ , m_7}{(1 + \cos \zeta)^{\frac{1}{3 } } \ , \widehat \sigma^{\frac{2}{3 } } } \ , \big(d \phi + \cos^2 \chi \,(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 ) \big ) \ , .
\label{e11hatdefn}\end{aligned}\ ] ] thus @xmath450 has regular coordinate components .
one might be concerned that the maxwell tensor has a singular source at @xmath445 because the @xmath433 is vanishing
. however , the frame components of the maxwell tensor are , in fact , regular .
the non - zero frame components in the compactified directions ( including @xmath433 ) are : @xmath451 f_{451011 } ~\sim~ & -\frac{2\,m_7\ , { \rm sign } ( 1- 2\cos\zeta)\ , \sin \zeta } { ( 1 + \cos \zeta)^\frac{2}{3 } \ , \ , \widehat \sigma^\frac{1}{3 } } \ , , \label{fasymp1b}\\[6 pt ] f_{56910 } \ , , f_{57810 } ~\sim~ & - \frac{2\,m_7\ , ( 1 + \cos \zeta \,\sin ^2\chi ) \ , \sin \zeta } { ( 1 + \cos \zeta)^\frac{2}{3 } \ , \ , \widehat \sigma^\frac{4}{3 } } \ , , \qquad f_{6789 } ~\sim~
\frac{2\,m_7 \sin \zeta } { ( 1 + \cos \zeta)^\frac{2}{3 } \ , \ , \widehat \sigma^\frac{1}{3 } } \,.\label{fasymp1c}\end{aligned}\ ] ] it is also useful to note that the electric part of the maxwell field is extremely simple @xmath452 here we simply take @xmath454 because the flow for @xmath455 simply involves reversing the sign of the internal components of the flux , @xmath450 .
one now has rather different asymptotics : @xmath456 @xmath457 & ~+~ \frac{64}{27 } \,\rho^2 \ , \big(d \phi + \frac{3\,\cos^2 \chi } { 2\,\widehat \sigma}\,(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 + \coeff{2}{3 } \ , d \phi ) \big)^2 \\[6pt ] & ~+~ \frac{4}{3 } \,m_7^{-2}\ , \big ( \ , d \chi^2 ~+~ \frac{3 } { 2\,\widehat \sigma } \,\cos^2 \chi \ , ds^2_{\ic\ip_2 } \\[6pt ] & \qquad\qquad ~+~ \frac{9 } { 4\ , \widehat \sigma^2 } \ , \sin^2 \chi \,\cos^2 \chi \,\big(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 + \coeff{2}{3 } \ , d \phi\big)\,\big)^2 \,\big ] \ , , \end{split}\ ] ] where , as before , @xmath458 note that the compact six - dimensional metric in ( [ 11met1 ] ) is simply the metric ( [ 6met1 ] ) specialized to @xmath423 and is therefore also conformally khler .
remarkably , for @xmath423 many of the components of @xmath205 vanish in the infra - red and we find that this limiting maxwell field is simply given by @xmath459 note that the space - time components parametrized by @xmath460 vanish in this limit and that @xmath461 is purely magnetic and lives entirely on the conformally khler six - manifold .
thus , for @xmath423 there are only m5 branes in the infra - red : the m2 branes have dissolved completely .
the first and rather remarkable surprise is that the warp factor , @xmath462 , in front of frames parallel to the m2-branes ( [ d11frames ] ) is not singular in the infra - red for @xmath428 . for @xmath463 , this warp factor
is expected to be singular because such a flow has no internal fluxes and the warp factor is then simply a power of the harmonic function describing m2 brane sources that have spread on the coulomb branch . however ( [ 11dasymp3 ] ) shows that there is no singularity for generic @xmath38 and for @xmath464 equation ( [ 11met1 ] ) shows that this warp factor actually vanishes .
thus there are no strongly singular sources of m2 branes in the infra - red .
the second surprise is that the internal six - dimensional manifold goes to a _ finite - sized _ conformally khler , six - dimensional manifold and this manifold is smooth at @xmath465 .
indeed , the only singularities are conical and occur at @xmath445 where the @xmath13 fiber defined by @xmath433 pinches off ( see ( [ r2met ] ) ) and at @xmath446 where the @xmath13 fiber defined by @xmath434 pinches off ( see ( [ 6met1 ] ) ) .
it is also evident from ( [ h0asymp1 ] ) and ( [ fasymp1a])([fasymp1c ] ) that the core of this holographic flow is populated by finite , smooth electric ( m2-brane ) and magnetic ( m5-brane ) fluxes .
thus there is evidently brane polarization and a geometric transition in which the m2 branes partially dissolve into smooth m5-brane fluxes leaving a finite sized `` bubble '' in the form of a six - dimensional khler manifold .
this is rather reminiscent of the kind of transition one finds in microstate geometries @xcite . to understand the brane content in the infra - red in more detail it is perhaps easiest to examine the projectors that define the supersymmetries .
these are given by , and .
define the rotated frames @xmath466 \gamma^{\widehat{10}}&~\equiv~\cos\omega\ , \gamma^{10}+\sin\omega\ , \gamma^{11}\ , , \qquad \gamma^{\widehat{11 } } ~\equiv~\cos\omega\ , \gamma^{11}-\sin\omega\ , \gamma^{10}\ , , \end{aligned}\ ] ] where @xmath467 and @xmath468 are functions that depend upon the flow . the details of these angles and how they flow are given in section [ subsec : solsuper ] and may also be found in @xcite . given the other projectors and the fact that @xmath469 , one can write as @xmath470 where @xmath471 is any one of the following @xmath472 this means that the flow represents m2 branes polarizing into three sets of m5 branes that have @xmath7 common directions , those of the m2 branes and one compactified direction , defined by @xmath473 .
this means that the directions transverse to the m5 branes are defined by @xmath474 , @xmath475 and four of the compact internal directions .
thus the brane wrapping is crucially determined by @xmath473 and hence by @xmath355 . for @xmath476 and @xmath106 ,
one has : @xmath477 thus @xmath478 , @xmath479 and so @xmath196 lies transverse to all the branes .
indeed , ( [ betair1 ] ) shows that @xmath480 involves only anti - m2 brane sources and so the conical singularity at this point is not altogether surprising .
the locus @xmath481 also defines the location of the residual m2 branes and of some of the m5 branes and thus another conical singularity is not surprising .
all the m5 branes have a common direction along @xmath434 , which is the hopf fiber in the khler metric ( [ 6met2 ] ) .
one rather interesting flow involves having @xmath482 at infinity .
this does not mean that @xmath483 all along the flow ; indeed ( [ iom2 ] ) takes the value @xmath484 on such a flow and this implies that as @xmath485 one must have @xmath486 .
what makes this flow interesting is that @xmath487 symmetry is restored in the infra - red . in particular , the metric ( [ 6met2 ] ) becomes precisely that of @xmath488 as described in @xcite , the situation is very different for @xmath476 . for @xmath489 and @xmath106 ,
one has simply flips the signs of the internal fluxes and is completely equivalent . ] : @xmath490 and @xmath491 we now have @xmath492 and so the m5 branes wrap @xmath493 while @xmath494 remains transverse to the branes .
more significantly , the m2-brane flux now vanishes entirely and all that remains is a very simple non - singular magnetic ( m5-brane ) flux ( [ irfsimp2 ] ) .
the limiting metric ( [ 11met1 ] ) is almost like that of @xmath495 where @xmath496 is the conformally khler metric .
the five - dimensional manifold that we label as @xmath497 is @xmath498 in poincar form with one spatial direction compactified and fibered over @xmath496 .
holographically it suggests that the ir phase is almost a cft except that one spatial direction has been `` put in a periodic '' box of some fixed scale and that some interactions have been turned on so that this direction becomes non - trivially fibered .
thus the ir phase is almost a cft fixed point .
finally , we would like to note that all the fluxes and most , if not all , of the metric in the ir limit are purely functions of @xmath196 .
this however , does not mean that these limits represent solutions to the equations of motion because the @xmath58 ( or @xmath47 ) dependence is critical to giving finite terms that survive in the ir limit of the equations of motion .
there are several natural generalizations of the results presented here .
the first and most obvious is to use a somewhat more general gauged supergravity ansatz .
our ansatz may be thought of as reducing to @xmath2 supergravity coupled to one vector and with a holomorphic superpotential , @xmath499 @xcite .
this can easily be generalized to @xmath2 supergravity coupled to three vector multiplets while still remaining within gauged @xmath4 supergravity .
this truncation was considered in @xcite and the holomorphic superpotential becomes @xmath500 where the @xmath501 are the complex scalars of the three vector multiplets .
our results here may be thought of as the special case with the three vector multiplets set equal and , in particular , @xmath502 .
as noted in @xcite , the uplift formulae will be far more complicated , but one expects that the infra - red limit will involve a more general khler manifold with a @xmath503 symmetry .
it may also have some non - trivial moduli in that the @xmath489 condition may simply become a constraint on the overall phase of @xmath504 .
these moduli would probably be related to the three distinct sets of m5-brane fluxes on the khler manifold .
we intend to investigate this further .
we have also attempted a much greater generalization in the spirit of @xcite .
one starts with the uplifted flow solution and introduces rotated frames that subsume the need for the rotation by @xmath353 : @xmath505 rather remarkably one can integrate these frames in our flow solution to find new variables , @xmath506 , so that @xmath507 one can also find explicit expressions for these new coordinates : @xmath508 combined with the observation implicit in the simple and very canonical form of @xmath475 given in ( [ e10hatsimp ] ) , one finds that our flow solution has a some extra structure and , in particular , the pairing of @xmath509 and the frame in @xmath475 in the supersymmetry projectors , along with the phase dependence of the supersymmetries is very suggestive of an underlying six - dimensional complex structure .
we therefore start with the @xmath506 coordinates and their associated frames .
we then take the metric and fluxes to have a completely general -invariant ansatz involving arbitrary functions of @xmath506 everywhere possible .
we also assume that the frame @xmath475 is universal and , in particular , take the metric ansatz to be of the form : @xmath510 e^5 & \eql h_0(u , v)^{-{1\over6 } } \ , h_2(u , v ) \ , dv \,,\qquad e^6 \eql h_0(u , v)^{-{1\over6 } } \ , h_3(u , v ) \ , v \ , d\theta \,,\\[6 pt ] e^7 & \eql \frac{1}{2}\ , h_0(u , v)^{-{1\over6 } } \ , h_3(u , v ) \ ,
v \ , \sin \theta \ , \sigma_1 \ , ; \qquad e^8 \eql \frac{1}{2}\ , h_0(u , v)^{-{1\over6 } } \ , h_3(u , v ) \ ,
, \sin \theta \ , \sigma_2 \,,\\ e^9 & \eql \frac{1}{2}\ , h_0(u , v)^{-{1\over6 } } \ , h_3(u , v ) \ , v \ , \sin \theta \,\cos \theta \ , \sigma_3 \ , ; \\[6 pt ] e^{10 } & \eql h_0(u , v)^{-{1\over6 } } \ , h_4(u , v ) \,\big(d\phi ~+~ \coeff{3}{2}\,(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 ) \big ) \ , ; \\[6 pt ] e^{11 } & \eql h_0(u , v)^{-{1\over6 } } \ , h_5(u , v ) \ , \big(d \phi + g(u , v ) \,\big(d\phi ~+~ \coeff{3}{2}\,(d\psi + \coeff{1}{2}\ , \sin^2 \theta\ , \sigma_3 ) \big)\big ) \ , , \end{split}\ ] ] where the @xmath511 and @xmath512 are , _ ab initio _ , undetermined functions .
for the supersymmetry projectors we take @xmath513 at the outset but retain @xmath514 .
we also assume that the supersymmetries have the same @xmath179- and @xmath180-dependence as in ( [ mphpseqs ] ) .
note that this implies that the supersymmetry only depends on the combination @xmath515 , which appears in @xmath434 . for the maxwell field
, we take the most general -invariant potential and choose a gauge in which all the components along @xmath516 vanish : @xmath517 & ~+~ p_2 \ , ( e^{6 } \wedge e^{9 } + e^{7 } \wedge e^{8 } ) \wedge e^{10 } ~+~ p_3 \ , ( e^{6 } \wedge e^{9 } + e^{7 } \wedge e^{8 } ) \wedge e^{11 } \ , , \end{split}\ ] ] where the @xmath460 and the @xmath270 s are also functions of @xmath506 .
the whole point is that this ansatz at least contains our uplifted flow solution and we want to use this more general structure to understand the underlying geometry and perhaps find more general supersymmetric solutions . to that end , one solves the supersymmetry variations to fix as many functions as possible . in @xcite
the whole problem reduced to determining a single `` master function '' from which every other flux and metric function was derived .
this master function itself satisfied a non - trivial , non - linear differential equation . in spite of the nice structure that we have discovered
, the same kind of procedure applied here does not lead to such a simple reduction .
we did , however , discover some rather general results that we will briefly summarize so as to give a flavor of what emerges . first , we find that , _ along the flow , _ the metric functions @xmath518 and @xmath519 are all fixed algebraically in terms of @xmath520 and @xmath521 .
moreover , the @xmath509 derivative of @xmath520 is simply related to @xmath520 and @xmath521 .
these conditions combine to reveal that the non - compact , eight - dimensional metric transverse to the m2-branes , @xmath522 , has six - dimensional foliations , defined by the @xmath523 , the coordinate @xmath509 and the frame @xmath524 , are necessarily a @xmath525-dependent family of khler metrics . by this
we mean that there is indeed a khler form , @xmath526 , on each leaf of the foliation but @xmath527 is proportional to @xmath528 : if @xmath525 were held constant , @xmath527 would indeed vanish .
thus the six - dimensional khler structure apparent in the ir actually descends from a family of such structures along the flow .
secondly , we find that the flux parametrized by @xmath529 is necessarily pure gauge and so we can set @xmath530 without loss of generality . the rest of the @xmath270 s satisfy a complicated system of equations that link them with @xmath354 and the remaining metric functions
. ultimately one can reduce the system to show that all the unknown functions are determined in terms of _ two _ unknown functions of @xmath506 , @xmath520 and a function that we will call @xmath461 .
the former satisfies an extremely complicated non - linear differential equation while @xmath461 is a pre - potential in that its derivatives determine some of the functions in the ansatz : the the @xmath509-derivative of @xmath461 gives the function @xmath512 in ( [ gend11frames ] ) while the @xmath525-derivative of @xmath461 gives @xmath531 .
the remarkable fact is that @xmath461 is _ harmonic _ ( annihilated by the laplacian ) in a metric that is conformal to @xmath522 .
the laplacian on @xmath522 explicitly involves @xmath520 and so the hamonicity of @xmath461 is far from simple to use in practice .
we have , of course , verified that the uplifted flows do indeed satisfy these conditions but so far have not been able to simplify and elucidate the general discussion to the degree that it is worthy of presentation in this paper .
the important bottom - line though is that every generalization we have considered shows the same structure for the eight - manifold transverse to the m2 branes : it consists of six - dimensional khler manifolds foliated over an @xmath447 base where one of the @xmath13 s acts as an isometry and the khler potentials depend upon the radial coordinate in this @xmath447 .
on a technical level , our results represent a highly non - trivial test of the recent results on uplifting gauged @xmath4 supergravity to m theory .
indeed , our discussion in section [ sect : general ] illustrates just how difficult it would be to construct the very symmetric class of flows we consider directly within eleven dimensions .
more broadly , we have , once again , seen how apparently very singular `` flows to hades '' in gauged supergravity can encode some very interesting physical flows and janus solutions when lifted to m theory . results like this illustrate why it is very important to understand how gauged supergravities are encoded in higher dimensional theories . the uplift formulae to eleven - dimensional supergravity have been well - studied and tested compared to the uplift of gauged @xmath1 supergravity in five dimensions to iib supergravity in ten dimensions .
one reason might be that for quite some time only uplift formulae for the internal metric and dilaton were known @xcite and those were inferred by analogy to the m - theory result rather than proven directly .
however , this has changed recently with various reformulations of type iib supergravity and the resulting uplift formulae for all fields @xcite .
while there are quite a number of interesting physical examples of iib flows , it is possible that there are others yet to be discovered because they look singular from the five - dimensional perspective . given that the dual theory in @xmath532 yang - mills theory , it would be interesting to examine such flows using the new anstze that are now available .
the limitation in using gauged supergravity is that one is restricted to a relatively small family of fields from the higher - dimensional perspective .
fortunately the fields one has are quite a number of the simplest relevant and marginal perturbations and so one can still probe interesting physics .
the limitation is most sorely felt when one tries to probe details and subtleties of the families of ir fixed points and for this gauged supergravity is too blunt an instrument . on the other hand
, the uplifts of gauged supergravity solutions can give invaluable insights into the geometric structures that underlie the more general classes of flow and thus enable broader , and perhaps more physically interesting solutions to be found .
this was evident for flows @xcite and so even though gauged supergravity sometimes does not describe the exact physics one wants , it can motivate and inform the search for physically interesting families of solutions . in this spirit
, we suspect that the families of flows and janus solutions considered here should admit interesting generalizations .
there are the generalizations within gauged supergravity as outlined at the beginning of section [ sect : general ]
. however , there should be families that involve a six - dimensional khler manifold fibered over a two - dimensional base with a @xmath13 isometry .
it will probably be very challenging to use merely this information to find general flow solutions .
however , we saw here that the fluxes also took on a relatively simple form and if one could understand the geometry underlying this one should be able to move towards the general class of solutions . from the physical perspective , the flows and janus solutions we have constructed are very interesting in that they involve m2 branes polarizing and dissolving into _ non - singular _ ( except for orbifolds ) distributions of m5 branes and for one choice of parameter , flowing to a higher - dimensional , almost - conformal fixed point @xcite .
indeed , it was shown in @xcite that such flows exist for any choice of parameter if one does not insist on supersymmetry .
apart from the interesting holographic interpretation of these flows , this kind of mechanism also underpins the microstate geometry program in which black holes are replaced with smooth , horizonless solitonic geometries .
( for reviews , see @xcite . )
the basic mechanism means that the black hole undergoes a phase transition ( driven by the chern - simons interactions ) in which the singular electric charge sources are replaced my smooth magnetic fluxes . in m
theory , this is realized by m2 charges being replaced by m5 fluxes .
there are vast families of supersymmetric examples of this in asymptotically flat backgrounds but so far there are no examples of this process ( supersymmetric , or not ) in asymptotically @xmath9 or @xmath498 space - times .
it is thus extremely helpful to have an example of such a transition and perhaps use it to understand how such a phase transition might occur more generally in asymptotically @xmath9 or @xmath498 space - times .
finally , there is the question of whether there are iib analogs of the flows and janus solutions studied here .
these would probably be flows in which d3 branes polarized into families of intersecting d5 branes while preserving @xmath533 supersymmetry in the field theory along the flow . such solutions might even `` flow up dimensions '' to give a compactified higher - dimensional field theories in the infra - red .
we would like to thank n. bobev , h. godazgar , m. godazgar , o. krger and h. nicolai for helpful discussions and correspondence .
this work was supported in part by doe grant de - sc0011687 .
we use the same conventions as in @xcite with the `` mostly plus metric '' and the eleven - dimensional equations of motion given by @xmath534 @xmath535 the maxwell equation can be rewritten in terms of forms as @xmath536 where @xmath537 is the hodge dual in eleven dimensions . in general
, we define the hodge dual of a @xmath538-form , @xmath355 , in @xmath539-dimensions by @xmath540 where @xmath541 then @xmath542 with the @xmath26 sign is for a positive definite metric and the @xmath27 sign for a minkowski signature mostly plus metric . for a @xmath543-form @xmath544 on @xmath545 with the warped product metric
, we have a convenient decomposition of the hodge dual : @xmath546 where @xmath547 and @xmath548 are , respectively , the dual on @xmath211 with respect to the four - dimensional part of the metric , @xmath19 , and the dual on @xmath213 with the internal metric , @xmath161 .
factoring out the warp factor , we have @xmath549 where @xmath550 is a @xmath270-form on @xmath211 and @xmath551 is the dual with respect to @xmath552 .
#1#2#3#4a_2#1^#2#3#4 using the 56-bein and with the @xmath28 gauge field set to zero , the @xmath553 composite gauge field of the @xmath4 theory , @xmath554}\eql -2\sqrt 2\,\left(u^{ij}{}_{ij}\partial_\mu v^{klij}-v^{ijij}\partial_\mu u^{kl}{}_{ij}\right)\,,\ ] ] has the following non - vanishing components : @xmath555 hence the kinetic action of the scalar fields is @xmath556 & \eql -3\,\partial_\mu\lambda\partial^\mu\lambda -{3\over 4}\sinh^2(2\lambda)\,\partial_\mu\zeta\partial^\mu\zeta \ , .
\end{split}\ ] ] similarly , for the @xmath89-tensors , @xmath557 and @xmath558}$ ] , we have : @xmath559 and @xmath560 \aa2134 & \eql \aa2156\eql \aa4123\eql \aa4356\eql \aa6125\eql \aa6345\eql { ( 1+z)\bar z\over ( 1-|z|^2)^{3/2}}\,,\\[6 pt ] \aa 1278 & \eql \aa3478\eql \aa 5678\eql-{(1+z)z\over ( 1-|z|^2)^{3/2}}\,,\\[6 pt ] \aa2178 & \eql \aa4378\eql \aa6578\eql { ( 1+z)z\over ( 1-|z|^2)^{3/2}}\,,\\[6 pt ] \aa7128 & \eql \aa7348\eql \aa7568\eql -{z+\bar z^2\over ( 1-|z|^2)^{3/2}}\,,\\[6 pt ] \aa8127 & \eql\aa8347\eql \aa 8567\eql { z+\bar z^2\over ( 1-|z|^2)^{3/2}}\ , .
\end{split}\ ] ] then the scalar potential is @xmath561
the calculation of the space - time part of the flux , @xmath216 , using the method employed in section [ subsec : stflux ] is quite involved even for much simpler solutions such as uplifts of stationary points . for the latter solutions only the freund - rubin part of the space - time flux is present , so that @xmath562 is proportional to the volume of the four - dimensional space - time , @xmath211 , where the proportionality constant is determined universally by the scalar potential of the four - dimensional theory @xcite .
this has been generalized recently in @xcite to uplifts of arbitrary solutions by including corrections proportional to derivatives of the scalar potential .
the new conjectured formula for the freund - rubin flux , @xmath563 , reads @xmath564\,.\ ] ] the @xmath565 tensor is proportional to the first variation of the potential , @xmath51 , along the noncompact generators of @xmath17 acting on the scalar coset , @xmath566 .
it is given by @xcite @xmath567m}-a_1{}^{m[i}a_{2\,m}{}^{jkl]}\,.\ ] ] the second tensor in is a self - dual contraction @xmath568 where @xmath569}\,.\ ] ] note that at a stationary point of the scalar potential , the @xmath570-tensor becomes anti - self - dual @xcite and hence the contraction terms in vanish . specializing the contraction in to the present solution we find @xmath571 then , using , and
, we obtain @xmath572\\[6 pt ] & \eql { m_7\over 3}\ , u\ , , \end{split}\ ] ] which agrees with the calculation of the space - time flux in section [ subsec : stflux ] .
the non - vanishing coefficients of the diagonal components of the ricci tensor , @xmath573 , as defined in : @xmath574 & \hspace{130 pt } + \sigma ^2 \left(3 \left(p^3+p\right)^2 + 3 p^2 x^4 - 2 \left(3 p^4 + 7 p^2 + 4\right ) x^2\right)\big]\\[10 pt ] \cals b_{1 } & \eql { \dx^2\over 6 \dx^{13/3 } \sigma^{8/3}\xi^2 } \big [ -8 p^2 x^2 + 16 p^2 \sigma x \\[6 pt ] & \hspace{130 pt}+\sigma ^2 \left(3 p^4 - 2 p^2 \left(3 x^2 + 1\right)+3 \left(x^2 - 1\right)^2\right)\big]\,,\\[10 pt ] \cals c_{1 } & \eql { p\dx\over 3 \dx^{13/3 } \sigma^{8/3}\xi^2 } \big[-4 \sigma
x \left(3 p^2+x^2 + 3\right)+8 \left(p^2 + 1\right ) x^2\\[6 pt ] & \hspace{130 pt}+\sigma ^2 \left(-3 p^4+p^2 \left(6 x^2 - 2\right)-3 x^4 + 10 x^2 + 1\right)\big]\,,\\[10 pt ] \cals d_{1 } & \eql { 2m_7^{-2}\over 3\dx^{4/3}\sigma^{8/3}}\,\big[\sigma ^2 \left(6 x^2-g^2 m_7 ^ 2 \left(3 p^2 + 3 x^2 + 4\right)\right)\\[6 pt ] & \hspace{130 pt}-2 \sigma x \left(g^2 m_7 ^ 2 + 2 p^2 + 2\right)-2 \left(p^2 + 1\right ) x^2 \big]\ , ; \end{split}\ ] ] @xmath575 & \hspace{180 pt}+\sigma ^2 \left(3 \left(p^3+p\right)^2 + 3 p^2 x^4 - 2 \left(p^2 + 1\right ) x^2\right)\big]\ , , \\[10 pt ] \cals b_{4 } & \eql { 1\over 3 \dx^{7/3}\sigma^{8/3}\xi^2}\big[-2 p^2 \sigma x \left(3 p^2 + 3 x^2 - 5\right)+2 \left(3 p^2 - 1\right ) p^2
x^2\\ & \hspace{200 pt } + \sigma ^2 \left(3 p^4 - 2 p^2 + 3 \left(x^2 - 1\right)^2\right ) \big]\\[10 pt ] \cals c_{4 } & \eql { 1\over 3 \dx^{10/3}\sigma^{8/3}\xi^2 } \big[2 \sigma x \left(3 p^2 \left(p^2+x^2\right)+x^2 - 3\right)-2 \left(3 p^4 + 2 p^2 - 1\right ) x^2\\[6 pt ] & \hspace{200 pt}+\sigma ^2 \left(-3 p^4 - 2 p^2 - 3 x^4 + 4 x^2 + 1\right ) \big]\,,\\[10 pt ] \cals d_{4 } & \eql { m_7^{-2}\over 3\dx^{4/3 } \sigma^{8/3}}\,\big [ 2 \sigma ^2 \left(-3 m_7 ^ 2 g^2 \left(p^2+x^2\right)- 4 m_7 ^ 2 g^2 + 6 x^2\right)\\[6 pt ] & \hspace{200 pt}-4 \sigma x \left(m_7 ^ 2 g^2 + 2 p^2 + 2\right)-4 \left(p^2 + 1\right ) x^2 \big]\ , ; \end{split}\ ] ] @xmath576 \cals b_5 & \eql { 4p^2 x ( \sigma -x)^2\over 3\dx^{10/3}\sigma^{8/3}\xi^2 } \ , , \\[10 pt ] \cals c_5 & \eql -{4p\over 3\dx^{10/3}\sigma^{8/3}\xi^2}\,(\sigma -x )
\big[\sigma \left(p^2+x^2 + 1\right)-2 \left(p^2 + 1\right ) x\big]\ , , \\[10 pt ] \cals d_5 & \eql \frac{2}{3 \sigma ^{8/3 } x^{4/3 } } \,\big[-2 \sigma x \left(g^2 - 4 m_7 ^ 2 \left(p^2 + 1\right)\right)+2 g^2 \sigma ^2+m_7 ^ 2 \left(4 p^2 + 1\right ) x^2\big]\ , ; \end{split}\ ] ] @xmath577 \cals b_6 & \eql -\frac{2 p^2 ( \sigma -x)^2}{3 \sigma ^{8/3 } x^{7/3 } \xi^2 } \,,\\[10 pt ] \cals c_6 & \eql { 2p\over 3\dx^{10/3}\sigma^{8/3}\xi^2}\,(\sigma -x ) \big[\sigma \left(p^2+x^2 + 1\right)-2 \left(p^2 + 1\right ) x\big ] \,,\\[10 pt ] \cals d_6 & \eql { 2\over 3 \dx^{4/3}\sigma^{8/3}}\,\big [ \sigma ^2 \left(m_7 ^ 2 \left(9 p^2 + 6\right)-g^2\right)\\[6 pt ] & \hspace{100 pt}+\sigma x \left(g^2 + 2 m_7 ^
2 \left(p^2 + 1\right)\right)+m_7 ^ 2 \left(p^2 + 1\right ) x^2\big ] \ , ; \end{split}\ ] ] @xmath578 \cals b_{10 } & \eql { 2\,p^2 x\over 3\dx^{10/3}\sigma^{8/3}\xi^2}\ , ( \sigma -x ) \left[\left(3 p^2 + 1\right ) x+\sigma \left(2 - 3 x^2\right)\right]\,,\\[10 pt ] \cals c_{10 } & \eql -{4p\over 3\dx^{10/3}\sigma^{8/3}\xi^2}\ , ( \sigma -x ) \left(\sigma \left(-\left(3 p^2 + 2\right ) x^2+p^2 + 1\right)+\left(3 p^4 + 4 p^2 + 1\right ) x\right ) \,,\\[10 pt ] \cals d_{10 } & \eql -{2 \over 3\dx^{4/3}\sigma^{8/3}}\,\big[2 \sigma x \left(g^2 - 4 m_7 ^ 2 \left(p^2 + 1\right)\right)-2 g^2 \sigma ^2-m_7 ^ 2 \left(4 p^2 + 1\right ) x^2 \big ] \ , ; \end{split}\ ] ] @xmath579 & \hspace{180pt}+{\sigma ^2\over 2 } \left(3 \left(p^3+p\right)^2 + 3 p^2 x^4 - 2 \left(p^2 + 1\right ) x^2\right ) \big ] \,,\\[10 pt ] \cals b_{11 } & \eql { 1\over 3\dx^{7/3}\sigma^{8/3}\xi^2}\,\big [ 2 \left(3 p^2 - 1\right ) p^2 x^2 - 2 p^2 \sigma x \left(3 p^2 + 3 x^2 - 5\right)\\[10 pt ] & \hspace{180pt}+\sigma ^2 \left(3 p^4 - 2 p^2 + 3 \left(x^2 - 1\right)^2\right ) \big ] \,,\\[10 pt ] \cals c_{11 } & \eql { 2p\over 3\dx^{10/3}\sigma^{8/3}\xi^2}\,\big [ 2 \sigma x \left(3 p^2 \left(p^2+x^2\right)+x^2 - 3\right)-2 \left(3 p^4 + 2 p^2 - 1\right ) x^2\\[10 pt ] & \hspace{180pt}+\sigma ^2 \left(-3 p^4 - 2 p^2 - 3 x^4 + 4 x^2 + 1\right ) \big ] \,,\\[10 pt ] \cals d_{11 } & \eql { 2\over 3\dx^{4/3}\sigma^{8/3}}\,\big[-\sigma ^2 \left(3 g^2 p^2-g^2 \left(3 x^2 + 1\right)+6 m_7 ^ 2 \left(x^2
+ 1\right)\right)\\[10 pt ] & \hspace{180pt}-4 \sigma x \left(m_7 ^ 2 \left(p^2 + 1\right)-g^2\right)-m_7 ^ 2 \left(2 p^2 - 1\right ) x^2 \big ] \ , .
\end{split}\ ] ] the non - vanishing coefficients of the off - diagonal components of the ricci tensor , @xmath288 , @xmath580 , as defined in : @xmath581 \cals b_{45 } & \eql - { 2m_7\tan\chi\,p\over \dx^{5/6 } \sigma^{8/3}\xi}\,(x-\sigma ) ( 2 \sigma + x)\ , ; \end{split}\ ] ] @xmath582 o. aharony , o. bergman , d. l. jafferis and j. maldacena , `` n=6 superconformal chern - simons - matter theories , m2-branes and their gravity duals , '' jhep * 0810 * , 091 ( 2008 ) [ arxiv:0806.1218 [ hep - th ] ] . a. khavaev , k. pilch and n. p. warner , `` new vacua of gauged n=8 supergravity in five - dimensions , '' phys .
b * 487 * , 14 ( 2000 ) [ hep - th/9812035 ] .
d. z. freedman , s. s. gubser , k. pilch and n. p. warner , `` renormalization group flows from holography supersymmetry and a c theorem , '' adv .
phys . * 3 * , 363 ( 1999 ) [ hep - th/9904017 ] .
k. pilch and n. p. warner , `` a new supersymmetric compactification of chiral iib supergravity , '' phys .
b * 487 * , 22 ( 2000 ) [ hep - th/0002192 ] .
k. pilch and n. p. warner , `` n=1 supersymmetric renormalization group flows from iib supergravity , '' adv .
* 4 * , 627 ( 2002 ) [ hep - th/0006066 ] .
n. halmagyi , k. pilch , c. romelsberger and n. p. warner , `` holographic duals of a family of n=1 fixed points , '' jhep * 0608 * , 083 ( 2006 ) [ hep - th/0506206 ] .
n. p. warner , `` some new extrema of the scalar potential of gauged @xmath583 supergravity , '' phys .
b * 128 * , 169 ( 1983 ) . c. h. ahn and j. paeng ,
`` three - dimensional scfts , supersymmetric domain wall and renormalization group flow , '' nucl .
b * 595 * ( 2001 ) 119 [ hep - th/0008065 ] .
c. h. ahn and k. woo , `` supersymmetric domain wall and rg flow from 4-dimensional gauged n=8 supergravity , '' nucl .
b * 599 * , 83 ( 2001 ) [ hep - th/0011121 ] .
c. h. ahn and k. s. woo , `` domain wall and membrane flow from other gauged d = 4 , n=8 supergravity .
part 1 , '' nucl .
b * 634 * , 141 ( 2002 ) [ hep - th/0109010 ] . c. h. ahn and t. itoh , `` an n = 1 supersymmetric g-2 invariant flow in m theory , '' nucl .
b * 627 * , 45 ( 2002 ) [ hep - th/0112010 ] . c. h. ahn and t. itoh , `` the eleven - dimensional metric for ads / cft rg flows with common su(3 ) invariance , '' nucl .
b * 646 * , 257 ( 2002 ) [ hep - th/0208137 ] . c. h. ahn and k. s. woo , `` domain wall from gauged d = 4 , n=8 supergravity .
part 2 , '' jhep * 0311 * , 014 ( 2003 ) [ hep - th/0209128 ] .
k. pilch and n. p. warner , `` n=2 supersymmetric rg flows and the iib dilaton , '' nucl .
b * 594 * , 209 ( 2001 ) [ hep - th/0004063 ] .
a. buchel , a. w. peet and j. polchinski , `` gauge dual and noncommutative extension of an n=2 supergravity solution , '' phys .
d * 63 * , 044009 ( 2001 ) [ hep - th/0008076 ] .
n. j. evans , c. v. johnson and m. petrini , `` the enhancon and n=2 gauge theory : gravity rg flows , '' jhep * 0010 * , 022 ( 2000 ) [ hep - th/0008081 ] .
c. n. pope and n. p. warner , `` a dielectric flow solution with maximal supersymmetry , '' jhep * 0404 * , 011 ( 2004 ) [ hep - th/0304132 ] .
i. bena and n. p. warner , `` a harmonic family of dielectric flow solutions with maximal supersymmetry , '' jhep * 0412 * , 021 ( 2004 ) [ hep - th/0406145 ] .
d. z. freedman , s. s. gubser , k. pilch and n. p. warner , `` continuous distributions of d3-branes and gauged supergravity , '' jhep * 0007 * , 038 ( 2000 ) [ hep - th/9906194 ] .
n. p. warner , `` holographic renormalization group flows : the view from ten - dimensions , '' class .
* 18 * , 3159 ( 2001 ) [ hep - th/0011207 ]
. s. s. gubser , `` curvature singularities : the good , the bad , and the naked , '' adv .
* 4 * , 679 ( 2000 ) [ hep - th/0002160 ] .
c. n. gowdigere and n. p. warner , `` holographic coulomb branch flows with n=1 supersymmetry , '' jhep * 0603 * , 049 ( 2006 ) [ hep - th/0505019 ] .
l. girardello , m. petrini , m. porrati and a. zaffaroni , jhep * 9812 * , 022 ( 1998 ) [ hep - th/9810126 ] . j. distler and f. zamora , `` nonsupersymmetric conformal field theories from stable anti - de sitter spaces , '' adv .
theor . math .
* 2 * , 1405 ( 1999 ) [ hep - th/9810206 ] .
n. bobev , k. pilch and n. p. warner , `` supersymmetric janus solutions in four dimensions , '' jhep * 1406 * , 058 ( 2014 ) [ arxiv:1311.4883 [ hep - th ] ] .
b. de wit and h. nicolai , `` n=8 supergravity , '' nucl .
b * 208 * , 323 ( 1982 ) .
m. j. duff , b. e. w. nilsson and c. n. pope , `` kaluza - klein supergravity , '' phys .
* 130 * , 1 ( 1986 )
. b. de wit and h. nicolai , `` @xmath584 supergravity with local su(8 ) invariance , '' nucl .
b * 274 * , 363 ( 1986 ) .
b. de wit and h. nicolai , `` the consistency of the s**7 truncation in d=11 supergravity , '' nucl .
b * 281 * ( 1987 ) 211 .
h. nicolai and k. pilch , `` consistent truncation of d = 11 supergravity on ads@xmath585 , '' jhep * 1203 * , 099 ( 2012 ) [ arxiv:1112.6131 [ hep - th ] ] . b. de wit and h. nicolai , `` deformations of gauged so(8 ) supergravity and supergravity in eleven dimensions , '' jhep * 1305 * , 077 ( 2013 ) [ arxiv:1302.6219 [ hep - th ] ] .
h. godazgar , m. godazgar and h. nicolai , `` testing the non - linear flux ansatz for maximal supergravity , '' phys .
d * 87 * , 085038 ( 2013 ) [ arxiv:1303.1013 [ hep - th ] ] .
h. godazgar , m. godazgar and h. nicolai , `` generalised geometry from the ground up , '' jhep * 1402 * , 075 ( 2014 ) [ arxiv:1307.8295 [ hep - th ] ] . h. godazgar , m. godazgar and h. nicolai , `` nonlinear kaluza - klein theory for dual fields , '' phys .
d * 88 * , no .
12 , 125002 ( 2013 ) [ arxiv:1309.0266 [ hep - th ] ] .
h. godazgar , m. godazgar , o. hohm , h. nicolai and h. samtleben , `` supersymmetric e@xmath586 exceptional field theory , '' jhep * 1409 * , 044 ( 2014 ) [ arxiv:1406.3235 [ hep - th ] ] . h. godazgar , m. godazgar , o. krueger and h. nicolai , `` consistent 4-form fluxes for maximal supergravity , '' arxiv:1507.07684 [ hep - th ] .
e. dhoker , j. estes and m. gutperle , `` ten - dimensional supersymmetric janus solutions , '' nucl .
b * 757 * , 79 ( 2006 ) [ hep - th/0603012 ] .
e. dhoker , j. estes and m. gutperle , `` exact half - bps type iib interface solutions .
i. local solution and supersymmetric janus , '' jhep * 0706 * , 021 ( 2007 ) [ arxiv:0705.0022 [ hep - th ] ] .
a. clark and a. karch , `` super janus , '' jhep * 0510 * , 094 ( 2005 ) [ hep - th/0506265 ] . m. -w .
suh , `` supersymmetric janus solutions in five and ten dimensions , '' jhep * 1109 * , 064 ( 2011 ) [ arxiv:1107.2796 [ hep - th ] ] .
e. dhoker , j. estes , m. gutperle and d. krym , `` exact half - bps flux solutions in m - theory .
i : local solutions , '' jhep * 0808 * , 028 ( 2008 ) [ arxiv:0806.0605 [ hep - th ] ] .
e. dhoker , j. estes , m. gutperle and d. krym , jhep * 0812 * , 044 ( 2008 ) [ arxiv:0810.4647 [ hep - th ] ] .
j. estes , r. feldman and d. krym , `` exact half - bps flux solutions in @xmath588 theory with d(2,1;@xmath589;0)@xmath590 symmetry : local solutions , '' phys .
d * 87 * , no . 4 , 046008 ( 2013 ) [ arxiv:1209.1845 [ hep - th ] ] .
c. bachas , e. dhoker , j. estes and d. krym , `` m - theory solutions invariant under @xmath591 , '' fortsch .
phys . * 62 * , 207 ( 2014 ) [ arxiv:1312.5477 [ hep - th ] ] .
n. bobev , n. halmagyi , k. pilch and n. p. warner , `` holographic , n=1 supersymmetric rg flows on m2 branes , '' jhep * 0909 * , 043 ( 2009 ) [ arxiv:0901.2736 [ hep - th ] ] . n. bobev , n. halmagyi , k. pilch and n. p. warner , `` supergravity instabilities of non - supersymmetric quantum critical points , '' class .
* 27 * , 235013 ( 2010 ) [ arxiv:1006.2546 [ hep - th ] ] .
g. lopes cardoso , g. dallagata and d. lust , `` curved bps domain wall solutions in five - dimensional gauged supergravity , '' jhep * 0107 * , 026 ( 2001 ) [ hep - th/0104156 ] .
g. lopes cardoso , g. dallagata and d. lust , `` curved bps domain walls and rg flow in five - dimensions , '' jhep * 0203 * , 044 ( 2002 ) [ hep - th/0201270 ] .
k. pilch and n. p. warner , `` n = 1 supersymmetric solutions of iib supergravity from killing spinors , '' hep - th/0403005 .
d. nemeschansky and n. p. warner , `` a family of m theory flows with four supersymmetries , ''
hep - th/0403006 .
i. bena and n. p. warner , `` bubbling supertubes and foaming black holes , '' phys .
d * 74 * , 066001 ( 2006 ) [ hep - th/0505166 ] .
p. berglund , e. g. gimon and t. s. levi , `` supergravity microstates for bps black holes and black rings , '' jhep * 0606 * , 007 ( 2006 ) [ hep - th/0505167 ] .
i. bena and n. p. warner , `` black holes , black rings and their microstates , '' lect .
notes phys .
* 755 * , 1 ( 2008 ) [ hep - th/0701216 ] .
d. z. freedman and s. s. pufu , `` the holography of @xmath461-maximization , '' jhep * 1403 * , 135 ( 2014 ) [ arxiv:1302.7310 [ hep - th ] ] .
k. lee , c. strickland - constable and d. waldram , `` spheres , generalised parallelisability and consistent truncations , '' arxiv:1401.3360 [ hep - th ] .
i. bena and n. p. warner , `` resolving the structure of black holes : philosophizing with a hammer , '' arxiv:1311.4538 [ hep - th ] .
b. de wit and h. nicolai , `` the parallelizing s(7 ) torsion in gauged @xmath583 supergravity , '' nucl .
b * 231 * , 506 ( 1984 ) . | we investigate a family of su(3)@xmath0u(1)@xmath0u(1)-invariant holographic flows and janus solutions obtained from gauged @xmath1 supergravity in four dimensions .
we give complete details of how to use the uplift formulae to obtain the corresponding solutions in m theory . while the flow solutions appear to be singular from the four - dimensional perspective , we find that the eleven - dimensional solutions are much better behaved and give rise to interesting new classes of compactification geometries that are smooth , up to orbifolds , in the infra - red limit .
our solutions involve new phases in which m2 branes polarize partially or even completely into m5 branes .
we derive the eleven - dimensional supersymmetries and show that the eleven - dimensional equations of motion and bps equations are indeed satisfied as a consequence of their four - dimensional counterparts . apart from elucidating a whole new class of eleven - dimensional janus and flow solutions , our work provides extensive and highly non - trivial tests of the recently - derived uplift formulae .
= 10000 * @xmath2 supersymmetric janus solutions and flows : * * from gauged supergravity to m theory * * krzysztof pilch , alexander tyukov and nicholas p. warner * department of physics and astronomy , university of southern california , los angeles , ca 90089 , usa [email protected] , [email protected] , [email protected] |
a medium composed of three - level atoms in @xmath0 configuration ( two ground states , @xmath1 and @xmath2 , dipole - coupled to an excited state @xmath3 , see fig .
[ fig : atomo ] ) can be rendered transparent using a technique known as electromagnetically induced transparency ( eit ) .
we will call the field mode coupling states @xmath2 and @xmath3 the probe field and the field mode coupling states @xmath1 and @xmath3 the control field . in the usual eit configuration
the probe field is weak compared with the control field , which , in turn , is in resonance with the transition frequency between states @xmath1 and @xmath3 .
we denote by @xmath4 the detuning of the probe field from resonance with respect to the transition frequency between states @xmath2 and @xmath3 .
the range of frequency detunings , @xmath4 , where absorption of the probe field by the medium is negligible is known as the transparency window ; its size is proportional to the control field .
the appearance of this transparency window using the control field is known as electromagnetically induced transparency ; see fig . [
fig : absorption ] for an example of the typical absorption spectrum of a probe field interacting with atoms showing eit . in ref .
@xcite the fundamentals of eit are explained in detail .
the probe field inside a medium composed of several three - level atoms exhibiting eit propagates without distortion with a velocity that is a function of the control field @xcite ; this can be used to slow down the probe velocity as much as desired @xcite . for an atom
, the atomic wave function has a dependence on its position ; when dealing with pulse propagation inside a medium composed of atoms , we can ignore this dependence when the pulse size is much larger than the variance of the atomic position ; in this case only the expectation value of the atomic position enters the description and we say that we treat the atoms as point - like . the theory developed in ref .
@xcite treats the atoms as point - like .
when the atoms that compose the medium are in a bose - einstein condensate , they can not longer be treated as point - like . in this case the condensate wave function together with its dynamics have a strong effect on the probe field propagation as is shown through numerical simulations in ref .
@xcite . to observe this effect
it is necessary that the time during which the pulse interacts with the medium is larger than the time scale characterizing the condensate wave function dynamics .
since a bose - einstein condensate has a well defined spatial phase relation , the system evolution does not necessarily decohere the information carried by the pulse , opening the possibility of using slow light and bec to manipulate optical information @xcite .
the main question we address in this paper is : how does the evolution of the condensate wave function modify the propagating pulse ? we address it in a form where analytical results can be obtained . to solve this question we have to solve the gross - pitaevskii equations , describing the condensate evolution , together with the maxwell equations , describing the field propagation .
we obtain a partial differential equation for the pulse propagation inside a bose - einstein condensate with atoms in @xmath0 configuration .
this equation is much simpler than the set of equations that are usually used to find the pulse propagation and the condensate wave function .
for some simple cases it can be solved analytically , showing explicitly how the different elements probe pulse , control field and wave functions interact and evolve .
the equation is obtained using a weak probe and approximations similar to those made in ref .
@xcite , but taking into account the condensate wave function . for a condensate wave function with a spatial profile that is initially constant ,
we write the solution in integral form .
the evolution for the field looks similar to the evolution of the wave function for atoms in state @xmath1 , but with its potential being described in a reference system that moves with the pulse ; we use this similarity for analyzing the pulse propagation in the medium .
a general overview of a three - level bose - einstein condensate interacting with two modes of the field is presented in section [ sec : model ] , where known approximations are used to express the model in a simplified way . in section [ sec : prop - equat - probe ] we obtain the propagation equation for the case of the probe pulse . in section [ sec : slow - light - inside ] we study how the pulse propagates inside the condensate showing eit and discuss some physical implications derived from the analytic solution .
concluding remarks are given in section [ sec : conclusion ] .
we only consider dynamics in the @xmath5 direction .
the hamiltonian of a three - level bose - einstein condensate interacting with two modes of light is given by @xmath6 with @xmath7 where @xmath8 are bosonic operators that create and destroy a particle , with mass @xmath9 , in the state @xmath10 , with @xmath11 , at position @xmath5 ; these obey commutation rules @xmath12 = \delta_{ij } \ , \delta ( x - x')\ , .\ ] ] the functions @xmath13 are the trapping potential for the atoms in state @xmath10 .
the constants @xmath14 are potentials due to elastic collisions between particles in the same level @xmath10 , @xmath15 are potentials due to elastic collisions between particles in level @xmath16 with particles in level @xmath10 .
we set the energy of the state @xmath17 as the zero of the energy scale .
the energy of atomic state @xmath18 is given by @xmath19 and the interaction hamiltonian reads @xmath20 where @xmath21 is proportional to the strength of the continuous control pulse , with wave number @xmath22 and frequency @xmath23 , @xmath24 is the incoming probe pulse , assumed to be centered at a frequency @xmath25 , and @xmath26 is the atom - field coupling . in order to calculate the system evolution we use the heisenberg representation , where the state vectors are time - independent and the evolution is given by time - dependent operators that obey the heisenberg equations of motion .
the heisenberg equations of motion for the bosonic operators are @xmath27\ , , \ ] ] substituting the hamiltonian @xmath28 given by eq .
( [ eq : hamiltonian ] ) we get the following differential equations for the bosonic operators @xmath29 here we have omitted the time and space dependence of the operators and variables and we assumed that @xmath30 .
we proceed now by writing the field operator as the sum of its expectation value plus fluctuations @xcite @xmath31 where @xmath32 is the identity operator and @xmath33 ( @xmath34 indicates the quantum expectation value ) is known as the mean field .
the gross - pitaevskii equations are partial differential equations where the fluctuations , @xmath35 , are not taken into account .
their solutions give an approximation to the evolution of the mean field , @xmath36 , whose squared norm gives the condensate spatial profile of the atoms in level @xmath16 .
the gross - pitaevskii equations are valid when the mean field dominates the fluctuations , this is the case when the interactions between atoms are weak and the system temperature is well below the transition temperature for bose condensation . using eq .
( [ eq : sum_cnumber_operator ] ) in eqs .
( [ eq : complete_system ] ) and neglecting fluctuations , we obtain a set of coupled gross - pitaevskii equations , they look similar to eqs .
( [ eq : complete_system ] ) , but with functions @xmath36 instead of operators @xmath37 .
we assume that the control field is in resonance with the dipole transition @xmath38 of atoms at rest , @xmath39 .
the dominant frequency of the probe field is @xmath40 , assumed to be in resonance with the dipole transition @xmath41 of atoms at rest , @xmath42 . in order to eliminate rapidly rotating terms
we make the following transformations : @xmath43 , @xmath44 , @xmath45 and @xmath46 .
all the atoms are initially in state @xmath2 , i.e. @xmath47 .
we use the weak probe approximation , @xmath48 , and assume that the wave functions @xmath49 can be expanded in powers of @xmath50 : @xmath51 the expansion limits the validity of our results to the case where the rabi frequency of the probe field is much weaker than the the rabi frequency of the control field , this is the standard approximation in eit based slow light physics . to zeroth order
we obtain @xmath52 and @xmath53 . at first order
we get @xmath54 where @xmath55 .
we inserted spontaneous decay from state @xmath3 , at rate @xmath56 , phenomenologically in eq .
( [ eq : psi0 ] ) .
since the probe pulse is weak , the population of level @xmath1 of the condensate is small , this is the reason that justify keeping only terms up to first order in @xmath57 in eq .
( [ eq : expansion_omega ] ) .
in the slowly varying approximation , the equation for the probe electromagnetic field becomes @xcite @xmath58 with @xmath59 being the slowly varying part of the electromagnetic field with @xmath60 .
to first order in @xmath57 , the dipole operator in the previous equation is @xmath61 eq . ( [ eq : expanionb ] ) implies @xmath62 we assume that @xmath63 reaches its stationary value very quickly compared with the time scale evolution of the other wave functions .
note that in the stationary regime the system reaches the electromagnetically induced transparency regime and the population of state @xmath3 is negligible .
these two considerations allow us to write eq .
( [ eq : psi0 ] ) as @xmath64 this expression is equivalent to eq . (
27 ) in ref @xcite .
we insert eq .
( [ eq : dip1 ] ) into eq .
( [ eqn : psi2psi0 ] ) and denote the condensate wave function of level @xmath2 , when there is no field , as @xmath65 obtaining @xmath66 substituting in the propagation equation we obtain @xmath67\right.\nonumber\\&&\left.+\frac{i \mathcal{\overline{e}}(x , t)\alpha(x , t)}{\hbar g(t ) } v_1+\frac{i \mathcal{\overline{e}}(x , t)u_{12}|\alpha(x , t)|^2\alpha(x , t)}{g(t ) } \right\}\ , , \nonumber\\\end{aligned}\ ] ] which is one of our main results .
the function @xmath68 is given by the solution of eq .
( [ eq : futuragp ] ) and does not depend on the field or the condensate wave functions of the other atomic levels . in the following section we take advantage of this to obtain an analytic expression for the solution in the particular case of an initially spatially uniform condensate .
this solution exemplifies how the different terms on the right side of eq .
( [ eq : progen ] ) contribute to the probe pulse evolution .
in this section , we study how the pulse probe propagates inside the condensate showing eit .
we will focus on the case where @xmath69 , with @xmath70 a constant , is a solution of the gross - pitaevsky equation for the state @xmath2 of the condensate .
spatially uniform condensates have been achieved experimentally @xcite .
the potential , @xmath71 , that generates this distribution , could be a square well ; it can also be a wide harmonic potential and the following discussion is valid in the region near the center of the potential , where there is little change in the density distribution @xcite .
this case avoids the effect that density change in atoms in level @xmath2 has on pulse propagation ; we are interested in the effect that state evolution of atoms in level @xmath1 has on the pulse .
this approximation allow us to obtain an analytic expression for the solution that completely characterizes both , the pulse propagation and the atomic state evolution .
defining @xmath72 , eq .
( [ eq : progen ] ) reads @xmath73 \mathcal{e}\left ( x , t\right)\ , .\end{aligned}\ ] ] to solve this equation we use a reference system that moves with the probe pulse . with substitution
@xmath74 we obtain @xmath75 where @xmath76 and @xmath77 the solution for the field shown in ( [ eq : solconst ] ) , includes some of the usual characteristic behavior of light traveling through media showing eit .
the pulse velocity depends on the value of @xmath21 ; as @xmath78 the pulse is stopped and , as can be seen from expression ( [ eq : dip1 ] ) , its information is transferred to the atoms @xcite .
note that if @xmath79 and the pulse is already stopped ( @xmath80 ) , the operator @xmath81 , defined in ( [ eq : fakehamiltonian ] ) , coincides with the hamiltonian operator of an atom trapped in the potential @xmath82 . as shown in relation ( [ eq : dip1 ] ) , the wave function of atoms in state @xmath1 is given by the probe field once the pulse is trapped ; then , while the pulse is trapped , the atomic wave function for atoms in state @xmath1 evolves in terms of its own hamiltonian .
as solution ( [ eq : solconst ] ) shows , the profile of it gets written in the pulse once it is released ( @xmath83 ) . if @xmath84 , the photons transfer momentum to the atoms that go from state @xmath2 to state @xmath1 .
this precisely appears in ( [ eq : fakehamiltonian ] ) as a displacement of the momentum operator by @xmath85 .
this effect has been used to move a pulse by physically moving the atoms in state @xmath1 , and then releasing the pulse once the atoms are in a new position @xcite .
it is clear from the solution that the pulse is modified as it travels through the medium , and in the process of stopping it ; the strength of this modification depends on @xmath21 and on the trapping potential for atoms in state @xmath1 .
the behavior described above is the expected behavior for trapping a pulse in a bec @xcite .
the advantage of the analytic expression for the field ( [ eq : solconst ] ) , is that it tells us how the pulse evolve : as if it were a quantum particle with a hamiltonian given by eq .
( [ eq : fakehamiltonian ] ) .
note that @xmath81 looks similar to the hamiltonian operator for atoms in state @xmath1 but multiplied by the factor @xmath86 and in a reference frame that moves with the pulse ; when the control field is zero , @xmath87 and @xmath81 coincides with the hamiltonian of atoms in state @xmath1 .
due to these facts we claim that : the way the probe pulse propagation is affected by the evolution of the wave function of atoms in state @xmath1 , is given by @xmath81 . with this interpretation ,
the coefficient factor @xmath88 in hamiltonian ( [ eq : fakehamiltonian ] ) , gives the strength of how the evolution of the atomic wave function affects the propagating pulse .
we will now explore some of the physical insights that solution ( [ eq : solconst ] ) exhibits .
the field , shown in ( [ eq : solconst ] ) , does not depend on @xmath89 : collisions between atoms in state @xmath1 does not affect the field propagation ; in the limit of weak probe field , the number of atoms transferred to state @xmath1 is small and collisions between them are negligible .
the field has a global phase which depends on @xmath90 ( see ( [ eq : solconst ] ) , ( [ eq : phidet ] ) ) , which in turn depends on term @xmath91 that characterizes the strength of collisions between atoms in state 2 with atoms in state 1 .
pulses that propagate in the medium with different control fields , @xmath21 , will have different phases ; the phase accumulation is maximum when the pulse is stopped , and it is negligible when @xmath21 is large . depending on which parameters are known , making these pulses interfere could be used to extract the value of @xmath91 , @xmath92 or @xmath26 .
if we make interfere a pulse that has been stopped and then released , with a pulse that propagates at constant velocity , we could extract information on the phase advance due to the process of trapping and releasing the pulse .
for intermediate cases ( @xmath93 ) the pulse is modified ; how it is modified depends on the wave function dynamics of atoms in state @xmath1 and on the coupling between this dynamics and the pulse , @xmath88 . for example
, if @xmath82 is constant inside the region the pulse propagates , the width of the wave function for state @xmath1 expands as a free particle , then the probe pulse width expands as if it were a free particle with mass @xmath94 . in this case ,
not taking into account that the field strength changes as a function of the control field , the evolution of the field , in the process of trapping it , would be similar to the evolution of a free particle wave function whose mass change with time .
the pulse propagation can be manipulated using @xmath82 .
for example , if @xmath95 , the center of the pulse ( in the moving frame ) will behave as if it were a harmonic oscillator with mass @xmath94 and frequency @xmath96 .
when a probe pulse of light propagates through a three - level bose - einstein condensate under conditions of electromagnetically induced transparency , its shape is modified by the evolution of the condensate wave function .
the strength of this modification is a function of the control field .
this is in contrast with the case where atoms are assumed point - like , where the probe pulse propagates without modification . we derived a partial differential equation
whose solution gives the probe pulse propagation inside a three - level bose - einstein condensate under conditions of electromagnetically induced transparency .
we obtained an analytic expression when the initial profile for the atoms in state @xmath2 is constant .
the solution shows explicitly how different parameters of the system trapping potential of the atoms , time dependent control field , scattering lengths , atom - field coupling affect the pulse propagation and the evolution of the condensate wave function .
the form of the solution for the probe propagation , expression ( [ eq : solconst ] ) , is similar to the formal solution of the schrodinger equation for the wave function of an atom in state @xmath1 , but with the hamiltonian given in a reference frame that moves with the pulse .
the solution allows us to see how to manipulate the pulse by choosing different trapping potentials for atoms in state @xmath1 and different control fields .
for example , if the potentials are zero , the width of the probe pulse expands in time as if it were a free quantum particle , but with an expansion rate which is a function of the control field .
the propagated pulse acquires a global phase that depends on the control field and the scattering length between atoms in state @xmath1 and @xmath2 .
this phase could be measured by performing interference experiments with pulses that propagate with different control fields .
we are confident that our results can help in the discussion of manipulating light memories using condensates .
we thank giovanna morigi and stefan rist for useful discussions . work supported in part by dgapa - unam grant papiit in103714 . | we obtain a partial differential equation for a pulse travelling inside a bose - einstein condensate under conditions of electromagnetically induced transparency .
the equation is valid for a weak probe pulse .
we solve the equation for the case of a three - level bec in @xmath0 configuration with one of its ground state spatial profiles initially constant .
the solution characterizes , in detail , the effect that the evolution of the condensate wave function has on pulse propagation , including the process of stopping and releasing it . |
although it is largely misunderstood and often misused , galaxy morphology is becoming a powerful tool for learning how galaxies form and evolve .
the _ hubble space telescope _ now allows us to study high - resolution ` deep fields ' of distant galaxies , yet finding a way to utilize this information is not obvious or simple .
the dominate paradigm for understanding galaxies is still largely based on the hubble sequence @xmath6 $ ] and its various revised forms .
hubble classifications are useful for describing the gross properties of nearby galaxies that are for the most part in quiescent star formation and dynamical states . even in this situation ,
hubble types correlate with physical properties , such as color and star formation rates , only in the _ mean _ [ 21 ] .
this is due to the fact that hubble types _ do not _ consider recent star formation in its classification , nor do interacting or merging properties factor into a hubble type . because the merger rate of galaxies in the local universe is @xmath7 3% [ 20 ] and
since recent star formation for nearby galaxies is also low [ 18 ] , the hubble sequence is effective in the mean for identifying nearby galaxy populations .
the subjective hubble morphological system is frankly inadequate at high redshift [ e.g. , 1,6 ] .
too many galaxies at high-@xmath1 fall in the catch - all `` irregular '' or `` peculiar '' class .
it is not yet understood if these unusual looking galaxies are true irregulars forming stars stochastically , or if they are starbursts triggered by some process , perhaps galaxy mergers . in this paper
, i will present preliminary work showing how high-@xmath1 galaxies seen in deep optical and nir images can be classified by understanding and quantifying star formation and merging properties based on their structures .
i also demonstrate how these tools can be used to perhaps trace the merger fraction of galaxies out to @xmath8 .
a truly physical morphology must correlate the way a galaxy ` looks ' , either through its structure or spectrum , with internal physical properties . the physical morphological system being developed by conselice et al .
[ 5,6,7,8 ] and bershady et al .
[ 2 ] thus far can distinguish the fundamental properties of star formation and galaxy interactions / mergers .
figure 1a shows the color - asymmetry diagram ( cad ) , a diagnostic tool used to calibrate these properties using nearby galaxies including the frei sample [ 12 ] supplemented by wiyn 3.5 m images of starburst galaxies .
the galaxies at @xmath9 are all consistent with merging systems based on their broad hi line profiles [ 8 ] .
the starbursts with @xmath10 , ngc 3310 and ngc 7678 , have well defined spiral patterns and whose starburst trigger was likely a bar instability , or in the case of ngc 3310 a minor merger in the distant past [ 8 ] .
the cad is therefore useful to quantify and trace the presence of not only interacting galaxies , but ones undergoing mergers . -0.75 in .
labeled galaxies are those consistent with an interaction or merger , with all galaxies with a@xmath11 having dynamical evidence for mergers [ 8 ] .
b. color - asymmetry diagram for galaxies in the hdf with a computed in the rest - frame b band.,title="fig:",scaledwidth=120.0% ] -4 in the cad is also useful for tracing star formation in galaxies , although high - frequency structure filtering techniques will likely turn out to be more useful .
the cad ` fiducial ' sequence , the dotted line in figure 1a is likely a star formation sequence .
as the star formation rate in a galaxy increases , it becomes bluer and more asymmetric due to the increase of localized star - forming regions .
it is important to note that even galaxies whose morphologies are dominated by star formation , such as the irregular ngc 4449 , do not have asymmetries that mimic the very high asymmetry produced by merging [ 7,8 ] .
although this calibration is limited by the small samples used , the fact remains that the only galaxies with @xmath9 are mergers .
this is consistent with other studies that demonstrate galaxies with the highest asymmetries are merging systems [ 23 ] .
do galaxies form by the merging of other galaxies ? how important is rapid collapse in the galaxy formation process [ 19 ] ?
the consensus is that merging is the dominate process , as shown in both semi - analytic and n - body modeling within cold dark matter dominated cosmologies [ 15,22,4 ] . finding a direct method of proving this beyond circumstantial evidence , such as evolving luminosity functions at high-@xmath1 [ 16 ] , remains a key observational challenge .
we can investigate this by the use of cads measured at high-@xmath1 .
several caveats must be stated before explaining the results of using quantitative features such as asymmetry at high-@xmath1 .
an advantage for astronomers studying galaxy evolution is the ability to study distant , and hence young galaxies .
thus we can directly compare properties of young and old galaxies to determine evolution .
this is easier said then done however , due to cosmological effects and biases . when measuring the same parameters at high and low-@xmath1 , zeroth - order features such as magnitudes and colors can be measured with some certainly by correcting for cosmological k - corrections and dimming .
higher order features , even simple measurements such as radii , are much harder to measure reliably . to robustly determine these features at high-@xmath1 _
requires _ simulating low-@xmath1 galaxies at high-@xmath1 to determine how parameters change when images are degraded in resolution , and have higher noise levels .
this is especially important for measuring morphological or structural features .
conselice et al .
[ 7 ] show that within certain limits , achievable with deep hst images , the asymmetry parameter measured on hdf galaxies with @xmath12 should reproduce the same values measured if the galaxies were nearby . for the low-@xmath1 hdf range @xmath13 ( figure 1b ) ,
there are some galaxies with asymmetries consistent with mergers .
there is also evidence for a possible correlation between asymmetry and color as seen in the nearby galaxy sample .
there are no symmetric red galaxies in this redshift range .
this is largely the result of a bias introduced by the small hdf field of view and the fact that the hdf was chosen to avoid any bright galaxies , that in the nearby universe are typically red . at the moderate redshift range @xmath14
, there are some red and symmetric galaxies ( figure 1b ) , and there is an increase in the presence of systems consistent with merging . at the highest redshift range ( @xmath15 )
, we see an even higher fraction of interactions , and no galaxies consistent with red ellipticals .
a fundamental aspect for understanding galaxy evolution is to constrain the history of galaxy merging in the universe .
galaxy pair studies indicate that a high fraction of distant galaxies are likely interacting [ e.g. , 20,17 ] .
dynamical pair fractions also reveal that the merger fraction of galaxies increases to at least @xmath16 with a rate @xmath17 ( 1+z)@xmath18 [ 17 ] .
dynamical pair studies are limited by several factors , the least of which is the inability to acquire spectroscopy for large numbers of faint galaxies .
perhaps a more important problem is the fact that these galaxies are not necessarily merging systems , but might be grazing interactions or even galaxies in orbits with several gyr merger time scales . despite this , dynamical pairs likely do measure some aspect of galaxy merging . using the asymmetry parameters on high-@xmath1 galaxies ,
it is possible to demonstrate how _ morphologically _ the merger fraction at each redshift evolves by directly measuring the fraction of galaxies with structures consistent with merging , as a function of redshift . doing this for galaxies in the hdf - n , we obtain figure 2 which shows the fraction of galaxies with rest - frame b - band asymmetries consistent with mergers as a function of @xmath1 , using the criteria : @xmath9 , m@xmath19 .
also plotted on this figure is the merger fractions found from dynamical pair studies by patton et al . [ 20 ] and le fvre et al . [
if we assume galaxies at these redshifts with @xmath9 are mergers , then the merger fraction increases as f = 0.03(1+z)@xmath4 [ 10 ] .
the merger fraction also potentially begins to decline at redshifts higher than @xmath5 [ 10 ] .
if the asymmetry parameter can indeed trace mergers out to these redshifts , this preliminary observation is scientifically exciting , and suggests that galaxy mergers are dominating the evolution of galaxies .
as is well known , the star formation rate of galaxies peaks at about this redshift [ 18,11 ] , as does the presence of active galaxies [ 3 ] .
are these effects and the mass assembly of galaxies driven by mergers ?
this morphological system suggests they are , demonstrating the power of galaxy classification when used and calibrated properly .
this work would not have been possible without the active contribution of many individuals , including the hdf - n nicmos go team , and especially mark dickinson , matt bershady , tamas budavari , jay gallagher , and anna jangren .
i also appreciate support from a nasa graduate student researchers program ( gsrp ) fellowship and from the stsci graduate student program .
abraham , r.g .
1996 , mnras , 279 , 47l bershady , m.a . ,
jangren , a. , & conselice , c.j .
2000 , aj , 119 , 264 5 boyle , b.j .
, & terlevich , r.j .
1998 , mnras , 293 , 49l cole , c. , lacey , c.g . ,
baugh , c.m , & frenk , c.s .
2000 , mnras , 319 , 168 conselice , c.j .
1997 , pasp , 109 , 1251 conselice , c.j . ,
& bershady , m.a . 1999 , in `` after the dark ages : when galaxies were young '' , ed .
holt & smith , p. 225
conselice , c.j . ,
bershady , m.a . , &
jangren , a. 2000a , apj , 529 , 886 conselice , c.j . , bershady , m.a . ,
& gallagher , j.s .
2000b , a&a , 354,886l conselice , c.j .
et al . 2000c , aj , 119 , 79 conselice , c.j . ,
bershady , m.a . , & dickinson , m. 2001 ( in prep ) ellis , r.s .
1997 , ara&a , 35 , 389 frei , z. , guhathakurta , p. , gunn , j. e. , tyson , j. a. 1996 , aj , 111 , 174 haynes , m.p . , & giovanelli , r. 1986 , apj , 306 , 466 hubble , e.p . 1926 , apj , 64 , 321 kauffmann , g. , colberg , j. , diaferio , a. , white , s. 1999 , mnras , 303 , 188 lacey , c. , guiderdoni , b. rocca - volmerange , b. , & silk , j. 1993 , apj , 402 , 15l le fvre , o. , et al .
2000 , mnras , 311 , 565l madau , p. , pozzetti , l. , & dickinson , m. 1998 , apj , 498 , 106 noguchi , m. 1999 , apj , 514 , 77 patton , d.r .
, et al . 1997 , apj , 475 , 29 roberts , m.s . & haynes , m.p .
1994 , ara&a , 32 , 115 somerville , r.s . ,
primack , j.r . , & faber , s.m .
2000 , astro - ph/0006364 wu , k. 1999 , ph.d .
thesis , university of california , santa cruz | galaxies have clearly evolved since the universe was 1 gyr old , but methods to trace and quantify this evolution are still in their infancy . in this paper
i demonstrate that with the careful use of a ` physical morphology ' it is possible to determine quantitatively how the process of galaxy evolution is occurring out to @xmath0 . using a system of parameters that traces star formation and galaxy interactions ,
i show how distinct galaxy populations at high-@xmath1 can be identified in deep high - resolution optical and nir images .
these tools are also used to measure a potential merger fraction of galaxies from @xmath2 .
if these methods are reliable , as is suggested by a local galaxy calibration , the merger fraction of galaxies scales @xmath3 ( 1+z)@xmath4 , peaks near @xmath5 , and declines thereafter .
i also discuss how this system is likely part of the ultimate physical classification of galaxies . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.